UNIVERSITY OF CALIFORNIA

Los Angeles

Stem Cell Engineered

Invariant Natural Killer T Cells

for Cancer Immunotherapy

A dissertation submitted in partial satisfaction of the

requirements for the degree Doctor of Philosophy

in Molecular Biology

by

Drake John Smith

2017

© Copyright by

Drake John Smith

2017 ABSTRACT OF THE DISSERTATION

Stem Cell Engineered

Invariant Natural Killer T Cells

for Cancer Immunotherapy

by

Drake John Smith

Doctor of Philosophy in Molecular Biology

University of California, Los Angeles, 2017

Professor Donald Barry Kohn, Chair

Cancer immunotherapy is a rapidly developing field that has already shown to be of great clinical value as evidenced by the success of engineered T cell therapies, such as chimeric antigen receptor (CAR) therapies in treating B cell leukemia and T cell receptor (TCR) therapies in treating melanoma, and by checkpoint inhibitor therapies, such as PD-1 and CTLA-4 antibodies, in treating a variety of cancers. This dissertation seeks to add to this growing knowledge base and carve out a niche in the discipline by utilizing a unique combination of immune cell type and method of delivery; a hematopoietic stem cell (HSC) can be genetically engineered using a viral vector in order to generate invariant natural killer T (iNKT) cells in vivo.

Several cancer immunotherapy clinical trials have already utilized iNKT cells either by infusion after expansion ex vivo and/or activation of the cells in vivo by dendritic cells loaded

ii with the synthetic ligand α-Galactosylceramide (α-GalCer). These trials have demonstrated that the treatments are well tolerated, and while some have shown promising anti-tumor immunity, most have yielded unsatisfactory results. This lack of clinical efficacy has been attributed to the cells’ very low and highly variable number in humans (0.001-1% in peripheral blood) and their rapid depletion after stimulation.

Many cancer immunotherapy treatments trend towards a phase of promising tumor regression followed by a disheartening cancer relapse. This may be due to several factors, but a major contributor specific to cancer immunotherapy is thought to be exhaustion of the therapeutic cells by the ex vivo expansion protocol. This protocol drives the therapeutic cells to expand and differentiate into terminally differentiated effector cells. While these effector cells have increased killing efficacy, it comes at the expense of a decreased life span and regeneration.

This would explain the initial regression mediated by the effector cells, followed by a relapse when the cells become exhausted. This issue can be addressed by utilizing viral vectors to genetically engineer hematopoietic stem cells to continually generate new therapeutic cells in vivo.

This dissertation lays the foundation for combining the genetic engineering of HSCs by viral transduction to generate iNKT cells to be used for cancer immunotherapy. Proof-of- principle experiments in mice and an expansion to the use of the humanized mouse model provided the necessary knowledge and tools for further development to be pursued. Ongoing and future studies aim to demonstrate anti-cancer efficacy in humanized mouse models in order to collect data for an application to utilize these HSC-engineered iNKT cells in a cancer immunotherapy clinical trial.

iii The dissertation of Drake John Smith is approved.

Lili Yang

Yvonne Y. Chen

Arnold I. Chin

Anna Wu Work

Donald Barry Kohn, Committee Chair

University of California, Los Angeles

2017

iv DEDICATION

This dissertation is dedicated to my grandfather,

John David Quinn, whom we lost to cancer in 2015.

v TABLE OF CONTENTS

Abstract of the Dissertation ii

Committee Page iv

Dedication Page v

List of Figures and Tables vii

Acknowledgements ix

Vita xii

Chapter 1: Engineering Immunotherapy 1

References 13

Chapter 2: Genetic engineering of hematopoietic stem cells to generate invariant 16 natural killer T cells (Smith et al., PNAS 2015)

References 22

Chapter 3: Propagating humanized BLT mice for the study of human immunology 26 and immunotherapy (Smith et al., Stem Cells and Dev 2016)

References 36

Chapter 4: Invariant natural killer T cells generated in vivo from genetically 38 engineered human hematopoietic stem cells (Smith et al., manuscript in preparation)

References 58

Chapter 5: Conclusions and future studies 60

References 63

vi LIST OF FIGURES AND TABLES

Chapter 1:

No figures.

Chapter 2:

Figure 2-1: Cloning of invariant natural killer T-cell receptor (iNKT TCR) 18

genes and construction of retroviral delivery vectors

Figure 2-2: Generation of functional iNKT cells through TCR gene 19

engineering of hematopoietic stem cells (HSCs)

Figure 2-3: Development of the HSC-engineered iNKT cells. B6-miNKT 20

and control B6-mock mice were analyzed for iNKT cell

development at 6–8 wk post HSC transfer

Figure 2-4: Protection from melanoma lung metastasis by the 21

HSC-engineered iNKT cells

Figure S2-1: Titration of the miNKT retroviral vector 24

Figure S2-2: Lineage differentiation of iNKT TCR-engineered HSCs 25

Chapter 3:

Figure 3-1: Generation of propagated BLT (proBLT) mice through 30

secondary transfer of bone marrow cells and human thymus

implants from primary BLT mice to naïve NSG mice

Figure 3-2: Reconstitution of multilineage human immune cells in proBLT 32

mice

Figure 3-3: Reconstitution of human thymus and human T cells in proBLT 33

mice

vii Figure 3-4: Inheritance of human immune cell genetic traits from BLT to 34

proBLT mice

Figure 3-5: Persistence of human immune cell gene modifications from 35

BLT to proBLT mice

Chapter 4:

Figure 4-1: Cloning of invariant natural killer T (iNKT) cell T cell 45

receptor (TCR) genes and construction of lentiviral delivery

vectors

Figure 4-2: Generation of functional iNKT cells through TCR gene 47

engineering of hematopoietic stem cells (HSCs)

Figure 4-3: Development of genetically engineered HSCs and iNKT cells 50

Figure 4-4: In vitro functionality of engineered iNKT cells 52

Chapter 5:

No figures.

viii ACKNOWLEDGEMENTS

First and foremost, I would like to thank my graduate mentor Dr. Lili Yang. She took a chance on me as her first graduate student as a biochemist who thought he could be an immunologist. She taught me everything I know about immunotherapy and her tireless pursuit in the name of science inspired me to push myself beyond my own limits as well. With her help, I have been able to meet and collaborate with some of the most influential minds in the field and I strive to have as great of an impact on the field as she has had in the years to come.

I want to retroactively thank my undergraduate mentor Dr. Joanie Hevel. She also took a chance on me as a freshman when I was straight off the farm and convinced I wanted to pursue a career in research. She taught me how to be a research scientist and to stay tenacious even when it feels like nothing is working. Her words of wisdom have aided in sustaining me through many weeks of negative results; her margarita recipe may have helped on occasion as well.

I need to thank all of the members of the Yang lab who I have worked with over the past few years. Their daily support of my research and training have made the pace of my progress possible. I would like to specifically thank my undergraduate student Levina Lin. She quickly became like a second pair of hands and was never afraid to call me out when I’d made a mistake, much to her enjoyment.

I must also thank the army of collaborators that I have been so fortunate to work with during my time at UCLA. Without their expertise and dedication, none of this work would have been possible. I want to specifically thank Jessica Scholes and Felicia Codrea of the BSCRC

FACS Core for mentoring me in the art of flow cytometry when I was a newly minted graduate student. I also wish to thank Debbie Posner of the CFAR Virology Core not only for her support of my research through providing and processing blood products, but for her support of my

ix personal journey through my PhD. Thank you for the words of advice and encouragement on my bad days.

I would like to thank the members of my committee for their critical evaluation of my research and their insightful suggestions that have helped me along the way. The advice I have received has aided in directing my research as well my career and life goals. Special thanks go to the chair of my committee, Don Kohn, who agreed to take me on for an early summer lab rotation when I first came to UCLA. I have seen his exemplary mentorship change the lives of many, including my own, and I hope to one day be able to do the same.

I have had the pleasure of serving as the President of the Biological Sciences Council for the past two years of my PhD. This has given me the opportunity to work with many motivated scientists who want to help make UCLA a better place for graduate students both now and in the future. Good luck to them as they take on the torch of slogging through bureaucracy.

Thank you to my friends for their continued support over the past four years, even when I wouldn’t see them for months at a time. Thanks to Brendan Barry for helping me get through my first year of graduate school and beyond as well as critical reading of this dissertation. Special thanks go to my first graduate mentor at UCLA, Eric Gschweng, who continually pushed me to be a better scientist and still does to this day. I strive to one day live up to his expectations and his example. Extra special thanks go to my best friend, Daniel

Sivalingam, who I could not have made it through graduate school without; thank you for dragging me away from the lab every once in a while to enjoy the great outdoors of California.

An eternal thank you goes to my family who has always supported my devotion to science and education: my father for instilling in me the value of a hard day’s labor and taking pride in your work, my mother for always being there with a kind word and some pertinent

x advice, my brother for nipping at my heels and keeping me motivated to set an example, and my sister for teaching me to embrace myself and relax a little sometimes.

Finally I thank my fiancée, Heesung Moon, who has supported me both in and outside of the lab during my PhD. Her unwavering dedication to her education is inspiring and has kept me going even when I wanted to quit. She has put up with my many late nights in lab and patiently taught me some much needed perspective. I’m beyond excited to begin the next part of our lives together, wherever it may be.

xi VITA

EDUCATION Doctorate (Ph.D.), Molecular Biology, University of California, Los Angeles, July 2013 - August 2017 (expected)

Bachelor of Science (B.S.), Biochemistry, Utah State University, August 2009 - May 2013

RESEARCH EXPERIENCE University of California, Los Angeles, Dr. Lili Yang, January 2014 - present Graduate Student Researcher • Utilized viral-mediated genetic engineering of stem cells to generate disease-targeting immune cells • Characterized functionality of engineered cells by flow cytometry and cytokine production • Assayed anti-cancer potential of engineered cells by xenograft tumor challenges • Led research team in performing pre-IND study of a new gene- and cell- based cancer immunotherapy treatment

Utah State University, Dr. Joanie Hevel, November 2009 - May 2013 Undergraduate Research Assistant • Utilized molecular cloning techniques to create bacterial expression vectors for proteins • Over-expressed protein in E. coli bacteria and purified using affinity chromatography • Performed both qualitative and quantitative radioactive assays to determine enzymatic activities of purified recombinant proteins • Introduced mutations into proteins and assayed the effects on enzymatic activity

PUBLICATIONS • Bo Li, Xi Wang, In Young Choi, Yu-Chen Wang, Siyuan Liu, Alexander T. Pham, Heesung Moon, Drake J. Smith, Dinesh S. Rao, Mark P. Boldin, and Lili Yang. “miR- 146a Modulates Autoreactive Th17 Cell Differentiation and Regulates Organ-Specific Autoimmunity” Journal of Clinical Investigation, September 5, 2017. Epub ahead of print. • Smith, Drake J., Levina J. Lin, Heesung Moon, Alexander T. Pham, Xi Wang, Siyuan Liu, Sunjong Ji, Valerie Rezek, Deanna M. Janzen, Sanaz Memarzadeh, Scott Kitchen, Dong Sung An, and Lili Yang. “Propagating Humanized BLT Mice for the Study of Human Immunology and Immunotherapy” Stem Cells and Development, December 15, 2016, 25(24): 1863-1873. doi:10.1089/scd.2016.0193. • Smith, Drake J., Siyuan Liu, Sunjong Ji, Bo Li, Jami McLaughlin, Donghui Cheng, Owen N. Witte, and Lili Yang. “Genetic Engineering of Hematopoietic Stem Cells to Generate Invariant Natural Killer T Cells” Proceedings of the National Academy of Sciences, January 20, 2015. doi:10.1073/pnas.1424877112.

xii AFFILIATIONS • American Society of Gene & Cell Therapy, 2016 - present Attend and present at annual meetings • Southern California Flow Cytometry Association, 2014 - present Attend and present at annual meetings • Engineering Immunity Consortium, 2014 - present Attend and present at annual meetings

AWARDS / HONORS / FUNDING • Midwinter Conference of Immunologists at Asilomar Ray Owen Poster Award, February 1, 2017 • University of California, Los Angeles Osslund Fellowship, April 2016 – July 2017 • University of California, Los Angeles Tumor Immunology Training Grant, July 2015 – July 2017 • Utah State University Maeser-Bauer Outstanding Graduating Senior in Biochemistry, April 22, 2013 • Utah State University Undergraduate Research Award, April 22, 2013 • National American Heart Association Summer Research Fellowship, June 2012 - August 2012 • Utah State University Undergraduate Research and Creative Opportunities Grant, January 2012 - May 2012 • National American Heart Association Summer Research Fellowship, June 2011 - August 2011 • Utah State University Smart Grant, January 2011 - May 2011 • Utah State University Dean’s List, January 2011 - May 2011 • Utah State University Dean’s List, August 2010 - December 2010 • Utah State University Presidential Scholarship, August 2009 - May 2013

SERVICE • University of California, Los Angeles Biological Sciences Council o President, 2015 - 2017 Chair Council meetings, represent the Council in campus affairs, and relay information to and from the Graduate Student’s Association, other Councils, and campus administrators. o Representative, 2013 - 2015 Represented the interests of graduate student constituents and served as a voting member on the Council. • University of California, Los Angeles Undergraduate Mentoring o Representative, 2015 - 2017 Serve on panels to educate undergraduate students about graduate school and provide advice about applications and how to succeed in graduate school. • University of California, Los Angeles Graduate Program in Bioscience Recruitment o Representative, 2014 - 2017 Escort and interview applicants to the graduate program and answer questions and address concerns from the applicants as a representative of UCLA graduate students.

xiii

CHAPTER 1:

Engineering Immunotherapy

1 A brief introduction to cancer immunotherapy

The field of cancer immunotherapy has seen a phase of explosive growth over the past few years and was even Science Magazine’s “Breakthrough of the Year” in 2013 [1]. However, the concept of harnessing the immune system to battle cancer is by no means a new concept. In

1891, Dr. William Coley, now known as the “Father of Immunotherapy,” began injecting pathogenic bacteria into the tumors of cancer patients after observing several cases of cancer patients going into spontaneous remission after developing bacterial skin infections [2]. While durable, complete remission was achieved in several types of malignancies utilizing this treatment, the lack of a known mechanism of action at the time and the inherent risk involved with injecting patients with pathogenic bacteria led to oncologists adopting surgery and radiotherapy as standard treatments. Almost a century later in 1976, bacteria would be utilized to treat cancer again when the tuberculosis vaccine “Bacille Calmette-Guérin” (BCG) was used to prevent recurrence of bladder cancer to great success and is even still in use today [3]. It is fascinating to imagine how much more advanced the field of immunotherapy might be today if these treatments had been continued in Dr. Coley’s time.

The year 1976 also saw the identification of the immunomodulatory cytokine interleukin 2 (IL-2), which less than a decade later was shown to be effective in treating cancer when administered to patients [4, 5]. Another decade later in the 1990s, the US Food and Drug

Administration (FDA) approved IL-2 for the treatment of multiple cancer types. There are also several other immunomodulatory molecules that have been utilized in the treatment of cancer to date. These molecules have the benefit of potently and directly activating the immune system, but can have devastating side effects in the form of inducing autoimmune reactions.

2 The 1990s also saw the cloning of the first tumor antigen, MZ2-E from melanoma, which allowed the development and eventual implementation of dendritic cell cancer vaccines

[6, 7]. The ability to manipulate tumor antigens is regarded as the “molecular divide” between the old methodologies of simply utilizing tumor immunology to treat cancer and the new methodologies of actually designing tumor immunology to treat cancer. Engineered antibodies designed to treat cancer by targeting a molecule on the surface of cancer cells and either directly inducing cell death via apoptosis or indirectly killing by recruiting immune cells to attack the tumor also came about during this time [8]. The first FDA approval for these types of antibodies came in 1997 for the treatment of non-Hodgkin’s lymphoma.

Current day immunotherapy still sees the development and continued use of these various treatments, but there are two techniques in particular that have gained considerable attention and resulted in the recognition of cancer immunotherapy as the “Breakthrough of the

Year” by Science Magazine in 2013. The first technique utilizes engineered antibodies deemed

“checkpoint inhibitors” that block the down-regulation of T cells [9-13]. By blocking this down- regulation, the T cells retain their activated status and are able to kill tumor cells. These treatments have the benefit of acting directly on immune cells and are not tumor-dependent, allowing them to be used for multiple cancer types. They do pose inherent risks to induce autoimmune reactions if the cells become overactive, but these side effects appear to be less severe and less common than previous immunomodulatory molecules. The second technique focuses on engineering T cells to increase their anti-tumor capacity [14-16]. This engineering of

T cells is mediated by viral vectors that introduce new genetic sequences into the T cell genome.

These genetic sequences code for either a chimeric antigen receptor (CAR) or T cell receptor

3 (TCR) that is specific for a target antigen expressed by cancer cells. These engineered receptors grant the T cells specificity for the tumor cells and allow them to target and kill.

Current T cell cancer immunotherapy: risks and limitations

In current T cell cancer immunotherapy blood is first drawn from the patient and lymphocytes are isolated. The cells are then stimulated before being transduced with a viral vector carrying the CAR or TCR gene [17-21]. The cells are subsequently expanded and then administered back to the patient. There are many variations and additions upon this protocol, but the core method is the same.

While this procedure has seen great success in treating and even curing some patients, it does come with its own set of risks and limitations. A major limitation comes in the form of the CAR or TCR targeting: this engineering allows the T cells to effectively kill the cancer cells expressing the target antigen, but the cells in a tumor are heterogeneous and will not all necessarily express this antigen [22, 23]. The cancer cells that down-regulate or lack the expression of the target antigen can escape the therapy causing the patient to relapse with a new variant of the cancer that is now refractory to the initial immunotherapy treatment.

Another limitation seen in some patients relates to long-term efficacy. A patient will receive treatment and exhibit cancer cell killing and tumor shrinkage, but then the tumor will rebound and the cancer will progress [24-26]. This failure to sustain a durable response could be caused by many factors (such as the antigen escape mentioned above), but a potential culprit specific to immunotherapy lies in the ex vivo expansion of the T cells in preparation to be administered. This expansion drives the differentiation of the T cells into effector T cells (TEFF).

While these TEFF cells are able to mediate efficient cell killing, they are not long-lived [27]. This

4 could explain the initial tumor killing response, followed by a relapse when the TEFF cells become exhausted and can no longer kill effectively.

There are several risks associated with this therapy as well, most stemming from either the viral transduction procedure or the therapeutic gene itself. The viral vector that is used to introduce the therapeutic gene into the T cell genome utilizes internal promoters to drive the expression of the new gene [28]. If this genome insertion occurs near an oncogene the altered expression could lead to the transduced T cell becoming cancerous itself [29, 30]. However, the viral vectors currently used in therapy have been vastly improved since their initial use in patients and new data suggest the risk of cancerous transformation is very low [31].

There are also risks associated with the products of the CAR or TCR genes. These risks arise from the possibility of off-target effects mediated by the therapeutic receptors. An example of these off-target effects manifests when an immunotherapy treatment for melanoma also causes skin rashes and loss of hair coloration. This occurs when the engineered T cells target the antigen presented by the cancer cells, but the healthy tissues in the patient also express this antigen and are can be targeted. These off-target effects can also occur when an antigen unrelated to the cancerous tissue just happens to be recognized by the therapeutic receptor and is targeted.

The engineered T cells already expressed an endogenous TCR with unknown specificity before being transduced and, with the activation mediated by the therapeutic receptor, the endogenous TCR can mediate off-target effects itself as well [32]. The endogenous TCR expression can have additional negative repercussions in TCR immunotherapy. This occurs when the alpha or beta chain of the therapeutic TCR mispairs with the alpha or beta chain of the endogenous TCR [33, 34]. This mispairing will decrease the number of TCRs targeting the

5 cancer antigen and thus lower the efficacy of the therapy, but of greater concern is the unpredictable specificity of the mispaired TCR leading to potential off-target effects.

A brief introduction to invariant natural killer T cells

Invariant natural killer T (iNKT) cells are a subset of conventional αβ T cells that utilize a semi-invariant T cell receptor and express select natural killer (NK) cell markers on their surface (CD161 in humans and NK1.1 in mice) [35]. The alpha chain (Vα24-Jα18 in humans and Vα14-Jα18 in mice) pairs with a limited selection of beta chains (Vβ11 in humans and Vβ8 and Vβ7 in mice) to form the semi-invariant TCR [36]. This invariant alpha chain and heavily biased beta chain pairing causes the TCR to recognize a specific set of antigen complexes. Unlike conventional TCRs that recognize peptide antigens presented by major histocompatibility complex (MHC) molecules, the iNKT cell TCR recognizes sphingolipid antigens presented by CD1d molecules [37]. CD1d is a non-polymorphic, MHC class I-like molecule expressed on antigen presenting cells (APC) and is responsible for presenting lipid antigen to T cells [38].

When an iNKT cell binds a sphingolipid antigen presented by a CD1d molecule on the surface of an APC it becomes activated and begins releasing massive amounts of cytokines [28].

This activation and cytokine release occurs on a much faster time scale than conventional T cells

(peaking at 3 days versus 7 days) and to a much higher degree (100 times more cytokine than conventional T cells). These iNKT cells are capable of releasing both Th1 cytokines, like IFN-γ, and Th2 cytokines, like IL-4 allowing multi-faceted immune cell activation [39]. While iNKT cells do have a limited ability to directly kill target cells, their major function is to act as master regulators and release cytokines to modulate the immune system.

6 Invariant natural killer T cells for cancer immunotherapy

Their ability to boost the activation of the immune system make iNKT cells a prime target for use in cancer immunotherapy treatments. They possess certain advantages over conventional CAR and TCR therapies as well. One such advantage is granted through their TCR binding interaction. Conventional T cell TCRs bind antigen presented by MHC molecules, which are restricted by human leukocyte antigen (HLA) matching. A new TCR must be designed in order for the therapy to be used in patients with varying HLA types. An iNKT cell TCR recognizes antigen presented by CD1d molecules, which are non-polymorphic [37]. The same iNKT cell TCR construct can be used in all patients regardless of their HLA type and binding and activation will still occur.

A conventional CAR or TCR therapy is designed against a specific cancer antigen.

This grants the engineered T cells the enhanced ability to kill tumor cells expressing this antigen, but leaves them unable to kill other tumor cells that lack this antigen. A specific therapeutic CAR or TCR can also only be used to treat the cancer that it was designed to recognize and must be redesigned and tested in order to recognize other cancer antigens. Cancer immunotherapy utilizing iNKT cells is not subject to these same restrictions. The therapeutic activity of iNKT cells is mediated through their activation of other immune effector cells, meaning the same treatment can be applied across multiple cancer types without the need for modification to the

TCR construct [40-43].

The major immune effectors cells activated against cancer by iNKT cells are NK cells and T cells [42, 44]. The activation of both of these cell types is crucial in order to mediate complete clearance of cancer in a patient. The cancer cells in a tumor are heterogeneous; the cells which present antigen are susceptible to T cell targeting and killing and the cells which

7 escape antigen targeting by downregulating the expression of their MHC class I molecules are susceptible to NK cell killing. An iNKT cell immunotherapy is able to activate both of these effector cell classes and mediate the clearance of multiple cancer cell types in a tumor thereby lowering the likelihood of a relapse.

In in vivo bioluminescent tracking experiments performed in our lab at UCLA (data not shown) we were able to demonstrate that iNKT cells are able to traffic to a tumor site after adoptive cell transfer. This tumor trafficking capacity of iNKT cells has also been observed in analysis of tumor samples from patients that had received iNKT cell immunotherapy[45]. The mechanism by which iNKT cells are able to achieve this tumor trafficking is unknown, but an increased expression of chemokine receptors on their cell surface is thought to likely play a role

[35]. This tumor homing is key in the use of iNKT cells for cancer immunotherapy. The iNKT cells home to the tumor microenvironment where the release of immunosuppressive cytokines by tumor-associated cells prevents killing by immune effector cells. A dendritic cell (DC) vaccine loaded with a potent synthetic antigen for iNKT cells, α-Galactosylceramide (α-GalCer), is injected and is able to bind and activate the iNKT cells [41]. The characteristic prolific release of inflammatory cytokines by the iNKT cells should overcome the immunosuppressive tumor microenvironment and render the tumor more susceptible to infiltration and killing by immune effector cells.

Current invariant natural killer T cell immunotherapy

Invariant natural killer T cells have already been utilized in several cancer immunotherapy clinical trials to date [41, 43, 46, 47]. One technique used to utilize iNKT cells for immunotherapy is through activation by a dendritic cell vaccine. The potent iNKT cell ligand

8 α-GalCer is loaded onto DCs and then injected into a patient to bind and activate the endogenous iNKT cells leading to the up-regulation of the immune system to fight cancer. The use of α-

GalCer has already been clinically approved and significantly lowers the barrier for future iNKT cell immunotherapy treatments. Another technique to utilize iNKT cells for therapy involves isolating lymphocytes from a patient’s blood and then expanding the iNKT cells ex vivo using the α-GalCer ligand [48]. The expanded iNKT cells are then infused back into the patient to fight the cancer. There have been select therapies that utilized both of these techniques in combination, some even including additional immune cell effector components.

It is important to note that the traditional method of TCR transduction of peripheral T cells is not applicable to iNKT cell immunotherapy. Simply transducing a T cell with an iNKT cell TCR will not grant it the desired therapeutic characteristics of an iNKT cell; these abilities are gained in the cell’s development in the thymus [35, 49]. While the iNKT cell immunotherapy treatments used to date have been well tolerated and some have shown some clinical efficacy, the results have fallen short of the perceived therapeutic potential [41]. The belief in the iNKT cell field is that the lack of clinical efficacy is due in part to the low number and high variability of iNKT cells in humans (0.001%-1% in peripheral blood) and their rapid depletion after stimulation [50]. There are not enough iNKT cells at the basal level to induce significant activation and propagate a meaningful immune response to fight the tumor [51].

Engineering stem cells for immunotherapy

A new addition to cellular immunotherapy, which recently received FDA-approval, was proposed in a clinical trial application by the lab of Dr. Antoni Ribas at UCLA. This therapy combines the canonical transduction of T cells with a therapeutic TCR and adds another

9 component by also transducing hematopoietic stem cells (HSC) with the therapeutic TCR [52].

The patient is treated with granulocyte colony-stimulating factor (G-CSF) and Plerixafor to mobilize the HSCs from the bone marrow into circulation in the peripheral blood [53]. Blood is collected from the patient and the stem cells are isolated. The cells are cultured very briefly ex vivo during which time they are transduced with the therapeutic TCR. The patient then receives a conditioning regiment to “make space” in the bone marrow niche before the transduced stem cells are infused [54].

This combinatorial therapy aims to address some of the long-term efficacy issues that have been observed with single treatment ex vivo expanded T cell therapies. The stem cells will home to the bone marrow and engraft to repopulate after the conditioning regiment [55]. The transduced stem cells are still able to differentiate without bias and generate the multiple lineages of functioning immune cells, as the therapeutic TCR cannot be expressed on the surface without the co-expression of CD3. When a transduced cell does commit to becoming a T cell, though, the therapeutic TCR is already being produced and is able to induce allelic exclusion and prevent the expression of an endogenous TCR [52, 56, 57]. This is an advantage over peripheral T cell transduction methods as it removes the complications that arise from the co-expression of non- therapeutic TCR chains. The transduced stem cells and their progeny will continue to generate new therapeutic T cells for the lifetime of the patient. These newly produced T cells will not be terminally differentiated from an ex vivo expansion protocol, as the T cells in traditional immunotherapies are, and should be able to exhibit efficacy over a longer time period.

10 Stem cell engineered iNKT cells for immunotherapy

The studies conducted by our lab aim to combine the therapeutic potential of iNKT cells with the advantages of transduced stem cell therapies. As mentioned previously, it is believed that the low number of circulating iNKT cells in a patient and their depletion after stimulation are the major hurdles to overcome in order to provoke meaningful iNKT cell immune responses against cancer [35, 50, 51]. As a subset of conventional αβ T cells, iNKT cells are also generated by HSCs in the bone marrow and educated in the thymus [39, 58]. HSCs would be transduced with a viral vector containing the alpha and beta chains of an iNKT cell TCR and then be infused back into a preconditioned patient, much like the Ribas lab clinical protocol. The

HSCs home to the bone marrow and are able to differentiate and lead to the typical immune cell lineages as the iNKT cell TCR is also unable to be expressed on the cell surface without the co- expression of CD3. When a transduced cell does commit to becoming a T cell, the therapeutic iNKT cell TCR has already been expressed and is able to induce allelic exclusion and prevent the rearrangement and expression of endogenous TCR alpha or beta chains [52, 56]. When the transduced T cell precursor enters the thymus the iNKT cell TCR expression on the cell surface is able to initiate the programming of the cell to become an iNKT cell and gain the desired therapeutic characteristics [35, 39].

The proportion of stem cells that are transduced with the viral vector can be titered to reach the desired level of circulating iNKT cells. This ability to manipulate the level of generated iNKT cells is critical as too few cells will not be able to propagate a meaningful immune response and too many cells will reduce the compartment of conventional T cells which will ultimately mediate the actual tumor cell killing. This ability to titer the level of iNKT cells is important from a safety standpoint as well as over activation could lead to an autoimmune

11 reaction. The basal number of iNKT cells only needs to be increased to a level which allows an activation to propagate and lead to a meaningful immune response. In essence our proposed stem cell transduction immunotherapy would prime a patient to then undergo iNKT cell activation therapy utilizing a dendritic cell vaccine loaded with the α-GalCer ligand. Chapter 2 details the proof-of-principle experiments performed in mice that allowed the foundation for further experiments and clinical reasoning to be made.

12 References

1. Couzin-Frankel, J., Breakthrough of the year 2013. Cancer immunotherapy. Science, 2013. 342(6165): p. 1432-3. 2. Decker, W.K. and A. Safdar, Bioimmunoadjuvants for the treatment of neoplastic and infectious disease: Coley's legacy revisited. Cytokine Growth Factor Rev, 2009. 20(4): p. 271-81. 3. Morales, A., D. Eidinger, and A.W. Bruce, Intracavitary Bacillus Calmette-Guerin in the treatment of superficial bladder tumors. J Urol, 1976. 116(2): p. 180-3. 4. Rosenberg, S.A., et al., Observations on the systemic administration of autologous lymphokine-activated killer cells and recombinant interleukin-2 to patients with metastatic cancer. N Engl J Med, 1985. 313(23): p. 1485-92. 5. Morgan, D.A., F.W. Ruscetti, and R. Gallo, Selective in vitro growth of T lymphocytes from normal human bone marrows. Science, 1976. 193(4257): p. 1007-8. 6. van der Bruggen, P., et al., A gene encoding an antigen recognized by cytolytic T lymphocytes on a human melanoma. Science, 1991. 254(5038): p. 1643-7. 7. Finn, O.J., Human tumor immunology at the molecular divide. J Immunol, 2007. 178(5): p. 2615-6. 8. Rudnicka, D., et al., Rituximab causes a polarization of B cells that augments its therapeutic function in NK-cell-mediated antibody-dependent cellular cytotoxicity. Blood, 2013. 121(23): p. 4694-702. 9. Leach, D.R., M.F. Krummel, and J.P. Allison, Enhancement of antitumor immunity by CTLA-4 blockade. Science, 1996. 271(5256): p. 1734-6. 10. Dong, H., et al., Tumor-associated B7-H1 promotes T-cell apoptosis: a potential mechanism of immune evasion. Nat Med, 2002. 8(8): p. 793-800. 11. Wolchok, J.D., et al., Nivolumab plus ipilimumab in advanced melanoma. N Engl J Med, 2013. 369(2): p. 122-33. 12. Topalian, S.L., et al., Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N Engl J Med, 2012. 366(26): p. 2443-54. 13. Sznol, M. and L. Chen, Antagonist antibodies to PD-1 and B7-H1 (PD-L1) in the treatment of advanced human cancer. Clin Cancer Res, 2013. 19(5): p. 1021-34. 14. June, C., et al., T-cell therapy at the threshold. Nat Biotechnol, 2012. 30(7): p. 611-4. 15. Pardoll, D.M., Immunology beats cancer: a blueprint for successful translation. Nat Immunol, 2012. 13(12): p. 1129-32. 16. Restifo, N.P., M.E. Dudley, and S.A. Rosenberg, Adoptive immunotherapy for cancer: harnessing the T cell response. Nat Rev Immunol, 2012. 12(4): p. 269-81. 17. Eshhar, Z., et al., Specific activation and targeting of cytotoxic lymphocytes through chimeric single chains consisting of antibody-binding domains and the gamma or zeta subunits of the immunoglobulin and T-cell receptors. Proc Natl Acad Sci U S A, 1993. 90(2): p. 720-4. 18. Porter, D.L., et al., Chimeric antigen receptor-modified T cells in chronic lymphoid leukemia. N Engl J Med, 2011. 365(8): p. 725-33. 19. Robbins, P.F., et al., Tumor regression in patients with metastatic synovial cell sarcoma and melanoma using genetically engineered lymphocytes reactive with NY-ESO-1. J Clin Oncol, 2011. 29(7): p. 917-24.

13 20. Brentjens, R.J., et al., Safety and persistence of adoptively transferred autologous CD19- targeted T cells in patients with relapsed or chemotherapy refractory B-cell leukemias. Blood, 2011. 118(18): p. 4817-28. 21. Pule, M.A., et al., Virus-specific T cells engineered to coexpress tumor-specific receptors: persistence and antitumor activity in individuals with neuroblastoma. Nat Med, 2008. 14(11): p. 1264-70. 22. Zellmer, V.R. and S. Zhang, Evolving concepts of tumor heterogeneity. Cell Biosci, 2014. 4: p. 69. 23. Meacham, C.E. and S.J. Morrison, Tumour heterogeneity and cancer cell plasticity. Nature, 2013. 501(7467): p. 328-37. 24. Kelderman, S., T.N. Schumacher, and J.B. Haanen, Acquired and intrinsic resistance in cancer immunotherapy. Mol Oncol, 2014. 8(6): p. 1132-9. 25. Sharma, P., et al., Primary, Adaptive, and Acquired Resistance to Cancer Immunotherapy. Cell, 2017. 168(4): p. 707-723. 26. Allen, C.T., et al., Anti-Tumor Immunity in Head and Neck Cancer: Understanding the Evidence, How Tumors Escape and Immunotherapeutic Approaches. Cancers (Basel), 2015. 7(4): p. 2397-414. 27. Gattinoni, L., et al., T memory stem cells in health and disease. Nat Med, 2017. 23(1): p. 18-27. 28. Astrakhan, A., et al., Ubiquitous high-level gene expression in hematopoietic lineages provides effective lentiviral gene therapy of murine Wiskott-Aldrich syndrome. Blood, 2012. 119(19): p. 4395-407. 29. Check, E., A tragic setback. Nature, 2002. 420(6912): p. 116-8. 30. Hacein-Bey-Abina, S., et al., A serious adverse event after successful gene therapy for X- linked severe combined immunodeficiency. N Engl J Med, 2003. 348(3): p. 255-6. 31. Carbonaro, D.A., et al., Preclinical demonstration of lentiviral vector-mediated correction of immunological and metabolic abnormalities in models of adenosine deaminase deficiency. Mol Ther, 2014. 22(3): p. 607-22. 32. Brownlie, R.J. and R. Zamoyska, T cell receptor signalling networks: branched, diversified and bounded. Nat Rev Immunol, 2013. 13(4): p. 257-69. 33. Thomas, S., H.J. Stauss, and E.C. Morris, Molecular immunology lessons from therapeutic T-cell receptor gene transfer. Immunology, 2010. 129(2): p. 170-7. 34. Shao, H., et al., TCR mispairing in genetically modified T cells was detected by fluorescence resonance energy transfer. Mol Biol Rep, 2010. 37(8): p. 3951-6. 35. Bendelac, A., P.B. Savage, and L. Teyton, The biology of NKT cells. Annu Rev Immunol, 2007. 25: p. 297-336. 36. Lee, P.T., et al., Distinct functional lineages of human V(alpha)24 natural killer T cells. J Exp Med, 2002. 195(5): p. 637-41. 37. Brigl, M. and M.B. Brenner, CD1: antigen presentation and T cell function. Annu Rev Immunol, 2004. 22: p. 817-90. 38. Castano, A.R., et al., Peptide binding and presentation by mouse CD1. Science, 1995. 269(5221): p. 223-6. 39. Godfrey, D.I. and S.P. Berzins, Control points in NKT-cell development. Nat Rev Immunol, 2007. 7(7): p. 505-18. 40. Crowe, N.Y., et al., Differential antitumor immunity mediated by NKT cell subsets in vivo. J Exp Med, 2005. 202(9): p. 1279-88.

14 41. Fujii, S., et al., NKT cells as an ideal anti-tumor immunotherapeutic. Front Immunol, 2013. 4: p. 409. 42. Vivier, E., et al., Targeting natural killer cells and natural killer T cells in cancer. Nat Rev Immunol, 2012. 12(4): p. 239-52. 43. Yamasaki, K., et al., Induction of NKT cell-specific immune responses in cancer tissues after NKT cell-targeted adoptive immunotherapy. Clin Immunol, 2011. 138(3): p. 255- 65. 44. Pilones, K.A., J. Aryankalayil, and S. Demaria, Invariant NKT cells as novel targets for immunotherapy in solid tumors. Clin Dev Immunol, 2012. 2012: p. 720803. 45. Nagato, K., et al., Accumulation of activated invariant natural killer T cells in the tumor microenvironment after alpha-galactosylceramide-pulsed antigen presenting cells. J Clin Immunol, 2012. 32(5): p. 1071-81. 46. Motohashi, S., et al., A phase I study of in vitro expanded natural killer T cells in patients with advanced and recurrent non-small cell lung cancer. Clin Cancer Res, 2006. 12(20 Pt 1): p. 6079-86. 47. Motohashi, S., et al., Anti-tumor immune responses induced by iNKT cell-based immunotherapy for lung cancer and head and neck cancer. Clin Immunol, 2011. 140(2): p. 167-76. 48. Watarai, H., et al., Methods for detection, isolation and culture of mouse and human invariant NKT cells. Nat Protoc, 2008. 3(1): p. 70-8. 49. Engel, I. and M. Kronenberg, Transcriptional control of the development and function of Valpha14i NKT cells. Curr Top Microbiol Immunol, 2014. 381: p. 51-81. 50. Berzins, S.P., M.J. Smyth, and A.G. Baxter, Presumed guilty: natural killer T cell defects and human disease. Nat Rev Immunol, 2011. 11(2): p. 131-42. 51. Molling, J.W., et al., Low levels of circulating invariant natural killer T cells predict poor clinical outcome in patients with head and neck squamous cell carcinoma. J Clin Oncol, 2007. 25(7): p. 862-8. 52. Vatakis, D.N., et al., Antitumor activity from antigen-specific CD8 T cells generated in vivo from genetically engineered human hematopoietic stem cells. Proc Natl Acad Sci U S A, 2011. 108(51): p. E1408-16. 53. DiPersio, J.F., et al., Plerixafor and G-CSF versus placebo and G-CSF to mobilize hematopoietic stem cells for autologous stem cell transplantation in patients with multiple myeloma. Blood, 2009. 113(23): p. 5720-6. 54. Tarantal, A.F., et al., Nonmyeloablative Conditioning Regimen to Increase Engraftment of Gene-modified Hematopoietic Stem Cells in Young Rhesus Monkeys. Mol Ther, 2012. 20(5): p. 1033-1045. 55. Morrison, S.J., N. Uchida, and I.L. Weissman, The biology of hematopoietic stem cells. Annu Rev Cell Dev Biol, 1995. 11: p. 35-71. 56. Yang, L. and D. Baltimore, Long-term in vivo provision of antigen-specific T cell immunity by programming hematopoietic stem cells. Proc Natl Acad Sci U S A, 2005. 102(12): p. 4518-23. 57. Gschweng, E.H., et al., HSV-sr39TK positron emission tomography and suicide gene elimination of human hematopoietic stem cells and their progeny in humanized mice. Cancer Res, 2014. 74(18): p. 5173-83. 58. Haraguchi, K., et al., Recovery of Valpha24+ NKT cells after hematopoietic stem cell transplantation. Bone Marrow Transplant, 2004. 34(7): p. 595-602.

15

CHAPTER 2:

Genetic engineering of hematopoietic stem cells

to generate invariant natural killer T cells

16 Genetic engineering of hematopoietic stem cells to generate invariant natural killer T cells Drake J. Smitha,b, Siyuan Liua,b, Sunjong Jia,b, Bo Lia,b, Jami McLaughlinb, Donghui Chenga, Owen N. Wittea,b,c,1, and Lili Yanga,b,1 aEli and Edythe Broad Center of Regenerative Medicine and Stem Cell Research, bDepartment of Microbiology, Immunology and Molecular Genetics, and cHoward Hughes Medical Institute, University of California, Los Angeles, CA 90095

Contributed by Owen N. Witte, December 29, 2014 (sent for review December 8, 2014; reviewed by Paula Cannon, Si-Yi Chen, and Drew M. Pardoll)

Invariant natural killer T (iNKT) cells comprise a small population of iNKT T-cell receptor (TCR) transgenic mice (5, 6) and iNKT αβ T lymphocytes. They bridge the innate and adaptive immune induced pluripotent stem (iPS) cell-derived transgenic mice systems and mediate strong and rapid responses to many diseases, (7) provide valuable tools to study iNKT cell biology in mice, including cancer, infections, allergies, and autoimmunity. How- but these methods are both costly and time-consuming. In ad- ever, the study of iNKT cell biology and the therapeutic applica- dition, approaches using transgenic mice have no direct clinical tions of these cells are greatly limited by their small numbers in application. As an alternative, a TCR-engineered HSC adoptive transfer strategy could overcome these limitations and become vivo (∼0.01–1% in mouse and human blood). Here, we report a new method to generate large numbers of iNKT cells in mice clinically applicable. Since its demonstration in mice in the early through T-cell receptor (TCR) gene engineering of hematopoietic 2000s, this HSC-engineered T-cell strategy has been widely used to successfully generate both mouse and human antigen-specific stem cells (HSCs). We showed that iNKT TCR-engineered HSCs conventional αβ T cells in multiple mouse and humanized mouse could generate a clonal population of iNKT cells. These HSC-engi- models (8–13). Human clinical trials testing this strategy for neered iNKT cells displayed the typical iNKT cell phenotype and treating melanoma are also ongoing (14). Based on these previous functionality. They followed a two-stage developmental path, first works and the scientific principle that iNKT cells follow a “TCR in thymus and then in the periphery, resembling that of endoge- instruction” development path similar to that of conventional αβ nous iNKT cells. When tested in a mouse melanoma lung metasta- T cells (15), we hypothesized that HSCs could be engineered sis model, the HSC-engineered iNKT cells effectively protected INFLAMMATION to express iNKT TCR genes and be programmed to develop IMMUNOLOGY AND mice from tumor metastasis. This method provides a powerful into clonal iNKT cells. In the present report, we demonstrated and high-throughput tool to investigate the in vivo development the feasibility of this new HSC-engineered iNKT cell approach in and functionality of clonal iNKT cells in mice. More importantly, mice and provided evidence to support its therapeutic potential in this method takes advantage of the self-renewal and longevity of amousemelanomalungmetastasismodel. HSCs to generate a long-term supply of engineered iNKT cells, thus opening up a new avenue for iNKT cell-based immunotherapy. Results Cloning of iNKT TCR Genes and Construction of Retroviral Delivery genetic engineering | hematopoietic stem cells | HSCs | Vectors. We used a robust and high-throughput single-cell TCR invariant natural killer T cells | iNKT cells cloning technology recently established in our laboratories to obtain iNKT TCR genes (Materials and Methods). Single iNKT nvariant natural killer T (iNKT) cells are a small population of cells were sorted from mouse spleen cells using flow cytometry based on a stringent collection of surface markers gated as αβ T lymphocytes highly conserved from mice to humans. Like lo + + + I CD3 mCD1d/PBS-57 TCR Vβ8 NK1.1 (Fig. 1A)(15).mCD1d/ conventional αβ T cells, iNKT cells are derived from hemato- PBS-57 is the tetramer reagent that specifically identifies iNKT poietic stem cells (HSCs) and develop in the thymus. However, they differ from conventional T cells in several important Significance aspects, including their display of NK cell markers, their rec- ognition of glycolipid antigens presented by the nonclassical monomorphic major histocompatibility complex (MHC) mole- This article describes a new method for generating large numbers cule CD1d, and their expression of semi-invariant T-cell recep- of invariant natural killer T (iNKT) cells in mice through genetic tors (identical α chains paired with a limited selection of β engineering of blood stem cells. iNKT cells are potent immune cells chains) (1, 2). Despite their small numbers in vivo (∼0.1–1% in that regulate many human diseases,includingcancer,infections, mouse blood and ∼0.01–1% in human blood), iNKT cells have allergies, and autoimmunity. However, both the study of iNKT cell been suggested to play important roles in regulating many dis- biology and the clinical application of iNKT cells have been greatly – eases, including cancer, infections, allergies, and autoimmunity hindered by their small numbers (∼0.01 1% in mouse and human (3). When stimulated, iNKT cells rapidly release a large amount blood). The method reported here provides a powerful new tool to study iNKT cell biology in a mouse model. It can also be applied of effector cytokines like IFN- and IL-4, both as a cell population γ to humans, opening a new avenue for iNKT cell-based immuno- and at the single-cell level. These cytokines then activate various therapy that has the potential to provide patients with thera- immune effector cells, such as natural killer (NK) cells and den- peutic levels of iNKT cells throughout life. dritic cells (DCs) of the innate immune system, as well as CD4 helper and CD8 cytotoxic conventional αβ T cells of the adaptive Author contributions: D.J.S., O.N.W., and L.Y. designed research; D.J.S., S.L., S.J., B.L., J.M., D.C., immune system via activated DCs (3, 4). Because of their unique and L.Y. performed research; D.J.S. and L.Y. analyzed data; and D.J.S. and L.Y. wrote the paper. activation mechanism, iNKT cells can attack multiple diseases Reviewers: P.C., University of Southern California; S.-Y.C., University of Southern California; independent of antigen and MHC restrictions, making them at- and D.M.P., Johns Hopkins University School of Medicine. tractive universal therapeutic agents (3, 4). Notably, because of the The authors declare no conflict of interest. capacity of effector NK cells and conventional αβ T cells to spe- Freely available online through the PNAS open access option. cifically recognize diseased tissue cells, iNKT cell-induced immune 1To whom correspondence may be addressed. Email: [email protected] or owenwitte@ reactions result in limited off-target side effects (3, 4). mednet.ucla.edu. Restricted by their extremely low numbers, both the study of This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. iNKT cells and their clinical applications have been challenging. 1073/pnas.1424877112/-/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.1424877112 PNAS | February 3, 2015 | vol. 112 | no. 5 | 1523–1528 17 TCRs (16). We included TCR Vβ8 staining to focus on the sion of a high-affinity iNKT TCR was selected for the follow-up + dominant Vβ8 population of mouse iNKT cells (1, 2). The studies and was denoted as the miNKT vector (Fig. 1 D and F). sorted single iNKT cells were then subjected to TCR cloning The control MIG vector that encodes an EGFP reporter gene (Fig. 1B). Several verified iNKT TCR α and β pairs were inserted was denoted as the Mock vector (Fig. 1 D and F) (10). into the murine stem cell virus (MSCV)-based retroviral vector to yield TCR gene delivery vectors (Fig. 1 C and D). Their Generation of Clonal iNKT Cells Through Genetic Engineering of HSCs. vector-mediated expressions were then tested in 293.T/mCD3, Following an established protocol (10), we performed miNKT- a stable cell line that has been engineered to express mouse CD3 transduced bone marrow transfer in B6 mice to generate the molecules that are required to support the surface display of recipient mice denoted as B6-miNKT (Fig. 2A). In brief, HSC- mouse TCRs (Fig. 1E). One vector that mediated high expres- enriched bone marrow cells harvested from donor B6 mice were cultured in vitro, transduced with either Mock or miNKT ret- roviral vectors, then separately transferred into irradiated re- cipient B6 mice. The recipient mice were allowed to reconstitute their immune system over the course of 6–8 wk, followed by analysis to determine the presence of iNKT cells. Similar to the previously reported conventional αβ TCR engineering approach (10), we obtained desirable titers of the newly constructed miNKT retroviral vector [∼0.5–1 × 106 infectious units (IFU)/mL; Fig. S1] and achieved high efficiencies of HSC transduction (routinely over 50% of the cultured bone marrow cells). Com- pared with the Mock-engineered recipient mice, denoted as B6- Mock, we observed a significant increase of iNKT cells in the B6- miNKT mice from thymus to peripheral tissues, suggesting the successful generation of HSC-engineered iNKT cells (Fig. 2 B and C). Through titrating the miNKT vector-transduced HSCs used for bone marrow transfer, we were able to control the in- crease of the iNKT cells from as high as 50% of the total αβ T cells, down to a desired level in the B6-miNKT mice (Fig. 2 D and E). The ability to regulate the number of engineered iNKT cells can be valuable for clinical applications of this HSC- engineered iNKT cell strategy. Study of the iNKT cells from the B6-miNKT mice revealed that these iNKT cells displayed atypicalphenotypeofmouseiNKTcellsinthattheyexhibited high expression of the NK1.1 marker, as well as a memory T- cell signature (CD62LloCD44hi) and a CD4+CD8− or CD4− CD8− coreceptor expression pattern (Fig. 2F) (17). Almost all of these iNKT cells showed positive staining for TCR Vβ8, in- dicating that they expressed the transgenic clonal iNKT TCR and suggesting that they were derived from the miNKT-engi- neered HSCs (Fig. 2F). The production of high levels of iNKT cells in the B6-miNKT mice persisted for up to 6 mo following the initial bone marrow transfer and also post secondary bone marrow transfer, highlighting the long-term effectiveness of this HSC-engineered iNKT cell strategy (Fig. 2 G and H).

Functionality of the HSC-Engineered iNKT Cells. We then analyzed the functionality of the HSC-engineered iNKT cells. When stimulated with the agonist glycolipid α-Galactosylceramide (α-GalCer) in vitro, the engineered iNKT cells proliferated vig- orously by over 20-fold in 5 d and produced large amounts of the effector cytokines IFN-γ and IL-4 (Fig. 2 I–K). When B6-miNKT mice were immunized with bone marrow-derived dendritic cells Fig. 1. Cloning of invariant natural killer T-cell receptor (iNKT TCR) genes (BMDCs) loaded with α-GalCer, the engineered iNKT cells and construction of retroviral delivery vectors. (A) Single iNKT cells were mounted a strong and rapid response in vivo, expanding close to sorted out from mouse spleen cells using flow cytometry based on a strin- 20-fold in 3 d (Fig. 2L). Notably, the in vivo expansion of these + + gent collection of surface markers (gated as CD3lomCD1d/PBS-57 TCR Vβ8 cells peaked at day 3 postimmunization, compared with 7 d NK1.1hi). A representative FACS plot is shown. mCD1d/PBS-57 indicates the postimmunization for conventional αβ T cells (18). This speedy tetramer reagent that specifically stains mouse iNKT TCRs. (B) Sorted single in vivo response is a signature of iNKT cells (17). These results iNKT cells were subjected to TCR cloning using a single-cell RT-PCR approach. indicate that the HSC-engineered iNKT cells are fully functional. Representative DNA gel pictures are presented showing the TCR α and β chain gene PCR products from five iNKT cells. (C)Representativesequencing Development of the HSC-Engineered iNKT Cells. Next, we analyzed results confirming the cloned single-cell iNKT TCR α and β chain genes. (D) the development of the HSC-engineered iNKT cells. iNKT cell Schematic representation of the retroviral vectors encoding either a control progenitors gated as TCR lomCD1d/PBS-57+ were detected in EGFP reporter gene (denoted as the Mock vector), or a pair of iNKT TCR β α the thymus of the B6-miNKT mice and were found to follow a and β chain genes (denoted as the miNKT vector). LTR indicates long-term repeats; IRES, internal ribosome entry sites; EGFP, enhanced green fluores- classic developmental path similar to that observed for endoge- cence protein; F2A, foot-and-mouth disease virus 2A sequence; and WRE, nous iNKT progenitor cells in the control B6-Mock mice (Fig. 3) − − woodchuck responsive element. (E) Schematic representation of the 293.T (15). These progenitor cells appeared as CD4 CD8 (DN), + + + − cell line that has been engineered to stably express mouse CD3 genes and so CD4 CD8 (DP), and CD4 CD8 (CD4 SP), corresponding with as to support the surface display of mouse TCRs (denoted as 293.T/mCD3). (F) an iNKT development from DN to DP, then to CD4 SP or back Representative FACS plots showing the expression of clonal iNKT TCRs in to DN cells (Fig. 3A). The expression of CD24, CD44, and 293.T/mCD3 cells transduced with the chosen miNKT vector. DX5 markers on iNKT progenitor cells further defined their

1524 | www.pnas.org/cgi/doi/10.1073/pnas.1424877112 Smith et al. 18 INFLAMMATION IMMUNOLOGY AND

Fig. 2. Generation of functional iNKT cells through TCR gene engineering of hematopoietic stem cells (HSCs). B6 mice receiving adoptive transfer of HSCs transduced with either the Mock retroviral vector (denoted as B6-Mock mice) or miNKT retroviral vector (denoted as B6-miNKT mice) were allowed to re- constitute their immune system in a duration of 6–8 wk, followed by analysis. The experiments were repeated at least three times, and representative results are presented. iNKT cells were detected as TCRβlomCD1d/PBS-57+ using flow cytometry. (A) Schematic representation of the experimental design to generate HSC-engineered iNKT cells in mice. (B and C) Increase of iNKT cells in B6-miNKT mice compared with that in the control B6-Mock mice. (B) FACS plots showing the detection of iNKT cells in various tissues. (C) Bar graphs showing the fold increase of percent iNKT cells in the indicated tissues. (D and E) Control of iNKT cell numbers in B6-miNKT mice through titrating the miNKT vector-transduced HSCs used for adoptive transfer. (D) FACS plots showing the detection of iNKT + cells in the spleens of various B6-miNKT recipient mice. Tc indicates the conventional αβ T cells (gated as TCRβ mCD1d/PBS-57−). (E)Bargraphsshowingthe percent iNKT of total αβ T cells in spleen. (F) Phenotype of the HSC-engineered iNKT cells. FACS plots are presented showing the surface markers of iNKT cells detected in the liver of B6-miNKT mice. (G and H)Long-termproductionofHSC-engineerediNKTcells.FACSplotsarepresentedshowingthedetectionof iNKT cells in the spleen of B6-miNKT mice for up to 6 mo after initial HSC adoptive transfer (G) and at 2 mo after secondary bone marrow transfer (BMT) (H). (I–K) Functionality of the HSC-engineered iNKT cells tested in vitro. Spleen cells of B6-miNKT mice were cultured in vitro in the presence of α-GalCer (100 ng/mL). (I) FACS plots showing the time-course proliferation of iNKT cells. (J) FACS plots showing the cytokine production in iNKT cells on day 3, as measured by in- tracellular cytokine staining. (K) ELISA analysis of cytokine production in the cell culture medium at day 3. Data are presented as mean of duplicate cultures ± SEM, *P < 0.01 (B6-miNKT samples compared with the corresponding B6-Mock controls). (L) Functionality of the HSC-engineered iNKT cells tested in vivo. B6-Mock or B6-miNKT mice were given i.v. injection of 1 × 106 bone marrow-derived dendritic cells (BMDCs) loaded with α-GalCer (denoted as BMDC/ α-GalCer) and then periodically bled to monitor iNKT cell responses. FACS plots are presented showing the change of iNKT cell frequencies in blood.

Smith et al. PNAS | February 3, 2015 | vol. 112 | no. 5 | 1525 19 Fig. 3. Development of the HSC-engineered iNKT cells. B6-miNKT and control B6-Mock mice were analyzed for iNKT cell development at 6–8wkpostHSC transfer. The experiments were repeated at least three times, and representative results are presented. iNKT cells were detected as TCRβlomCD1d/PBS-57+ using flow cytometry. (A and B) FACS plots showing the characteristic development of iNKT cells in thymus. (C) FACS plots showing the maturation of iNKT cells in the periphery measured by the up-regulation of the NK1.1 marker. Comparisons of iNKT cells from thymus and periphery (liver) are shown. (D and E) FACS plots and bar graphs showing the exclusion of nontransgenic TCR expression on the HSC-engineered iNKT cells. Comparisons of iNKT and conventional αβ T (Tc) cells from the liver of B6-Mock or B6-miNKT mice are shown. Pan-TCR Vα panel includes Vα2, Vα3.2, and Vα8.3, whereas pan-TCR Vβ panel includes Vβ3, Vβ4, Vβ5, Vβ6, Vβ11, and Vβ13. N.D., not detected. development in thymus into four stages: Stage 1 (CD24+CD44− transgenic iNKT TCRs intracellularly, and these TCRs are not DX5−), Stage 2 (CD24−CD44−DX5−), Stage 3 (CD24−CD44+ functional. In addition to generating iNKT cells, our results show DX5−), and Stage 4 (CD24−CD44+DX5+) (15). Similar to their that TCR-engineered HSCs can also differentiate into all other endogenous counterparts, HSC-engineered iNKT cell progeni- blood cell lineages analyzed, including B cells (gated as CD19+), tors detected in the thymus of B6-miNKT mice followed a de- macrophages (gated as CD3−CD19−F4/80+), myeloid cells (gated velopmental path from Stages 1–4 (Fig. 3B). In addition to their as CD3−CD19−CD11b+), and granulocytes (gated as CD3−CD19− development in thymus to gain TCR expression (Control Point 1), Gr-1+)(Fig. S2). iNKT cells also differ from conventional αβ T cells in that they need to undergo an additional maturation step in the periphery to Antitumor Capacity of the HSC-Engineered iNKT Cells. Finally, we acquire the expression of NK1.1 (Control Point 2) (15). In B6- studied the cancer therapy potential of the HSC-engineered miNKT mice, iNKT cells detected in the periphery did up-regulate iNKT cells. B6-miNKT mice and control B6-Mock mice were NK1.1 expression compared with iNKT cells detected in the thy- challenged with B16.F10 melanoma cells through i.v. injections mus, similar to that observed for endogenous iNKT cells in the and analyzed for lung metastasis 2 wk later (Fig. 4A). Exper- control B6-Mock mice (Fig. 3C). imental mice received immunization with either unloaded Overexpression of prerearranged αβ TCR genes in HSCs or α-GalCer–loaded BMDCs (denoted as BMDC/none or has been shown to induce allelic exclusion and block the rear- BMDC/α-GalCer, respectively) on day 3 post tumor challenge to rangements of endogenous TCR genes in the resulting conven- boost iNKT cell activities and to mimic a therapeutic vaccination tional αβ T cells (13). Study of the iNKT cells generated in the treatment (Fig. 4A). Monitoring of the HSC-engineered iNKT B6-miNKT mice revealed that these cells expressed the trans- cells in the B6-miNKT mice showed that these cells actively + genic TCR (Vβ8 ), but not the other TCR Vβ chains analyzed in responded to tumor challenge, evidenced by their expansion our experiment (Fig. 3 D and E). In particular, these HSC- from ∼1.5% to ∼7% in blood (Fig. 4B). In comparison, endog- engineered iNKT cells did not express the TCR Vβ7 used by enous iNKT cells in the control B6-Mock mice also responded to ∼10% of endogenous iNKT cells (Fig. 3 D and E). Analysis of tumor challenge, but their limiting starting number (<0.2%) only TCR α chain expression also showed an exclusion of other TCR allowed them to reach ∼1.7% in blood (Fig. 4B). We observed Vα expression on the engineered iNKT cells (Fig. 3D). These a significant protection from lung metastasis in the B6-miNKT results suggest that the iNKT TCR-engineered HSCs give rise to mice compared with that in the control B6-Mock mice, as evi- clonal iNKT cells that only express the transgenic iNKT TCRs, denced by the reduction of both the number and size of tumor likely through an allelic exclusion mechanism during iNKT cell nodules (Fig. 4 C–E). Inclusion of a BMDC/α-GalCer immuni- development in thymus. zation further expanded the HSC-engineered iNKT cells (up to We also studied the lineage differentiation of iNKT TCR- ∼30% in blood; Fig. 4B). However, no significant further re- engineered HSCs. By detecting intracellular expression of duction of lung tumor nodules was observed (Fig. 4C). We + transgenic iNKT TCRs (gated as Vβ8intra ), TCR-engineered speculate that this may be due to a “saturation” of the antitumor HSCs and their progeny cells could be tracked (Fig. S2). Notably, capacity of iNKT cell-induced effector cells like NK cells and because only T cells express the CD3 molecules that are required tumor-specific conventional αβ T cells that were limiting in mice. to support the surface display of TCRs and their signaling, the Total clearance of tumor metastasis likely requires combination other cells that lack CD3 molecules can only express the therapy such as combining with adoptive transfer of additional

1526 | www.pnas.org/cgi/doi/10.1073/pnas.1424877112 Smith et al. 20 INFLAMMATION IMMUNOLOGY AND

Fig. 4. Protection from melanoma lung metastasis by the HSC-engineered iNKT cells. B6-miNKT and control B6-Mock mice were given i.v. injection of 0.5–1 × 106 B16.F10 melanoma cells on day 0 and analyzed for melanoma lung metastasis on day 14. On day 3, experimental mice received i.v. injection of 1 × 106 BMDCs either unloaded or loaded with α-GalCer (denoted as BMDC/none or BMDC/α-GalCer, respectively) to mimic a therapeutic vaccine treatment. The experiments were repeated twice (5–7micepergroup),andrepresentativeresultsarepresented.(A) Schematic representation of the experimental design to study the cancer therapy potential of the HSC-engineered iNKT cells in the B16 melanoma lung metastasis mouse model. (B) FACS plots showing the ex- pansion of iNKT cells in the blood of experimental mice in response to tumor challenge and BMDC/α-GalCer vaccination. (C–F)Analysisofmelanomalung metastasis in the experimental mice on day 14. (C) Enumeration of lung tumor nodules. Data are presented as mean ± SEM, *P < 0.01 (B6-miNKT samples compared with corresponding B6-Mock controls). (D)Photosoflungshowingmelanomametastasis.(E) Immunohistology analysis showing the H-E staining of lung sections. Metastatic tumor nodules are indicated by arrows. Bars: 1,000 μm (40× magnification); 500 μm (100× magnification). (F)Oneimageoflungfrom a representative B6-miNKT mouse showing the detection of pale, depigmented tumor nodules. effector cells. Notably, depigmentation of tumor nodules was generate iNKT cells within as few as 6 wk (Fig. 2). Most im- observed in high numbers in the B6-iNKT mice (Fig. 4F). Key portantly, unlike transgenic mouse technologies, this method can molecules in the pigment synthesis pathway are a major class of be applied to humans through gene-modified CD34+ cell trans- tumor antigens for melanoma, and mutating or down-regulating fer and therefore has direct translational potential (20). these molecules are common strategies by which melanoma In our study, we showed that the HSC-engineered mouse tumor cells escape immune attack, often leading to depigmen- iNKT cells followed a classical iNKT cell development path, tation (19). The presence of a large fraction of depigmented Check Point 1 in the thymus to gain iNKT TCR expression tumor nodules in the B6-miNKT mice therefore suggests a and Check Point 2 in the periphery to gain NK1.1 expression strong immune response against these tumors, presumably in- (Fig. 3). They also displayed a typical iNKT cell phenotype + duced by the HSC-engineered iNKT cells through activation of (TCRβlomCD1d/PBS-57hiNK1.1hiCD62LloCD44hiCD4 /−CD8−) antitumor NK and conventional αβ T cells (Fig. 4F) (3, 4). and exhibited full iNKT cell functionality with potent and fast Discussion response to antigen stimulation, both in vitro and in vivo (Fig. 2). In this report, we describe a new method of generating large These findings confirm the new HSC-engineered iNKT cell numbers of iNKT cells in mice through iNKT TCR gene engi- method as a powerful tool to study mouse iNKT cell biology. By neering of HSCs. Compared with existing iNKT TCR transgenic increasing the iNKT cells to a suitable level, their development mouse technology and iNKT iPS cell-derived transgenic mouse and function in health and disease conditions can be easily technology, this new method is cost-effective and high-through- monitored. In particular, this method allows us to generate large put. It is easy to implement through a standard retrovirus- numbers of clonal iNKT cells, thus enabling the investigation of transduced bone marrow transfer and has a fast turnover to similarities and differences between individual iNKT cell clones.

Smith et al. PNAS | February 3, 2015 | vol. 112 | no. 5 | 1527 21 For example, by studying the antigen recognition and functional can only be acquired during iNKT cell development, leaving differentiation of single iNKT cell clones, critical clues might be HSC engineering the sole approach to produce functional engi- revealed to increase understanding of the origins of various iNKT neered iNKT cells. cell subsets with distinct functions, such as those iNKT cell subsets biased to produce Th1, Th2, or Th17 effector cytokines (3). The Materials and Methods flexibility of this method also allows the convenient generation of The full description of materials and methods is provided in SI Materials iNKT cells of different genomic backgrounds at a fast pace and an and Methods. affordable cost, allowing examination of the functions of desig- Single-Cell iNKT TCR Cloning. The cloning method was performed based on an nated genes for regulating iNKT cell biology (8). established protocol (24), with several modifications. Details are provided in The therapeutic potential of this HSC-engineered iNKT cell SI Materials and Methods. approach is also promising. A broad range of applications, fast and strong responses, and the clinical availability of a potent HSC Isolation, Transduction, Adoptive Transfer, and Secondary Bone Marrow stimulatory reagent α-GalCer make iNKT cells attractive thera- Transfer. The procedures were reported previously (10) and are provided in SI peutic targets (4). In the past 2 decades, a series of iNKT cell- Materials and Methods. based clinical trials have been conducted, mainly targeting cancer (4, 21). A recent trial reported encouraging antitumor immunity in Statistical Analysis. Student’s two-tailed t test was used for paired compar- patients with head and neck squamous cell carcinoma, attesting to isons. Data are presented as mean ± SEM, unless otherwise indicated. P < the potential of iNKT cell-based immunotherapies (22). How- 0.01 was considered significant. ever, most trials yielded unsatisfactory results (4, 21). Overall, these trials all worked through the direct stimulation or ex vivo ACKNOWLEDGMENTS. We thank the University of California, Los Angeles expansion of patients’ endogenous iNKT cells, thus yielding (UCLA), animal facility for providing animal support, the UCLA Broad Stem Cell Research Center (BSCRC) FACS Core for providing flow cytometry only short-term, limited clinical benefits to a small number of support, the UCLA Translational Pathology Core Laboratory (TPCL) for patients. The low frequency and high variability of iNKT cells providing immunohistology support, the Laboratory of Donald Kohn, in in humans (∼0.01–1% in blood), as well as the rapid depletion particular, Eric Gschweng and Michael Kaufman, for providing technical of these cells poststimulation, are considered to be the major help, the National Institutes of Health Tetramer Core Facility for providing stumbling blocks limiting the success of these trials. However, the mCD1d/PBS-57 tetramer reagents, Donald Kohn and Gay Crooks for insightful discussion, and Pin Wang for critical reading of this manuscript. if successfully applied to humans, the reported new HSC- This work was supported by a UCLA BSCRC Innovation Award (to L.Y.), a engineered iNKT cell approach has the potential to provide UCLA SPORE in Prostate Cancer Career Development Award (NIH P50 patients with a lifelong supply of therapeutic levels of iNKT cells, CA092131, to L.Y.), a Concern Foundation Stem Cell Research Award (to taking advantage of the longevity and self-renewal of HSCs (23), L.Y.), a California Institute for Regenerative Medicine (CIRM) Basic Biology V thus eliminating a key barrier against current iNKT cell-based Exploratory Concepts Award (RB5-07089, to L.Y.), and a National Institutes of Health (NIH) Director’s New Innovator Award (DP2 CA196335, to L.Y.). immunotherapies. It is worthy to note that simply engineering O.N.W. is an investigator of the Howard Hughes Medical Institute and is par- conventional αβ T cells with iNKT TCR genes will not convert tially supported by the Eli and Edythe Broad Center of Regenerative Medicine these cells into iNKT cells. The unique functions of iNKT cells and Stem Cell Research.

1. Taniguchi M, Harada M, Kojo S, Nakayama T, Wakao H (2003) The regulatory role of 13. Giannoni F, et al. (2013) Allelic exclusion and peripheral reconstitution by TCR Valpha14 NKT cells in innate and acquired immune response. Annu Rev Immunol 21: transgenic T cells arising from transduced human hematopoietic stem/progenitor 483–513. cells. Mol Ther 21(5):1044–1054. 2. Bendelac A, Savage PB, Teyton L (2007) The biology of NKT cells. Annu Rev Immunol 14. Baltimore D, Witte ON, Yang L, Economou J, Ribas A (2010) Overcoming barriers to – 25:297 336. programming a therapeutic cellular immune response to fight melanoma. Pigment 3. Berzins SP, Smyth MJ, Baxter AG (2011) Presumed guilty: Natural killer T cell defects Cell Melanoma Res 23(2):288–289. and human disease. Nat Rev Immunol 11(2):131–142. 15. Godfrey DI, Berzins SP (2007) Control points in NKT-cell development. Nat Rev Im- 4. Vivier E, Ugolini S, Blaise D, Chabannon C, Brossay L (2012) Targeting natural killer munol 7(7):505–518. cells and natural killer T cells in cancer. Nat Rev Immunol 12(4):239–252. 16. Liu Y, et al. (2006) A modified alpha-galactosyl ceramide for staining and stimulating 5. Bendelac A, Hunziker RD, Lantz O (1996) Increased interleukin 4 and immunoglobulin E production in transgenic mice overexpressing NK1 T cells. JExpMed184(4): natural killer T cells. J Immunol Methods 312(1-2):34–39. 1285–1293. 17. Watarai H, Nakagawa R, Omori-Miyake M, Dashtsoodol N, Taniguchi M (2008) 6. Taniguchi M, et al. (1996) Essential requirement of an invariant V alpha 14 T cell Methods for detection, isolation and culture of mouse and human invariant NKT cells. antigen receptor expression in the development of natural killer T cells. Proc Natl Nat Protoc 3(1):70–78. Acad Sci USA 93(20):11025–11028. 18. Yang L, et al. (2012) miR-146a controls the resolution of T cell responses in mice. JExp 7. Ren Y, et al. (2014) Generation of induced pluripotent stem cell-derived mice by re- Med 209(9):1655–1670. programming of a mature NKT cell. Int Immunol 26(10):551–561. 19. Ramirez-Montagut T, Turk MJ, Wolchok JD, Guevara-Patino JA, Houghton AN (2003) 8. Yang L, Qin XF, Baltimore D, Van Parijs L (2002) Generation of functional antigen-spe- Immunity to melanoma: Unraveling the relation of tumor immunity and autoim- cific T cells in defined genetic backgrounds by retrovirus-mediated expression of TCR munity. Oncogene 22(20):3180–3187. cDNAs in hematopoietic precursor cells. Proc Natl Acad Sci USA 99(9):6204–6209. 20. Kohn DB, Pai SY, Sadelain M (2013) Gene therapy through autologous trans- 9. Arnold PY, Burton AR, Vignali DA (2004) Diabetes incidence is unaltered in glutamate plantation of gene-modified hematopoietic stem cells. Biol Blood Marrow Transplant decarboxylase 65-specific TCR retrogenic nonobese diabetic mice: Generation by 19(1, Suppl):S64–S69. retroviral-mediated stem cell gene transfer. J Immunol 173(5):3103–3111. 21. Pilones KA, Aryankalayil J, Demaria S (2012) Invariant NKT cells as novel targets for 10. Yang L, Baltimore D (2005) Long-term in vivo provision of antigen-specific T cell im- munity by programming hematopoietic stem cells. Proc Natl Acad Sci USA 102(12): immunotherapy in solid tumors. Clin Dev Immunol 2012:720803. 4518–4523. 22. Yamasaki K, et al. (2011) Induction of NKT cell-specific immune responses in cancer 11. Bettini ML, Bettini M, Vignali DA (2012) T-cell receptor retrogenic mice: A rapid, tissues after NKT cell-targeted adoptive immunotherapy. Clin Immunol 138(3):255–265. flexible alternative to T-cell receptor transgenic mice. Immunology 136(3):265–272. 23. Morrison SJ, Uchida N, Weissman IL (1995) The biology of hematopoietic stem cells. 12. Vatakis DN, et al. (2011) Antitumor activity from antigen-specific CD8 T cells gener- Annu Rev Cell Dev Biol 11:35–71. ated in vivo from genetically engineered human hematopoietic stem cells. Proc Natl 24. Smith K, et al. (2009) Rapid generation of fully human monoclonal antibodies specific Acad Sci USA 108(51):E1408–E1416. to a vaccinating antigen. Nat Protoc 4(3):372–384.

1528 | www.pnas.org/cgi/doi/10.1073/pnas.1424877112 Smith et al. 22 Supporting Information

Smith et al. 10.1073/pnas.1424877112 Materials and Methods murine TCR genomic segments [the international ImMuno- Mice and Materials. C57BL/6J (B6) mice were purchased from the GeneTics information system (IMGT), www.imgt.org]. The selected Jackson Laboratory. Six- to ten-week-old females were used for iNKT TCR α and β pair cDNAs were then synthesized as a single all experiments unless otherwise indicated. All animal experi- bicistonic gene, with codon optimization and a F2A sequence ments were approved by the Institutional Animal Care and Use linking the TCRα and TCRβ cDNAs to enable their coex- Committee of the University of California, Los Angeles. pression (GenScript). α-Galactosylceramide (α-GalCer, KRN7000) was purchased from Avanti Polar Lipids; lipopolysaccharides (LPS) and 5-fluoro- The 293.T/mCD3 Stable Cell Line. HEK293.T human embryonic uracil (5-FU) from Sigma; recombinant murine IL-3, IL-6 and stem kidney epithelial cells (ATCC) were stably transduced with a cell factor (SCF) from PeproTech; and polybrene from Millipore. lentiviral vector (3) coexpressing all four chains of mouse CD3 Fluorochrome-conjugated mCD1d/PBS-57 tetramer reagents complex (CD3γ, CD3δ, CD3e, and CD3ζ), through linking the were provided by the NIH Tetramer Core Facility (Emory four cDNAs with three different 2A sequences (F2A, foot-and- University, Atlanta, GA). Fixable Viability Dye eFluor455UV mouth disease virus 2A; P2A, porcine teschovirus-1 2A; and was purchased from affymetrix eBioscience. T2A, Thosea asigna virus 2A). The transduced cells were then transiently transfected with an MOT1 vector encoding a mouse Antibodies and Flow Cytometry. Fluorochrome-conjugated anti- CD8 TCR, using a standard calcium precipitation procedure (1). Single cells supporting the high surface expression of OT1 TCRs bodies specific for mouse CD3, CD4, CD8, CD19, CD11b, CD24, + + CD62L, CD44, DX5, F4/80, Gr-1, TCRβ, TCR Vβ7, TCR Vβ8, (gated as CD3 TCR Vβ5 ) were sorted out using flow cytometry and TCR Vα8.3 were purchased from BioLegend; for mouse and grown into single-cell clones. A stable, single-cell clone that NK1.1, IFN-γ, IL-4, TCR Vα2, TCR Vα3.2, TCR Vβ3, TCR lost OT1 TCR expression, but retained the capacity to support mouse TCR surface expression, was selected and designated as Vβ4, TCR Vβ5, TCR Vβ6, TCR Vβ11, and TCR Vβ13, from BD Biosciences. Fc Block (anti-mouse CD16/32) was purchased the 293.T/mCD3 stable cell line. from BD Biosciences. Cells were stained as previously described Mock and miNKT Retroviruses. Mock (MIG) retroviral vector was (1) and analyzed using an LSRFortessa flow cytometer (BD reported previously (1). miNKT retroviral vector was constructed Biosciences). FlowJo software was used to analyze the data. by inserting the synthetic bicistronic gene (iNKT TCRα-F2A- ELISA. The ELISAs for detecting mouse cytokines were performed TCRβ) into the MIG vector, replacing the IRES-EGFP segment. following a standard protocol from BD Biosciences. The capture Retroviruses were made using HEK293.T cells, following a stan- dard calcium precipitation protocol as previously described (1). and biotinylated antibody pairs for detecting mouse IFN-γ and IL-4 were also purchased from BD Biosciences. The streptavidin– HSC Isolation, Transduction, Adoptive Transfer, and Secondary Bone HRP conjugate and mouse IFN- and IL-4 Single-Use ELISA γ Marrow Transfer. The procedures were reported previously (1). In Ready-Set-Go (RSG) Standards were purchased from affymetrix brief, B6 mice were treated with 5-fluorouracil (250 μg per gram eBioscience. The 3,3′,5,5′-Tetramethylbenzidine (TMB) substrate body weight). Five days later, bone marrow (BM) cells were was purchased from KPL. The samples were analyzed for ab- harvested and cultured for 4 d in BM cell culture medium con- sorbance at 450 nm using an Infinite M1000 microplate reader taining recombinant murine IL-3 (20 ng/mL), IL-6 (50 ng/mL), (Tecan). and SCF (50 ng/mL). On days 2 and 3, BM cells were spin- infected with retroviruses supplemented with 8 g/mL of poly- Single-Cell iNKT TCR Cloning. The single-cell iNKT TCR RT-PCR μ brene, at 770 × g, 30 °C for 90 min. On day 4, BM cells were was performed based on an established protocol (2), with certain collected and i.v. injected into B6 recipients that had received modifications. iNKT cells were sorted from mouse spleen cells 6 lo 1,200 rads of total body irradiation (∼1–2 × 10 transduced BM based on a stringent forum of surface markers (CD3 mCD1d/ cells per recipient). For secondary BM transfer, fresh total BM PBS-57+TCR V 8+NK1.1hi)usingaFACSAriaIIflowcyto- β cells harvested from the primary BM recipients were i.v. injected meter (BD Biosciences) (lo, low; hi, high). Single cells were into secondary B6 recipient mice that had received 1,200 rads of sorted directly into PCR plates containing cell lysis buffer. The 6 total body irradiation (∼10 × 10 total BM cells per recipient). plates were then immediately flash frozen and stored at −80 °C The BM recipient mice were maintained on the combined anti- until use. Upon thawing, the cell lysate from each cell was split in biotics sulfmethoxazole and trimethoprim oral suspension (Septra; half on the same PCR plate and processed directly into iNKT Hi-Tech Pharmacal) in a sterile environment for 6–8 wk until TCR cloning for both α and β chain genes using a OneStep RT- analysis or use for further experiments. PCR kit (QIAGEN), following the manufacturer’s instructions and using the iNKT TCR gene-specific primers. These primers Bone Marrow-Derived Dendritic Cell Generation, Antigen Loading, were designed to amplify the ∼200 bps spanning the CDR3 re- and Mouse Immunization. B6 mouse BMDCs were generated gions of the iNKT TCR α and β chain cDNAs and were cus- from BM cell cultures and matured with LPS as described pre- tomer-synthesized by Integrated DNA Technologies (IDT): for viously (1). The LPS-matured BMDCs were then cultured at TCRα (FW primer: 5′-GGG AGA TAC TCA GCA ACT CTG 37 °C in a 6-well plate at 10 × 106 cells/well/2 mL BMDC culture GAT AAA GAT GC -3′; BW primer: 5′- CCA GAT TCC ATG medium containing 5 μg/mL of α-GalCer for 2 h, with gentle GTT TTC GGC ACA TTG -3′) and for TCRβ (FW: 5′- GGA shaking every 30 min. The α-GalCer–loaded BMDCs were then GAT ATC CCT GAT GGA TAC AAG GCC TCC -3′; BW: 5′- washed twice with PBS and used to immunize mice through i.v. GGG TAG CCT TTT GTT TGT TTG CAA TCT CTG -3′). injection (∼1 × 106 BMDCs/mouse). Verified sequences (productive germline Vα14-Jα18-Cα assem- bly for TCRα and Vβ8-D/J/N-Cβ assembly for TCRβ) were used In vitro iNKT Cell Functional Assays. Spleen cells containing iNKT to construct the complete cDNA sequences encoding the TCR α cells were cultured in vitro in a 24-well plate at 2 × 106 cells per and β chains from a single cell, based on information about well in regular mouse lymphocyte culture medium, with or

Smith et al. www.pnas.org/cgi/content/short/1424877112 1of3 23 without the addition of α-GalCer (100 ng/mL), for 5 d. On days 3 with α-GalCer. On day 14, mice were euthanized, and their lungs and 5, cells were collected and assayed for iNKT cell expansion were harvested and analyzed for melanoma metastasis by counting using flow cytometry, and the cell culture supernatants were tumor nodules under a Zeiss Stemi 2000-CS microscope (Carl collected and assayed for effector cytokine (IFN-γ and IL-4) Zeiss AG) at 10× magnification. Representative lungs were also production by ELISA. On day 3, some cells were also treated analyzed by immunohistology. with 4 μL/6 mL BD GolgiStop for 4–6 h and then assayed for intracellular cytokine production using flow cytometry via in- Immunohistology. Lung tissues collected from the experimental tracellular staining using the BD Cytofix/Cytoperm Fixation/ mice were fixed in 10% neutral-buffered formalin and embedded Permeabilization Kit (BD Biosciences). in paraffin for sectioning (5 μm thickness), followed by hema- toxylin and eosin staining using standard procedures (UCLA In vivo iNKT Cell Functional Assay. Mice were immunized with 6 Translational Pathology Core Laboratory, Los Angeles, CA). α-GalCer–loaded BMDCs through i.v. injection (∼1 × 10 BMDCs per mouse) and then periodically bled to monitor the in The sections were imaged using an Olympus BX51 upright mi- vivo iNKT cell responses using flow cytometry. croscope equipped with an Optronics Macrofire CCD camera (AU Optronics) at 40× and 100× magnifications. The images were B16 Melanoma Lung Metastasis Mouse Model. Mice that received i.v. analyzed using Optronics PictureFrame software (AU Optronics). injection of 0.5–1 × 106 B16.F10 melanoma cells were allowed to develop lung metastasis over the course of 2 wk (4). On day 3 Statistical analysis. Student’s two-tailed t test was used for paired post tumor challenge, the experimental mice received i.v. in- comparisons. Data are presented as mean ± SEM, unless oth- jection of 1 × 106 BMDCs that were either unloaded or loaded erwise indicated. P < 0.01 was considered significant.

1. Yang L, Baltimore D (2005) Long-term in vivo provision of antigen-specific T cell im- 3. Yang L, et al. (2008) Engineered lentivector targeting of dendritic cells for in vivo munity by programming hematopoietic stem cells. Proc Natl Acad Sci USA 102(12): immunization. Nat Biotechnol 26(3):326–334. 4518–4523. 4. Fujii S, Shimizu K, Kronenberg M, Steinman RM (2002) Prolonged IFN-gamma-pro- 2. Smith K, et al. (2009) Rapid generation of fully human monoclonal antibodies specific ducing NKT response induced with alpha-galactosylceramide-loaded DCs. Nat Immunol to a vaccinating antigen. Nat Protoc 4(3):372–384. 3(9):867–874.

Fig. S1. Titration of the miNKT retroviral vector. The 293.T/mCD3 cells were transduced with the titrated volume of indicated virus supernatants. Three days later, virus-mediated expression of mouse TCRs was measured using flow cytometry. Representative FACS plots showing the detection of mouse TCRs on cell surface are presented. Note mouse CD3 (mCD3) molecules only display on cell surface in complex with the transgenic mouse TCRs, therefore, they can be used as an indicator of transgenic TCR expression. The results show comparable titers of the miNKT and MOT1 retroviral vectors, estimated as ∼0.5–1 × 106 IFU/mL (infectious units per milliliter). Mock, the control retroviral vector encoding an EGFP reporter gene; miNKT, the retroviral vector encoding a selected pair of mouse iNKT TCR α and β chain genes; MOT1, the retroviral vector encoding the α and β chain genes of OT1 TCR, a mouse conventional αβ TCR specific for chicken ovalbumin (1).

Smith et al. www.pnas.org/cgi/content/short/1424877112 2of3

24 Fig. S2. Lineage differentiation of iNKT TCR-engineered HSCs. B6-miNKT and control B6-Mock mice were analyzed for the presence of iNKT TCR-engineered cells at 6–8 wk post HSC transfer. The experiments were repeated at least three times, and representative FACS plots (A)andbargraphs(B)areshown.En- + gineered cells were detected by intracellular staining of the transgenic TCRβ chain (gated as TCR Vβ8intra ). Comparison analysis of the spleen cells of B6-miNKT and B6 control mice is presented. N.D., not detected.

Smith et al. www.pnas.org/cgi/content/short/1424877112 3of3

25

CHAPTER 3:

Propagating humanized BLT mice for the study

of human immunology and immunotherapy

26 STEM CELLS AND DEVELOPMENT ORIGINAL RESEARCH REPORT Volume 00, Number 00, 2016 Ó Mary Ann Liebert, Inc. DOI: 10.1089/scd.2016.0193

Propagating Humanized BLT Mice for the Study of Human Immunology and Immunotherapy

Drake J. Smith,1,2 Levina J. Lin,1 Heesung Moon,1 Alexander T. Pham,1 Xi Wang,1 Siyuan Liu,1 Sunjong Ji,1 Valerie Rezek,3–5 Saki Shimizu,5,6 Marlene Ruiz,5,6 Jennifer Lam,5,6 Deanna M. Janzen,6,7 Sanaz Memarzadeh,3,7–9 Donald B. Kohn,1,3,7,10 Jerome A. Zack,1,3,5,7 Scott G. Kitchen,3–5 Dong Sung An,5,6 and Lili Yang1,3,7,8

The humanized bone marrow-liver-thymus (BLT) mouse model harbors a nearly complete human immune system, therefore providing a powerful tool to study human immunology and immunotherapy. However, its application is greatly limited by the restricted supply of human CD34+ hematopoietic stem cells and fetal thymus tissues that are needed to generate these mice. The restriction is especially significant for the study of human immune systems with special genetic traits, such as certain human leukocyte antigen (HLA) haplotypes or monogene deficiencies. To circumvent this critical limitation, we have developed a method to quickly propagate established BLT mice. Through secondary transfer of bone marrow cells and human thymus implants from BLT mice into NSG (NOD/SCID/IL-2Rg-/-) recipient mice, we were able to expand one primary BLT mouse into a colony of 4–5 proBLT (propagated BLT) mice in 6–8 weeks. These proBLT mice reconstituted human immune cells, including T cells, at levels comparable to those of their primary BLT donor mouse. They also faithfully inherited the human immune cell genetic traits from their donor BLT mouse, such as the HLA- A2 haplotype that is of special interest for studying HLA-A2-restricted human T cell immunotherapies. Moreover, an EGFP reporter gene engineered into the human immune system was stably passed from BLT to proBLT mice, making proBLT mice suitable for studying human immune cell gene therapy. This method provides an opportunity to overcome a critical hurdle to utilizing the BLT humanized mouse model and enables its more widespread use as a valuable preclinical research tool.

Keywords: humanized BLT mice, human immunology, human immunotherapy, propagating, CD34, HSC

Introduction (IL-2Rc) gene that is the common signaling subunit of the receptors for multiple cytokines (IL-2, IL-4, IL-7, IL-9, IL- mall animal models that harbor a human immune 15, and IL-21) and is, thus, involved in the development of Ssystem and support human immune cell development lymphoid and natural killer (NK) cells [1–3]. At present, are valuable tools for the study of human immunology and NOD/SCID/IL-2Rg-/- (NSG) or BALB/c/RAG2-/-/IL-2Rg-/- the preclinical development of human immunotherapy. In the (BRG) mice are the standard recipients in generation of past decades, various humanized mouse models were de- humanized mice, because they are deficient of murine T cells, veloped toward enabling the long-term engraftment of a B cells, and NK cells. NSG mice have shown stronger en- complete human immune system [1–3]. As the first step, graftment, presumably due to their expression of a signal immunodeficient mouse strains were identified or developed regulatory protein alpha (SIRP-a) gene that is more homol- that accept xenogeneic transplanted human immune cells. ogous to the human SIRP-a gene and that delivers a more These mouse strains carry spontaneous or targeted mutations effective ‘‘do-not-eat-me’’ signal resulting in decreased of immune regulatory genes such as the Prkdc gene (spon- phagocytosis of engrafted human cells [4,5]. As the second taneous mutation in SCID or NOD/SCID mice) or RAG1/ step, various human immune cell xenograft protocols were RAG2 genes that are involved in the formation of antigen developed. The initial protocol involves injecting human receptors on B and T cells, or the IL-2 receptor gamma-chain peripheral blood mononuclear leukocytes (PBL) into SCID

1Department of Microbiology, Immunology and Molecular Genetics; 2Molecular Biology Interdepartmental PhD Program; 3Jonsson Comprehensive Cancer Center, David Geffen School of Medicine; 4Department of Medicine; 5AIDS Institute; 6School of Nursing; 7Eli and Edythe Broad Center of Regenerative Medicine and Stem Cell Research; 8Molecular Biology Institute; 9Department of Obstetrics and 10 Gynecology; Department of Pediatrics, Division of Hematology/Oncology, University of California, Los Angeles, California. 1 27 2 SMITH ET AL. mice. The resulting Hu-PBL-SCID mice contain limited traits, such as certain human leukocyte antigen (HLA) hap- lineages of human immune cells and do not support de novo lotypes or immune monogene deficiencies [19–22]. human immune cell development [6]. Transfer of human To circumvent this critical limitation, we have developed CD34+ hematopoietic stem and progenitor cells (HSPCs) into a method to quickly propagate established BLT mice newborn NSG or BRG mice results in long-term engraftment without the need of additional human tissues. We hypoth- of CD34+ cells and reconstitution of multilineage human esized that human CD34+ cells engrafted into the bone immune cells [7,8]. However, a key limitation of such HIS marrow of a primary BLT mouse retained their HSC po- (human immune system) mice is the aberrant development of tential and could repopulate a human immune system in human T cells in the mouse thymus that affects the MHC multiple naı¨ve NSG mice through secondary bone marrow selection and functional maturation of human T cells [7,8]. transfer; meanwhile, the human thymus organoid estab- Human T cells generated in these HIS mice are selected on lished in a primary BLT mouse maintained a human thymus mouse MHC molecules, making them incapable of recog- structure and could be split and transplanted into the sec- nizing human antigen-presenting cells (APCs) and mounting ondary recipient NSG mice to provide a human thymus human HLA-restricted T cell responses [7,8]. The issue of T microenvironment supporting proper human T cell devel- cell education was circumvented by another protocol wherein opment. In the present article, we demonstrated the feasi- transfer of human fetal liver and thymus tissue under the bility of this new method to expand a single primary BLT kidney capsule of recipient SCID mice gave rise to a human mouse into a colony of 4–5 proBLT (propagated BLT) mice organoid that was capable of supporting proper human T cell in 6–8 weeks, and we provided evidence to support the development [9]. Although human T cells are abundant in potential research value of these proBLT mice. these SCID-hu mice, they are mainly confined to the organoid and, thus, the reconstitution of the other organs is poor [9]. Materials and Methods The most comprehensive transfer protocol combined the strengths of these early protocols and involves the transfer of Mice and materials + human CD34 cells into the most supportive NSG recipient SCID tm1Wjl -/- NOD.Cg-Prkdc Il2rg /SzJ (NOD/SCID/IL-2Rg , mice, as well as implanting human fetal thymus and liver NSG) mice were purchased from the Jackson Laboratory tissue under the kidney capsule of these mice [10,11]. The and maintained in the animal facilities at the University of resulting humanized mice, named bone marrow-liver-thymus California, Los Angeles (UCLA). Six- to 10-week-old females (BLT) mice, support the long-term engraftment and systemic were used for all experiments, unless otherwise indicated. All reconstitution of a nearly complete human immune system, animal experiments were approved by the Institutional Animal including multilineage human adaptive and innate immune Care and Use Committee of UCLA. cells consisting of T cells, B cells, NK cells, dendritic cells, X-VIVO-15 cell culture medium was purchased from and macrophages [10,11]. Importantly, human immune cells Lonza. Recombinant human Flt3 ligand, stem cell factor developed in BLT mice, especially T cells, are functional, (SCF), thrombopoietin (TPO), IL-3, and Fixable Viability and they have shown productive responses to skin xenografts Dye eFluor506 were purchased from affymetrix eBioscience. and various viral/bacterial infections [10–13]. RetronectinÒ was purchased from Clontech. Because it supports the development and maintenance of a nearly complete and functional human immune system, the humanized BLT mouse model is a promising tool to study Antibodies and flow cytometry human hematopoiesis and immune cell activities under Fluorochrome-conjugated antibodies that were specific for healthy and disease conditions [1,2]. It is particularly useful human CD45, TCRab, CD11b, CD11c, CD14, CD19, CD56, for studies of human immunodeficiency virus (HIV) infec- and HLA-A2 were purchased from BioLegend; those spe- tion because of the high frequencies of human T cells in the cific for human CD34 were purchased from BD Biosciences. lymphoid and mucosal tissues of BLT mice, as well as be- Human Fc Receptor Blocking Solution (TruStain FcXÔ) cause of the proper maturation status and lineage differen- was purchased from BioLegend, whereas mouse Fc Block tiation of these human T cells [14–16]. To date, studies using (anti-mouse CD16/32) was purchased from BD Biosciences. BLT mice have generated valuable knowledge in many as- Cells were stained as previously described and analyzed by pects of HIV infection, including prevention, mucosal using an MACSQuant Analyzer 10 flow cytometer (Miltenyi transmission, HIV-specific innate and adaptive immunity, Biotec) [23]. FlowJo software was used to analyze the data. viral latency, and novel anti-retroviral and immune-based therapies for suppression and reservoir eradication [14–17]. Tissue processing for flow cytometry analysis The humanized BLT mouse model is also ideal for the study of hematopoietic stem cell (HSC)- and T cell-based im- For flow cytometry analysis, all tissues were processed munotherapies, because of the long-term engraftment of into mononuclear cells (MNCs) and lysed of red blood cells human HSCs and T cells in BLT mice [10,11]. We and (RBCs). Blood and bone marrow samples were directly others have utilized BLT mice for the preclinical develop- lysed with Tris-buffered ammonium chloride (TAC) buffer, ment of gene-modified HSC-based immunotherapies for following a standard protocol (Cold Spring Harbor Proto- treating cancer and HIV [18,19]. Despite its potential as a cols). Spleens were smashed against a 70 mm cell strainer valuable research tool, the application of BLT mice is greatly (Corning) to prepare single cells and then lysed with TAC. limited by the restricted supply of human CD34+ cells and Livers were cut into small pieces by using a pair of scissors, human fetal liver and thymus tissues that are required to smashed against a 70 mm cell strainer to prepare single cells, generate these mice. The restriction is especially problematic and passed through a 33% Percoll gradient isolation (Sigma) for the study of human immune systems with special genetic to remove hepatocytes, followed by TAC lysis.

28 PROPAGATING HUMANIZED BLT MICE 3

Human CD34+ cells and thymus tissues Immunohistology Human fetal liver CD34+ HSPCs, as well as fetal thymus Primary human fetal thymus tissue or human thymus tissues, were obtained from the CFAR Gene and Cellular organoid implants dissected out from the experimental BLT Therapy Core Laboratory at UCLA, without identification or proBLT mice were fixed in 10% neutral-buffered formalin information under federal and state regulations. CD34+ cells and embedded in paraffin for sectioning (4 mm thickness), were sorted from fetal liver cells through magnetic-activated followed by hematoxylin and eosin (H/E) staining or anti- cell sorting by using a Direct CD34 Progenitor Cell Isolation body staining (for human CD45 or CD3) by using standard Kit (Miltenyi Biotec) following the manufacturer’s in- procedures (UCLA Translational Pathology Core Labora- structions [18]. The purity of CD34+ cells was more than tory). The sections were imaged by using an Olympus BX51 97%, as evaluated by flow cytometry. upright microscope equipped with an Optronics Macrofire CCD camera (AU Optronics) at 40 · ,100· ,and400· mag- pMND-EGFP lentiviruses nifications. The images were analyzed by using Optronics PictureFrame software (AU Optronics). pMND-EGFP lentiviral vector was constructed by insert- ing an EGFP reporter gene into the lentivector that contains Statistical analysis the MND retroviral LTR U2 region as an internal promoter [24]. VsVg-pseudotyped pMND-EGFP lentiviruses were Student’s two-tailed t test was used for paired compari- produced by using HEK293.T cells, following a standard sons. Data are presented as mean – SEM, unless otherwise calcium precipitation protocol and an ultracentrifugation indicated. P < 0.05 was considered significant. concentration protocol as previously described [25]. Results Generation of humanized BLT Mice Generation of BLT mice Humanized BLT mice were generated as previously de- scribed, with certain modifications [10,11]. In brief, human We generated humanized BLT mice by following previ- CD34+ cells were cultured for no more than 48 h in X-VIVO- ously established procedures, with certain modifications 15 cell culture medium containing recombinant human Flt3 (Materials and Methods section; Fig. 1A). In brief, cryo- + ligand (50 ng/mL), SCF (50 ng/mL), TPO (50 ng/mL), and IL- preserved human fetal liver CD34 cells were thawed and 3 (20 ng/mL) in nontissue culture-treated plates coated with cultured for no more than 48 h in X-VIVO-15 medium Retronectin. Viral transduction, when applicable, was per- containing recombinant human Flt3 ligand, SCF, TPO, and formed at 24 h by adding concentrated pMND-EGFP lenti- IL-3. Viral transduction, when applicable, was performed at viruses directly to the culture medium. At around 48 h, CD34+ 24 h by adding concentrated lentiviral vectors directly to the + cells were collected and i.v. injected into NSG recipient mice culture medium. At around 48 h, CD34 cells were collected -/- ( 0.5–1 106 CD34+ cells per recipient) that had received and i.v. injected into NOD/SCID/IL-2Rgc (NSG) recipi- * · 6 + 270 rads of total body irradiation. Then, 1–2 fragments of ent mice (*0.5–1 · 10 CD34 cells per recipient) that had 3 received 270 rads of total body irradiation. On the same day, human fetal thymus (*1mm ), as well as donor-matched 3 fetal liver CD34- cells, when available (*4.5 · 106), were 1–2 fragments of human fetal thymus (*1 mm ), as well as donor-matched fetal liver CD34- cells, when available implanted under the kidney capsule of each recipient NSG 6 mouse. The mice were maintained on trimethoprim/sulf- (*4.5 · 10 ), were implanted under the kidney capsule of methoxazole (TMS) chow in a sterile environment for 8–12 each recipient NSG mouse. The resulting BLT mice were weeks until analysis or use for further experiments. allowed to reconstitute a human immune system, whereas periodic bleedings were performed to monitor the presence + Generation of Humanized proBLT Mice of human immune cells (gated as hCD45 , Fig. 1B, D). We started to detect human immune cells in BLT mice at 4 Humanized proBLT (propagated BLT) mice were gener- weeks post human tissue transplantation. The levels of hu- ated through adoptive transfer of bone marrow cells and man immune cells gradually increased over time and then human thymus organoid implant fragments from BLT mice peaked and stabilized at around week 12 (Fig. 1B). In our into secondary recipient NSG mice that had received 270 experiments, we routinely obtained *30–80% human im- rads of total body irradiation. On the day of transfer, one mune cell reconstitution in the blood of BLT mice. Data primary BLT mouse was dissected. Total bone marrow cells from a representative experiment showing *40% human harvested from the femur and tibia of the BLT mouse were immune cell reconstitution are presented (Fig. 1B, D). In our split equally and i.v. injected into 4–5 recipient NSG mice. studies, we found that transfer of fetal liver (either tissue On average, about 40–50 · 106 total bone marrow cells were fragments or CD34- cells) was optional for making BLT harvested from each primary BLT donor mouse, whereas mice. We also found that human CD34+ cells and human about 10 · 106 total BLT bone marrow cells were given to fetal thymus used for making BLT mice did not need to be each NSG recipient mouse. Meanwhile, human thymus or- donor matched, making it flexible to use CD34+ cells iso- ganoid was dissected out from the kidney capsule of a BLT lated from fetal liver, cord blood, adult bone marrow, or G- mouse, cut into pieces of *1 mm3, and surgically implanted CSF-mobilized adult peripheral blood. The reconstitution under the kidney capsule of the secondary recipient NSG efficiency was ranked as the following: fetal liver CD34+ mice (1–2 human thymus organoid pieces per recipient). The cells > cord blood CD34+ > adult CD34+ cells (data not mice were maintained on TMS chow in a sterile environment shown). In this article, data from BLT mice made with fetal for 8–12 weeks until analysis or use for further experiments. liver CD34+ cells (either donor matched or -unmatched with

29 4 SMITH ET AL.

FIG. 1. Generation of proBLT mice through secondary transfer of bone marrow cells and human thymus implants from primary BLT mice to naı¨ve NSG mice. The experiments were repeated over six times. Representative results are presented. (A) Schematic representation of the experimental design to generate humanized BLT and proBLT mice. (B) Scatter plot showing the time-course detection of human CD45+ cells (gated as hCD45+ of total MNCs) in the blood of BLT mice post human tissue transplantation (n = 5). Data are presented as individual mouse measurements and mean – SEM of all ex- perimental mice at the indicated time points. (C) Scatter plot showing the time course detection of human CD45+ cells (gated as hCD45+ of total MNCs) in the blood of proBLT mice post secondary transfer of bone marrow cells and human thymus implants from a single primary BLT donor mouse (n = 4–5). Data are presented as individual mouse measurements and mean – SEM of all experimental mice at the indicated time points. (D) Representative FACS plots showing the presence of human CD45+ cells in the blood of proBLT mice at 6 weeks post secondary transfer (n = 5). NSG, NOD/SCID/IL-2Rg-/- mice; BLT, human bone marrow-liver-thymus implanted NSG mice; proBLT, propagated BLT mice; BM, bone marrow; MNCs, mononuclear cells. fetal thymus tissue) are presented. Notably, when nondonor weeks post primary human tissue transplant were used as matched CD34+ cells and fetal thymus are used to produce donor mice. Total bone marrow cells harvested from one BLT BLT mice, such BLT mice can still be valuable tools to mouse (*40–50 · 106 cells) were split equally and i.v. injected study human immune cell development, but certain pre- into 4–5 recipient NSG mice (*10 · 106 cells per recipient). cautions need to be taken when using these mice to study Meanwhile, human thymus organoid dissected out from the human T cell immunity. In these BLT mice, T cell function BLT mouse was cut into pieces of *1 mm3, and it was then may be affected by the HLA mismatch between human T surgically implanted under the kidney capsule of the recipient cells developed in these mice (selected on fetal thymus NSG mice (1–2 human thymus organoid pieces per recipient). HLAs) and human APCs generated in these mice (derived The resulting proBLT mice were allowed to reconstitute a from CD34+ cells and, therefore, carrying their HLAs). For human immune system, whereas periodic bleedings were certain research, a partial HLA match (e.g., HLA-A2+) may performed to monitor the generation of human immune cells still allow for the study of defined types of T cell responses (gated as hCD45+; Fig. 1C, D). Interestingly, human immune (e.g., HLA-A2-restricted T cell response). cells were reconstituted more quickly in proBLT mice com- pared with those in BLT mice, peaking and stabilizing at Generation of propagated BLT mice around week 6 (Fig. 1C). The faster human immune cell re- constitution in proBLT mice was likely due to their inheritance Humanized propagated BLT (proBLT) mice were gener- of both early and intermediate human hematopoietic progen- ated through secondary transfer of bone marrow cells and itor cells, as well as mature human immune cells, from the human thymus organoid fragments harvested from BLT mice bone marrow of donor BLT mice. Notably, because the bone into naı¨ve recipient NSG mice that had received 270 rads of marrow cells and human thymus organoid implant harvested total body irradiation (Fig. 1A). Primary BLT mice at 12 from a single BLT mouse could be used to generate 4–5

30 PROPAGATING HUMANIZED BLT MICE 5 proBLT mice, these proBLT mice contain a human immune opment [10,11]. In our experiments, these human thymus system including a human T cell educational microenviron- organoids could grow to a size of *20–200 mm3. The im- ment genetically identical to that of their donor BLT mouse age of a representative organoid is presented in Fig. 3A. and, thus, can be considered ‘‘clonal’’ (Fig. 1A). Post stabi- When generating proBLT mice, such an organoid was cut lization, the human immune cell reconstitution levels are into *1 mm3 fragments and 1–2 pieces of these fragments consistent among individual ‘‘clonal’’ proBLT mice and were implanted under the kidney capsule of each second- similar to those of the primary BLT donor mouse (Fig. 1C, D). ary NSG recipient mouse (Materials and Methods section). Therefore, the proBLT approach allows the expansion of es- In proBLT mice, secondary human thymus organoids were tablished BLT mice to a large homogeneous colony in a short observed and grown to a size similar to that of the primary period of time, while also maintaining the ‘‘clonal’’ nature of BLT mice. An immunohistology study revealed that both the engrafted human immune system. These features are es- the BLT and proBLT human thymus organoids displayed a pecially attractive for studies that are large scale and require typical human thymus structure similar to that of the pri- specific genetic traits of human immune cells such as MHC mary human fetal thymus, comprising a cortex region that haplotypes or immune monogene deficiencies [19–22]. could support the positive selection of human thymocytes for HLA recognition, and a medulla region that could sup- Reconstitution of multilineage human immune cells port the negative selection of thymocytes for autoreactive T in proBLT mice cell depletion (Fig. 3B, C). These human thymus organoids were populated with developing human thymocytes that had To study the human immune cell reconstitution in undergone T cell receptor (TCR) selection, evidenced by the proBLT mice, we performed a systemic analysis of these positive immunochemical staining for human CD45 and mice in comparison with their ‘‘parental’’ BLT mouse. Data CD3 markers (Fig. 3B). Interestingly, in the medulla of from a representative analysis are presented in Fig. 2. High these human thymus organoids, we also observed abundant percentages of human immune cells were detected in all numbers of Hassall’s corpuscles (HCs), structures that are immune-homing tissues of proBLT mice, including pe- unique to human thymus and have been implicated in reg- ripheral blood ( 80%), central lymphoid organs such as * ulating the development of human FoxP3+ regulatory T cells bone marrow ( 80%) and spleen ( 80%), and immune * * (Fig. 3C) [26]. These results indicate the reconstitution of a regulatory organs such as liver ( 90%), at levels similar to > proper human thymus microenvironment in both BLT and those of their primary donor BLT mouse (Fig. 2A). Lineage proBLT mice. analysis revealed the reconstitution of a nearly complete Next, we analyzed the development and reconstitution of human immune system in these proBLT mice, including human T cells in proBLT mice. Similar to that in the human adaptive immune cells such as TCR + T cells and CD19+ ab thymus organoid of BLT mice, human thymocytes (gated as B cells, as well as innate immune cells such as CD56+ hCD45+) in the human thymus organoid of proBLT mice natural killer (NK) cells, CD11b+ myeloid cells, CD11c+ expressed rearranged human TCR receptors (stained as dendritic cells, and CD14+ monocytes/macrophages, with a ab hTCR +), and they seemed to follow a classical human T composition similar to that of the primary BLT mouse ab cell developmental path from DP (gated as CD4+CD8+) to (Fig. 2B–D). Notably, in the primary BLT mice, we detected CD4 or CD8 SP (gated as CD4+CD8- or CD4-CD8+, re- high numbers of human CD34+ HSPCs (gated as hCD45+ spectively) stages (Fig. 3D) [27]. Large numbers of mature Lin-hCD34+) that were enriched in the bone marrow human T cells (gated as hCD45+hTCR +) were detected in (comprising 10% of total hCD45+ cells) but not in other ab * various peripheral tissues of proBLT mice, including blood, tissues such as liver and spleen (comprising 0.4% of total < spleen, bone marrow, and liver (Figs. 2 and 3). These T cells hCD45+ cells) (Fig. 2B). This observation suggests that the comprised both CD4+ helper and CD8+ cytotoxic T cell initial transplants of human fetal liver CD34+ HSPCs were subsets, at a ratio similar to that observed in the primary able to home to the proper HSC niche in the recipient NSG BLT mice (Fig. 3E, G, H). Therefore, proBLT mice are able mice and expand, while still maintaining their characteristic to support the proper development and systemic reconsti- longevity and multi-potential to repopulate a nearly complete tution of human T cells, at levels comparable to those of the human immune system. Encouragingly, a similarly high per- primary BLT mice. centage of human CD34+ HSPCs (*10%) were detected in the bone marrow of proBLT mice, indicating the ability of these human CD34+ HSPCs to survive the secondary bone marrow Inheritance of human immune cell genetic traits transfer and to repopulate the secondary NSG recipient mice, from BLT to proBLT mice which was key to the success of the proBLT method (Fig. 2C). Certain genetic traits, such as HLA haplotypes that play Reconstitution of human thymus and human T cells important roles in regulating the development and func- in proBLT mice tionality of human T cells, are critical for the study of hu- man immunity [28]. For example, our knowledge of The most attractive feature of the BLT model is its ca- antigen-specific human T cell responses to viral infections pacity to support long-term systemic reconstitution of and cancers are largely based on the studies of HLA-A2- properly matured human T cells, which benefits from the restricted T cell reactions [29]. Humanized BLT mice that presence of an authentic human thymus component [10,11]. harbor a human immune system of HLA-A2 haplotype Post insertion under the kidney capsule of NSG recipient (denoted as BLTA2+) are valuable tools for studying HLA- mice, the implanted human fetal thymus fragments A2-restricted T cell responses and developing T cell-based (*1 mm3 in size; 1–2 pieces per implantation site) grow into immunotherapies. However, the supply of human CD34+ human thymus organoids that support human T cell devel- cells and fetal thymus tissues of HLA-A2 haplotype needed

31 FIG. 2. Reconstitution of multilineage human immune cells in proBLT mice. The experiments were repeated at least three times. Representative results are presented from BLT and proBLT mice at 8–12 weeks post the primary human tissue or secondary BLT tissue transfer (n = 4–5 per experimental group). All tissues were processed into mononuclear cells and lysed of red blood cells for flow cytometry analysis (Materials and Methods section). (A) FACS plots showing the detection of human CD45+ cells (gated as hCD45+ cells of total mononuclear cells) in various lymphoid and immune-regulatory tissues of BLT and proBLT mice. (B) FACS plots showing the detection of multilineage human immune cells (pregated on human CD45+ cells) in various lymphoid and immune-regulatory tissues of BLT mice. Note the homing of human CD34+ cells to the bone marrow of BLT mice. (C) FACS plots showing the detection of multilineage human immune cells (pregated on human CD45+ cells) in the bone marrow of proBLT mice. Note the persistence of human CD34+ cells in proBLT mice post secondary transfer of BLT mice bone marrow cells. (D) Bar graphs showing the comparison of human immune cell composition in the bone marrow of BLT and proBLT mice. Data are presented as mean – SEM. CD34+(34+): CD34+ hematopoietic stem and progenitor cells (gated as hCD45+Lin-hCD34+); T, T cells (gated as hCD45+hTCRab+); B, B cells (gated as hCD45+hCD19+); NK, natural killer cells (gated as hCD45+hCD56+); My, myeloid cells (gated as hCD45+hCD11b+); + + + + Mo, monocytes/macrophages (gated as hCD45 hCD14 ); DC, dendritic cells (gated as hCD45 hCD11c ).

326 PROPAGATING HUMANIZED BLT MICE 7

FIG. 3. Reconstitution of human thymus and human T cells in proBLT mice. The experiments were repeated at least three times. Representative results are presented from BLT and proBLT mice at 8–12 weeks post the primary human tissue or secondary BLT tissue transfer (n = 4–5 per experimental group). (A) Representative image of a human thymus implant growing on a mouse kidney. Data from a representative BLT mouse is shown. (B) Immunohistology analysis showing the H/ E staining and antibody staining of sections of primary human fetal thymus or human thymus organoids from BLT or proBLT mice. Bars: 1,000 mm (40 · magnification); 500 mm (100 · magnification). (C) Histological analysis showing the H/E staining of representative human thymus organoid sections. Data from representative BLT and proBLT mice are shown. Note the presence of cortex and medulla regions that are typical of human thymus, as well as Hassall’s corpuscles (HCs) that are unique to human thymus. Bars: 500 mm (100 · magnification); 5,000 mm (400 · magnification). (D) FACS plots showing the detection of human CD45+ cells in human thymus organoid implants from BLT and proBLT mice. Analysis of human TCRab expression (pregated on human CD45+ cells) and CD4/CD8 co-receptor expression (pregated on hCD45+hTCRab+ cells) are presented. DP, CD4+CD8+ double-positive thymocytes; CD4+, CD4 single-positive thymocytes; CD8+, CD8 single- positive thymocytes. (E) FACS plots showing the detection of human T cells (gated as hCD45+hTCRab+) and their CD4/ CD8 subsets (further gated as CD4+ or CD8+) in the blood of BLT and proBLT mice. (F–H) Scatter plots showing the presence of human T cells (gated as hCD45+hTCRab+; F) and their CD4 (further gated as CD4+; G) and CD8 (further gated as CD8+; H) subsets in the blood of BLT and proBLT mice. Data are presented as individual mouse measurements and mean – SEM of all mice in an experimental group. H/E, hematoxylin and eosin; ns, not significant. to produce such BLTA2+ mice are especially limiting. We immune system, including the human T cell compartment proposed to overcome this hurdle by maximizing the utili- that expressed HLA-A2. zation of the limited supply of HLA-A2+ human tissues, through expanding a small number of established BLTA2+ A2+ Persistence of human immune cell gene mice into a large colony of proBLT mice without the modifications from BLT to proBLT mice need for additional primary human tissues. As shown in Fig. 4, the resulting proBLTA2+ mice faithfully inherited the Humanized BLT mice are potent tools that are used to study HLA-A2 haplotype genetic trait and reconstituted a human gene-modified human immune cell therapies, especially gene-

33 8 SMITH ET AL.

FIG. 4. Inheritance of human immune cell genetic traits from BLT to proBLT mice. HLA-A2+ human fetal liver CD34+ cells and matching fetal thymus were used to generate the primary BLT mice (denoted as BLTA2+), which were then utilized to generate the secondary proBLT mice (denoted as proBLTA2+). The experiments were repeated at least three times. Representative results are presented from BLT and proBLT mice at 8–12 weeks post the primary human tissue or secondary BLT tissue transfer (n = 4–5 per experimental group). (A) FACS plots showing the inheritance of HLA-A2 haplotype on reconstituted human immune cells (pregated as hCD45+) from the spleen of BLTA2+ and proBLTA2+ mice. HLA-A2- and HLA-A2+ PBMCs were included as controls. PBMC, peripheral blood mononuclear cells from healthy human donors. T, human T cells (gated as hCD45+hTCRab+). (B) Scatter plot showing the detection of human T cells (gated as hCD45+hTCRab+) in the blood of BLTA2+ and proBLTA2+ mice. Data are presented as individual mouse measurements and mean – SEM of all mice in an experimental group. ns, not significant. modified HSC therapies for treating diseases such as cancer, immune system was stably passed from BLT to proBLT HIV, or primary immune deficiencies [30–34]. To evaluate mice, making them suitable for studying gene-modified hu- whether the proBLT approach may be useful for such appli- man immune cell gene therapies, especially gene-modified cations, we transduced human CD34+ cells with a lentivector HSC therapies (Fig. 5). Therefore, these proBLT mice can pMND-EGFP encoding an enhanced green fluorescence be considered expanded ‘‘clones’’ of the established primary (EGFP) reporter gene, and we then used these genetically BLT mice. Through bypassing the need for additional human modified CD34+ cells to generate BLT mice (Fig. 5A). In the CD34+ cells and fetal thymus tissues, the proBLT approach resulting BLT mice (denoted as BLTEGFP), we observed high provides an opportunity to overcome a critical hurdle to uti- expression of the EGFP transgene in a high portion of human lizing the BLT humanized mouse model and enables its more immune cells (gated as hCD45+EGFP+ comprising *60% of widespread use as a valuable preclinical research tool. total hCD45+ cells; Fig. 5B). proBLT mice generated from There are two key factors that make the proBLT approach these BLTEGFP mice, denoted as proBLTEGFP, repopulated a successful. One is the persistence of human CD34+ HSPCs human immune system that persistently expressed high levels through secondary bone marrow transfer in NSG recipient of EGFP transgene with the appearance of hCD45+ cells mice; the other is the regeneration of a functional human starting from 4 weeks post secondary bone marrow transfer thymus structure in the secondary NSG recipient mice (Fig. 5B, C). Transgene expression was detected in multilineages through implanting the human thymus organoids harvested of human immune cells, including CD4+ helper and CD8+ cy- from the primary BLT mice. Despite their original sources totoxic human T cells (gated as hCD45+hTCRab+CD4+ or (e.g., fetal liver, cord blood, adult bone marrow, or adult G- hCD45+hTCRab+CD8+, respectively; Fig. 5D). Moreover, CSF-mobilized peripheral blood), human CD34+ cells post the levels of human immune cell gene modifications were adoptive transfer always preferentially homed to the bone very consistent among individual proBLT mice and were marrow of NSG mice, likely because NSG mouse bone similar to those of the primary BLTEGFP mice (*60% of total marrow provides a nurturing environment that supports the hCD45+ cells; Fig. 5C). Therefore, proBLT mice are suitable engraftment and long-term maintenance of these cells to support large-scale studies of gene-modified human im- (Fig. 2B) [4]. The ability of BLT bone marrow-experienced mune cell therapies. human CD34+ HSPCs to survive the secondary bone mar- row transfer, while maintaining their longevity and multi- Discussion potential, validates the supporting function of the NSG mouse bone marrow niche. Meanwhile, the ability of the In this article, we describe a new method for propagating human thymus organoids generated in BLT mice to develop humanized BLT mice for the study of human immunology into a functional human thymus structure in the proBLT and immunotherapy. Through secondary transfer of bone mice is intriguing (Fig. 3). A human thymus graft is critical marrow cells and human thymus implants from BLT mice for the proper development, functional maturation, and sys- into naı¨ve NSG recipient mice, we were able to expand one temic reconstitution of human T cells in humanized mice [9]. primary BLT mouse into a colony of 4–5 proBLT mice in 6– In addition to thymocytes that are derived from HSPCs, a 8 weeks (Fig. 1). These proBLT mice reconstituted human functional human thymus comprises thymic epithelial/den- immune cells, including T cells, at levels comparable to dritic cells that mediate the positive and negative selections those of their primary BLT donor mice (Figs. 2 and 3). They of human T cells, as well as the programming of special also faithfully inherited the human immune genetic traits human T cell sublineages such as CD4+CD25+FoxP3+ reg- from their donor BLT mice, such as the HLA-A2 haplotype ulatory T cells (Tregs) [27]. In our experiments, we observed that is of special interest for studying antigen-specific T cell a typical cortex/medulla structure in human thymus orga- activities and T cell-based immunotherapies (Fig. 4). More- noids harvested from both BLT and proBLT mice, suggest- over, an EGFP reporter gene engineered into the human ing that human thymocytes and thymic epithelial/dendritic

34 PROPAGATING HUMANIZED BLT MICE 9

FIG. 5. Persistence of human immune cell gene modifications from BLT to proBLT mice. Human CD34+ cells transduced with pMND-EGFP lentiviruses were used to generate the primary BLT mice modified with the EGFP reporter gene (denoted as BLTEGFP), which were then utilized to generate the secondary proBLT mice (denoted as proBLTEGFP). The experiments were repeated at least three times. Representative results are presented (n = 4–5 per experimental group). (A) Schematic of the pMND-EGFP lentiviral vector. DLTR, self-inactivating long-term repeats; MNDU3, internal promoter derived from the MND retroviral LTR U3 region; C, packaging signal with the splicing donor and splicing acceptor sites; RRE, rev-responsive element; cPPT, central polypurine tract; WPRE, woodchuck responsive element; EGFP, enhanced green fluorescence protein reporter gene. (B) Representative FACS plots showing the detection of EGFP reporter gene expression in human immune cells (gated as hCD45+) from the blood of BLTEGFP and proBLTEGFP mice at 8 weeks post the primary human tissue or secondary BLT tissue transfer. (C) Scatter plot showing the time course detection of EGFP reporter gene expression in human immune cells (gated as hCD45+) from the blood of proBLTEGFP mice post secondary BLT tissue transfer (n = 4). Data are presented as individual mouse measurements and mean – SEM of all experimental mice at the indicated time points. Note the stable and persistent expression of EGFP transgene from the primary BLTEGFP to the secondary proBLTEGFP mice. (D) Representative FACS plots of blood from proBLTEGFP mice at 8 weeks post secondary BLT tissue transfer, showing the detection of high EGFP transgene expression in various reconstituted human immune cells (gated as hCD45+hTCRab+/-), in particular in human T cells (gated as hCD45+hTCRab+) and their CD4/CD8 subsets (further gated as CD4+ or CD8+). cells were present and organized properly in these organoids pro-proBLT mice did not yield satisfactory results. Recon- (Fig. 3B). In particular, we observed abundant numbers of stitution of human immune cells in the pro-proBLT mice HCs in the medulla of these human thymus organoids (Fig. 3C). was low and variable. It seemed that the human CD34+ cells HCs are structures unique to human thymus, formed from had exhausted their long-term potential post the tertiary eosinophilic type VI epithelial reticular cells arranged con- bone marrow transfer, a phenomenon similar to but more centrically, and have been implicated in the development of severe than that has been observed for mouse HSCs post human Tregs [26]. Thymic stromal lymphopoietin expressed series bone marrow transfers in the mouse model. It has by thymic epithelial cells within the HCs has been indicated to been well recognized in the mouse model that series of bone activate thymic CD11+ dendritic cells that then mediate the marrow transfers induce stress and impair the longevity of secondary positive selection of human Tregs [26]. Taken to- HSCs. Furthermore, in the BLT humanized mouse model, gether, these observations verified the presence of an au- the mouse bone marrow environment does not provide the thentic human thymus environment in proBLT mice that is optimal support for the long-term maintenance of human responsible for the functional reconstitution of human T cells HSCs. Supplementing the primary BLT mice and proBLT in these mice. mice with human cytokines that are important for human Based on the success of generating proBLT mice, it is HSC maintenance, such as Flt3 ligand, SCF, TPO, and IL-3, intriguing to propose that the tertiary transfer of bone may improve the longevity of engrafted human CD34+ cells marrow cells and human thymus organoid harvested from and allow for the further propagation of proBLT mice. If the proBLT mice into NSG mice may further expand the possible, such a preclinical animal model will allow the BLT colony size by another four- to five-fold, allowing the continuous ‘‘passage’’ of human immune cells in living propagation of a single primary BLT mouse to *5 pro-BLT animals for an extended period of time, therefore maxi- mice in 6–8 weeks and then to *25 pro-proBLT mice in mizing the research value of a limited supply of primary 3–4 months. Unfortunately, our initial attempts to generate human immune cells and tissues.

35 10 SMITH ET AL.

proBLT mice are particularly valuable for the study of D.J.S. is a predoctoral fellow supported by the UCLA Tumor human immune systems with special genetic traits. These Immunology Training Grant (T32 CA009056). This work genetic traits often pose particular restrictions on the supply was supported by an UCLA Center for AIDS Research Grant of human tissues that can be used. One example is the HLA (NIH/NIAID AI028697, to J.Z.), a National Institutes of haplotype. For instance, the study of HLA-A2-restricted T Health (NIH) Director’s New Innovator Award (DP2 cell responses is of special interest for cancer immunother- CA196335, to L.Y.), a STOP CANCER Research Career apy research, whereas the study of HLA-B57-restricted T Development Award (to L.Y.), a GTSN Challenge Award cell responses is of special interest for HIV latency research for Lethal Prostate Cancer (to L.Y.), and a CIRM 2.0 Part- [35,36]. Generation of BLT mice for such studies requires nering Opportunity for Translational Research Projects HLA-A2+ or HLA-B57+ human CD34+ HSPCs as well as Award (TRAN1-08533, to L.Y.). HLA haplotype-matched fetal thymus tissues. Another ex- ample is monogene deficiencies of the human hematopoietic Author Disclosure Statement system. For instance, adenosine deaminase (ADA) defi- ciency is of interest for studying human ADA-deficient se- No competing financial interests exist. vere combined immunodeficiency (SCID), whereas b-globin deficiency is of interest for studying human sickle cell dis- ease [20–22]. Over the past decades, gene-corrected autol- References ogous HSC transfer has become a promising therapy for 1. Shultz LD, MA Brehm, JV Garcia-Martinez and DL these monogene deficiency-induced diseases and inspires Greiner. (2012). Humanized mice for immune system in- further investigations [32]. Humanized animal models are vestigation: progress, promise and challenges. Nat Rev valuable tools for the preclinical development of these Immunol 12:786–798. therapies, but their application is greatly limited by the small 2. Rongvaux A, H Takizawa, T Strowig, T Willinger, EE number of CD34+ cells that can be collected from patients, Eynon, RA Flavell and MG Manz. (2013). Human hemato- which are often too few to engraft enough animals for lymphoid system mice: current use and future potential for meaningful studies [20–22]. By expanding established BLT medicine. Annu Rev Immunol 31:635–674. mice and reducing the need for additional primary human 3. Theocharides AP, A Rongvaux, K Fritsch, RA Flavell and tissues, the proBLT approach provides an attractive solution MG Manz. (2016). Humanized hemato-lymphoid system to overcome this critical hurdle and makes the humanized mice. Haematologica 101:5–19. BLT mouse model suitable for the preclinical study of such 4. Brehm MA, A Cuthbert, C Yang, DM Miller, P DiIorio, J hematopietic stem cell-based gene therapies. Laning, L Burzenski, B Gott, O Foreman, et al. (2010). Parameters for establishing humanized mouse models to Despite its valuable research potential, the BLT model still study human immunity: analysis of human hematopoietic has its own limitations and can be improved further to make it stem cell engraftment in three immunodeficient strains of more representative of the human immune system in terms of mice bearing the IL2rgamma(null) mutation. Clin Immunol the composition of various lineages of immune cells and their 135:84–98. functions. These next-generation humanized mouse models 5. Takenaka K, TK Prasolava, JC Wang, SM Mortin-Toth, S utilize recipient mice such as NSG or BRG that are further Khalouei, OI Gan, JE Dick and JS Danska. (2007). Poly- genetically modified to allow for enhanced human immune morphism in Sirpa modulates engraftment of human he- cell reconstitution. Such modifications include the deficiency matopoietic stem cells. Nat Immunol 8:1313–1323. of murine c-Kit gene that supports improved human CD34+ 6. Mosier DE, RJ Gulizia, SM Baird and DB Wilson. (1988). HSPC engraftment [37,38], the addition of human transgenes Transfer of a functional human immune system to mice with such as SIRPa to improve overall human hematopoietic cell severe combined immunodeficiency. Nature 335:256–259. engraftment [39], or knock-in of human immune regulatory 7. Traggiai E, L Chicha, L Mazzucchelli, L Bronz, JC Pif- genes such as TPO, IL-3, GM-CSF (granulocyte-macrophage faretti, A Lanzavecchia and MG Manz. (2004). Develop- colony-stimulating factor), and M-CSF (macrophage colony- ment of a human adaptive immune system in cord blood stimulating factor) that promote the development and func- cell-transplanted mice. Science 304:104–107. tion of human monocytes, macrophages, and NK cells [40]. 8. Chicha L, R Tussiwand, E Traggiai, L Mazzucchelli, L The proBLT approach should also be applicable to expand Bronz, JC Piffaretti, A Lanzavecchia and MG Manz. (2005). BLT mice that are produced with these advanced recipient Human adaptive immune system Rag2-/-gamma(c)-/- mice. mice and support the studies of a human immune system that Ann N Y Acad Sci 1044:236–243. more closely resembles the human situation. 9. McCune JM, R Namikawa, H Kaneshima, LD Shultz, M Lieberman and IL Weissman. (1988). The SCID-hu mouse: murine model for the analysis of human hematolymphoid Acknowledgments differentiation and function. Science 241:1632–1639. 10. Melkus MW, JD Estes, A Padgett-Thomas, J Gatlin, PW The authors are grateful to the University of California Denton, FA Othieno, AK Wege, AT Haase and JV Garcia. Los Angeles (UCLA) animal facility for providing animal (2006). Humanized mice mount specific adaptive and innate support, the UCLA Translational Pathology Core Laboratory immune responses to EBV and TSST-1. Nat Med 12:1316– (TPCL) for providing immunohistology support, and Dr. 1322. Gay Crooks for providing flow cytometry reagents. They 11. Lan P, N Tonomura, A Shimizu, S Wang and YG Yang. also thank Dr. Rachel Steward and the FPA Women’s (2006). Reconstitution of a functional human immune Health, and the UCLA AIDS Institute/CFAR Virology Core/ system in immunodeficient mice through combined human Gene and Cell Therapy Core/Humanized Mouse Core for fetal thymus/liver and CD34+ cell transplantation. Blood providing human cells/tissues and humanized mice services. 108:487–492.

36 PROPAGATING HUMANIZED BLT MICE 11

12. Wege AK, MW Melkus, PW Denton, JD Estes and JV 27. Spits H. (2002). Development of alphabeta T cells in the Garcia. (2008). Functional and phenotypic characterization human thymus. Nat Rev Immunol 2:760–772. of the humanized BLT mouse model. Curr Top Microbiol 28. Choo SY. (2007). The HLA system: genetics, immunology, Immunol 324:149–165. clinical testing, and clinical implications. Yonsei Med J 13. Brainard DM, E Seung, N Frahm, A Cariappa, CC Bailey, 48:11–23. WK Hart, HS Shin, SF Brooks, HL Knight, et al. (2009). 29. Tscharke DC, NP Croft, PC Doherty and NL La Gruta. Induction of robust cellular and humoral virus-specific (2015). Sizing up the key determinants of the CD8(+) T cell adaptive immune responses in human immunodeficiency response. Nat Rev Immunol 15:705–716. virus-infected humanized BLT mice. J Virol 83:7305–7321. 30. Rossi JJ, CH June and DB Kohn. (2007). Genetic therapies 14. Marsden MD and JA Zack. (2015). Studies of retroviral against HIV. Nat Biotechnol 25:1444–1454. infection in humanized mice. Virology 479–480:297–309. 31. Gschweng E, S De Oliveira and DB Kohn. (2014). He- 15. Karpel ME, CL Boutwell and TM Allen. (2015). BLT matopoietic stem cells for cancer immunotherapy. Immunol humanized mice as a small animal model of HIV infection. Rev 257:237–249. Curr Opin Virol 13:75–80. 32. Kuo CY and DB Kohn. (2016). Gene Therapy for the 16. Garcia JV. (2016). In vivo platforms for analysis of HIV Treatment of Primary Immune Deficiencies. Curr Allergy persistence and eradication. J Clin Invest 126:424–431. Asthma Rep 16:39. 17. Tager AM, M Pensiero and TM Allen. (2013). Recent ad- 33. Wang CX and PM Cannon. (2016). The clinical applica- vances in humanized mice: accelerating the development of tions of genome editing in HIV. Blood 127:2546–2552. an HIV vaccine. J Infect Dis 208 Suppl 2:S121–S124. 34. Petz LD, JC Burnett, H Li, S Li, R Tonai, M Bakalinskaya, 18. Shimizu S, P Hong, B Arumugam, L Pokomo, J Boyer, N EJ Shpall, S Armitage, J Kurtzberg, et al. (2015). Progress Koizumi, P Kittipongdaja, A Chen, G Bristol, et al. (2010). toward curing HIV infection with hematopoietic cell A highly efficient short hairpin RNA potently down- transplantation. Stem Cells Cloning 8:109–116. regulates CCR5 expression in systemic lymphoid organs in 35. Kershaw MH, JA Westwood and PK Darcy. (2013). Gene- the hu-BLT mouse model. Blood 115:1534–1544. engineered T cells for cancer therapy. Nat Rev Cancer 13: 19. Vatakis DN, RC Koya, CC Nixon, L Wei, SG Kim, P 525–541. Avancena, G Bristol, D Baltimore, DB Kohn, et al. (2011). 36. Zaunders J and D van Bockel. (2013). Innate and Adaptive Antitumor activity from antigen-specific CD8 T cells Immunity in Long-Term Non-Progression in HIV Disease. generated in vivo from genetically engineered human hema- Front Immunol 4:95. topoietic stem cells. Proc Natl Acad Sci U S A 108:E1408– 37. Cosgun KN, S Rahmig, N Mende, S Reinke, I Hauber, C E1416. Schafer, A Petzold, H Weisbach, G Heidkamp, et al. 20. Carbonaro DA, L Zhang, X Jin, C Montiel-Equihua, S (2014). Kit regulates HSC engraftment across the human- Geiger, M Carmo, A Cooper, L Fairbanks, ML Kaufman, mouse species barrier. Cell Stem Cell 15:227–238. et al. (2014). Preclinical demonstration of lentiviral vector- 38. McIntosh BE, ME Brown, BM Duffin, JP Maufort, DT mediated correction of immunological and metabolic ab- Vereide, Slukvin, II and JA Thomson. (2015). Non- normalities in models of adenosine deaminase deficiency. irradiated NOD,B6.SCID Il2rgamma-/- Kit(W41/W41) Mol Ther 22:607–622. (NBSGW) mice support multilineage engraftment of hu- 21. Hoban MD, GJ Cost, MC Mendel, Z Romero, ML Kaufman, man hematopoietic cells. Stem Cell Rep 4:171–180. AV Joglekar, M Ho, D Lumaquin, D Gray, et al. (2015). 39. Strowig T, A Rongvaux, C Rathinam, H Takizawa, C Correction of the sickle cell disease mutation in human he- Borsotti, W Philbrick, EE Eynon, MG Manz and RA Fla- matopoietic stem/progenitor cells. Blood 125:2597–2604. vell. (2011). Transgenic expression of human signal regu- 22. Romero Z, F Urbinati, S Geiger, AR Cooper, J Wherley, latory protein alpha in Rag2-/-gamma(c)-/- mice improves ML Kaufman, RP Hollis, RR de Assin, S Senadheera, et al. engraftment of human hematopoietic cells in humanized (2013). Beta-globin gene transfer to human bone marrow mice. Proc Natl Acad Sci U S A 108:13218–13223. for sickle cell disease. J Clin Invest 123:3317–3330. 40. Rongvaux A, T Willinger, J Martinek, T Strowig, SV Gearty, 23. Smith DJ, S Liu, S Ji, B Li, J McLaughlin, D Cheng, ON LL Teichmann, Y Saito, F Marches, S Halene, et al. (2014). Witte and L Yang. (2015). Genetic engineering of hema- Development and function of human innate immune cells in topoietic stem cells to generate invariant natural killer T a humanized mouse model. Nat Biotechnol 32:364–372. cells. Proc Natl Acad Sci U S A 112:1523–1528. 24. Giannoni F, CL Hardee, J Wherley, E Gschweng, S Se- Address correspondence to: nadheera, ML Kaufman, R Chan, I Bahner, V Gersuk, et al. Lili Yang, PhD (2013). Allelic exclusion and peripheral reconstitution by Department of Microbiology, Immunology TCR transgenic T cells arising from transduced human he- and Molecular Genetics matopoietic stem/progenitor cells. Mol Ther 21:1044–1054. M/C 957243 25. Yang L, H Yang, K Rideout, T Cho, KI Joo, L Ziegler, A University of California Elliot, A Walls, D Yu, D Baltimore and P Wang. (2008). Los Angeles, CA 90095 Engineered lentivector targeting of dendritic cells for in vivo immunization. Nat Biotechnol 26:326–334. E-mail: [email protected] 26. Watanabe N, YH Wang, HK Lee, T Ito, YH Wang, W Cao and YJ Liu. (2005). Hassall’s corpuscles instruct dendritic Received for publication July 1, 2016 cells to induce CD4+CD25+ regulatory T cells in human Accepted after revision September 8, 2016 thymus. Nature 436:1181–1185. Prepublished on Liebert Instant Online XXXX XX, XXXX

37

CHAPTER 4:

Invariant natural killer T cells generated in vivo from genetically engineered human hematopoietic stem cells

38 INTRODUCTION

A new era of cancer immunotherapy

Cancer immunotherapy aims to harness and enhance the native power of the human immune system to fight cancer. Compared to other cancer treatments, such as surgery, radiotherapy, and chemotherapy, immunotherapy has the potential benefits of fewer side effects, longer protection, and broader application to various forms of cancer. Attracted by these features, generations of immunologists and clinicians have been working hard since the 1990s to develop immunotherapy in the fight against cancer. Significant progress has been made over the past decade, including the utilization of adoptive TIL (tumor-infiltrating-lymphocyte) therapy for treating melanoma; the approval of Provenge (a dendritic cell-based adoptive cell therapy) for treating prostate cancer; the success of CD19-specific CAR (chimeric antigen receptor) modified adoptive T cell therapy in treating B-ALL (B-cell Acute Lymphocytic Leukemia); and the impressive tumor regression induced by blocking PD-1 and PD-L1 (molecular “brakes” regulating immune responses) in patients with melanoma, kidney cancer, and lung cancer [8, 9].

These breakthroughs mark a new era of cancer immunotherapy, highlight its promise as the next generation of cancer medicine, and set the stage for the development of more innovative immunotherapies for cancer.

Invariant natural killer T cells for cancer immunotherapy: opportunities and challenges

Invariant natural killer T (iNKT) cells are a small population of αβ T lymphocytes highly conserved from mice to humans [28]. Like conventional αβ T cells, iNKT cells differentiate from hematopoietic stem cells (HSCs) and develop in the thymus, but these cells also differ from conventional T cells in several important ways: they recognize glycolipid antigens presented by

39 the non-classical monomorphic MHC (major histocompatibility complex) molecule CD1d; they display select natural killer cell markers; and they express semi-invariant T cell receptors

(identical α chains paired with a limited selection of β chains), thus gaining the name invariant natural killer T cells. These cells have several unique features that make them exceedingly attractive agents for cancer immunotherapy: they can be stimulated by the synthetic ligand α- galactosylceramide (α-GalCer), which is already approved for use in the clinic; post-stimulation, they can immediately release a vast amount of cytokines, mainly IFN-γ, to induce anti-tumor responses mediated by natural killer (NK) cells and cytotoxic T lymphocytes (CTLs); and they can target multiple cancers independent of tumor-antigen restriction. This targeting feature defines iNKT cells as general therapeutic agents for many cancers.

Invariant natural killer T cells acquire their potent functions through a unique 2-stage development path: stage 1 to gain iNKT cell T cell receptor (TCR) expression in the thymus, and stage 2 to gain functional maturation in the periphery that is associated with the up-regulation of select NK cell markers [32]. Functionally matured iNKT cells display a memory T cell phenotype and can react immediately to stimulations; they even have pre-synthesized and stored cytokine mRNA ready for use [34]. At the single cell level, an iNKT cell can produce over 100 times more cytokine than a conventional T cell. It is worth noting that engineering conventional

T cells to express an iNKT cell TCR does not impart the special capacities of an iNKT cell; these functions can only be acquired through the unique iNKT cell development path taken in the thymus.

Invariant natural killer T cells have the remarkable capacity to target multiple types of cancer independent of tumor antigen and MHC restrictions [34]. The iNKT cells recognize glycolipid antigens presented by non-polymorphic CD1d, which frees them from MHC

40 restriction. Although the natural ligands of iNKT cells remain to be decisively identified, it is suggested that iNKT cells can recognize certain conserved glycolipid antigens derived from many tumor tissues. The cells can be stimulated through recognizing these glycolipid antigens that are either directly presented by CD1d+ tumor cells, or indirectly cross-presented by tumor infiltrating macrophages or dendritic cells (DCs) in the case of CD1d- tumors. Thus, iNKT cells can respond to both CD1d+ and CD1d- tumors. Most attractively, iNKT cells can be stimulated by the synthetic ligand α-GalCer [28]. Previous clinical trials have demonstrated the safety and efficacy of using dendritic cells loaded with α-GalCer to stimulate iNKT cells in vivo, thus providing a powerful clinical tool to activate iNKT cell anti-tumor responses and bypassing the need for tumor antigen recognition [37].

Post stimulation with tumor antigens or α-GalCer loaded dendritic cells, iNKT cells can employ multiple mechanisms to attack tumor cells [34, 35]. The iNKT cells can directly kill

CD1d+ tumor cells through cytotoxic action, but their most potent anti-tumor capabilities come from their immune adjuvant effects. Invariant natural killer T cells remain quiescent prior to stimulation, but after stimulation they immediately produce large amounts of cytokines, mainly

IFN-γ. IFN-γ activates natural killer cells to target and kill MHC-negative tumor cells, while simultaneously activating dendritic cells that then stimulate CTLs to target and kill MHC- positive tumor target cells. Therefore, iNKT-cell - induced anti-tumor immunity can effectively target multiple types of cancer independent of tumor antigen and MHC restrictions, therefore effectively blocking tumor immune escape and minimizing the chance of tumor recurrence. It is worth noting that because of the tumor recognition capacity of NK cells and CTLs, iNKT cell induced anti-tumor immunity will only target cancerous, but not healthy, tissues [34, 35].

41 Attracted by the potent and broad anticancer functions of iNKT cells, scientists and clinicians have conducted a series of clinical trials utilizing iNKT cells to treat various forms of cancer, ranging from solid tumors to blood cancers [34, 37]. These clinical trials have utilized α-

GalCer alone, α-GalCer loaded dendritic cells, ex vivo expanded iNKT cells, or a combination of these therapies and shown these treatments to be safe and well tolerated. Several recent trials reported encouraging anti-tumor immunity in patients with non-small cell lung cancer and head and neck squamous cell carcinomas, attesting to the potential of iNKT cell-based immunotherapies [34]. However, most other trials yielded unsatisfactory results. Overall, these trials have all worked through the direct stimulation or ex vivo expansion of a patient’s endogenous iNKT cells, thus yielding only short-term and limited clinical benefits to a small number of patients. The extremely low frequencies of iNKT cells in cancer patients (~0.001-

0.1% in blood), as well as the rapid depletion of these cells post stimulation, are considered the major factors limiting the success of these trials [43]. In order to unleash the full potential of iNKT cells for cancer immunotherapy, innovative therapies that can overcome these limitations are in high demand.

The rise of stem cell-engineered immunotherapy (SEI)

Hematopoietic stem cells (HSCs), also commonly known as blood stem cells, reside in the bone marrow and give rise to all blood lineage cells, including immune cells. HSCs have the remarkable capacity to live as long as their hosts, to self-renew, and to constantly replenish immune cells [48]. These unique properties make HSCs ideal for engineering long-term immunotherapy.

42 For over a decade, our research team has been intrigued by the promise of engineering hematopoietic stem cells for immunotherapy. In early 2000, Dr. Lili Yang proposed a project to generate anti-cancer T cells through the genetic modification of HSCs with tumor-specific TCR genes. This study was highly successful and was able to demonstrate for the first time that TCR gene-modified HSCs could develop into tumor-specific T cells, and that these stem cell- engineered T cells could effectively target tumors to provide life-long protection against cancer in mice [49]. A similar study was performed in the BLT humanized mouse model and demonstrated that the principle could be applied in human cells as well [45]. A clinical trial proposed by Dr. Antoni Ribas at UCLA, which recently received FDA-approval, will utilize the stem cell engineered T cell therapy in patients with melanoma.

SEI provides an ideal solution to overcome the limitations of the current iNKT cell-based immunotherapies by providing patients with a lifelong supply of therapeutic levels of stem cell- engineered iNKT cells. In this project, we propose to develop a pre-clinical model of stem cell- based iNKT cell therapy. This model will then be used to collect the necessary safety and efficacy data in order to apply for a clinic trial to utilize the engineering of stem cells to generate iNKT cells to treat cancer in patients.

Clinical Rationale

There is compelling evidence to suggest a significant role of iNKT cells in tumor surveillance in mice, in which iNKT cell defects predispose mice to develop cancer and the adoptive transfer or stimulation of iNKT cells can provide protection against cancer [35, 43]. In humans, iNKT cell frequency is decreased in patients with solid tumors (including melanoma, colon, lung, breast, and head and neck cancers) and blood cancers (including leukemia, multiple

43 myeloma, and myelodysplastic syndroms), while increased iNKT cell numbers are associated with a better clinical prognosis [43]. Moreover, there are interesting associative studies suggesting that iNKT cells promote graft-versus-leukemia effects and suppress graft-versus-host disease after HSC transplantation [52, 53]. There are also instances wherein the administration of

α-GalCer loaded dendritic cells and ex vivo expanded iNKT cells has led to promising clinical benefits in patients with lung cancer and head and neck cancer, although the increases of iNKT cells have been transient and the clinical benefits have been short-term, likely due to the short lifetime of iNKT cells post stimulation [34, 36]. Thus, it is plausible to propose that increasing iNKT cells to therapeutic levels throughout the life of cancer patients via HSC engineering would provide patients with the best chance to exploit the full potential of iNKT cells to battle their diseases.

RESULTS AND DISCUSSION

Cloning of iNKT cell TCR Genes and Construction of Lentiviral Delivery Vectors

We used a robust and high-throughput single-cell TCR cloning technology recently established in our laboratories to obtain iNKT cell TCR genes (Materials and Methods). Single iNKT cells were sorted from healthy donor peripheral blood mononuclear cells (PBMCs) using flow cytometry based on a stringent collection of surface markers gated as human TCRαβ+TCR

Vα24-Jα18+TCR Vβ11+CD161+ (Fig. 1A) [32]. We included TCR Vβ11 staining to focus on the dominant Vβ11+ population of human iNKT cells [28]. The sorted single iNKT cells were then subjected to TCR cloning (Fig. 1B). Several verified iNKT TCR α and β chains, linked by a P2A self-cleaving sequence to allow for co-expression, were inserted into an MND-based lentiviral vector to yield TCR gene delivery vectors (Fig. 2A). One vector that mediated high expression of

44

a high-affinity iNKT cell TCR was selected for the follow-up studies. An EGFP gene linked by an F2A self-cleaving sequence was included in another vector to allow for robust tracking of transduced cells (Fig. 2A) [54].

Generation of functional iNKT cells through TCR gene engineering of HSCs

Humanized BLT mice were generated following previously established procedures, with certain modifications (Materials and Methods, Fig. 2A). Briefly, cryopreserved human CD34+ cells were thawed and cultured for no more than 48 hours in X-VIVO-15 medium containing recombinant human Flt3 ligand, SCF, TPO, and IL-3. Viral transduction was performed at 24 hours by adding concentrated lentiviral vectors directly to the culture medium. At around 48 hours, CD34+ cells were collected and i.v. injected into NOD/SCID/IL-2Rγc-/- (NSG) recipient mice (~0.5-1 x106 CD34+ cells per recipient) that had received 270 rads of total body irradiation.

During the same period 1-2 fragments of cryo-recovered human fetal thymus (~1mm3), as well as donor-matched fetal liver CD34- cells when available, were implanted under the kidney capsule of each recipient NSG mouse. The resulting BLT mice were allowed to reconstitute a human immune system while periodic bleedings were performed to monitor the presence of

45 human immune cells (gated as hCD45+), human T cells (gated as hCD45+ hTCRαβ+), and human iNKT cells (gated as hCD45+ hTCRαβ+ hTCR Vα24Jα18+).

We were able to detect human immune cells in the BLT control (without transduction) and BLT-iNKT mice at the week 6 bleeding time-point after human tissue transplant (Fig 2B and

2C). The human immune cell reconstitution gradually increased until it peaked and stabilized at week 12. There was a high percentage of human T cells already detected at the week 6 time- point that continued to increase until stabilizing at week 12 (Fig 2B and 2C). Human iNKT cells were not detected in the BLT-iNKT mice until 12 weeks after human tissue transplant and gradually increased and stabilized by the 16 week time-point (Fig 2B and 2C). Significant numbers of human iNKT cells were not detected in the control BLT mice by flow cytometry at any time-point.

There is an obvious discrepancy in the data: the pre-rearranged iNKT cell TCR inserted into the transduced CD34+ cell genome should cause the generation of iNKT cells before conventional T cells. We have observed iNKT cell generation preceding conventional T cell production when generating BLT mice utilizing fetal liver CD34+ cells (data not shown). The presence of such a high percentage of T cells at the week 6 time-point is also unprecedented, as

T cell generation does not generally increase to appreciable levels until 8 weeks post human tissue transplant. We hypothesize that the T cells detected at the week 6 time-point are not generated by the stem cell graft, but are actually thymocytes from the fetal thymus implant that have temporarily expanded in the peripheral blood. These cells are replaced by the immune cells generated from the stem cell graft by week 12 as evidenced by the presence of iNKT cells at that time-point.

46 47

In order to determine systemic immune reconstitution in these humanized BLT mice immune-homing tissues were collected and processed to isolate lymphocytes. The lymphocytes were then stained and analyzed by flow cytometry to assay the generation of engineered iNKT cells (Fig. 2D). Human iNKT cells were detected in all tested tissues collected from BLT-iNKT mice. Significant levels of iNKT cells were not detected in any of the tested tissues of control

BLT mice (Fig. 2E).

Engineered iNKT cell levels were highest in the liver, the tissue with the highest endogenous level of iNKT cells, which makes a strong case that the engineered iNKT cells are capable of tissue-specific homing. Significant numbers of iNKT cells were also detected in the peripheral blood and spleen as a compartment of αβ T lymphocytes. Engineered iNKT cells could also be detected in the bone marrow (BM), although at a much lower level. It is important to note the presence of iNKT cells in the human thymus implant (Thy) of these mice.

This indicates that the engineered iNKT cells are undergoing selection and education; this is critical as the special characteristics of iNKT cells are gained during their development in the thymus through the expression and binding of their semi-invariant TCR.

48 Development of engineered HSCs

The ability of these engineered HSCs to regenerate and self-renew was assayed by performing bone-marrow transplants along with human thymus implant transplants into recipient pre-conditioned NSG mice in order to generate proBLT mice [55]. These proBLT mice were allowed to reconstitute a human immune system and bled to monitor the continued generation of human iNKT cells (Fig. 3A). The data shown indicate the continued generation of stable levels of iNKT cells granted from the “parental” BLT-iNKT mice (Fig. 3B).

The continued generation of iNKT cells by the engineered HSCs indicates the transduction of these cells and the expression of the iNKT cell TCR does not negatively affect their survival. Furthermore, the ability of the engineered HSCs to reconstitute and differentiate in proBLT mice is promising as this will be the protocol of an HSC-engineered iNKT cell immunotherapy treatment. The ability to generate iNKT proBLT mice also greatly aids in the collection of further data, reducing the need to generate as many BLT-iNKT mice, which requires precious tissues and lentivirus.

Development of engineered iNKT cells

The capability of the iNKT cell TCR to induce iNKT cell generation in transduced cells was assayed by generating BLT-iNKT mice with a lentiviral vector including an EGFP reporter gene (Fig. 2A). These mice were bled and the immune cell lineages present in the peripheral blood in either GFP- or GFP+ cells were determined by flow cytometry (Fig. 3C and 3D). The data indicate the expression of the iNKT cell TCR is fully capable of instructing transduced cells to become iNKT cells.

49 50

In order to further probe the phenotype of these engineered iNKT cells they were stained and analyzed by flow cytometry in order to determine the surface expression of key immune markers (Fig. 3E). The expression of the human natural killer (NK) cell marker CD161 is a sign of iNKT cell maturation and is present on a significant proportion of the engineered iNKT cells.

The higher ratio of CD45RO/CD45RA expression as compared to conventional T cells is also indicative of the increased memory phenotype of the iNKT cell population. The presence of CD4 single-positive, CD8 single-positive, as well as double-negative subpopulations within the iNKT cell compartment is also a surface phenotype of educated iNKT cells in the periphery (Fig. 3E and 3F). The ability of the transduced iNKT cell TCR to induce allelic inclusion and prevent the expression of endogenous TCR α or β chains was assayed by staining with an antibody for the engineered iNKT cell TCR Vβ11 chain as well as with a kit (Materials and Methods) that stains for a wide variety of common TCR Vβ chains (use of the kit excluded the use of “Tube G” which included an antibody for TCR Vβ11) (Fig 3E and 3G). The data show that the engineered iNKT cells lack the surface expression of the probed TCR Vβ chains and are positive for the engineered Vβ11 chain.

51 In vitro functionality of engineered iNKT cells

To determine the functional capabilities of engineered iNKT cells the liver and bone marrow were collected from BLT-iNKT mice and processed to isolate lymphocytes. These cells were cultured in vitro with or without the addition of α-GalCer (100ng/mL) for 3 days. The cells

52 were collected and stained for analysis by flow cytometry (Fig. 4A and 4B). The data show the ability of the engineered iNKT cells to robustly expand in vitro in response to α-GalCer. Some cells were treated with GolgiStop for the last 4-6 hours of culture before being collected.

These cells were collected and permeabilized (Materials and Methods) before being stained for intracellular cytokine production and analyzed by flow cytometry (Fig. 4C and 4D). The data show the ability of engineered iNKT cells to produce both IFN-γ and IL-4 in response to α-

GalCer stimulation.

CONCLUSIONS AND FUTURE STUDIES

The data disclosed in this chapter show some of the progress made on this project thus far. The successful generation of the BLT-iNKT mouse model will be critical for future experiments as well as further development of the project for a future clinical trial application.

The characterization of the development and surface phenotype of the engineered iNKT cells is also important in order to validate the TCR engineering of HSCs as an approach to generating iNKT cells. Finally the functionality of the engineered iNKT cells is crucial for their use as a cancer immunotherapy treatment.

Future experiments will aim to demonstrate the anti-cancer potential of the engineered iNKT cells utilizing both in vitro and in vivo tumor challenge models. The success of these experiments will create the tools necessary in order to collect the pre-clinical data necessary in order to apply for a clinical trial.

53 MATERIALS AND METHODS

Mice and Materials

NOD.Cg-PrkdcSCIDIl2rgtm1Wjl/SzJ (NOD/SCID/IL-2Rγ-/-, NSG) mice were purchased from the

Jackson Laboratory and maintained in the animal facilities at the University of California, Los

Angeles (UCLA). Six- to ten-week-old females were used for all experiments unless otherwise indicated. All animal experiments were approved by the Institutional Animal Care and Use

Committee of UCLA. X-VIVO-15 cell culture medium was purchased from Lonza.

Recombinant human Flt3 ligand, stem cell factor (SCF), thrombopoietin (TPO), and IL-3 were purchased from Peprotech. Phytohemagglutinin (PHA) was purchased from Sigma-Aldrich.

Fixable Viability Dye eFluor506 was purchased from affymetrix eBioscience. Retronectin® was purchased from Clontech.

Antibodies and Flow Cytometry

Fluorochrome-conjugated antibodies specific for human CD45, TCRαβ, CD161, CD4, CD8,

CD45RA, CD45RO, IFN- γ, and IL-4 were purchased from BioLegend. Fluorochrome- conjugated antibodies specific for human CD34 and TCR Vα24-Jα18 were purchased from BD

Biosciences. A fluorochrome-conjugated antibody specific for human Vβ11 and the IOTest®

Beta Mark TCR V beta Repertoire Kit (“Tube G” was not used in staining as it included an antibody for Vβ11) was purchased from Beckman-Coulter. Fc Receptor Blocking Solution

(TruStain FcX™) was purchased from BioLegend, while mouse Fc Block (anti-mouse CD16/32) was purchased from BD Biosciences. Cells were stained as previously described [49] and analyzed using a MACSQuant Analyzer 10 flow cytometer (Miltenyi Biotec). FlowJo software was used to analyze the data.

54 Single-Cell iNKT TCR Cloning

The single-cell iNKT TCR RT-PCR was performed based on an established protocol [56], with certain modifications. Human iNKT cells were sorted from healthy donor PBMCs based on a stringent forum of surface markers (hTCRαβ+hTCR Vα24-Jα18+ hTCR Vβ11+CD161+) using a

FACSAria II flow cytometer (BD Biosciences). Single cells were sorted directly into PCR plates containing cell lysis buffer. The plates were then immediately flash frozen and stored at −80 °C until use. Upon thawing, the cell lysate from each cell was split in half on the same PCR plate and processed directly into iNKT TCR cloning for both α and β chain genes using a OneStep

RTPCR kit (QIAGEN), following the manufacturer’s instructions and using the iNKT TCR gene-specific primers. These primers were designed to amplify the ∼200 bps spanning the CDR3 regions of the iNKT TCR α and β chain cDNAs and were customer- synthesized by Integrated

DNA Technologies (IDT): for TCRα (FW primer: 5′-GGG AGA TAC TCA GCA ACT CTG

GAT AAA GAT GC -3′; BW primer: 5′- CCA GAT TCC ATG GTT TTC GGC ACA TTG -3′) and for TCRβ (FW: 5′- GGA GAT ATC CCT GAT GGA TAC AAG GCC TCC -3′; BW: 5′-

GGG TAG CCT TTT GTT TGT TTG CAA TCT CTG -3′). Verified sequences (productive germline Vα24-Jα18-Cα assembly for TCRα and Vβ11-D/J/N-Cβ assembly for TCRβ) were used to construct the complete cDNA sequences encoding the TCR α and β chains from a single cell, based on information about human TCR genomic segments [the international ImMuno-

GeneTics information system (IMGT), www.imgt.org]. The selected iNKT TCR α and β pair cDNAs were then synthesized as a single bicistonic gene, with codon optimization and an F2A sequence linking the TCRα and TCRβ cDNAs to enable their co-expression (GenScript).

55 Mock and hiNKT Lentiviruses

Mock pMND-EGFP lentiviral vector was constructed by inserting an EGFP reporter gene into the lentivector that contains the MND retroviral LTR U2 region as an internal promoter [57]. The hiNKT lentiviral vectors were constructed by inserting either the synthetic bicistronic gene

(iNKT TCRα-F2A-TCRβ) or synthetic tricistronic gene (iNKT TCRα-F2A-TCRβ-P2A-EGFP) into the mock lentiviral vector replacing the EGFP reporter gene. VsVg-pseudotyped lentiviruses were produced using HEK293.T cells, following a standard calcium precipitation protocol and an ultracentrifugation concentration protocol as previously described [58].

Human CD34+ Cells and Thymus Tissues

Human fetal liver CD34+ hematopoietic stem and progenitor cells, as well as fetal thymus tissues, were obtained from the CFAR Gene and Cellular Therapy Core Laboratory at UCLA, without identification information under federal and state regulations. Granulocyte colony- stimulating factor (G-CSF) mobilized leukopaks from healthy donors were purchased from

Cincinnati Children's Hospital Medical Center (CCHMC) and HemaCare. CD34+ cells were sorted from fetal liver cells and G-CSF mobilized blood through magnetic-activated cell sorting using a Direct CD34 Progenitor Cell Isolation Kit (Miltenyi Biotec) following the manufacturer’s instructions [59]. The purity of CD34+ cells was more than 97% as evaluated by flow cytometry.

Generation of Humanized BLT Mice

Humanized BLT (bone marrow-liver-thymus) mice were generated as previously described, with certain modifications. In brief, human CD34+ cells were cultured for no more than 48 hours in X-

56 VIVO-15 cell culture medium containing recombinant human Flt3 ligand (50 ng/mL), SCF (50 ng/mL), TPO (50 ng/mL), and IL-3 (10 ng/mL) in non-tissue culture-treated plates coated with

Retronectin®. Viral transduction, when applicable, was performed at 24 hours by adding concentrated lentiviruses directly to the culture medium. At around 48 hours, CD34+ cells were collected and i.v. injected into NSG recipient mice (~0.5-1 x 106 CD34+ cells per recipient) that had received 270 rads of total body irradiation. 1-2 fragments of human fetal thymus (~1mm3), as well as donor-matched fetal liver CD34- cells when available, were implanted under the kidney capsule of each recipient NSG mouse. The mice were maintained on water containing trimethoprim/sulfmethoxazole (TMS) in a sterile environment for 8-12 weeks until analysis or use for further experiments.

In Vitro iNKT Cell Functional Assays

Spleen, liver, and bone marrow cells containing iNKT cells collected from humanized BLT- iNKT mice were cultured in vitro in a 24-well plate at 2 x 106 cells per well in regular human lymphocyte culture medium, with or without the addition of PHA or α-GalCer (100 ng/mL), for

3 days. On day 3 cells were collected and assayed for iNKT cell expansion using flow cytometry, some cells were also treated with 4 µL/6 mL of BD GolgiStop for 4–6 hours and then assayed for intracellular cytokine production using flow cytometry via intracellular staining using the BD

Cytofix/Cytoperm Fixation/ Permeabilization Kit (BD Biosciences).

Statistical Analysis

Data are presented as mean ± SEM, unless otherwise indicated.

57 References

1. Restifo, N.P., M.E. Dudley, and S.A. Rosenberg, Adoptive immunotherapy for cancer: harnessing the T cell response. Nat Rev Immunol, 2012. 12(4): p. 269-81. 2. Sznol, M. and L. Chen, Antagonist antibodies to PD-1 and B7-H1 (PD-L1) in the treatment of advanced human cancer. Clin Cancer Res, 2013. 19(5): p. 1021-34. 3. Bendelac, A., P.B. Savage, and L. Teyton, The biology of NKT cells. Annu Rev Immunol, 2007. 25: p. 297-336. 4. Godfrey, D.I. and S.P. Berzins, Control points in NKT-cell development. Nat Rev Immunol, 2007. 7(7): p. 505-18. 5. Fujii, S., et al., NKT cells as an ideal anti-tumor immunotherapeutic. Front Immunol, 2013. 4: p. 409. 6. Pilones, K.A., J. Aryankalayil, and S. Demaria, Invariant NKT cells as novel targets for immunotherapy in solid tumors. Clin Dev Immunol, 2012. 2012: p. 720803. 7. Vivier, E., et al., Targeting natural killer cells and natural killer T cells in cancer. Nat Rev Immunol, 2012. 12(4): p. 239-52. 8. Berzins, S.P., M.J. Smyth, and A.G. Baxter, Presumed guilty: natural killer T cell defects and human disease. Nat Rev Immunol, 2011. 11(2): p. 131-42. 9. Morrison, S.J., N. Uchida, and I.L. Weissman, The biology of hematopoietic stem cells. Annu Rev Cell Dev Biol, 1995. 11: p. 35-71. 10. Yang, L. and D. Baltimore, Long-term in vivo provision of antigen-specific T cell immunity by programming hematopoietic stem cells. Proc Natl Acad Sci U S A, 2005. 102(12): p. 4518-23. 11. Vatakis, D.N., et al., Antitumor activity from antigen-specific CD8 T cells generated in vivo from genetically engineered human hematopoietic stem cells. Proc Natl Acad Sci U S A, 2011. 108(51): p. E1408-16. 12. de Lalla, C., et al., Invariant NKT cell reconstitution in pediatric leukemia patients given HLA-haploidentical stem cell transplantation defines distinct CD4+ and CD4- subset dynamics and correlates with remission state. J Immunol, 2011. 186(7): p. 4490-9. 13. Rubio, M.T., et al., Early posttransplantation donor-derived invariant natural killer T- cell recovery predicts the occurrence of acute graft-versus-host disease and overall survival. Blood, 2012. 120(10): p. 2144-54. 14. Yamasaki, K., et al., Induction of NKT cell-specific immune responses in cancer tissues after NKT cell-targeted adoptive immunotherapy. Clin Immunol, 2011. 138(3): p. 255- 65. 15. Koya, R.C., et al., Kinetic phases of distribution and tumor targeting by T cell receptor engineered lymphocytes inducing robust antitumor responses. Proc Natl Acad Sci U S A, 2010. 107(32): p. 14286-91. 16. Smith, D.J., et al., Propagating Humanized BLT Mice for the Study of Human Immunology and Immunotherapy. Stem Cells Dev, 2016. 25(24): p. 1863-1873. 17. Smith, K., et al., Rapid generation of fully human monoclonal antibodies specific to a vaccinating antigen. Nat Protoc, 2009. 4(3): p. 372-84. 18. Giannoni, F., et al., Allelic exclusion and peripheral reconstitution by TCR transgenic T cells arising from transduced human hematopoietic stem/progenitor cells. Mol Ther, 2013. 21(5): p. 1044-54.

58 19. Yang, L., et al., Engineered lentivector targeting of dendritic cells for in vivo immunization. Nat Biotechnol, 2008. 26(3): p. 326-34. 20. Shimizu, S., et al., A highly efficient short hairpin RNA potently down-regulates CCR5 expression in systemic lymphoid organs in the hu-BLT mouse model. Blood, 2010. 115(8): p. 1534-44.

59

CHAPTER 5:

Conclusions and Future Studies

60 Cancer immunotherapy has established itself as viable treatment option and while it is now still only given to patients who have failed previous therapies this could soon change as immunotherapies continue to show greater efficacy and improved safety [1-9]. There are numerous biotechnology companies, with more starting up each year, that are researching new technologies and have multiple cancer immunotherapy treatments in varying phases of clinical development. This large industry-backed support of cancer immunotherapy research and development highlights its potential to radically change, if not replace entirely, the current cancer treatment methodology.

The studies covered in this dissertation detail the timeline of a cancer immunotherapy treatment from its inception and initial testing through to the collection of preclinical data in preparation for a clinical trial application [10, 11, Smith et al., manuscript in preparation]. The proposed treatment combines the therapeutic potential of iNKT cells to be used as a universal cellular therapy for cancer with the long-lived and durable responses that can be achieved when utilizing stem cells to generate new therapeutic immune cells.

This universal potential for iNKT cells is granted through two avenues: their mechanism of activation and mode of efficacy. The activation of iNKT cells occurs through the binding of the invariant TCR with a sphingolipid loaded CD1d molecule [12]. CD1d is a non- polymorphic MHC class I-like molecule and is not restricted to HLA matching meaning the same iNKT cell TCR can bind and receive activation from APCs regardless of their HLA type

[13-15]. This allows a single TCR construct to be used in all patients and does not limit the therapy to individuals of a certain HLA type. The anti-cancer efficacy of iNKT cells is not mediated by the cells themselves, but through their ability to potently mobilize the other effector cells of the immune system against cancer [12, 16-18]. The activation of iNKT cells can be

61 mediated through a synthetic ligand loaded onto a dendritic cell vaccine; this allows iNKT cell immunotherapy to be utilized against varying cancer types regardless of cancer antigen. These combined features set the stage for a standardized iNKT cell immunotherapy protocol and reagents to be used in a wide array of cancer treatments.

The advantages of utilizing engineered stem cells for cancer immunotherapy can be applied to iNKT cells as described in this dissertation as well as CAR and TCR immunotherapy protocols. This can be achieved through the methodology currently being initiated by the Ribas lab at UCLA in which stem cells are collected from the patient, very briefly cultured ex vivo during which time they are transduced with a viral vector carrying the therapeutic genes, and then infused back into the patient. The cells home to the bone marrow, engraft and repopulate, and are able to continue generating new therapeutic cells in vivo for the lifetime of the patient

[19-21]. Engineered stem cells can also be utilized entirely ex vivo where instead of being infused back into the patient after transduction they are grown in specific culture conditions that foster the generation and development of therapeutic cells [22]. These ex vivo-produced cells are then infused into the patient to exhibit anti-cancer efficacy.

All of the aforementioned cancer immunotherapy techniques can be combined to some degree to yield increased efficacy as well. Combinatorial therapies could yield increased targeting and killing capacity, but could also improve the safety of therapies by increasing specificity of targeted killing [23-28]. There is great potential to combine these engineered cellular therapies with blocking antibody treatments as well [29, 30]. The increased capacity of the engineered cells to target and kill cancer combined with the blocking antibody’s ability to prevent the down-regulation of these engineered cells in the tumor microenvironment allows the therapies to work to the greatest effect.

62 References

1. Emens, L.A., et al., Cancer immunotherapy trials: leading a paradigm shift in drug development. J Immunother Cancer, 2016. 4: p. 42. 2. Linardou, H. and H. Gogas, Toxicity management of immunotherapy for patients with metastatic melanoma. Ann Transl Med, 2016. 4(14): p. 272. 3. Hamid, O., et al., Safety and tumor responses with lambrolizumab (anti-PD-1) in melanoma. N Engl J Med, 2013. 369(2): p. 134-44. 4. Alatrash, G., et al., Cancer immunotherapies, their safety and toxicity. Expert Opin Drug Saf, 2013. 12(5): p. 631-45. 5. Zhou, L., et al., The efficacy and safety of immunotherapy in patients with advanced NSCLC: a systematic review and meta-analysis. Sci Rep, 2016. 6: p. 32020. 6. Grupp, S.A., et al., Chimeric antigen receptor-modified T cells for acute lymphoid leukemia. N Engl J Med, 2013. 368(16): p. 1509-18. 7. Motohashi, S., et al., A phase I study of in vitro expanded natural killer T cells in patients with advanced and recurrent non-small cell lung cancer. Clin Cancer Res, 2006. 12(20 Pt 1): p. 6079-86. 8. Motohashi, S., et al., Anti-tumor immune responses induced by iNKT cell-based immunotherapy for lung cancer and head and neck cancer. Clin Immunol, 2011. 140(2): p. 167-76. 9. Yamasaki, K., et al., Induction of NKT cell-specific immune responses in cancer tissues after NKT cell-targeted adoptive immunotherapy. Clin Immunol, 2011. 138(3): p. 255- 65. 10. Smith, D.J., et al., Genetic engineering of hematopoietic stem cells to generate invariant natural killer T cells. Proc Natl Acad Sci U S A, 2015. 112(5): p. 1523-8. 11. Smith, D.J., et al., Propagating Humanized BLT Mice for the Study of Human Immunology and Immunotherapy. Stem Cells Dev, 2016. 25(24): p. 1863-1873. 12. Bendelac, A., P.B. Savage, and L. Teyton, The biology of NKT cells. Annu Rev Immunol, 2007. 25: p. 297-336. 13. Castano, A.R., et al., Peptide binding and presentation by mouse CD1. Science, 1995. 269(5221): p. 223-6. 14. Teitell, M., et al., Nonclassical behavior of the mouse CD1 class I-like molecule. J Immunol, 1997. 158(5): p. 2143-9. 15. Brigl, M. and M.B. Brenner, CD1: antigen presentation and T cell function. Annu Rev Immunol, 2004. 22: p. 817-90. 16. Vivier, E., et al., Targeting natural killer cells and natural killer T cells in cancer. Nat Rev Immunol, 2012. 12(4): p. 239-52. 17. Fujii, S., et al., NKT cells as an ideal anti-tumor immunotherapeutic. Front Immunol, 2013. 4: p. 409. 18. Crowe, N.Y., et al., Differential antitumor immunity mediated by NKT cell subsets in vivo. J Exp Med, 2005. 202(9): p. 1279-88. 19. Vatakis, D.N., et al., Antitumor activity from antigen-specific CD8 T cells generated in vivo from genetically engineered human hematopoietic stem cells. Proc Natl Acad Sci U S A, 2011. 108(51): p. E1408-16.

63 20. Yang, L. and D. Baltimore, Long-term in vivo provision of antigen-specific T cell immunity by programming hematopoietic stem cells. Proc Natl Acad Sci U S A, 2005. 102(12): p. 4518-23. 21. Shaw, K.L., et al., Clinical efficacy of gene-modified stem cells in adenosine deaminase- deficient immunodeficiency. J Clin Invest, 2017. 127(5): p. 1689-1699. 22. Seet, C.S., et al., Generation of mature T cells from human hematopoietic stem and progenitor cells in artificial thymic organoids. Nat Methods, 2017. 14(5): p. 521-530. 23. Mondino, A., G. Vella, and L. Icardi, Targeting the tumor and its associated stroma: One and one can make three in adoptive T cell therapy of solid tumors. Cytokine Growth Factor Rev, 2017. 24. Zhang, E. and H. Xu, A new insight in chimeric antigen receptor-engineered T cells for cancer immunotherapy. J Hematol Oncol, 2017. 10(1): p. 1. 25. Chang, Z.L. and Y.Y. Chen, CARs: Synthetic Immunoreceptors for Cancer Therapy and Beyond. Trends Mol Med, 2017. 23(5): p. 430-450. 26. Zah, E., et al., T Cells Expressing CD19/CD20 Bispecific Chimeric Antigen Receptors Prevent Antigen Escape by Malignant B Cells. Cancer Immunol Res, 2016. 4(6): p. 498- 508. 27. Oberschmidt, O., S. Kloess, and U. Koehl, Redirected Primary Human Chimeric Antigen Receptor Natural Killer Cells As an "Off-the-Shelf Immunotherapy" for Improvement in Cancer Treatment. Front Immunol, 2017. 8: p. 654. 28. Bollino, D. and T.J. Webb, Chimeric antigen receptor-engineered natural killer and natural killer T cells for cancer immunotherapy. Transl Res, 2017. 29. Chen, N., et al., CAR T-cell intrinsic PD-1 checkpoint blockade: A two-in-one approach for solid tumor immunotherapy. Oncoimmunology, 2017. 6(2): p. e1273302. 30. Durgan, K., et al., Targeting NKT cells and PD-L1 pathway results in augmented anti- tumor responses in a melanoma model. Cancer Immunol Immunother, 2011. 60(4): p. 547-58.

64