Over the past 15 years, cancer therapy has been revolutionized through the development and use of antibodies that block T-cell inhibition via immune checkpoint, such as CTLA-4 and PD-1.
Dario A. Vignali, PhD
Over the past 15 years, cancer therapy has been revolutionized through the development and use of antibodies that block T-cell inhibition via immune checkpoints, such as CTLA-4 and PD-1. CTLA-4 was the first targeted immune checkpoint to lead to enhanced T-cell responses in patients with melanoma.1AntiPD-1 antibodies were subsequently introduced and led to the most success to date.
Patients treated with immune checkpoint inhibitors have experienced a variety of immune-related adverse events (AEs).2Furthermore, just a subset of patients with cancer benefit from these drugs, with most patients developing acquired resistance. As a result, investigators have designed studies to evaluate the efficacy of antiPD-1 antibodies combined with surgery, radiation, chemotherapy, or targeted therapies. Patients treated with both PD-1 and CTLA-4 antibodies have exhibited increased therapeutic efficacy and just moderately increased AEs compared with those seen in patients treated with CTLA-4 blockade alone.3
“PD-1 and CTLA-4 have been the focus of a lot of attention over the past 5 to 10 years or so, but we know that there are many other inhibitory molecules that control immune responses in cancer,” said Dario A. Vignali, PhD, the Frank Dixon Chair in Cancer Immunology, professor and vice chair of Immunology at the University of Pittsburgh in Pennsylvania and coleader of the Cancer Immunology Program at the University of Pittsburgh Medical Center Hillman Cancer Center. “They have been the focus of a lot of interest to ask what’s next in terms of potential targets that might derive some therapeutic benefit, whether on their own or in combination with PD-1 and CTLA-4.”
The benefit of PD-1 blockade may be augmented by blocking other inhibitory receptors. Lymphocyte activation gene 3 (LAG-3), T-cell immunoglobulin and mucin domaincontaining protein 3 (TIM-3), and T-cell immunoglobulin and immunoreceptor tyrosine-based inhibitory motif (ITIM) domains (TIGIT) are promising candidates for boosting T-cell responses alone or in combination with other immune checkpoint inhibitors. “Obviously, we have TIGIT, TIM-3, and others, but becauseLAG-3has been in the clinic for a while, we are likely to get data regarding its efficacy first,” said Vignali.
About 10 therapeutic agents targetLAG-3(CD233), the third checkpoint inhibitor that has been targeted in the clinic.LAG-3was discovered nearly 30 years ago and is, in many ways, a classical inhibitory receptor.
LAG-3is expressed on activated CD4-positive/CD8- positive T cells and can also be expressed on other cell types, including regulatory T cells, natural killer (NK) cells, subpopulations of B cells, and plasmacytoid dendritic cells.
The impact of inhibitory receptors’ expression on regulatory T cells is not fully understood or appreciated as a modulating factor in response to immunotherapy. Both PD-1 andLAG-3have been shown to be highly expressed on regulatory T cells.
“We published a paper in Science Immunology late last year,4showing thatLAG-3can limit the suppressive activity of regulatory T cells,” Vignali said. “In the tumor microenvironment, where regulatory T cells are already superpotent, their presence or absence may not necessarily [affect] tumor growth in an untreated patient, but it may have an impact during immunotherapy. You can imagine that if inhibitory activity was increased as a consequence of blocking inhibitory receptors on regulatory T cells, it could result in an important resistance mechanism that gives rise to certain patients not responding effectively to immunotherapy. This is an aspect of responsiveness to immunotherapy treatment that has not been fully examined.”
Like many inhibitory receptors,LAG-3is dynamically modulated during a normal immune response; however, cells exhibiting an exhausted phenotype in cancer expressLAG-3for longer periods, as seen with PD-1 and CTLA-4. Therefore, the expression of these receptors is often concordant with cells that also express PD-1 in many mouse tumor models and human solid tumors.
The expression of LAG-3 is very sensitive to metalloprotease-mediated shedding. The metalloproteases, a disintegrin and metalloproteinase domain-containing protein (ADAM) 10 and ADAM17, have been shown to help LAG-3 cleave to the cell surface, which can control cell surface expression and any downstream signaling.5Therefore, the expression of these metalloproteases is also relevant for understanding the cell surface expression of LAG-3.
Vignali and colleagues have designed studies to determine the importance of cell surface shedding in modulating the activity of LAG-3. “We have generated a mouse model in which we can selectively express a nonsheddable version of LAG-3,” he said. “We have made mutations that make it resistant to metalloprotease-mediated shedding, and we can selectively control its expression in a cell typerestricted manner. For instance, we can conditionally express a noncleavable LAG-3 just on CD8-positive cells. Its expression would be controlled by the promoter, but when it is expressed, it would be a nonsheddable form rather than the sheddable form.”
This model enables Vignali and colleagues to investigate the impact on biology, the immune response to tumors, and the response to immunotherapy. “It actually results in mice that are profoundly resistant to immunotherapy, even though we have only made this really small change,” Vignali said. “I think that’s quite an interesting aspect of its biology.” He also noted that few studies in human tumor microenvironments have focused on understanding the impact of these changes on the efficacy of therapeutics and nonresponsiveness to antiPD-1 immunotherapy.
The function of LAG-3, TIM-3, and TIGIT tends to be modulated and driven through binding with a ligand. It has been unclear how this dynamic affects LAG-3’s functionality, partially because of the originally described ligand for LAG-3, which is major histocompatibility complex (MHC) class II. CD4- positive cells interact with antigen-presenting cells that express MHC class II, so the T-cell receptor can bind and recognize its ligand.
On the other hand, CD8-positive cells may not encounter MHC class II while engaging their target cell. In this scenario, it is difficult to understand how LAG-3 activity can be mediated in the absence of its ligand, leading to the hypothesis that LAG-3 may be more important for CD4-positive than CD8-positive cells.
However, data have demonstrated thatLAG-3is important for controlling both CD4-positive and CD8-positive T-cell functionality.
It is unclear how LAG-3 functions if its ligand is not there. Experts in the field have hypothesized that additional ligands have yet to be identified.
Although 2 lectins (carbohydrate-binding proteins) have been described as additional ligands to LAG-3, because these proteins are notoriously “sticky,” it is difficult to confirm whether they are true ligands. Vignali and colleagues have made observations that propose another potential ligand, but more work is needed to understand the importance of these additional ligands.
“This has been a complete black box ever since its discovery,” Vignali said, commenting on how LAG-3 works. “We know it’s an inhibitory receptor, so it turns cells off and limits their functionality, but we do not know how.”
Many inhibitory receptors have motifs that provide clues to their functionality. CTLA-4 and PD-1 have tyrosine-based motifs, which are known to bind to and recruit phosphatases, leading to the dephosphorylation of downstream signaling pathways and the limitation of signaling activities. Investigators know that LAG-3 is an inhibitory receptor that may somehow interfere with the T-cell receptor.
Although a few studies have demonstrated thatLAG-3enters the immunological synapse, where, along with many other molecules, it engages with an antigen-presenting cell or a target cell as many molecules do the same, no information has been gleaned regarding which molecules LAG-3may interact with or disrupt.
To that end, Vignali and colleagues conducted high-resolution microscopy and biochemistry studies usingLAG-3mutants.
“We have found that LAG-3 interacts and modulates the functionality of the T-cell receptor directly a by modulating coreceptor function in a unique and novel way,” he said.
“The interesting take-home message is that LAG-3 seems to be more of a signal disruptor than a signaling molecule, which is unique,” Vignali said.
“Its mode of action seems to be very different [from that of] PD-1. There is the general school of thought that optimal combinatorial therapy may be obtained by targeting molecules that have quite distinct modes of action rather than molecules that function in a similar way. That may underlie some of the studies in mouse models and emerging data from the clinic, which [support] mechanistically why combining LAG-3 and PD-1 provides a synergistic benefit.”
Another inhibitory receptor of interest is TIM-3, which is constitutively activated on innate immune cells. Its expression on T cells is associated with activated and terminally differentiated states.7-10Several proposed ligands for TIM-3 include phosphatidylserine, high-mobility group box 1 chromosomal protein, galectin-9, and carcinoembryonic antigen-related cell adhesion molecule 1; however, it is unclear whether all these molecules are bona fide ligands of TIM-3.
Studies have reported that despite the lack of classical activating or inhibitory signaling molecules, the cytoplasmic domain of TIM-3 can mediate intracellular signaling in both T and myeloid cells. T-cell dysfunction and exhaustion have been shown to be prevented through the association ofHLA-Bassociated transcript 3 with the cytoplasmic domain of TIM-3.11TIM-3 has also been shown to promote Akt/mammalian target of rapamycin (mTOR) signaling and to be essential for optimal effector T-cell responses.12
Unlike CTLA-4, PD-1, and LAG-3, which preferentially interact with cell surface molecules expressed on antigen-presenting cells, antigen presenting cell-expressed membrane-bound ligands for TIM-3 have not been reported. In a recent study, TIM-3positive cells in breast cancer samples were found to be of myeloid rather than T-cell origin, suggesting that therapeutic approaches targeting TIM-3 may have a strong effect on macrophages, dendritic cells, and other antigen-presenting cells.13
T-cell reporter systems have been used to assess coinhibition mechanisms and the efficacy of immune checkpoint inhibitors. There are commercially available systems for evaluating antibodies against CTLA-4, PD-1, and LAG-3; however, no validated system for evaluating antibodies against TIM-3 has been reported to date.
Recently, investigators described how an antibody targeting human TIM-3 functioned as an agonist and led to the promotion of CD8 T-cell differentiation via activation of mTOR complex 1.14To understand the role of TIM-3 in human T-cell responses, further studies are needed to elucidate whether functionally active antibodies against TIM-3 may act as agonists or antagonists.
Another coinhibitory pathway implicated in the limitation of T-cell responses and that may be promising for cancer immunotherapy is TIGIT. Like many other inhibitory receptors, TIGIT is expressed on regulatory T cells and affects both developmental and suppressive functions.
TIGIT has also been shown to be expressed on activated NK and T cells. It competes with the activating receptor DNAX accessory molecule-1 (DNAM-1; CD226) for binding with nectin and nectin-like ligands, even though CD155 is its main ligand.
TIGIT is upregulated and coexpressed with PD-1 by CD8-positive tumor-infiltrating lymphocytes, with studies demonstrating that dual PD-1/TIGIT blockade leads to enhanced antitumor CD8-positive T-cell responses and tumor regression.15,16
CD155 is also the main ligand for CD96, which is constitutively expressed on resting NK cells. When TIGIT and CD155 are engaged, human NK-cell toxicity and cytokine production are inhibited by counterbalancing DNAM-1mediated activation, which can be reversed via antibody-mediated blockade of TIGIT.17
In mouse models, the binding of CD96 to CD155 inhibits the production of interferon-alpha by NK cells, which can be promoted via antibody blockade of CD96, alone or in combination with antiCTLA-4 or anti–PD-1, leading to improved tumor control.18Further studies are needed to explain why both TIGIT and CD96 are required to counteract DNAM-1mediated NK-cell activation and whether blockade of TIGIT and/or CD96 modulates NK-cell effector function and clinical responses in patients with cancer.
In a recent study of patients with melanoma, regulatory T cells increased the expression of TIGIT and decreased the expression of CD226, its competing costimulatory receptor, compared with CD4-positive effector T cells.19This eventually led to an increased TIGIT:CD226 ratio, and regulatory T cells failed to upregulate CD226 following T-cell activation. Patients with a high TIGIT:CD226 ratio had increased frequencies of regulatory T cells in tumors and poor clinical outcomes following immune checkpoint blockade. These findings demonstrate that regulatory T-cell suppression may be counteracted in cancer patients who receive TIGIT blockade in combination with novel therapies that activate CD226 in regulatory T cells.
The Future of Immune Checkpoints
Although immune checkpoint inhibitor therapy has been very successful in altering the therapeutic landscape of cancer, treating patients with combinations of these drugs may have superior effects. The promising results from studies on the cotargeting of PD-1 and CTLA-4 and from preclinical models investigating the combination of antiPD-1 antibodies with other immune checkpoint inhibitors have led to the investigation of additional combination strategies for enhancing antitumor responses in cancer patients. Ongoing clinical trials are assessing the combination of anti–PD-1 antibodies with antibodies against LAG-3, TIM-3, and TIGIT, which may result in synergistic effects and improved patient outcomes.