Checkpoint Inhibitors: Applications for Autoimmunity

Abstract

To limit excessive T cell-mediated inflammatory responses, the immune system has a milieu of inhibitory receptors, called immune checkpoints. Cancer cells have evolved to seize those inhibitory pathways and to prevent T cell-mediated killing of tumor cells. Therefore, immune checkpoint inhibitors (ICI) consisting of blocking antibodies against these receptors present an exciting avenue in the fight against cancer. The last decade has seen the implementation of ICI against a variety of cancer indications that have improved the overall anti-tumor responses and patient survival. However, inflammatory toxicities and autoimmunity are a significant adverse event of ICI therapies. In this review, we will discuss the biology of immune checkpoints, highlight research strategies that may help reduce the incidence of immune-related adverse events associated with ICI therapies, and also suggest investigational approaches to manipulate immune checkpoints to treat primary autoimmune disorders.

Introduction

T cell activation is initiated by the recognition of antigens presented by major histocompatibility complex (MHC_Ag) proteins to the T cell receptor (TCR). These antigens can be derived from non-self (i.e., microbial), self-tissues, or tumors. Following T cell activation, the combination of the inflammatory environment and the molecular properties of the antigen lead to epigenetic and metabolic changes associated with T cell differentiation. In both CD4⁺ and CD8⁺ T cells, antigen-specific naïve T cells differentiate into memory T cell subsets with different lineage-specific functions [1]. In order to limit excessive inflammation following antigen clearance and to prevent self-reactive T cell responses (thus avoiding autoimmunity), tight regulation of immune homeostasis is required. One of the mechanisms to achieve that is by inhibitory receptors collectively called immune checkpoints. Immune checkpoints are expressed primarily on T cells and, following recognition of their ligands on antigen-presenting cells (APCs) (Fig. 1), transmit inhibitory signals that prevent excessive cellular responses. Nevertheless, cancers have evolved to usurp these immuno-regulatory mechanisms in order to prevent T cell-mediated recognition and clearance of tumor cells. They achieve this by upregulating a milieu of immune checkpoint ligands so that TCR recognition of tumor-MHC_Ag complexes would prevent T cell anti-tumor responses and limit tumor cell killing. New generation cancer therapies that block immune checkpoints, called immune checkpoint inhibitors (ICI), and augment anti-tumor T cell responses have been successfully implemented for the treatment of variety of human cancers. Currently, the US Food and Drug Administration (FDA) has approved several blocking antibodies to three immune checkpoint targets consisting of cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) and programmed cell death-1 (PD-1) that are expressed on activated T cells, and PD-1 ligand 1 (PD-L1), which is expressed on APC and tumor cells.

Fig. 1.

Immune checkpoints and their ligands. Immune checkpoint receptors expressed in T cells (red) bind to their ligands expressed in APCs, including tumor cells (green). The intracellular portion of immune checkpoint receptors contains different binding motifs (red squares). These motifs recruit effector molecules, such as the tyrosine phosphatases SHP-1 and SHP-2, which dephosphorylate key proteins in the TCR signaling cascade; the T cell inhibitory kinase Csk, which down modulates T cell activation by phosphorylating the inhibitory Y₅₀₅ of Lck; and the adaptor proteins Grb2 and PAG, which recruit other effector molecules that prevent the propagation of TCR signaling

Immune tolerance is a broad term that describes the immune system’s ability to maintain a state of unresponsiveness to harmless antigens such as our body’s microbial flora and our own cells and tissues. As expected, targeting the immuno-regulatory mechanisms that enforce immune tolerance with ICI has been associated with the development of serious immune-related adverse events (irAE), including life-threatening autoimmunity [2•]. Low-grade 1–2 irAE are associated with dermatologic, endocrine, and gastrointestinal toxicities, while grade 3–5 irAE are associated with severe life-threatening neurologic, cardiac, and hepatic toxicities [3]. In addition, ICI therapies may cause irreversible life-long autoimmune diseases. This review will focus on the functional immunobiology of immune checkpoints in T cells in light of the growing concern posed by the manifestation of irAE and autoimmune inflammation in cancer patients who receive ICI therapies. We will attempt to highlight other research areas that should be considered for the development of safer next-generation anti-cancer therapies aimed at reducing the incidence of irAE. The rational for pharmacological interventions of checkpoints in order to treat primary autoimmunity will be discussed.

Biology of Immune Checkpoints

Immune checkpoints can be classified broadly into two groups based on their contribution to immune tolerance assessed by the degree of immune toxicities observed when signaling via these receptors is abrogated by pharmacological (therapeutic modulation) or genetic (gene-knockdown transgenic animal models) approaches [4]. CTLA-4 and PD-1 are key members of the first group which includes immune checkpoints critical to the maintenance of immune tolerance. Anti-cancer therapies consisting of blocking humanized monoclonal antibodies against these critical immune checkpoints have improved the prognosis and survival of patients with several cancers [5]. Despite their efficacy, large proportion of the patients does not respond to these therapies. Moreover, up to 25% of patients undergoing anti-PD-1 treatment and as high as 85% of patients receiving anti-CTLA-4 develop irAE, including tissue-specific autoimmunity [2•]. It is therefore surprising how limited our knowledge is regarding the signaling pathways triggered by immune checkpoints and their T cell subset-intrinsic and tissue site-specific molecular functions. To improve the efficacy of future generation anti-cancer therapies targeting immune checkpoints, it is therefore imperative to broaden our understanding of the signaling partners and molecular pathways that they trigger.

CTLA-4 is the oldest member of the group of immune checkpoints and also the first inhibitory receptor that was targeted for anti-cancer therapy. In conventional T cells (T_CONV), CTLA-4 is largely retained in intracellular vesicles [6], but its expression is constitutively high in regulatory T cells (T_REG) [7]. The mechanisms regulating its surface expression are not well understood, but the cytoplasmic tail of CTLA-4 has been shown to regulate CTLA-4 localization in ganglioside (GM1)-containing lipid rafts within the immunological synapse [8]. CTLA-4 internalization is regulated by the μ2 chain of the clathrin adaptor protein 2 (AP-2) complex that binds to the cytoplasmic tail of CTLA-4 via Y₂₀₁VKM sequence [9]. At the T cell surface, CTLA-4 competes with the receptor CD28 for binding to CD80 or CD86 expressed on APC (including some tumor cells), due to its higher affinity to those ligands compared to CD28 [10, 11]. Following ligation with either CD80 or CD86, CTLA-4 is able to promote their internalization via transendocytosis, which limits their expression on APC and subsequent availability for T cell co-stimulation via CD28 [12••]. In addition, CTLA-4 has been shown to control T cell adhesion and to promote T cell motility [13, 14], thereby interfering with T cell-APC interactions, in possibly T cell subset-specific manner [15, 16]. CTLA-4 ligation has been shown to inhibit T cell activation by down modulating NF-kB, NFAT, and AP-1 as well as IL-2 production in TCONV [17]. While the inhibitory function of CTLA-4 has been well recognized, the scientific community has not reached a consensus regarding the signaling pathways triggered by CTLA-4. Earlier studies in T_CONV have suggested that CTLA-4 ligation recruits the tyrosine phosphatase Src homology 2 (SH2) domain-containing protein tyrosine phosphatase (SHP-2) [18] and the serine/threonine protein phosphatase 2 (PP2A) [19] to its cytoplasmic domain, which mediate the de-phosphorylation of key signaling proteins downstream of the TCR signaling cascade. Follow-up reports could not demonstrate a direct interaction between either SHP-1 or SHP-2 [20•] and questioned the role of PP2A as mediator of CTLA-4 inhibitory function since a mutant lacking the PP2A binding site showed an enhanced inhibitory capacity [21]. A recent report by Kong et al. demonstrated that in T_REG, the cytoplasmic tail of CTLA-4 associates with the protein kinase C isoform PKCν and promoted cellular motility by recruiting the GIT2-aPIX-PAK signaling complex [22]. Collectively, the elusive nature of CTLA-4 signaling under-scores the importance to investigate the cell subset- and tissue site-specific signaling pathways that maybe engaged by CTLA-4 in order to carry out its inhibitory functions.

Similarly, to CTLA-4, PD-1 has been well recognized as a critical immune checkpoint and anti-cancer therapies designed to block PD-1 or its ligand PD-L1 have shown significant success for a variety of cancer indications. PD-1 is highly expressed on the surface of activated central (T_CM), effector memory (T_EM), and follicular helper T cells (T_FH), and low to moderately expressed in T_REG and terminally differentiated effector T cells (T_EMRA) [23]. The cytoplasmic tail of PD-1 contains an immunoreceptor tyrosine-based switch motif (ITSM) and an immunoreceptor tyrosine-based inhibitory motif (ITIM), which in human PD-1 are phosphorylated at Y₂₄₈ and Y₂₂₃, respectively, following PD-1 ligation with either PD-L1 or the higher-affinity ligand PD-L2. Tyrosine phosphorylation of the ITSM and ITIM recruits the tyrosine phosphatase SHP-2, which in T_CONV dephosphorylates key members of the TCR signaling complex [24•]. Wei et al. have demonstrated that the extent of T cell inhibition by PD-1 is proportional to the density of surface PD-1 expression. The authors show that very low levels of PD-1 are sufficient to prevent TNFα and IL-2 production as well as T cell proliferation, and medium low levels of PD-1 inhibit IFNγ production and cytotoxicity while high PD-1 surface expression is required to inhibit macrophage inflammatory protein 1 beta production [25]. Consequently, Jiang et al. recently revealed that self-reactive T cells calibrate their surface PD-1 expression based on the affinity of their TCR for self-antigen, with higher-affinity T cell clones expressing the highest level of PD-1, thus enforcing peripheral tolerance [26]. In addition to its role as a rheostat of T cell responses [27], PD-1 also inhibits glycolysis, which is the main metabolic process utilized by T cells during differentiation, and promotes fatty acid oxidation, thus preventing T cell effector development [28]. In addition, PD-1 inhibits T cells proliferation by arresting T cell cycle progression through the G1 phase by inhibiting Akt and Ras signaling pathways [29]. Finally, PD-1 induces T cell exhaustion by upregulating basic leucine transcription factor, ATF-like (BATF) to inhibit T cell functions and promotes exhaustion by inhibiting the transcription factor AP-1 [30]. As a result, PD-1 is considered a canonical exhaustion marker that is constitutively expressed following prolonged antigen exposure and accordingly blocking PD-1 function in effector CD8⁺ T cells restores their cytotoxic function [30]. Nevertheless, studies in T_FH and T_REG have challenged this paradigm since both subsets express surface PD-1 and in the case of T_REG also high CTLA-4, but their functions are not inhibited. It is yet unclear whether other signaling partners or functional downstream pathways may underlie these differences in T_FH and T_REG cells, and whether they are differentially mediated by PD-L1 or PD-L2.

Unlike CTLA-4 and PD-1, immune checkpoints that fall under the second category of inhibitory receptors (as mentioned earlier) mainly contribute to the maintenance of immune tolerance. These include lymphocyte activation gene-3 (LAG-3), T cell immunoglobulin-3 (TIM-3), T cell immunoglobulin and ITIM domain (TIGIT), V-domain Ig suppressor of T cell activation (VISTA), leukocyte-associated Ig-like receptor (LAIR-1), and B- and T-lymphocyte associated (BTLA) (Fig. 1 and 2). While clinical trials designed to explore the anti-cancer efficacy of those immune checkpoints alone or in combination therapies are underway, time is required to collect and process the data before we can evaluate their efficacy.

Fig. 2.

Timeline of immune checkpoint discovery. The figure shows a timeline of inhibitory receptor discovery. The stacked bars correspond to the number of PubMed published studies in each year. *Number of publications until July 6, 2017

LAG-3 belongs to the immunoglobulin superfamily and is structurally similar to CD4. It is expressed on the surface of activated T_CONV and T_REG. It engages MHC class II with a higher affinity than CD4 and inhibits the proliferation and cytokine secretion from CD4⁺ and CD8⁺ T cells, and NK cells. Given that CD8⁺ T cells and NK do not recognize MHC class II, the possibility of a second LAG-3 ligand has been actively explored. A study by Xu et al. has suggested that LSECtin, a DC-SIGN family member, which is expressed in the liver and in many tumor cells, may be the second ligand for LAG-3 [31]. Although the signaling pathways and the intracellular partners of LAG-3 have not been identified, it has been demonstrated that K₄₆₈ of the KIEELE motif in its intracellular domain is critical for its inhibitory function in CD4⁺ T cells [32].

The TIM family gene locus in humans contains three members, namely TIM-1, TIM-3, and TIM-4. Of those, TIM-3 has received much attention as an inhibitory receptor that is crucial for the negative regulation of IFNγ producing CD4⁺ T helper 1 cells (T_H1) as well as cytotoxic CD8⁺ T cells. TIM-3 is expressed on the surface of activated T_H1 cells, cytotoxic CD8⁺ T cells, T_REG, and some innate immune cells. Galectin-9 [33], a C-type lectin, and Ceacam-1 [34], which is also expressed by T cells, have been identified as the two ligands that bind to TIM-3. The cytoplasmic tail of TIM-3 contains two conserved tyrosine residues (Y₂₅₆ and Y₂₆₃), which provide a docking site for human leucocyte antigen B-associated transcript 3 (Bat3) and the Src family tyrosine kinase Fyn, which transmit activating and inhibitory signals, respectively. When TIM-3 is not ligated, Y₂₅₆ and Y₂₆₃ are not phosphorylated and serve as the binding site for Bat3. This obstructs the SH2 domain-binding sites within the TIM-3 cytoplasmic tail and recruits the catalytically active form of Lck, thereby enhancing TCR signaling [35••]. On the contrary, following ligation by either ligand, Y₂₅₆ and Y₂₆₃ become phosphorylated, which releases the bound Bat3, and the SH2 domain-binding sites of TIM-3 become available for binding by SH2 domain-containing Src kinases such as Fyn. Fyn has been shown to phosphorylate another transmembrane protein, phosphoprotein associated with glycosphingolipid microdomains (PAG) [36]. PAG is an adaptor protein that recruits the Src kinase Csk to the proximity of the TCR signalosome, which subsequently phosphorylates Lck at its inhibitory tyrosine (Y₅₀₅) and leads to the suppression of the TCR signaling cascade. Nevertheless, it is yet unclear how the balance between Bat3 and Fyn, as well as the two TIM-3 ligands, affects TIM-3 signaling, function, and contribution to overall effector T cell responses.

TIGIT is also expressed on activated T cells and has a T_REG-intrinsic function. Similarly, to CTLA-4, TIGIT competes with the co-stimulatory receptor CD226 (PTA1) for binding to CD155 (PVR). In addition, CD112 (PVRR2) was also shown to bind TIGIT, albeit with a much lower affinity [37]. Following TIGIT ligation in NK cells, Lck and Fyn phosphorylate its ITIM motif (Y₂₃₁) and immunoglobulin tail tyrosine (ITT)-like motif (Y₂₂₅), which subsequently recruits the SH2-containing inositol 5′-phosphatase (SHIP-1) via the adaptor protein Grb2 to abrogate PI3K and MAPK downstream signaling pathways [38•]. The phosphorylated ITT-like motif also binds to β-arrestin 2 and interferes with NFκβ signaling also through SHIP-1 recruitment [39]. In T cells, microarray analysis of TIGIT stimulation with agonistic antibody revealed that it down modulates key TCR signaling pathways [40]. Nevertheless, despite its clear role as an inhibitory receptor, TIGIT upregulates the anti-apoptotic molecule Bcl-xL as well as the expression of the receptors for IL-2, IL-7, and IL-15, which are all critical for T cell proliferation and maintenance of T cell homeostasis [40].

VISTA, also known as programmed death-1 homolog (PD-1H), is the youngest member of the immune checkpoint receptor family [41••, 42]. Mouse data suggests that it is exclusively expressed in hematopoietic cells and within this compartment myeloid cells have the highest expression [43]. In humans, surface VISTA expression is similar between CD4⁺ and CD8⁺ T cells. VISTA contains one SH2 domain and three SH3 domains as well as a putative protein kinase C binding site [43]. While the signaling pathways regulated by VISTA are still unknown, its inhibitory function in T cell activation has been well documented with both agonist anti-VISTA antibodies as well as VISTA-Ig fusion protein [41••, 44]. These observations suggest that VISTA may serve as a ligand as well as a receptor in T cells and APC, but its respective binding partner has not been identified yet.

LAIR-1 is a protein that is expressed by the majority of human immune cells as a surface receptor or a soluble protein [45] and is able to inhibit human T cells by binding with high affinity to collagens. LAIR-1 has two ITIM-like domains in its cytoplasmic tail and recruits SHP-1, Csk, and possibly SHP-2 [46] to mediate its inhibitory functions. Similarly to PD-1, the level of LAIR-1 inhibition in T cells is proportional to the level of its surface expression. Interestingly, naïve B and T cells express the highest levels of this receptor, suggesting a direct regulatory role for LAIR-1 in naïve lymphocytes [47]. However, the signaling pathways triggered by LAIR-1 in different immune cell subsets are poorly understood and need yet to be investigated.

Many immune cells also express BTLA and its level on naïve T cells is low but increases following T cell activation [48]. Unlike other inhibitory receptors, BTLA is not expressed on T_REG but it is highly expressed on exhausted T cells following prolonged antigen stimulation [48]. The cytoplasmic portion of BTLA contains one membrane proximal tyrosine within an YDND motif, which is a Grb2 binding site, and two membrane distal tyrosines within one ITIM and one ITSM motif. Following interaction with its ligand, herpes virus entry mediator (HVEM), tyrosine phosphorylation of both ITIM and ITSM is required to further employ the tyrosine phosphatases SHP-1 and SHP-2 [49]. Interestingly, Grb2 binding has been associated with the recruitment of PI3K and the activation of the PI3K-Akt pathway and overall enhanced cell survival [50]. Underlying the cell-intrinsic and tissues site-specific properties of immune checkpoints, previous study has demonstrated that BTLA is important for maintaining immune tolerance at mucosal sites [51] and prevented invariant natural killer T cell (iNKT)-mediated liver injury in a Concanavalin A model of hepatitis [52]. Nevertheless, the downstream pathways affected by BTLA-HVEM ligation are not yet fully known, and given the dual role of this immune checkpoint in promoting cell survival while eliciting T cell anergy and its potential application as an anti-cancer target, it is critical to better understand its downstream signaling partners.

Immune Checkpoints in Autoimmunity

The main function of immune checkpoints is to prevent excessive T cell inflammatory responses and organ-specific autoimmunity. Therefore, it is not surprising that genetic mutations within coding and non-coding regions that affect the expression or function of immune checkpoints (or their ligands) predispose to clinical autoimmunity. Furthermore, while antibody blockade of immune checkpoints augments anti-tumor immunity, it also exacerbates organ-specific autoimmunity in genetically susceptible animal models as well as in cancer patients receiving immune checkpoint blockade therapies. This evidence suggests that enhancing the inhibitory function of immune checkpoints by their ligands should be beneficial in different autoimmune diseases, while their therapeutic blockade (with monoclonal antibodies) has proven advantageous in augmenting anti-viral and anti-tumor T cell responses. Given the different degree of immune checkpoint contribution to immune tolerance and their cell-intrinsic functions, it is important to gain better insight into the molecular pathways triggered by immune checkpoints and their cell- and tissue-specific contribution to immune regulation.

CTLA-4 is a master immune checkpoint, and germ line CTLA-4 knockout mice develop fatal lymphoproliferative disorder within the first month of life [53, 54]. Conditional CTLA-4 deletion in FoxP3⁺ T_REG alone is sufficient to cause systemic lymphoproliferation and organ-specific autoimmunity [55]. Similarly, patients with heterozygous loss-of-function mutations in CTLA-4 or mutations within the LRBA gene, which affects CTLA-4 trafficking, develop overt lymphoproliferative disorder in multiple non-lymphoid organs parallel with lymphadenopathy, splenomegaly, autoantibody-mediated cytopenias, and organ-specific autoimmunity [56, 57]. These observations underlie the critical role of CTLA-4 in central and peripheral tolerance early in life.

Earlier studies evaluating the inhibitory function of immune checkpoints have used transgenic knockout mouse models that precipitate the receptor deletion from birth. It is now clear that the expression and function of immune checkpoints is tissue- and cell-type specific and may have different functional consequences once immune tolerance has been established. All of these factors need to be considered for the design of therapies targeting these receptors. To investigate the consequences of T_REG-specific CTLA-4 ablation in adult-hood, Paterson et al. generated a mouse model of conditional CTLA-4 deletion by crossing floxed CTLA-4^fl/fl with tamoxifen-inducible Cre recombinase expressing strain [58••]. Interestingly, tamoxifen administration and conditional CTLA-4 deletion in adult mice did not result in systemic organ-specific autoimmunity or in enhanced anti-tumor immunity. In fact, CTLA-4 deletion resulted in an increase of activated T_CONV and T_REG cells. CTLA-4-deficient T_REG cells were functionally suppressive and used compensatory CTLA-4-independent mechanisms to inhibit autoimmune inflammation by producing anti-inflammatory cytokines and an increase in their surface expression of PD-1 and LAG-3. An earlier study also showed increased ex vivo proliferation and normal suppressive functions in T_REG from cancer patients undergoing anti-CTLA-4 treatment [59].

Similar study that evaluated the consequences of LAG-3 ablation in T_REG demonstrated that LAG-3 limited T_REG proliferation but did not have an effect on their suppressive properties and as a result, increased LAG-3 expression in T_REG contributed to the exacerbation of autoimmune diabetes [60]. Nevertheless, constitutive LAG-3 deletion from T cells accelerated the development of autoimmune diabetes in NOD mice demonstrating its contribution to the maintenance of peripheral tolerance. Two additional studies also demonstrated that while PD-1 was not required for T_REG suppressive function, PD-1 deletion in T_REG resulted in enhanced T_REG proliferation [61, 62] and suppressive function [61].

Taken together, these studies suggest that immune checkpoints may exert profound regulatory control on T_REG homeostasis that restricts their expansion and activation. As a consequence, non-depleting antibody blockade of CTLA-4, PD-1, and LAG-3 may favor the expansion of suppressive T_REG cells that will limit the re-invigorated anti-tumor-specific T_CONV [63, 64], which is highly undesirable in anti-cancer treatments. While universal antibody blockade of these receptors in autoimmune diseases is unfavorable, targeting other molecules within T_REG-intrinsic pathways regulating the expression and signaling via these receptors may reveal new therapeutic avenues aimed at restoring immune tolerance in autoimmunity.

Current therapeutic agents directed against PD-1 or CTLA-4 rely on non-depleting antibodies that block the signaling via these receptors in all T cells. As a result, the beneficial effect of cytotoxic anti-tumor responses maybe counteracted by enhanced suppressive T_REG function as discussed above as well as inhibition of the immune tolerance-inducing properties of PD-1 and CTLA-4 in auto-reactive T_CONV. In fact, it is be-coming clear that immune checkpoints have intrinsic cell subset- and tissue-specific properties that maybe exploited to improve the efficacy of both anti-cancer therapies and treatments aimed at promoting their inhibitory functions in autoimmune diseases.

Checkpoint Inhibitors and Immune-Related Adverse Events

Given that blocking anti-CTLA-4 and anti-PD-1 antibodies have now been implemented as anti-cancer treatments, the majority of studies investigating the toxicities associated with ICI in humans come from clinical trials utilizing these drugs. Autoimmunity related to ICI is primarily due to abrogated immune tolerance mediated via these receptors in self-reactive T_CONV. It is well established that negative selection prunes but does not eliminate self-reactive T cells, and self-reactive T cells can be detected in the circulation of healthy people [65•, 66]. Peripheral T_REG [67] as well as surface expression of immune checkpoints prevents the activation and clonal expansion of these self-reactive T cells [26, 66]. As may be expected, this control of immune homeostasis is lost in a significant proportion of cancer patients receiving anti-CTLA-4 and to a lesser extent anti-PD-1 therapies [68]. As a result, autoimmunity is a significant irAE associated with ICI anti-cancer treatments. In fact, recent case reports of cancer patients who developed type I diabetes following anti-PD-1 antibody treatment confirmed the presence of islet-specific CD8⁺ T cells in the peripheral blood of those patients [69, 70]. Similarly, increased infiltration of CD8⁺ T cells and reduced FoxP3⁺ T cells were reported following anti-PD-1 therapy in a patient who developed lymphocytic myocarditis [71].

The technological advent of whole genome sequencing technologies has allowed us to carry out a plethora of genome-wide association studies, which have investigated the relationship between genetic polymorphisms within the genes encoding immune checkpoints and tissue site-specific inflammation. These data are a useful guideline to the undesirable adverse events that maybe observed following ICI therapies, especially in light of new treatment regiments combining the blockade of two or more immune checkpoints. As mentioned earlier, polymorphisms in the genes encoding CTLA-4 and PD-1 have been strongly associated with different autoimmune diseases [72, 73]. Studies have also demonstrated that mutations within CTLA-4 and TIM-3 are associated with increased susceptibility to asthma and allergy, respectively [74, 75], although more research is required to confirm those observations. Separately, each immune checkpoint may have differential contribution to the immune homeostasis in different tissue sites. Consequently, for the implementation and therapeutic use of current and new generation ICI, it will be imperative to consider their cell subset- and tissue site-specific functions.

Conclusions

Blocking antibody therapies against immune checkpoints have demonstrated durable clinical responses for a variety of cancers. Nevertheless, a large proportion of these patients does not respond to treatment or develops resistance to those therapies [76, 77, 78••, 79]. Additionally, ICI therapies result in a variety of irAE and in some cases life-long autoimmunity. To overcome those challenges, we need to gain a better understanding of the molecular pathways underlying immune checkpoint signaling that will allow us not only to design better anti-cancer therapies but also to utilize this knowledge in the design of suitable therapies targeting immune checkpoints in autoimmunity.

Few underexplored avenues associated with immune checkpoints are critical to better understand their molecular signaling networks. First, immune checkpoint receptors have different T cell subset-specific distribution [23]. This likely reflects their overall contribution to cell subset-specific functions. Second, the data is scarce regarding the spatiotemporal distribution of immune checkpoint ligands in different tissues, including different tumor types, as well as the signaling pathways regulating their expression, both of which are crucial for the design and implementation of next-generation therapies targeting immune checkpoints or their associated signaling pathways. This will reveal the phenotypic immune checkpoint ligand landscape of different tumors that may be a useful surrogate to be considered for the efficacy of ICI therapies. In addition, the temporal regulation of immune checkpoints and their ligands expression during the course of the inflammatory response will be informative for the design of combined treatment regimens. The tissue-specific contribution of immune checkpoints may also point to the irAE of such therapies. Third, given the critical role of immune checkpoints in the prevention of autoimmunity, it is important to translate our current observations that we have collected from ICI therapies in cancer to understand the molecular mechanisms of autoimmunity as well as for the design of therapeutic strategies. Agonistic manipulation of immune checkpoint pathways or signaling molecules may be explored for different autoimmune diseases, especially in the context of tissue-specific contribution of immune checkpoints to the maintenance of tissue tolerance. Finally, in a joint effort, cancer biologists and immunologists need to combine knowledge in order to extract mechanistic information that will serve as a springboard for the design of new treatments.