Clear cell renal cell carcinoma (KIRC) represents 80% of kidney malignancies and is among the most common malignancies worldwide.1 Although overall 5-year survival rates are approximately 75%, treatment of advanced and metastatic KIRC remains a challenge due to limited therapy options. For years, resistance to chemotherapy and radiation has been a major issue rendering surgical cytoreductive procedure the only effective therapy for advanced and metastatic KIRC. Development and clinical implementation of novel systemic therapies, such as multitarget tyrosine kinase inhibitors, has improved the long-term survival of advanced stage KIRC and recently brought into question the therapy regime with cytoreductive surgical therapy.2 The development of immune checkpoint inhibitors targeting the cytotoxic T-lymphocyte associated protein 4 (CTLA-4) or the programmed cell death 1 (PD-1) receptor/PD-1 ligand 1 (PD-L1)-axis have led to improved clinical outcomes in advanced and metastatic disease stages across various malignancies such as melanoma or non-small cell lung cancer.3 4 Recent studies have also demonstrated efficacy in targeting PD-1/PD-L1 and CTLA-4 in patients with metastatic KIRC.5 6 Despite the great success of immunotherapies targeting the PD-1/PD-L1 axis, many patients do not respond or ultimately develop resistance under treatment. As a result, other inhibitory receptors are being investigated as alternative immune therapy targets.7 One of these potential targets is the immune checkpoint lymphocyte activating 3 (LAG3, LAG-3, CD223).8 9

The regulation of LAG3 in KIRC on an epigenetic level, however, is largely unknown and might involve DNA methylation. DNA methylation is an important epigenetic mechanism involved in many fundamental biological processes, that is, differentiation (including T cell differentiation), X chromosome inactivation, and imprinting. Methylation of the gene promoter region is frequently associated with transcriptional silencing while gene body methylation is often a hallmark of transcriptionally active genes.10 11 Elucidation of the epigenetics of immune-checkpoint receptor genes might help in understanding mechanisms of response and resistance to the respective inhibitors, ultimately paving the way for the development of predictive biomarkers.

Methylation has been shown to regulate expression of PD-1 and PD-L1 in various malignancies.12–16 We therefore hypothesized that LAG3 expression might also be regulated epigenetically via DNA methylation. Understanding the epigenetic regulation of LAG3 expression is of major interest, as it could be useful for the stratification of patients who may benefit from immunotherapeutic LAG3 inhibition.

The results shown here are partly based on data generated by The Cancer Genome Atlas Research Network (TCGA, http://cancergenome.nih.gov/). The KIRC TCGA cohort comprises in total n=741 samples consisting of n=533 tumor samples and n=208 normal adjacent tissues (NAT). For the majority of the KIRC cohort, clinical follow-up and overall survival (OS) data were available (n=534, mean follow-up period 3.68 years, range 0–12.4 years). The mean age at initial diagnosis was 61.1 (range 26–90). For validation, a second cohort comprising n=118 formalin-fixed and paraffin-embedded KIRC tumors from patients treated at the University Hospital Bonn was included (UHB cohort). The mean follow-up period was 4.6 years (range 0–14.1 years), and the mean age at initial diagnosis was 62.7 (range 28–84).

The genomic organization of LAG3 is depicted in figure 1. The LAG3 gene encodes for two protein coding transcripts sharing the same transcription start site. Transcript LAG3-202 represents a truncated version of the full-length transcript LAG3-201 lacking the last three exons. An extended promoter and its flanks are predicted in the region of the transcription start site. Binding sites of the transcriptional repressor CTCF (CCCTC-binding factor) are predicted within the promoter and downstream from LAG3. In total, we investigated 16 CpG sites probed by beads from the Infinium HumanMethylation450 BeadChip. CpG sites 1–12 are located in the promoter and promoter flank region, CpG site 2 is found in the promoter-embedded CTCF binding site, beads 13 and 14 target CpG sites within the gene body downstream from LAG3-202 and CpGs 15 and 16 are situated in the downstream CTCF binding site. We further developed qMSP assays that target promoter CpG sites 4 and 8. In total, we analyzed a cohort comprised of n=318 patients with KIRC and n=24 NAT obtained from TCGA, a cohort of n=118 KIRCs from the University Hospital Bonn (UHB cohort), isolated PBMCs (n=25), and human KIRC cell lines (n=12) with regard to LAG3 gene methylation and expression.

Despite the resounding success of multiple antibody-based therapies that inhibit the PD-1/PD-L1 axis, there are numerous nonresponding patients. Since this axis represents only one of a large number of inhibitory immune checkpoints, investigation of other potential targets that can be blocked to enhance the anticancer immune response is warranted. One of these targets is the immune checkpoint receptor, LAG3.

In the present study, we investigated DNA methylation at single CpG site resolution within the LAG3 gene. We correlated LAG3 methylation with transcriptional activity, protein expression, most frequent somatic mutations, overall patient survival, and tumor immune microenvironment in KIRC tissue. Our results suggest a strong epigenetic regulation of LAG3 via promoter methylation. Molecular, clinicopathological, and immunogenicity correlates with LAG3 promoter DNA methylation.

Distinct binding sites for transcriptional repressor CTCF are predicted within the promoter and downstream from LAG3. CTCF is known to be heavily involved in transcriptional regulation of regulation of chromatin architecture.28 Differential methylation of the CTCF binding sites within LAG3 correlates with its expression; however, further functional investigations are needed to clarify the mechanism of this differential regulation.

LAG3 is expressed on activated T cells, natural killer cells, B cells, and plasmacytoid dendritic cells and binds with high affinity to MHC class II receptors. Research has shown that LAG3 negatively regulates cellular proliferation, activation, and homeostasis of T cells, similar to the function of CTLA-4 and PD-1. Furthermore, it has been demonstrated that LAG3 plays a role in regulatory T cell (Treg) suppressive function and maintenance of CD8⁺ T cell exhaustion during chronic viral infection.29 30 In addition, LAG3 is involved in the process of maturation and activation of dendritic cells.

In an oncological context, LAG3 inhibition activates effector T cells, similar to PD-1 inhibitors, and additionally inhibits suppressive Tregs. This combined effect has the potential to enhance the antitumor response of the tumor-infiltrating immune cells.8 9 Hence, LAG3 is considered as a potential target in immunotherapy. Currently, clinical trials evaluating the safety and efficacy of multiple therapeutic monoclonal LAG3 antibodies such as relatlimab (Bristol-Myers Squibb) or MK-4280 (Merck/MSD Sharp & Dohme) alone (ClinicalTrials.gov Identifier: NCT01968109) or in combination with nivolumab and/or ipilimumab (NCT01968109, NCT03459222, NCT02996110) in KIRC are ongoing.

Clinical studies evaluating the efficacy of immune checkpoint inhibitors targeting the PD-1/PD-L1 axis have proposed different potential markers such as tumor-intrinsic immune checkpoint expression, the intensity of intratumoral CD8⁺ T cell infiltrates, or an enhanced interferon-γ signature to be predictive for therapy response.31–35 Currently, immunohistochemically determined receptor status is the only widely used biomarker prior immune checkpoint inhibition. However, the use of PD-L1 IHC as a predictive biomarker is confounded by multiple difficulties, such as the usage of different antibodies and expression scores, as well as interlaboratory and interobserver variability.36 The use of methylation as a biomarker has several advantages: DNA methylation is a stable epigenetic modification which is not as dynamic as mRNA or protein expression. In addition, it is chemically stable and can also be measured investigator-independently. Moreover, a quantitative measurement is also possible in small sample amounts (microdissected cells, biopsies), a situation often faced in the clinical setting.37 38 DNA methylation status is a particular promising biomarker for immunotherapies as demonstrated in previous reports.12–16

Results of the present study suggest that LAG3 gene expression in KIRC is strongly regulated epigenetically, that is, LAG3 methylation correlates with LAG3 mRNA expression. In particular, hypomethylation of the promoter region negatively correlates with enhanced mRNA expression in the KIRC TCGA cohort whereas two methylation sites, one in the body of the LAG3 gene, and one downstream within a putative binding site of the transcriptional repressor CTCF, positively correlate with mRNA expression. These findings on epigenetic regulation are consistent with our findings in KIRC tumor samples (UHB cohort). Overall, LAG3 promoter hypomethylation was concordantly associated with LAG3 protein expression. Results from IHC demonstrated that LAG3 is expressed intrinsically by a subset of KIRC tumor cells and sparsely by immune infiltrates. This LAG3-expressing KIRC subset was characterized by adverse clinical outcome, high immune cell infiltration, and LAG3⁺ immune cell infiltration indicating those tumors as immunologically “hot tumors.” Consistent with our observations, distinct populations of phenotypically exhausted CD8⁺ T cells have previously been shown to exist in highly immune-infiltrated KIRC tumors.39 Furthermore, soluble LAG3, which is known to function as a dendritic cell activator, was also recently shown to be associated with advanced KIRC stage.40 Tumor-specific LAG3 upregulation might be an immune evasion phenotype similarly to PD-L1 upregulation or upregulation of other immune checkpoints. This hypothesis needs further functional evaluation. However, the finding of KIRC cell-intrinsic LAG3 expression is contradictory to a recent study on immune checkpoints in the microenvironment of primary and metastatic KIRC that revealed LAG3 expression by tumor infiltrating immune cells in a small subset of KIRC but not by the tumor itself.27 Previously, the immune checkpoint PD-1, which was thought to be solely expressed on immune cells, was also found to be expressed cancer cell-intrinsically in different tumor entities including renal cancers.41 Furthermore, it has already been shown that LAG3 is solely expressed on immune cells and in the central nervous system.42 Additionally, we found LAG3 expression in proximal tubules of NAT, the site in which malignant transformation occurs for KIRC,43 44 and in the urothelium of the renal pelvis. While the finding of cancer cell-intrinsic LAG3 expression in KIRC is interesting, further validation and functional analyses are required.

Based on data on LAG3 DNA methylation shown here, LAG3 methylation status could be used as an observer-independent predictive biomarker for stratification of patients who could benefit from LAG3 or other immune checkpoint inhibition therapy. However, this is speculative and further prospective studies are needed to evaluate the extent to which therapy response of LAG3 inhibition is predicted by promoter hypomethylation. Therefore, we recommend integration of LAG3 promoter methylation in biomarker programs of clinical trials to evaluate the efficacy of immune checkpoint inhibition in patients with KIRC.

Correlations of LAG3 mRNA expression and methylation in the KIRC TCGA cohort, comprising whole tumor tissue, NAT, PBMC, and KIRC cell lines showed significant differences between sample type. This suggests a cell type-specific epigenetic regulation of LAG3. For example, the methylation status of CpG site 13 showed a differential correlation with LAG3 mRNA in PBMCs compared with KIRC cell lines. This could be a hint for an alternative LAG3 promoter that indicates a leukocyte-specific expression of a particular splice variant. LAG3 variants that display different effects on immunological microenvironments have already been described.45 Widespread tissue-specific and cell type-specific methylation patterns in diverse biological processes are already described, highlighting the complexity of epigenetics.46

In KIRC, the most frequently mutated genes comprise chromatin modifier genes as PBRM1 and SETD2.47 48 Interestingly, recent studies have shown associations between therapy response to ICI and PBRM1 mutation status in KIRC and mesothelioma.47 49 Genomic alterations in SETD2 have already been associated with an altered activation of the interferon signaling pathway. Our data also showed a differential LAG3 methylation status in the context of mutations within chromatin modifier genes.

Further, our data demonstrate that promoter hypomethylation predicts unfavorable OS in KIRC. This finding highlights clinical significance. However, the main focus of our study is on the epigenetic regulation of LAG3 via DNA methylation in KIRC. LAG3 promoter hypomethylation is associated with increased immune cell infiltration and an interferon-γ signature. In the context of PD-1/PD-L1 directed immune checkpoint inhibition, interferon-γ signature is suggested as a critical driver for antitumoral immune efficacy and thereby a potential biomarker for ICI therapy response.35 Using GEO datasets, LAG3 methylation in KIRC cell lines and PBMCs was investigated independently. Our results showed that compared with tumor cell lines, PBMCs are hypomethylated in the LAG3 promoter region. The KIRC TCGA data are based on whole tumor tissue comprising tumor cells, stroma, and immune infiltrates. LAG3 promoter hypomethylation could thereby serve as a surrogate biomarker for high immune cell infiltration. However, a limitation of this data is that only isolated immune cells from peripheral blood of healthy individuals were examined. In this context, further studies are needed to clarify whether LAG3 methylation in the tumor is a surrogate biomarker for immunogenicity, that is, infiltration by immune cells and an interferon-γ signature, or whether LAG3 expression of the tumor cells themselves represents a more aggressive phenotype. One aspect of this is highlighted by our results from IHC staining, which showed that LAG3-expressing tumors are associated with poorer survival.

Our results suggest an epigenetic regulation of LAG3 in KIRC by DNA methylation. Understanding the epigenetic regulation of LAG3 in KIRC might help to develop biomarkers for the stratification of patients eligible for anti-LAG3 immunotherapies. Our study provides rationale to include LAG3 methylation testing as a potential predictive biomarker into the biomarker program of clinical trials with anti-LAG3 immune checkpoint inhibitors.