Lymphocyte-activation gene 3 (LAG-3) is a coreceptor found on activated T-lymphocytes activated B-lymphocytes and natural killer (NK) cells. It is closely related to CD4 where it shares multiple common and divergent features. It contains specific binding sites with high affinity to major histocompatibility complex (MHC) Class II and functions as an inhibitor of T-cell signalling. Tumour-infiltrating lymphocytes with high LAG-3 expression have been found in many solid tumours including ovarian cancer, melanoma, colorectal cancer and haematological malignancies including Hodgkin and diffuse large B-cell lymphoma. LAG-3 antagonism has been demonstrated to restore the anti-tumourigenic function of T-cells in vivo, however, mechanistic knowledge remains relatively poorly defined. As other immune checkpoint inhibitors have transformed the management of difficult to treat cancers, such as melanoma, it is hoped that LAG-3 might have the same potential. This review will explore LAG-3 modulation as an anticancer therapy, highlighting recent clinical developments.
- medical oncology
- regional perfusion
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Lymphocyte-activation gene 3 (LAG-3; also known as CD223) was first identified over 30 years ago.1 Although broadly expressed and selectively transcribed in activated natural killer (NK) and T-lymphocytes, it was recognised as encoding a protein that had key features linking it to CD4, and therefore was thought to play a role in regulating the immune response. LAG-3 joins a class of genes, such as programmed cell death protein 1 (PD-1) and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), which are also receptors regulating immune checkpoints and involved in the recognition of novel antigens expressed in disease, as self or otherwise.
Following the recent successes in monoclonal antibodies (mAbs) targeting these immune checkpoints to initiate or abrogate the immune response (eg, ipilimumab, as well as the older abatacept), there is a large amount of interest in developing novel immunotherapies that act on the body’s own response to antigens expressed in diseases, including cancer.
The LAG-3 gene spans ~6.6 kb and includes eight exons. Interestingly, its chromosomal location is adjacent to the gene for CD4 on the distal part of the short arm of chromosome 12. The LAG-3 gene encodes a type I transmembrane protein, which is made up of 498 amino acids, and shares significant structural homology with CD4, although it shares less than 20% similarity at the primary amino acid sequence. In addition, the LAG-3 protein can be grouped within the immunoglobulin superfamily containing four extracellular Ig regions, with one variable (V type) and three constant (C type) Ig-like domains (see figure 1).
It has been noted that there are similarities in the exon/intron organisation, as well as sharing internal sequence homologies between LAG-3 and the CD4 coreceptor between domains 1 and 2, and between domains 2 and 4, indicating that LAG-3 may have arisen from a common ancestor. However, it is the differences between the two genes which are important and may give rise to important functional variances between their mechanisms.
Of particular interest is the membrane distal D1 domain which houses an extra loop comprising 30 amino acids (see figure 1) and encoded by exon 3 in a region that in CD4 seems to be important for interactions with class II major histocompatibility complex (MHC) molecules and HIV gp120.2 This proline-rich loop is constitutively expressed on LAG-3 positive lymphocytes and allows it to bind with higher affinity to MHC Class II than CD4 alone.3
Comparison of the intracellular domain with CD4 also shows an absence of a binding site for the tyrosine kinase p56Lck, which facilitates signal transduction downstream of the T-cell receptor. Instead, definable motifs within LAG-3 have been found conserved in many other mammalian species, which indicates some importance in the cytoplasmic domain for its function.4 Three definable motifs have been specifically identified including a potential serine phosphorylation site (S454), a conserved ‘KIEELE’ motif with no homology to any other known protein and an unusual glutamic acid-proline repetitive sequence. Mutant cell work using a murine T-cell hybridoma line shows that a single lysine residue (K468) within the ‘KIEELE’ motif appears to be essential for LAG-3 activity.5
LAG-3 receptors have been found expressed on activated CD4 +and CD8+T cells (including T regulatory cells), and are also present on the surface of NK cells and invariant NK T-cells, activated B lymphocytes6 and plasmacytoid dendritic cells.7 LAG-3, CTLA-4 and PD-1 function as negative regulators of T-cell expansion and homoeostasis, as shown using in vivo inhibition of these proteins by antibody blockade. Similarly, these proteins show a mixed cellular distribution, predominantly retained in intracellular compartments, with nearly half the cellular content of LAG-3 in close proximity to microtubules and recycling endosomes which facilitates rapid translocation after cell activation.8
LAG-3 functions in T-cells specifically via association with the CD3/T-cell receptor (TCR) complex, resulting in decreased T-cell proliferation and cytokine synthesis.9 This downregulation, rather than termination through apoptosis, may allow the immune system to generate memory responses. This is induced by MHC class II binding in both CD4 +and CD8+T cell subsets.10
Interestingly, antibodies that do not block LAG-3 to MHC-II binding can nonetheless promote T-cell functions, and this supports the evidence that there remain other ligands for LAG-3 that also contribute to immune regulation. This includes fibrinogen-like protein 1,11 which is upregulated in human solid tumours (normally expressed in the liver and pancreas) and galectin-3,11 which has been identified as modulating CD4 +and CD8+T cell response.12 LAG-3 expression on conventional T effector cells is controlled by cleavage by two transmembrane metalloproteases, ADAM10 and ADAM17. The resultant cleavage product, soluble LAG-3 (sLAG-3) has a short half-life and does not seem to have any global function.13 It is possible that selective agonists of LAG-3 cleavage could be another potential candidate for therapeutic intervention.14
LAG-3 expression in malignancy
The immune system plays an important role in removing abnormal and malignant cells.15 Upregulated expression of inhibitory receptors are pivotal to balance costimulatory receptor activity and limit T-cell activation, preventing autoimmunity and tissue damage.16 Malignant cells may hijack immune checkpoint mechanisms to protect against antitumouricidal responses elicited by CD4 +and CD8+T cells16. Expression of immune checkpoints, such as PD-1 and LAG-3 correlate with intratumoural T-cell dysfunction in patients.4 17 Furthermore, tumours may recruit regulatory T-cells (Tregs) to further reduce immune activation.18 19
Expression of LAG-3 has been identified predominantly on infiltrating immune cells by immunohistochemistry or next generation sequencing in a range of different tumour types.20 High expression of LAG-3 can be found in tumour infiltrating T-cells in ovarian cancer, Hodgkin Lymphoma, melanoma, non-small cell lung cancer (NSCLC), diffuse large B-cell lymphoma, follicular lymphoma, head and neck squamous cell carcinoma (HNSCC), correlating alongside other immune checkpoint receptors with an aggressive tumour phenotype and overall poor prognosis.21–33 It also plays a role in chronic viral infection such as hepatitis B virus and therefore indirectly in viral-induced cancers, such as hepatocellular carcinoma.31 34 However, paradoxically, studies in breast cancer, gastric cancer and oesophageal squamous cell carcinoma have shown opposite results, suggesting further studies investigating the use of LAG-3 as a putative cancer biomarker are required.35–37
In these circumstances, LAG-3 expression can be described as a marker of T-cell exhaustion, a process characterised by progressive loss of function and sustained inhibitory receptor expression which commonly develops under conditions of antigen persistence. These also include expression of T-cell immunoglobulin and mucin domain-containing protein 3 (TIM-3) and T-cell immunoreceptor with Ig and ITIM domains, which are also being considered as targets for immune checkpoint inhibitors (ICPIs) in their own right. T-cell exhaustion was first identified in chronic viral infection and more recently in tumourigenesis characterised by dysfunctional CD8 +T cells, alterations in CD4 +T cell cytokine secretion and inhibition of Treg expression.38 39 However, there are important differences in antigen exposure and metabolic constraints between chronic infection and the processes involved in the tumour microenvironment, meaning that the processes underlying each disease require specific investigation.40 What remains clear is that this hyporesponsiveness can be transiently restored by PD-1 blockade, and there may be potential synergistic effects in combination with LAG-3 immunotherapy to allow immune cells to exert antitumour effects.
LAG-3 contributes to ‘immune escape’ in carcinogenesis, mirroring observations seen with other ICPIs, such as anti-PD-L1 and anti-CTLA-4.4 16 Furthermore, upregulation is likely to contribute to the acquired resistance seen in patients treated with ICPIs who initially respond.17 18 Preclinical studies have demonstrated that blockade of LAG-3 strongly supports anticancer immune responses, significantly reducing tumour growth in several murine models compared with controls.26 41 These findings suggest that antagonism of LAG-3 may be a promising therapeutic target for novel cancer immunotherapy development.
Notably, LAG-3 blockade appears to more effective when combined with other anti-cancer therapies, and particular synergy has been noted in combination with other ICPIs. Concomitant blockade or genetic deletion of both PD-1 and LAG-3 has demonstrated increased anticancer efficacy preclinically, providing further validation for clinical investigation.42
Clinical modulation of LAG-3
The success of other ICPIs and the rationale of targeting LAG-3 has fuelled the development of therapies to modulate this pathway, despite lack of functional understanding.43 LAG-3 modulation is primarily focused on exploring anticancer efficacy, either as a monotherapy or in combination with other ICPIs, cytotoxic chemotherapy or other novel agents.20 33 Most of the combination studies combine LAG-3 antagonism with an anti-PD-L1 or anti-PD-1 drug. In addition, LAG-3 modulation is being explored in autoimmune conditions, with efficacy demonstrated in diseases such as ulcerative colitis and psoriasis.44
Currently, over 100 clinical trials are ongoing (at date of censoring) with at least 15 novel molecular and biological agents in clinical development (see table 1). The majority (57%) are antagonistic mAbs against LAG-3, but other agents include LAG-3 depleting antibodies, sLAG-3 fusion protein and bispecific mAbs (BMAs). Despite the wealth of favourable pre-clinical data demonstrating anticancer efficacy, as monotherapy or in combination, only a limited amount of this work has been recapitulated in the clinical setting thus far.
Antagonistic LAG-3 mAbs block the interaction between MHC class II molecules and LAG-3, attenuating downstream signalling.33 The first LAG-3 mAb to enter clinical evaluation was relatimab (BMS-986016). A phase 1/2 study (NCT01968109) to assess the tolerability of relatimab with nivolumab (anti-PD-1) in patients with advanced melanoma with progression on prior anti-PD-1/PD-L1 therapy showed an overall response rate (ORR) of 11.5%.45 Subanalysis has demonstrated the ORR in patients with LAG-3 expression of >1% on tumour infiltrating lymphocytes (TILs) was over three times greater (18% ORR) than patients with LAG-3 negative TILs (5% ORR), irrespective of PD-L1 status. A further phase 2 study (NCT03470922) of relatimab with nivolumab in treatment-naïve melanoma patients showed a significantly longer progression-free survival (10.1 months) compared with nivolumab monotherapy (4.6 months).46 However, an increase in treatment-related adverse events according to the Common Terminology Criteria for Adverse Events grades 3 and 4 was seen in the combination (18.4%) versus nivolumab monotherapy (9.7%).
An alternative LAG-3 mAb, Ieramilimab (LAG-525) has been investigated in a phase 1/2 dose-escalation study (NCT02460224) in combination with the investigational anti-PD-1 inhibitor spartalizumab (PDR001). Durable responses were reported in 9.9% (10/121) of patients with a variety of tumours, including mesothelioma and triple negative breast cancer.47 Biopsies from the patients with breast cancer demonstrated an overall trend in conversion of immune-cold towards an immune-activated profile, suggesting LAG-3 blockade may modulate the tumour microenvironment favourably.
LAG-3 fusion proteins
Eftilagimod alpha (IMP321) a soluble recombinant LAG-3 fusion protein, comprising the extracellular region of LAG-3 and the Fc portion of human IgG.20 It binds to a subset of MHC class II molecules, mediates antigen-presenting cell activation, followed by CD8 +T cell activation. Several clinical trials have investigated efficacy with variable results.
In melanoma, a phase 1 study (NCT02676869) of eftilagimod alpha in combination with pembrolizumab (anti-PD-1) demonstrated an ORR of 33% and 50% in pembrolizumab refractory and anti-PD-1 naïve subgroups, respectively.48 Positive results have also been reported in NSCLC and HNSCC, both in combination with pembrolizumab.49 50 However, in hormone positive breast cancer a phase 2 trial (NCT02614833) did not prolong overall progression-free survival when used as an adjunct to the cytotoxic chemotherapy drug paclitaxel.51
Several first-in-class BMAs targeting LAG-3 and other immune targets are being investigated clinically. BMAs target two different antigens simultaneously and are effective alternatives to combining two or more different therapies. Due to the marked therapeutic synergy demonstrated when combining LAG-3 and other ICPIs, the development of BMAs is an attractive therapeutic strategy.
Preclinical studies have demonstrated similar efficacy to combination therapy.52 Current agents in development include FS-118 (targeting LAG-3/PD-L1), R07247669 (targeting LAG-3/PD-L1), tebitelimab/MGD013 (targeting LAG-3/PD-1) and XmAb22841 (targeting LAG-3/CTLA-4). Several other agents are in preclinical development.33
LAG-3 is a protein which shares particular similarities with the CD4 coreceptor in structure but is present only on activated T-cells and other lymphocytes. As a marker of immune exhaustion, downregulation of LAG-3 as an anticancer therapy is progressing rapidly through clinical development and holds potential to become a third class of ICPIs. However, important unanswered questions remain relating to identification of important downstream signalling mechanisms. Deeper mechanistic understanding of the LAG-3 modulation pathway may lead to more efficacious therapies and the identification of biomarkers that would facilitate the stratification of patients responsive to LAG-3 blockade.
Take home messages
Lymphocyte-activation gene 3 (LAG-3) is a protein member of the Ig superfamily, closely related to CD4 and is a co-receptor for MHC class II.
It is found expressed on activated CD4 +and CD8+T cells, NK cells and invariant NK T-cells, activated B lymphocytes and plasmacytoid dendritic cells.
Although its mechanism is yet to be fully elucidated, a proline-rich loop present on domain 1 is necessary for binding with high affinity to MHC class II.
Blockade of LAG-3 strongly supports anticancer immune responses with over 100 clinical trials in clinical development.
While the majority are classical monoclonal antibodies against LAG-3, other novel agents include LAG-3 depleting antibodies, soluble LAG-3 fusion proteins and bispecific monoclonal antibodies.
Figures produced with BioRender.
Handling editor Runjan Chetty.
MPL and DSKL contributed equally.
Contributors Conception or design: MPL and DSKL. Acquisition, analysis or interpretation of data: MPL and DSKL. Drafting the work or revising it critically for important intellectual content: NEA, AEF and JK. Final approval of the version to be published: MPL, DSKL, NEA, JK and AEF.
Funding This study was funded by Royal College of Surgeons of England.
Competing interests JK has received honoraria from Clovis Oncology, Tesaro and AstraZeneca, outside the submitted work. None of the aforementioned companies had any input into the study design, data collection, reporting or preparation of the manuscript.
Provenance and peer review Commissioned; internally peer reviewed.