To affinity and beyond: Harnessing the T Cell receptor for cancer immunotherapy

Abstract

T cell adoptive therapies for immune-mediated regression of cancers have attracted a great deal of recent attention. Clinical results are glamorous, yet much remains to be uncovered behind the basic science that allows us to engineer T cells and T cell receptors (TCRs) for clinical use. We discuss the development of TCRs for therapeutic use in the context of thymic selection toward central tolerance and we review therapies based on tumor infiltrating lymphocytes (TILs), endogenous antigen specific TCRs, and engineered TCRs. Further we discuss the development of low and high affinity TCRs and the extent to which each challenges central tolerance. Current results suggest that adaptation of TCR engineering of moderate affinity TCRs coupled with co-regulatory and stimulatory molecules may be the safest and most efficacious road for TCR development aimed at tumor abolition.

Keywords:

• affinity
• adoptive cell therapy
• cancer
• co-receptor
• T cell receptor
• tumor

Abbreviations

TCR	=	T cell receptor
CDR	=	complementarity determining region
MHC	=	major histocompatibility complex
SLEC	=	short-lived effector cell
mTEC	=	medullary thymic epithelial cell
TSA	=	tissue-specific self-antigen
AIRE	=	autoimmune regulator
TIL	=	tumor infiltrating lymphocyte
TAA	=	tumor-associated antigen
CTA	=	cancer testis antigen

Introduction

T cells are a highly attractive immune cell subset for cancer immunotherapy. The intrinsic properties of T cells coupled to clinical developments in engineering them have made T cells potent anti-tumor agents. Optimal activation of T cells depends on 3 signals: TCR engagement (signal 1), co-stimulation (signal 2) and an inflammatory stimulus (signal 3) delivered via cytokine. Here we focus on the characteristics of signal 1 that lead to effective tumor clearance as well as the potential for amplification of signal 1 through modulation of signal 2. T cell receptors (TCRs) have the capacity for exquisite antigenic specificity and this can be employed toward tumor antigens.¹ Upon tumor-antigen recognition T cells impart potent cytotoxicity that can aid tumor cell death. Further, T cell memory function allows for cellular persistence to antigenic epitopes. Manipulation and amplification of the aforementioned T cell attributes poses a powerful anti-cancer agent.² A challenge to adoptive cellular therapy with T cells is that tumor antigens are often non-mutated self-antigens and development of high affinity TCRs to tumor antigens faces the paradigm of central tolerance. Moderate affinity TCRs aimed at tumor antigen harbor less potential for toxic responses and may be as effective at mediation of anti-tumor activity, when enhanced by co-receptor and co-stimulatory molecules, as high affinity TCRs.

Composition

The T cell receptor (TCR) on a majority of T cells is composed of 2 chains, α and β. In this review, T cells bearing γ and δ receptors known as γδ T cells will not be discussed. The chains converge at the cell surface to form a heterodimeric receptor that has exquisite antigen specificity. Variability in the α and β chain sequences allows for 1 × 10¹³ individual TCRs after thymic selection with the potential to recognize a similarly diverse set of antigenic epitopes.³ TCR precision for antigen is endowed by 3 complementarity determining regions (CDRs) that comprise each TCR chain. Somatic recombination, or V(D)J recombination, is the process of random gene rearrangement within each hypervariable region of the CDR that allows for vast TCR diversity. CDR1 and CDR2 segments are positioned to make primary contact with the major histocompatibility complex (MHC) during antigen presentation.⁴ CDR3 is situated to interact with the central region of presented antigen concomitant with the fact that CDR3 shows the greatest variability of gene recombination.⁵

Kinetics

Though the structure of the TCR determines the peptide sequence that is recognized, the added parameter of TCR signal strength and subsequent outcome, on account of affinity for antigen, is under investigation. Several models exist that attempt to decipher the precise strength of a TCR signal generated upon peptide-MHC (pMHC) interaction.⁶ It was first proposed that the dissociation rate (K_d) (K_d = [TCR][pepMHC]/[TCR:pepMHC]) of TCR complexed with pMHC was determinant of the magnitude of T cell activation.^7,8 Interaction of TCR:pMHC with sufficiently long dwell time leads to an intracellular signaling cascade that promotes Ca²⁺ release associated with T cell activation.^9,10 A second proposal marked the idea that the number of TCR-pMHC interactions, as a function of affinity for antigen and ligand density, was determinant of T cell activation levels.¹¹ Further work showed that TCR affinity for pep:MHC may regulate the type of signal generated and subsequent progeny produced.¹² High affinity TCRs interact long enough with pep-MHC to allow an asymmetric generation of daughter T cells and subsequent development of short-lived effector cells (SLECs). In contrast, low affinity TCRs were not in contact with APCs at the time of cell division and offspring lacked SLEC capacity, resultant in a feeble immune response.¹³

Signals

The downstream signals that are generated by TCR engagement are complex and continue to be revealed. Recent data suggest that the early event of actin remodeling is critical for ZAP70 and LCK phosphorylation and induction of Ca²⁺ flux.¹⁴ The very early TCR signal cascade leads to perturbations in Ca²⁺ at the level of the endoplasmic reticulum.¹⁵ For the CD4 T cell lineage differential Ca²⁺ signals are a determinant of lineage commitment.¹⁶ Ca²⁺-mediated de-phosphorylation of NFAT and its subsequent nuclear translocation are early hallmarks of T cell activation.¹⁷ Taken together, very early signaling events have been shown to be graded in intensity and may provide a plausible mechanism to target for enhanced levels of T cell activation that circumvent the TCR altogether.

Downstream measurements that affect T cells engineered for the tumor microenvironment are cytokine and cytotoxic granule release resultant in tumor cell killing as well as ability to differentiate into an effector cell. Study of the immunological synapse has shown the link between early TCR driven molecular signaling through ZAP70 and initiation of cytotoxic granule polarization and release.¹⁸ The ability to initiate a cytotoxic response as measured by granule release may be a property of high affinity TCR interactions.¹⁹ Elucidation of these data is critical for the development of tenable T cells raised against tumor antigen.

II. Thymic selection: The natural process

Central tolerance

Selection of a high affinity TCR specific for tumor antigen is a dichotomous pursuit. On one hand high affinity TCRs produce strong signals that enable the aforementioned traits of T cells to flourish. On the other hand, a high affinity TCR to tumor antigen may, and has, caused adverse reaction to self-antigens (tumor-antigen). Central tolerance encompasses the mechanisms in the thymus that are in place to insure lack of T cell reaction to self-peptide upon entry into the periphery.²⁰ Clonal deletion, elimination of TCRs that present with high affinity to self peptide, and clonal diversion, conversion of moderate to high affinity TCRs to regulatory T cell (Tregs) lineage are the key processes that impact central tolerance.^21,22 Events that lead to central tolerance are key to the development of TCRs for adoptive cell therapy.

Clonal deletion

In the thymus cortical and medullary thymic epithelial cells present fragments of self-peptide loaded into the groove of MHC class I or II molecules directly after processing in the endoplasmic reticulum and endocytic vesicles, respectively.^23,24 Through this process the TCRα chain continually rearranges, with CDR3 region having the most promiscuity, and CDR1 and CDR2 poised to stabilize the interaction with MHC.²⁵ The affinity hypothesis posits that TCRs with low affinity for self-antigen are selected to survive and are primed for interaction with foreign antigen (positive selection), while TCRs with high affinities for self-peptide are sentenced to death (clonal deletion) or enter an anergic state.^26-28

Thymic Signals for TCR Development

Advances have begun to elucidate the signals that lead to survival or anergy/death to define the TCR repertoire available in the periphery, yet more research is needed. As methods for antigen-specific T cell selection and TCR engineering proceed, it is important to keep in mind that very specialized APCs present to naturally tune TCRs. Medullary thymic epithelial cells (mTECs) are specialized to present tissue specific self-antigens (TSAs).^29,30 mTECs express the transcription factor AIRE (autoimmune regulator) and AIRE promotes presentation of a large variety of TSAs to TCRs. AIRE induces chemokine expression in the thymus to direct cellular localization.³¹ The importance of specialized cells for antigen presentation to mature affinity specific TCRs is highlighted by the fact that AIRE deficient mice show altered maturation of thymocytes and incidentally develop spontaneous multi-organ disease. ³²

The signals that are orchestrated within a T cell upon TCR sampling in the thymus are under intense investigation. The zipper model proposes that dwell time of pMHC complexed to TCR-CD3 leads to positive or negative selection. A dwell time of greater than 4 seconds effectively “zippers” the TCR together with its co-receptor to allow complete phospohorylation of CD3 ITAMs and subsequent negative selection due to strong affinity TCR. A low affinity TCR will have a dwell time of less than 4 seconds and incomplete or transient activation of ITAMs will occur, ultimately inducing a survival signal, or positive selection.³³

Downstream of early TCR signal events, the kinetics of Ca²⁺ flux are essential to appropriate TCR selection. Ca²⁺ stores and ERK work in tandem at the level of the ER. Strong TCR signals result in Ca²⁺-mediated robust activation of pERK and subsequent negative selection. Positive selection is associated with less Ca²⁺ flux and low-level pERK activation that leads to a sustained response.³⁴ The protein themis is expressed at high levels in the thymus and was recently shown to regulate Ca²⁺ flux as an intricate player in TCR selection.³⁵ The precise mechanism of how high affinity and low affinity TCRs are discriminated between in the thymus remains unknown and further discovery of this process is important to the field of TCR engineering for immunotherapy.²⁰

Clonal Diversion: A Treg is Born

The process of clonal deletion is incomplete and TCRs with low affinities for self-antigen “escape” into the periphery with the potential to create autoimmune responses. Tregs are a specialized set of CD4/Foxp3⁺ T cells endowed with capacity to regulate auto-reactive T cells and comprise the heart of peripheral tolerance.³⁶ Tregs must undergo their own clonal selection in the thymus and this process is dubbed clonal diversion. TCR specificity for self-peptide plays a role in the formation of Foxp3⁺ cells. Early studies with transgenic TCRs specific to foreign antigen show that Tregs developed intra-thymically only when foreign antigen was present.^37,38 More recent work has used transgenic TCRs cloned from naturally occurring Tregs to show that self-reactive TCRs are required for direction of Foxp3⁺ cell formation in the thymus and that a selection niche is also at play.^21,22

Separate signals?

Downstream signals, pERK and Ca²⁺ flux, in regards to the zipper model that mediate Treg formation need to be addressed.³³ One study used a TCR signal strength reporter mouse to show that cells that underwent clonal deletion had greater expression of a fluorescent reporter protein than did Tregs.³⁹ These data agree with the prevailing idea in the field that Treg TCR affinity is intermediate to positively selected TCRs and those targeted for deletion. Interestingly, Tregs in the periphery do not flux Ca²⁺ efficiently, similar to a cell with anergic properties, or produce IL-2, a downstream product of Ca²⁺ driven NFAT activation.⁴⁰ As Treg specific tissue-restricted antigens are discovered, it will be important to elucidate the TCR:pMHC affinities and resultant signals that occur to drive the Treg lineage. This information will aid in development of high affinity TCRs for tumor eradication and elucidate requirements for self-antigen signals in order to develop novel therapeutics that enhance TCR-mediated signals.

Peripheral Tolerance

In the context of cancer Tregs mediate cancer-induced inflammation in a protective manner.⁴¹ In contrast, Tregs are key players in tumorigenic tolerance and growth as suppressors of cytotoxic immunity.⁴² It is critical to grasp Treg-mediated immune suppression of T cells that have escaped central tolerance in order to appropriately design T cells armed to overcome tumorigenic tolerance. Tregs have several modes of suppression and we will highlight those with potential to target effector cell TCRs in the tumor microenvironment.

Tregs inhibit TCR Induced Proliferation of Effector T Cells

Several groups have demonstrated that Tregs inhibit activation of effector T cells upon TCR activation with CD3/CD28 in the presence and absence of APCs. Further studies saw rapid inhibition of cytokine transcription (IL-2) in conventional T cells.^43,44 A recent report elegantly demonstrated a TCR driven mechanism for these aforementioned data in which pre-activated Tregs co-cultured with conventional T cells led to immediate suppression of effector T cell signaling events through inhibition of Ca²⁺-store release and subsequent de-phosphorylation of NFAT. Further, Ca²⁺-independent events in conventional T cells were not affected.⁴⁵

CTLA4

Cytotoxic T lymphocyte-associated protein 4 (CTLA4) is a co-inhibitory molecule expressed constitutively at the cell surface of activated Tregs. The mechanisms of effector T cell suppression that occur via CTLA4 upregulation on Tregs are indirect and mediated by CTLA4 interaction with APCs.⁴⁶ CTLA4 expression causes down modulation of co-stimulatory molecules CD80 and CD86 on dendritic cells (DCs) and this inhibition can be recused by blockade of CTLA4. Treg-CTLA4 conditioned APCs cause impaired proliferation of effector T cells.⁴⁷ Further, decreased glutathione synthesis in DCs mediated by Treg-CTLA4 leads to a higher redox potential at the immunological synapse and results in an unfavorable environment for effector T cell survival and proliferation.^48,49

Abrogation of CTLA4 expression in mice leads to fatal multi-organ tissue destruction due to hypo-proliferative lymphocytes and indicates the importance of CTLA4 as a negative regulator upon T lymphocyte activation.⁵⁰ Treg-specific deletion of CTLA4 leads to loss of Treg-mediated mechanisms of suppression.⁵¹ Blockade of CTLA4 is an approved FDA therapy that aims to reverse tumor-mediated tolerance of tumor infiltrating lymphocytes (TILs) and has resulted in increased survival times in metastatic melanoma patients when compared to traditional treatments. Not surprisingly, CTLA4 blockade results in immune-related adverse events in metastatic melanoma patients, namely hepatitis, enterocolitis, hypophysitis and uveitis. These adverse events can be attributed to loss of peripheral regulatory immunity through impairment of endogenous CTLA4-Treg-mediated control of auto-reactive T cell response to self-antigen.⁵²

PD-1

Programmed cell death 1/ Programmed cell death ligand-1 (PD-1/PD-L1) pathway blockade similarly aims to abrogate immune checkpoint signals to enhance effector T cell cytotoxicity in the tumor microenvironment.⁵³ PD-1 is induced by TCR signals and is highly expressed on antigen-experienced T cells at sites of inflammation in tissues. Further, while PD-1 deletion in mice results in lupus-like autoimmune disease, mice do not develop a lethal lympho-proliferative phenotype as seen in CTLA4 deficient animals.⁵⁴ Along these lines, PD-1 expression has proven critical for the development and function of inducible Tregs more so than natural Tregs.⁵⁵ Blockade of PD-1 in mice with chronic viral infection was able to reverse PD-1-mediated CD8 T cell exhaustion and restore viral control associated with virus-specific T cell function.⁵⁶ These data present PD-1 as a favorable target to reverse TIL exhaustion, or tumor-tolerance, in the tumor microenvironment. Indeed, clinical trial data suggest that PD-1 blockade enhances patient survival through restoration of TIL function with limited immune-related toxicities in patients.⁵³ Further, recent repots also identify PD-1 as a novel biomarker for highly tumor-reactive CD8 T cells in metastatic melanoma patients.⁵⁷

III. TCRs present

Here we present an overview of endogenous and engineered TCRs that have been used clinically for treatment. We discuss selection of TCR affinities and highlight adverse trial outcomes associated with central tolerance. Lastly we discuss ways to augment T cell activation and killing capacity in tumor.

A. Endogenous T cells

Manipulation of the endogenous T cell pool for tumor-antigen specificity poses the question: Where T cells? Antigen-specific T cell therapies and TIL therapies are based on the concept that antigen-specific T cells are present within a patient's T cell pool and can be expanded to target and kill tumor.⁵⁸ Key studies led to the proof that T cells specific for tumor antigen could be isolated from the periphery. ∼2% of peripheral CD8 T cells were specific for metastatic melanoma tumor antigen MART-1.^59,60 A second rich source of T cells proved to be the tumor itself. A T cell line established from metastatic melanoma tumor was cultured and re-infused, along with high-dose IL-2, into the subject to result in cancer regression. The precise T cell line was found to be specific to melanoma tumor-antigen gp100.⁶¹ These striking data are the result of methodologies developed to identify endogenous cytotoxic T cells with TCR specific for tumor antigen.

Peripheral Antigen-Specific T Cells

While ∼2% of T cells in peripheral circulation showed MART-1 tumor-antigen specificity, the true percentage for the majority of cancer types of tumor antigen-specific T cells is likely 10–20-fold lower.^62,63 Thus, the methodologies for isolation of antigen-specific T cells are intricate. Firstly, T cells from peripheral blood must be stimulated for selection with one of several antigen presentation methods. Next, clonal selection or cell sorting to target the antigen-specific T cell is undertaken. Finally, cytokine cocktails and CD3 stimulation are used for expansion of the antigen-specific T cell population and subsequent reinfusion to the subject.^58,64 Expansion of antigen-specific T cells from the periphery has drawbacks that will hinder entry into mainstream cancer care. Firstly, paucity of antigen-specific T cells in the peripheral pool makes this method analogous to “finding a needle in a haystack.” Secondly, methodology for the steps of clonal stimulation and T cell selection need to be centralized.

TILs

The method of TIL isolation, culture, and reinfusion is promising and has been refined toward centralized methodology. The nature of TILs is that they are found at the source of antigenic response, thus the potential is enhanced to culture antigen-specific T cells.⁶⁵ Along these lines, TILs have potent killing capacity when compared to lymphokine-activate killer cells and this is a likely property of natural activation to tumor-antigen at tumor site.⁶⁶ The young TIL method begins with dissociation of tumor tissue and culture of TILs with IL-2 for 4–5 weeks to expand the T cell population. Further, CD8 T cells are enriched via magnetic bead sort and re-infused to a patient after he is conditioned with lymphodepleting chemotherapy regimen.⁶⁷ Clinical drawbacks of TIL are largely associated with the lymphodepletion phase that must occur. First, patients must be hospitalized in acute care setting, which is costly. Second, engraftment of a highly active T cell clone to immune-depleted patients may lead to loss of normal immune repertoire and regeneration of skewed immune responses. ⁶⁸

Low Affinity TCRs Challenge Central Tolerance

Whether tumor-antigen specific T cells are harvested from the periphery or tumor, each method proposes to expand TCRs specific for self-antigen, and thus challenges the paradigm of central tolerance. Early data from metastatic melanoma patients treated with immune-modulatory therapies led to the observation that patients experienced autoimmune vitiligo.⁶⁹ Specific data showed that MART1 tumor-antigen specific CD8 T cells, isolated and expanded from subject peripheral blood, mounted an autoimmune response and caused damage to melanocytes that expressed MART1.⁷⁰ Similarly, TILs expanded from metastatic melanoma patients produced uveitis or vitiligo in 5 of 13 patients treated.⁶⁸ The aforementioned adverse events were non-lethal and treatable. Application of endogenous-derived T cells in cancer types other than melanoma will be important to determine safety of these therapies.

Low-Affinity TCRs

It is unexpected that self-reactive T cells can be found in the periphery or migrated to the tumor at all, given that clonal deletion aims to rid the body of such T cells. Further, clonal diversion posits that TCRs with moderate affinity for self-antigen will be redirected toward Treg lineage. Indeed, the majority of tumor antigens discussed above that have led to severe autoimmune responses in adoptive T cell therapy settings are found in the thymus and presented by mTECs during T cell selection.⁷¹ Thus, endogenous T cells that are selected and expanded for T cell therapies are truly low affinity TCRs for self-antigen, as high affinity TCRs have undergone clonal deletion (Fig. 1).

Figure 1. Graduated increase in TCR affinity also increases with predicted tumor regression and autoimmune risk (or challenge to central tolerance). In accordance, peripherally isolated antigen-specific T cells are predicted to be slightly less potent for tumor regression given that TILs develop in the tumor microenvironment under more significant selection pressures than peripheral T cells. Therapeutically manipulated TCRs have paramount antigenic specificity and predicted anti-tumor activity, and challenge central tolerance to the greatest degree.

A recent study in TCR transgenic mice examined the affinity of a TCR specific for a tissue-restricted antigen. A sub-population of transgenic T cells escaped negative selection and was found in the periphery, although they possessed high affinity for self-restricted antigen. Further, these cells were able to activate and cause unchecked autoimmune responses in the face of infection.⁷² These results challenge the belief that negative selection eliminates highly self-specific TCRs and show that endogenous tumor-antigen specific T cells may circulate in the periphery and possess reactive potential to self-antigen when provoked.

The posit that high affinity TCRs have profoundly enhanced functions as compared to low affinity TCRs comes from the in vitro setting where TCR:pMHC affinities may be unusually strong. Further, antigens that have been analyzed have not been true-self antigens and this may skew results.⁷³ A microbial infection model was used to delineate the kinetics of response for individual T cell clones of varying TCR affinities. Surprisingly, T cells with low affinity TCRs became activated, proliferated, and generated memory CD8 T cell responses at levels equal to high affinity clones. However, TCR affinity for ligand did impact the timing of T cell expansion, extravasation from lymphoid organs, and onset of contraction phase.⁷⁴ Further, it was shown that low affinity virus-specific T cells efficiently mediate viral response and clearance.⁷⁵ In vitro, NY-ESO-specific TCRs of various affinities were generated to assess the role of TCR affinity on CD8 T cell activation. Intermediate TCR-pMHC parameters conferred maximum CD8 T cell responsiveness as measured by calcium flux, chemokine/cytokine secretion, and cytotoxicity.⁷⁶ In line with these findings, various affinity TCRs to gp100 human melanoma antigen were assessed in vivo to determine the optimal TCR:pMHC strength for anti-tumor CD8 T cell activity. Strikingly, high affinity TCRs with greater pERK, Ca²⁺ flux, and cytokine release did not lead to greater anti-tumorigenicity. Instead, a plateau existed where maximum cytotoxicity and autoimmune reactivity were reached at the level of moderate TCR affinity to tumor antigen.⁷³ Further, TCR transgeneic mice with recognition for tissue-restricted TRP-1 melanoma antigen were generated. Two lines of mice differed in their affinity for TRP-1 antigen (high or low affinity TCRs) as evidenced by tetramer dissociation rate and peptide responsiveness. Surprisingly, the high and low affinity TCRs generated showed equivalent anti-tumor activity as assessed by CD8 cytotoxic parameters.⁷⁷ Taken together these data present a salient picture that calls into question the relevance of high affinity TCRs for adoptive immunotherapy.

B. Genetically Engineered T cells

Endogenous low affinity TCRs obtained from patient peripheral blood or TILs require higher levels of pMHC to be present for the number of peptide:ligand interactions to be sufficient for activation to occur.⁷⁸ Further, low-affinity TCRs require increased CD8 co-receptor stabilization, DC-derived co-stimulation, and blockade of co-inhibitory receptors to enhance activation and function.⁷³ In keeping with the zipper model, co-stimulatory modifications may allow increased dwell time of low affinity TCRs and result in full activation of T cells. Thus, the potential for high affinity TCRs engineered to combat tumor antigen is relevant. Many possibilities exist with the advantageous strategy of engineering T cells to express a pre-specified TCR with high affinity for tumor antigen.

High Affinity TCRs

High affinity TCRs are effective therapies that may improve anti-tumor cytotoxicity as the number of peptide:ligand interactions necessary for T cell activation is minimized and the dwell time for signal activation is optimal. TCR genes that encode α and β chains can be derived from highly effective TILs or transgenic mice that are engineered to express human MHC.⁷⁹ Retroviral or lentiviral transduction strategies are applied to introduce TCR genes into the T cell.⁸⁰ Early technical problems for TCR based therapies were low levels of expression of transferred TCR at the cell surface and mispairings of host α and transferred β chains (or vice versa).⁸¹ Recent advances in TCR engineering resolved these difficulties. Knock down of endogenous TCRs has increased expression of transferred TCRs at the T cell surface. Introduction of cysteine residue technology allowed for transferred α and β chains to pair appropriately.⁸⁰

Which Antigen?

An obvious benefit of TCR engineering is the ready availability of a high affinity tumor-antigen specific TCR. However, this technology begs the question: Which antigen? A requisite for tumor-associated antigen (TAA) is that it is expressed exclusively by tumor cells and at levels sufficient to evoke a T cell response. Thus far, differentiation antigens and cancer testis antigens (CTAs) have been attractive targets. RT-PCR data demonstrated lack of expression of CTAs in normal tissues and upregulation of CTAs was found in tumor tissues of various origins.⁸² Further, CTAs have been shown to evoke favorable effector T cell responses in cancer patients when CTA-specific peptides were loaded to DCs and vaccinated to metastatic melanoma patients.⁸³ Taken together, the aforementioned data set describes motives behind tumor antigen selection for development of high affinity TCRs.

High Affinity TCRs Challenge Central Tolerance

High affinity TCRs against differentiation antigens MART-1, gp100 and CTAs: MAGE-A1, MAGE-A3, and NY-ESO1 have been developed and placed in clinical trials. Induced autoimmune responses due to introduction of high affinity self-specific TCRs have been the hamartia of clinical trials thus far. Of 8 clinical trials assessed reviewed by Kuert et al, 5 report therapy-associated toxicities and 3 of 8 trials reported lethal toxicities likely owing to specificity of transferred TCRs for tissue-restricted antigen.⁸⁴ 2 of 3 studies with lethal toxicities were targeted to MAGE-A3 antigen. Autopsy of patients showed dysregulated brain function and pointed to a previously unappreciated expression of MAGE-family proteins in brain tissue.⁸⁵ Histological analysis confirmed strong cross-reactivity of the MAGE-A3 specific TCR with brain-associated MAGE protein. Cardiac tissue from patients with lethal toxicities associated with heart failure showed no expression of MAGE protein in cardiac tissues; however, it was determined that the high affinity TCR cross-reacted with the cardiac protein of striated-muscle, titin.⁸⁶

Differential results in clinical trials regarding efficacy and toxicities between low affinity and high affinity TCRs have brought to light several critical thoughts. First, it is key to note that low affinity TCRs are able to cause tumor regression when harvested and expanded from the tumor microenvironment. Just as mTECs express AIRE, it is likely that APCs in the tumor microenvironment play an underappreciated role in maturation of low affinity TCRs to TAAs that endow cytotoxic T cells with not yet understood specificities. Second, demonstration that increasing TCR affinities plateau with regard to tumor-regression capacity and ability to induce autoimmunity lite the thoughts that we may not be adapted to, or stand to benefit from, introduction of genetically manipulated high affinity TCRs.

Patient-Specific Antigen Selection

Recent data suggest that patient-specific selection of mutated self-antigens from tumor DNA may be a successful strategy to overcome the treatment barrier of central tolerance. A novel approach allowed identification of mutated gene products from whole exome sequencing of patient-specific tumor. Peptides synthesized from mutated DNA sequences were presented by patient-specific tumor lines to patient derived TILs and response was measured by IFN-g production. Strikingly, in 3 metastatic melanoma patients tested, the top several isolated patient-specific mutated tumor antigens produced a strong cytotoxic response when incubated with patient-specific TILs.⁸⁷ This approach was further bolstered when regression of metastatic cholangiocarcinoma was seen upon treatment with patient-specific TILs specific for endogenous mutated tumor antigen.^88,89

A key consideration in the aforementioned studies is that mutated antigens among patients and between cancer types differed greatly. Thus, patient-specific TIL therapies may hold more impact than high affinity TCRs that are aimed to treat large patient populations.

IV. Enhanced TCR Efficacy

Whether future trials prove low or high affinity TCRs to be effective in cancer regression, the role of TCR signal enhancement through co-receptor stabilization, engagement of co-stimulatory molecules, and diversion of inhibitory signals promise to augment the capacity of the TCR. Here we highlight the current understanding of how accessory molecules strengthen TCR signals.

CD8 Co-receptor

Through direct binding of invariant domains of MHC I, the CD8 co-receptor contributes to extracellular interactions with pMHC that enhances T cell signals during antigen presentation.^90,91 Blockade of CD8 activity through disruption of microdomain organization at the T cell surface led to decreased binding efficiency and association kinetics with pMHCI.⁹² pMHC molecules with impaired CD8 binding capacity were used to assess the kinetics of TCR:pMHC interactions in live cytotoxic T cells. Low affinity TCR:pMHC binding and dissociation were substantially enhanced by the CD8 co-receptor. Less of a contribution was measured when high affinity TCRs were tested.⁹¹ In vivo data assessed anti-tumor responses of CD8 T cells with varied TCR affinities to melanoma antigen gp100. The CD8 co-receptor enhanced anti-tumor activity measured by pERK activation, Ca²⁺ flux, and cytokine release of low affinity TCRs. ⁷³ Together these data highlight role of CD8 for TCR stabilization and enhanced activation for low affinity TCRs.

Tumor Tolerance of TCRs

High affinity CD8 T cells in the tumor microenvironment become tolerized and lose function. Antigen-specific T cells that were infused into melanoma patients upregulated inhibitory receptors and lacked proliferative capacity resultant in trial failure. ⁹³ Murine studies that employ high and low affinity transgenic TCRs to examine antigen-specific tumor regression demonstrate that high affinity TCRs become tolerized in the tumor microenvironment through loss of cytotoxic cytokine production and tumor killing capacity. Low affinity TCRs do not show significant loss of function as compared to high affinity TCRs, yet low affinity tumor-antigen specific TCRs were not capable of significant anti-tumor activity.^94,95 Further, in a T cell transgenic mouse model of prostate cancer, prostate-antigen specific CD8 T cells that infiltrated the tumor microenvironment became tolerized and converted to a suppressive phenotype as measured by TGF-β production.⁹⁶ It is interesting to note that a fate of high affinity TCRs to self-antigen in thymic selection is diversion to the suppressive Treg lineage. Thus, a mechanism of tumor tolerance that remains to be explored is diversion of high affinity TCRs to suppressor cells that aid tumor progression.

Methods are underway to overcome tumor tolerance and re-establish TCR tumor-antigen specific killing. CTLA-4 and PD-1 are negative regulators of CD8 T cell activity and are upregulated on the cell surface in association with functional exhaustion. Blockade of CTLA-4 is an approved therapy to promote reversion of tumor tolerance mechanisms. Unfortunately, CTLA-4 obstruction results in profound autoimmunity. Targeting of PD-1 to enhance blunted T cell function may be more feasible and clinical trials are underway. ⁹⁷ Secondary methods to revert T cell exhaustion in the tumor microenvironment are amplification of co-stimulatory molecules CD28, CD134, CD137.⁹⁸ Insertion into third generation of chimeric antigen receptor constructs has been achieved and clinical trials that employ this technology have positive results and show enhanced T cell persistence and function.⁸⁰ Technologies that have been developed to deliver high affinity TCR genes to T cells will be effective to modulate enhanced co-stimulatory molecule expression as well.

Concluding Remarks

As adoptive T cell therapies progress, basic research that explores signals that mediate thymic selection is necessary. Along these lines, the cooperative signals between mTECs and T cells that determine low, moderate, and high affinity TCRs need to be elucidated. Further, signal requirements downstream of the TCR that pertain to modulation of pERK and Ca²⁺ harbor the potential to amplify T cell activity in TCR-independent manner. Low to moderate affinity TCRs are proving to be as efficacious and safer than high affinity TCRs. Recent reports suggest that patient-specific antigen selection may be key to T cell mediated tumor regression. Future research efforts that focus on amplification of T cell effector capacity of low to moderate affinity TCRs directs toward patient-specific tumor antigens will be critical to progress research and treatment.

Disclosure of Potential Conflicts of Interest

No conflicts of interest were disclosed.

Acknowledgments

We thank Brian Riesenberg, Caroline Wallace, and EJD for thoughtful insight and critical reading of the manuscript. We thank the Li laboratory and M&I Department for continued support.

Jessica Thaxton is funded by Department of Defense BCRP fellowship W81×1H.