Engineered Technologies to Study the Role of Mechanical Signals in Human Lymphoma Growth and Therapeutic Response

ENGINEERED TECHNOLOGIES TO STUDY THE ROLE OF MECHANICAL SIGNALS IN HUMAN LYMPHOMA GROWTH AND THERAPEUTIC RESPONSE

A Dissertation Presented to the Faculty of the Graduate School of Cornell University In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

by Apoorva May 2018

ACKNOWLEDGEMENT

I am very thankful to my advisor Ankur for his mentorship, support and constant encouragement. His passion and enthusiasm for research are contagious and inspiring. I appreciate his engaging, hands-on and proactive approach to research. I also appreciate Ankur’s patience, and how he went out of his ways to help me on several occasions-- Anything good I do, will be because of him.

All this would not have been possible without Marcia Sawyer. Thank you to Marcia for all her unconditional support, encouragement and help. I am glad that when PhD gets tough, Marcia is around to talk to. Last but not least, I appreciate how she kept all the grad students informed and organized.

For the phosphorylation studies, Prof. Cosgrove and Alex made significant contribution to the experiments and analysis. I am grateful to Prof. Cosgrove for inspiring discussions, feedback and access to his lab facility. For the computational analysis and microfluidic experiments, I often visited Prof. Kirby. I thank him for his time and ideas. I am also grateful to my other committee members, Prof. Archer and Prof. Shepherd, for their feedback and guidance.

When I joined Singh lab, I had absolutely zero knowledge of biomedical engineering. Field and Alberto were the only students in the lab. They were both colleagues as well as friends. Field and Alberto contributed immensely in getting me started from zero, and I am truly thankful to them. During next three years of joining the lab I witnessed Singh lab growing into the most diverse, energetic and fun lab. It was a pleasure to work with or alongside of Matt, Regan, Jessica and Shiv. I am especially grateful to Kristine for assisting me in countless number of assays and cultures. I also thank her for editing and feedback on first and fourth chapter of this dissertation.

My interaction with the entrepreneurship community at Cornell has been a rewarding and enriching experience. I thoroughly enjoyed all the hackathons and summits organized by Cornell entrepreneurship community. Thank you to Ken Rother, Steven

Gall, Brad Treat, Tom Schryver and Felix Litvinsky for teaching me on every step of this journey, and making this experience remarkable.

Jason and Pankaj were my companions for everything. I cherish my true friendship with J&P, I appreciate them sticking with me through my thick and thin. This part of my Cornell experience would always hold a special place in my memory. This journey would also be memorable because of Abhishek, Rina, Igor, Ritesh and Pooja. The many interesting talks I had with Akanksha on random topics at random places in random hours were integral to this journey. Akanksha l, you were my companions through “ups and downs” and I cherish our little “roller coaster ride”.

Ma, Papa, Mama and Tullu are always there for me with their unquestioned and unconditional support. As I always wonder- How could I thank them. I don’t know. Maybe I can’t. And so, I won’t.

ENGINEERED TECHNOLOGIES TO STUDY THE ROLE

OF MECHANICAL SIGNALS IN HUMAN LYMPHOMA GROWTH AND

THERAPEUTIC RESPONSE

Apoorva, Ph.D.

Cornell University, 2018

Diffuse Large B cell lymphoma (DLBCL) is the most common lymphoma representing ~30% of all B cell Non-Hodgkin lymphomas. DLBCL is a heterogeneous disease associated with a variety of clinical presentations and genetic diversity (1). The standard combination chemotherapy, Rituximab (R)-

CHOP (doxorubicin, vincristine, prednisone, mechlorethamine) has been the frontline therapy for years, and still a significant percentage of DLBCL patients are not cured (2). Activated B cell-like (ABC) DLBCL is the most chemo-resistant

DLBCL subtype to R-CHOP with a 5-year overall survival as low as 40% vs 80% for germinal center B cell-like (GCB) DLBCL (1, 3). A myriad of independently predictive biomarkers for resistance have been identified, including gene expression signature, stromal signatures, epigenetic silencing of specific genes, and specific somatic mutations (3, 4). But none is sufficient to predict resistance in a given patient and few are helpful in guiding the selection of targeted therapies.

Therefore, new treatments and treatment-specific biomarkers are needed to improve clinical outcome of ABC-DLBCL.

It is becoming increasingly evident that tumor microenvironment is an active participant in the progression and pathogenesis of lymphoma (5, 6). DLBCLs

originate and reside in lymphoid tissues subjected characteristic lymphatic and vascular fluid forces, extracellular matrices, and cell-cell interactions with a variety of stromal and immune cells. In the context of ABC-DLBCL, which are the most chemoresistant subtypes, the hallmark ABC-DLBCL mutations result in constitutive activation of B cell receptor (BCR) pathways. Hence these pathways are emerging as a source of therapeutic targets for the treatment of these tumors.

However to date, existing BCR pathway inhibitors such as those targeting

Bruton’s tyrosine kinase (BTK) are active in a limited subset of patients and only for a short duration (few months). Therefore, there is a need to understand factors that modulate BCR. My dissertation focuses on three components of lymphoma microenvironment, namely (a) fluid shear stress and mass transport, (b) tissue stiffness, and (c) vascularization. To study the role of lymphatics-mediatedshear stress in DLBCLs, we engineered an integrated cell culture micro-reactor platform with micron-scale high resistance channels that recapitulates fluid flow velocities, pressure, and shear stresses developed in subcapsular sinuses of lymph nodes. The findings suggest that lymphatic-grade shear stress increases DLBCL cell proliferation and reduces chemotherapeutic responsiveness across DLBCL subtypes. The interplay of α4β1-integrin, CD20, and B cell receptors guides

DLBCLs to respond differentially to fluid shear stress and nutrient mass transport, by altering the phosphorylation of downstream signaling pathways. The dissertation next focuses on the role of mechanical stiffness of the malignant lymphoid tissues. The DLBCLs are commonly characterized by enlargement and palpable stiffness of lymph nodes. However, there are no studies on the

quantification of tissue stiffness in lymphomas and on the role, it can play in tumor progression or therapeutic response. We first determined the strain energy density of freshly isolated healthy and tumorous mouse lymph nodes, as well as lymphoma tissue from a patient using a micropipette aspiration technique, previously applied to quantify the local stiffness measurement of soft tissues (7,

8). We next developed stiffness-matched hydrogels to elucidate the role of lymphoid tissue stiffness on proliferation, phenotype, and therapeutic response of

DLBCLs. The final part of the dissertation focuses on a cellular component of

DLBCL microenvironment – the endothelial cells, which play an important role in angiogenesis. It is now increasingly recognized that DLBCLs are promoted by two angiogenic modes, firstly by autocrine signals via self-expressed VEGFR and

VEGF, and secondly by paracrine signals of endothelial progenitors of the microenvironment (9). In this dissertation we present an engineered bio-artificial organoid platform to study the interaction between DLBCL and endothelial cells, the angiogenic component of lymphoma microenvironment. We demonstrate that two integrins α4β1 and αvβ3, expressed by both DLBCLs and endothelial cells, modulate the VEGF, as well as CD20 and B cell receptor expressed by DLCBL cells. We anticipate that our engineered technology will be useful in study of lymphoma biology, and discovery and optimization of new class of therapeutics.

Key word: Lymphoma, Organoids, Integrins, Micro-reactors, Angiogenesis

REFERENCES:

1. Roschewski M, Staudt LM, Wilson WH. Diffuse large B-cell lymphoma- treatment approaches in the molecular era. Nat Rev Clin Oncol. 2014;11(1):12-23. Epub 2013/11/13. doi: 10.1038/nrclinonc.2013.197. PubMed PMID: 24217204. 2. Friedberg JW. Relapsed/refractory diffuse large B-cell lymphoma. Hematology Am Soc Hematol Educ Program. 2011;2011:498-505. Epub 2011/12/14. doi: 10.1182/asheducation-2011.1.498. PubMed PMID: 22160081. 3. Lenz G, Wright G, Dave SS, Xiao W, Powell J, Zhao H, Xu W, Tan B, Goldschmidt N, Iqbal J, Vose J, Bast M, Fu K, Weisenburger DD, Greiner TC, Armitage JO, Kyle A, May L, Gascoyne RD, Connors JM, Troen G, Holte H, Kvaloy S, Dierickx D, Verhoef G, Delabie J, Smeland EB, Jares P, Martinez A, Lopez-Guillermo A, Montserrat E, Campo E, Braziel RM, Miller TP, Rimsza LM, Cook JR, Pohlman B, Sweetenham J, Tubbs RR, Fisher RI, Hartmann E, Rosenwald A, Ott G, Muller- Hermelink HK, Wrench D, Lister TA, Jaffe ES, Wilson WH, Chan WC, Staudt LM, Lymphoma/Leukemia Molecular Profiling P. Stromal gene signatures in large-B-cell lymphomas. N Engl J Med. 2008;359(22):2313-23. doi: 10.1056/NEJMoa0802885. PubMed PMID: 19038878. 4. Sehn LH, Gascoyne RD. Diffuse large B-cell lymphoma: optimizing outcome in the context of clinical and biologic heterogeneity. Blood. 2015;125(1):22-32. doi: 10.1182/blood-2014-05-577189. PubMed PMID: WOS:000350809500009. 5. Scott DW, Gascoyne RD. The tumour microenvironment in B cell lymphomas. Nat Rev Cancer. 2014;14(8):517-34. Epub 2014/07/11. doi: 10.1038/nrc3774. PubMed PMID: 25008267. 6. Burger JA, Wiestner A. Targeting B cell receptor signalling in cancer: preclinical and clinical advances. Nat Rev Cancer. 2018;18(3):148-67. Epub 2018/01/20. doi: 10.1038/nrc.2017.121 nrc.2017.121 [pii]. PubMed PMID: 29348577. 7. Aoki T, Ohashi T, Matsumoto T, Sato M. The pipette aspiration applied to the local stiffness measurement of soft tissues. Ann Biomed Eng. 1997;25(3):581-7. Epub 1997/05/01. PubMed PMID: 9146811. 8. Buskohl PR, Gould RA, Butcher JT. Quantification of embryonic atrioventricular valve biomechanics during morphogenesis. J Biomech. 2012;45(5):895-902. Epub 2011/12/16. doi: 10.1016/j.jbiomech.2011.11.032. PubMed PMID: 22169154; PMCID: 3535469. 9. Ruan J, Hajjar K, Rafii S, Leonard JP. Angiogenesis and antiangiogenic therapy in non-Hodgkin's lymphoma. Ann Oncol. 2009;20(3):413-24. Epub 2008/12/18. doi: 10.1093/annonc/mdn666. PubMed PMID: 19088170; PMCID: PMC2733074.

BIOGRAPHICAL SKETCH

Apoorva is a native of Hasanpur (Bihar), India. He was born in 1986 to Kiran and Chandrabhanu. He has a little brother, Ityalam. Apoorva did his schooling from DAV public school in his home town. After finishing his highschool and pre-college, Apoorva moved to Chennai in 2004, where he pursued B.Tech in Materials Engineering from IITM. Following that Apoorva lived in Bangalore from 2008 to 2011. He studied MS in mechanical engineering at IISc Bangalore. Between 2011-2017, Apoorva studied Ph.D. in Mechanical and Aerospace Engineering at Cornell University.

TABLE OF CONTENTS

Topic Page 1 1 Engineered Bio-synthetic Lymphoma Microenvironment 1 Abstract 2 1.1 Introduction 6 1.2 Lymphoma microenvironment 7 1.3 Molecular origin of DLBCLs 9 1.4 Therapeutic landscape for diffuse large B cell lymphomas: 11 1.5 Lymphoma microenvironment and its immune component in the context of emerging immunotherapy for DLBCL 13 1.5.1 Lymphatic-grade low fluid flow differentially mechanomodulates signaling in genetically diverse human B cell lymphomas

14 1.5.2 Lymph node stiffness mimicking hydrogels regulate human B cell lymphoma growth and cell surface receptor expression in a molecular subtype-specific manner 15 1.5.3 3D integrin ligand presentation modulates the vascularization in human B cell lymphoma in a molecular subtype-specific manner 15 Conclusion 16 References 23 2 How Biophysical Forces Regulate Human B Cell Lymphomas 23 Summary 24 2.1 Introduction

25 2.2 Results and Discussion 26 2.2.1 Ex vivo lymphoma micro-reactor with high resistance channels 27 2.2.2 Fluid flow enhances proliferation in DLBCL and regulates drug- induced apoptosis in lymphoma cells 29 2.2.3 Fluid flow upregulates the surface expression of B cell receptors (BCR) and integrin receptors in ABC-DLBCL, similar to xenografted human tumors in mice 32 2.2.4 Fluid flow differentially regulates phosphorylation of complementary signaling pathways in ABC-DLBCLs 33 2.2.5 Understanding the contributions of fluid shear stress and mass transport in ABC-DLBCL surface receptor expression 34 2.2.6 Mechanistic model of fluid flow-mediated mechanomodulation of ABC-DLBCLs. 35 2.2.7 The CD79B complex in BCR cross-talks with integrins under fluid flow conditions to mechanomodulate human DLBCLs.

2.3 Conclusion 36 37 2.4 Experimental Procedures 37 2.5 Cell culture in lymphoma micro-reactor 37 2.5.1 Human cell line xenografts in NSG mice 38 2.5.2 Inducible shRNA mediated CD79B knockdown 38 2.5.3 Statistical analysis 38 Acknowledgements 39 Author Contributions 39 Declaration of Interests 40 References 57 Supplemental Information

80 3 Lymph Node Stiffness Mimicking Hydrogels regulate Human B Cell Lymphoma Growth and Cell Surface Receptor Expression in a Molecular Subtype-Specific Manner 80 Abstract 81 3.1 Introduction 3.2 Materials and methods 84

3.2.1 Tg(IghMyc)22Bri (Eµ-myc) mouse model of human lymphoma 84

3.2.2 Human B and T cell lymphoma lines 84

3.2.3 Cell encapsulation, organoid fabrication and rheology 84

3.2.4 Micropipette aspiration method to determine soft tissue stiffness 85

3.2.5 Cellular analyses of lymphomas in organoids 86

3.2.6 Drug-induced apoptosis studies in organoids 87

3.3 Statistical analysis 87

4 Results and Discussion 87

4.1 Change in Physiological and Mechanical Properties of Primary Lymphoma 87 Tissues 4.2 Biomaterials-based engineered lymphoma organoids with controlled stiffness 89

4.3 Biomaterials-based engineered lymphoma organoids with controlled stiffness 90 regulate the proliferation of human lymphomas and survival of primary murine lymphomas 4.4 Stiffness promotes drug resistance against standard chemotherapy as well as 92 novel chemotherapy-free epigenetic drugs Conclusion 95

95 Acknowledgements 96 References 4. Integrins modulate the synergy between vascularization and Human B Cell Lymphoma in a Molecular Subtype-Specific Manner 108 Abstract

xii

109 4.1 Introduction 111 4.2 Materials and methods 4.2.1 Human B cell lymphoma and endothelial lines 112

4.2.2 Cell encapsulation, organoid fabrication and vascularization 112

4.2.3 Cellular analyses of lymphomas in organoids 112

4.2.4 Integrin-inhibition studies in organoids 113

4.2.5 Statistical analysis 113

4.3 Results and Discussion 113

4.3.1 The presence of lymphoma cells upregulates the proliferation of endothelial cells 113 4.3.2 Synthetic hydrogel-based engineered lymphoma organoids 114 with tunable integrin ligand presentation 4.3.3 3D integrin engagement modulates release of growth factor 115 by lymphomas and development of vascular network 4.3.4 Endothelial cells engage v 3 integrins to adhere to the ECM, 116 and 4 1 to adhere to DLBCLs and to form clusters 4.3.5 Integrins v 3 and 4 1 modulate the migration of 117 endothelial cells and DLBCLs 4.4 Conclusion 118

Acknowledgements 118

References 118

5. Future Work 147

xiii

LIST OF FIGURES

Figure 1.1 Lymphoma microenvironment is the host niche composed of extracellular matrix, 3 homo-and-hetero-typic cells, physical factors such as topography, and biomolecular

factors such as cytokines which create a cooperative environment for the growth and progression of lymphomas. Figure 2.2 GCB and ABC types of diffuse large B cell lymphomas arise at two different stages 6 immune reaction and have different sets of mutations except BCL-2.

Figure 3.3 Current therapeutic landscape of small molecule drugs for diffuse large B cell 10 lymphomas.

Figure 2.1 Effect of fluid shear stress on proliferation and cytokine secretion in human 51 DLBCLs.

Figure 2.2 Effect of fluid shear stress on drug response in human DLBCLs. 53

Figure 2.3 Effect of fluid flow on integrin and B cell receptor expression in human ABC- and 55 GCB-DLBCLs.

Figure 2.4 Effect of fluid flow on phosphorylation of intracellular signaling proteins. 57

Figure 2.5 Understanding the contributions of fluid shear stress and mass transport in 58 mechanomodulation of DLBCLs.

Figure 2.6 Proposed mechanistic model and effect of inhibiting signaling pathways. 59

xiv

Figure 2.7 The CD79B complex cross-talk with fluid flow in mechanomodulation of 60 DLBCLs.

Figure 2.S.1 Computational modeling of lymphoma micro-reactors, Related to Figure 1. 68

Figure 2.S.2 Computational modeling of multi-shear micro-reactor, Related to Figure 1.A) 69

Figure 2.S.3 Design and operating mechanism of cross-flow micro-reactors, Related to Figure 1. 71

Figure 2.S.4 Effect of fluid flow on protein secretion in CD79A and CD79B BCR mutated ABC 74 DLBCL, determined using ELISA, Related to Figure 1.

Figure 2.S.5 Effect of fluid flow on BCR and Integrin expression in GCB-DLBCL and ABC- 75 DLBCL in 2D and 3D, Related to Figure 3.

Figure 2.S.6 Effect of fluid shear stress on phosphorylation of intracellular signaling protein 76 pAKT in ABC-DLBCL and GCB-DLBCL, Related to Figure 4.

Figure 2.S.7 Inhibition of signaling pathways and CD79B to mechanistically understand the effect 78 of fluid flow in DLBCL mechanomodulation, Related to Figure 5, 6, and 7. A) Media

viscosity measurement, Related to Figure 5. Figure 3.1 Change in lymph node and spleen size and weight with lymphoma tumorigenesis. 99

Figure 3.2 Mechanical characterization of healthy versus tumorous lymph nodes and spleen, 100

and hydrogel organoids.

Figure 3.3 Lymphoid stiffness regulates proliferation of mature B cell lymphomas. 101

Figure 3.4 Lymphoid tissue stiffness modulates drug resistance in mature B cell lymphomas 102 treated with the conventional chemotherapy drug vincristine or histone deacetylase

inhibitor Panobinostat. Figure 3.5 Effect of lymphoid tissue stiffness on B cell lymphoma markers and B cell receptor. 103

Figure 3.6 lymphoid tissue stiffness differentially regulates the integrin β1 expression in 104 genetically diverse B cell lymphomas.

Figure 3.7 Effect of lymphoid tissue stiffness on integrin β1 and IgM BCR expression in 105 primary mouse B cell lymphomas.

Figure 3.8 Lymphoid tissue stiffness modulates actomyosin contractility in human B cell 105 lymphomas.

Figure 4.1 Effect of DLBCLs on Proliferation of HUVECs. 117

Figure 4.2 Synthetic hydrogel-based engineered lymphoma organoids with tunable integrin 119 ligand presentation as a model for angiogenesis.

Figure 4.3 3D integrin engagement modulates release of growth factor by lymphomas and 120 development of vascular network.

Figure 4.4 Endothelial cells engage v 3 integrins to adhere to the ECM, and 4 1 to adhere 122 to DLBCLs and to form clusters.

xvi

Figure 4.5 Integrins v3 and 41 modulate the migration of endothelial cells and DLBCLs. 123

xvii

LIST OF ABBREVIATIONS

1 DLBCL Diffuse Large B Cell Lymphoma 2 ABC Activated B Cell Lymphoma 3 GCB Germinal Cell Lymphoma 4 RGD Arg-Gly-Asp Peptide 5 REDV Arg–Glu–Asp–Val Peptide 6 DOX Doxorubicin 7 HDACi Histone deacetylase inhibitors 8 IMDM Iscove's Modified Dulbecco's media 9 RPMI Roswell Park Memorial Institute medium 10 VEGF Vascular Endothelial Growth Factor 11 VEGFR Vascular Endothelial Growth Factor Receptor

xviii

LIST OF SYMBOLS

1 V CD61

2 3 CD51

3 4 CD29

4 1 CD49D

xix

PREFACE Diffuse Large B cell lymphoma (DLBCL) is the most common type of lymphoma that affects 7 out of 100,000 people every year. The standard treatment consists of a combination chemotherapy called R-CHOP (Rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone) which has been the primary regimen for lymphoma for more than 20 years. Currently, a significant percentage of DLBCL patients undergo relapse (~40%) and die with lymphoma. Despite improved understanding of genetics and pathophysiology of DLBCLs, the factors responsible for relapse and subsequent drug-resistance are poorly understood. It is becoming increasingly evident that tumor microenvironment is an active participant in the progression and pathogenesis of lymphoma, and often this transduction is via mechanical signals. DLBCLs originate and reside in lymphoid tissues subjected characteristic lymphatic-grade fluid shear stresses, extracellular matrices, and heterogeneous cell-cell interactions. This dissertation researches these three components of lymphoma microenvironment, namely (a) fluid shear stress, (b) tissue stiffness, and (c) vascularization. To study the role of lymphatic grade shear stress in DLBCLs, we present an integrated cell culture micro-reactor platform with micron- scale high resistance channels that recapitulates fluid flow velocities, pressure, and shear stresses developed in lymph nodes. The findings suggest that lymphatic-grade shear stress increases DLBCL cell proliferation and reduces chemotherapeutic responsiveness, in a DLBCL subtype and mutational dependent manner. The interplay of α4β1-integrin, CD20, and B cell receptors guides DLBCLs to respond differentially to fluid shear stress and nutrient mass transport, by altering the phosphorylation of downstream signaling pathways. The DLBCLs are also characterized by enlargement and palpable stiffness of neoplastic lymphoid tissue. However there are no reported research on the role of tissue stiffness in lymphomas. This dissertation presents a

tunable ex vivo engineered organoid to elucidate the role of lymphoid tissue stiffness on DLBCLs. By varying the composition of silica nanoparticle in this hydrogel we recapitulate the stiffness of healthy and neoplastic lymph nodes. We demonstrate that DLBCLs grow and respond to therapeutics differentially in soft and high stiffness microenvironments depending the subtype of DLBCLs. DLBCL microenvironment consists of vascular network that supplies nutrients and oxygen to tumor. It is now increasingly recognized that DLBCLs are promoted by two angiogenic modes, firstly by autocrine signals via self-expressed VEGFR and VEGF, and secondly by paracrine signals of endothelial progenitors of the microenvironment. In this dissertation we present an engineered bio-artificial organoid based vascularization platform to study the role of angiogenic component of DLBCL microenvironment. We demonstrate that two integrins α4β1 and αvβ3 are expressed by both DLBCLs and endothelial cells, and the activity these integrins modulates the VEGF, as well as CD20 and B cell receptor expressed by DLCBL cells. We anticipate that our engineered technology will be useful in study of lymphoma biology, and discovery and optimization of new class of therapeutics.

xxi Chapter 1 Engineered Bio-synthetic Diffuse Large B Cell Lymphoma Microenvironment

Fnu Apoorva1 and Ankur Singh1*

1Sibley School of Mechanical and Aerospace Engineering, College of Engineering,

Cornell University, Ithaca, NY 14853, USA

ABSTRACT

Diffuse large B cell lymphomas (DLBCLs) are the most common type of lymphomas accounting for ~30% of the diagnosed cases of lymphoma. The current primary treatment for DLBCLs is a combination chemotherapy regimen that shows differential therapeutic response depending on the molecular subtype of DLBCL. Unfortunately, this standard therapy has not changed over past two decades. As a result, we witnessed a dismal improvement in the survival rates of lymphoma patients, and ~40% of patients still relapse and remain incurable. This high relapse rate underscores the need to investigate the origin of drug resistance and discover more specific treatments. DLBCLs are characterized by highly orchestrated interaction between malignant B cells and their constitutive microenvironment. Numerous studies on DLBCLs have implicated tumor associated stromal cells in the development of drug resistance. The small molecule therapeutics, such as lenalidomide and bevacizumab, which target tumor microenvironment have shown promising results during their clinical trials. It is becoming increasingly evident from these investigations that tumor microenvironment plays a pivotal role in the progression and prognosis of DLBCLs. A physiologically relevant in-vitro model of tumor microenvironment gives a well-defined platform for cancer research in contrast to the complex and highly variable in vivo model. However, currently there are no physiologically relevant microenvironment models of DLBCLs.

1. INTRODUCTION

An estimate by NCI suggests that 2.1 percent of men and women in US will be diagnosed with non-Hodgkin lymphoma at some point during their lifetime (9).

1 Considered 5th most common cancer in US, lymphoma is a heterogeneous group of immune cells related malignancies which originate in lymphoid tissues. As per WHO classification system there are more than 60 different types of lymphomas (10, 11). Diffuse Large B cell lymphoma (DLBCL) is the most common and aggressive type of lymphoma which accounts for ~30% of the diagnosed lymphoma cases. Based on morphology, genetics and clinical outcome DLBCLs are classified into two major types, namely Activated B Cell type (ABC-type), and Germinal Center B Cell type (GCB- type) (12-15). ~20% of DLBCLs still remain unclassified (16).

For past two decades RCHOP, a combination chemotherapy consisting of Rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone, has been the cornerstone therapy for all types of DLBCL. Even to this day ~40% of DLBCL patients are either refractory to RCHOP or relapse post-treatment (17). The ABC and GCB types of DLBCL show differential response to RCHOP therapy. GCB-DLBCL patients respond positively to RCHOP in 75% cases, whereas ABC-DLBCL has a dismal success rate as low as 35% (4). The mechanism through which DLBCLs are resistant to current therapies are unknown but may be linked to the particular spectrum of somatic mutations in these tumors, which are in concert with complex growth signals provided by the lymphoid tumor microenvironment.

It is becoming increasingly evident that tumor microenvironment plays an important role in the progression and prognosis of DLBCLs (18). The highly orchestrated bidirectional communication between tumor cells and its complex milieu of microenvironment determines the synergetic evolution of host microenvironment and tumor progression, and subsequently the therapeutic outcome (19). The biological origin of refractory response of some of the DLBCLs could be explained by understanding these reciprocal interactions between the DLBCL cells and their host microenvironment.

The paradigm of all lymphoma studies are two dimensional cultures, which lack the physiologically relevant characteristics of lymphoma microenvironment. While the

2 mammalian tissues are characterized by three-dimensional hierarchical arrangement of cells within the extracellular matrix, two dimensional cultures lack this microenvironment. As a result, the biochemical and biomechanical cues in two dimensional cultures are not representative of a mammalian tissue. This explains why clinical researchers often miss the biological subtleties when using a two-dimensional culture for their study (20-22). 3D culture systems such as cellular spheroid and hydrogel-scaffold based tumor cultures more accurately reflected the clinical gene expression than their two dimensional counterparts (23, 24). The Xenograft models of DLBCLs have been found to be even more realistic representative of the disease, often leading to breakthrough earlier impossible to achieve using the convention two dimensional cultures (25-27). Chapuy et.al generated and characterized patient derived xenografts (PDXs) of DLBCLs (28). The report by Chapuy et.al shows that PDX models reflect immunophenotypic, transcriptional, genetic and functional heterogeneity of lymphoma. This study also demonstrated the significance of PDXs in the functional analyses of proximal BCR pathway inhibition. Unfortunately, PDXs and xenografts in general are expensive, time consuming, and often require specialized expertise which make unsuitable for first-line drug screening and mechanistic study. This underscores the need of a simple but physiologically relevant bio-synthetic engineered model of lymphoma, to study the role of microenvironment, its pathology and signaling pathways. Such engineered models of lymphoma would serve as a more a suitable preclinical screening platform facilitating efficient translation of lymphoma research into novel therapeutic regimens.

3 Here we describe the critical components of lymphoma microenvironment, and its significance with respect to the current therapeutic landscape, and summarize the relevant engineered lymphoid platforms developed in recent past. While there are many components of DLBCL microenvironment, as shown in Figure-1, in this dissertation I focus on three important components of DLBCL microenvironment, namely (a) biophysical interaction within the lymphoma microenvironment, (b) cell-matrix interactions, and (c) cell-cell interactions. We present three engineered biosynthetic microenvironment models of lymphomas and elucidate the interplay between DLBCLs

Figure 1. Lymphoma microenvironment is the host niche composed of extracellular matrix, homo- and-hetero-typic cells, physical factors such as topography, and biomolecular factors such as cytokines which create a cooperative environment for the growth and progression of lymphomas. and its microenvironment. We hope these physiologically relevant biosynthetic

4 microenvironment models of DLBCLs will serve for better understanding of DLBCL oncology and discovery of novel therapeutics.

1.1 Lymphoma Microenvironment

The tumor microenvironment is the tumor-cooperative host niche resulting from the complex organization of extracellular matrix, tumor associated stromal cells, infiltrated immune cells, and biomolecules such as cytokines and chemokines that supports the survival and growth of tumor (29). During progression of cancer the tumor microenvironment undergoes a profound change which shifts the tissue homeostasis in favor of tumor. Over an extended period of time this altered microenvironment is known to transform the molecular signature of stromal and immune cells as well (30). This altered microenvironment promotes progression of cancer by transducing pro-survival factors into tumor.

Approximately two-third of the DLBCLs arise in lymphatic region (31). The inception of DLBCL is characterized by massive proliferation of malignant B cells within the lymphoid tissue causing an alteration in the surrounding extracellular matrix (ECM) which includes enhancement of stiffness and vascularity of the ECM. The stiffness of ECM determines the extracellular and intracellular force within the microenvironment (32, 33). The vascularity of ECM dictates the fluid force and biomolecular transport (34-40). As DLBCL tumor grows, it’s microenvironment undergoes complex transformation, changing the physical forces within the lymphoid ECM as well as intra and inter biomolecular signals. This evolution of microenvironment is simultaneously accompanied by infiltration and molecular modification of immune cells and stromal cells such as fibroblasts and endothelial cells (41). Over all DLBCL microenvironment is characterized as a dynamically evolving milieu of tumor, infiltered cells, ECM, metabolic factors, secreted biomolecules and physical forces, as shown in Error! Reference source not found..

Numerous oncological studies show physiologically closer models of microenvironment, such as use of hydrogel based three-dimensional culture of cancer,

5 successfully induce real-tumor-type histological morphology which is usually not observed in a conventional two-dimensional culture (21, 22, 42-44). These three- dimensional microenvironment models also showed a more realistic drug response than their two-dimensional counterparts, demonstrating that three-dimensional tissue cultures abridge the gap between conventional two-dimensional cell culture and the in- vivo models. In the context of lymphomas, Singh and his colleagues demonstrated that three dimensional organoids functionalized with specific integrins mimicked a lymphoid neoplasm-like heterogeneous microenvironment which promoted proliferation, clustering and drug resistance in lymphoma cells (45). This underscores the relevance of development of a suitable biosynthetic microenvironment model that recapitulates the physio-spatial organization of host niche of lymphomas is the key to these challenges.

Never the less, there are several challenges involved in development of microenvironment targeting therapeutics. Despite improved genetic and molecular knowledge of lymphoma, our understanding of lymphoma microenvironment is still poor. A better understanding of lymphoma microenvironment would also facilitate discovery of novel therapeutics. It is becoming increasingly evident that targeting lymphoma microenvironment could be an alternative to chemo-treatment. While the standard therapy has as high as ~40% relapse rate among DLBCL patients, recent investigations of microenvironment targeting therapeutics showed promising results.

The reciprocal interaction between DLBCL and its microenvironment often determines the course of canonical and non-canonical signaling pathways such B-Cell Receptor (BCR). For example, chronic BCR can be simulated by the presence of latent microbial or auto-antigen in the lymphoid follicles that leads to development and expansion of malignant B cells, as observed in the context of Helicobacter Pylori infection. To further corroborate the role of microenvironment in BCR, Apoorva et.al and Field et.al reported that integrin signaling could modulate the BCR as well (33, 45). Thus the complex cross talk between the microenvironment and neoplastic cells plays an important role not only in pathogenesis, but also in disease progression and therapeutic resistance. These findings have opened up new avenues for microenvironment checkpoint-based regimen

6 for Lymphoma.

Another hallmark of cancer is altered proteolysis that remodels the extracellular matrix to allow unregulated tumor growth and invasion, and eventually metastasis. Lymphoma patients often demonstrate significantly higher levels of IL-6, MMP-2 and MMP-9 in peripheral blood. IL6, MMP-2 and MMP-9 modulate proteolysis with in lymphoma microenvironment by facilitating the degradation of extracellular matrix. The primary aim of a micro-engineered lymphoma microenvironment is to recapitulates the complex extracellular architecture of lymphoma and provides right biophysical and biomolecular cues to the tumor cells. The 2D mono-culture is the paradigm for in-vitro study of pathophysiology of cancer and therapeutic discovery. This conventional method involves culturing of a given type of cells on a 2D substrate such as TCP. Recent studies have shown that 2D monocultures often lead to loss of key cellular processes.

1.2 Molecular origin of Heterogeneity in Tumor Cells and Its Host Microenvironment in DLBCL

7 The current lymphoma therapeutic development recognizes three different molecular subtypes of diffuse large B cell lymphomas: (a) ABC lymphomas (b) GCB lymphomas (c) neither ABC/GCB, depending on the mutation, morphology and the stage of immune reaction that presents tumor (Figue-2). The GCB-DLBCL is presented when aberrant

Figure 2. GCB and ABC types of diffuse large B cell lymphomas arise at two different stages immune reaction and have different sets of mutations except BCL-2. mutation occurs at the stage of germinal center reaction during differentiation of centroblasts to centrocytes. GCB-DLBCLs are characterized by BCL-2/BCL-6, EZH2 and MYC mutations. The ABC-DLBCLs are presented when mutation happens during transformation of centrocytes to plasma B cells. These tumors are characterized by mutation in BCL-2, CARD11, MYD88 and CD79A/B. DLBCLs which have MYC rearrangement as well as one of the two BCL2 or BCL6 rearrangements are called “double-hit” lymphomas. Similarly, if all three of MYC, BCL-2 and BCL-6 arrangements are presented, it is called a “triple-hit” lymphomas. The double-hit and triple-hit DLBCLs are more aggressive and are associated with poor prognosis.

8 1.2.1 Genetic and Clinical Heterogeneity in DLBLs and It’s Host Microenvironment

The clinical and genetic heterogeneity of tumor cells and their host microenvironment is the hallmark of lymphoma. The molecular heterogeneity nature of DLCBL combined with the manifested through distinct morphology, gene expression and epigenetics within of the microenvironment. This heterogeneity of tumor cells and microenvironment poses significant challenge in the understanding of the biology of lymphoma, its prognosis and therapeutics (46). The recent discoveries substantiate that clinical outcome of RCHOP treatment in DLBCLs is corelated to the molecular profile of DLBCLs and their host microenvironment (46-48). Reddy et.al reported exome sequencing and transciptome sequencing of 1001 DLBCL patients identifying 150 driver oncogenes of DLBCL (47). The unbiased CRISPR screening of DLBCL cell lines characterized the functional impact of driver oncogenes and elucidate the connects between genomics, CRISPR hits and clinical observation. Previously various research focusing on molecular profile of DLBCL associated stromal cells proved a strong correlation between DLBCL microenvironment and clinical outcome (41, 49). These studies corroborate the need to understand the microenvironment of DLBCLs to be able to develop molecularly targeted therapies.

1.2.2 Intra-tumoral Heterogeneity in DLBLs

The single cell RNA-seq and intratumoral histology suggest that refractory forms of Lymphomas are characterized by high degree of intratumoral heterogeneity (50). It is becoming increasingly evident that high level of genetic variation within the tumor microenvironment acts as a precursor for evolutionary selection of clinically resistant form of tumor. The intratumor heterogeneity also determines the course dysregulated epigenetics. The complex epigenetic dysregulation, for example, alter the phenotype of tumor associated stromal cells such as fibroblasts, immune cells and endothelial cells. (51)

1.3 Therapeutic Landscape for Diffuse Large B cell Lymphomas:

9 The therapeutic targets for ABC-DLBCLs are mostly BCR pathways or pathways downstream of BCR. These therapies include protease inhibitors such as Bortezomib (reference) which are aimed at suppressing NFκB pathway. Interestingly, while monotherapy with Bortezomib failed to treat the relapsed/refractory DLBCLs, the combination therapy consisting of Bortezomib and EPOCH (etoposide, vincristine, doxorubicin, cyclophosphamide, and prednisone) showed a remarkable success with significantly higher response in as high as 83% of refractory ABC patients. The same combination therapy for refractory GCB-DLBCL patients showed only 13% success. There are several immunomodulatory agents currently under investigation for treatment of lymphomas. These immunomodulatory drugs are structural and functional analogs of thalidomide. One such example is lenalidomide which targets DLBCL microenvironment, and has shown promising results. The preclinical studies demonstrated that lenalidomide functions in DLBCLs via several mechanisms; (a) inhibition of production of pro-tumor-growth cytokines such as TNFα, (b) arrest of angiogenesis by inhibiting VEGF release, (c) inducing G0/G1 cell cycle arrest, and (d) decreasing NFκB, activity. During phase II trial lenalidomide positively responded 39% of refractory/relapsed patients. Chronically activated BCR is a common observation (~23%) among ABC-DLBCLs. This motivated development of numerous BCR targeting therapies such as SYK inhibitors (fostamatinib) and BTK inhibitors (ibrutinib). For GCB-DLBCLs most of the novel therapeutics are aimed at containing BCL-2. The BCL-2 inhibiting compounds currently under investigation are ABT-737, ABT-262, and ABT-199.

2. Engineered Microenvironment for Diffuse Large B Cell Lymphoma

DLBCLs originate and reside in lymphoid tissues subjected to characteristic lymphatic-grade fluid shear stresses, extracellular matrices, and heterogeneous cell- cell interactions. This dissertation researches these three components of lymphoma microenvironment, namely (a) fluid shear stress, (b) lymphoid tissue stiffness, and (c) endothelial induced vascularization. In order to delineate the effect of these three components of lymphoma microenvironment, we present three micro-engineered platforms. Firstly, in order to study the role of lymphatic grade shear stress in

10 DLBCLs, a microfluidic based cell culture micro-reactor platform has been presented. Secondly, we present an ex vivo engineered organoid that recapitulates the mechanical properties of a malignant lymphoid tissue. And lastly, we present a hydrogel based platform with the capability of tunable three dimensional presentation of integrin peptides that recapitulates the vascularization aspect of DLBCLs.

The micro-reactor platform consists of micron-scale high resistance channels to recapitulate fluid flow velocities, pressure and shear stress developed in lymph nodes. The findings suggest that lymphatic-grade shear stress increases DLBCL cell proliferation and reduces chemotherapeutic responsiveness, in a DLBCL subtype and mutational dependent manner. The interplay of α4β1-integrin, CD20, and B cell receptors guides DLBCLs to respond differentially to fluid shear stress and nutrient mass transport a DLBCL-subtype dependent manner.

The second engineered system we present is a micro-engineered gelatin and silica nanoparticle based organoids, and we elucidate the role of lymphoid tissue stiffness in DLBCLs. By varying the composition of silica nanoparticle in this hydrogel we recapitulate the stiffness of healthy and neoplastic lymph nodes. We demonstrate that DLBCLs proliferate and respond to therapeutics differentially in soft and high stiffness microenvironments depending the subtype of DLBCLs.

The third engineered system we present is an engineered bio-artificial organoid that recapitulates vascularization in DLBCLs. These organoids are made from 4-arm PEGMAL which can be tethered with variable amount of RGD and REDV integrin peptides in order to control the 3D presentation of integrins to the DLBCLs. The results that two integrins α4β1 and αvβ3 play an important role in angiogenesis in DLBCLs. The 3D presentation of integrins α4β1 and αvβ3 modulate angiogenesis by regulating the release of VEGF by DLBCLs.

2.0 Lymphoma microenvironment and its immune component in the context of emerging immunotherapy for DLBCL

The unique tumor environment of the lymphatic system, especially its immune components need special consideration when pursuing immune-based therapeutic

11 approaches for lymphoma. Beguelin et.al recently reported a study of germinal center reaction of B cells in an ex-vivo 3D organoids and an in-vivo mice model. This study demonstrated that synthetic 3D B cell follicular organoids consisting of RGD motifs and 40LB stromal cells could mimic and modulate the phenotypic and transcriptional regulatory markers of germinal center reactions and germinal center B cells. This 3D synthetic organoid showed higher proliferation of germinal center B cells, and a physiologically relevant gene expression of hallmark genes of GC B cells, such as EZH2, BCL6 and Aicda. The analysis of sequencing data also indicated that indels and missense mutations in synthetic 3D organoid GC B cells were comparable to in- vivo GC B cells from mice, where B cells in convectional 2D culture underwent class switch recombination that suppressed indels and missense mutations. EZH2, BCL6 and Aicda genes play an important role in DLBCL pathobiology. EZH2 is implicated in tumorigenesis and poor prognosis of DLBCL. McCabe et.al reported that EZH2 inhibition could be effective therapeutic strategy for DLBCLs, and using a highly selective EZH2 inhibitor, GSK126, demonstrated its therapeutic efficacy in a DLBCL xenograft mice. BCL6 is another important transcriptional factor which has been extensively studied in the context of Lymphoma. BCL6 plays an important role in both B-cell and T-cell associated antibody response. In the case of DLBCL, BCL6 often prevents terminal differentiation of B-cells by suppressing PRDM11. Similarly, AICDA has been implicated in lymphomagenesis and epigenetic heterogeneity and plasticity in DLBCL. Teater et.al reported that AICDA overexpression caused aggressive DLBCL growth and an elevated cytosine methylation heterogeneity in BCL2 driven murine lymphomas. These studies on relevance of synthetic 3D organoids to follicular B cells and germinal center reactions

12 Figure 3. Current therapeutic landscape of small molecule drugs for diffuse large B cell lymphomas.

2.1 Lymphatic-Grade Low Fluid Flow Differentially Mechanomodulates Signaling in Genetically Diverse Human B Cell Lymphomas

Over the past 3 years, our research group has engineered an artificial lymphoma- supportive tissue using synthetic building blocks and has created technologies that

13 provide plug-and-play modularity to address patient-specific needs. We have developed a modular and tunable 3D microscale lymphoid organoids made from bio-adhesive polymeric hydrogels that incorporate bio-ligands which function as a pro-survival signals and support cells found in lymph nodes. Other features of the organoids include cell-mediated remodeling of the 3D niche as found in native tumors. A unique aspect of the technology is that pro-survival signals are patient-specific, an aspect no existing lymphoma culture platform can provide. Next, to better mimic the conditions found in patient’s body, we have integrated fluid flow characteristics unique to lymph nodes. This aspect is achieved by integrating a high-resistance microfluidic platform with organoids. Our preliminary results without the integrated organoids indicate that simple flow of lymphatics-mimicking flow induces remarkable and significant differences in proliferation of lymphoma cells as compared to static conditions. Importantly, fluid flow results in reduced drug uptake than what scientists and companies have predicted and therefore has a differential effect of the death of lymphoma than classic tests. Therefore, our experimental therapeutics technologies appear to be on the right path to provide prognostic values and increase “predictive power” of pre-clinical studies for drugs in development, and importantly, will allow a faster and more rational screening and translation of therapeutic regimens.

2.2 Lymph Node Stiffness Mimicking Hydrogels Regulate Human B Cell Lymphoma Growth and Cell Surface Receptor Expression in a Molecular Subtype-Specific Manner

Diffuse large B cell lymphoma (DLBCL) is the most common type of non-Hodgkin's lymphoma, with multiple molecular subtypes. The activated B cell-like DLBCL subtype accounts for roughly one-third of all the cases and has an inferior prognosis. There is a need to develop better class of therapeutics that could target molecular pathways in resistant DLBCLs; however this requires DLBCLs to be studied in representative tumor microenvironments. The pathogenesis and progression of lymphoma has been mostly studied from the point of view of genetic alterations and intracellular pathway dysregulation. By comparison, the importance of lymphoma microenvironment in which these malignant cells arise and reside has not been studied in as much detail. We

14 have recently elucidated the role of integrin signaling in lymphomas and demonstrated that inhibition of integrin-ligand interactions abrogated the proliferation of malignant cells in vitro and in patient-derived xenograft. Here we demonstrate the role of lymph node tissue stiffness on DLBCL in a B cell molecular subtype specific manner. We engineered tunable bio-artificial hydrogels that mimicked the stiffness of healthy and neoplastic lymph nodes of a transgenic mouse model and primary human lymphoma tumors. Our results demonstrate that molecularly diverse DLBCLs grow differentially in soft and high stiffness microenvironments, which further modulates the integrin and B cell receptor expression level as well as response to therapeutics. We anticipate that our findings will be broadly useful to study lymphoma pathobiology and discover new class of therapeutics that target B cell tumors in physical environments. 2.3 3D Integrin Ligand Presentation Modulates the Vascularization in Human B Cell Lymphoma in a Molecular Subtype-Specific Manner

In chapter 3 I present an engineered hydrogel based vascularization platform where one can modulate the specificity and density integrin ligand presentation to tumor to study the effect of integrins on angiogenic factors in a neoplastic lymph nodes. I also present a microfluidic conduit system to study the migration of cells in response to angiogenic factors produced by DLBCLs and HUVECs. Recent studies have shown the significance of tumor associated endothelial cells in pathogenesis and treatment of lymphoma. It is increasingly becoming evident that tumor microenvironment actively participates in the initiation and progression of lymphoma, by promoting vascular network formation by these endothelial cells. Unfortunately, antiangiogenic lymphoma therapies such as bevacizumab have reaped poor outcome because of lack of mechanistic insight into lymphoma tumor microenvironment. I demonstrate that both DLBCLs and endothelial cells employ two sets of integrins 4 1 and v 3, and surface receptor VEGFR to establish a dysregulated VEGF based autocrine and paracrine signaling pathways that promotes vascular network formation and lymphoma progression in a synergistic manner. In summary, these data indicate that integrins may present a novel antiangiogenic target for lymphoma therapy. 3. CONCLUSION:

15 The development of a functional biosynthetic model of lymphoma for investigation of pathobiology and drug discovery necessitates incorporating physiologically relevant components of the lymphoma microenvironment. Such models of lymphomas abridge the gap between two-dimensional cultures and their in-vivo counterparts. This report presented three engineered bio-synthetic models of lymphoma cancer, which individually recapitulated three key physiological characteristics of lymphoma microenvironment, namely (a)lymphatic grade flow, (b)lymphoid stiffness, and (c) endothelial induced vascularization, though many other parameters need to be considered. The three different models of lymphomas repeatedly identified the role of integrins in mediating the interplay between DLBCLs and its microenvironment. These insights endorse the increasing significance of engineered in-vitro models of lymphoma microenvironment for studying the biology of lymphoma, probing of therapeutic targets and testing potential therapeutic interventions. These models of lymphomas models allow a faster and more rational screening and translation of therapeutic regimens.

References: 1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2018. CA Cancer J Clin. 2018;68(1):7-30. Epub 2018/01/10. doi: 10.3322/caac.21442. PubMed PMID: 29313949. 2. Leonard JP, Martin P, Roboz GJ. Practical Implications of the 2016 Revision of the World Health Organization Classification of Lymphoid and Myeloid Neoplasms and Acute Leukemia. J Clin Oncol. 2017;35(23):2708-15. Epub 2017/06/28. doi: 10.1200/JCO.2017.72.6745. PubMed PMID: 28654364. 3. Swerdlow SH, Campo E, Pileri SA, Harris NL, Stein H, Siebert R, Advani R, Ghielmini M, Salles GA, Zelenetz AD, Jaffe ES. The 2016 revision of the World Health Organization classification of lymphoid neoplasms. Blood. 2016;127(20):2375-90. Epub 2016/03/17. doi: 10.1182/blood-2016-01-643569. PubMed PMID: 26980727; PMCID: PMC4874220. 4. Shipp M, Tamayo P, Angelo M, Ray T, Reich M, Neuberg D, Last K, Aster J, Mesirov J, Lister A, Golub TR. Diffuse large B cell lymphoma outcome prediction by

16 gene expression profiling. Blood. 2000;96(11):222a-a. PubMed PMID: WOS:000165256100949. 5. Shipp MA, Ross KN, Tamayo P, Weng AP, Kutok JL, Aguiar RCT, Gaasenbeek M, Angelo M, Reich M, Pinkus GS, Ray TS, Koval MA, Last KW, Norton A, Lister TA, Mesirov J, Neuberg DS, Lander ES, Aster JC, Golub TR. Diffuse large B-cell lymphoma outcome prediction by gene-expression profiling and supervised machine learning. Nat Med. 2002;8(1):68-74. doi: DOI 10.1038/nm0102-68. PubMed PMID: WOS:000173056900033. 6. Scott DW, Wright GW, Williams M, Lih J, Walsh W, Jaffe ES, Rosenwald A, Campo E, Chan WC, Connors JM, Smeland EB, Mottok A, Braziel RM, Ott G, Delabie J, Tubbs RR, Cook JR, Weisenburger DD, Greiner TC, Glinsmann-Gibson BJ, Fu K, Staudt LM, Gascoyne RD, Rimsza LM. Determining Cell-Of-Origin Subtypes In Diffuse Large B-Cell Lymphoma Using Gene Expression Profiling On Formalin-Fixed Paraffin-Embedded Tissue - An LLMPP Project. Blood. 2013;122(21). PubMed PMID: WOS:000331385000102. 7. Scott DW, Wright GW, Williams PM, Lih CJ, Walsh W, Jaffe ES, Rosenwald A, Campo E, Chan WC, Connors JM, Smeland EB, Mottok A, Braziel RM, Ott G, Delabie J, Tubbs RR, Cook JR, Weisenburger DD, Greiner TC, Glinsmann-Gibson BJ, Fu K, Staudt LM, Gascoyne RD, Rimsza LM. Determining cell-of-origin subtypes of diffuse large B-cell lymphoma using gene expression in formalin-fixed paraffin- embedded tissue. Blood. 2014;123(8):1214-7. doi: 10.1182/blood-2013-11-536433. PubMed PMID: WOS:000335838100021. 8. Schmitz R, Wright GW, Huang DW, Johnson CA, Phelan JD, Wang JQ, Roulland S, Kasbekar M, Young RM, Shaffer AL, Hodson DJ, Xiao W, Yu X, Yang Y, Zhao H, Xu W, Liu X, Zhou B, Du W, Chan WC, Jaffe ES, Gascoyne RD, Connors JM, Campo E, Lopez-Guillermo A, Rosenwald A, Ott G, Delabie J, Rimsza LM, Wei KTK, Zelenetz AD, Leonard JP, Bartlett NL, Tran B, Shetty J, Zhao Y, Soppet DR, Pittaluga S, Wilson WH, Staudt LM. Genetics and Pathogenesis of Diffuse Large B-Cell Lymphoma. New Engl J Med. 2018;378(15):1396-407. doi: 10.1056/NEJMoa1801445. PubMed PMID: WOS:000429718500006.

17 9. Sehn LH, Gascoyne RD. Diffuse large B-cell lymphoma: optimizing outcome in the context of clinical and biologic heterogeneity. Blood. 2015;125(1):22-32. doi: 10.1182/blood-2014-05-577189. PubMed PMID: WOS:000350809500009. 10. Abramson JS, Palomba L, Gordon LI, Lunning M, Arnason J, Forero-Torres A, Albertson TM, Exton VS, Sutherland C, Xie B, Snodgrass S, Siddiqi T. Transcend NHL 001: Immunotherapy with the CD19-Directed CAR T-Cell Product JCAR017 Results in High Complete Response Rates in Relapsed or Refractory B-Cell Non-Hodgkin Lymphoma. Blood. 2016;128(22). PubMed PMID: WOS:000394452500061. 11. Onea AS, Jazirehi AR. CD19 chimeric antigen receptor (CD19 CAR)-redirected adoptive T-cell immunotherapy for the treatment of relapsed or refractory B-cell Non- Hodgkin's Lymphomas. Am J Cancer Res. 2016;6(2):403-24. PubMed PMID: WOS:000374216800022. 12. Lesokhin AM, Ansell SM, Armand P, Scott EC, Halwani A, Gutierrez M, Millenson MM, Cohen AD, Schuster SJ, Lebovic D, Dhodapkar M, Avigan D, Chapuy B, Ligon AH, Freeman GJ, Rodig SJ, Cattry D, Zhu L, Grosso JF, Bradley Garelik MB, Shipp MA, Borrello I, Timmerman J. Nivolumab in Patients With Relapsed or Refractory Hematologic Malignancy: Preliminary Results of a Phase Ib Study. J Clin Oncol. 2016;34(23):2698-704. Epub 2016/06/09. doi: 10.1200/JCO.2015.65.9789. PubMed PMID: 27269947; PMCID: PMC5019749 online at www.jco.org. Author contributions are found at the end of this article. 13. Burger JA, Kipps TJ. CXCR4: a key receptor in the crosstalk between tumor cells and their microenvironment. Blood. 2006;107(5):1761-7. Epub 2005/11/05. doi: 10.1182/blood-2005-08-3182. PubMed PMID: 16269611. 14. Quail DF, Joyce JA. Microenvironmental regulation of tumor progression and metastasis. Nat Med. 2013;19(11):1423-37. doi: 10.1038/nm.3394. PubMed PMID: WOS:000326920300020. 15. Whiteside TL. The tumor microenvironment and its role in promoting tumor growth. Oncogene. 2008;27(45):5904-12. Epub 2008/10/07. doi: 10.1038/onc.2008.271. PubMed PMID: 18836471; PMCID: PMC3689267.

18 16. Bhowmick NA, Neilson EG, Moses HL. Stromal fibroblasts in cancer initiation and progression. Nature. 2004;432(7015):332-7. Epub 2004/11/19. doi: 10.1038/nature03096. PubMed PMID: 15549095; PMCID: PMC3050735. 17. Vitolo U, Seymour JF, Martelli M, Illerhaus G, Illidge T, Zucca E, Campo E, Ladetto M, Committee EG. Extranodal diffuse large B-cell lymphoma (DLBCL) and primary mediastinal B-cell lymphoma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2016;27(suppl 5):v91-v102. Epub 2016/07/06. doi: 10.1093/annonc/mdw175. PubMed PMID: 27377716. 18. Yu H, Mouw JK, Weaver VM. Forcing form and function: biomechanical regulation of tumor evolution. Trends Cell Biol. 2011;21(1):47-56. Epub 2010/09/28. doi: 10.1016/j.tcb.2010.08.015. PubMed PMID: 20870407; PMCID: PMC3014395. 19. Apoorva FNU, Tian YF, Pierpont TM, Bassen DM, Cerchietti L, Butcher JT, Weiss RS, Singh A. Award Winner in the Young Investigator Category, 2017 Society for Biomaterials Annual Meeting and Exposition, Minneapolis, MN, April 05-08, 2017: Lymph node stiffness-mimicking hydrogels regulate human B-cell lymphoma growth and cell surface receptor expression in a molecular subtype-specific manner. J Biomed Mater Res A. 2017;105(7):1833-44. Epub 2017/02/09. doi: 10.1002/jbm.a.36031. PubMed PMID: 28177577. 20. Ruan J, Hajjar K, Rafii S, Leonard JP. Angiogenesis and antiangiogenic therapy in non-Hodgkin's lymphoma. Ann Oncol. 2009;20(3):413-24. Epub 2008/12/18. doi: 10.1093/annonc/mdn666. PubMed PMID: 19088170; PMCID: PMC2733074. 21. Ruan J, Luo M, Wang C, Fan L, Yang SN, Cardenas M, Geng H, Leonard JP, Melnick A, Cerchietti L, Hajjar KA. Imatinib disrupts lymphoma angiogenesis by targeting vascular pericytes. Blood. 2013;121(26):5192-202. Epub 2013/05/02. doi: 10.1182/blood-2013-03-490763. PubMed PMID: 23632889; PMCID: PMC3695363. 22. Reddy ST, van der Vlies AJ, Simeoni E, Angeli V, Randolph GJ, O'Neil CP, Lee LK, Swartz MA, Hubbell JA. Exploiting lymphatic transport and complement activation in nanoparticle vaccines. Nat Biotechnol. 2007;25(10):1159-64. Epub 2007/09/18. doi: 10.1038/nbt1332. PubMed PMID: 17873867.

19 23. Swartz MA, Kaipainen A, Netti PA, Brekken C, Boucher Y, Grodzinsky AJ, Jain RK. Mechanics of interstitial-lymphatic fluid transport: theoretical foundation and experimental validation. J Biomech. 1999;32(12):1297-307. Epub 1999/11/24. PubMed PMID: 10569708. 24. Triacca V, Guc E, Kilarski WW, Pisano M, Swartz MA. Transcellular Pathways in Lymphatic Endothelial Cells Regulate Changes in Solute Transport by Fluid Stress. Circ Res. 2017;120(9):1440-52. Epub 2017/01/29. doi: 10.1161/CIRCRESAHA.116.309828. PubMed PMID: 28130294. 25. Swartz MA, Berk DA, Jain RK. Transport in lymphatic capillaries. I. Macroscopic measurements using residence time distribution theory. Am J Physiol. 1996;270(1 Pt 2):H324-9. Epub 1996/01/01. doi: 10.1152/ajpheart.1996.270.1.H324. PubMed PMID: 8769768. 26. Berk DA, Swartz MA, Leu AJ, Jain RK. Transport in lymphatic capillaries. II. Microscopic velocity measurement with fluorescence photobleaching. Am J Physiol. 1996;270(1 Pt 2):H330-7. Epub 1996/01/01. doi: 10.1152/ajpheart.1996.270.1.H330. PubMed PMID: 8769769. 27. Lenz G, Wright G, Dave SS, Xiao W, Powell J, Zhao H, Xu W, Tan B, Goldschmidt N, Iqbal J, Vose J, Bast M, Fu K, Weisenburger DD, Greiner TC, Armitage JO, Kyle A, May L, Gascoyne RD, Connors JM, Troen G, Holte H, Kvaloy S, Dierickx D, Verhoef G, Delabie J, Smeland EB, Jares P, Martinez A, Lopez-Guillermo A, Montserrat E, Campo E, Braziel RM, Miller TP, Rimsza LM, Cook JR, Pohlman B, Sweetenham J, Tubbs RR, Fisher RI, Hartmann E, Rosenwald A, Ott G, Muller- Hermelink HK, Wrench D, Lister TA, Jaffe ES, Wilson WH, Chan WC, Staudt LM, Pr LLMP. Stromal Gene Signatures in Large-B-Cell Lymphomas. New Engl J Med. 2008;359(22):2313-23. doi: DOI 10.1056/NEJMoa0802885. PubMed PMID: WOS:000261270300003. 28. Yamada KM, Cukierman E. Modeling tissue morphogenesis and cancer in 3D. Cell. 2007;130(4):601-10. Epub 2007/08/28. doi: 10.1016/j.cell.2007.08.006. PubMed PMID: 17719539.

20 29. Weigelt B, Ghajar CM, Bissell MJ. The need for complex 3D culture models to unravel novel pathways and identify accurate biomarkers in breast cancer. Adv Drug Deliv Rev. 2014;69-70:42-51. Epub 2014/01/15. doi: 10.1016/j.addr.2014.01.001. PubMed PMID: 24412474; PMCID: PMC4186247. 30. McKinnon DD, Domaille DW, Cha JN, Anseth KS. Biophysically defined and cytocompatible covalently adaptable networks as viscoelastic 3D cell culture systems. Adv Mater. 2014;26(6):865-72. Epub 2013/10/16. doi: 10.1002/adma.201303680. PubMed PMID: 24127293; PMCID: PMC4582033. 31. Tibbitt MW, Anseth KS. Hydrogels as extracellular matrix mimics for 3D cell culture. Biotechnol Bioeng. 2009;103(4):655-63. Epub 2009/05/28. doi: 10.1002/bit.22361. PubMed PMID: 19472329; PMCID: PMC2997742. 32. Kloxin AM, Lewis KJ, DeForest CA, Seedorf G, Tibbitt MW, Balasubramaniam V, Anseth KS. Responsive culture platform to examine the influence of microenvironmental geometry on cell function in 3D. Integr Biol (Camb). 2012;4(12):1540-9. Epub 2012/11/10. doi: 10.1039/c2ib20212c. PubMed PMID: 23138879; PMCID: PMC3928973. 33. Tian YF, Ahn H, Schneider RS, Yang SN, Roman-Gonzalez L, Melnick AM, Cerchietti L, Singh A. Integrin-specific hydrogels as adaptable tumor organoids for malignant B and T cells. Biomaterials. 2015;73:110-9. Epub 2015/09/26. doi: 10.1016/j.biomaterials.2015.09.007. PubMed PMID: 26406451; PMCID: PMC4623826. 34. Zhang J, Grubor V, Love CL, Banerjee A, Richards KL, Mieczkowski PA, Dunphy C, Choi W, Au WY, Srivastava G, Lugar PL, Rizzieri DA, Lagoo AS, Bernal- Mizrachi L, Mann KP, Flowers C, Naresh K, Evens A, Gordon LI, Czader M, Gill JI, Hsi ED, Liu QQ, Fan A, Walsh K, Jima D, Smith LL, Johnson AJ, Byrd JC, Luftig MA, Ni T, Zhu J, Chadburn A, Levy S, Dunson D, Dave SS. Genetic heterogeneity of diffuse large B-cell lymphoma. P Natl Acad Sci USA. 2013;110(4):1398-403. doi: 10.1073/pnas.1205299110. PubMed PMID: WOS:000314453900055. 35. Reddy A, Zhang J, Davis NS, Moffitt AB, Love CL, Waldrop A, Leppa S, Pasanen A, Meriranta L, Karjalainen-Lindsberg ML, Norgaard P, Pedersen M, Gang

21 AO, Hogdall E, Heavican TB, Lone W, Iqbal J, Qin Q, Li G, Kim SY, Healy J, Richards KL, Fedoriw Y, Bernal-Mizrachi L, Koff JL, Staton AD, Flowers CR, Paltiel O, Goldschmidt N, Calaminici M, Clear A, Gribben J, Nguyen E, Czader MB, Ondrejka SL, Collie A, Hsi ED, Tse E, Au-Yeung RKH, Kwong YL, Srivastava G, Choi WWL, Evens AM, Pilichowska M, Sengar M, Reddy N, Li S, Chadburn A, Gordon LI, Jaffe ES, Levy S, Rempel R, Tzeng T, Happ LE, Dave T, Rajagopalan D, Datta J, Dunson DB, Dave SS. Genetic and Functional Drivers of Diffuse Large B Cell Lymphoma. Cell. 2017;171(2):481-94 e15. Epub 2017/10/07. doi: 10.1016/j.cell.2017.09.027. PubMed PMID: 28985567; PMCID: PMC5659841. 36. Rosenwald A, Wright G, Chan WC, Connors JM, Campo E, Fisher RI, Gascoyne RD, Muller-Hermelink HK, Smeland EB, Staudt LM, Pr LLMP. The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma. New Engl J Med. 2002;346(25):1937-47. doi: DOI 10.1056/NEJMoa012914. PubMed PMID: WOS:000176266200002. 37. Meyer PN, Fu K, Greiner T, Smith L, Delabie J, Gascoyne R, Ott G, Rosenwald A, Braziel R, Campo E, Vose J, Lenz G, Staudt L, Chan W, Weisenburger DD. The Stromal Cell Marker SPARC Predicts for Survival in Patients With Diffuse Large B- Cell Lymphoma Treated With Rituximab. Am J Clin Pathol. 2011;135(1):54-61. doi: 10.1309/Ajcpjx4bjv9nlqhy. PubMed PMID: WOS:000285538700006. 38. McGranahan N, Swanton C. Clonal Heterogeneity and Tumor Evolution: Past, Present, and the Future. Cell. 2017;168(4):613-28. Epub 2017/02/12. doi: 10.1016/j.cell.2017.01.018. PubMed PMID: 28187284. 39. Marks DL, Olson RL, Fernandez-Zapico ME. Epigenetic control of the tumor microenvironment. Epigenomics. 2016;8(12):1671-87. Epub 2016/10/05. doi: 10.2217/epi-2016-0110. PubMed PMID: 27700179; PMCID: PMC5289035.

CHAPTER 2 How Biophysical Forces Regulate Human B Cell Lymphomas FNU Apoorva1, Alexander M. Loiben2, Shivem B. Shah2, Alberto Purwada2, Lorena Fontan3, Rebecca Goldstein3, Brian J. Kirby1,3, Ari M. Melnick3, Benjamin D. Cosgrove2, and Ankur Singh1,2,4,5*

1Sibley School of Mechanical and Aerospace Engineering, Cornell University, Ithaca 14853, USA 2Nancy E. and Peter C. Meinig School of Biomedical Engineering, Cornell University, Ithaca 14853, USA 3Division of Hematology/Oncology, Department of Medicine, Weill Cornell Medical College, New York, NY 10021, USA 4Englander Institute for Precision Medicine, Weill Cornell Medical College, New York, NY 10021, USA

(Cell reports 23 (2), 499-511) Summary

The role of microenvironment-mediated biophysical forces in human lymphomas remains elusive. Diffuse large B cell lymphomas (DLBCLs) are heterogeneous tumors, which originate from highly proliferative germinal center B cells. These tumors, their associated neo-vessels, and lymphatics presumably expose cells to particular fluid flow and survival signals. Here, we show that fluid flow enhances proliferation and modulates response of DLBCLs to specific therapeutic agents. Fluid flow upregulates surface expression of B cell receptors (BCRs) and integrin receptors in subsets of ABC- DLBCLs with either CD79A/B mutations or WT BCRs, similar to what is observed with xenografted human tumors in mice. Fluid flow differentially upregulates signaling targets, such as SYK and p70S6K, in ABC-DLBCLs. By selective knockdown of CD79B and inhibition of signaling targets, we provide mechanistic insights into how fluid flow mechanomodulates BCRs and integrins in ABC-DLBCLs. These findings

23 redefine microenvironment factors that regulate lymphoma-drug interactions and will be critical for testing targeted therapies.

1. Introduction

Diffuse Large B cell lymphoma (DLBCLs) are lymphoproliferative tumors arise from proliferative immune cells in lymphoid tissues. Gene expression profiling has enabled DLBCLs to be sub-classified into germinal center B cell (GCB) DLBCL and activated B cell (ABC) DLBCL subtypes (52-56). The current therapy involves a chemo- immunotherapy regimen containing CHOP (cyclophosphamide, hydroxyldaunomycin [doxorubicin], oncovin [vincristine], and prednisone) with Rituximab (a chimeric anti- CD20 IgG1 monoclonal antibody). However, a significant percentage of DLBCL patients are not cured by this treatment (2). ABC-DLBCL is the most chemo-resistant DLBCL subtype with a 5-year overall survival as low as 45% versus 80% for germinal center B cell-like (GCB) DLBCL (1, 3). Understanding factors that promote resistance to drug therapy and identifying new therapeutic targets are important to improve clinical outcome of DLBCL patients.

Targeting hallmark pathways of ABC-DLBCL, such as those downstream of the chronically activated B cell receptor (BCR) signaling (53, 56-58), has the potential to impact a broad cross-section of ABC-DLBCL patients (58, 59). Importantly, even when chronically activated by somatic mutations, the BCR pathway still needs signals from the microenvironment to drive cell survival, and yet extracellular factors that regulate BCR signaling remain less understood. The BCR is a transmembrane protein complex composed of heavy-chain and light-chain immunoglobulins (Igs), CD79A/Igα and CD79B/Igβ (60). ABC-DLBCLs commonly manifest somatic mutation of components in the BCR pathway, such as CD79A/B (~20% of ABC-DLBCLs) (56), CARD11 (~10%) (61), and several others. Proposed therapeutic strategies for ABC-DLBCL target proteins signaling downstream of the BCR pathway, including kinase inhibitors targeting Spleen tyrosine kinase (SYK), and Bruton’s tyrosine kinase (BTK), among others (52, 57). However, the pattern of response to BCR targeted therapies varies

24 according to mutations present in a given ABC-DLBCL. For example, a SYK shRNA suppresses the growth of ABC-DLBCL cell line, HBL-1(56), which expresses a CD79B mutation in the IgM BCR. In contrast, SYK shRNA is less effective in the ABC-DLBCL cell line, OCI-LY10, with a CD79A mutation. This underscores the need for better understanding the regulators of BCR signaling in heterogeneous subclasses of ABC- DLBCLs, under growth conditions that mimic tumor microenvironment.

A major impediment in the field is that, unlike other solid tumors, the importance of the physical nature of the lymphoma microenvironment has not been studied in detail (5). We have recently shown that the crosstalk between lymphoid tissue’s extracellular matrix , stiffness, and integrins on lymphoma cells are critical for tumor cell survival and signaling, both in vitro and in vivo (45, 62, 63). Once DLBCL cells seed a lymphoid tissue, malignant B cells proliferate steadily, causing massive distortion, enlargement and vascularization of these tumor-seeded tissues as shown by us (35) and others (34). The increased vascularization and lymphatic flow presumably expose the lymphoma cells to fluid flow, i.e. fluid shear stress (tangential forces on cell surface) and nutrient mass transport, which supports their survival, proliferation, and response to drugs. Here we have endeavored to determine how fluid forces regulate human lymphomas. Because the fluid shear stress values in enlarged, vascularized lymphoma tissues are also unknown, the current study focuses on uncovering the biological impact of biophysical forces at the peak fluid shear stress of 0.8 – 1.3 dynes/cm2, reported earlier to be applicable in the subcapsular sinus lumen of lymphoid tissues (64). Conventional tissue culture plates, which do not induce shear stress and large bioreactors (shear stress of 6- 12 dynes/cm2) are not suitable to study the effect low fluid shear stresses in the range of 0.8-1.3 dynes/cm2. Here, using an ex vivo lymphoma micro-reactor platform and inhibiting various key signaling proteins and knocking down a CD79 complex, we demonstrate, for the first time to our knowledge, that fluid shear stress alone or in combination with nutrient mass transport can act as an important biophysical stimulus on DLBCLs by influencing the BCR and integrin signaling, with consequences on cell proliferation, differentially dependent on heterogeneous DLBCL subclasses.

2. Results and Discussion

25 2.1 Ex vivo lymphoma micro-reactor with high-resistance channels

Our primary goal was to characterize how DLBCLs respond to fluid forces reported earlier in the subcapsular sinus lumen of lymphoid tissues. Therefore, we designed our micro-reactors to achieve a fluid shear stress of ~ 0.3-1.3 dynes/cm2 (Figure 1A). Our secondary goal was to maintain the pressure difference in the micro-reactor close to the physiological pressure difference of 1 mm Hg (or 133 Pa), previously reported between paracortex and afferent vessels (64). In addition, our goal was to ensure that fluid flow was uniformly distributed such that almost all of the cells in the micro-reactor were exposed to the same biophysical forces. An addition micro-reactor design criteria was to maintain DLBCL cells in 50% conditioned media through media recirculation (45).

To achieve our goal, we considered four micro-reactor designs (Figure 1B and Supplementary Figure S1-3), and computationally modeled them to compare across above-mentioned criterion. Engineering details and analysis of the computational results are included in Supplementary Methods, Supplementary Figure S1-S3, and Supplementary video 1. Of these four micro-reactors, two of the micro-reactors were chosen from previous reports where they showed utility in studying adherent cell types and hematopoietic lineages (65, 66). Our own two micro-reactors were a modification of recently described work (67). These were designed to include a cell culture chamber that was connected to media chamber by 5 µm wide high-resistance narrow channels, which dramatically reduced media flow velocity into the cell chamber (Figure 1B, C). One of our micro-reactors (called side flow) had media flow channel perpendicular to the high-resistance channel (Figure 1C, D) and the other micro-reactor (called cross flow) had media flow chamber aligned with the high-resistance channel (Supplementary Figure S3). Based on computational analysis, we determined that the side flow micro-reactors (Figure 1B) resulted in uniformly distributed flow patterns and shear stress (0.3 – 1.3 dynes/cm2) with approximately 100% of the cell culture area exposed to a uniform fluid flow (Figure 1E). Importantly, the pressure difference across the side-flow micro-reactor was 100 Pa, which is very close to the physiological value of 1 mm Hg (or 133 Pa) (64). These two computational outputs benchmarked the superiority of our side flow micro-reactor against alternative micro-reactor platforms

26 tested in this study, where only 25 – 66% area of the micro-reactor was exposed to uniform shear forces and the pressure difference was markedly below 1 mmHg (2-20 Pa, Figure 1E and Supplementary Figure S1E).

In the side-flow micro-reactors, cells were introduced into the central chamber using the “Cell Inlet” port (Figure 1B). The media was flown through the “Media Inlet” port and perfused into the central chamber as indicated by purple arrows in Figure 1B. The media was recirculated and diluted by connecting the “Media Inlet” and “Media Outlet” ports by means of a rotary pump (Figure 1C). The central cell culture chamber was connected to the circulating media channel (width 20 µm) through high-resistance perfusion channels (width 5 µm), which prevented the 8-10 µm cells from leaving the central cell chamber (Figure 1D). Dimensions of the central chamber were length 3 cm, height 100 µm, and width 0.6 cm. To prevent the top of cell culture chamber from collapsing, we included 40 µm wide PDMS micro-pillars, separated by near 800 µm distance within the central chamber of the micro-reactor (Figure 1C, D), which did not significantly affect the flow rate. Since DLBCLs are loosely adherent cells and conventionally cultured in tissue culture plates without any substrates, the micro- reactors were not pre-coated with any extracellular matrices. However, these cells can still adhere to serum proteins, such as fibronectin (68), using integrin receptors. When DLBCL cells were cultured in these micro-reactors (without substrates) over 24 hr, lymphoma cells were uniformly distributed throughout the micro-reactor and did not accumulate against the micro-reactor wall (Figure 1F). The cells in the micro-reactor showed excellent viability, as qualitatively determined using a live-dead assay (Figure 1F).

2.2 Fluid flow enhances proliferation of human DLBCL cells and regulates drug-induced apoptosis in lymphoma cells

We hypothesized that in contrast to static growth conditions, fluid flow, either as shear stress alone or in combination with mass transport, will modulate proliferation of lymphoma cells through mechanical stimulation of surface receptors such as integrins. Our secondary hypothesis was that fluid flow-mediated changes in integrin stimulation will modulate BCR through outside-in crosstalk. We first examined the role of fluid

27 flow on the growth of BCR-dependent ABC-DLBCL cell lines (Figure 1G) with CD79A mutations (OCI-LY10) and CD79B mutations (HBL-1). We exposed OCI- LY10 and HBL-1 to a fluid flow (i.e. shear stress of ~1.3 dyne/cm2). We observed significantly higher proliferation that started as early as day 3 in flow conditions as compared to static cultures and continued to grow with time (Figure 1H).

Although DLBCL cell lines are not traditionally cultured with immune or stromal supports, researchers have explored the use of follicular dendritic cells in lymphoma cultures as immune-stromal support (69). Therefore, we incorporated human tonsil derived follicular dendritic cells (45, 69) into these cultures (Figure 1I). A 48 hr cell proliferation analysis confirmed that even in the presence of follicular dendritic cells, both OCI-LY10 and HBL-1 proliferated more under flow conditions than static conditions (Figure 1J), suggesting that micro-reactors are suitable for mono- and-co- cultures of DLBCLs.

We next determined whether fluid flow changes the abundance of cytokines secreted by OCI-LY10 and HBL-1 cells in the media, which could have autocrine and paracrine effects. As indicated in Figure 1K, we observed a significant increase in TNFα, and GM-CSF in media from OCI-LY10 cells grown under flow conditions for 72 hr. In contrast, there was no significant change in aforementioned proteins in media from HBL-1 cells grown under flow conditions. There was modest or no difference in the remaining tested cytokines (IFNγ, IL1A, IL17A, IL1B, IL4, IL6, IL8, and IL12) (Supplementary Figure S4). Several of these cytokines play an important and widespread role in cancer (70-75).

Since DLBCLs proliferate better in flow conditions, we hypothesized that the healthier status of cells under flow conditions will reduce the apoptosis of ABC- DLBCLs when exposed to anti-lymphoma drugs. We first determined the uptake of the drug doxorubicin by DLBCLs under static and flow conditions using flow cytometry (Doxorubicin is fluorescesent (76)). These experiments were performed by culturing DLBCLs in micro-reactors for 72 hr, followed by 24 hr of exposure to the drug. We did not observe significant differences in uptake of the drug (Figure 2A). Both HBL-1 and OCI-LY10 cells exhibited a significant decrease in apoptosis (Annexin-V+) when

28 exposed to doxorubicin under flow conditions (Figure 2B-D). Finally, we tested whether the reduced apoptosis could be extended to other classes of anti-cancer therapeutics, and evaluated the histone deacetylate inhibitor (HDACi, Panobinostat) response under flow conditions. As illustrated in Figure 2E, a similar pattern of reduced apoptosis was observed under flow conditions as compared to static conditions. Collectively these results suggest that drug-induced DLBCL apoptosis is affected by the presence of fluid flow.

2.3 Fluid flow upregulates the surface expression of B cell receptors (BCR) and integrin receptors in ABC-DLBCL, similar to xenografted human tumors in mice

The BCR is vital for the growth of ABC-DLBCL through downstream pathways and abrogation of BCR signaling results in proliferation arrest of ABC-DLBCL (53, 56). Hence targeting of BCR signaling pathways has emerged as a promising therapeutic approach for these tumors (53, 59, 77, 78). To determine whether flow conditions affect the level of expression of the BCR, which could affect its signaling, we measured the levels of IgM BCR, by flow cytometry. As indicated in Figure 3A (top panel) and 3B, in HBL-1 ABC-DLBCLs (CD79B mutation), cell surface expression of IgM was significantly increased by 3.8 fold under flow conditions. By contrast, in OCI-LY10 ABC-DLBCLs (CD79A mutation), there was no change under flow conditions. Notably, OCI-LY3 ABC-DLBCL cells, which have WT CD79A/B, also exhibited significantly increased IgM expression under flow conditions. However, no differences were observed in the GCB-DLBCL cell line OCI-LY1. Collectively, these findings suggest that true levels of expression of the BCR are under-appreciated in 2D static cultures and the BCR expression is differentially upregulated among subsets of ABC- DLBCLs.

Although the majority of B cell lymphomas express surface CD20 receptor, only 76% patient response to CHOP plus rituximab(79) and 50% to rituximab alone (80, 81). The differences in response have been attributed to decreased CD20 expression observed in some patients treated with rituximab (81). Therefore, understanding mechanisms that regulate CD20 expression is important. We determined the effect of

29 fluid shear stress on CD20 expression by exposing the cells to either static or flow conditions. As indicated in Figure 3C, fluid flow induced 3.5-fold higher upregulation of CD20 on the surface of OCI-LY10 cells (p<0.05). In contrast, we observed only 1.3- fold upregulation of CD20 in HBL-1 cells.

Integrins have been shown to play important roles in transmitting mechanical stimuli received from fluid shear stress into intracellular signals in a wide variety of cells (82, 83). In lymphoid cells, integrins can function as pro-survival signals (62, 68), and upregulation of α4β1 integrins in DLBCLs provides a protective pro-survival signal (45, 84). We therefore determined the role of fluid shear stress on the expression of α4β1 integrins on ABC-DLBCL cells. Flow induced a significant 1.7-fold higher expression of α4 subunits in OCI-LY10 and a 4.3-fold increase in HBL-1 (Figure 3A (lower panel) and 3D). No differences were observed in OCI-LY3 and OCI-LY1 cells. HBL-1 cells also exhibited a significant 1.3-1.5-fold increase in the β1 integrin subunit under flow conditions (p<0.05) (Figure 3E and Supplementary Figure S5A). These findings suggest that the expression levels of integrin mechanoreceptors on ABC-DLBCL cells are dependent on fluid forces.

To determine whether changes in integrin expression were observed in vivo, we performed a direct comparison of HBL-1 and OCI-LY10 cell lines to their xenografted tumors in immunocompromised NSG (NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ) mice, a model of choice for cancer xenograft modeling. Xenografts were harvested when tumors grew, and BCR, integrin α4, and integrin β1 levels were evaluated to compare with their levels on cells before implantation (representing static conditions). As indicated in Figure 3F, tumors developed in NSG mice implanted with OCI-LY10 xenografts, often well vascularized. Flow cytometric analysis showed higher mean fluorescence intensity in integrin α4 signal in xenografted cells (Figure 3F, bottom panel). Quantitative analysis of the surface expression level of IgM (as a measure of the BCR), integrin α4, and integrin β1 on HBL-1 xenografts under static (in vitro conditions) versus in vivo conditions show a similar pattern of differential expression as was observed in our micro-reactors (Figure 3F (right panel) and Figure 3G). These results clearly demonstrate that lymphoma micro-reactors are

30 more similar to tumors in vivo, therefore confirming the power and importance of a microfluidic-based model system.

We were intrigued by the question of what these integrin receptors bind to in a tissue culture system that is not pre-coated with an extracellular matrix. Although lymphomas are loosely adherent cells, in our experiments we did not see these cells moving under flow conditions. This led us to hypothesize that there was weak but significant adhesion between DLBCLs and the underlying glass substrate in the micro-reactors, likely due to the serum proteins deposited from the media. Since β1 is a common integrin subunit involved in dimerization with multiple alpha subunits, including α4 and α5, we determined whether β1 integrin was involved in adhesion of DLBCL cells to the underlying surface. When integrin β1 was blocked using a monoclonal anti-β1 antibody, HBL-1 ABC-DLBCL cells accumulated at the edges and a few high-resistance channels of the micro-reactor through which media is supposed to exit the cell culture chamber (Figure 3H, Supplementary Video 2 and 3). In contrast, in the absence of blocking anti-β1 integrin antibody, the cells remained in their initial position, suggesting that β1 integrin was a surface receptor involved in adhesion to the floor of the micro-reactors. Unlike HBL-1, blocking β1 integrin did not markedly affect the adhesion of OCI-LY10 under flow conditions despite the fact that both HBL-1 and OCI-LY10 express similar levels of integrin β1 (Supplementary Figure S5B and C). Finally, we also tested whether α5 integrin was expressed by these cells, and observed a very modest increase in HBL-1 cells grown in flow conditions, but not in OCI-LY3 and OCI-LY10 (Figure 3I), suggesting α5 integrin may not play an important role in our observations.

We also determined whether the pattern of increase or decrease in these surface receptors under fluid flow was maintained under 3D extracellular matrix rich conditions, which involve different relative tangent and normal shear stresses. We encapsulated ABC-DLBLCs in Matrigel™ matrix inside the micro-reactors and exposed them to the same shear stress (1.3 dynes/cm2) as in the 2D micro-reactors. The change in BCR and integrin α4 followed similar patterns as in conditions without a matrix (Supplementary Figure S5D). However, we observed consistent upregulation of β1 subunit, which could be attributed to the engagement of β1 integrins by

31 Matrigel™, causing their upregulation (85). Although we chose Matrigel™ for this study because of its extensive use by others in the cancer field, we have previously shown that ABC-DLBCL subtypes have a heterogeneous need for integrin subtypes for survival and Matrigel™ may not be the ideal matrix (45, 86). Therefore, future studies will focus on understanding the role of fluid flow and ligands specific to the integrin receptors found on each DLBCL subtype.

2.4 Fluid flow differentially regulates phosphorylation of complementary signaling pathways in ABC-DLBCLs

To understand the mechanism of signaling through the integrin and BCR pathways induced by fluid flow, we exposed ABC and GCB-DLBCLs to fluid flow (1.3 dynes/cm2) and examined phosphoprotein levels of various downstream proteins using multiplexed bead-based Luminex assays (87). A GCB-DLBCL cell line, OCI-LY1, was used as a negative control. Since we observed that surface expression of BCR is differentially modulated in ABC-DLBCL cells under flow conditions, we determined the phosphorylation of immediate downstream molecules LYN (p-Lyn) and SYK (p- SYK), which are critical for activation of the BCR. We found that p-LYN and p-SYK exhibited >2-fold increase in HBL-1 due to fluid flow exposure, whereas in OCI-LY10, there was no change due to fluid flow (Figure 4A-B). p-LYN was also increased in OCI-LY3 cells, however, no significant increase in p-SYK was observed. We further observed no change in p-LYN and p-SYK in OCI-OCI-LY1 cells under flow conditions. Upregulation of p-LYN and p-SYK with HBL-1 (CD79B mutant) but not in OCI-LY10 (CD79A mutant) is in agreement with our observations in Figure 3 that there is significant upregulation of the BCR in HBL-1, and not in OCI-LY10. Our observations are further aligned with the finding by Staudt and colleagues (56) that shRNA targeting SYK suppresses the growth of ABC-DLBCL line HBL-1 but not OCI-LY10, OCI-LY3 or GCB-DLBCL lines. Typically, BCR activation can trigger downstream AKT signal, which activates mTOR by inhibiting the tuberous sclerosis complex, leading to p70S6K activation. We observed elevated levels of p-AKT in these cells and p-p70SK expression level was increased by a significant >3-fold in HBL-1 when exposed to fluid flow, whereas OCI-LY10 showed a marginal decrease or no change with fluid flow

32 (Figure 4C and Supplementary Figure S6).

We next explored the impact of fluid flow on the MAPK signaling pathway (88, 89), which is also linked to the BCR receptor pathway (Figure 4D-F). The phosphorylation of p38, and the downstream target cJUN was increased >3-fold in HBL-1 and 2-fold in OCI-LY3, when exposed to the fluid flow. In contrast, OCI-LY10 and OCI-LY1 showed a marginal decrease or no change with fluid flow. While OCI-LY3 manifested a small but significant increase in p-ERK levels under flow conditions, no such increase was observed in OCI-LY10 and HBL-1 cells.

2.5 Understanding the contributions of fluid shear stress and mass transport in ABC-DLBCL surface receptor expression

The differences seen in the expression of the surface receptors can be attributed to the effect of fluid shear stress (i.e. more forces on cell receptors from the flow), or increased mass transport (i.e. faster removal of secreted factors and metabolic products because the flow carries them away), or both. We created matched experiments with culture solutions with different transport properties to establish that shear stress at the cell surface plays a prominent role in generating the differences seen in BCR, integrin, and CD20 expression. We added dextran to the culture solution to increase the solution viscosity and thus increase the tangential stress on cells in flow (Figure 5A-E). Adding dextran reduces the rate at which species diffuse toward or away from the cell (90). The addition of 1% dextran increased the viscosity by 10 times (Supplementary Figure S7A) and is expected to have decreased solute molecular diffusivity by a similar amount. We then performed the following comparisons: 1) compared static culture with and without dextran, where the addition of dextran reduced solute molecular diffusivity and is expected to reduce transport of secreted factors and metabolic products, and 2) compared flow culture with and without dextran, in which we increased the viscosity of flow media and reduced flow by 10 times to maintain the same tangential shear stress (1.3 dynes/cm2). For the flow-culture comparison, the flow with dextran had lower diffusion of species than that without dextran because the dextran reduced solute diffusivity. Further, the flow with dextran had lower convection of species than that without dextran because for the same tangential shear stress a lower velocity flow was

33 used.

Under these defined conditions, we found that there was no effect of static versus flow versus flow (viscous) on α4 integrin mechanoreceptor expression in HBL-1 cells (Figure 5A). This suggests that the increase in α4 integrin in HBL-1 cells is primarily due to fluid shear stress and not diffusive or convective-diffusive mass transport. In contrast, OCI-LY10 cells showed a significant decrease in α4 integrin expression under viscous flow conditions as compared to regular flow conditions. Because fold-change values in flow (viscous) conditions were still higher than under static conditions, we concluded that the effect in OCI-LY10 is a combined contribution from mass transport and fluid shear stress. We did not see differences in α4 integrin expression in static vs static (viscous) for OCI-LY3 or HBL-1(Figure 5D).

The cell surface expression of the remaining markers CD20, β1 integrin, and BCR (Figure 5A-D) showed a significant decrease in surface expression in OCI-LY10 and HBL-1 cells when cultured under flow (viscous) conditions as compared to flow conditions. Because expression under flow (viscous) values were still higher than static conditions, we concluded that the effect in HBL-1 and OCI-LY10 were a combinatorial effect of mass transport and fluid shear stress for these receptors. The only exception here is the expression of integrin β1 in OCI-LY10, which showed an increase in surface expression under viscous conditions, the reasons for which are not yet clear and will require further investigation. We further compared static (viscous) versus static conditions, and determined that the there was no significant difference in IgM/BCR and β1 integrin expression levels in HBL-1 (Figure 5E).

2.6 Mechanistic model of fluid flow-mediated mechanomodulation of ABC- DLBCLs.

Based on our findings in previous sections, we propose a mechanomodulation model in HBL-1 ABC-DLBCLs. Here, fluid flow upregulates the cell surface expression of IgM, integrin α4 and β1, which in turn increases phosphorylation of LYN, SYK, p38, c-JUN, and p70S6K, and affects cell proliferation (Figure 6A). In contrast, fluid flow does not upregulate IgM and integrin signaling in OCI-LY10 and other complementary

34 mechanism may be responsible for its proliferation. To prove our hypothesis, we blocked integrin receptors by using an anti-α4β1 antibody over 48 hr and observed statistically significant ~50% decrease in proliferation of HBL-1 but no decrease in OCI-LY10 cells (Figure 6B). These findings indeed support the hypothesis that fluid flow mechanomodulates integrin receptors to control HBL-1 proliferation.

To further understand the role of IgM BCR and the interplay of signaling molecules that are upregulated in HBL-1 ABC-DLBCLs, we first inhibited SYK by using two complementary SYK inhibitors R406 or PRT062607. The dose of R406 and PRT062607 was chosen at maximum inhibitory dose, which is capable of inhibiting the activity of signaling target but remains non-toxic (cell survival 94 ± 4 %, Supplementary Figure S7B). As indicated in Figure 6C, inhibition of SYK (R406, 200 nM) prevented an increase in HBL-1 proliferation under flow conditions, suggesting that SYK signaling regulates cell proliferation. Inhibition of SYK did not reduce the integrin β1 expression levels under flow conditions (Figure 6D), suggesting that the limiting of proliferation was attributable to the blocking of SYK signaling, independent of any effect in modulating integrin expression levels. We further verified our results by using PRT062607 (91) at 5 nM dose. The anti-SYK activity of PRT062607 is at least 80-fold greater than for other kinases (92). As shown in Figure 6E, cell proliferation and expression of integrin α4 and β1 was similar between R406 and PRT062607. We next inhibited p70S6K using a mechanistic target of rapamycin (mTOR) inhibitor at 0.5 nM dose (maximum inhibitory dose while still maintaining ~95% viability). As indicated in Figure 6F, inhibition of mTOR prevented an increase HBL-1 proliferation under flow conditions, suggesting that p70S6K signaling also regulates cell proliferation. The mTOR inhibitor did not reduce the IgM expression under flow conditions (Figure 6G and Supplementary Figure S7C), suggesting that the limited proliferation was attributable to the blocking of downstream p70S6K signaling and not a reduction in IgM BCR expression.

2.7 The CD79B complex in BCR cross-talks with integrins under fluid flow conditions to mechanomodulate human DLBCLs.

Finally, we tested whether the CD79B protein in HBL-1 is a target of fluid flow

35 based mechanomodulation and whether it modulates integrin-BCR crosstalk. To study this, we used doxycycline inducible shRNAs against CD79B (or Luciferase as control), which were validated using western blot analysis (Figure 7A, Supplementary Figure S7D). When exposed to doxycycline to induce the shRNA, the cells expressed green fluorescent protein GFP (Figure 7B, Supplementary Figure S7E), and the induced shRNA reduced expression of CD79B (Figure 7C). Under flow conditions, doxycycline inducible shRNA knockdown of CD79B in HBL-1 cells led to significantly decreased proliferation of total cells as well as GFP+ cells as compared to HBL-1 cells transduced with the control shLuc (Figure 7D). Furthermore, under flow conditions, knockdown of the CD79B significantly reduced integrin β1 expression to the levels of static conditions with CD79B knockdown (Figure 7E). In contrast, induction of shRNA against luciferase in HBL-1 cells as a control did not reduce integrin β1 expression. In the same studies, under flow conditions, knockdown of the CD79B did not reduce upstream IgM/BCR expression levels (Figure 7F). Lastly, we observed a significant reduction of pSYK and pLYN expression in as well as GFP+ cells treated with shRNA against CD79B, but not with shRNA against Luciferase (Figure 7G, H). The decrease in pSYK and pLYN correlates with decrease in cell number when CD79B is knocked- out, and confirm that the increase in tumor growth under fluid flow conditions is partially mediated by downstream activation of SYK and LYN, as proposed in the model in Figure 6. Knocking down CD79B in HBL-1 cells under static conditions did not affect pSYK levels (Supplementary Figure S7F). These new findings provide mechanistic insight into the cross-talk of fluid flow, integrin β1, and CD79B, and may explain how heterogeneous subtypes of ABC-DLBCLs are differentially influenced by fluid flow and mass transport.

3. Conclusion

This study highlights how complex biophysical forces in lymphomas differentially regulate heterogeneous B cell lymphoma subtypes and the expression levels of surface receptors, as well as downstream signaling proteins, which otherwise remain comparable under static culture conditions. Our current findings underscore the need

36 for a lymphoma micro- reactor platform that is more biologically similar to the situation in vivo than a simple static 2D culture, to better understand the biology and predict the response of lymphomas to drugs. While it has been shown that CD79A/B mutations do not enhance BCR clustering, nor are they better at inducing nuclear factor kB (NF-kB) activity than their wild-type counterparts, our studies demonstrate that fluid flow has a differential effect on ABC-DLBCL subsets. Future studies will explore whether the differences observed could also be attributed to CD79A/B mutations themselves or other downstream mutations, such as those in CARD11. Future investigations will also determine the effect of knocking down/disrupting multiple pathways and understanding the role of fluid flow in new classes of therapeutic inhibitors (such as those targeting BCR signaling pathway). In the current study, we used fluid shear stresses reported for subcapsular sinuses in lymphoid tissues; however, more detailed measurements of fluid flow and associated stresses in lymphoma tissues need to be performed in a tumor- specific manner, followed by integration into the micro- reactors to model the effect outside of the body. Understanding the fluid flow-mediated effect on lymphomas could provide mechanistic insight into why some DLBCLs are sensitive to treatment while others are refractory and, importantly, allow a faster and more rational translation of targeted therapeutic regimens.

4. Experimental Procedures

4.1 Cell culture in lymphoma micro-reactors

In all studies, we loaded 100,000 DLBCL cells per micro-reactor. In studies that involved human tonsil-derived follicular dendritic cells, co-cultures were seeded at 1:1 ratio, with total cell number being 200,000 per micro-reactor. Static conditions were chosen as standard 2D non-treated tissue culture plates, which are routinely used in lymphoma research. Remaining details of cell lines, flow cytometry and phosphoprotein levels are included in Supplementary Methods.

4.2 Human cell line xenografts in NSG mice: NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice (6-8 weeks) were implanted with OCI-LY10 and HBL-1 xenografts. Bilateral tumor grafts (106 cell/injection) were implanted into the

37 subcutaneous pockets (retronucal and caudal regions) using a syringe with Matrigel. The cut edges were sealed with a suture of fibrin glue. Tumor development was monitored and mice were sacrificed at pre-determined time point. All studies were performed in compliance with the Institutional Animal Care and Use Committee (IACUC) approval at Cornell University.

4.3 Inducible shRNA mediated CD79B knockdown: shRNAs were delivered by lentivirus infection, which were produced by transfection of 293T cells with the LT3GEPIR vector (93). Infected HBL1 cells were selected by puromycin treatment (1µg/mL). Doxycycline (1µg/mL) was used to induce the expression of the indicated shRNAs along with GFP that was used to monitor shRNA expression. Mature antisense sequences of shRNA used to knockdown CD79B and Luciferase were: shCD79B#744: 5’-TTCAGCTGTGCCAAGGTGCTGA-3’, shLuciferase: 5’- TTAATCAGAGACTTCAGGCGGT-3’, respectively. Negative control included cells not treated with Doxycycline. The knockdown of CD79B was validated by western blot analysis. For micro-reactor experiments, after 48 hr of treatment with Doxycycline, cells were exposed to flow or static conditions for 72 hr and Doxycycline concentration was maintained during the course of the experiment. The phospho-SYK (Tyr525/526) (C87C1; #2710) and phospho-LYN (Tyr507) (#2731) were purchased from Cell Signaling Technology, Inc. A donkey anti-rabbit IgG H&L (Alexa Fluor® 647; abcam, ab150075) secondary antibody against these primary antibodies was used for detecting expression levels of pSYK and pLYN.

4.4 Statistical analysis: Statistical analysis was performed using a two-tailed unpaired t-test or a 1-way ANOVA with Tukey’s posthoc test or a 2-way ANOVA with Bonferroni’s correction. Statistical analysis for each plot is described in the figure legend. Statistical analyses were performed using GraphPad Prism software and a p- value of less than 0.05 was considered significant. All values are reported as mean ± SEM. Data presented in this manuscript is a triplicate experiment representative of at least 3 independent experiments, except the NSG mice study.

Acknowledgements

38 The authors acknowledge partial financial support from the National Institutes of Health (1R33CA212968-01, A.S, 1R01AI132738-01A1, A.S, and R00AG042491, B.D.C), Department of Defense Career Development Award (W81XWH-17-1- 0215, A.S.), and the National Science Foundation (CBET-1511914, A.S and B.K). Authors acknowledge Graduate Assistantships in Areas of National Need from the US Department of Education (A.M.L and S.B.S). This work was performed in part at Cornell NanoScale Facility, an NNCI member supported by NSF Grant ECCS- 1542081. Authors acknowledge Scott Butler, Rishi Puri, and Dr. Kristy Richards of Cornell’s Progressive Assessment of Therapeutics (PATh) program for assisting with the xenograft models. We are thankful to Prof. Avery August (Cornell University) for providing feedback on revised manuscript. Opinions, interpretations, conclusions and recommendations are those of the authors and are not endorsed by the funding agencies.

Author contributions

F.A. conducted all experiments, collected data, performed data analysis, and wrote the manuscript with A.S. A.M.L and B.D.C performed phosphorylation analysis; S.B.S and A.P assisted with experiments. B.K advised on computational simulations. L.F, R.G, A.M.M provided reagents. A.S developed the concept and with F.A. contributed to the planning and design of the project.

Declaration of Interests

The authors declare no competing interests.

39 REFERENCES Alizadeh, A.A., Eisen, M.B., Davis, R.E., Ma, C., Lossos, I.S., Rosenwald, A., Boldrick, J.C., Sabet, H., Tran, T., Yu, X., et al. (2000). Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403, 503-511.

Apoorva, F.N.U., Tian, Y.F., Pierpont, T.M., Bassen, D.M., Cerchietti, L., Butcher, J.T., Weiss, R.S., and Singh, A. (2017). Lymph node stiffness-mimicking hydrogels regulate human B-cell lymphoma growth and cell surface receptor expression in a molecular subtype-specific manner. J Biomed Mater Res A 105, 1833-1844.

Blonska, M., Zhu, Y., Chuang, H.H., You, M.J., Kunkalla, K., Vega, F., and Lin, X. (2015). Jun-regulated genes promote interaction of diffuse large B-cell lymphoma with the microenvironment. Blood 125, 981-991.

Brower, V. (2015). Ibrutinib promising in subtype of DLBCL. Lancet Oncol 16, e428.

Burger, J.A., and Wiestner, A. (2018). Targeting B cell receptor signalling in cancer: preclinical and clinical advances. Nat Rev Cancer.

Cayrol, F., Diaz Flaque, M.C., Fernando, T., Yang, S.N., Sterle, H.A., Bolontrade, M., Amoros, M., Isse, B., Farias, R.N., Ahn, H., et al. (2015). Integrin alphavbeta3 acting as membrane receptor for thyroid hormones mediates angiogenesis in malignant T cells. Blood 125, 841-851.

Coiffier, B., Lepage, E., Briere, J., Herbrecht, R., Tilly, H., Bouabdallah, R., Morel, P., Van Den Neste, E., Salles, G., Gaulard, P., et al. (2002). CHOP chemotherapy plus rituximab compared with CHOP alone in elderly patients with diffuse large-B-cell lymphoma. N Engl J Med 346, 235-242.

Cosgrove, B.D., Alexopoulos, L.G., Hang, T.C., Hendriks, B.S., Sorger, P.K., Griffith, L.G., and Lauffenburger, D.A. (2010). Cytokine-associated drug toxicity in human hepatocytes is associated with signaling network dysregulation. Mol Biosyst 6, 1195- 1206.

Dai, B., Zhao, X.F., Hagner, P., Shapiro, P., Mazan-Mamczarz, K., Zhao, S., Natkunam, Y., and Gartenhaus, R.B. (2009). Extracellular signal-regulated kinase positively

40 regulates the oncogenic activity of MCT-1 in diffuse large B-cell lymphoma. Cancer research 69, 7835-7843.

Davis, R.E., Brown, K.D., Siebenlist, U., and Staudt, L.M. (2001). Constitutive nuclear factor kappaB activity is required for survival of activated B cell-like diffuse large B cell lymphoma cells. J Exp Med 194, 1861-1874.

Davis, R.E., Ngo, V.N., Lenz, G., Tolar, P., Young, R.M., Romesser, P.B., Kohlhammer, H., Lamy, L., Zhao, H., Yang, Y., et al. (2010). Chronic active B-cell- receptor signalling in diffuse large B-cell lymphoma. Nature 463, 88-92.

Fellmann, C., Hoffmann, T., Sridhar, V., Hopfgartner, B., Muhar, M., Roth, M., Lai, D.Y., Barbosa, I.A., Kwon, J.S., Guan, Y., et al. (2013). An optimized microRNA backbone for effective single-copy RNAi. Cell Rep 5, 1704-1713.

Fontan, L., and Melnick, A. (2013). Molecular pathways: targeting MALT1 paracaspase activity in lymphoma. Clin Cancer Res 19, 6662-6668.

Fontan, L., Yang, C., Kabaleeswaran, V., Volpon, L., Osborne, M.J., Beltran, E., Garcia, M., Cerchietti, L., Shaknovich, R., Yang, S.N., et al. (2012). MALT1 small molecule inhibitors specifically suppress ABC-DLBCL in vitro and in vivo. Cancer Cell 22, 812- 824.

Frank, N.Y., Margaryan, A., Huang, Y., Schatton, T., Waaga-Gasser, A.M., Gasser, M., Sayegh, M.H., Sadee, W., and Frank, M.H. (2005). ABCB5-mediated doxorubicin transport and chemoresistance in human malignant melanoma. Cancer Res 65, 4320- 4333.

Friedberg, J.W. (2011). Relapsed/refractory diffuse large B-cell lymphoma. Hematology Am Soc Hematol Educ Program 2011, 498-505.

Geng, H., Hurtz, C., Lenz, K.B., Chen, Z., Baumjohann, D., Thompson, S., Goloviznina, N.A., Chen, W.Y., Huan, J., LaTocha, D., et al. (2015). Self-enforcing feedback activation between BCL6 and pre-B cell receptor signaling defines a distinct subtype of acute lymphoblastic leukemia. Cancer Cell 27, 409-425.

Goldstein, R.L., Yang, S.N., Taldone, T., Chang, B., Gerecitano, J., Elenitoba-Johnson,

41 K., Shaknovich, R., Tam, W., Leonard, J.P., Chiosis, G., et al. (2015). Pharmacoproteomics identifies combinatorial therapy targets for diffuse large B cell lymphoma. J Clin Invest 125, 4559-4571.

Hodge, D.R., Hurt, E.M., and Farrar, W.L. (2005). The role of IL-6 and STAT3 in inflammation and cancer. Eur J Cancer 41, 2502-2512.

Jafarnejad, M., Woodruff, M.C., Zawieja, D.C., Carroll, M.C., and Moore, J.E., Jr. (2015). Modeling Lymph Flow and Fluid Exchange with Blood Vessels in Lymph Nodes. Lymphat Res Biol 13, 234-247.

Kawano, M., Hirano, T., Matsuda, T., Taga, T., Horii, Y., Iwato, K., Asaoku, H., Tang, B., Tanabe, O., Tanaka, H., et al. (1988). Autocrine generation and requirement of BSF- 2/IL-6 for human multiple myelomas. Nature 332, 83-85.

Kuppers, R. (2005). Mechanisms of B-cell lymphoma pathogenesis. Nat Rev Cancer 5, 251-262.

Lam, L.T., Wright, G., Davis, R.E., Lenz, G., Farinha, P., Dang, L., Chan, J.W., Rosenwald, A., Gascoyne, R.D., and Staudt, L.M. (2008). Cooperative signaling through the signal transducer and activator of transcription 3 and nuclear factor- {kappa}B pathways in subtypes of diffuse large B-cell lymphoma. Blood 111, 3701- 3713.

Lecault, V., Vaninsberghe, M., Sekulovic, S., Knapp, D.J., Wohrer, S., Bowden, W., Viel, F., McLaughlin, T., Jarandehei, A., Miller, M., et al. (2011). High-throughput analysis of single hematopoietic stem cell proliferation in microfluidic cell culture arrays. Nat Methods 8, 581-586.

Lee, D.Y., Li, Y.S., Chang, S.F., Zhou, J., Ho, H.M., Chiu, J.J., and Chien, S. (2010). Oscillatory flow-induced proliferation of osteoblast-like cells is mediated by alphavbeta3 and beta1 integrins through synergistic interactions of focal adhesion kinase and Shc with phosphatidylinositol 3-kinase and the Akt/mTOR/p70S6K pathway. The Journal of biological chemistry 285, 30-42.

Lenz, G., Davis, R.E., Ngo, V.N., Lam, L., George, T.C., Wright, G.W., Dave, S.S.,

42 Zhao, H., Xu, W., Rosenwald, A., et al. (2008a). Oncogenic CARD11 mutations in human diffuse large B cell lymphoma. Science 319, 1676-1679.

Lenz, G., Wright, G., Dave, S.S., Xiao, W., Powell, J., Zhao, H., Xu, W., Tan, B., Goldschmidt, N., Iqbal, J., et al. (2008b). Stromal gene signatures in large-B-cell lymphomas. N Engl J Med 359, 2313-2323.

Li, D., Tang, T., Lu, J., and Dai, K. (2009). Effects of flow shear stress and mass transport on the construction of a large-scale tissue-engineered bone in a perfusion bioreactor. Tissue Eng Part A 15, 2773-2783.

McLaughlin, P., Grillo-Lopez, A.J., Link, B.K., Levy, R., Czuczman, M.S., Williams, M.E., Heyman, M.R., Bence-Bruckler, I., White, C.A., Cabanillas, F., et al. (1998). Rituximab chimeric anti-CD20 monoclonal antibody therapy for relapsed indolent lymphoma: half of patients respond to a four-dose treatment program. J Clin Oncol 16, 2825-2833.

Nakao, Y., Kimura, H., Sakai, Y., and Fujii, T. (2011). Bile canaliculi formation by aligning rat primary hepatocytes in a microfluidic device. Biomicrofluidics 5, 22212.

Ngo, V.N., Young, R.M., Schmitz, R., Jhavar, S., Xiao, W., Lim, K.H., Kohlhammer, H., Xu, W., Yang, Y., Zhao, H., et al. (2011). Oncogenically active MYD88 mutations in human lymphoma. Nature 470, 115-119.

Nguyen, T.K., Jordan, N., Friedberg, J., Fisher, R.I., Dent, P., and Grant, S. (2010). Inhibition of MEK/ERK1/2 sensitizes lymphoma cells to sorafenib-induced apoptosis. Leuk Res 34, 379-386.

Poincloux, R., Collin, O., Lizarraga, F., Romao, M., Debray, M., Piel, M., and Chavrier, P. (2011). Contractility of the cell rear drives invasion of breast tumor cells in 3D Matrigel. Proc Natl Acad Sci U S A 108, 1943-1948.

Roschewski, M., Staudt, L.M., and Wilson, W.H. (2014). Diffuse large B-cell lymphoma-treatment approaches in the molecular era. Nat Rev Clin Oncol 11, 12-23.

Ruan, J., Hajjar, K., Rafii, S., and Leonard, J.P. (2009). Angiogenesis and antiangiogenic therapy in non-Hodgkin's lymphoma. Ann Oncol 20, 413-424.

43 Ruan, J., Luo, M., Wang, C., Fan, L., Yang, S.N., Cardenas, M., Geng, H., Leonard, J.P., Melnick, A., Cerchietti, L., et al. (2013). Imatinib disrupts lymphoma angiogenesis by targeting vascular pericytes. Blood 121, 5192-5202.

Scott, D.W., and Gascoyne, R.D. (2014). The tumour microenvironment in B cell lymphomas. Nat Rev Cancer 14, 517-534.

Singh, A., Brito, I., and Lammerding, J. (2018). Beyond Tissue Stiffness and Bioadhesivity: Advanced Biomaterials to Model Tumor Microenvironments and Drug Resistance. Trends Cancer.

Spurgeon, S.E., Coffey, G., Fletcher, L.B., Burke, R., Tyner, J.W., Druker, B.J., Betz, A., DeGuzman, F., Pak, Y., Baker, D., et al. (2013). The selective SYK inhibitor P505- 15 (PRT062607) inhibits B cell signaling and function in vitro and in vivo and augments the activity of fludarabine in chronic lymphocytic leukemia. J Pharmacol Exp Ther 344, 378-387.

Tian, Y.F., Ahn, H., Schneider, R.S., Yang, S.N., Roman-Gonzalez, L., Melnick, A.M., Cerchietti, L., and Singh, A. (2015). Integrin-specific hydrogels as adaptable tumor organoids for malignant B and T cells. Biomaterials 73, 110-119.

Tjin, E.P., Groen, R.W., Vogelzang, I., Derksen, P.W., Klok, M.D., Meijer, H.P., van Eeden, S., Pals, S.T., and Spaargaren, M. (2006). Functional analysis of HGF/MET signaling and aberrant HGF-activator expression in diffuse large B-cell lymphoma. Blood 107, 760-768.

Tzima, E., del Pozo, M.A., Shattil, S.J., Chien, S., and Schwartz, M.A. (2001). Activation of integrins in endothelial cells by fluid shear stress mediates Rho-dependent cytoskeletal alignment. The EMBO journal 20, 4639-4647.

Wilson, W.H., Young, R.M., Schmitz, R., Yang, Y., Pittaluga, S., Wright, G., Lih, C.J., Williams, P.M., Shaffer, A.L., Gerecitano, J., et al. (2015a). Targeting B cell receptor signaling with ibrutinib in diffuse large B cell lymphoma. Nat Med 21, 922-926.

Wilson, W.H., Young, R.M., Schmitz, R., Yang, Y., Pittaluga, S., Wright, G., Lih, C.J., Williams, P.M., Shaffer, A.L., Gerecitano, J., et al. (2015b). Targeting B cell receptor

44 signaling with ibrutinib in diffuse large B cell lymphoma. Nature medicine 21, 922-926.

Wojciechowski, W., Li, H., Marshall, S., Dell'Agnola, C., and Espinoza-Delgado, I. (2005). Enhanced expression of CD20 in human tumor B cells is controlled through ERK-dependent mechanisms. Journal of immunology 174, 7859-7868.

Yang, Y., Shaffer, A.L., 3rd, Emre, N.C., Ceribelli, M., Zhang, M., Wright, G., Xiao, W., Powell, J., Platig, J., Kohlhammer, H., et al. (2012). Exploiting synthetic lethality for the therapy of ABC diffuse large B cell lymphoma. Cancer Cell 21, 723-737.

Yee, C., Biondi, A., Wang, X.H., Iscove, N.N., de Sousa, J., Aarden, L.A., Wong, G.G., Clark, S.C., Messner, H.A., and Minden, M.D. (1989). A possible autocrine role for interleukin-6 in two lymphoma cell lines. Blood 74, 798-804.

Yoon, S.O., Zhang, X., Freedman, A.S., Zahrieh, D., Lossos, I.S., Li, L., and Choi, Y.S. (2010). Down-regulation of CD9 expression and its correlation to tumor progression in B lymphomas. Am J Pathol 177, 377-386.

Yu, W., Qu, H., Hu, G., Zhang, Q., Song, K., Guan, H., Liu, T., and Qin, J. (2014). A microfluidic-based multi-shear device for investigating the effects of low fluid-induced stresses on osteoblasts. PLoS One 9, e89966.

Figure 2.2 Effect of fluid shear stress on proliferation and cytokine secretion in human DLBCLs. (A) Table represents desirable properties in a micro-reactor to mimic biophysical conditions in lymphoma microenvironment. (B) Schematic represents

46 layout of the side flow micro-reactor. Media enters the outer circulation channels through “Media Inlet” port and perfuse into the central cell culture chamber through high-resistance channel (arrows). Cell suspension is added through “Cell Inlet” port into the middle chamber of the micro-reactor. Black arrowheads represent media recirculation between inlet and outlet ports. (C) Side flow micro-reactor with dimensions and relevant design features. (D) Micrographs of the micro-reactor with 20 µm wide media flow channel and 5 µm wide high-resistance channel. The central chamber where cells are cultured is shown to contain equally spaced PDMS pillars. (E) Table summarizing the COMSOL results. (F) Left Panel: A photograph of red food dye loaded side flow micro-reactor. Right Panel: Immunofluorescence image of entire side flow micro-reactor with Calcein (green, live) and Ethidium Homodimer I (red, dead) stained ABC-DLBCL over 48 hr. (G) Table representing cell lines with DLBCL subtypes and BCR mutation. (H) Proliferation of HBL-1 and OCI-LY10 cells under static or flow conditions (1.3 dynes/cm2) maintained over 5 days. Proliferation was quantiﬁed with MTS assay, where an increased absorbance corresponds to increased proliferation (Mean ± S.E.M, n = 3, *p < 0.05, **P<0.01, ***P<0.001 as compared to static conditions, 2-way ANOVA with Bonferroni’s correction). (I) Immunofluorescence images indicating distribution of follicular dendritic cells and ABC-DLBCL line HBL-1 inside a side flow micro-reactor over 48 hr (Scale = 20 µm). (J) Proliferation of HBL-1 and OCI-LY10 cells under static or flow conditions (1.3 dynes/cm2) maintained over 5 days, in the presence of follicular dendritic cells. (Mean ± S.E.M, n = 3, **P<0.01, ***P<0.001 as compared to static conditions, 1-way ANOVA with Tukey’s test). (K) Scatter plots represent ELISA analysis of GM-CSF and TNFα secreted into the culture media by DLBCLs under static and flow conditions (Mean ± S.E.M, n = 3, **P<0.01 compared to static, two-tailed unpaired t-test). Each plot is representative of n =3 independent repeats. See also Figure S1-S4.

Figure 2. Effect of fluid shear stress on drug response in human DLBCLs. (A) Doxorubicin uptake over 36 hr under static versus flow conditions (1.3 dynes/cm2) maintained over 3 days. (Mean ± S.E.M, n = 3, *p < 0.05 as compared to static conditions, 2-way ANOVA with Bonferroni’s correction). Mean fluorescence intensity (MFI). (B-C) Scatter plots represent % apoptosis (B) and MFI (C) in HBL-1 ABC-DLBCLs exposed to doxorubicin. HBL-1 cells were cultured for 72 hr in micro-reactors and exposed to 1 µM doxorubicin for 24 hr. Fluid shear stress was maintained at 1.3 dynes/cm2 for the entire 96 hr. HBL-1 cells were stained with fluorescent-tagged anti-IgM antibody and Annexin V-FITC for apoptosis analysis (Mean ± S.E.M, n = 3, *p < 0.05 compared to corresponding 0 µM doxorubicin group, #p < 0.05 compared to static 1 µM group, 2-way ANOVA with Bonferroni’s correction). (D) Effect of fluid flow on doxorubicin (1 µM) response in

48 OCI-LY10 ABC-DLBCL, cultured in static versus flow condition. Fluid shear stress was maintained at 1.3 dynes/cm2 for the entire 96 hr. Mean ± S.E.M, n=3, *p < 0.05 compared to corresponding 0 µM group, #p < 0.05 compared to static 1 µM group (2- way ANOVA with Bonferroni’s correction). (E) Scatter plots represent MFI in ABC- DLBCLs exposed to Histone deacetylase inhibitor (HDACi) panobinostat. Cells were cultured for 72 hr in micro-reactors and exposed to HDACi for 24 h. Fluid shear stress was maintained at 1.3 dynes/cm2 for the entire 96 hr. Mean ± S.E.M, n=3, *p < 0.05 compared to corresponding 0 µM group, #p < 0.05 compared to static 1 µM group (2-way ANOVA with Bonferroni’s correction). Each plot is representative of n =3 independent repeats.

Figure 3. Effect of fluid flow on integrin and B cell receptor expression in human ABC- and GCB-DLBCLs. (A) Histogram overlays represent ﬂow cytometry analysis of the surface expression level of IgM B cell receptor (BCR, top panel)

50 and integrin α4 (bottom panel) in ABC-DLBCL cell lines OCI-LY3, OCI-LY10 and HBL-1, and GCB-DLBCL cell line OCI-LY1 under static or flow conditions. (B- E) Scatter plots represent flow cytometry analysis of the surface expression level of IgM BCR (B), CD20 (C), integrin α4 (D), and integrin β1 (E) in DLBCLs under static or flow conditions. Fluid shear stress was maintained at 1.3 dynes/cm2 for the entire 4 days. (Mean ± S.E.M, n = 3, *p < 0.05 compared to static group, #p < 0.05 compared to the other cell lines, 2-way ANOVA with Bonferroni’s correction). Mean fluorescence intensity (MFI). (F) Image represents NSG mice implanted with OCI-LY10 xenografts. After 4 weeks of engraftment, tumors were harvested for analysis. Histogram overlays represent flow cytometry analysis of the surface expression level of integrin α4 (bottom panel). Scatter plots (top right) represent flow cytometry analysis of the surface expression level of IgM BCR, integrin α4, and integrin β1 in static versus in vivo HBL-1 xenografts. (Mean ± S.E.M, n = 3, *p < 0.05 **p < 0.005, ***p < 0.001 compared to static group, two-tailed unpaired t-test). (G) Scatter plots represent flow cytometry analysis in static versus in vivo OCI-LY10 xenografts. (Mean ± S.E.M, n = 3, *p < 0.05 **p < 0.005, ***p < 0.001 compared to static group, two-tailed unpaired t-test). (H) Micrograph representing 2D culture of ABC-DLBCLs under flow conditions (1.3 dynes/cm2) with or without integrin β1 blocking. Letter P represents pillars in the micro- reactor. Images are representative of n =3 independent repeats. Scale bar 40 µm. (I) Scatter plot represents flow cytometry analysis of the surface expression level of integrin α5 in OCI-LY3, HBL-1, and OCI-LY10. Fluid shear stress was maintained at 1.3 dynes/cm2 for the entire 4 days. (Mean ± S.E.M, n = 3, ***p < 0.001 compared to static group, 1-way ANOVA with Tukey’s test). Each plot is representative of n =3 independent repeats. See also Figure S5.

Figure 4. Effect of fluid flow on phosphorylation of intracellular signaling proteins. GCB-DLBCL (OCI-LY1) and ABC-DLBCL (OCI-LY3, OCI-LY10, HBL-1) cells were exposed to flow rates inducing 0 or 1.3 dynes/cm2 shear stress for 72 hr. Levels of the phosphoproteins p-LYN (Tyr507), p-SYK (Tyr352), p-ERK 1/2 (Thr185/Tyr187), p- cJUN (Ser73), p-p70 S6K (Thr389/Thr412), and p-p38 (Thr180/Tyr182) were measured by multiplex Luminex assays. Data were normalized to β-tubulin levels (by sample) and then fold-change normalized to static cultures (by cell line). Dots from n = 3 biological replicates with line at mean. * is significant at p < 0.05. Fluid shear stress was maintained at 1.3 dynes/cm2 for the entire 4 days. These experiments were repeated 3 times. See also Figure S6.

Figure 5. Understanding the contributions of fluid shear stress and mass transport in mechanomodulation of DLBCLs. (A-D) Fold change in the surface receptor expression of IgM BCR, integrin, and CD20 using dextran-based viscous media (Mean ± S.E.M, n = 3, *p < 0.05 compared to all groups for (A-C) and *p < 0.05 compared to static for (D), 1-way ANOVA with Tukey’s test). Mean fluorescence intensity (MFI). (E) Understanding the effect of static (viscous) condition on surface receptor expression of IgM BCR and β1 integrin in HBL-1 ABC-DLBCL using a

53 dextran-based viscous media (Mean ± S.E.M, n = 3, *p < 0.05 compared to static, two-tailed unpaired t-test). All results represent 4 days experiments with or without flow. Each plot is representative of n =3 independent repeats. See also Figure S7.

Figure 6. Proposed mechanistic model and effect of inhibiting signaling pathways. (A) Schematic represents the proposed model of fluid shear stress and mass transport-based mechanosignaling of HBL-1 ABC-DLBCL with CD79B mutation. Red arrows represent increased expression level of specified surface marker. Red circles represent nutrient mass transport. (B) Dot plots represent growth percentage normalized to static conditions without any blocking antibodies (Mean ± S.E.M, n = 5, *p < 0.05 compared to all groups, 2-way ANOVA with

54 Bonferroni’s correction). Integrin receptors were blocked with α4β1 antibody over 48 h with recirculating media in micro-reactors. (C-D) Effect of SYK inhibition (R406, 200 nM) on (C) HBL-1 cell growth and (D) integrin β1 over 48 hr. (Mean ± S.E.M, n = 3, *p < 0.05 compared to all groups, 1-way ANOVA with Tukey’s test). (E) Comparative effects of two SYK inhibitors (R406, 200 nM; PRT062607; 5 nM) on cell growth over 48 hr. (Mean ± S.E.M, n = 3, two-tailed unpaired t-test was performed to compare Flow+R406 and Flow+PRT062607). (F-G) Effect of mTOR inhibition on (F) cell growth, and (G) IgM BCR over 48 hr. (Mean ± S.E.M, n = 3, *p < 0.05 compared to all groups, 1-way ANOVA with Tukey’s test). ‘ns’ represent p > 0.05 In all experiments with fluid flow, the fluid shear stress was maintained at 1.3 dynes/cm2 for the entire duration. Each plot is representative of n =3 independent repeats. See also Figure S7.

Figure 7. The CD79B complex cross-talk with fluid flow in mechanomodulation of DLBCLs. (A) Western blot indicating knockdown of CD79B in HBL-1 using a doxycycline inducible shRNA against CD79B (shRNA744) or Luciferase (shRNALuc). Induced shRNA encoded expression of GFP in cells. Tubulin served as control. (B) Scatter plots represents flow cytometry gating and expression of GFP in HBL-1 cells, exposed to doxycycline and fluid flow for 72 hr (C) Dot plots represent total cell percentage and GFP+ cell percentage, each indicating the effect of CD79B knockdown under flow conditions. Mean ± S.E.M, n = 3, *p < 0.05 compared to all groups. (D-E) Dot plots represent integrin β1 and IgM expression in GFP+ cells, indicating the crosstalk of CD79B mutation with these surface proteins under fluid flow conditions. (Mean ± S.E.M, n =

56 3, *p < 0.05 compared to all groups). (F-G) Dot plots represent p-SYK and p-LYN expression in GFP+ cells, suggesting increase in tumor growth under fluid flow conditions is further mediated by downstream p-SYK and p-LYN. (Mean ± S.E.M, n = 3, *p < 0.05 compared to all groups). In all experiments, a 1-way ANOVA with Tukey’s test was used for statistical comparison. In all experiments with fluid flow, the fluid shear stress was maintained at 1.3 dynes/cm2 for the entire duration. These experiments were repeated 3 times. See also Figure S7.

1.1 Supplemental Information

Document S1. Supplemental Experimental Procedures and Figures S1–S7

Supplementary Video 1 Bead distributions in micro-reactors, Related to Figure 1. Flow of green fluorescent beads (radius of 3 µm) that showed bead distributions in devices

Supplementary Video 2, Effect of blocking integrin β1, Related to Figure 3. When integrin β1 was blocked using a monoclonal anti-β1 antibody, HBL-1 DLBCL cells accumulated at the edges and a few outflowing high resistance channels of the device. Supplementary Video 3, Effect of not blocking integrin β1, Related to Figure 3. When integrin β1 was not blocked using a monoclonal anti-β1 antibody, HBL-1 DLBCL cells remained in place.

Supplementary Experimental procedures

Micro-reactor fabrication and operation, and computational simulations

Micro-reactors with high-resistance narrow channels were micro-fabricated with SU8- 2100 (Microchem) resin on silicon wafer using soft-lithography technique (94). Single layer poly (dimethylsiloxiane) (PDMS, sylguard 184, Dow Corning) was casted onto the mold and bonded to a glass substrate using oxygen plasma (PDC-32G plasma

57 cleaner, Harrick Plasma). All micro-reactor fabrication processes were performed at the Cornell NanoScale Science and Technology Facility. The width of the cell culture chamber was 0.6 cm, length was 3 cm, and the height of the chamber was 100 µm and was significantly larger than small lymphoma cells. This height was chosen to prevent any stacks of cells cluster to block the height of the cannel as well as maintain laminar flow (width >> height). To prevent the top of the micro-reactor from collapsing, PDMS posts with diameter 40 µm and separated by 800 µm were incorporated in the micro- reactor design. The width of the high resistance channel was 5 µm and the width of outer channel was 20 µm. The micro-reactor consisted of three ports, two of which were on the circulatory channels for media entrance and exit. In order to achieve the circulatory flow on the micro-reactor, the entry and exit ports were coupled to a 0.5 mm Polytetrafluoroethylene (PTFE) tube by means of a rotary pump (Watson Marlow 2056). The third port was a cell chamber through which cells were introduced. The high resistance channels which connected circulatory channel and middle cell chamber functioned as perfusion channels, which prevented the cells from leaving the central chamber without affecting the exchange of media between circulatory channel and the main cell chamber. The micro-reactor was flushed with ethanol to wet the interior and sterilize the micro-reactor channels prior to use. The computational flow analysis of all micro-reactors without extracellular matrix was done using multi-physics computer software COMSOL, using a 2D solution of the Laplace equation for pressure ( ). This model served as a suitable approximation for the Navier-Stokes solution both for open-channel Hele-Shaw flow (for those experiments performed without matrix; separation of variables, parabolic solution in z, Laplace equation solution in x and y). Computational validation of side-flow lymphoma micro-reactor was performed using SPHERO™ FITC Surface Labeled Particles with 3 µm diameter.

For proof of-concept 3D extracellular matrix studies, Matrigel™-cell suspension was infused in the micro-reactors to form a 3:1 diluted Matrigel™ network in the culture chamber. Briefly, a 75 µL of Growth Factor Reduced Matrigel™ (Corning, #354230) was mixed with 25 µL of HBL-1 or LY10 in media. Micro-reactors were kept on ice to maintain cold temperatures and Matrigel/cell mixture was injected using a 1mL syringe

58 with a 25G needle into the micro-reactor. Prior to starting the flow experiments, the micro-reactor was incubated in a cell culture incubator for 10 min at 37°C to cure Matrigel into a gel. To estimate the flow distribution and the stresses in the 3D matrigel matrix, the transport equations were solved with a hydraulic permeability model (where velocity was assumed independent of z and the xy distribution of velocity was solved with the 2D Laplace equation), and the normal and shear stress in the gel were estimated by use of a Brinkman analysis in the gel. The matrigel porosity (epsilon=0.6) and Darcy’s permeability (k=1.52x10-14 m2), were estimated guided by previously reported work (95, 96). The shear stress experienced by cells encapsulated within the matrigel was estimated using the 3D interstitial flow model reported earlier (97). The hydraulic resistance of the test section and the side channels are effectively in parallel, thus as the hydraulic resistance of the test section increased with the addition of the gel, the fraction of the flow rate that permeates the test section is significantly reduced upon gelation. Therefore in order to achieve the shear stress that cells experienced within the gel-less micro-reactor, the total flow rate through the micro-reactor with Matrigel™ was increased to 1.2 m/sec. Although a Brinkman analysis shows that the forces caused by the gradient of normal stresses exceed the forces caused by the shear stresses, the shear stresses are the putative mechanotransductive mechanisms and were the focus of the calculations. We estimated the shear stress experienced by cells in the absence of Matrigel™ using a surface-area-averaged shear stress D(shear)/4*π*a2, where D is the drag (4*pi*µ*U*a), a is the radius of cell, and U is the velocity that would be present at cell center if the cell were absent. This value is physically meaningful in fluid and gels because it is equal to the surface-area-averaged shear stress on the sphere. For cells near a wall in a Hele-Shaw flow, D_shear/4*pi*a^2 is approximated by U*µ/a, which is the average shear stress (excluding normal stresses) a rigid sphere would experience in an infinite uniform flow of velocity U; although it is less physically meaningful, U*µ/a is also equal to what the local fluid shear stress would be if the cell were absent. For cells in matrigel in our system (where the mean gel pore size is small relative to the cell radius), D_shear/4*π*a2 is approximated by U*µ/sqrt(k), which scales with the shear stresses expected at sub-pore scales but is not related to the macroscopic velocity

59 gradients generated by the gel simulation

Cell culture and drug response

In this study we used five cell lines (a) human tonsil-derived Follicular Dendritic Cells (immortalized cells, HK) (45), (b) HBL1, human ABC-DLBCL cell line with mutation in CD79B along with WT CARD11 (45, 98), (c) OCI-LY10, a human ABC-DLBCL cell line with mutation in CD79A with WT CARD11 (45, 98), OCI-LY3, a human ABC- DLBCL with WT BCR and mutant CARD11, and LY1, a human GCB-DLBCL that is BCR and CARD11 independent. HK and HBL1 were cultured with RPMI 1640 media supplemented with 10% FBS, 1% Glutamax, 10 mM HEPES buffer and 1% penicillin/streptomycin (45). Ly1, OCI-LY3, and OCI-LY10, were cultured with Iscove's Modified Dulbecco's media (IMDM) supplemented with 20% FBS and 1% penicillin/streptomycin, as reported by others (99, 100). All cell lines were checked for mycoplasma. The antibodies used in various studies include anti Human IgM – APC (SA-DA4), anti-Human CD20 – FITC (2H7), anti-Human CD19 – PerCP-Cy5.5 (HIB19), anti-Human CD49d – PE (9F10), anti-Human CD29 – APC (TS2/16) all purchased from eBiosciences. All antibodies were used at all at 1:500 dilutions. FITC- Annexin V was purchased from Biotium. The proliferation study was conducted with cell proliferation assay CellTiter 96® AQueous (promega) (45). In all studies, we loaded 100,000 tumor cells per micro-reactor. In studies that involved FDCs, HK cells were co-seeded at 1:1 ratio, with total cell number being 200,000 per micro-reactor. Static conditions were chosen as standard 2D non-treated tissue culture plates, which are routinely used in lymphoma research. A TCPS platform was chosen instead of the micro-reactor for static conditions because of the lack of availability of fresh nutrients to cells away from the media inlet. The total volume in static TCPS cultures was 1 mL. To match this volume, total volume in a micro-reactor was also 1 mL with 18 µL in the cell culture chamber and remaining in the tubing and outer circulation chamber. The flow experiments were performed for 72 hr unless otherwise noted, drug treatment for 24 hr, and 48 hr for antibody treatment. From a micro-reactor we isolated ~ 100,000 cells. Flow cytometry was done on 5000 cells to accommodate various assays.

Multiplexed phosphoprotein assays

60 Phosphoprotein levels were quantified by multiplexed bead-based Luminex assays, essentially as described previously (87). ABC-DLBCL or GCB-DLBCL cells were exposed to flow rates of 0 (“static”; standard non-treated TCPS) or 3 µL/min (or 1.3 dynes/cm2 shear stress) for 72 hr as described above. Cells were placed on ice and culture medium was removed and centrifuged at 1000g for 4 minutes at 4°C to pellet non-adherent cells. Adherent cells and pelleted non-adherent cells were lysed in a lysis buffer containing Milliplex MAP Lysis Buffer (EMD Millipore, Billerica, MA; #43- 040), Protease Inhibitor Cocktail III (EMD Millipore, #535140, 1:200 dilution) and phenylmethylsulfonyl fluoride (Amresco, Solon, OH, #M145, 2 mM) for 15 min at 4°C. Lysates were clarified by centrifugation at 15,000g for 10 min at 4°C. Clarified lysates were analyzed using a bicinchonicic assay (Thermo Fisher Scientific, Waltham, MA, #23235) to determine the total protein concentration. A micro-reactor without cells was maintained, lysed, and analyzed to calculate the protein contribution independent of the cells and estimate the cellular protein concentration. Milliplex MAPmate assays (EMD Millipore) were used to quantify the following phosphoproteins: p-Akt (Ser473, #46- 677MAG), p-c-Jun (Ser73, #46-622MAG), p-IκBα (Ser32, #46-643MAG), p-ERK1/2 (Thr185/Tyr187, #46-602MAG), p-p70 S6K (Thr389/Thr412, #46-629MAG), p-MEK1 (Ser222, #46-670MAG), and p-p38 (Thr180/Tyr182, #46-610MAG). p-LYN (Y507, #171V50032M) and p-SYK (Y352, # 171V50040M) were purchased from Bio-Rad. Assays were performed in multiplex and using the Milliplex MAPmate Cell Signaling Buffer & Detection Kit (EMD Millipore, #48-602MAG) per the manufacturer’s protocol. Primary (capture) beads and secondary detection antibodies were diluted 1:80 from stock. Each phosphoprotein assay was individually validated using positive control lysates for protein loading concentrations of 2-20 µg/well (data not shown). The optimal protein loading concentration for multiplexing was determined to be 6 µg/well. Milliplex MAPmate multiplexed phosphoprotein assays were conducted per manufacturer’s recommendations on a Bio-Plex 200 instrument (Bio-Rad). Assay technical replicates were averaged for each biological sample. Then, for each phosphoprotein assay, these raw fluorescence data was normalized the fluorescence values for Milliplex MAPmate β-tubulin (#46-713MAG). Two assay technical

61 replicates were performed on lysates from each of n = 3 biological replicates and no replicates were omitted. For each biological sample under flow, normalized fluorescence value was related to the static flow conditions for the same cell line to determine the relative fold-change in normalized phosphoprotein level due to the flow condition. To compare between cell lines for each condition and analyte, a Student’s t- test was performed with α = 0.05.

Effects of fluid shear stress and mass transport using media viscosity

The goal of this study was to understand the contributions of fluid shear stress and mass transport in mechanomodulation of DLBCLs. This was achieved by changing the viscosity of the media while keeping the same shear stress. To keep the fluid shear stress same (1.3 dynes/cm2) and eliminate mass transport, the media was supplemented with 1% Dextran (Sigma-Aldrich, Molecular Weight 2,000,000 Da) to increase the viscosity by 10-fold, while at the same time flow rate was reduced by 10-fold (i.e. 0.13 µL/min). This was done because for a given shear stress the flowrate is inversely proportional to viscosity. In the static case, changing the viscous mobility will also change the molecular diffusion. Therefore, the static to static/viscous comparisons provides information as to the role of diffusion in the observations. Viscosity measurements were taken at 25°C using a Sine-wave Vibro Viscometer SV series (SV-10, A&D, Ltd.). BCR and integrin expression was determined as described earlier using anti-human IgM (SA- DA4), anti-human CD20 (2H7), anti-human CD49d (9F10), and anti-human CD29 (TS2/16) all purchased from eBiosciences. Results were plotted as fold change normalized to static control.

Inhibition of SYK and mTOR

For pathway inhibition studies, all experiments were performed by continuously culturing cells in SYK inhibitors (R406 (catalog# S2194 or PRT062607 (catalog#S8032) from Selleckchem) or mTOR inhibitor (Selleckchem, catalog# S1039) for 48 hr, after which cells were analyzed using flow cytometry as before or supernatant was analyzed for lactate metabolic activity. To determine an appropriate concertation of SYK inhibitors, a dose response curve was generated and 200 nM R406 and 5 nM

62 PRT062607 was selected.

63 Results of the computational modeling of lymphoma micro-reactor: To understand the fluid shear stress pattern experienced by the cells cultured within the cell culture chambers of the previously reported micro-reactors (65, 66) and the newly designed lymphoma micro-reactors, we performed COMSOL-based computational simulations Supplementary Figure S1-S3) using a 2D solutions of the Laplace equation for pressure ( ). This model served as a suitable approximation for the Navier- Stokes solution both for open-channel Hele-Shaw flow (for those experiments performed without matrix; separation of variables, parabolic solution in z, Laplace equation solution in x and y). For the flow in matrix, a hydraulic permeability was used (separation of variables, no z-dependence, Laplace equation solution in x and y) and our estimations of stresses in 3D gels come from a Brinkman analysis in the gel (see section 2.3 for detail). We determined the fluid shear stress distribution pattern in a previously reported microwell micro-reactor suitable for culturing non-adherent hematopoietic cells (66), when exposed to a fluid shear stress of 1.3 dynes/cm2. This system is capable of applying low shear stress, however it had a gradient of shear stress in the XY and YZ planes (Supplementary Figure S1B, S1D, S2A). In the microwell micro-reactor, the variation in shear stress, in the plane of cell culture where most cells are expected to accumulate, was between 0.005 to 1.3 dynes/cm2, accounting for only 66 % of total cell culture area available for shear stress distribution in the range of 0.3 – 1.3 dynes/cm2. When a shear stress condition of 1.3 dynes/cm2 was applied, the pressure difference across the micro-reactor was approximately 2-3 Pa, determined from the COMSOL modeling and was considerably below the physiological pressure difference of 1 mm Hg (or 133 Pa) (64) (Supplementary Figure S1E). The pressure difference across the micro-reactor describes the spatial range of the area-averaged normal component of the surface stress. Although this is not a key parameter in the experiments, we wished to maintain pressure difference inside the micro-reactor to be within the physiological range. The stresses generated by the pressure gradients (both tangent stresses and gradients in normal stress) are the mechanotransductive stimuli driving our study, and we use shear stress to indicate this throughout the study as these tangential and normal stresses are proportional to shear stress. We next considered a parallel plate multi-shear

64 micro-reactor reported elsewhere (65) and capable of inducing low shear stress in one of the four connected chambers (therefore permitting 25% area use with uniform shear stress in the range of 0.3 – 1.3 dynes/cm2 (Supplementary Figure S1E and Supplementary Figure S2B-D)). When COMSOL modeling was performed on these multi-shear micro-reactors, the pressure difference across the micro-reactor was computed to be only 4 Pa (Supplementary Figure S1E), therefore markedly lower than 1 mm Hg (or 133 Pa) (64).

As an alternative, our single-layer side-flow design resulted in uniformly distributed flow patterns and shear stress (0.3 – 1.3 dynes/cm2) with approximately 100% area available for cell culture with uniform shear stress in the tested range (Supplementary Figure S1A, C, E). Importantly, the pressure difference across the micro-reactor was 100 Pa which is very close to 133 Pa (64). These two computational outputs benchmark our micro-reactors against existing microfluidic platforms. In this design, the first five high-resistance channels are a source of media input into the cell culture chamber and next five high-resistance channels function as sinks, so the flow velocity increases from left to right for the first five channels and then decreases for the remaining five channels. By this unique design, the flow is relatively uniform across the chamber and therefore cells experience nearly uniform shear stress in both 2D and 3D cultures. Finally, our cross-flow micro-reactor (as opposed to side-flow above) also resulted in a uniform velocity and shear stress distribution in 66% of the area (Supplementary Figure S2A,B), the design induced a 20 Pa pressure difference across the micro-reactor and therefore was not considered for future studies, but may be relevant for other biomedical applications with adherent cells.

For side-flow lymphoma micro-reactors, the velocity in the cell culture chamber was 0.0005 m/s, which was markedly lower than inlet velocity of 0.085 m/s (Supplementary Figure S1 and S3A-C) and this velocity resulted in a shear stress of ~1.3 dyne/cm2. Quantitative analysis of the shear stress inside the cell chamber demonstrated that the side-flow micro-reactors resulted in six-fold lower average shear stress than the cross-flow micro-reactors (Supplementary Figure S3D and E), while at the same time side-flow micro-reactor maintained physiological pressure difference.

65 The flow rate that resulted in this shear stress was within the mass transport convective regime (101) with Peclet number >1 (Supplementary Figure S3F). We also determined the nutrient (media component) distribution profiles in side-flow micro- reactors using the scalar transport equation and the aforementioned Laplace equation solution and validated these results using green fluorescent beads (radius of 3 µm) that showed bead distributions (Supplementary Figure S3G,H and Supplementary video 1), similar to simulations.

Figure S1. Computational modeling of lymphoma micro-reactors, Related to Figure 1. (A) COMSOL-based computed velocity flow profile, shear stress, and pressure distribution in the cell culture chamber of the side-flow lymphoma micro-reactor and (B) COMSOL-based shear stress profile in the cell culture chamber of a two-layer microwell perfusion micro-reactor. Color scale represents the magnitude of velocity, shear stress, or pressure distribution at any point under steady state conditions. (C-D) COMSOL-based quantitative shear stress analysis in side-flow (C) versus microwell (D) micro-reactors. (E) Table summarizing the

67 COMSOL computed inlet velocity, pressure difference across micro-reactor, and area with uniform shear stress within the 0.3 -1.3 dynes/cm2 range.

Figure S2. Computational modeling of multi-shear micro-reactor, Related to Figure 1.A) Computed velocity flow profile, shear stress, and pressure distribution in the cell culture chamber of a previously reported two-layer microwell perfusion micro-

68 reactor and, B-C) Computed velocity flow profile, shear stress, and pressure distribution in the cell culture chamber of a previously reported multi-shear micro-reactor. Color scale represents the magnitude of velocity, shear stress, or pressure distribution at any point under steady state. D) Shear Stress in X and Y direction in multi-shear micro- reactor.

C D

69 E F

G H

Figure S3. Design and operating mechanism of cross-flow micro-reactors, Related to Figure 1. A) Color scale represents the magnitude of velocity, shear stress, and pressure under steady state conditions. B) Computed shear stress flow profile through the x-axis and y-axis of cross-flow micro-reactors. (C-F) Numerical models of flow distribution and experimental validation, Related to Figure 1. (C) Quantitative flow rate analysis in the side-flow micro-reactor as a function of channel number and distance across the walls of micro-reactor with high resistance channels, (D) Shear stress distribution in cell culture chamber as a function of inlet velocity as determined from computational modeling, (E) Pressure distribution in cell culture chamber as a function of inlet velocity as determined from computational modeling, (F) Péclet number as a function of

70 shear stress in the side-flow micro-reactor. (G-H) Flow patterns in side-flow micro-reactors, Related to Figure 1. G) Media infusion simulation in the side- flow micro-reactor. H) Experimental validation of flow patterns in culture chamber using 3 µm fluorescent beads.

1.6 Flow 1.6 Flow ** ** Static 1.6 Flow 1.6 Static ** Static 1.4 Static 1.4 ** Flow 1.4 1.4 1.2 1.2 1.2 1.2

Absorption 450 Absorption 1.0

Absorption 450 Absorption 1.0 Absorption 450 Absorption 1.0 450 Absorption 1.0 0.8 0.8 0.8 0.8

TNF (LY10) TNF (HBL1) GM-CSF (LY10)GM-CSF (HBL1) IFN (LY10) IFN (HBL1) IL1 (LY10) IL1 (HBL1)

1.6 ** Static 1.6 1.6 1.6 Flow Static Static Static 1.4 Flow Flow Flow 1.4 1.4 1.4

1.2 1.2 1.2 1.2

Absorption 450 Absorption 1.0 Absorption 450 Absorption 450 Absorption 1.0 1.0 450 Absorption 1.0

0.8 0.8 0.8 0.8

IL17 (LY10) IL17 (HBL1) IL6 (LY10) IL6 (HBL1) IL8 (LY10) IL8 (HBL1) IL12(LY10) IL12 (HBL1) Figure S4. Effect of fluid flow on protein secretion in CD79A and CD79B BCR mutated ABC DLBCL, determined using ELISA, Related to Figure 1. Scatter plots represent ELISA analysis of IL8 and IL12 secreted into the culture media by DLBCLs under static and flow conditions (Mean ± S.E.M, n = 3, *p < 0.05 compared to static, 2-way ANOVA with Bonferroni’s correction).

71 A

C B

HBL-1 LY10

Unit Area

40 µm Integrin β1

72 IgM D 2.5 *** Static 2.0 Flow

1.5

1.0

0.5 Fold Change in MFI in Change Fold

(Relative to Static Culture) Static to (Relative 0.0

LY3 LY10

Integrin a4 Integrin b1 2.0 6 *** Static *** Static Flow Flow 1.5 4 ***

1.0 2 *** 0.5 Fold Change in MFI in Change Fold Fold Change in MFI in Change Fold (Relative to Static Culture) Static to (Relative (Relative to Static Culture) Static to (Relative 0.0 0

LY3 LY3 LY10 LY10

Figure S5. Effect of fluid flow on BCR and Integrin expression in GCB-DLBCL and ABC-DLBCL in 2D and 3D, Related to Figure 3. A) Histogram overlays represent flow cytometry analysis of the surface expression level of integrin β1 in ABC-DLBCL cell lines LY3, LY10, and HBL-1, and LY1 GCB-DLBCL under static or flow conditions. B) Histogram overlays represent flow cytometry analysis of the surface expression level of integrin β1 in ABC-DLBCL cell lines LY10, and HBL-1, and LY1 GCB-DLBCL under static conditions. C) Micrograph representing 2D culture of LY10 under flow conditions (1.3 dynes/cm2) with or without integrin β1 blocking. Scale bar 40 µm. D) Effect of fluid flow on DLBCLs in 3D hydrogel. Scatter plots represent flow cytometry analysis of IgM BCR and α4 and β1 in ABC-DLBCL cell lines LY10 and HBL-1 MFI: Mean fluorescent intensity. (Mean ± S.E.M, n = 3, ***p < 0.0001 compared to static group, 2-tailed unpaired t-test).

Figure S6. Effect of fluid shear stress on phosphorylation of intracellular signaling protein pAKT in ABC-DLBCL and GCB-DLBCL, Related to Figure 4. GCB- DLBCL (LY1) and ABC-DLBCL (LY3, LY10, HBL-1) cells were exposed to flow rates inducing 0 or 1.3 dynes/cm2 shear stress for 72 hr. Levels of the phosphoproteins p-AKT (Ser473) was measured by multiplex Luminex assays. Data were normalized to β-tubulin levels (by sample) and then fold-change normalized to static cultures (by cell line). Dots from n = 3 biological replicates from two repeats with line at mean.

74 15

A B 10

centipoise 5

0 0 1 0.5 0.25 Dextran % C D

F E

75 PRT 100

50 % Survival % 25

0 0.0 0.5 1.0 1.5 2.0 2.5 log (nM) log(inhibitor) vs. normalized response Best-fit values LogIC50 1.330 IC50 21.37 Std. Error LogIC50 0.05629

76 Figure S7: Inhibition of signaling pathways and CD79B to mechanistically understand the effect of fluid flow in DLBCL mechanomodulation, Related to Figure 5, 6, and 7. A) Media viscosity measurement, Related to Figure 5. Media viscosity measurement as a function of dextran weight percentage. (Mean ± S.E.M, n = 3, all samples are statistically significant from each other, 1-way ANOVA with Tukey’s posthoc test) B). Standard Curve for SYK inhibitor, Related to Figure 6. Log inhibitor versus survival plot for HBL-1 when exposed to SYK inhibitor PRT062607. (Mean ± S.E.M, n = 3, non-linear regression fit). C) Effect of mTOR on IgM BCR expression, Related to Figure 6. Flow cytometry histogram overlay of IgM expression level in cells exposed to static versus flow versus flow with mTOR inhibitor. Representative of n = 3. D) Fold change in western blot signal, Related to Figure 7. Fold change was calculated by first using the following relationship: Signal(CD79B)*Signal(GFP)/Signal(Actin); followed by normalizing to shLuciferase. E) Micro-reactors with HBL-1 cells with doxycycline induced shRNA that makes them express GFP (green). Representative of n =3. Scale bar 40 µm. F) Dot plots represent pSYK expression in GFP+ cells, suggesting increase in tumor growth under fluid flow conditions is further mediated by downstream pSYK. (Mean ± S.E.M, n = 3, *p < 0.05 compared to all groups). In all experiments, a 1-way ANOVA with Tukey’s posthoc test was used for statistical comparison. In all experiments with fluid flow, the fluid shear stress was maintained at 1.3 dynes/cm2 for the entire duration. These experiments were repeated 3 times.

REFERENCES

Atencia, J., and Beebe, D.J. (2005). Controlled microfluidic interfaces. Nature 437, 648- 655.

77 1206.

Hashwah, H., Schmid, C.A., Kasser, S., Bertram, K., Stelling, A., Manz, M.G., and Muller, A. (2017). Inactivation of CREBBP expands the germinal center B cell compartment, down-regulates MHCII expression and promotes DLBCL growth. Proc Natl Acad Sci U S A.

Jafarnejad, M., Woodruff, M.C., Zawieja, D.C., Carroll, M.C., and Moore, J.E., Jr. (2015). Modeling Lymph Flow and Fluid Exchange with Blood Vessels in Lymph Nodes. Lymphat Res Biol 13, 234-247.

McCarty, W.J., and Johnson, M. (2007). The hydraulic conductivity of Matrigel. Biorheology 44, 303-317.

Ng, C.P., and Pun, S.H. (2008). A perfusable 3D cell-matrix tissue culture chamber for in situ evaluation of nanoparticle vehicle penetration and transport. Biotechnol Bioeng 99, 1490-1501.

Pedersen, J.A., Lichter, S., and Swartz, M.A. (2010). Cells in 3D matrices under interstitial flow: effects of extracellular matrix alignment on cell shear stress and drag forces. Journal of biomechanics 43, 900-905.

Pfeifer, M., Grau, M., Lenze, D., Wenzel, S.S., Wolf, A., Wollert-Wulf, B., Dietze, K., Nogai, H., Storek, B., Madle, H., et al. (2013). PTEN loss defines a PI3K/AKT pathway-dependent germinal center subtype of diffuse large B-cell lymphoma. Proc Natl Acad Sci U S A 110, 12420-12425.

T Clozel, S.Y., R Elstrom, W Tam, P Martin, M Kormaksson, S Banerjee, A Vasanthakumar, B Culjkovic, D Scott, S Brennan, M Leser, R Shaknovich, A Chadburn, F Tabbo, L Godley, R Gascoyne, K Borden, G Inghirami, J Leonard, A Melnick, L Cerchietti (2013). Mechanism-Based Epigenetic Chemosensitization Therapy of

78 Diffuse Large B Cell Lymphoma. Cancer Discovery 3, 1002-1019.

Xia, Y.N., and Whitesides, G.M. (1998). Soft lithography. Angew Chem Int Edit 37, 550-575.

79 CHAPTER 3 Lymph Node Stiffness Mimicking Hydrogels Regulate Human B Cell Lymphoma Growth and Cell Surface Receptor Expression in a Molecular Subtype-Specific Manner

FNU Apoorva1, Ye F. Tian1, Timothy M. Pierpont2, David M. Bassen3, Leandro Cerchietti4, Jonathan T. Butcher3, Robert S. Weiss2, Ankur Singh1

1Sibley School of Mechanical and Aerospace Engineering, College of Engineering, Cornell University, Ithaca, New York, USA,

2Department of Biomedical Sciences, College of Veterinary Medicine, Cornell University, Ithaca, New York, USA,

3Meinig School of Biomedical Engineering, College of Engineering, Cornell University, Ithaca, New York, USA,

4Division of Hematology and Medical Oncology, Weill Cornell Medical College of Cornell University, New York, NY 10065, USA

(J. Biomed. Mater. Res., Part A 105, 1833-1844)

Abstract Diffuse large B cell lymphoma (DLBCL) is the most common type of non-Hodgkin's lymphoma, with multiple molecular subtypes. The activated B cell-like DLBCL subtype accounts for roughly one-third of all the cases and has an inferior prognosis. There is a need to develop better class of therapeutics that could target molecular pathways in resistant DLBCLs; however this requires DLBCLs to be studied in representative tumor microenvironments. The pathogenesis and progression of lymphoma has been mostly studied from the point of view of genetic alterations and intracellular pathway dysregulation. By comparison, the importance of lymphoma microenvironment in which these malignant cells arise and reside has not been studied in as much detail. We have recently elucidated the role of integrin signaling in lymphomas and demonstrated

80 that inhibition of integrin-ligand interactions abrogated the proliferation of malignant cells in vitro and in patient-derived xenograft. Here we demonstrate the role of lymph node tissue stiffness on DLBCL in a B cell molecular subtype specific manner. We engineered tunable bio-artificial hydrogels that mimicked the stiffness of healthy and neoplastic lymph nodes of a transgenic mouse model and primary human lymphoma tumors. Our results demonstrate that molecularly diverse DLBCLs grow differentially in soft and high stiffness microenvironments, which further modulates the integrin and B cell receptor expression level as well as response to therapeutics. We anticipate that our findings will be broadly useful to study lymphoma biology and discover new class of therapeutics that target B cell tumors in physical environments.

Keywords: B cell receptor, organoids, integrin, germinal center, tissue stiffness, biomechanics, extracellular matrix, chemoresistance, contractility

1. Introduction

Despite advances in therapeutics and diagnostics, many mature B cell lymphomas remain incurable. This is partly attributable to the genetic heterogeneity of B cell lymphomas as well as the complex growth and survival signaling provided by the biophysical, biochemical, and cellular components of the lymphoid tumor microenvironment. Diffuse Large B cell lymphoma (DLBCL) is the most common lymphoma, representing ~30 -40% of all B cell Non Hodgkin lymphomas, and gene expression profiling has allowed sub-classification into germinal center B cell (GCB) and activated B cell (ABC) DLBCL subtypes (52-56, 61). The standard combination chemotherapy, CHOP (doxorubicin, vincristine, prednisone, mechlorethamine) was the frontline therapy for >30 years. Ritxumab (R-CHOP) was the only major improvement in spite of expenditure of substantial resources on clinical trials, and still a significant percentage of DLBCL patients are not cured. Activated B cell-like (ABC) DLBCL is the most chemo-resistant DLBCL subtype to R-CHOP with a 5-year overall survival as low as 30-45% versus 80% for germinal center B cell-like (GCB) DLBCL (1, 3, 53). Improved therapies are needed for all DLBCLs but most urgently for ABC-DLBCLs,

81 which are the most chemo-resistant . A myriad of independently predictive biomarkers for resistance have been identified, including ABC and GCB DLBCL expression signature, CD5 positivity, double-hit/MYC-BCL2, stromal signatures, epigenetic silencing of specific genes, and specific somatic mutations (3, 4). But none is sufficient to predict resistance in a given patient and few are helpful in guiding selection of targeted therapies. Therefore, new treatments and treatment-specific biomarkers are needed to improve clinical outcome of both ABC and GCB DLBCL patients. Targeting hallmark ABC-DLBCL pathways, such as those controlled by the transmembrane protein B cell receptor (BCR), have the potential to impact a broad cross-section of ABC-DLBCL patients. However, rational clinical translation requires understanding the factors that modulate expression level of BCR in lymphomas.

The microenvironment in tumors often differs from that in normal tissue, exhibiting altered composition and density of ECM proteins, stromal and immune cells, and growth factors. This altered microenvironment has been implicated as contributing to cancer development and progression for a variety of tumors (102-104). However the role of the lymphoid tumor microenvironment, where lymphomas arise and reside, remains unclear. A major bottleneck in the field is that it remains unclear how combinations of biophysical and/or biomolecular signals that exist in close spatial and temporal order across the lymphoma tumor microenvironment affect ABC-DLBCL and GCB-DLBCL growth and response to therapies. This is attributed to the lack of a lymphoma tumor microenvironment-mimicking engineered tissue model systems, since classic 2D cultures lack the growth and survival signaling provided by the tumor microenvironment. The tumor microenvironment of B cell lymphomas contains highly variable numbers of immune cells, stromal cells, blood vessels and an altered extracellular matrix (5). In the recent years, we and others have shown that the crosstalk between lymphoma cells and the extracellular matrix via integrin molecules is important for their survival and chemo-resistance (62, 105, 106). Specifically, in our effort to understand the role of lymphoma microenvironment, we have recently shown that the crosstalk between the lymphoma microenvironment and integrin αvβ3 is critical for human T cell lymphoma survival, both in vitro and in vivo (62). On the other hand, we

82 discovered that human ABC-DLBCLs have a differential need for integrin α4β1 and αvβ3 for survival (45), in a stromal follicular dendritic cell-dependent manner. We hypothesize that extracellular matrix stiffness is another aspect of the lymph node biophysical microenvironment that can influence lymphoma fate.

For many cancers, increased extracellular matrix stiffness is associated with induction and progression of malignant phenotypes (102, 107). In line with these observations and others, several bioengineered scaffolds have been developed to study the role of mechanical stiffness in tumor survival and progression. A recent study evaluated the role of 3D matrix stiffness on proliferation and drug sensitivity of human myeloid leukemia, which represent liquid tumors originating from bone marrow (108). However, to date, there are no reported studies on the role of tissue stiffness in lymphomas which arise and for the most part, reside in lymphoid tissues such as lymph nodes. Despite the enlargement of neoplastic lymphoid tissue and the palpable stiffness typically used as an initial screen, the role of increased lymph node stiffness in lymphoma remains unclear. The challenge in performing these studies is lack of tunable ex vivo engineered systems that present survival signals to lymphomas and limited understanding of lymphoid tissue stiffness.

Here we quantify healthy and neoplastic lymph node (and spleen) tissue biomechanics using micropipette aspiration. This technique measures local tissue distention due to applied vacuum pressure, analogous to a uniaxial tensile test, as demonstrated by us using finite element analysis (109, 110). We demonstrate that lymphoid tumors have increased tissue stiffness as compared to healthy lymph nodes and spleen. We correlated this data with independent rheology measurements, and engineered hydrogel-based organoids with physiological or pathologically stiff matrix. We here demonstrate design of biomaterials-based hydrogels with tunable mechanical stiffness for growth modulation of B cell lymphomas and discover that lymphomas respond differentially to soft and high stiffness microenvironments in a molecular subtype dependent manner. Engineered hydrogel-based lymphoma organoids using our previously reported immune organoids (111) demonstrated that lymphoid-mimicking stiffness can act as an important biophysical stimulus on lymphoid tumor cells by

83 influencing BCR and integrin expression levels, with consequences on cell survival and drug resistance, dependent on the particular DLBCL molecular subtype. This is the first study to our knowledge that demonstrates the role of lymphoid tissue stiffness on the pro-survival surface receptors and drug response in genetically diverse B cell lymphomas. 2. Materials and methods

2.1 Tg(IghMyc)22Bri (Eµ-myc) mouse model of human lymphoma

Tg(IghMyc)22Bri (Eµ-myc) mice on the C57BL/6J strain background were purchased from the Jackson Laboratory and maintained as hemizygotes. All animals used in this study were handled in accordance with federal and institutional guidelines, under a protocol approved by the Cornell University Institutional Animal Care and Use Committee (IACUC). Mice were monitored regularly and sacrificed upon reaching humane endpoint criteria.

2.2 Human B and T cell lymphoma lines

Lymphoma cell lines that represent a spectrum of DLBCL were used in the current study. These cell lines included: HBL-1 (ABC-DLBCL, IgM BCR dependent), LY-10 (ABC-DLBCL, IgM BCR dependent), LY-3 (ABC-DLBCL, IgM BCR independent), and LY-1 (GCB-DLBCL, IgM BCR independent). Human tonsil derived follicular dendritic cells (FDC) were used as supporting stromal cells for human lymphomas (45). All cells except LY-10 were cultured in RPMI 1640 media supplemented with 10 % fetal bovine serum and 1 % penicillin/streptomycin. Ly-10 cells was cultured in Iscove's Modified Dulbecco's media (IMDM) supplemented with 20 % fetal bovine serum and 1% penicillin/streptomycin. The subsequent passage of these lymphoma culture involved use of 20 % conditioned media (45). De-identified, freshly isolated primary human follicular B cell lymphoma tissue from lymph node of an untreated patient was made available by Dr. Leandro Cerchietti at Weill Cornell Medical College of Cornell University, according to Institutional Review Board guidelines.

2.3 Cell encapsulation, organoid fabrication and rheology

A stock solution of 6 % (weight/volume) gelatin (Sigma Aldrich) in RPMI/IMDM and

84 stock suspension of 4 % Silicate nanoparticles (Southern Clay Products Inc, diameter ~25 nm and thickness ~ 1 nm) was freshly prepared in sterile deionized water. We have previously shown that mixing gelatin and silicate nanoparticles results in the formation of a stable hydrogel network (111, 112). Lymphoma cells (12,000 per 12 µL hydrogel) with and without FDCs (6,000 per 12 µL hydrogel) were added to gelatin stock solution. The FDCs were mitotically inhibited through incubation in cell culture medium containing 0.01 mg/mL Mitomycin C at 37 ºC for 45 min in complete media conditions prior to the encapsulation. The cells were rinsed twice with 10 mL of 1X PBS before usage in the experiments. The two stock solutions were mixed together using the repeated pipetting method, as described earlier by us (111, 112), to achieve a cross- linked 12 µL hydrogel organoids with (a) 2% gelatin and 2% nanoparticle, (b) 2.5% gelatin and 2% nanoparticle, and (c) 3% gelatin and 2% nanoparticle compositions. Organoids were fabricated in non-treated 96-well plates, cured for 5 min before adding growth media, and media was replaced every 3 days. For mechanical characterization using a rheometer, hydrogels were analyzed using Paar Physica MCR 300 rheometer with 50 mm flat-plate geometry under a strain sweep (1–10% strain) condition at a fixed frequency of 0.1 Hz and temperature of 37 °C.

2.4 Micropipette aspiration method to determine soft tissue stiffness

Lymphoid tissue and hydrogel mechanical properties were quantified using our well characterized micropipette aspiration method (109, 110, 113). Isolated murine or human tissues or hydrogels were placed directly in phosphate-buffered saline (PBS) and a drawn glass capillary micropipette (rp≥35 µm) was placed adjacent to the sample surface. Vacuum pressure was incrementally applied via silicone tubing and calibrated by manometer. Previous strain history was mitigated by preconditioning with ~20 cycles of low pressurization (<1 Pa). The lowest pressure that ensured stable contact with the micropipette was taken as zero pressure. Monotonically increasing pressure loads were then applied quasitatically, with images captured at each increment at 150×magnification using a Zeiss Discovery v20 stereo microscope (Spectra Services, Inc.). The aspirated length was measured using calibrated images in NIH ImageJ. An experimental “stretch ratio”, λ=(L+rp)/p was defined by normalizing the aspirated

85 length to the pipette radius as previously described (109, 110). The ΔP versus λ curves were fit using the axial Cauchy stress for a uniaxial load of an incompressible material with an assumed exponential material law, specifically,

------(1)

The fitting parameters α and C were determined by minimizing the sum of the errors squared between Eq.1 and the ΔP versus λ data curves. The C parameter was modified using a scale factor determined from Finite Element simulation, reported earlier by us,

Cmod=γ(α)C (109). Strain energy density was evaluated as a metric for comparing mechanical testing data as previously described (109, 110). Data is presented with representative stress response curves and strain energy density values presented as mean ± SEM with n=3. Statistical comparisons were performed using ANOVA with Tukey post-tests or Student’s t-test, *p<0.05.

2.5 Cellular analyses of lymphomas in organoids

Lymphoma proliferation studies were performed using a CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega), as reported earlier by us (45, 114, 115). Typically, on day 1, 3, and 5, media was removed and organoids were washed once with PBS and CellTiter reagent was added as per manufacturer’s recommendation, and analyzed after 4 hr using a Biotek H1 Hybrid plate reader. For flow cytometry analysis, cells were harvested from organoids by means of enzymatic degradation with a collagenase solution (25 U/mL, Worthington Biosciences) for 6 hr. The degraded organoid suspension was passed through a 70 µm filter (BD Falcon) to separate the cells from any polymeric debris. The harvested cells were washed twice to remove the remaining collagenase and stained with antibodies against B cell marker CD19, BCR IgM, integrins CD29 (β1) and CD49D (α4) at 1:500 dilutions for 1 hr at 4 °C. The antibodies used for flow cytometry studies were Anti Human IgM (APC), Anti Human CD20 (FITC), Anti Human CD19 (FITC), Anti Human CD49D (APC), and Anti Human CD29 (APC), and all were purchased from eBiosciences. The stained cells were washed

86 twice, re-suspended in equal volume of FACS buffer and analyzed using a BD Accuri C6 flow cytometer. FlowJo software was used to analyze data. For the Rho-ROCK pathway inhibition studies, lymphoma organoids were cultured in 1 nM Y-27632 reconstituted lymphoma growth media for 48 hr, followed by proliferation analysis.

2.6 Drug-induced apoptosis studies in organoids

All drug studies were performed by allowing cells to proliferate in organoids with varying stiffness or 2D cultures for 3 days under normal growth conditions followed by addition of the therapeutics. We studied two classes of drugs: a conventional chemotherapeutic vincristine (0.5 µM), which inhibits microtubule polymerization, and Panobistant (50 nM), a Histone Deacetylase Inhibitor (HDACi). After 48 hr of drug exposure, cells were harvested and stained with Annexin V-FITC (Biotium) and CD19 B cell markers. Stained cells were quantified for percent apoptosis using Accuri C6 flow cytometer (BD Biosciences) and analyzed using FlowJo software.

2.7 Statistical analysis

The statistical analysis of variance (Tukey's test for 1-way ANOVA or Bonferroni correction for 2-way ANOVA) was carried using GraphPad Prism software. A p-value of less than 0.05 was considered significant. Two tail t-tests were performed for comparison between two groups. All experiments were reproduced and performed in triplicates unless otherwise mentioned. The quantities are reported in the form of Mean ± S.E.M.

3. RESULTS AND DISCUSSION

3.1 Change in Physiological and Mechanical Properties of Primary Lymphoma Tissues

To characterize changes in the stiffness of lymphoid tissues during tumorigenesis, we first determined the change in the size of lymph nodes and spleen harvested from the Eµ-myc mouse model of human Burkitt’s lymphoma (116). Oncogenesis in these mice is driven by the presence of a transgene containing the c-Myc gene under the control of the immunoglobulin heavy chain enhancer (Eµ) and c-Myc promoter. In these animals,

87 c-MYC is overexpressed in B lymphoid cells, resulting in hyperproliferation of the pre- B cell population prenatally as well as spontaneous formation of lymphoma. Nearly all Eµ-myc mice eventually develop pre-B or B-cell lymphomas, with approximately 50% of animals developing neoplasms by 15-20 weeks of age. As indicated in Figure 1A and B, both inguinal and cervical lymph node showed marked enlargement as compared to a healthy lymph node (~10 fold larger size; 0.3 cm with tumor-bearing mice; p<0.05). Similarly, spleen, another common lymphoid tissue for lymphomagenesis, also indicated a significant increase in tissue size as compared to healthy spleen (Figure 1A, B; p<0.05). We further quantified the fold change in weight percentage of these enlarged tumors and observed ~25 fold increase in lymph node weight and ~8 fold increase in spleen (Figure 1C; p<0.05).

Previously reported mechanical properties of lymph node and spleen are technique and sample preparation dependent (117, 118) and therefore careful measurement of both lymphoid tissue and bioengineered synthetic matrices must be performed. For example, Thomas and colleagues elegantly analyzed lymph nodes from B16 melanoma–bearing mice with respect to tumor stage and observed that lymph node metastasis of melanoma was associated with alterations in lymph node extracellular matrix content, increased intranodal pressures and increased lymph node tissue stiffness and viscoelasticity (117). The technique utilized in this work was indentation methods and yielded values in the range of 100-150 kPa. In contrast, magnetic resonance elastography techniques have reported stiffness of lymphoid tissues in the range of 1.5 – 3 kPa (118). Likewise, ultrasound elastography and indentation tests were used by Yuen et al to determine mechanical properties of porcine cervical lymph nodes however stiffness change in abnormal lymph nodes were not determined using these methods (119). To characterize the mechanical stiffness of lymphoid tissues we used a micropipette aspiration method, reported earlier by us (109, 110). This technique quantifies the elongation of tissues aspirated inside the pipette in response to applied vacuum pressure (Figure 2A). The strain energy density was chosen as a measure of nonlinear tissue stiffness because prior studies have shown this measurement to be less variable for soft tissues than the effective modulus (109, 110). Indeed, attempts to

88 quantify lymphoid tissue stiffness using rheology approaches did not provide meaningful results because the tissue was not fluidic (data not shown). As indicated in Figure 2B,C, the strain energy density for neoplastic mouse lymph node was a significant two-fold higher than the healthy mouse lymph node (p<0.5). We next compared the stiffness of a primary human lymphoma tissue and observed that the stiffness was comparable to neoplastic mouse tissue (p>0.05). No healthy human lymph node was used in the study because of the unavailability of a healthy donor.

3.2 Biomaterials-based engineered lymphoma organoids with controlled stiffness

Bioengineered in vitro models of several cancer types (e.g. melanoma, breast cancer, and lung cancer) have correlated soft or rigid extracellular matrix stiffness with malignant phenotypes (120, 121). These results highlight that stiffness-induced alterations in cell proliferation is dependent on niche and tumor type. Mechanical properties such as substrate rigidity have been found to influence the proliferation of human T cells, with softer substrates resulting in higher naïve T cell proliferation when stimulated ex vivo (122). However, it remains unknown whether the tissue stiffness of lymph nodes has any effect on lymphomas.

Based on the observations of lymphoid tissue strain energy density, we engineered hydrogel-based lymphoma organoids with strain energy densities in the range of 1- 2 J/m3 and co-encapsulating lymphomas with supporting human tonsil derived follicular dendritic cells. These hydrogels were engineered by reinforcing gelatin with silicate nanoparticles, a hydrogel platform previously reported by us to support the survival, growth, and differentiation of primary B cells (111, 112). As indicated in Figure 2D, hydrogels fabricated using 2.5% gelatin and reinforced with 2% nanoparticles showed strain energy density comparable to primary human and mouse lymphoma tissue. Although lymphoid tissues could not be analyzed using a rheometer, hydrogel literature has widely reported storage modulus as a stiffness measurement (111, 123-125). We therefore determined the storage and loss modulus of these same set of organoids and observed a similar trend with stiffness storage modulus (defined as stiffness hereafter) as was observed with strain energy density, except that the

89 measurement yielded storage modulus values in the range of 1500 – 3000 Pa (Figure 2E), which corresponded to 1 – 2 J/m3 strain energy density using the aspiration method (Figure 2C). This range of stiffness is similar to values reported for lymphoid tissue stiffness using magnetic resonance elastography (118). In addition, we have previously reported that gelatin-nanoparticle hydrogels in this stiffness range support viability and spreading of mammalian cells (111, 112). For the remainder of the studies, this storage modulus (G') was used as a reference for hydrogel organoid stiffness and to study the role of tissue stiffness on lymphoma growth and protein expression levels.

3.3 Biomaterials-based engineered lymphoma organoids with controlled stiffness regulate the proliferation of human lymphomas and survival of primary murine lymphomas

We first determined the role of tissue stiffness on human and mouse lymphoma growth. These studies were performed with and without stromal support. For human lymphomas, follicular dendritic cells were used as a stromal support as reported earlier by us (45). We hypothesized that the stiffness of lymphoid tissue will modulate the proliferation of lymphoma cells through mechanical stimulation of surface receptors such as integrins and BCR, and possibly through actin-mediated contractility which has previously been shown to be influenced by extracellular matrix stiffness (126-128). In these studies, we first examined the role of organoid stiffness on the growth of IgM BCR-dependent ABC-DLBCL cell lines (Figure 3A) (98) and IgM-BCR independent ABC-DLBCL line OCI-LY3. In both ABC-DLBCL subtypes, we observed increased growth in all organoids over 5 days of culture and the growth was significantly higher than conventional 2D cultures of lymphoma (Figure 3A, p<0.05). Importantly, ABC- DLBCLs grown in organoids with medium tissue stiffness (2000 Pa) showed the maximum proliferation as compared to lower stiffness (1500 Pa) and higher stiffness (3000 Pa) organoids. Importantly, the stiffness of the organoid with maximum proliferation was comparable to that of human lymphoma tissue and mouse neoplastic lymph nodes, suggesting that tissue stiffness supports lymphoma growth. When GCB- DLBCL cells, which are IgG-BCR dependent and not IgM-BCR dependent, were cultured in the lymphoma organoids, we observed no differences in proliferation rate

90 between 2D and 3D organoid groups (Figure 3B), suggesting GCB-DLBCL are either less dependent or independent of lymphoid tissue stiffness. Stiffness-dependent proliferation of ABC-DLBCL was further confirmed with another human ABC-DLBCL line HBL-1 (Figure 3C), which demonstrated significantly higher proliferation over 4 days of culture in organoids with 2000 Pa stiffness (p<0.05). We observed no differences in survival of ABC or GCB-DLBCLs in 3D organoids as compared to 2D cultures, as indicated by live dead staining in Figure 3D. These studies further highlight an important observation that there exists an optimum tissue stiffness at which lymphomas proliferate, which is distinct from the observations with tumors of non- lymphoid origin such as mammary tumors where increasing stiffness correlates with tumor growth (104). The reasons for these observations are not entirely clear and may be dependent on other phenotypic changes in these cells. Therefore, future ex vivo and in vivo studies are needed to understand the mechanism behind lower proliferation in 3000 Pa organoids.

We next determined the role of another lymphoma tumor microenvironment component, follicular dendritic cells, on the growth of ABC-DLBCL and GCB-DLBCL in organoids with similar mechanical stiffness to patient tissue. In ABC-DLBCL, presence of growth-arrested follicular dendritic cells supported significantly higher lymphoma proliferation as compared to organoids that did not contain follicular dendritic cells (Figure 3E, p<0.05). These differences were not observed in 2D co- cultures. Unlike ABC-DLBCLs, GCB-DLBCLs did not show any significant differences between organoids and 2D cultures, with or without follicular dendritic cells (Figure 3D). These studies further indicate that GCB-DLBCLs are less dependent on lymphoid tumor microenvironment than ABC-DLBCL.

We next determined the role of 2000 Pa hydrogel stiffness on the survival of primary Eµ-myc lymphomas in organoids as compared to 2D co-cultures. Previous studies have shown that bone marrow stromal cells are critical for the survival of primary Eµ-myc lymphomas (129). We screened HS-5 bone marrow stromal cells which sustain hematopoiesis (130-132), our previously reported 40LB stromal cells that present mouse CD40 Ligand and B cell activating factors (111, 112, 133), JAWSII

91 mouse dendritic cells (134, 135), and human follicular dendritic cells HK (45, 62) for sustaining the short term survival of primary Eµ-myc lymphomas. Only HS-5 sustained the survival of primary Eµ-myc lymphomas (data not shown) and therefore we used HS-5 stromal cells in 3D organoids. As indicated in Figure 3F, 2000 Pa hydrogel stiffness resulted in significantly higher survival of Eµ-myc cells as compared to 2D cells (p<0.05). These findings suggest that matching the appropriate soft tissue stiffness is important for prolonging the survival and growth of primary and established lymphomas.

3.4 Stiffness promotes drug resistance against standard chemotherapy as well as novel chemotherapy-free epigenetic drugs

Since most studies to date have focused on drug-lymphoma interactions under 2D conditions, we hypothesized that increased proliferation under the effect of lymphoid tissue stiffness will reduce the apoptosis of ABC-DLBCLs when exposed to anti- lymphoma drugs. In separate experiments, we exposed ABC- and GCB-DLBCLs to either the conventional chemotherapy drug vincristine or the epiegentic Histone Deacetylase inhibitor (HDACi) Panobinostat. HDACi are a new class of epigenetic drugs that have shown promise in early-phase clinical trials for the treatment of hematological malignancies. For example, in a phase II B study (136), 24% of refractory cutaneous T cell lymphoma (CTCL) patients who had originally failed conventional therapies had a complete or partial response to vorinostat, an FDA approved HDACi. Another drug of interest is Panobinostat that has been tested in hematological malignancies such as acute myeloid leukemia (AML) (137) and multiple myeloma (138, 139). Panobinostat has recently entered into phase II clinical trial with other hematological malignancies. We have previously shown that Panobinostat induces less apoptosis when lymphomas are treated in a tumor-specific microenvironment, such as in the presence of pro-survival integrin binding ligands (45); however, not much is known about the drug and its mechanism of action in B cell lymphomas in the context of tissue stiffness.

As indicated in Figure 4A,B, a significant 20-40% lower Annexin-V staining (apoptosis) was observed in HBL-1 cells in 3D organoids of varying stiffness as

92 compared to 2D growth conditions when exposed to either vincristine of Panobinostat. Organoids with tissue stiffness (2000 Pa) comparable to human lymphoma tissue stiffness demonstrated maximum resistance to apoptosis within organoid groups (Resistance: 1500 Pa < 2000 Pa > 3000 Pa). Since HBL-1 is IgM BCR-dependent ABC- DLBCL, we determined the effect of drug on a BCR-independent ABC-DLBCL in order to understand the relationship between drug response and BCR signaling. Unlike HBL-1 ABC-DLBCLs, BCR independent LY3 showed the lowest apoptosis in the softest organoids, although all organoids demonstrated significant lower apoptosis than in 2D growth conditions (Resistance: 1500 Pa > 2000 Pa = 3000 Pa). Similar to ABC- DLBCLs, we observed reduced apoptosis in GCB-DLBCL OCI-LY1 cells grown in 3D organoids; however we did not observe any difference between the soft and stiff organoids (Resistance: 1500 Pa = 2000 Pa = 3000 Pa).

These observations support the proliferation results where ABC-DLBCL proliferation was dependent on organoid stiffness, whereas GCB-DLBCL proliferation was independent of stiffness. We do not anticipate drug diffusion to be responsible for these differences because these organoids have been previously used for growth factor delivery and antibody staining (111). The general reduction in apoptosis observed in 3D microenvironment as compared to classic 2D cultures are consistent with our previous reported work, where similar drug uptake was observed under both 3D and 2D conditions (45). Instead, we hypothesize that reduced apoptosis in the 3D organoid microenvironment can be attributed to differential regulation of pro-survival signals such as integrin receptors and BCR on these cells. To test this hypothesis, we evaluated the expression level of hallmark markers CD19 and BCR on the surface of cells either grown in 2D or in 3D organoids with varying stiffness. As indicated in Figure 5A, IgM- dependent ABC-DLBCL (HBL-1) upregulated the expression level of CD19 in 3D organoids as compared to 2D cultures, and maximum CD19 expression level was observed at 2000 Pa stiffness. Similar observations were made with IgM-independent ABC-DLBCL cell line LY-3. In contrast, while GCB-DLBCL upregulated the expression of CD19 in 3D, no differences were observed among the three hydrogel stiffness values that were studied. Importantly, IgM BCR expression was highly

93 dependent on stiffness in ABC-DLBCL (Figure 5B). In HBL-1 ABC-DLBCLs, 2000 Pa stiff hydrogels demonstrated maximum upregulation of IgM-BCR whereas in BCR independent LY-3, 2000 Pa stiffness significantly reduced IgM expression level between the three organoid stiffness. There was no effect of stiffness in LY-1 GCB- DLBCL, and all 2D cultures had lower expression levels of IgM-BCR.

This observation is important because in ABC-DLBCLs, IgM BCR is chronically activated and has implications in lymphoma survival and progression (52, 60). The BCR is a transmembrane protein complex located on the outer surface of healthy and malignant B cells (60). It is a heterodimer composed of heavy-chain and light-chain immunoglobulins (Igs), CD79A/Igα and CD79B/Igβ. ABC-DLBCLs are commonly associated with mutations of components in the BCR pathway, such as CD79A/B (~20% of ABC-DLBCLs) (56), CARD11 (~10%) (61), and several others. Targeting hallmark pathways of ABC-DLBCL, such as those activated by the BCR, has the potential to impact a broad cross-section of ABC-DLBCL patients (58, 59).

Proposed therapeutic strategies for ABC-DLBCL target proteins that signal downstream of the BCR pathway, including kinase inhibitors targeting Bruton’s tyrosine kinase (BTK), mitogen-activated protein kinases (MAPKs), and AKT (52). These signaling mediators intersect with integrins such as α4β1, and other mechano- transduction pathways (140-142) which can be activated through adhesion to extracellular matrix and tissue stiffness. Therefore, we next determined the expression level of integrin β1 in human lymphomas and primary mouse lymphomas. As indicated in Figure 6, integrin β1 mean fluorescent intensity was significantly higher in 2000 Pa organoids in both IgM-dependent and –independent ABC-DLBCLs (p<0.05). In contrast, in GCB-DLBCL integrin β1 was higher in the softest organoid (1500 Pa) and not in 2000 Pa organoids. Since previous studies have shown that integrin β1 supports survival and growth of ABC-DLBCLs, this current study suggests that the stiffness of lymphoid tissue can upregulate the expression level of mechanoreceptor integrin β1 in ex vivo cultures. We further observed an increase in integrin β1 expression in primary Eu-Myc lymphomas as well, as compared to 2D co-cultures, a finding that supports the need for modular 3D lymphoma tissue constructs with tunable stiffness for supporting

94 primary and established lymphoma cultures. We used the 2000 Pa stiffness to compare differences in expression levels of IgM BCR in primary lymphoma tissues when cultured in 3D versus 2D. We cultured Eu-Myc tumors with HS-5 in 2D co-cultures and 3D organoids, and observed a significant increase in IgM-BCR levels Figure 7, suggesting that 3D organoids modulate the expression levels of surface markers in primary mouse tumors.

Stiffness and deformation are strongly regulated by actomyosin contractility (128). Because our results indicated that the integrin β1 in lymphomas were regulated by matrix stiffness and since previous studies have shown that integrin β1 is linked to the actin cytoskeleton via association with proteins such as talin and vinculin (143-145), we hypothesized that the enhanced growth of lymphomas in 2000 Pa organoids could be attributed to integrin-actin-mediated contractility. We therefore inhibited actin- myosin contractility using the Rho-ROCK inhibitor Y-27632 (140, 146). As indicated in Figure 8, inhibiting contractility did not affect ABC-DLBCLs in 2D cultures, but significantly reduced proliferation in 2000 Pa 3D organoids. This effect was not observed in GCB-DLBCL, which suggests that contractility mediated proliferation is specific to ABC-DLBCL.

Conclusion

Our results are the first evidence that lymphoma survival, proliferation, drug response, and BCR signaling is influenced by lymphoid tissue stiffness in a molecular subtype dependent manner. These results emphasize the role of biomaterials based 3D tissues with tumor matched stiffness and the importance of lymphoid tissue stiffness in malignant B cell tumors. In the past, lymphoid tissue stiffness and mechanosensing has been largely ignored in mechanistic study lymphoma progression and its therapeutic evaluation of B cell lymphomas ex vivo. Therefore, investigation of chemotherapeutics and upcoming new classes of therapeutic agents that target epigenetic, matrix, or angiogenesis pathways should consider the potential confounding interactions between the lymphoid tumor microenvironment and cell-molecular level. Future studies will determine the independent role of tissue bioadhesive signals, stiffness, and stress relaxation in the proliferation and drug response in lymphomas.

95 Acknowledgments

The authors acknowledge financial support from the National Institutes of Health and the National Cancer Institute (1R21CA185236-01 (AS) and R01CA163255-05 (RSW)). The authors thank Dr. Jude Phillip at Weill Cornell Medical College of Cornell University for handling primary human lymphoma tissue, and Yashira Negron Abril for assistance with Eu-Myc mouse genotyping. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies.

REFERENCES

1. Fontan L, Melnick A. Molecular pathways: targeting MALT1 paracaspase activity in lymphoma. Clin Cancer Res. 2013;19(24):6662-8. 2. Fontan L, Yang C, Kabaleeswaran V, Volpon L, Osborne MJ, Beltran E, et al. MALT1 small molecule inhibitors specifically suppress ABC-DLBCL in vitro and in vivo. Cancer Cell. 2012;22(6):812-24. 3. Alizadeh AA, Eisen MB, Davis RE, Ma C, Lossos IS, Rosenwald A, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000;403(6769):503-11. 4. Davis RE, Brown KD, Siebenlist U, Staudt LM. Constitutive nuclear factor kappaB activity is required for survival of activated B cell-like diffuse large B cell lymphoma cells. J Exp Med. 2001;194(12):1861-74. PMCID: PMC2193582. 5. Davis RE, Ngo VN, Lenz G, Tolar P, Young RM, Romesser PB, et al. Chronic active B-cell-receptor signalling in diffuse large B-cell lymphoma. Nature. 2010;463(7277):88-92. PMCID: 2845535. 6. Lenz G, Davis RE, Ngo VN, Lam L, George TC, Wright GW, et al. Oncogenic CARD11 mutations in human diffuse large B cell lymphoma. Science. 2008;319(5870):1676-9. 7. Lenz G, Wright G, Dave SS, Xiao W, Powell J, Zhao H, et al. Stromal gene signatures in large-B-cell lymphomas. N Engl J Med. 2008;359(22):2313-23. 8. Roschewski M, Staudt LM, Wilson WH. Diffuse large B-cell lymphoma- treatment approaches in the molecular era. Nat Rev Clin Oncol. 2014;11(1):12-23. 9. Sehn LH, Gascoyne RD. Diffuse large B-cell lymphoma: optimizing outcome

96 in the context of clinical and biologic heterogeneity. Blood. 2015;125(1):22-32. 10. Chaudhuri O, Koshy ST, Branco da Cunha C, Shin JW, Verbeke CS, Allison KH, et al. Extracellular matrix stiffness and composition jointly regulate the induction of malignant phenotypes in mammary epithelium. Nature materials. 2014. 11. Paszek MJ, DuFort CC, Rossier O, Bainer R, Mouw JK, Godula K, et al. The cancer glycocalyx mechanically primes integrin-mediated growth and survival. Nature. 2014;511(7509):319-25. PMCID: 4487551. 12. Paszek MJ, Zahir N, Johnson KR, Lakins JN, Rozenberg GI, Gefen A, et al. Tensional homeostasis and the malignant phenotype. Cancer Cell. 2005;8(3):241-54. 13. Scott DW, Gascoyne RD. The tumour microenvironment in B cell lymphomas. Nat Rev Cancer. 2014;14(8):517-34. 14. Damiano JS, Dalton WS. Integrin-mediated drug resistance in multiple myeloma. Leukemia & lymphoma. 2000;38(1-2):71-81. 15. Naci D, El Azreq MA, Chetoui N, Lauden L, Sigaux F, Charron D, et al. alpha2beta1 integrin promotes chemoresistance against doxorubicin in cancer cells through extracellular signal-regulated kinase (ERK). J Biol Chem. 2012;287(21):17065-76. PMCID: 3366820. 16. Cayrol F, Diaz Flaque MC, Fernando T, Yang SN, Sterle HA, Bolontrade M, et al. Integrin alphavbeta3 acting as membrane receptor for thyroid hormones mediates angiogenesis in malignant T cells. Blood. 2015;125(5):841-51. 17. Tian YF, Ahn H, Schneider RS, Yang SN, Roman-Gonzalez L, Melnick AM, et al. Integrin-specific hydrogels as adaptable tumor organoids for malignant B and T cells. Biomaterials. 2015;73:110-9. 18. Seo BR, Bhardwaj P, Choi S, Gonzalez J, Andresen Eguiluz RC, Wang K, et al. Obesity-dependent changes in interstitial ECM mechanics promote breast tumorigenesis. Sci Transl Med. 2015;7(301):301ra130. 19. Shin J, Mooney DJ. Extracellular matrix stiffness causes systematic variations in proliferation and chemosensitivity in myeloid leukemias. P Natl Acad Sci USA. 2016. 20. Buskohl PR, Gould RA, Butcher JT. Quantification of embryonic atrioventricular valve biomechanics during morphogenesis. Journal of biomechanics.

97 2012;45(5):895-902. PMCID: 3535469. 21. Buskohl PR, Sun MJ, Thompson RP, Butcher JT. Serotonin potentiates transforming growth factor-beta3 induced biomechanical remodeling in avian embryonic atrioventricular valves. PLoS One. 2012;7(8):e42527. PMCID: 3412853. 22. Purwada A, Jaiswal MK, Ahn H, Nojima T, Kitamura D, Gaharwar AK, et al. Ex vivo engineered immune organoids for controlled germinal center reactions. Biomaterials. 2015;63:24-34. PMCID: 4490011. 23. Purwada A, Singh A. Immuno-engineered Organoids for Regulating the Kinetics of B cell Development and Antibody Production. Nature Protocols (In press). 2017;doi:10.1038/nprot.2016.157. 24. Butcher JT, McQuinn TC, Sedmera D, Turner D, Markwald RR. Transitions in early embryonic atrioventricular valvular function correspond with changes in cushion biomechanics that are predictable by tissue composition. Circ Res. 2007;100(10):1503- 11. 25. Singh A, Qin H, Fernandez I, Wei J, Lin J, Kwak LW, et al. An injectable synthetic immune-priming center mediates efficient T-cell class switching and T-helper 1 response against B cell lymphoma. Journal of controlled release : official journal of the Controlled Release Society. 2011;155(2):184-92. 26. Singh A, Nie H, Ghosn B, Qin H, Kwak LW, Roy K. Efficient modulation of T- cell response by dual-mode, single-carrier delivery of cytokine-targeted siRNA and DNA vaccine to antigen-presenting cells. Mol Ther. 2008;16(12):2011-21. 27. Harris AW, Pinkert CA, Crawford M, Langdon WY, Brinster RL, Adams JM. The E mu-myc transgenic mouse. A model for high-incidence spontaneous lymphoma and leukemia of early B cells. J Exp Med. 1988;167(2):353-71. PMCID: 2188841. 28. Rohner NA, McClain J, Tuell SL, Warner A, Smith B, Yun Y, et al. Lymph node biophysical remodeling is associated with melanoma lymphatic drainage. FASEB J. 2015;29(11):4512-22. PMCID: 4608913. 29. Hirsch S, Guo J, Reiter R, Papazoglou S, Kroencke T, Braun J, et al. MR elastography of the liver and the spleen using a piezoelectric driver, single-shot wave- field acquisition, and multifrequency dual parameter reconstruction. Magn Reson Med.

98 2014;71(1):267-77. 30. Wing-Han Yuen Q, Zheng YP, Huang YP, He JF, Chung-Wai Cheung J, Ying M. In-vitro Strain and Modulus Measurements in Porcine Cervical Lymph Nodes. Open Biomed Eng J. 2011;5:39-46. PMCID: 3103907. 31. Tan Y, Tajik A, Chen J, Jia Q, Chowdhury F, Wang L, et al. Matrix softness regulates plasticity of tumour-repopulating cells via H3K9 demethylation and Sox2 expression. Nat Commun. 2014;5:4619. PMCID: 4133791. 32. Tilghman RW, Cowan CR, Mih JD, Koryakina Y, Gioeli D, Slack-Davis JK, et al. Matrix rigidity regulates cancer cell growth and cellular phenotype. PLoS One. 2010;5(9):e12905. PMCID: 2944843. 33. O'Connor RS, Hao X, Shen K, Bashour K, Akimova T, Hancock WW, et al. Substrate rigidity regulates human T cell activation and proliferation. J Immunol. 2012;189(3):1330-9. PMCID: 3401283. 34. Yan C, Pochan DJ. Rheological properties of peptide-based hydrogels for biomedical and other applications. Chemical Society reviews. 2010;39(9):3528-40. PMCID: 3104857. 35. Hutson CB, Nichol JW, Aubin H, Bae H, Yamanlar S, Al-Haque S, et al. Synthesis and characterization of tunable poly(ethylene glycol): gelatin methacrylate composite hydrogels. Tissue Eng Part A. 2011;17(13-14):1713-23. PMCID: 3118706. 36. Patel RG, Purwada A, Cerchietti L, Inghirami G, Melnick A, Gaharwar AK, et al. Microscale Bioadhesive Hydrogel Arrays for Cell Engineering Applications. Cell Mol Bioeng. 2014;7(3):394-408. 37. Reichl EM, Ren Y, Morphew MK, Delannoy M, Effler JC, Girard KD, et al. Interactions between myosin and actin crosslinkers control cytokinesis contractility dynamics and mechanics. Curr Biol. 2008;18(7):471-80. PMCID: 2361134. 38. Butcher DT, Alliston T, Weaver VM. A tense situation: forcing tumour progression. Nat Rev Cancer. 2009;9(2):108-22. PMCID: 2649117. 39. Swaminathan V, Mythreye K, O'Brien ET, Berchuck A, Blobe GC, Superfine R. Mechanical stiffness grades metastatic potential in patient tumor cells and in cancer cell lines. Cancer Res. 2011;71(15):5075-80. PMCID: 3220953.

99 40. T Clozel SY, R Elstrom, W Tam, P Martin, M Kormaksson, S Banerjee, A Vasanthakumar, B Culjkovic, D Scott, S Brennan, M Leser, R Shaknovich, A Chadburn, F Tabbo, L Godley, R Gascoyne, K Borden, G Inghirami, J Leonard, A Melnick, L Cerchietti. Mechanism-Based Epigenetic Chemosensitization Therapy of Diffuse Large B Cell Lymphoma. Cancer Discovery. 2013;3(9):1002-19. 41. Dierks C, Grbic J, Zirlik K, Beigi R, Englund NP, Guo GR, et al. Essential role of stromally induced hedgehog signaling in B-cell malignancies. Nat Med. 2007;13(8):944-51. 42. Roecklein BA, Torok-Storb B. Functionally distinct human marrow stromal cell lines immortalized by transduction with the human papilloma virus E6/E7 genes. Blood. 1995;85(4):997-1005. 43. Graf L, Iwata M, Torok-Storb B. Gene expression profiling of the functionally distinct human bone marrow stromal cell lines HS-5 and HS-27a. Blood. 2002;100(4):1509-11. 44. Mhyre AJ, Marcondes AM, Spaulding EY, Deeg HJ. Stroma-dependent apoptosis in clonal hematopoietic precursors correlates with expression of PYCARD. Blood. 2009;113(3):649-58. PMCID: 2628371. 45. Nojima T, Haniuda K, Moutai T, Matsudaira M, Mizokawa S, Shiratori I, et al. In-vitro derived germinal centre B cells differentially generate memory B or plasma cells in vivo. Nat Commun. 2011;2:465. 46. Purwada A, Tian YF, Huang WS, Rohrbach KM, Deol S, August A, et al. Self- Assembly Protein Nanogels for Safer Cancer Immunotherapy. Adv Healthc Mater. 2016;5(12):1413-9. 47. Ali OA, Huebsch N, Cao L, Dranoff G, Mooney DJ. Infection-mimicking materials to program dendritic cells in situ. Nat Mater. 2009;8(2):151-8. 48. Duvic M, Vu J. Update on the treatment of cutaneous T-cell lymphoma (CTCL): Focus on vorinostat. Biologics. 2007;1(4):377-92. PMCID: 2721288. 49. Mandawat A, Fiskus W, Buckley KM, Robbins K, Rao R, Balusu R, et al. Pan- histone deacetylase inhibitor panobinostat depletes CXCR4 levels and signaling and exerts synergistic antimyeloid activity in combination with CXCR4 antagonists. Blood.

100 2010;116(24):5306-15. 50. San-Miguel JF, Hungria VT, Yoon SS, Beksac M, Dimopoulos MA, Elghandour A, et al. Panobinostat plus bortezomib and dexamethasone versus placebo plus bortezomib and dexamethasone in patients with relapsed or relapsed and refractory multiple myeloma: a multicentre, randomised, double-blind phase 3 trial. Lancet Oncol. 2014;15(11):1195-206. 51. Richardson PG, Schlossman RL, Alsina M, Weber DM, Coutre SE, Gasparetto C, et al. PANORAMA 2: panobinostat in combination with bortezomib and dexamethasone in patients with relapsed and bortezomib-refractory myeloma. Blood. 2013;122(14):2331-7. 52. Kuppers R. Mechanisms of B-cell lymphoma pathogenesis. Nat Rev Cancer. 2005;5(4):251-62. 53. Brower V. Ibrutinib promising in subtype of DLBCL. Lancet Oncol. 2015;16(9):e428. 54. Wilson WH, Young RM, Schmitz R, Yang Y, Pittaluga S, Wright G, et al. Targeting B cell receptor signaling with ibrutinib in diffuse large B cell lymphoma. Nature medicine. 2015;21(8):922-6. 55. Dumbauld DW, Lee T, Singh A, Scrimgeour J, Gersback CA, Zamir EA, et al. How vinculin regulates force transmission. Proc Natl Acad Sci U S A. 2013:In Press. 56. Singh A, Suri S, Lee T, Chilton JM, Cooke MT, Chen W, et al. Adhesion strength-based, label-free isolation of human pluripotent stem cells. Nat Methods. 2013;10(5):438-44. PMCID: 3641175. 57. Puklin-Faucher E, Sheetz MP. The mechanical integrin cycle. J Cell Sci. 2009;122(Pt 2):179-86. 58. Brakebusch C, Wennerberg K, Krell HW, Weidle UH, Sallmyr A, Johansson S, et al. Beta1 integrin promotes but is not essential for metastasis of ras-myc transformed fibroblasts. Oncogene. 1999;18(26):3852-61. 59. Lahlou H, Muller WJ. beta1-integrins signaling and mammary tumor progression in transgenic mouse models: implications for human breast cancer. Breast cancer research : BCR. 2011;13(6):229. PMCID: 3326542.

101 60. Michael KE, Garcia AJ. Cell adhesion strengthening: measurement and analysis. Methods Cell Biol. 2007;83:329-46. 61. Coyer SR, Singh A, Dumbauld DW, Calderwood DA, Craig SW, Delamarche E, et al. Nanopatterning reveals an ECM area threshold for focal adhesion assembly and force transmission that is regulated by integrin activation and cytoskeleton tension. J Cell Sci. 2012;125(21):5110-23.

Figure Legend

Figure 1. Change in lymph node and spleen size and weight with lymphoma tumorigenesis. A) Photographs representing healthy and neoplastic lymph nodes and spleen from Eµ-myc mice. Animals were sacrificed prior to showing clinical signs (healthy) or upon reaching health endpoints (lymphoma). B) Change in the size of lymph node and spleen isolated from Eµ-myc mice pre-and-post lymphomagenesis. N =5; Mean ± S.E.M; p<0.05, C) Fold change in the weight % of lymph node and spleen isolated from Eµ-myc mice pre-and-post lymphomagenesis. Weight % was initially

102 measured relative to body weight and then normalized to healthy lymph node or spleen, respectively. N =5; Mean ± S.E.M; p<0.05.

Figure 2. Mechanical characterization of healthy versus tumorous lymph nodes and spleen, and hydrogel organoids. A) Photograph showing the mouse tumorous lymph node being aspirate using the micropipette aspiration method. The pipette radius, rp, and the aspirated length, L are indicated. B). Representative pipette test data for a lymph node. C) Strain energy density was calculated from the area under P vs λ curve. Bar graph represents strain energy density of healthy and neoplastic lymph nodes from Eµ-myc mice, and a human lymphoid tumor. D) Storage and loss modulus of hydrogels fabricated using gelatin (denoted G) at the indicated concentration (2, 2.5, or 3%) and silicate nanoparticles (denoted N) at a concentration of 2%. E) Comparison of storage modulus and strain energy density of hydrogel organoid as a function of gelatin weight %. The red oval dotted circle indicates the stiffness range for healthy lymph nodes, blue represents diseased lymph node, and green represents neoplastic splenic tissue stiffness range. Lymph node and spleen were freshly isolated from Eµ-myc mice pre-and-post lymphomagenesis.

103

Figure 3. Lymphoid stiffness regulates proliferation of mature B cell lymphomas. (A-C) Proliferation of ABC-DLBCL lines (LY-10, LY-3, HBL-1) and GCB-DLBCL line LY-1 in hydrogel organoids, co-cultured with follicular dendritic cells (FDCs). Proliferation was quantified with MTS assay, with FDCs pre-treated with mitomycin C (Mean ± S.E.M, n = 3, *p < 0.05). (D) Fluorescence microscopy projections represent live/dead population of ABC- and GCB-DLBCLs encapsulated within 2000 Pa hydrogels. Images are representative of n = 3. Scale bar 10 µm. (E) Proliferation of ABC-DLBCL and GCB-DLBCL in 2000 Pa hydrogel organoids, co-cultured with or without follicular dendritic cells (FDCs). Proliferation was quantified with MTS assay, with FDCs pre-treated with mitomycin C (Mean ± S.E.M, n = 3, *p < 0.05). (F) Proliferation of primary Eu-myc lymphomas in 2000 Pa hydrogel organoids or 2D, co-

104 cultured with HS-5 bone marrow stromal cells. Proliferation was quantified with MTS assay, with HS-5 cells pre-treated with mitomycin C (Mean ± S.E.M, n = 3, *p < 0.05)

Figure 4. Lymphoid tissue stiffness modulates drug resistance in mature B cell lymphomas treated with the conventional chemotherapy drug vincristine or histone deacetylase inhibitor Panobinostat. (A) Bar graphs represent % apoptosis in B cell lymphomas exposed to vincristine. ABC- or GCB-DLBCL cell lines HBL-1, LY- 3, and LY-1 were cultured for 3 days in 2D or 3D organoids and exposed to 1 mM vincristine for 48 hr. B cell lymphomas were stained with APC-IgM BCR and Annexin V-FITC for apoptosis analysis (Mean ± S.E.M, n = 3, *p < 0.05 compared to all 3D groups and #p < 0.05 compared to all other 3D groups). (B) Bar graphs represent % apoptosis in IgM BCR dependent ABC-DLBCL line HBL-1 exposed to HDACi Panobinostat. HBL-1 was cultured for 3 days in 2D cultures or 3D organoids and exposed to 50 nM Panobinostat for 24 h. B cell lymphomas were stained with CD19 and Annexin V-FITC for apoptosis analysis (Mean ± S.E.M, n = 3, *p < 0.05 compared to all 3D groups and #p < 0.05 compared to all other 3D and 2D groups).

105

Figure 5. Effect of lymphoid tissue stiffness on B cell lymphoma markers and B cell receptor. (A-B) Bar graphs represent flow cytometry analysis (mean fluorescence intensity (MFI)) of (A) CD19 and (B) IgM surface markers in ABC- and GCB-DLBCL cells cultured in 2D condition or 3D organoids (all with FDCs). Mean ± S.E.M, n = 3, *p < 0.05, ANOVA. and #p < 0.05 compared to all other 3D groups.

Figure 6. lymphoid tissue stiffness differentially regulates the integrin β1 expression in genetically diverse B cell lymphomas. Bar graphs represent flow cytometry analysis (mean fluorescence intensity (MFI)) of integrin β1 surface markers in ABC- and GCB-DLBCL cells cultured in 2D condition or 3D organoids (all with FDCs). Mean ± S.E.M, n = 3, *p < 0.05, ANOVA and and #p < 0.05 compared to all other 3D groups.

106

Figure 7. Effect of lymphoid tissue stiffness on integrin β1 and IgM BCR expression in primary mouse B cell lymphomas. Bar graphs represent flow cytometry analysis (mean fluorescence intensity (MFI)) of integrin β1 and IgM BCR surface markers in primary Eu-myc lymphomas cultured in 2D condition or 3D organoids (all with HS-5). Mean ± S.E.M, n = 3, *p < 0.05, t-test.

Figure 8. Lymphoid tissue stiffness modulates actomyosin contractility in human B cell lymphomas. Bar graphs represent cell proliferation analysis of ABC-DLBCL or GCB-DLBCL cells cultured in 2D conditions or 3D organoids, in the absence or presence of ROCKi (all with follicular dendritic cells). Mean ± S.E.M, n = 3, *p < 0.05, t-test.

107 CHAPTER 4

Integrins modulate the interaction between endothelial cells and Human B Cell Lymphoma in a Molecular Subtype-Specific Manner

FNU Apoorva1, Lu Ling2, Kristine Lai1, Claudia Fischbach2, Ankur Singh1

1Sibley School of Mechanical and Aerospace Engineering, College of Engineering, Cornell University, Ithaca, New York, USA,

2Meinig School of Biomedical Engineering, College of Engineering, Cornell University, Ithaca, New York, USA,

Abstract

Diffuse large B cell lymphoma (DLBCL) is a heterogeneous group of lymphoma cancer that accounts for 30% of non-Hodgkin's lymphoma. The activated B cell (ABC) subtype of DLBCL, which constitutes one-third of the DLBCL, is the most refractory form of lymphoma associated with poor clinical outcome (52-56, 61). It is increasingly becoming evident that tumor microenvironment actively participates in the initiation and progression of lymphoma, by promoting vascular network formation by endothelial cells. Recent studies have shown the significance of tumor associated endothelial cells in pathogenesis and treatment of lymphoma. Unfortunately antiangiogenic lymphoma therapies such as bevacizumab have reaped poor outcome because of lack of mechanistic insight into lymphoma microenvironment and associated angiogenesis. We recently elucidated the role of integrins in transduction of B cell Receptor signaling, and how these signaling implicate mechanical stiffness and lymphatic grade shear stress in lymphoma. Here we demonstrate the role of integrins in angiogenesis in DLBCLs. We present an engineered hydrogel based DLBCL organoid where one can modulate the integrin signature of tumor microenvironment. Such an organoid model is apt to study the cell-cell interactions within DLBCL microenvironment, and could potentially develop into an DLBCL vascularization platform to study the angiogenic factors in a neoplastic lymph nodes. The current study focuses on DLBCL-endothelial interaction

108 and demonstrate that both DLBCLs and endothelial cells employ two sets of integrins 4 1 and v 3, and surface receptor VEGFR to establish dysregulated VEGF based autocrine and paracrine signaling pathways that promotes the proliferation of endothelial cells and lymphoma progression in a synergistic manner. Taken together, these data indicate that integrins may present a novel antiangiogenic target for lymphoma therapy. Keywords: B cell receptor, organoids, integrin, angiogenesis, VEGF, extracellular matrix, integrin inhibitors

1. Introduction

DLBCL is a pathophysiologically heterogeneous group of B cell lymphoproliferative disease that accounts for 30% of non-Hodgkin Lymphoma. Despite the advancement in understanding of molecular pathogenesis and phenotypic heterogeneity of DLBCLs, the standard care for any type of DLBCL is R-CHOP. Activated B Cell (ABC) lymphoma is an aggressive form of DLBCL for which patients 5-year survival is as low as 30-45%, and it has not improved much in past 20 years (1, 3, 53). Considering the relatively high occurrence rate of ABC-DLBCLs and subsequent high relapse rate, it needs an immediate attention.

Recent studies show that development of DLBCL neoplasm accompanied by angiogenesis is a prognostic signature of a less favorable clinical outcome of RCHOP. It is evident from recent investigations that tumor microenvironment actively participates in the initiation and progression of lymphoma. The microenvironment of tumor constitutes of extracellular matrix and a large variety of infiltrating non-malignant cells such as stromal cells, immune cells and endothelial cells. Angiogenesis is the formation of new blood vessels by endothelial cells that supplies neoplasm with nutrients and oxygen. Angiogenesis has been extensively studied in the context of solid tumor, and several novel anti-angiogenic therapy have been discovered which suppress the growth of established tumors by targeting tumor vasculature. Recent studies underscore the significance of angiogenesis in pathogenesis and therapy of lymphoma. Currently an anti-VEGF monoclonal antibody bevacizumab is in phase-II clinical trial,

109 which heretofore has shown only limited efficacy in NHL.

The tumor microenvironment modulates angiogenesis through a set of stimulatory and inhibitory factors, and endothelial cell surface receptors. Some of the extensively studied angiogenic factors are Vascular Endothelial Growth Factors (VEGF) such as VEGF-A, VEGF-B, VEGF-C, VEGF-D) and Placenta Growth Factors (PlGF). These growth factors transduce angiogenic signals to endothelial cells by binding with their receptor tyrosine kinases, namely VEGFR-1 and VEGFR-2. From several past in-vitro and in-vivo studies of solid neoplasm it is evident that integrins of endothelial cells regulate cell growth, survival and migration during angiogenesis. Avraamides et.al presented a comprehensive list of integrins and their respective roles in pathological angiogenesis, and emphasized on role of heterodimers 4 1 and v 3. These integrins when in ligation state they promote cell growth, homing and assembly of endothelial cells, but their unligated state or their antagonization with an integrin antagonist induces apoptosis of endothelial cells and subsequently supression of tumor via p53 enhancement. FischBach et.al presented a bio-artificial culture platform to elucidated the significance of 3D integrin engagement in agiogenesis. It is evident from these findings that 3D tumor microenvironment cue and 3D integrin egagement regulates secretion of VEGF and IL-8 by camcer cells, which subsequently modulates agiogenesis and cancer progression.

Eunice et.al demonstrated that lymphoma cells secrete VEGF and express VEGFR-1 and VEGFR-2, and presented a potential VEGF targeting anti-angiogenic therapy for lymphoma. Ruan et.al showed that Platlet-derived growth factor-type BB (PDGF-BB) is also implicated in vascular maturation and stabilization, and demonstrated the role of PDGFRβ inhibitor imatinib mesylate in targeting angiogenesis and attenuating lymphoma. Recently, Tian et.al investigated the role of crosstalk between lymphoma cells and the extracellular matrix via integrin molecules in lymphoma growth and chemo-resistance (62, 105, 106). Specifically, Tian et.al demonstrated that the crosstalk between the lymphoma microenvironment and integrin α4β1 and αvβ3 is critical for human DLBCLs survival (45).

Despite significant evidence for implication of integrins in agiogenesis in solid

110 neoplasm, there is no reported study to this date on the role of 3D integrin engagement in angiogenesis in lymphoma. We hypothesize that 3D integrin engagement regulates endothelial and lymphoma growth by modulating the secretion of VEGF and surface expression of VEGFR by both lymphomas and endothelial cells. We present a hydrogel based bioartifical platform with tunable integrin ligand presentation capability to study angiogenesis in lymphoma. We also demonstrate that 3D integrin engagement transduces the cross talk between endothelial cells and lymphoma cells via VEGF and surface expression of VEGFR in ABC-DLBCLs. This cross talk differentially promotes the growth of lymphoma cells and endothelial cells in a molecular subtype dependent manner, which we believe could potentially impact vascularization and disease progression in DLBCL patients. We anticipate that these findings would lead to better understanding of angiogenesis in DLBCL, and help in development of integrin antagonist based novel anti-angiogenic lymphoma therapy.

Materials and methods

1.1 Human B cell lymphoma and endothelial lines

We used a set of five lymphoma cell lines that would capture the molecular diversity of DLBCLs. These cell lines are: HBL-1 (ABC-DLBCL, IgM BCR dependent), LY-10 (ABC-DLBCL, IgM BCR dependent), LY-7 (GCB-DLBCL, IgM BCR independent), LY-3 (ABC-DLBCL, IgM BCR independent), and LY-1 (GCB-DLBCL, IgM BCR independent). human umbilical vein endothelial cells (HUVEC, Lonza) were used as vascular cells for human lymphomas (45). All lymphoma cells except LY-10 were cultured in RPMI 1640 media supplemented with 10 % fetal bovine serum and 1 % penicillin/streptomycin. Ly-10 cells was cultured in Iscove's Modified Dulbecco's media (IMDM) supplemented with 20 % fetal bovine serum and 1% penicillin/streptomycin. The subsequent passage of these lymphoma culture involved use of 20 % conditioned media (45). The HUVEC cells at low passage (p < 6) were cultured in Bio-Whittaker medium 199 (M199) supplemented with endothelial cell growth supplement (ECGS, Millipore, Billerica, MA), 20% FBS, 1% P/S, 2 mM Glutamax, and 5 U/mL heparin.

111 1.2 Cell encapsulation, organoid fabrication and vascularization

The organoids encapsulating lymphoma cells with 7.5% (w/v) PEGMAL macromer concentration were fabricated using PEGMAL (Laysan bio, Inc., > 90% purity) and thiolated cross-linkers. PEGMAL macromers were initially functionalized with thiolated bioadhesive peptides RGD or REDV with 4:1, 4:0.75, 4:0.5, 4:0.25, and 4:0 MAL-to-peptide molar ratio. MMP-degradable peptide VPM and nondegradable dithiothreitol (DTT) thiolated cross-linkers were combined at a 50:50 molar ratio, and added to macromere for crosslinking in 4:1.5 macromer-to-cross-linker molar ratio. All components were prepared using PBS++ solution with pH 7.4 and 1% HEPES. Lymphoma cells were suspended in the PEGMAL macromer solution prior to cell encapsulation. Each organoid was formed in the middle of a well of a nontreated 96 well plate by placing 5 µL of functionalized PEGMAL macromere and injecting 5 µL of cell-containing cross-linker solution macromer droplet, flowed by quick mixing using a pipette. The resulting hydrogel organoids were set by curing them for 15 min at 37° C inside cell culture incubator. Fresh 50:50 solution of M199 and RPMI 1640 (or IMDM for LY10) and supplemented with 10% FBS (20% for LY10), and 1% P/S was then added to the organoid culture.

1.3 Cellular analyses of lymphomas in organoids

Lymphoma proliferation studies were performed using a CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega) (45, 114, 115). For this study media was removed from TCP wells, and organoids were washed with PBS. Following that CellTiter reagent was added as per manufacturer’s protocol, and resulting supernatant was analyzed after 4 hr using a Biotek H1 Hybrid plate reader. For flow cytometry analysis, organoids were degraded by enzymatic degradation for 6 hrs using collagenase solution (25 U/mL, Worthington Biosciences), and cells were harvested using a 70 µm cell strainer (BD Falcon). The harvested cells were washed with PBS and stained with antibodies against B cell marker CD19, BCR IgM, integrins CD29 (β1) and CD49D (α4) at 1:500 dilutions for 1 hr at 4 °C. The antibodies used for flow cytometry studies were Anti Human IgM (APC), Anti Human CD20 (PE CY7), Anti Human CD49D (FITC), and Anti Human CD29 (PE), Anti Human CD51/61 (FITC), Anti Human CD31

112 (FITC) and all were purchased from eBiosciences. The stained cells were re-suspended in FACS buffer and analyzed using a BD Accuri C6 flow cytometer. FlowJo software was used to analyze data.

1.4 Integrin-inhibition studies in organoids

The integrin inhibition studies on lymphoma cells were performed by allowing cells to proliferate in organoids for 3 days under normal growth conditions followed by addition of the integrin inhibitors. We studied two classes of integrin inhibitors: (a) cilegitide (αvβ3 inhibitor) and (b) BIO1211 (α4β1 inhibitor). After 48 hr of drug exposure, cells were harvested and stained with for flowcytometer analysis of biomarkers, and supernatant was used to run ELISA. In order to study the effect of integrin inhibition on angiogenesis morphology, co-culture of HUVEC and lymphomas were treated with two integrin inhibitors right at the inception of culture, and fluorescent microscopy was done at an intermittent period of 24 hr for 3 days.

1.5 Statistical analysis

We carried statistical analysis of variance by Tukey's test (1-way ANOVA) or Bonferroni correction (2-way ANOVA) method using GraphPad Prism software. The groups which have p-value of less than 0.05 were considered significantly different. All experiments were carried in triplicates unless otherwise mentioned. The quantities reported in this manuscript are in the form of Mean ± S.E.M.

2. RESULTS AND DISCUSSION

2.1 The presence of lymphoma cells upregulates the proliferation of endothelial cells

In order to understand the role of lymphoma tumor in angiogenesis, we cultured HUVEC cells in supernatant obtained from various lymphoma culture. The endothelial cells manifested rapid proliferation in lymphoma supernatant in a lymphoma subtype dependent manner. As indicated in Figure 1A and B, culture of endothelial cells in IM199 media conditioned with supernatant from lymphomas, showed remarkably higher proliferation as compared to no-supernatant counterpart, and this could be

113 observed as soon as on third day of starting the culture. While ABC-DLBCLs proliferated to ~4 fold p<0.05 compared to their no supernatant counterpart, GCBs proliferated at a relative slower rate, to ~2 fold p<0.05 in comparison to culture without supernatant. Similarly, ABC lymphomas, in the presence of endothelial cells in 2D culture, proliferate at ~2X rate compared to stand alone culture of lymphomas, as shown in (Figure 1C, D; p<0.05). Cocco et.al reported an investigation on DLBCLs and follicular lymphoma in which conditioned media lymphomas stimulated strong angiogenic response in a CAM assay. Brandvold et.al demonstrated that MYC over expression in chicken lymphocytes generates an angiogenic phenotype both in-vivo as well as in-vitro. Brandvold et.al also reported that conditioned media from the culture of MYC transformed lymphocytes of chicken strongly simulated the proliferation of endothelial cells as high as 5 fold compared to a culture without conditioned media. (Ribatti et.al and Brandvold et.al). In yet another reported experiment by Bao et.al, it was observed that conditioned media from stem-cell like glioma cells significantly increased migration and growth of endothelial cells. Many previously reported studies implicated VEGF secretion by cancer cells as the prime pro-angiogenic factor. Gratzinger et.al conducted immunohistochemistry on DLBCL patient tumor and found that 60% of these samples stained positive for VEGF. Kuramoto et.al and Ship et.al reported that often DLBCLs are associated with overexpressed VEGF gene expression. This motivated us to characterize the VEGF secretion by lymphoma cells. In order to capture the entire spectrum of DLBCL, we measures VEGF level in the culture media of various lymphoma cell lines. Figure 1D presents the VEGF release profile of different cell lines, and as shown, lymphomas secrete VEGF in varying amount depending on the molecular subtype of DLBCLs.

2.2 Synthetic hydrogel-based engineered lymphoma organoids with tunable integrin ligand presentation

It has been previously reported that 3D culture and integrin engagement has profound effect on angiogenesis in various types of cancer. But to this date role of 3D culture and integrin in lymphoma has not been extensively studied. We hypothesized that integrin specificities and integrin density would determine the extent of induction of pro-survival

114 factors. In the spirit of understanding the effect of 3D culture and integrin engagement on DLBCLs and endothelial cells we synthesized an engineered hydrogel organoids of PEGMAL with varying density of RGD and REDV motives. We observed that 3D integrin liganded organoids differentially stimulated the proliferation of DLBCLs in a molecular subtype and integrin density dependent way, as shown in Figure 2A. HBL- 1, which is an ABC-DLCBCL cell line, responded strongly to RGD. The 3D organoids of HBL-1 grafted with RGD always facilitated enhanced proliferation, and the extent of proliferation was correlated to the RGD ligand density; organoids with 4:1 macromer to RGD ligand showed as high as ~2 fold more proliferation than organoids with no ligand. The organoids with REDV motif showed higher proliferation than their 2D counterpart, but the effect of density of REDV motif on cell growth was insignificant. As shown in Figure 2A, a similar observation was made for LY-10. LY-3 which is a BCR independent ABC-DLBCL cell line, while 3D organoids proliferated more than its 2D counterpart, increasing integrin ligand density did not have a stimulating affect. Similarly, for GCB cell lines LY-1 and LY-7, the 3D organoids with RGD ligand fared better in terms of proliferation, but we did not observe any significance of increasing ligand density. Fischbach et.al investigated a synthetic ECM of alginate hydrogel that contained RGD motives at density matching in-vivo tumor condition to study angiogenesis. Interestingly, Fishbach observed that 3D culture and matching integrin density promoted cancer cells proliferation, and secreted progiogenic factors such as VEGF, IL-10 and bFGF at a markedly higher concentration than in 2D culture. DLBCLs and immune cells in general secrete VEGF. We observed that 3D culture of DLBCLs were characterized by higher VEFG level in the culture when compared to their 2D counterparts. We hypothesized that enhanced proliferation of VEGF promoted the growth of endothelial cells, which subsequently upregulated the growth of DLBCLs.

2.3 3D integrin engagement modulates release of growth factor by lymphomas and development of vascular network

We next determined the role of endothelial cells on DLBCLs. We cocultured endothelial cells and DLBCLs together, and found that presence of endothelial cells, promoted proliferation of both ABC and GCB type of DLBCLs. The observation that HUVECs

115 proliferated upto ~5 fold compared to fresh media, motivated us to hypothesize that lymphomas released growth factor VEGF which promoted proliferation of HUVECs. We also observed 3D organoids enhanced the proliferated of endothelial cells even more, as shown Figure 3C. We hypothesized next that 3D integrin engagement modulates VEGF secretion by lymphomas. In order to test our hypothesis we investigated VEGF release by a spectrum of DLBCL organoids. The results indicate that integrin specificity and density of integrin ligand presentation correlates to the amount of VEGF released by the orgnoids. Both GCB and ABC- DLBCL organoids secreted more VEGF, as we increased the density of RGD ligand in the organoids. In LY-1 (GCB-DLCBCL) we did not observe any significant difference between 2D and 3D VEGF release, unless density of RGD was at least 75% of the PEG macromer. In case of REDV presenting organoids, we noticed a similar trend, VEGF released from LY-1 organoids matched that of 2D culture unless REDV density was above 75% of the macromer. The 3D ABC-organods released more VEGF that 2D, but in a composition independent manner.

2.4 Endothelial cells engage v 3 integrins to adhere to the ECM, and 4 1 to adhere to DLBCLs and to form clusters

We next investigated if blocking of integrins on lymphomas had any effect on their VEGF secretion. We first conducted integrin blocking study in 2D cultures. We observed that inhibition of αvβ3 using cilengitide affected VEGF secretion only in ABC-DLBCLs with chronically activated BCR. The VEGF release in GCB-DLBCL (LY-1 and LY-7) and BCR independent ABC-DLBCL (LY-3). Interestingly, inhibition of α4β1 by BIO1211 (α4β1 inhibitor) also depreciated VEGF release for all types of DLBCL except LY-1 (GCB) and LY-3 (BRC independent ABC).

The coculture of HUVECs and DLBCLs were characterized by cluster formation of lymphomas on above HUVECs. Interestingly, these clusters of DLBCLs were resilient enough that they could not be dismantled even after multiple washing by PBS. This motivated us to ask if integrins played any role in physical interaction between HUVECs and DLBCLs. We treated HUVECs and DLBCLs with cilengitide ( v 3 inhibitor) for 30 mins, and cocultured them over TCP. We observed that HUVECs treated with

116 cilengitide, as low as 1 µM, did not adhere to TCP and lost their spread morphology, and subsequently died over next 24 hrs. The DLBCLs treated with cilengitide did not undergo cluster formation, and ~80% cells dies within 24 hrs of starting the culture. In order to elucidate the role of 4 1 we treated HUVECs and DLBCLs with BIO1211 (α4β1 inhibitor), and coculutred them over TCP. We observed that BIO1211 treated HUVECs could adhere and spread over TCP. In contrast DLBCLs did no more adhere to BIO1211 treated HUVECs and no cluster formation was observed. Vice-versa, treatment of DLBCLs with BIO1211 resulted in loss of cluster formation by DLBCLs above HUVECs. These observations concluded that different set of integrins are involved HUVEC-ECM and HUVEC-DLBCL interaction. While HUVECs engage αVβ3 integrins with ECM in order to adhere and spread, they engage α4β1 integrins to form the physical association with DLBCLs.

2.5 Integrins v 3 and 4 1 modulate the migration of endothelial cells and DLBCLs

We next studied migration of lymphomas and endothelial cells. We hypothesized that migration of endothelial cells is modulated by integrins presentation by ECM to lymphomas. We employed a microfluidic conduit system with three ports, first housed 2D control lymphomas, second contained HUVECs, and third port cultured 3D ligand specific organoids. We measured the relative migration of HUVEC cells from HUVEC port to other two, and the migration of endothelial and lymphoma cells, as shown in. We observed that organoids presenting RGD and REDV ligand to lymphomas promoted migration of HUVEC cells. While we observed a significant difference between migration of endothelial cells as a result of 2D and 3D culture of lymphomas, we did not observe a significant correlation between ligand density and migration though. We also observed migration of lymphomas towards endothelial cells, that resulted in cluster formation of lymphomas away from three respective ports. This observation is important, since it demonstrates that HUVECs promotes cluster formation of DLBCLs in a three-step process; firstly endothelial cells and lymphomas both release factors to induce migration of lymphomas and endothelial cells, secondly they engage integrin 4 1 in order to form a physical adhesion between lymphomas and endothelial cells, and

117 thirdly release of growth factors and integrin transduced pro-survival signal promote the proliferation of DLBCLs.

3. CONCLUSION

Despite past studies implicating integrins in angiogenesis in many different cancer, integrins role in lymphoma angiogenesis has been largely ignored. This report presents a 3D biosynthetic synthetic model of DLBCL to study the cell-cell interaction between lymphomas and endothelial cells. The scaffold has a facile integrin ligand presentation capability that allows to probe role of integrin specificity and its 3D ligand presentation on lymphomas and endothelial cells. Taken together our results suggest that 3D integrin engagement differentially modulates the growth and migration of endothelial cells in lymphoma in a molecular subtype dependent manner, by differentially regulating VEGF and VEGFR. We anticipate that this 3D biosynthetic model of DLBCL could potentially evolve into an DLBCL angiogenesis platform and enable discovery of novel antiangiogenic lymphoma therapy based on integrin antagonists.

Acknowledgments

The authors acknowledge financial support from the National Institutes of Health and the National Cancer Institute (1R21CA185236-01 (AS) and R01CA163255-05 (RSW)). The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies.

REFERENCES

118 3. Lenz G, Wright G, Dave SS, Xiao W, Powell J, Zhao H, Xu W, Tan B, Goldschmidt N, Iqbal J, Vose J, Bast M, Fu K, Weisenburger DD, Greiner TC, Armitage JO, Kyle A, May L, Gascoyne RD, Connors JM, Troen G, Holte H, Kvaloy S, Dierickx D, Verhoef G, Delabie J, Smeland EB, Jares P, Martinez A, Lopez-Guillermo A, Montserrat E, Campo E, Braziel RM, Miller TP, Rimsza LM, Cook JR, Pohlman B, Sweetenham J, Tubbs RR, Fisher RI, Hartmann E, Rosenwald A, Ott G, Muller- Hermelink HK, Wrench D, Lister TA, Jaffe ES, Wilson WH, Chan WC, Staudt LM, Lymphoma/Leukemia Molecular Profiling P. Stromal gene signatures in large-B-cell lymphomas. N Engl J Med. 2008;359(22):2313-23. doi: 10.1056/NEJMoa0802885. PubMed PMID: 19038878. 4. Sehn LH, Gascoyne RD. Diffuse large B-cell lymphoma: optimizing outcome in the context of clinical and biologic heterogeneity. Blood. 2015;125(1):22-32. doi: 10.1182/blood-2014-05-577189. PubMed PMID: WOS:000350809500009. 5. Scott DW, Gascoyne RD. The tumour microenvironment in B cell lymphomas. Nat Rev Cancer. 2014;14(8):517-34. Epub 2014/07/11. doi: 10.1038/nrc3774. PubMed PMID: 25008267. 6. Burger JA, Wiestner A. Targeting B cell receptor signalling in cancer: preclinical and clinical advances. Nat Rev Cancer. 2018;18(3):148-67. Epub 2018/01/20. doi: 10.1038/nrc.2017.121 nrc.2017.121 [pii]. PubMed PMID: 29348577. 7. Aoki T, Ohashi T, Matsumoto T, Sato M. The pipette aspiration applied to the local stiffness measurement of soft tissues. Ann Biomed Eng. 1997;25(3):581-7. Epub 1997/05/01. PubMed PMID: 9146811. 8. Buskohl PR, Gould RA, Butcher JT. Quantification of embryonic atrioventricular valve biomechanics during morphogenesis. J Biomech. 2012;45(5):895-902. Epub 2011/12/16. doi: 10.1016/j.jbiomech.2011.11.032. PubMed PMID: 22169154; PMCID: 3535469. 9. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2018. CA Cancer J Clin. 2018;68(1):7-30. Epub 2018/01/10. doi: 10.3322/caac.21442. PubMed PMID: 29313949.

119 10. Leonard JP, Martin P, Roboz GJ. Practical Implications of the 2016 Revision of the World Health Organization Classification of Lymphoid and Myeloid Neoplasms and Acute Leukemia. J Clin Oncol. 2017;35(23):2708-15. Epub 2017/06/28. doi: 10.1200/JCO.2017.72.6745. PubMed PMID: 28654364. 11. Swerdlow SH, Campo E, Pileri SA, Harris NL, Stein H, Siebert R, Advani R, Ghielmini M, Salles GA, Zelenetz AD, Jaffe ES. The 2016 revision of the World Health Organization classification of lymphoid neoplasms. Blood. 2016;127(20):2375-90. Epub 2016/03/17. doi: 10.1182/blood-2016-01-643569. PubMed PMID: 26980727; PMCID: PMC4874220. 12. Shipp M, Tamayo P, Angelo M, Ray T, Reich M, Neuberg D, Last K, Aster J, Mesirov J, Lister A, Golub TR. Diffuse large B cell lymphoma outcome prediction by gene expression profiling. Blood. 2000;96(11):222a-a. PubMed PMID: WOS:000165256100949. 13. Shipp MA, Ross KN, Tamayo P, Weng AP, Kutok JL, Aguiar RCT, Gaasenbeek M, Angelo M, Reich M, Pinkus GS, Ray TS, Koval MA, Last KW, Norton A, Lister TA, Mesirov J, Neuberg DS, Lander ES, Aster JC, Golub TR. Diffuse large B-cell lymphoma outcome prediction by gene-expression profiling and supervised machine learning. Nat Med. 2002;8(1):68-74. doi: DOI 10.1038/nm0102-68. PubMed PMID: WOS:000173056900033. 14. Scott DW, Wright GW, Williams M, Lih J, Walsh W, Jaffe ES, Rosenwald A, Campo E, Chan WC, Connors JM, Smeland EB, Mottok A, Braziel RM, Ott G, Delabie J, Tubbs RR, Cook JR, Weisenburger DD, Greiner TC, Glinsmann-Gibson BJ, Fu K, Staudt LM, Gascoyne RD, Rimsza LM. Determining Cell-Of-Origin Subtypes In Diffuse Large B-Cell Lymphoma Using Gene Expression Profiling On Formalin-Fixed Paraffin-Embedded Tissue - An LLMPP Project. Blood. 2013;122(21). PubMed PMID: WOS:000331385000102. 15. Scott DW, Wright GW, Williams PM, Lih CJ, Walsh W, Jaffe ES, Rosenwald A, Campo E, Chan WC, Connors JM, Smeland EB, Mottok A, Braziel RM, Ott G, Delabie J, Tubbs RR, Cook JR, Weisenburger DD, Greiner TC, Glinsmann-Gibson BJ, Fu K, Staudt LM, Gascoyne RD, Rimsza LM. Determining cell-of-origin subtypes of

120 diffuse large B-cell lymphoma using gene expression in formalin-fixed paraffin- embedded tissue. Blood. 2014;123(8):1214-7. doi: 10.1182/blood-2013-11-536433. PubMed PMID: WOS:000335838100021. 16. Schmitz R, Wright GW, Huang DW, Johnson CA, Phelan JD, Wang JQ, Roulland S, Kasbekar M, Young RM, Shaffer AL, Hodson DJ, Xiao W, Yu X, Yang Y, Zhao H, Xu W, Liu X, Zhou B, Du W, Chan WC, Jaffe ES, Gascoyne RD, Connors JM, Campo E, Lopez-Guillermo A, Rosenwald A, Ott G, Delabie J, Rimsza LM, Wei KTK, Zelenetz AD, Leonard JP, Bartlett NL, Tran B, Shetty J, Zhao Y, Soppet DR, Pittaluga S, Wilson WH, Staudt LM. Genetics and Pathogenesis of Diffuse Large B-Cell Lymphoma. New Engl J Med. 2018;378(15):1396-407. doi: 10.1056/NEJMoa1801445. PubMed PMID: WOS:000429718500006. 17. Roschewski M, Staudt LM, Wilson WH. Diffuse large B-cell lymphoma- treatment approaches in the molecular era. Nat Rev Clin Oncol. 2014;11(1):12-23. Epub 2013/11/13. doi: 10.1038/nrclinonc.2013.197. PubMed PMID: 24217204. 18. Burger JA, Kipps TJ. CXCR4: a key receptor in the crosstalk between tumor cells and their microenvironment. Blood. 2006;107(5):1761-7. Epub 2005/11/05. doi: 10.1182/blood-2005-08-3182. PubMed PMID: 16269611. 19. Quail DF, Joyce JA. Microenvironmental regulation of tumor progression and metastasis. Nat Med. 2013;19(11):1423-37. doi: 10.1038/nm.3394. PubMed PMID: WOS:000326920300020. 20. Abbott A. Cell culture: biology's new dimension. Nature. 2003;424(6951):870- 2. Epub 2003/08/22. doi: 10.1038/424870a. PubMed PMID: 12931155. 21. Yamada KM, Cukierman E. Modeling tissue morphogenesis and cancer in 3D. Cell. 2007;130(4):601-10. Epub 2007/08/28. doi: 10.1016/j.cell.2007.08.006. PubMed PMID: 17719539. 22. Weigelt B, Ghajar CM, Bissell MJ. The need for complex 3D culture models to unravel novel pathways and identify accurate biomarkers in breast cancer. Adv Drug Deliv Rev. 2014;69-70:42-51. Epub 2014/01/15. doi: 10.1016/j.addr.2014.01.001. PubMed PMID: 24412474; PMCID: PMC4186247.

121 23. Hirschhaeuser F, Menne H, Dittfeld C, West J, Mueller-Klieser W, Kunz- Schughart LA. Multicellular tumor spheroids: an underestimated tool is catching up again. J Biotechnol. 2010;148(1):3-15. Epub 2010/01/26. doi: 10.1016/j.jbiotec.2010.01.012. PubMed PMID: 20097238. 24. Beguelin W, Rivas MA, Calvo Fernandez MT, Teater M, Purwada A, Redmond D, Shen H, Challman MF, Elemento O, Singh A, Melnick AM. EZH2 enables germinal centre formation through epigenetic silencing of CDKN1A and an Rb-E2F1 feedback loop. Nat Commun. 2017;8(1):877. Epub 2017/10/14. doi: 10.1038/s41467-017-01029- x. PubMed PMID: 29026085; PMCID: PMC5638898. 25. Monti S, Chapuy B, Takeyama K, Rodig SJ, Hao Y, Yeda KT, Inguilizian H, Mermel C, Currie T, Dogan A, Kutok JL, Beroukhim R, Neuberg D, Habermann TM, Getz G, Kung AL, Golub TR, Shipp MA. Integrative analysis reveals an outcome- associated and targetable pattern of p53 and cell cycle deregulation in diffuse large B cell lymphoma. Cancer Cell. 2012;22(3):359-72. Epub 2012/09/15. doi: 10.1016/j.ccr.2012.07.014. PubMed PMID: 22975378; PMCID: PMC3778921. 26. Tentler JJ, Tan AC, Weekes CD, Jimeno A, Leong S, Pitts TM, Arcaroli JJ, Messersmith WA, Eckhardt SG. Patient-derived tumour xenografts as models for oncology drug development. Nat Rev Clin Oncol. 2012;9(6):338-50. Epub 2012/04/18. doi: 10.1038/nrclinonc.2012.61. PubMed PMID: 22508028; PMCID: PMC3928688. 27. Day CP, Merlino G, Van Dyke T. Preclinical mouse cancer models: a maze of opportunities and challenges. Cell. 2015;163(1):39-53. Epub 2015/09/26. doi: 10.1016/j.cell.2015.08.068. PubMed PMID: 26406370; PMCID: PMC4583714. 28. Chapuy B, Cheng H, Watahiki A, Ducar MD, Tan Y, Chen L, Roemer MG, Ouyang J, Christie AL, Zhang L, Gusenleitner D, Abo RP, Farinha P, von Bonin F, Thorner AR, Sun HH, Gascoyne RD, Pinkus GS, van Hummelen P, Wulf GG, Aster JC, Weinstock DM, Monti S, Rodig SJ, Wang Y, Shipp MA. Diffuse large B-cell lymphoma patient-derived xenograft models capture the molecular and biological heterogeneity of the disease. Blood. 2016;127(18):2203-13. Epub 2016/01/17. doi: 10.1182/blood-2015-09-672352. PubMed PMID: 26773040; PMCID: PMC4859195.

122 29. Whiteside TL. The tumor microenvironment and its role in promoting tumor growth. Oncogene. 2008;27(45):5904-12. Epub 2008/10/07. doi: 10.1038/onc.2008.271. PubMed PMID: 18836471; PMCID: PMC3689267. 30. Bhowmick NA, Neilson EG, Moses HL. Stromal fibroblasts in cancer initiation and progression. Nature. 2004;432(7015):332-7. Epub 2004/11/19. doi: 10.1038/nature03096. PubMed PMID: 15549095; PMCID: PMC3050735. 31. Vitolo U, Seymour JF, Martelli M, Illerhaus G, Illidge T, Zucca E, Campo E, Ladetto M, Committee EG. Extranodal diffuse large B-cell lymphoma (DLBCL) and primary mediastinal B-cell lymphoma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2016;27(suppl 5):v91-v102. Epub 2016/07/06. doi: 10.1093/annonc/mdw175. PubMed PMID: 27377716. 32. Yu H, Mouw JK, Weaver VM. Forcing form and function: biomechanical regulation of tumor evolution. Trends Cell Biol. 2011;21(1):47-56. Epub 2010/09/28. doi: 10.1016/j.tcb.2010.08.015. PubMed PMID: 20870407; PMCID: PMC3014395. 33. Apoorva FNU, Tian YF, Pierpont TM, Bassen DM, Cerchietti L, Butcher JT, Weiss RS, Singh A. Award Winner in the Young Investigator Category, 2017 Society for Biomaterials Annual Meeting and Exposition, Minneapolis, MN, April 05-08, 2017: Lymph node stiffness-mimicking hydrogels regulate human B-cell lymphoma growth and cell surface receptor expression in a molecular subtype-specific manner. J Biomed Mater Res A. 2017;105(7):1833-44. Epub 2017/02/09. doi: 10.1002/jbm.a.36031. PubMed PMID: 28177577. 34. Ruan J, Hajjar K, Rafii S, Leonard JP. Angiogenesis and antiangiogenic therapy in non-Hodgkin's lymphoma. Ann Oncol. 2009;20(3):413-24. Epub 2008/12/18. doi: 10.1093/annonc/mdn666 mdn666 [pii]. PubMed PMID: 19088170; PMCID: 2733074. 35. Ruan J, Luo M, Wang C, Fan L, Yang SN, Cardenas M, Geng H, Leonard JP, Melnick A, Cerchietti L, Hajjar KA. Imatinib disrupts lymphoma angiogenesis by targeting vascular pericytes. Blood. 2013;121(26):5192-202. Epub 2013/05/02. doi: 10.1182/blood-2013-03-490763. PubMed PMID: 23632889; PMCID: PMC3695363.

123 36. Reddy ST, van der Vlies AJ, Simeoni E, Angeli V, Randolph GJ, O'Neil CP, Lee LK, Swartz MA, Hubbell JA. Exploiting lymphatic transport and complement activation in nanoparticle vaccines. Nat Biotechnol. 2007;25(10):1159-64. Epub 2007/09/18. doi: 10.1038/nbt1332. PubMed PMID: 17873867. 37. Swartz MA, Kaipainen A, Netti PA, Brekken C, Boucher Y, Grodzinsky AJ, Jain RK. Mechanics of interstitial-lymphatic fluid transport: theoretical foundation and experimental validation. J Biomech. 1999;32(12):1297-307. Epub 1999/11/24. PubMed PMID: 10569708. 38. Triacca V, Guc E, Kilarski WW, Pisano M, Swartz MA. Transcellular Pathways in Lymphatic Endothelial Cells Regulate Changes in Solute Transport by Fluid Stress. Circ Res. 2017;120(9):1440-52. Epub 2017/01/29. doi: 10.1161/CIRCRESAHA.116.309828. PubMed PMID: 28130294. 39. Swartz MA, Berk DA, Jain RK. Transport in lymphatic capillaries. I. Macroscopic measurements using residence time distribution theory. Am J Physiol. 1996;270(1 Pt 2):H324-9. Epub 1996/01/01. doi: 10.1152/ajpheart.1996.270.1.H324. PubMed PMID: 8769768. 40. Berk DA, Swartz MA, Leu AJ, Jain RK. Transport in lymphatic capillaries. II. Microscopic velocity measurement with fluorescence photobleaching. Am J Physiol. 1996;270(1 Pt 2):H330-7. Epub 1996/01/01. doi: 10.1152/ajpheart.1996.270.1.H330. PubMed PMID: 8769769. 41. Lenz G, Wright G, Dave SS, Xiao W, Powell J, Zhao H, Xu W, Tan B, Goldschmidt N, Iqbal J, Vose J, Bast M, Fu K, Weisenburger DD, Greiner TC, Armitage JO, Kyle A, May L, Gascoyne RD, Connors JM, Troen G, Holte H, Kvaloy S, Dierickx D, Verhoef G, Delabie J, Smeland EB, Jares P, Martinez A, Lopez-Guillermo A, Montserrat E, Campo E, Braziel RM, Miller TP, Rimsza LM, Cook JR, Pohlman B, Sweetenham J, Tubbs RR, Fisher RI, Hartmann E, Rosenwald A, Ott G, Muller- Hermelink HK, Wrench D, Lister TA, Jaffe ES, Wilson WH, Chan WC, Staudt LM, Pr LLMP. Stromal Gene Signatures in Large-B-Cell Lymphomas. New Engl J Med. 2008;359(22):2313-23. doi: DOI 10.1056/NEJMoa0802885. PubMed PMID: WOS:000261270300003.

124 42. McKinnon DD, Domaille DW, Cha JN, Anseth KS. Biophysically defined and cytocompatible covalently adaptable networks as viscoelastic 3D cell culture systems. Adv Mater. 2014;26(6):865-72. Epub 2013/10/16. doi: 10.1002/adma.201303680. PubMed PMID: 24127293; PMCID: PMC4582033. 43. Tibbitt MW, Anseth KS. Hydrogels as extracellular matrix mimics for 3D cell culture. Biotechnol Bioeng. 2009;103(4):655-63. Epub 2009/05/28. doi: 10.1002/bit.22361. PubMed PMID: 19472329; PMCID: PMC2997742. 44. Kloxin AM, Lewis KJ, DeForest CA, Seedorf G, Tibbitt MW, Balasubramaniam V, Anseth KS. Responsive culture platform to examine the influence of microenvironmental geometry on cell function in 3D. Integr Biol (Camb). 2012;4(12):1540-9. Epub 2012/11/10. doi: 10.1039/c2ib20212c. PubMed PMID: 23138879; PMCID: PMC3928973. 45. Tian YF, Ahn H, Schneider RS, Yang SN, Roman-Gonzalez L, Melnick AM, Cerchietti L, Singh A. Integrin-specific hydrogels as adaptable tumor organoids for malignant B and T cells. Biomaterials. 2015;73:110-9. Epub 2015/09/26. doi: 10.1016/j.biomaterials.2015.09.007. PubMed PMID: 26406451; PMCID: PMC4623826. 46. Zhang J, Grubor V, Love CL, Banerjee A, Richards KL, Mieczkowski PA, Dunphy C, Choi W, Au WY, Srivastava G, Lugar PL, Rizzieri DA, Lagoo AS, Bernal- Mizrachi L, Mann KP, Flowers C, Naresh K, Evens A, Gordon LI, Czader M, Gill JI, Hsi ED, Liu QQ, Fan A, Walsh K, Jima D, Smith LL, Johnson AJ, Byrd JC, Luftig MA, Ni T, Zhu J, Chadburn A, Levy S, Dunson D, Dave SS. Genetic heterogeneity of diffuse large B-cell lymphoma. P Natl Acad Sci USA. 2013;110(4):1398-403. doi: 10.1073/pnas.1205299110. PubMed PMID: WOS:000314453900055. 47. Reddy A, Zhang J, Davis NS, Moffitt AB, Love CL, Waldrop A, Leppa S, Pasanen A, Meriranta L, Karjalainen-Lindsberg ML, Norgaard P, Pedersen M, Gang AO, Hogdall E, Heavican TB, Lone W, Iqbal J, Qin Q, Li G, Kim SY, Healy J, Richards KL, Fedoriw Y, Bernal-Mizrachi L, Koff JL, Staton AD, Flowers CR, Paltiel O, Goldschmidt N, Calaminici M, Clear A, Gribben J, Nguyen E, Czader MB, Ondrejka SL, Collie A, Hsi ED, Tse E, Au-Yeung RKH, Kwong YL, Srivastava G, Choi WWL,

125 Evens AM, Pilichowska M, Sengar M, Reddy N, Li S, Chadburn A, Gordon LI, Jaffe ES, Levy S, Rempel R, Tzeng T, Happ LE, Dave T, Rajagopalan D, Datta J, Dunson DB, Dave SS. Genetic and Functional Drivers of Diffuse Large B Cell Lymphoma. Cell. 2017;171(2):481-94 e15. Epub 2017/10/07. doi: 10.1016/j.cell.2017.09.027. PubMed PMID: 28985567; PMCID: PMC5659841. 48. Rosenwald A, Wright G, Chan WC, Connors JM, Campo E, Fisher RI, Gascoyne RD, Muller-Hermelink HK, Smeland EB, Staudt LM, Pr LLMP. The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma. New Engl J Med. 2002;346(25):1937-47. doi: DOI 10.1056/NEJMoa012914. PubMed PMID: WOS:000176266200002. 49. Meyer PN, Fu K, Greiner T, Smith L, Delabie J, Gascoyne R, Ott G, Rosenwald A, Braziel R, Campo E, Vose J, Lenz G, Staudt L, Chan W, Weisenburger DD. The Stromal Cell Marker SPARC Predicts for Survival in Patients With Diffuse Large B- Cell Lymphoma Treated With Rituximab. Am J Clin Pathol. 2011;135(1):54-61. doi: 10.1309/Ajcpjx4bjv9nlqhy. PubMed PMID: WOS:000285538700006. 50. McGranahan N, Swanton C. Clonal Heterogeneity and Tumor Evolution: Past, Present, and the Future. Cell. 2017;168(4):613-28. Epub 2017/02/12. doi: 10.1016/j.cell.2017.01.018. PubMed PMID: 28187284. 51. Marks DL, Olson RL, Fernandez-Zapico ME. Epigenetic control of the tumor microenvironment. Epigenomics. 2016;8(12):1671-87. Epub 2016/10/05. doi: 10.2217/epi-2016-0110. PubMed PMID: 27700179; PMCID: PMC5289035. 52. Fontan L, Melnick A. Molecular pathways: targeting MALT1 paracaspase activity in lymphoma. Clin Cancer Res. 2013;19(24):6662-8. doi: 10.1158/1078- 0432.CCR-12-3869. PubMed PMID: 24004675. 53. Fontan L, Yang C, Kabaleeswaran V, Volpon L, Osborne MJ, Beltran E, Garcia M, Cerchietti L, Shaknovich R, Yang SN, Fang F, Gascoyne RD, Martinez-Climent JA, Glickman JF, Borden K, Wu H, Melnick A. MALT1 small molecule inhibitors specifically suppress ABC-DLBCL in vitro and in vivo. Cancer Cell. 2012;22(6):812- 24. Epub 2012/12/15. doi: 10.1016/j.ccr.2012.11.003

S1535-6108(12)00483-7 [pii]. PubMed PMID: 23238016.

126 54. Alizadeh AA, Eisen MB, Davis RE, Ma C, Lossos IS, Rosenwald A, Boldrick JC, Sabet H, Tran T, Yu X, Powell JI, Yang L, Marti GE, Moore T, Hudson J, Jr., Lu L, Lewis DB, Tibshirani R, Sherlock G, Chan WC, Greiner TC, Weisenburger DD, Armitage JO, Warnke R, Levy R, Wilson W, Grever MR, Byrd JC, Botstein D, Brown PO, Staudt LM. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000;403(6769):503-11. Epub 2000/02/17. doi: 10.1038/35000501. PubMed PMID: 10676951. 55. Davis RE, Brown KD, Siebenlist U, Staudt LM. Constitutive nuclear factor kappaB activity is required for survival of activated B cell-like diffuse large B cell lymphoma cells. J Exp Med. 2001;194(12):1861-74. PubMed PMID: 11748286; PMCID: PMC2193582. 56. Davis RE, Ngo VN, Lenz G, Tolar P, Young RM, Romesser PB, Kohlhammer H, Lamy L, Zhao H, Yang Y, Xu W, Shaffer AL, Wright G, Xiao W, Powell J, Jiang JK, Thomas CJ, Rosenwald A, Ott G, Muller-Hermelink HK, Gascoyne RD, Connors JM, Johnson NA, Rimsza LM, Campo E, Jaffe ES, Wilson WH, Delabie J, Smeland EB, Fisher RI, Braziel RM, Tubbs RR, Cook JR, Weisenburger DD, Chan WC, Pierce SK, Staudt LM. Chronic active B-cell-receptor signalling in diffuse large B-cell lymphoma. Nature. 2010;463(7277):88-92. Epub 2010/01/08. doi: 10.1038/nature08638. PubMed PMID: 20054396; PMCID: 2845535. 57. Burger JA, Wiestner A. Targeting B cell receptor signalling in cancer: preclinical and clinical advances. Nat Rev Cancer. 2018. Epub 2018/01/20. doi: 10.1038/nrc.2017.121 nrc.2017.121 [pii]. PubMed PMID: 29348577. 58. Wilson WH, Young RM, Schmitz R, Yang Y, Pittaluga S, Wright G, Lih CJ, Williams PM, Shaffer AL, Gerecitano J, de Vos S, Goy A, Kenkre VP, Barr PM, Blum KA, Shustov A, Advani R, Fowler NH, Vose JM, Elstrom RL, Habermann TM, Barrientos JC, McGreivy J, Fardis M, Chang BY, Clow F, Munneke B, Moussa D, Beaupre DM, Staudt LM. Targeting B cell receptor signaling with ibrutinib in diffuse large B cell lymphoma. Nature medicine. 2015;21(8):922-6. Epub 2015/07/21. doi: 10.1038/nm.3884. PubMed PMID: 26193343.

127 59. Brower V. Ibrutinib promising in subtype of DLBCL. Lancet Oncol. 2015;16(9):e428. Epub 2015/07/28. doi: 10.1016/S1470-2045(15)00192-8. PubMed PMID: 26212869. 60. Kuppers R. Mechanisms of B-cell lymphoma pathogenesis. Nat Rev Cancer. 2005;5(4):251-62. Epub 2005/04/02. doi: 10.1038/nrc1589. PubMed PMID: 15803153. 61. Lenz G, Davis RE, Ngo VN, Lam L, George TC, Wright GW, Dave SS, Zhao H, Xu W, Rosenwald A, Ott G, Muller-Hermelink HK, Gascoyne RD, Connors JM, Rimsza LM, Campo E, Jaffe ES, Delabie J, Smeland EB, Fisher RI, Chan WC, Staudt LM. Oncogenic CARD11 mutations in human diffuse large B cell lymphoma. Science. 2008;319(5870):1676-9. doi: 10.1126/science.1153629. PubMed PMID: 18323416. 62. Cayrol F, Diaz Flaque MC, Fernando T, Yang SN, Sterle HA, Bolontrade M, Amoros M, Isse B, Farias RN, Ahn H, Tian YF, Tabbo F, Singh A, Inghirami G, Cerchietti L, Cremaschi GA. Integrin alphavbeta3 acting as membrane receptor for thyroid hormones mediates angiogenesis in malignant T cells. Blood. 2015;125(5):841- 51. Epub 2014/12/10. doi: 10.1182/blood-2014-07-587337. PubMed PMID: 25488971. 63. Apoorva FNU, Tian YF, Pierpont TM, Bassen DM, Cerchietti L, Butcher JT, Weiss RS, Singh A. Lymph node stiffness-mimicking hydrogels regulate human B-cell lymphoma growth and cell surface receptor expression in a molecular subtype-specific manner. J Biomed Mater Res A. 2017;105(7):1833-44. Epub 2017/02/09. doi: 10.1002/jbm.a.36031. PubMed PMID: 28177577. 64. Jafarnejad M, Woodruff MC, Zawieja DC, Carroll MC, Moore JE, Jr. Modeling Lymph Flow and Fluid Exchange with Blood Vessels in Lymph Nodes. Lymphat Res Biol. 2015;13(4):234-47. Epub 2015/12/20. doi: 10.1089/lrb.2015.0028. PubMed PMID: 26683026; PMCID: 4685511. 65. Yu W, Qu H, Hu G, Zhang Q, Song K, Guan H, Liu T, Qin J. A microfluidic- based multi-shear device for investigating the effects of low fluid-induced stresses on osteoblasts. PLoS One. 2014;9(2):e89966. Epub 2014/03/04. doi: 10.1371/journal.pone.0089966

PONE-D-13-46453 [pii]. PubMed PMID: 24587156; PMCID: 3937402.

128 66. Lecault V, Vaninsberghe M, Sekulovic S, Knapp DJ, Wohrer S, Bowden W, Viel F, McLaughlin T, Jarandehei A, Miller M, Falconnet D, White AK, Kent DG, Copley MR, Taghipour F, Eaves CJ, Humphries RK, Piret JM, Hansen CL. High- throughput analysis of single hematopoietic stem cell proliferation in microfluidic cell culture arrays. Nat Methods. 2011;8(7):581-6. Epub 2011/05/24. doi: 10.1038/nmeth.1614 nmeth.1614 [pii]. PubMed PMID: 21602799. 67. Nakao Y, Kimura H, Sakai Y, Fujii T. Bile canaliculi formation by aligning rat primary hepatocytes in a microfluidic device. Biomicrofluidics. 2011;5(2):22212. Epub 2011/07/30. doi: 10.1063/1.3580753

011191BMF [pii]. PubMed PMID: 21799718; PMCID: 3145237. 68. Blonska M, Zhu Y, Chuang HH, You MJ, Kunkalla K, Vega F, Lin X. Jun- regulated genes promote interaction of diffuse large B-cell lymphoma with the microenvironment. Blood. 2015;125(6):981-91. Epub 2014/12/24. doi: 10.1182/blood- 2014-04-568188 blood-2014-04-568188 [pii]. PubMed PMID: 25533033; PMCID: 4319238. 69. Yoon SO, Zhang X, Freedman AS, Zahrieh D, Lossos IS, Li L, Choi YS. Down- regulation of CD9 expression and its correlation to tumor progression in B lymphomas. Am J Pathol. 2010;177(1):377-86. Epub 2010/06/23. doi: 10.2353/ajpath.2010.100048

S0002-9440(10)60093-4 [pii]. PubMed PMID: 20566742; PMCID: 2893680. 70. Hodge DR, Hurt EM, Farrar WL. The role of IL-6 and STAT3 in inflammation and cancer. Eur J Cancer. 2005;41(16):2502-12. Epub 2005/10/04. doi: S0959- 8049(05)00709-4 [pii]

10.1016/j.ejca.2005.08.016. PubMed PMID: 16199153. 71. Yee C, Biondi A, Wang XH, Iscove NN, de Sousa J, Aarden LA, Wong GG, Clark SC, Messner HA, Minden MD. A possible autocrine role for interleukin-6 in two lymphoma cell lines. Blood. 1989;74(2):798-804. Epub 1989/08/01. PubMed PMID: 2787680.

129 72. Kawano M, Hirano T, Matsuda T, Taga T, Horii Y, Iwato K, Asaoku H, Tang B, Tanabe O, Tanaka H, et al. Autocrine generation and requirement of BSF-2/IL-6 for human multiple myelomas. Nature. 1988;332(6159):83-5. Epub 1988/03/03. doi: 10.1038/332083a0. PubMed PMID: 3258060. 73. Lam LT, Wright G, Davis RE, Lenz G, Farinha P, Dang L, Chan JW, Rosenwald A, Gascoyne RD, Staudt LM. Cooperative signaling through the signal transducer and activator of transcription 3 and nuclear factor-{kappa}B pathways in subtypes of diffuse large B-cell lymphoma. Blood. 2008;111(7):3701-13. Epub 2007/12/28. doi: blood- 2007-09-111948 [pii]

10.1182/blood-2007-09-111948. PubMed PMID: 18160665; PMCID: 2275028. 74. Ngo VN, Young RM, Schmitz R, Jhavar S, Xiao W, Lim KH, Kohlhammer H, Xu W, Yang Y, Zhao H, Shaffer AL, Romesser P, Wright G, Powell J, Rosenwald A, Muller-Hermelink HK, Ott G, Gascoyne RD, Connors JM, Rimsza LM, Campo E, Jaffe ES, Delabie J, Smeland EB, Fisher RI, Braziel RM, Tubbs RR, Cook JR, Weisenburger DD, Chan WC, Staudt LM. Oncogenically active MYD88 mutations in human lymphoma. Nature. 2011;470(7332):115-9. doi: 10.1038/nature09671. PubMed PMID: 21179087. 75. Yang Y, Shaffer AL, 3rd, Emre NC, Ceribelli M, Zhang M, Wright G, Xiao W, Powell J, Platig J, Kohlhammer H, Young RM, Zhao H, Xu W, Buggy JJ, Balasubramanian S, Mathews LA, Shinn P, Guha R, Ferrer M, Thomas C, Waldmann TA, Staudt LM. Exploiting synthetic lethality for the therapy of ABC diffuse large B cell lymphoma. Cancer Cell. 2012;21(6):723-37. Epub 2012/06/16. doi: 10.1016/j.ccr.2012.05.024

S1535-6108(12)00218-8 [pii]. PubMed PMID: 22698399; PMCID: 4059833. 76. Frank NY, Margaryan A, Huang Y, Schatton T, Waaga-Gasser AM, Gasser M, Sayegh MH, Sadee W, Frank MH. ABCB5-mediated doxorubicin transport and chemoresistance in human malignant melanoma. Cancer Res. 2005;65(10):4320-33. Epub 2005/05/19. doi: 65/10/4320 [pii]

10.1158/0008-5472.CAN-04-3327. PubMed PMID: 15899824.

130 77. Wilson WH, Young RM, Schmitz R, Yang Y, Pittaluga S, Wright G, Lih CJ, Williams PM, Shaffer AL, Gerecitano J, de Vos S, Goy A, Kenkre VP, Barr PM, Blum KA, Shustov A, Advani R, Fowler NH, Vose JM, Elstrom RL, Habermann TM, Barrientos JC, McGreivy J, Fardis M, Chang BY, Clow F, Munneke B, Moussa D, Beaupre DM, Staudt LM. Targeting B cell receptor signaling with ibrutinib in diffuse large B cell lymphoma. Nat Med. 2015;21(8):922-6. Epub 2015/07/21. doi: 10.1038/nm.3884 nm.3884 [pii]. PubMed PMID: 26193343. 78. Goldstein RL, Yang SN, Taldone T, Chang B, Gerecitano J, Elenitoba-Johnson K, Shaknovich R, Tam W, Leonard JP, Chiosis G, Cerchietti L, Melnick A. Pharmacoproteomics identifies combinatorial therapy targets for diffuse large B cell lymphoma. J Clin Invest. 2015;125(12):4559-71. Epub 2015/11/04. doi: 10.1172/JCI80714

80714 [pii]. PubMed PMID: 26529251; PMCID: 4665772. 79. Coiffier B, Lepage E, Briere J, Herbrecht R, Tilly H, Bouabdallah R, Morel P, Van Den Neste E, Salles G, Gaulard P, Reyes F, Lederlin P, Gisselbrecht C. CHOP chemotherapy plus rituximab compared with CHOP alone in elderly patients with diffuse large-B-cell lymphoma. N Engl J Med. 2002;346(4):235-42. Epub 2002/01/25. doi: 10.1056/NEJMoa011795

346/4/235 [pii]. PubMed PMID: 11807147. 80. McLaughlin P, Grillo-Lopez AJ, Link BK, Levy R, Czuczman MS, Williams ME, Heyman MR, Bence-Bruckler I, White CA, Cabanillas F, Jain V, Ho AD, Lister J, Wey K, Shen D, Dallaire BK. Rituximab chimeric anti-CD20 monoclonal antibody therapy for relapsed indolent lymphoma: half of patients respond to a four-dose treatment program. J Clin Oncol. 1998;16(8):2825-33. Epub 1998/08/15. doi: 10.1200/JCO.1998.16.8.2825. PubMed PMID: 9704735. 81. Wojciechowski W, Li H, Marshall S, Dell'Agnola C, Espinoza-Delgado I. Enhanced expression of CD20 in human tumor B cells is controlled through ERK-

131 dependent mechanisms. Journal of immunology. 2005;174(12):7859-68. Epub 2005/06/10. PubMed PMID: 15944291. 82. Tzima E, del Pozo MA, Shattil SJ, Chien S, Schwartz MA. Activation of integrins in endothelial cells by fluid shear stress mediates Rho-dependent cytoskeletal alignment. The EMBO journal. 2001;20(17):4639-47. Epub 2001/09/05. doi: 10.1093/emboj/20.17.4639. PubMed PMID: 11532928; PMCID: 125600. 83. Lee DY, Li YS, Chang SF, Zhou J, Ho HM, Chiu JJ, Chien S. Oscillatory flow- induced proliferation of osteoblast-like cells is mediated by alphavbeta3 and beta1 integrins through synergistic interactions of focal adhesion kinase and Shc with phosphatidylinositol 3-kinase and the Akt/mTOR/p70S6K pathway. The Journal of biological chemistry. 2010;285(1):30-42. Epub 2009/11/06. doi: 10.1074/jbc.M109.010512. PubMed PMID: 19889638; PMCID: 2804177. 84. Tjin EP, Groen RW, Vogelzang I, Derksen PW, Klok MD, Meijer HP, van Eeden S, Pals ST, Spaargaren M. Functional analysis of HGF/MET signaling and aberrant HGF-activator expression in diffuse large B-cell lymphoma. Blood. 2006;107(2):760-8. Epub 2005/09/29. doi: 2005-05-1929 [pii]

10.1182/blood-2005-05-1929. PubMed PMID: 16189274. 85. Poincloux R, Collin O, Lizarraga F, Romao M, Debray M, Piel M, Chavrier P. Contractility of the cell rear drives invasion of breast tumor cells in 3D Matrigel. Proc Natl Acad Sci U S A. 2011;108(5):1943-8. Epub 2011/01/20. doi: 10.1073/pnas.1010396108

1010396108 [pii]. PubMed PMID: 21245302; PMCID: 3033302. 86. Singh A, Brito I, Lammerding J. Beyond Tissue Stiffness and Bioadhesivity: Advanced Biomaterials to Model Tumor Microenvironments and Drug Resistance. Trends Cancer. 2018. doi: https://doi.org/10.1016/j.trecan.2018.01.008. 87. Cosgrove BD, Alexopoulos LG, Hang TC, Hendriks BS, Sorger PK, Griffith LG, Lauffenburger DA. Cytokine-associated drug toxicity in human hepatocytes is associated with signaling network dysregulation. Mol Biosyst. 2010;6(7):1195-206. Epub 2010/04/03. doi: 10.1039/b926287c. PubMed PMID: 20361094; PMCID: 2943488.

132 88. Nguyen TK, Jordan N, Friedberg J, Fisher RI, Dent P, Grant S. Inhibition of MEK/ERK1/2 sensitizes lymphoma cells to sorafenib-induced apoptosis. Leuk Res. 2010;34(3):379-86. Epub 2010/02/02. doi: 10.1016/j.leukres.2009.07.013. PubMed PMID: 20117835; PMCID: 3150480. 89. Dai B, Zhao XF, Hagner P, Shapiro P, Mazan-Mamczarz K, Zhao S, Natkunam Y, Gartenhaus RB. Extracellular signal-regulated kinase positively regulates the oncogenic activity of MCT-1 in diffuse large B-cell lymphoma. Cancer research. 2009;69(19):7835-43. Epub 2009/10/01. doi: 10.1158/0008-5472.CAN-09-1606. PubMed PMID: 19789340; PMCID: 2839501. 90. Li D, Tang T, Lu J, Dai K. Effects of flow shear stress and mass transport on the construction of a large-scale tissue-engineered bone in a perfusion bioreactor. Tissue Eng Part A. 2009;15(10):2773-83. Epub 2009/02/20. doi: 10.1089/ten.TEA.2008.0540. PubMed PMID: 19226211. 91. Geng H, Hurtz C, Lenz KB, Chen Z, Baumjohann D, Thompson S, Goloviznina NA, Chen WY, Huan J, LaTocha D, Ballabio E, Xiao G, Lee JW, Deucher A, Qi Z, Park E, Huang C, Nahar R, Kweon SM, Shojaee S, Chan LN, Yu J, Kornblau SM, Bijl JJ, Ye BH, Ansel KM, Paietta E, Melnick A, Hunger SP, Kurre P, Tyner JW, Loh ML, Roeder RG, Druker BJ, Burger JA, Milne TA, Chang BH, Muschen M. Self-enforcing feedback activation between BCL6 and pre-B cell receptor signaling defines a distinct subtype of acute lymphoblastic leukemia. Cancer Cell. 2015;27(3):409-25. Epub 2015/03/12. doi: 10.1016/j.ccell.2015.02.003

S1535-6108(15)00053-7 [pii]. PubMed PMID: 25759025; PMCID: 4618684. 92. Spurgeon SE, Coffey G, Fletcher LB, Burke R, Tyner JW, Druker BJ, Betz A, DeGuzman F, Pak Y, Baker D, Pandey A, Hollenbach SJ, Sinha U, Loriaux MM. The selective SYK inhibitor P505-15 (PRT062607) inhibits B cell signaling and function in vitro and in vivo and augments the activity of fludarabine in chronic lymphocytic leukemia. J Pharmacol Exp Ther. 2013;344(2):378-87. Epub 2012/12/12. doi: 10.1124/jpet.112.200832 jpet.112.200832 [pii]. PubMed PMID: 23220742; PMCID: 3558816.

133 93. Fellmann C, Hoffmann T, Sridhar V, Hopfgartner B, Muhar M, Roth M, Lai DY, Barbosa IA, Kwon JS, Guan Y, Sinha N, Zuber J. An optimized microRNA backbone for effective single-copy RNAi. Cell Rep. 2013;5(6):1704-13. Epub 2013/12/18. doi: 10.1016/j.celrep.2013.11.020

S2211-1247(13)00687-6 [pii]. PubMed PMID: 24332856. 94. Xia YN, Whitesides GM. Soft lithography. Angew Chem Int Edit. 1998;37(5):550-75. doi: Doi 10.1002/(Sici)1521-3773(19980316)37:5<550::Aid- Anie550>3.0.Co;2-G. PubMed PMID: WOS:000072649300001. 95. Ng CP, Pun SH. A perfusable 3D cell-matrix tissue culture chamber for in situ evaluation of nanoparticle vehicle penetration and transport. Biotechnol Bioeng. 2008;99(6):1490-501. Epub 2007/10/31. doi: 10.1002/bit.21698. PubMed PMID: 17969174; PMCID: 2796126. 96. McCarty WJ, Johnson M. The hydraulic conductivity of Matrigel. Biorheology. 2007;44(5-6):303-17. Epub 2008/04/11. PubMed PMID: 18401073; PMCID: 2562030. 97. Pedersen JA, Lichter S, Swartz MA. Cells in 3D matrices under interstitial flow: effects of extracellular matrix alignment on cell shear stress and drag forces. Journal of biomechanics. 2010;43(5):900-5. Epub 2009/12/17. doi: 10.1016/j.jbiomech.2009.11.007

S0021-9290(09)00643-5 [pii]. PubMed PMID: 20006339. 98. T Clozel SY, R Elstrom, W Tam, P Martin, M Kormaksson, S Banerjee, A Vasanthakumar, B Culjkovic, D Scott, S Brennan, M Leser, R Shaknovich, A Chadburn, F Tabbo, L Godley, R Gascoyne, K Borden, G Inghirami, J Leonard, A Melnick, L Cerchietti. Mechanism-Based Epigenetic Chemosensitization Therapy of Diffuse Large B Cell Lymphoma. Cancer Discovery. 2013;3(9):1002-19. 99. Hashwah H, Schmid CA, Kasser S, Bertram K, Stelling A, Manz MG, Muller A. Inactivation of CREBBP expands the germinal center B cell compartment, down- regulates MHCII expression and promotes DLBCL growth. Proc Natl Acad Sci U S A. 2017. Epub 2017/08/24. doi: 201619555 [pii]

10.1073/pnas.1619555114

134 1619555114 [pii]. PubMed PMID: 28831000. 100. Pfeifer M, Grau M, Lenze D, Wenzel SS, Wolf A, Wollert-Wulf B, Dietze K, Nogai H, Storek B, Madle H, Dorken B, Janz M, Dirnhofer S, Lenz P, Hummel M, Tzankov A, Lenz G. PTEN loss defines a PI3K/AKT pathway-dependent germinal center subtype of diffuse large B-cell lymphoma. Proc Natl Acad Sci U S A. 2013;110(30):12420-5. Epub 2013/07/11. doi: 10.1073/pnas.1305656110

1305656110 [pii]. PubMed PMID: 23840064; PMCID: 3725065. 101. Atencia J, Beebe DJ. Controlled microfluidic interfaces. Nature. 2005;437(7059):648-55. Epub 2005/09/30. doi: 10.1038/nature04163. PubMed PMID: 16193039. 102. Chaudhuri O, Koshy ST, Branco da Cunha C, Shin JW, Verbeke CS, Allison KH, Mooney DJ. Extracellular matrix stiffness and composition jointly regulate the induction of malignant phenotypes in mammary epithelium. Nature materials. 2014. Epub 2014/06/16. doi: 10.1038/nmat4009. PubMed PMID: 24930031. 103. Paszek MJ, DuFort CC, Rossier O, Bainer R, Mouw JK, Godula K, Hudak JE, Lakins JN, Wijekoon AC, Cassereau L, Rubashkin MG, Magbanua MJ, Thorn KS, Davidson MW, Rugo HS, Park JW, Hammer DA, Giannone G, Bertozzi CR, Weaver VM. The cancer glycocalyx mechanically primes integrin-mediated growth and survival. Nature. 2014;511(7509):319-25. Epub 2014/07/18. doi: 10.1038/nature13535. PubMed PMID: 25030168; PMCID: 4487551. 104. Paszek MJ, Zahir N, Johnson KR, Lakins JN, Rozenberg GI, Gefen A, Reinhart- King CA, Margulies SS, Dembo M, Boettiger D, Hammer DA, Weaver VM. Tensional homeostasis and the malignant phenotype. Cancer Cell. 2005;8(3):241-54. Epub 2005/09/20. doi: S1535-6108(05)00268-0 [pii]

10.1016/j.ccr.2005.08.010. PubMed PMID: 16169468. 105. Damiano JS, Dalton WS. Integrin-mediated drug resistance in multiple myeloma. Leukemia & lymphoma. 2000;38(1-2):71-81. doi: 10.3109/10428190009060320. PubMed PMID: 10811449. 106. Naci D, El Azreq MA, Chetoui N, Lauden L, Sigaux F, Charron D, Al-Daccak R, Aoudjit F. alpha2beta1 integrin promotes chemoresistance against doxorubicin in

135 cancer cells through extracellular signal-regulated kinase (ERK). J Biol Chem. 2012;287(21):17065-76. doi: 10.1074/jbc.M112.349365. PubMed PMID: 22457358; PMCID: 3366820. 107. Seo BR, Bhardwaj P, Choi S, Gonzalez J, Andresen Eguiluz RC, Wang K, Mohanan S, Morris PG, Du B, Zhou XK, Vahdat LT, Verma A, Elemento O, Hudis CA, Williams RM, Gourdon D, Dannenberg AJ, Fischbach C. Obesity-dependent changes in interstitial ECM mechanics promote breast tumorigenesis. Sci Transl Med. 2015;7(301):301ra130. Epub 2015/08/21. doi: 10.1126/scitranslmed.3010467. PubMed PMID: 26290412. 108. Shin J, Mooney DJ. Extracellular matrix stiffness causes systematic variations in proliferation and chemosensitivity in myeloid leukemias. P Natl Acad Sci USA. 2016. doi: 10.1073/pnas.1611338113. 109. Buskohl PR, Gould RA, Butcher JT. Quantification of embryonic atrioventricular valve biomechanics during morphogenesis. Journal of biomechanics. 2012;45(5):895-902. Epub 2011/12/16. doi: 10.1016/j.jbiomech.2011.11.032. PubMed PMID: 22169154; PMCID: 3535469. 110. Buskohl PR, Sun MJ, Thompson RP, Butcher JT. Serotonin potentiates transforming growth factor-beta3 induced biomechanical remodeling in avian embryonic atrioventricular valves. PLoS One. 2012;7(8):e42527. Epub 2012/08/11. doi: 10.1371/journal.pone.0042527. PubMed PMID: 22880017; PMCID: 3412853. 111. Purwada A, Jaiswal MK, Ahn H, Nojima T, Kitamura D, Gaharwar AK, Cerchietti L, Singh A. Ex vivo engineered immune organoids for controlled germinal center reactions. Biomaterials. 2015;63:24-34. Epub 2015/06/16. doi: 10.1016/j.biomaterials.2015.06.002. PubMed PMID: 26072995; PMCID: PMC4490011. 112. Purwada A, Singh A. Immuno-engineered Organoids for Regulating the Kinetics of B cell Development and Antibody Production. Nature Protocols (In press). 2017;doi:10.1038/nprot.2016.157. 113. Butcher JT, McQuinn TC, Sedmera D, Turner D, Markwald RR. Transitions in early embryonic atrioventricular valvular function correspond with changes in cushion

136 biomechanics that are predictable by tissue composition. Circ Res. 2007;100(10):1503- 11. Epub 2007/05/05. doi: CIRCRESAHA.107.148684 [pii]

10.1161/CIRCRESAHA.107.148684. PubMed PMID: 17478728. 114. Singh A, Qin H, Fernandez I, Wei J, Lin J, Kwak LW, Roy K. An injectable synthetic immune-priming center mediates efficient T-cell class switching and T-helper 1 response against B cell lymphoma. Journal of controlled release : official journal of the Controlled Release Society. 2011;155(2):184-92. Epub 2011/06/29. doi: 10.1016/j.jconrel.2011.06.008. PubMed PMID: 21708196. 115. Singh A, Nie H, Ghosn B, Qin H, Kwak LW, Roy K. Efficient modulation of T- cell response by dual-mode, single-carrier delivery of cytokine-targeted siRNA and DNA vaccine to antigen-presenting cells. Mol Ther. 2008;16(12):2011-21. PubMed PMID: 18813280. 116. Harris AW, Pinkert CA, Crawford M, Langdon WY, Brinster RL, Adams JM. The E mu-myc transgenic mouse. A model for high-incidence spontaneous lymphoma and leukemia of early B cells. J Exp Med. 1988;167(2):353-71. Epub 1988/02/01. PubMed PMID: 3258007; PMCID: 2188841. 117. Rohner NA, McClain J, Tuell SL, Warner A, Smith B, Yun Y, Mohan A, Sushnitha M, Thomas SN. Lymph node biophysical remodeling is associated with melanoma lymphatic drainage. FASEB J. 2015;29(11):4512-22. Epub 2015/07/17. doi: 10.1096/fj.15-274761 fj.15-274761 [pii]. PubMed PMID: 26178165; PMCID: 4608913. 118. Hirsch S, Guo J, Reiter R, Papazoglou S, Kroencke T, Braun J, Sack I. MR elastography of the liver and the spleen using a piezoelectric driver, single-shot wave- field acquisition, and multifrequency dual parameter reconstruction. Magn Reson Med. 2014;71(1):267-77. Epub 2013/02/16. doi: 10.1002/mrm.24674. PubMed PMID: 23413115. 119. Wing-Han Yuen Q, Zheng YP, Huang YP, He JF, Chung-Wai Cheung J, Ying M. In-vitro Strain and Modulus Measurements in Porcine Cervical Lymph Nodes. Open Biomed Eng J. 2011;5:39-46. Epub 2011/06/07. doi: 10.2174/1874120701105010039

137 TOBEJ-5-39 [pii]. PubMed PMID: 21643424; PMCID: 3103907. 120. Tan Y, Tajik A, Chen J, Jia Q, Chowdhury F, Wang L, Zhang S, Hong Y, Yi H, Wu DC, Zhang Y, Wei F, Poh YC, Seong J, Singh R, Lin LJ, Doganay S, Li Y, Jia H, Ha T, Wang Y, Huang B, Wang N. Matrix softness regulates plasticity of tumour- repopulating cells via H3K9 demethylation and Sox2 expression. Nat Commun. 2014;5:4619. Epub 2014/08/08. doi: 10.1038/ncomms5619 ncomms5619 [pii]. PubMed PMID: 25099074; PMCID: 4133791. 121. Tilghman RW, Cowan CR, Mih JD, Koryakina Y, Gioeli D, Slack-Davis JK, Blackman BR, Tschumperlin DJ, Parsons JT. Matrix rigidity regulates cancer cell growth and cellular phenotype. PLoS One. 2010;5(9):e12905. Epub 2010/10/05. doi: 10.1371/journal.pone.0012905 e12905 [pii]. PubMed PMID: 20886123; PMCID: 2944843. 122. O'Connor RS, Hao X, Shen K, Bashour K, Akimova T, Hancock WW, Kam LC, Milone MC. Substrate rigidity regulates human T cell activation and proliferation. J Immunol. 2012;189(3):1330-9. Epub 2012/06/27. doi: 10.4049/jimmunol.1102757 jimmunol.1102757 [pii]. PubMed PMID: 22732590; PMCID: 3401283. 123. Yan C, Pochan DJ. Rheological properties of peptide-based hydrogels for biomedical and other applications. Chemical Society reviews. 2010;39(9):3528-40. Epub 2010/04/28. doi: 10.1039/b919449p. PubMed PMID: 20422104; PMCID: 3104857. 124. Hutson CB, Nichol JW, Aubin H, Bae H, Yamanlar S, Al-Haque S, Koshy ST, Khademhosseini A. Synthesis and characterization of tunable poly(ethylene glycol): gelatin methacrylate composite hydrogels. Tissue Eng Part A. 2011;17(13-14):1713-23. Epub 2011/02/11. doi: 10.1089/ten.TEA.2010.0666. PubMed PMID: 21306293; PMCID: 3118706. 125. Patel RG, Purwada A, Cerchietti L, Inghirami G, Melnick A, Gaharwar AK, Singh A. Microscale Bioadhesive Hydrogel Arrays for Cell Engineering Applications. Cell Mol Bioeng. 2014;7(3):394-408. Epub 2014/10/21. doi: 10.1007/s12195-014- 0353-8. PubMed PMID: 25328548.

138 126. Reichl EM, Ren Y, Morphew MK, Delannoy M, Effler JC, Girard KD, Divi S, Iglesias PA, Kuo SC, Robinson DN. Interactions between myosin and actin crosslinkers control cytokinesis contractility dynamics and mechanics. Curr Biol. 2008;18(7):471- 80. Epub 2008/03/29. doi: 10.1016/j.cub.2008.02.056

S0960-9822(08)00243-1 [pii]. PubMed PMID: 18372178; PMCID: 2361134. 127. Butcher DT, Alliston T, Weaver VM. A tense situation: forcing tumour progression. Nat Rev Cancer. 2009;9(2):108-22. Epub 2009/01/24. doi: 10.1038/nrc2544 nrc2544 [pii]. PubMed PMID: 19165226; PMCID: 2649117. 128. Swaminathan V, Mythreye K, O'Brien ET, Berchuck A, Blobe GC, Superfine R. Mechanical stiffness grades metastatic potential in patient tumor cells and in cancer cell lines. Cancer Res. 2011;71(15):5075-80. Epub 2011/06/07. doi: 10.1158/0008- 5472.CAN-11-0247

0008-5472.CAN-11-0247 [pii]. PubMed PMID: 21642375; PMCID: 3220953. 129. Dierks C, Grbic J, Zirlik K, Beigi R, Englund NP, Guo GR, Veelken H, Engelhardt M, Mertelsmann R, Kelleher JF, Schultz P, Warmuth M. Essential role of stromally induced hedgehog signaling in B-cell malignancies. Nat Med. 2007;13(8):944-51. Epub 2007/07/17. doi: nm1614 [pii]

10.1038/nm1614. PubMed PMID: 17632527. 130. Roecklein BA, Torok-Storb B. Functionally distinct human marrow stromal cell lines immortalized by transduction with the human papilloma virus E6/E7 genes. Blood. 1995;85(4):997-1005. Epub 1995/02/15. PubMed PMID: 7849321. 131. Graf L, Iwata M, Torok-Storb B. Gene expression profiling of the functionally distinct human bone marrow stromal cell lines HS-5 and HS-27a. Blood. 2002;100(4):1509-11. Epub 2002/08/20. PubMed PMID: 12184274. 132. Mhyre AJ, Marcondes AM, Spaulding EY, Deeg HJ. Stroma-dependent apoptosis in clonal hematopoietic precursors correlates with expression of PYCARD. Blood. 2009;113(3):649-58. Epub 2008/10/24. doi: 10.1182/blood-2008-04-152686 blood-2008-04-152686 [pii]. PubMed PMID: 18945969; PMCID: 2628371.

139 133. Nojima T, Haniuda K, Moutai T, Matsudaira M, Mizokawa S, Shiratori I, Azuma T, Kitamura D. In-vitro derived germinal centre B cells differentially generate memory B or plasma cells in vivo. Nat Commun. 2011;2:465. Epub 2011/09/08. doi: 10.1038/ncomms1475. PubMed PMID: 21897376. 134. Purwada A, Tian YF, Huang WS, Rohrbach KM, Deol S, August A, Singh A. Self-Assembly Protein Nanogels for Safer Cancer Immunotherapy. Adv Healthc Mater. 2016;5(12):1413-9. doi: 10.1002/adhm.201501062. PubMed PMID: ISI:000379550400004. 135. Ali OA, Huebsch N, Cao L, Dranoff G, Mooney DJ. Infection-mimicking materials to program dendritic cells in situ. Nat Mater. 2009;8(2):151-8. PubMed PMID: 19136947. 136. Duvic M, Vu J. Update on the treatment of cutaneous T-cell lymphoma (CTCL): Focus on vorinostat. Biologics. 2007;1(4):377-92. Epub 2007/12/01. PubMed PMID: 19707308; PMCID: 2721288. 137. Mandawat A, Fiskus W, Buckley KM, Robbins K, Rao R, Balusu R, Navenot JM, Wang ZX, Ustun C, Chong DG, Atadja P, Fujii N, Peiper SC, Bhalla K. Pan-histone deacetylase inhibitor panobinostat depletes CXCR4 levels and signaling and exerts synergistic antimyeloid activity in combination with CXCR4 antagonists. Blood. 2010;116(24):5306-15. Epub 2010/09/03. doi: 10.1182/blood-2010-05-284414 blood-2010-05-284414 [pii]. PubMed PMID: 20810927. 138. San-Miguel JF, Hungria VT, Yoon SS, Beksac M, Dimopoulos MA, Elghandour A, Jedrzejczak WW, Gunther A, Nakorn TN, Siritanaratkul N, Corradini P, Chuncharunee S, Lee JJ, Schlossman RL, Shelekhova T, Yong K, Tan D, Numbenjapon T, Cavenagh JD, Hou J, LeBlanc R, Nahi H, Qiu L, Salwender H, Pulini S, Moreau P, Warzocha K, White D, Blade J, Chen W, de la Rubia J, Gimsing P, Lonial S, Kaufman JL, Ocio EM, Veskovski L, Sohn SK, Wang MC, Lee JH, Einsele H, Sopala M, Corrado C, Bengoudifa BR, Binlich F, Richardson PG. Panobinostat plus bortezomib and dexamethasone versus placebo plus bortezomib and dexamethasone in patients with relapsed or relapsed and refractory multiple myeloma: a multicentre, randomised,

140 double-blind phase 3 trial. Lancet Oncol. 2014;15(11):1195-206. Epub 2014/09/23. doi: 10.1016/S1470-2045(14)70440-1

S1470-2045(14)70440-1 [pii]. PubMed PMID: 25242045. 139. Richardson PG, Schlossman RL, Alsina M, Weber DM, Coutre SE, Gasparetto C, Mukhopadhyay S, Ondovik MS, Khan M, Paley CS, Lonial S. PANORAMA 2: panobinostat in combination with bortezomib and dexamethasone in patients with relapsed and bortezomib-refractory myeloma. Blood. 2013;122(14):2331-7. Epub 2013/08/21. doi: 10.1182/blood-2013-01-481325 blood-2013-01-481325 [pii]. PubMed PMID: 23950178. 140. Dumbauld DW, Lee T, Singh A, Scrimgeour J, Gersback CA, Zamir EA, Fu J, Chen C, Curtis JE, Craig SW, Garcia AJ. How vinculin regulates force transmission. Proc Natl Acad Sci U S A. 2013:In Press. 141. Singh A, Suri S, Lee T, Chilton JM, Cooke MT, Chen W, Fu J, Stice SL, Lu H, McDevitt TC, Garcia AJ. Adhesion strength-based, label-free isolation of human pluripotent stem cells. Nat Methods. 2013;10(5):438-44. Epub 2013/04/09. doi: 10.1038/nmeth.2437. PubMed PMID: 23563795; PMCID: 3641175. 142. Puklin-Faucher E, Sheetz MP. The mechanical integrin cycle. J Cell Sci. 2009;122(Pt 2):179-86. Epub 2009/01/02. doi: 10.1242/jcs.042127. PubMed PMID: 19118210. 143. Brakebusch C, Wennerberg K, Krell HW, Weidle UH, Sallmyr A, Johansson S, Fassler R. Beta1 integrin promotes but is not essential for metastasis of ras-myc transformed fibroblasts. Oncogene. 1999;18(26):3852-61. Epub 1999/08/13. doi: 10.1038/sj.onc.1202770. PubMed PMID: 10445848. 144. Lahlou H, Muller WJ. beta1-integrins signaling and mammary tumor progression in transgenic mouse models: implications for human breast cancer. Breast cancer research : BCR. 2011;13(6):229. Epub 2012/01/24. doi: 10.1186/bcr2905 bcr2905 [pii]. PubMed PMID: 22264244; PMCID: 3326542.

141 145. Michael KE, Garcia AJ. Cell adhesion strengthening: measurement and analysis. Methods Cell Biol. 2007;83:329-46. Epub 2007/07/07. doi: S0091-679X(07)83014-7 [pii]

10.1016/S0091-679X(07)83014-7. PubMed PMID: 17613315. 146. Coyer SR, Singh A, Dumbauld DW, Calderwood DA, Craig SW, Delamarche E, Garcia AJ. Nanopatterning reveals an ECM area threshold for focal adhesion assembly and force transmission that is regulated by integrin activation and cytoskeleton tension. J Cell Sci. 2012;125(21):5110-23. doi: Doi 10.1242/Jcs.108035. PubMed PMID: ISI:000312984300018.

Figure Legend

Figure 4.1 Effect of DLBCLs on Proliferation of HUVECs. A) Prominent blood vessels in a neoplastic lymph node from 2 months old Eu-Myc mice. B) Proliferation of HUVECs in the presence of conditioned media from DLBCLs. Fold change in the MTS

142 absorption. N =5; Mean ± S.E.M; p<0.05. C) Proliferation of monoculture of ABC- DLBCL lines (LY-10, LY-3, HBL-1) and GCB-DLBCL line (LY-1 and LY-7) N =5; Mean ± S.E.M; p<0.05, D) Proliferation of ABC- DLBCL lines (LY-10, LY-3, HBL-1) and GCB-DLBCL line (LY-1 and LY-7) as a result of coculture with HUVECs N =5; Mean ± S.E.M; p<0.05

Figure 4.2 Synthetic hydrogel-based engineered lymphoma organoids with tunable

143 integrin ligand presentation as a model for angiogenesis. A) Schematic to describe the hydrogel synthesis method, B). Schematic of our angiogenesis platform describing the relative arrangement of DLBCL organoids, Collagen ECM and HUEVCs, C) Effect of integrin ligand specificity and its density on Proliferation of ABC-DLBCL lines (LY- 10, LY-3, HBL-1) and GCB-DLBCL line (LY-1 and LY-7) organoids as a result of N =5; Mean ± S.E.M; p<0.05, D) Effect of integrin ligand specificity and its density on surface receptors of DLBCL N =5; Mean ± S.E.M; p<0.05.

144 Figure 4.3 3D integrin engagement modulates release of growth factor by lymphomas and development of vascular network. (A-C) VEGF release profile of ABC-DLBCL lines (LY-10, LY-3, HBL-1) and GCB-DLBCL line LY-1 in organoids, Mean ± S.E.M, n = 3, *p < 0.05, (D) Bar graph showing reduction in VEGF secretion level as a result of inhibition of v 3 and 4 1. Mean ± S.E.M, n = 3, *p < 0.05

Figure 4.4 Endothelial cells engage v 3 integrins to adhere to the ECM, and 4 1 to adhere to DLBCLs and to form clusters. (A) Microscopic images showing the adhesion and spreading of HUVECs with and without v 3 and 4 1 integrin inhibitors. (B) Bar graphs representing apoptosis of B cell lymphomas and HUVECs as result of inhibition Mean ± S.E.M, n = 3, *p < 0.05, (C) Bar graphs represent reduction in cluster size of DLBCLs as a result of integrin inhibition Mean ± S.E.M, n = 3, *p < 0.05.

145

Figure 4.5. Integrins v3 and 41 modulate the migration of endothelial cells and DLBCLs. (A) Microscopic images of vascular network formation by endothelial cells induced by RGD presenting HBL-1 organoids. The images show infiltration of HUVECs into HBL-1 organoid on third day of starting seeding of cells. (B) Schematic of the microfluidic conduit system to study the migration of HUVECs and DLBCLs, and (C) Microscopic images showing protruding HUVECs under chemotactic gradient created by DLBCL organoids. The difference in migration towards 2D vs 3D DLBCL organoids is visible by as soon as 12 hours (D) Visible difference between proliferation of HUVECs inside conduit after 48 hrs; conduit facing 3D organoids proliferate higher than the conduit connecting 2D DLBCL.

146 Chapter 5 Future Work

Moving forward, we propose to combine three innovations, employ a lymphoid organoid with suitable lymph node mimicking mechanical properties within microfluidic model of lymphatic system along with various cellular component of lymphomas microenvironment to present a holistic engineer lymphoma-on-chip that would serve as a surrogate lymphoid-and-patient-specific neoplasm allowing for efficient and accurate screening of therapeutic drugs for lymphoma. Such a comprehensive biosynthetic model would possess the desired qualitied such as modularity, tunability, easy to use, and ability to perform long term cultures. Currently we lack a suitable biosynthetic angiogenesis model for DLBCL. Anti- angiogenic drugs have played an important role in advancing cancer therapy. The 3D organoid model presented in chapter 4 of this thesis demonstrated its application in study of angiogenic factors. This organoid model could potentially be tailored into a bio-synthetic model of DLBCL vascularization. The current landscape of lymphoma therapy is diverse, and it includes anti-lymphoma drugs ranging from chemotherapeutics to those whose mechanism of action is non- cytotoxic and microenvironment dependent, such as epigenetic and metabolic agents, which require long term survival for reprogramming and therefore cannot be tested using conventional 2D culture platforms. This necessitates to improve upon the biosynthetic models presented here to make them suitable for testing a wide range of

147