ANTI-CD8 MONOCLONAL ANTIBODY-MEDIATED DEPLETION ALTERS THE PHENOTYPE AND FUNCTIONALITY OF SURVIVING CD8+ T CELLS
by
ERIC WILLIAM CROSS B.S., University of Minnesota, 2008
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment
of the requirements for the degree of Doctor of Philosophy
Immunology Program 2018
ii This thesis for the Doctor of Philosophy degree by
Eric Cross
has been approved for the Immunology Program
by
Raul Torres, Chair Eric Clambey Jordan Jacobelli
Mario Santiago Cara Wilson Ross Kedl, Advisor
iii Cross, Eric (Ph.D., Immunology)
Anti-CD8-monoclonal antibody-mediated depletion alters the phenotype and functionality of surviving CD8+ T cells
Thesis directed by Professor Ross Kedl
ABSTRACT
It is common practice for researchers to use antibodies to remove a specific cell type to infer its function. However, it is difficult to completely eliminate a cell type and there is often limited or no information as to how the cells which survive depletion are effected. This is particularly important for CD8+ T cells for two reasons. First, they are more
resistant to mAb-mediated depletion than other lymphocytes. Second, targeting either the CD or CD chain could induce differential effects. I show here that administration of two commonly used depleting-mAbs, targeting either the CD or CD subunits, in vivo leaves behind a population of CD8+ T cells with altered CD8 levels and retention of bound depleting-mAb, increasing the need for caution when identifying them. These surviving CD8+ T cells are able to participate in the immune response to immunization or infection. The resulting antigen-specific CD8+ T cells at the peak of the response adopt different phenotypes dependent on the depleting-mAb used and method of challenge. Further, I demonstrate that these depleting-mAbs alter the cytotoxic function of surviving CD8+ T cells in line with their phenotypic markers in vivo, yet use of these mAbs in vitro, exposing CD8+ T cells only at the time of the assay, has a different effect on their function. I
demonstrate that there is a similar dose response to both anti-CD8 mAbs on the expression of activation markers within the first few days of stimulation in vitro. However, only the anti-CD8 mAb is able to significantly increase CD25 expression, glycolysis, and
iv mitochondrial respiration – indicative of a mechanism behind the differences seen between these anti-CD8 mAbs. The impact of anti-CD8 antibodies on CD8+ T cell phenotype and function indicates the need to carefully consider the use of these, and possibly other "depleting" antibodies, as they could significantly complicate the interpretation of results or change the outcome of an experiment. These observations could impact how modulation of CD8+ T cell activation is pursued for immunotherapy. By demonstrating the feasibility to target CD8, the TCR coreceptor, to either raise or lower the threshold for activation to combat autoimmunity or increase tumor responses respectively.
The form and content of this abstract are approved. I recommend its publication. Approved: Ross Kedl
v TABLE OF CONTENTS
CHAPTER
I. )NTRODUCT)ON AND BACKGROUND………1
Introduction………...1
Cellular Immunity………...2
Humoral Immunity……….12
The Antibody as a Tool……….19
Scope of Project……….…22
II. MATERIALS AND METHODS……….26
III. CD8+ T CELLS THAT SURVIVE ANTIBODY-MEDIATED DEPLETION CAN PARTICIPATE IN IMMUNE RESPONSES………33
Introduction………....33
Results………34
Discussion……….42
IV. ANTI-CD8 MONLONAL ANTIBODIES CAN ALTER THE ACTIVATION, PHENOTYPE, AND FUNCTIONALITY OF CD8+ T CELLS………..47
Introduction………47
Results………48
Discussion……….69
V. FUTURE DIRECTIONS AND DISCUSSION..………76
Future Directions……….76
Discussion……….86
vi
APPENDIX……….
THE POTENTIAL FOR CD8+ T CELL HELP FOR HUMORAL IMMUNITY………
Introduction………
Materials and Methods………..
Results……….113
Discussion……….134 Conclusion………. REFERENCES...………142
vii LIST OF TABLES
TABLES
1-1 Summary of relative serum levels and functionality of human antibody isotypes…. 1-2 Mouse IgG isotype effector functions and relative affinities for Fc receptors.……… 1-3 Properties of commonly used anti-murine CD8 monoclonal antibodies………..
viii LIST OF FIGURES
FIGURES
1-1 T cell activation, differentiation, and memory………..3
1-2 B cell activation changes the B cell receptor from a membrane receptor to a secreted antibody….………... 1-3 Milestones of early monoclonal antibody development……….. 1-4 Model of CD8:anti-CD8 binding and effect on MHC-I interaction………..… 3-1 CD8+ T cells are resistant to mAb-mediated depletion and more difficult to identify post-depletion………... 35
3-2 CD8+ dendritic cells are not affected by depleting anti-CD8 mAbs………..39
3-3 CD8+ T cells that survive depletion participate in immune responses………. 41
3-4 Serum clearance of CD8 depleting mAbs………..43 4-1 CD8+ T cells that survive depleting mAb treatment acquire distinct
differentiation phenotypes……… 4-2 Antibody isotype does not alter the CD8+ T cell response……… 4-3 Precursor frequency is not responsible for phenotypic differences between
anti-CD and – treated CD + T cells……….. 4-4 CD8+ T cells surviving depletion differentially localize dependent on mAb
used for depletion and type of challenge……….. 4-5 Anti-CD surviving CD + T cells are more protective against viral
infection than anti-CD surviving CD + T cells……….. 4-6 Anti-CD8 mAbs have direct effects on target recognition and indirect
ix
effects on cytotoxic function………. 4-7 CD69 regulation changes with anti-CD8 mAb treatment in a
dose-dependent manner………. 4-8 CD25 regulation changes with anti-CD8 mAb treatment in a
dose-dependent manner………. 4-9 Anti-CD mAb treatment promotes proliferation while suppressing
activation markers……….. 4-10 CD8+ T cell metabolism is boosted by anti-CD during activation……… 5-1 Model of hypothesized data from potential experiments using anti-CD ………. 5-2 Model of hypothesized data from potential experiments using anti-CD ………. 5-3 IL-2 signaling can influence multiple facets of CD8+ T cell metabolism………
1 CHAPTER I
INTRODUCTION AND BACKGROUND Introduction
Understanding of the immune system has been a source of significant advancement in both health and for the means to ask questions about living processes. Around 1000A.D. variolation was likely practiced for the first time in China and India with subsequent utilization in Turkey and Africa before it spread to Europe1, 2. With this process, healthy
people were exposed to infected tissue to protect them from re-exposure and limit the chances that others get infected in the first place. Now, this practice is done safely with attenuated pathogens or subunits thereof and called vaccination. Edward Jenner brought about vaccination in 1796 by substituting a much safer virus, cowpox, as the means to protect from a related, but deadly virus, smallpox. Its impact on quality of life is on par with mass infrastructure projects such as sanitation and water treatment and perhaps only dwarfed by the development of industrial/chemical nitrogen fixation that has allowed the Earth to support billions more people with conventional crops3. Vaccination is an amazing
feat that is possible because organisms as far back evolutionarily as jawed fishes have the distinct advantage of an adaptive branch of the immune system. Adaptive refers to the organism s ability to respond more quickly and more robustly to a pathogen that it had previously encountered. This property provides a large selective advantage and its importance to higher/complex organisms such as ourselves is evident from how quickly ---
Parts of this chapter were reprinted with the permission of the American Physiological Society. Parts of this chapter were originally published in Pennock, N. et al. ” T cell responses: naive to memory and everything in between”. Adv Physiol Educ 4 (2013)4.
2 our immune system is still evolving. The adaptive branch consists of lymphocytes primarily represented by the conventional CD4+ and CD8+ T cells as well as the B cell, representing cellular and humoral immunity respectively. These two types of immunity complement each other in protecting the host and are discussed below.
The immune system has contributed in another, less obvious, way to our well-being, through our understanding and utilization of antibodies to our own benefit. Antibodies have allowed a means to address all sorts of biological questions. Their specificity and affinity for whatever they are raised against allow the identification of how other molecules fit into pathways by many means including immunoprecipitation, enzyme-linked
immunosorbent assays, neutralization, direct agonism or antagonism, etc. These methods have been used in virtually every field of biological study. Further, antibodies have recently stepped into the limelight as clinical therapeutics. The work herein describes a select pair of antibodies against a specific molecule, CD8, and their use as a tool for removing CD8 expressing T cells. After, I will focus on effects these antibodies have secondarily on the T cells that survive and discuss how those results indicate that these antibodies may be used as potential therapeutics. Prior to this, we need to first establish how the immune system creates antibodies and the purpose and function of the CD8+ T cell.
Cellular Immunity T cell activation
Producing a T cell that is capable of mediating immune protection first requires activation of the naïve T cell. This involves coordinated interactions between a number of molecules on the T cell and an antigen-presenting cell (APC), a cell that bears an antigenic peptide derived from the infectious agent non-covalently bound to a major
3 Figure 1-1. T cell activation, differentiation, and memory. (A) Depiction of the
activation CD and CD T cells by professional APCs. The primary stimulus, signal , comes from presentation of cognate peptide in the context of MHC to the TCR of T cells and the proximal signaling molecules are shown. CD80/86 interaction with CD28 primarily constitutes signal . B Summary of some of the important factors involved in
differentiation, the balance of which promote effector/exhaustion, memory, or unfit/anergic T cells.
4 histocompatibility complex (MHC) class I or class II molecule (Fig. 1-1A). The T cell
receptor (TCR) is composed of two chains ( and ), which recognize the peptide antigen only when it is bound in the context of an appropriate class I or class II MHC. On the T cell, the TCR associates with a complex of membrane proteins collectively known as CD3 (composed of -, -, -, and -subunits). Each TCR also associates with either a CD4 or CD8 coreceptor, depending on the type of T cell. These two molecules bind to MHC (class I for CD8 and class II for CD4), further stabilizing the interaction between the T cell and APC5.
The cytosolic region of this complex is responsible for propagating an intracellular signal subsequent to TCR ligation.
Subsequent to recognition of the cognate peptide and MHC by a specific TCR, the T cell and APC undergo actin-mediated membrane reorganization, facilitating the grouping of these TCRs on the cell surface and the formation of the immunological synapse6, 7. Besides
the TCR, other relevant molecules (costimulatory and/or adhesion) are also recruited to the site of the TCR-MHC interaction, forming a large multimolecular structure known the supramolecular activation complex (SMAC). This complex consists of a focal point of signaling molecules (cSMAC) surrounded by a ring of adhesion molecules (pSMAC)8. This
arrangement promotes both prolonged and stronger intracellular interactions and the appropriate spatial ordering of all the different TCR/coreceptor/costimulatory molecules9.
The grouping of TCR/peptide/MHC within the cSMAC results in the phosphorylation of CD3 components by Src family kinases LCK – recruited by the CD4 or CD8 coreceptor – and FYN (Figure 1-1A)10, 11, 12, 13, 14. This phosphorylation recruits and activates -chain-associated
protein kinase 70 (ZAP-70)15, which, in turn, phosphorylates linker for activation of T cells
5 responsible for the remainder of the downstream signaling after TCR ligation17. The
culmination of TCR and concomitant costimulatory signals (as discussed below) collectively induces a transcriptional program resulting in robust IL-2
production/secretion, an autocrine and paracrine factor that stimulates T cells to proliferate18, 19, 20, 21.
Simply stimulating a T cell with its cognate antigen alone does not lead to activation but instead results in a T cell refractory to further stimulus22, 23. The discovery of this
hyporesponsive state, known as anergy, led to the hypothesis that T cell activation requires additional input to become fully activated. Indeed, it was eventually discovered that
blockade of CD28 (or its ligands CD80 and CD86 on APCs) during T cell-APC interactions resulted in an anergic T cell phenotype24. More extensive studies have revealed a number
of aspects of the intracellular signaling that are responsible for rescuing T cells from their anergic state, generally traversing through phosphatidylinositol 3-kinase (PI3K) and phospholipase C- PLC ) and by the generation of a Ca2+ flux25, 26, 27, 28.
The identification of CD28 as the primary costimulatory pathway for T cell activation confirmed the two-signal model of T cell activation (Figure 1-1A). However, numerous lines of evidence have suggested that, although CD28 ligation is a necessary second signal, other membrane-bound and/or membrane-soluble inflammatory signals are necessary to achieve complete T cell activation as well. Collectively, the data seem to
indicate a role for inflammatory cytokine (secreted mediators) in directing the differentiation of the stimulated T cell into an effector that is appropriate for the immunological insult being addressed29, 30. Likewise, the data reflect a general role for
6 their appropriate ligand on APCs (CD70, OX-40L, and 41BBL respectively) to promote the survival of proliferating cells through their differentiation process and into memory cells31, 32. The identification of these costimulatory signals has also provided mechanistic insights
as to the connection between the innate and adaptive arms of immunity. Few (and
sometimes none) of the costimulatory ligands described above are found on the surface of resting, immature APCs, i.e., an APC unstimulated by pathogen-associated molecule
patterns (PAMPS), including but not limited to microbial components, or proinflammatory mediators produced by other innate immune or stromal cells33. Thus, in the steady state, T
cell interactions with a specific antigen on these resting APCs results in anergy and immune tolerance, a process that appears to be responsible for eliminating self-reactive T cells to antigens expressed only in the periphery and thereby preventing autoimmunity34.
However, when an APC becomes activated by sensing pathogens or inflammation through one or more cytokine and/or innate pattern recognition receptors, the various
costimulatory ligands are then expressed, allowing T cell activation, proliferation, and differentiation35. Thus, the production of innate inflammatory signals and mediators is a
necessary prelude to the effective transition to an adaptive response.
Finally, T cells also express an array of inhibitory receptors, helping to fine tune the eventual response of the T cell. These inhibitory receptors can act to both limit
costimulatory molecule ligation and signaling. A good example is cytotoxic T lymphocyte antigen (CTLA)-4, an inhibitory molecule expressed on activated T cells that both produces intracellular phosphatase activity that dampens downstream signaling of the TCR and CD28, and also acts as a competing receptor for CD80 and CD86. In fact, CTLA-4 actually has higher affinity for CD86 binding than does CD28, suggesting the importance of being
7 able to control how inflammatory an immune response is36. A number of other inhibitory
receptors have been identified, including programmed cell death-1 (PD-1), lymphocyte activation gene 3 (LAG-3), and V-domain Ig suppressor of T cell activation (VISTA), and blockade of their function using monoclonal antibodies is being successfully exploited clinically for the purposes of augmenting immunity against various cancers37.
Honing in on CD8+ T cells, their significance as an immune sentinel was first
demonstrated by their ability to kill cells that were engineered to produce, but not secrete or transport to the membrane, an influenza nucleoprotein, meaning the CD8+ T cell was able to recognize what was being made within the cells38, 39, 40. Thanks to the expression of
MHC-I on the surface of all nucleated cells, and the loading of peptides derived from throughout the cytosol on MHC-I, the CD8+ T cell can see inside cells from nearly all tissues. As an example, when a stromal cell is infected by a virus the cell translates viral proteins within the cytosol that can then be loaded on MHC-I to present to already activated T cells, but stromal cells cannot engulf antigens from the environment for
presentation nor express the costimulatory molecules necessary for naïve T cell activation. The APC considered most important to CD8+ T cell activation is the CD8+ dendritic cell (DC) because it is able to present exogenous antigens – proteins taken up from the
environment – to CD8+ T cells, as well as provide necessary costimulation. While the CD8 molecule on conventional T cells is a heterodimer of an and chain, the CD molecule expressed on DCs is a homodimer of two -chains – the significance of this alpha-only expression on DCs is yet to be fully established. The process by which these DCs present exogenous antigens to CD8+ T cells, termed cross-presentation, begins with the uptake of pathogens/proteins into endosomal compartments from the environment. Subsequently,
8 antigenic components are largely thought to be shuttled by unknown mechanisms to the cytosol and targeted to the immunoproteosome to be degraded into peptide fragments of 8 or 9 amino acid residues. These peptide fragments are then shuttled to the endoplasmic reticulum with the help of transporter associated with antigen processing (TAP) and are ultimately loaded onto the peptide binding cleft of MHC-I that is then transported to the cell surface for interaction with CD8+ T cells41, 42, 43. The CD8+ DC can thus present antigens
from its environment as well as from within itself to activate CD8+ T cells in the appropriate context of costimulation and inflammatory cytokines.
CD8+ T cell differentiation and memory
CD8+ T cells function by scanning, and if necessary, killing infected cells in most tissues. In the presence of )FN and )L-12 cytokines, CD8+ T cells differentiate into effectors and are often referred to as cytotoxic T lymphocytes (CTLs) that produce )FN and TNF themselves. )f after activation they recognize their cognate antigen presented on MHC-I, CTLs can kill through a couple mechanisms. First, they can release granules filled with cytotoxic proteins called granzymes and perforin to lyse a target cell. Alternately, the binding of Fas (expressed by the activated CD8+ T cell) to Fas ligand (expressed on
stressed/damaged cells) can induce the apoptosis of Fas ligand+ cells. CTLs are often considered short-lived effector cells (SLECs) to differentiate them from memory-precursor effector cells (MPECs), CD8+ T cells during the primary response that are less cytolytic and upregulate markers, such as IL- R , that bona fide memory T cells express.
A T cell response typically peaks ~7–15 days after initial antigen stimulation in terms of numbers and effector differentiation. For a productive response, this peak corresponds roughly to the eradication of the pathogen. A few days past the peak of the
9 response, 90–95% of antigen-specific T cells then undergo apoptosis, leaving behind a pool of T cell clones with a memory for the pathogen they had taken part in eliminating. Once formed, subsets of memory cells can survive for decades [the half-life of memory T cells is ~8–15 yr44], providing protection for the better part of a lifetime. Further, over the course
of the response, these T cells are instilled with a range of phenotypes and functionalities. Compared to their naïve counterparts, these memory T cells have less stringent
requirements for subsequent activation via antigenic and costimulatory receptors, an increased proliferative potential, and a more rapid effector response. In addition, due to changes in chemokine sensing and adhesive properties, memory cells can traffic through both secondary lymphoid organs (SLOs; i.e. spleen and lymph nodes) and peripheral tissues, giving them access to tissues poorly accessed by naïve (peripheral tissues) or effector (SLO) T cells. Collectively, these functions produce an in situ response to reinfection in a fraction of the time taken by the primary response.
For both CD4 and CD8 T cells, there are two main subclasses of memory cells: central-memory (TCM) and effector-memory (TEM) T cells (Figure 1-1B). TCM cells are
commonly defined phenotypically as expressing high levels of the IL-7 receptor (CD127), high levels of adhesion markers such as CD44 and CD62L, low levels of the surface marker killer cell lectin-like receptor subfamily G member 1 (KLRG-1), and high levels of the
chemokine/homing receptor C-C chemokine receptor type 7 (CCR7). Furthermore, TCM cells
are functionally characterized by their increased potential for proliferation after antigen reencounter. TEM cells phenotypically contrast with TCM cells in that they generally express
low levels of CD62L, low levels of IL-7 receptor CD , high levels of KLRG-1, and are deficient in CCR7. As their name implies, TEM cells display rapid effector function
10 (granzyme B and IFN production) but a limited proliferative potential. The high
expression of CD62L and CCR7 by TCM cells allow for preferential homing to SLOs (which
constitutively produce the CCR7 ligands CCL19 and CCL21), where they are well situated to protect from a systemic infection and seed the peripheral tissues with new effector cells after stimulation.
Metabolism is intricately intertwined with T cell activation and fate
A growing body of evidence suggests that not only is the metabolism of a T cell linked to effector and memory fates, but metabolism is actually an active player in determining those fates and their functionality. Naïve T cells and most normal,
non-cancerous differentiated cells take advantage of the efficient energy production from their mitochondria. During the rapid proliferation of T cells following activation, glycolysis with lactic acid fermentation is thought to be utilized to meet the high energy demands and preserve biomolecules. Much of this initial boost to glycolysis comes from CD28 signaling – as well as other costimulatory receptors of the tumor necrosis factor receptor superfamily (TNFRSF) – and is partly why signal is necessary for T cell activation28, 45, 46, 47, 48, 49.
Additionally, the TCR signal itself is tied directly into initiating the transcription factors (IRF4, HIF1 , AP4, Myc, and the sterol regulatory binding proteins (SREBPs)) necessary to begin gene expression for a variety of metabolic processes including glycolysis,
glutaminolysis, and lipid biosynthesis50, 51, 52, 53, 54, 55. This phenomenon of depending on
glycolysis is not completely understood, despite being first described in 1924 by Otto Warburg et al. He called this aerobic glycolysis – referring to relying on glucose even in the presence of oxygen – and it has since been termed the Warburg effect56, 57. Interestingly,
11 them. In general, aerobic glycolysis drives T cells toward an effector phenotype, then as demand for clonal expansion decreases, T cells switch to a mix of catabolic processes, the TCA cycle, and ultimately oxidative phosphorylation that support the continued survival of memory cells as it does for naïve cells58, 59. In support of this, pharmacological treatment
with rapamycin and metformin treatment skews the T cell fate decision toward memory by inhibiting mTOR and the respiratory chain complex 1 respectively60, 61, 62. However, under
normal conditions, much of this metabolic control is thought to be fine-tuned by cytokine signaling. Perhaps the most significant example is the dichotomy of IL-2 and IL-15 signaling that lead to the SLEC phenotype/reliance on glycolysis and the MPEC phenotypes/reliance on mitochondrial respiration respectively58, 62. The breadth of metabolism s influence has
expanded beyond just T cells to the immune system in general, and its influence over how the immune system interacts with pathogens and tumors is increasingly being
prioritized63.
Particularly for cancer, there has been much interest in how to control metabolism as tumors are generally thought to be dependent on aerobic glycolysis. Additionally, there are many drugs already in use to target metabolism that are safer compared to other current therapy options such as tyrosine-kinase inhibitors or traditional
DNA-intercalating/damaging agents. The problem with using a small molecule drug to target, say glycolysis, is that our own immune system, particularly T cells, also relies on aerobic glycolysis. This is the same conundrum that is encountered with the use of traditional chemotherapy that damages the DNA of dividing cells, both the tumor and the immune system responding against it are rapidly dividing. Thus, it may be more efficacious to target the host s immune response directly versus the tumor when manipulating these pathways.
12 Serendipitously, we now know that some of the breakthrough immunotherapy treatments that target T cells to increase anti-tumor responses, work in part by altering their
metabolism. One of the negative regulators of T cell activation, PD-1, partly drives the bioenergetic insufficiencies of T cells often found within the tumor microenvironment or chronic infection, termed exhaustion64. For years now, there has been an approved
anti-PD-1 mAb and we are now beginning to understand why it is a successful treatment. Part of the efficacy of anti-PD-1 derives from its ability to block metabolic regulation by PD-1. It was recently shown that PD-1 can affect metabolism indirectly by recruiting SHP-2,a Src homology 2 (SH2) domain containing non-transmembrane protein tyrosine phosphatase, which dephosphorylates CD28 and therefore downregulates CD28-mediated increases in glycolytic machinery, including the glucose transporter-1 (GLUT1) and mTOR47, 65.
Furthermore, the transcriptional coactivator PGC , which supports both glycolytic machinery and mitochondrial biogenesis, is downregulated by PD-1 signaling, thus impacting mitochondrial respiration as well64. Overall, CD8+ T cells responding in the
tumor microenvironment see a promotion of normal effector metabolism and regain cytotoxic function64, 66. CTLA-4, another negative regulator of T cell activation, has an
approved blocking mAb, that can also inhibit glucose metabolism, but it is as yet unclear by what mechanism67. It will be exciting to see the continued unveiling of how these complex
pathways intertwine and how they can be taken advantage of for therapeutic use. Humoral Immunity
B cell activation
While many laymen imagine the immune system in terms of cells that directly kill pathogens or infected cells, the truth is much of our immunity and most of our vaccines
13 work due to the antibodies generated by B cells. Just as the TCR acts as an antigen receptor for T cells, the B cell receptor (BCR) acts in a similar capacity. Structurally, the BCR includes 4 polypeptides consisting of 2 identical light chains and transmembrane heavy chains that are joined to form a Y-shaped molecule with two-binding sites. The binding sites are at the end of what is termed the Fab fragment consisting of a light chain and heavy chain that have a constant domain and a variable domain distal to the Fc, which is made up of two constant domains (Figure1-2). The sites responsible for binding on the Fab fragment are called the complimentarity-determining regions (CDRs), which are
polymorphic/hypervariable in amino acid sequence allowing for immense clonal diversity of B cells and thus an ability to recognize a large range of potential antigens. Like the TCR, the BCR also requires accessory transmembrane polypeptide chains )g & )g that have immunoreceptor tyrosine activating motifs (ITAM) that act similar to CD3 on T cells. )ndeed, )g & )g recruit the Src family kinases LYN and SYK that activate a multimolecular signalosome including BTK and PLC , responsible for the downstream signaling. After infection or immunization, whole pathogens or products derived from them bind BCRs directly as soluble units or are presented by APCs and cross-link them inducing an activating signal. In conjunction with CD40 stimulation via helper CD4+ T cells, the B cell becomes fully stimulated and undergoes rapid proliferation and differentiation.
Stimulation can then lead to the most distinguishing aspect of a BCR, its downregulation from the surface and ability to be secreted – termed an antibody.
Many of the daughter cells will increase in size and granularity, becoming plasma cells responsible for secreting copious amounts of antibody into the bloodstream. Some
14 Figure 1-2. B cell activation changes the B cell receptor from a membrane receptor to a secreted antibody. (Left) B cell receptor and a depiction of the proximal associated signaling molecules. (Right) Upon activation the B cell receptor RNA transcript undergoes differential splicing to remove the transmembrane, hydrophobic C-terminus and instead include a hydrophilic C-terminus allowing secretion of the antibody. H = heavy chain, L = light chain, small black hashes = disulfide bonds.
15 daughter B cells will instead enter a specialized niche called the germinal center (GC) that allows for Fc-isotype class-switching, affinity maturation via hypermutation, and
differentiation into plasma cells or memory B cells that will proliferate and differentiate into new effectors when the pathogen/antigen is encountered again. There are two key cell types that support the GC reaction: i. follicular-helper CD4+ T cells that express the surface receptors CXCR5, PD-1, and ICOS and secrete IL-21 to support GC B cell maturation,
survival, as well as promote affinity maturation and ii. follicular-dendritic cells that retain protein antigens (in the form of immune complexes with antibodies) on their surface for long periods to continually act as an antigen-depot and help select for B cells that do not recognize self68, 69, 70. The net effect of the GC reaction is the production of populations of B
cells with a higher affinity for the pathogen/antigen encountered that produce an antibody isotype profile that is most able to control that pathogen. Therefore after an immune response the host has generated T and B lymphocyte populations better geared toward control of the offending agent should it be encountered again.
Generation and function of antibody
Newly activated B cells can differentially splice the BCR RNA transcript to include a hydrophilic C-terminus that allows export of the BCR out of the cell71. Secreted antibodies
are then able to bind pathogens or pathogen-derived molecules and perform effector functions largely dictated by the Fc domain of the antibody. Of course, one aspect of
secreted antibody that is independent of the Fc domain is its potential ability to neutralize a microbe or toxin by coating it, making it unable to interact with the host and thus less pathogenic. There are 5 isotypes of antibody that correspond to their heavy chain
16 effector functions are summarized in Table 1-1. Naïve B cells express IgM as its BCR (along with IgD) which is unique in its ability to form a pentamer with itself, by incorporating a centralized J-chain polypeptide. Pentameric IgM thus has 10 antigen binding sites
providing high avidity. The multimer configuration is more suitable for recruiting C1q, a component of the complement pathway that can result in the direct lysis of a pathogen s membrane by initiating formation of the membrane attack complex (MAC). Most B cells that class-switch will adopt an IgG isotype that represents most of the antibody in our serum. However, IgG is further divided into subclasses and differs somewhat between species: in humans (1, 2, 3, 4), mice (1, 2a, 2b, 2c, 3), and rats (1, 2a, 2b, 2c). Isotype-specific IgG functions are dependent on their affinities for binding Fc Rs that are expressed on different cell types, see Table 1-2 for a description of relative affinities of mouse IgG isotype for Fc Rs and their functions. Further, the choice to class-switch to a certain isotype
depends on the inflammatory environment and in particular the type of helper CD4+ T cell the response generates. As an example, with parasitic infections T helper type-2 CD4+ T cells (TH2) are generated that have a cytokine profile that includes IL-4, skewing the
antibody response toward )gE. )gE can engage the Fc R and stimulate mast cells and basophils that are specialized to control parasitic infections. IgA is made as a dimer, held together by a J-chain (like pentameric IgM). It is unique in that it is secreted across barriers, such as into the gastrointestinal tract and sinuses where it can control microbial
populations that may pass these mucosal surfaces into the host. IgA is enveloped by a secretory polypeptide that is thought to protect it from degradation within such
environments. IgD is expressed on mature, naïve B cells and is thought to mainly function as a BCR, as its presence in serum is minimal.
17 Table 1-1. Summary of relative serum levels and functionality of human antibody isotypes. CDC = complement-dependent cytotoxicity, ADCC = antibody-dependent cell-mediated cytotoxicity, ADCP = antibody-dependent cell-cell-mediated phagocytosis.
18 Table 1-2. Mouse IgG isotype effector functions and relative affinities for Fcγ
receptors. CDC = complement-dependent cytotoxicity, ADCC = antibody-dependent cell-mediated cytotoxicity (NK cell and monocyte driven), ADCP = antibody-dependent cellular phagocytosis (monocyte/macrophage, neutrophil, dendritic cell driven).
*For IgG2a and IgG2c isotypes: inbred mouse strains with the Igh1-b allele have IgG2c isotype instead of IgG2a
**relative binding to murine Fc Rs of both murine and rat )gG isotypes
19 Most simply, the coating of a pathogen or a toxin by antibodies can neutralize it, making it unable to interact with the host and thus less pathogenic. Further, the Fc of an antibody bound to a pathogen can be recognized by Fc receptors – expressed on innate phagocytic cells that promote engulfment/phagocytosis as well as on B cells and natural killer (NK) cells. Some Fc domains are also able to interact with C1q, a component of the complement pathway that can result in the direct lysis of a pathogen s membrane by initiating formation of the membrane attack complex (MAC). See Table 1 for a summary of the effector functions of the Fc-isotypes.
The Antibody as a Tool
Few scientific discoveries have had as much of an impact on the biological sciences as the generation of antibodies against specific molecules of interest, particularly the advent of the means to generate monoclonal antibodies (mAb) using hybridomas. This technique was first described in 1975 by Georges Köhler and Cesar Milstein for which the Nobel Prize in Physiology or Medicine was awarded72. By fusing a B cell clone with a
myeloma cell line, the resulting hybridoma cell line is immortalized and also secretes the mAb of interest. The subsequent development of mAbs as a tool proceeded with haste and is summarized in Figure 1-3. Briefly, a crucial limitation of the use of mAbs was quickly realized due to hybridomas being of a different species than the intended host and an immune response is generated against it73, 74. Thus, there has been much effort to
circumvent this by creating chimeric mAbs in which the Fc portion or even the entire sequence beside the CDR residues responsible for antigen recognition is replaced at the DNA level with that of another species, such as human. More recently, the field has begun to engineer more minute aspects of the antibody to achieve longer serum half-life by altering
20 Figure 1-3. Milestones of early monoclonal antibody development. Beginning with the description of the hybridoma technique – the means to create mAbs – the development of mAbs as a tool has been brisk. Depicted are the notable milestones that led to the first fully human mAb FDA approved for clinical use. After 2002, there has been continued
advancement in more fine-tuned engineering to increase serum half-life, increase the affinity of previous clones, covalently attach drugs for cell/tissue-specific targeting, etc.
21 either the isoelectric point, increasing binding to the neonatal FcR, and/or PEGylation75.
Furthermore, there has been a plethora of alterations to change or boost specific functions including, increased C1q recruitment and complement activation, addition of sialylated glycans to enhance inflammation reduction, covalent attachment of radioactive elements or therapeutic drugs for targeted delivery, and incorporating different Fab fragments on the same backbone to bind dual targets at once75, 76, 77.
The specificity and affinity innate to mAbs created a means to: robustly delineate and classify types of cells and their lineage, reliably assay for molecules of interest in vitro and ex vivo, remove cell types from an organism, increase specificity of drug delivery, and influence signaling by cross-linking target molecules or acting as an agonist/antagonist directly. Unfortunately, a particular mAb does not accomplish these feats in isolation. As an example, a mAb utilized to deliver an attached therapeutic to a particular cell type may complicate the intended goal by physically impeding binding of the target to its normal binding partners or promote activation of the target by stabilizing the target s interaction with its ligand. Either scenario could impact the delivered drug s efficaciousness. (opefully, these scenarios would be studied and likely become evident if the mAb is to be used
clinically, however, as a tool for basic/pre-clinical research we are often ignorant of the unintended consequences of these reagents.
Indeed, when we think about mAbs being designed to agonize or antagonize a pathway or to induce a certain kind of signal, the particular molecule it binds and even the specific epitope is important. As an example, the anti-IL-2 mAbs JES6-1 and S4B6 have both been used to both block and promote IL-2 signaling. Only recently have we learned that JES6-1 binds IL-2 in such a way that its interaction with IL- R CD is blocked whereas
22 S4B6 blocks the interaction with IL- R CD , encouraging either regulatory or effector T cell development78.
Scope of Project
For this thesis I will first compare two different anti-CD8 mAbs that deplete CD8+ T cells, one targeting the -subunit and another the -subunit. These are two of the original anti-CD8 clones described shortly after the advent of hybridomas in 1979 by Ledbetter and Herzog79. 53-6.7 (anti-CD was chosen because it has several, seemingly unique
properties. This anti-CD mAb clone is purported to hold CD8 in a high-affinity
conformation with MHC-I and increase MHC-I-multimer binding/staining80, 81. Further, it
has been demonstrated to promote TCR signaling/calcium flux81, 82 as well as induce the
secretion of chemokine without TCR or pharmacologic stimuli83, 84, 85. Interestingly, it
accomplishes at least some of these feats without cross-linking82. There have even been
some studies that have used this mAb clone to block CD , assuming its association with TCR was negated, however, as previously mentioned this clone stabilizes TCR:peptide-MHC-I interaction indicating the conclusions generated were potentially wrong or made under false pretenses86, 87, 88, 89, 90. Together, these data suggest that the 53-6.7 mAb has the
potential to induce CD8-signaling directly as well as cooperate with the TCR for increased cell activation and that the true function of this mAb and the CD8+ T cells exposed to it needs to be more carefully documented. 53-5.8 (anti-CD instead behaves in an opposing manner. This anti-CD mAb s binding is known to substantially decrease MHC-I-multimer binding to CD8+ T cells81. The CD8 subunits are also known to bind different protein
partners, indicating the potential for different downstream activities through the binding (agonism/antagonism) of either subunit. Thus, these mAbs were chosen because they
23 represented a good chance to study potential differential effects on the surviving CD8+ T cell population after depletion. See Figure 1-4 for a model of the anti-CD8 mAbs used herein and see Table 1-3 for a summary of common anti-murine CD8 antibodies.
In chapter III I show that these widely available and substantially used depleting-mAbs leave behind a population amenable to study. I then show that these mAbs impede the binding of one another, despite targeting different subunits and differentially induce CD8 internalization making the remaining population more difficult to discern. Further, these CD8+ T cells surviving depletion retain bound mAb throughout a primary challenge, likely due to a long half-life of the antibodies in serum. In chapter IV I elucidate further the effect these mAbs have directly on those CD8+ T cells resistant to depletion. I hypothesize that CD8+ T cells that survive depletion with anti-CD would be more fully
differentiated/effector-like, whereas anti-CD surviving CD + T cells would be less differentiated/unfit, relative to each other and untreated, control CD8+ T cells. Below, I show that these survivors adopt a different phenotypic profile, localization, and functional attributes dependent on the mAb used. Further, what I hypothesized turns out to be more nuanced. The depleting mAb and the inflammatory environment affect the phenotype of the depletion-surviving, responding CD8+ T cells as well as the concentration of anti-CD8 mAb the cells are exposed to. Collectively, the data suggest that these mAbs, or similar acting clones, may be useful tools for immunomodulatory therapy by virtue of altering the perceived TCR signal or potentially transducing its own signal in isolation.
24 Table 1-3. Properties of commonly used anti-murine CD8 monoclonal antibodies.
25 Figure 1-4. Model of CD8:anti-CD8 binding and effect on MHC-I interaction. The two monoclonal antibodies used herein bind primarily to the CD subunit - . or CD subunit (53-5.8). 53-6.7 has been predicted to hold CD8 in a high affinity conformation for MHC-I binding, increasing MHC-I binding to CD8:TCR. 53-5.8 has been shown to
significantly reduce CD8 interaction with MHC-I by binding to the epitope responsible for MHC-I binding, reducing MHC-I binding to CD8:TCR.
26 CHAPTER II
MATERIALS AND METHODS Materials and Methods Mice and reagents
All experiments were performed with 5-12 week old C57BL/6 mice of either sex, purchased from Jackson Laboratories. OT1 are a TCR transgenic mouse specific for the SIINFEKL peptide derived from ovalbumin (amino acids 257-254) in the context of H-2Kb
and were bred and housed at the University of Colorado vivarium. All animal protocols were approved by the Institute of Animal Care and Use Committees of the University of Colorado. Antibodies used for depletion, rat anti-CD -6.72) was either purchased from Bio X Cell or made in house and rat anti-CD 53-5.8) was purchased from Bio X Cell. Isotype control antibodies, rat clones TNP6A7 (IgG2a) and 2A3 (IgG1) were also purchased from Bio X Cell. Fluorochrome-conjugated antibodies used for flow cytometry include MHC-II (M5/114.15.2) [1:600]; CD4 (GK1.5), CD -6. , CD -5.8), B220 (RA3-6B2), CD44 (IM7) [1:500]; CD -2C11), CD25 (PC61.5), CD45.1 (A20), CD45.2 (104), CD62L (MEL-14), CD69 (H1.2F3), CD127 (SB/199), CXCR3 (CXCR3-117) and KLRG1 (2F1/KLRG1) [1:300] all purchased from Biolegend and Goat anti-Rat IgG [1:800] from Jackson
Immunoresearch Laboratories. Fluorochrome-conjugated antibodies used for microscopy include CD169 (MOMA-1, BioRad), IgM (RMM-1, in house), and CD45.1 (A20, Biolegend). SIINFEKL and altered peptide ligands used herein were all synthesized by the University of Colorado Protein Production Shared Resource facility.
27 OT1 adoptive transfer and assessing depletion-surviving CD8+ T cell phenotype and function
OT1 T cells were isolated from whole splenocytes by CD8-negative magnetic
selection (to a purity of at least 90%; BioLegend) and 106 cells were adoptively transferred,
unless otherwise noted, into CD45-congenic recipient mice by tail vein injection. The following day 250-500µg of depleting antibody was delivered intraperitoneally. For subunit-vaccinations, 100µg whole ovalbumin (Sigma), 50µg poly(I:C) (Sigma), and 50µg anti-CD40 (FGK4.5, made in house or from Bio X Cell) suspended in PBS was given
intravenously and assessed 7 days later unless otherwise stated. For infectious challenge, 107 PFU of Vaccinia virus expressing ovalbumin was given intravenously and assessed 5
days later unless otherwise stated. Spleens and lymph nodes harvested were macerated with glass slides, red blood cells lysed with Ammonium-Chloride-Potassium buffer (155mM ammonium chloride, 10mM potassium bicarbonate and 0.1mM Ethylenediaminetetraacetic acid) and stained with fluorochrome-conjugated antibodies to determine phenotype of transferred OT1 T cells.
Serum rat IgG ELISA
To assess clearance of depleting rat IgG antibodies, mice were injected i.p. with 500µg of either anti-CD8 (IgG2a; clone 53-6.7), anti-CD8 (IgG1; clone 53-5.8), IgG2a isotype control (clone TNP6A7), or IgG1 isotype control (clone 2A3). At 3hours, 4days, 7days, and 14days after injection blood was collected from tail vein into Sarstedt Z-gel serum capture tubes. After blood coagulation at room temperature the tubes were centrifuged in a microfuge for 10,000rpm for 5min and refrigerated until use.
28 Corning Immulon 4 HBX ELISA plates were coated overnight at 4∘C with goat anti-rat IgG (Southern Biotech) at 10µg/mL in PBS. Plates were then washed once (PBS with 0.1% Tween-20; Thermo Fisher) and blocked with 1%bovine serum albumin in PBS for at least an hour at 37∘C. In separate round bottom 96-well plates, serum was added to the top row at a dilution of 1:10 in duplicate wells and diluted serially 1:3 down the plate. 50µL of all wells were then transferred to the ELISA plates after washing 3 times. Serum antibodies were allowed to bind overnight at 4∘C. Plates were washed 5 times and to detect captured rat IgG 50uL of Goat anti-rat IgG conjugated to alkaline phosphatase (AP) (Southern Biotech) was added at a dilution of 1:2000 in PBS and incubated 45min at 37∘C. After 5 washes, plates were developed by addition of 100ul of AP substrate
[para-Nitrophenylphosphate tablets (Thermo Fischer) dissolved into 1M diethanolamine buffer (5X stock solution from Thermo Fischer) to a concentration of 1mg/mL] and incubated at 37∘C. Absorbance values were read at 405 nm (VersaMax ELISA reader; MDS Analytical Technologies) when plates were sufficiently developed.
Confocal microscopy
For imaging, spleens and lymph nodes were harvested from mice and fixed on ice for 30min in 1% PFA with 3% sucrose in PBS. Tissue was subsequently incubated on ice with 20% sucrose in PBS for 30-60min. Tissue samples were then frozen in Tissue-Tek OCT compound (Sakura) using dry ice. A Leica Cryostat was used to cut 5-7µm sections for staining. Sections were imaged using a Zeiss LSM 700 confocal microscope at x10 magnification. Images were analyzed using Imaris or Zen Blue software. The white pulp was delineated by IgM and MOMA-1 staining and surfaces created with Imaris software or regions with Zen software. The red pulp was considered all area outside white pulp
29 surfaces/regions. To identify CD45.1+ OT1 T cells within sections, all channels and
surfaces/regions were hidden beside the channel of CD45.1 staining. For Imaris the spot function was utilized, assuming an approximate 10µm diameter of T cells, to enumerate CD45.1+ OT1 T cells. The surfaces were then distance transformed to create a separate channel whereby spots could be quantified as within the surface (white pulp) or not (red pulp). For Zen software all CD45.1+ OT1 T cells were manually enumerated, blinded to all other channels and regions, the regions were then revealed and the total number within all white pulp regions and those outside, in the red pulp, counted.
Vaccinia virus plaque assay
OT1 T cells were transferred and mice infected as described above. Five or 7 days post-challenge, ovaries from infected mice were harvested and homogenized in 5-10mL PBS. Serial dilutions were made and added in duplicate onto 24-well plates containing 1.25x105 Vero cells seeded the day before. Plates were incubated at 37∘C for 3 days to allow
plaque formation. After incubation, media was aspirated and cells were washed with PBS and fixed with 0.5mL 10% buffered formalin per well for 5min at room temperature. The formalin was then aspirated and plates were washed with PBS prior to the addition of 0.5 mL of 0.1% gentian violet (Ricca Chemical Company) per well and incubated 5min at room temperature. The crystal violet solution was aspirated and the plates were allowed to air dry. Viral titer in ovaries was determined by counting plaques and back calculating the number of infectious vaccinia particles per ovary pair.
Cytotoxicity assays
OT1 T cells were isolated, adoptively transferred, and the host mice treated with depleting mAbs as before. Spleens and lymph nodes were harvested and CD8+T cells
30 purified by CD8-negative selection as described above. OT1 T cells were then isolated by flow-assisted cell sorting by CD45-congenic staining. OT1 T cells were then added to cultures of B16 tumor cells expressing RFP in a 96-well plate at a 1:1 effector to target ratio. Incucyte active caspase-3/7 dye was added to the media as a means to quantify dying cells. Plates were then imaged using the IncuCyte ZOOM (Essen Bioscience). Curves of the total death or ratio of dead to live tumor target cells over time were generated by
quantifying 2-4 fields/well taken every 1-2hours. To assess the impact of mAbs on
TCR:peptide-MHC-I binding we cultured OT1 splenocytes with SIINFEKL peptide (2µg/mL) for 2 days followed by washing and resuspended the splenocytes in media supplemented with IL-2 (25U/mL) for 3 days. For the IncuCyte assay, OT1 T cells were isolated using the Biolegend CD8 negative magnetic selection kit at plated at an effector to target ratio of 10:1. Anti-CD8 mAbs were then added to the media at various concentrations during the cytotoxicity assay.
CD8+ T cell proliferation and early effects of anti-CD8 mAbs
CD8+ T cells from unmanipulated mice were purified using a BioLegend CD8 negative magnetic selection kit with a purity of at least 95% (as determined by flow cytometry) and stained with either Cell Trace Violet (Thermo Fischer) or
5(6)-carboxyfluorescein N-hydroxysuccinimidyl ester (CFSE; Sigma) proliferation dye for 15min in a 37∘C water bath and quenched/washed 2X and subsequently cultured with 5% FBS containing RPMI. The labeled cells were then plated at 1-1.5x105/well on a 96-well plate
previously coated overnight with anti-CD clone -2C11; Tonbo Biosciences) at 1-2µg/mL in PBS. Costimulation was provided in the media at 1µg/mL with anti-CD28
31 (37.51; BioLegend). Either anti-CD or – were added to the media at concentrations from 1ng-10µg/mL. Three days later cells were harvested for flow cytometric analysis.
For peptide stimulation experiments, CD8+ T cells were enriched magnetically with a BioLegend CD8 T cell negative-selection kit. OT1 splenocytes were cultured with 5% FBS containing RPMI in a 24-well cell culture plate. The immunodominant SIINFEKL (pOVA), SIYNFEKL (pY3), SIIQFEKL (pQ4), and the irrelevant SSIEFARL (pHSV-1) peptides were added to the media at 1µg/mL with or without anti-CD8 antibody at 1µg/mL and allowed to activate/proliferate. Three days later cells were harvested for flow cytometric analysis. Metabolic flux assay
CD8+ T cells from unmanipulated mice were purified using a BioLegend CD8 negative magnetic selection kit and plated at 0.5-1x106/well on a 24-well plate previously
coated overnight with anti-CD 145-2C11; Tonbo Biosciences) at 1-2µg/mL in PBS. Costimulation was provided in the media at 1µg/mL with anti-CD28 (37.51; BioLegend). Either anti-CD or – were added to the media at 10ng/mL. The CD8+ T cells were harvested 2-3 days later, seeded at 2x105 per well onto a 96-well poly-D-lysine coated
Seahorse plate in Seahorse XF RPMI with L-glutamine with or without dextrose. Cells were allowed to equilibrate for at least 1hr at 37∘C and 0%CO2 prior to assay. Oxygen
consumption rate and extracellular acidification rate were measured using a Seahorse XFe analyzer Agilent Technologies following the manufacturer s instructions in the Seahorse XF Mito Stress Test manual. Oligomycin (ATP-synthase inhibitor; final
concentration (FC) = 1µM), FCCP (uncouples the mitochondrial membranes; FC = 1µM), and antimycin A + rotenone (electron transport chain inhibitors; FC = 0.5µM each) were injected sequentially into wells with dextrose containing media to determine measures of
32 respiration. Similarly, cells seeded into glucose free media were assayed for glycolytic measures throughout injections of dextrose (10mM), oligomycin (ATP-synthase inhibitor; FC = 1µM), and 2-deoxyglucose (glycolysis inhibitor; FC = 50mM). All drugs were provided as part of manufacturer kits or purchased from Sigma and diluted in Seahorse XF media as recommended in the Seahorse XF Mito Stress Test manual.
Statistical analysis
Graphpad Prism 7 software was used for all statistical analyses. For Fig 4B
localization data and Fig metabolism data the Student s two-tailed, unpaired t-test was used. For all other data one-way ANOVA was applied and if a significant difference was determined multiple-comparisons of means was used to generate a p-value. For
cytotoxicity experiments, one-way ANOVA was applied to the area under the curve values generated from the time course curves. All error bars represent mean +/-SEM. * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001, non-significant differences were not displayed for simplicity.
33 CHAPTER III
CD8+ T CELLS THAT SURVIVE ANTIBODY-MEDIATED DEPLETION CAN PARTICIPATE IN IMMUNE RESPONSES
Introduction
Monoclonal antibodies have been used for decades with certain assumptions of their behavior, yet unintended consequences of their use discovered much later. As an example, both anti-Asialo-GM1 and anti-NK1.1 mAbs deplete natural killer cells from the blood efficiently, however, only anti-NK1.1 removes these cells from peripheral tissues effectively 91. As blood is frequently used as a convenient proxy to determine depletion
efficacy, it can appear that all natural killer cells are depleted after anti-Asialo-GM1 treatment when they are not. Anti-Asialo-GM1 has also been found to remove basophils, another unintended consequence of its use 92. Similarly, an anti-Gr1 mAb, commonly used
to deplete neutrophils, also removes memory CD8+ T cells that could change how results are interpreted93. The use of depleting mAb reagents most successfully in the clinic will
require an understanding of how the remaining population is altered. This is already a concern in the case of transplant rejection, where the approved treatments to deplete T cells can eventually lead to transplant rejection that is caused by effector-memory T cells that failed to be depleted and have undergone robust homeostatic expansion94.
Here we chose to focus on CD8+ T cells for two reasons. First, there are two main target molecules for depletion, CD and CD , whereby targeting of either could result in differences in how we perceive the survivors. Second, the mAbs commonly used to deplete CD8+ T cells often leave behind a population amenable to study95, 96, 97, 98, 99, 100, 101. Herein
34 significantly less able to deplete CD8+ T cells relative to an anti-CD8 , clone 53-5.8.
Further, these mAbs remain bound to the surviving CD8+ T cells and alter how they would be identified/accounted for. By blocking each other s subsequent binding and differentially inducing CD8 downregulation/internalization these mAbs make it more difficult to detect this surviving population by conventional staining and flow cytometry. This is made more significant by the long half-life of these antibodies in serum. These results have a potential impact on experiments whose interpretation assumes the absence of CD8+ T cells. The significance of or magnitude of surviving CD8+ T cells is likely largely hidden as many reports do not explicitly say how or if the surviving population is ever identified97, 102, 103, 104, 105, 106, 107, 108 and/or what mAb and in what amount was used to deplete88, 109, 110, 111, 112, 113.
Results
CD8+ T cells that survive depletion by mAb treatment differentially internalize CD8 and retain mAb on their surface
It is not uncommon practice to deplete and verify the efficacy of depletion via flow cytometry with a fluorochrome-labeled mAb of the same clone96, 101, 102, 114. Therefore we
sought to determine if antibody remains on a target cell after its use for depletion. This is important to assess, as it could interfere with the determination of depletion efficiency and may alter their activation. A low dose of 10µg of either anti-CD , anti-CD , or µg of each together were administered to unmanipulated mice i.p and then splenocytes were
harvested for staining the next day. A low dose of either mAb was able to significantly reduce CD8+ T cells (gated as B220- CD4- CD3+) within the spleen with anti-CD having the greatest depletion, similar to what has been previously shown (Figure 3-1A)79. The
35 Figure 3-1. CD8+ T cells are not completely removed by mAb-mediated depletion and are more difficult to identify post-depletion.
36 Figure 3-1. CD8+ T cells are not completely removed by mAb-mediated depletion and are more difficult to identify post-depletion. (A-C) Naïve mice were treated i.p. with a total of 10µg of: anti-CD only, anti-CD only, or µg of both. Splenocytes were isolated and stained the next day. CD8+ T cells were considered CD3+ CD4- B220-. (A) CD8+ T cells as a percentage of lymphocytes and total numbers were calculated. (B) Remaining CD8+ T cell surface bound depleting mAb was determined by staining with Goat anti-Rat IgG. (C) Levels of unbound CD and CD on surviving CD + T cells was determined by staining with the same clones used to deplete. (D) Unmanipulated splenocytes were kept on ice and stained with either anti-CD or – and then stained with the other. Light gray line
represents B220+ cells. Data representative of at least 3 independent experiments, with 3-4 mice per group. One-way ANOVA was applied and if a significant difference was
determined multiple-comparisons of means was used to generate a p-value. Error bars represent mean +/-SEM. * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001, non-significant differences were not displayed for simplicity.
37 combination treatment did not increase depletion efficiency, potentially due to decreased amount of the more effective mAb being used. To test if CD8 was accessible to further staining after depleting mAb treatment, CD and CD were stained with the same clones used for depletion (Figure 3-1C). We could detect bound depleting-rat IgG on CD8+ T cells of both anti-CD and anti-CD treated mice but not untreated or isotype control treated mice (Figure 3-1B and data not shown). A substantial increase in retention of surface bound mAb was noted for anti-CD treated over anti-CD or combined treated animals. Thus, CD engagement likely results in increased internalization of CD relative to CD engagement. This is further supported by an equivalent drop in CD staining when anti-CD or – is used to deplete. anti-CD was readily detectable in all groups, yet anti-anti-CD also resulted in lowered CD staining indicating some level of CD internalization and/or steric hindrance of anti-CD blocking binding of anti-CD . To test for steric hindrance, splenocytes were stained with either anti-CD or – alone on ice for and then the reciprocal anti-CD8 mAb was similarly used to stain (Figure 3-1D). Surprisingly, both mAbs were able to impede the subsequent staining of the other, showing clearly separate
populations when overlaid. Combining this steric hindrance with mAb-mediated CD8 internalization, the surviving population may easily be mischaracterized as CD8 negative and remain undetected. Thus, many of these commonly used mAb reagents need to be tested thoroughly for a given system to be sure depletion is sufficient and the ability to verify depletion is adequate.
In addition to T cells, CD8 is expressed as a CD8 / homodimer on dendritic cells that occupy the T cell zones, the periarteriolar lymphoid sheath of the spleen and
38 activation of naïve T cells4. Therefore, it is important to ensure that these CD8+ DCs were
not also depleted or perturbed by anti-CD8 mAb treatment meant to deplete CD8+ T cells. To assess this mice were injected i.p. with 100μg of anti-CD , anti-CD , or a combination of IgG2a and IgG1 isotype control mAbs and simultaneously immunized. Flow cytometry was done on whole splenocytes the next day. I saw that despite binding of anti-CD mAb to DCs there was no significant change in their numbers (Figure 3-2A and B) or their expression of the costimulatory markers CD86 and CD80 that support CD8+ T cell activation (Figure 3-2C and D). Thus, the anti-CD mAb does not affect CD + DCs and therefore any effect seen with its use is likely resultant from direct action on CD8 expressed on T cells themselves.
CD8+ T cells that survive depleting mAb treatment are present throughout primary immune responses and retain bound anti-CD8 & - mAb.
To begin to understand how bound depleting mAb could affect an immune response I first sought to address whether mAbs remains bound throughout a primary challenge. To address this 106 CD45.1+ OT1 CD8+ T cells (specific for the SIINFEKL peptide of ovalbumin
protein) were adoptively transferred into CD45.2+ C57BL/6 mice. Using transferred OT1 T cells ensured a population of antigen-specific CD8+ T cells post-immunization/infection remained after depletion that could be identified with a congenic CD45 marker. The next day mice were treated with a high dose (500µg) of either anti-CD or anti-CD i.p. or left untreated. For these experiments I chose a high dose of depletion mAb, in line with
previous reports95, 99, 115, to ensure that the maximum proportion of CD8+ T cells were
exposed to a saturating amount of mAb. The following day the depleted mice were either immunized against whole ovalbumin (OVA) with poly(I:C) and anti-CD40 mAb as adjuvants
39 Figure 3-2. CD8+ dendritic cells are not affected by depleting anti-CD8 mAbs. (A-D) C57BL/6 mice were treated i.p. with 100μg anti-CD , anti-CD , or a combination of isotype control mAbs and numbers of DCs (gated as MHC-II+ CD11c+) or levels of select costimulatory molecules determined by flow cytometry. (A) Representative flow plots of total DCs (top) and of DCs separated into either CD11b+ or CD8+ (bottom). (B) Numbers of total DCs. (C) Surface expression of CD86 of CD8+ DCs. (D) Surface expression of CD80 of CD8+ DCs. A & B are representative data from 3 independent experiments and C & D are representative data from 2 independent experiments, with 3 mice per group with each mouse represented by a point. One-way ANOVA was applied and no significant differences among the means were noted.
40 or infected with vaccinia virus expressing ovalbumin (VV-OVA). Seven days after
immunization or infection, both anti-CD and – surviving OT T cells had readily detectable mAbs bound, confirmed by staining directly for rat IgG (Figure 3-3A).
Interestingly, the levels of expression/retainment were different between anti-CD and – dependent on whether the mice were immunized or infected. This may be important in regards to potential direct effects of bound anti-CD8, including signaling and modulation of TCR:peptide-MHC-I interaction that could continually impact these CD8+ T cells and alter their behavior. As expected, both anti-CD and – treated groups had drastically fewer OT1 T cells during the primary challenge to either vaccinia or immunization (Figure 3-3B). Generally, there were less survivors of the anti-CD treatment consistent with the
efficiency of depletion prior to challenge (Figure 3-1A). After immunization, a larger difference in surviving OT1 T cells trended within the LNs compared to the spleen. These results confirmed that bound mAbs are present throughout a response. Furthermore, the differences in total surviving cell number between the spleen and LNs across challenge types suggested both depletion and challenge strategies can affect T cell localization.
Another factor to consider for how differences in CD8+ T cell phenotype may be arising from anti-CD8 mAb treatment is how long those antibodies remain in circulation. This is of particular interest for long term experiments past the primary response. I
demonstrated that anti-CD8 mAb is bound to CD8+ T cells 7 days into an immune response (Figure 3-3), however, I do not know if those mAbs remain bound or alternately, remain in the circulation available to continually rebind, or a combination of both. To address this I i.p. injected 500µg of either anti-CD or – and their corresponding isotype controls (IgG2a and IgG1 respectively) and tracked their serum titers over the course of 2 weeks
41 Figure 3-3. CD8+ T cells that survive depletion are present throughout immune responses. (A-B) 106 CD45.1+ OT1 T cells were transferred i.v. into CD45.2+ mice, treated
i.p. with 500µg of anti-CD or – the next day, and immunized or infected. Spleens and lymph nodes were collected on day 5 post-infection or on day 7 post-immunization. CD8+ T cells were considered as CD45.1+ B220-. (A) Surface bound depleting mAb on OT1s was determined by staining with Goat anti-Rat IgG. (B) Remaining OT1 T cells as a percentage of lymphocytes and total numbers were calculated per spleen or pooled lymph nodes. Data are representative of at least 3 independent experiments, with 3-4 mice per group each mouse represented by a point. One-way ANOVA was applied and if a significant difference was determined multiple-comparisons of means was used to generate a p-value. Error bars represent mean +/-SEM. * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001,
42 (Figure 3-4). Surprisingly, the anti-CD8 mAbs persisted in circulation nearly as long as their isotype controls. This is contrast to other reports that show a more limited serum
persistence of mAbs 116. However, the anti-CD did drop more precipitously between
and 14 days after injection than either anti-CD or isotype control mAbs. CD8+ T cells that survive depletion will thus likely have anti-CD8 mAbs continually bound and alter the TCR signal strength and potentially signal directly themselves over the course of and past the primary immune response.
Discussion
Despite decades of work done with lymphocyte depleting mAbs, there is a
surprising dearth of information as to how these reagents work and what unintended side effects occur from their use. Herein, I demonstrate that CD8+ T cells that survive mAb-mediated depletion can retain the depleting mAb on their surface, alter the surface expression of CD8, and are able to participate in immune responses. Further, these
depleting anti-CD8 mAbs, despite actively binding targets, are stable and likely to remain in serum for the duration of at least the primary immune response. This highlights the
importance of knowing how well depleting reagents and protocols used within a given experimental system work and that it may be worthwhile to track the surviving population throughout the experiment.
As shown above, how the remaining CD8+ T cell population appears after depletion is dependent on several parameters. The clone or subunit targeted of a depleting mAb used can substantially differ in its effect on CD8 internalization as well as interfere with
subsequent staining of CD8, both of which can prevent remaining CD8+ T cells from being accounted for. The ability of one mAb to block binding of another even when it is targeting
43 Figure 3-4. Serum clearance of CD8 depleting mAbs. 500µg was injected i.p. into
C57BL/6 mice and serum isolated from blood taken at 3 hours, 4, 7, and 14 days after injection. EC50 was determined by sandwich ELISA using a Goat anti-Rat IgG antibody for
capture and isotype-specific Goat anti-Rat antibodies conjugated to alkaline phosphatase for detection. Data are expressed as the arithmetic mean. Data are from one experiment with four mice per anti-CD8 mAb treatment group or 3 mice per isotype control treatment group. 3 h ou r s 4 d ay s 7 d ay s 14 da ys 0 1 0 0 4 0 0 0 6 0 0 0 8 0 0 0 1 0 0 0 0 S e r u m c l e a r a n c e T i m e p o s t - i n j e c t i o n E C 5 0 I g G 2 a i s o t y p e a n t i - C D 8 I g G 1 i s o t y p e a n t i - C D 8
