The coinhibitory receptor lymphocyte activation gene 3 (LAG-3) is an immune checkpoint molecule that negatively regulates T cell activation, proliferation, and homeostasis. Blockade or deletion of LAG-3 in autoimmune-prone backgrounds or induced-disease models has been shown to exacerbate disease. We observed significantly fewer LAG-3+ CD4 and CD8 T cells from subjects with relapsing-remitting multiple sclerosis (RRMS) and type 1 diabetes. Low LAG-3 protein expression was linked to alterations in mRNA expression and not cell surface cleavage. Functional studies inhibiting LAG-3 suggest that in subjects with RRMS, LAG-3 retains its ability to suppress T cell proliferation. However, LAG-3 expression was associated with the expression of markers of apoptosis, indicating a role for low LAG-3 in T cell resistance to cell death. In T cells from subjects with RRMS, we observed a global dysregulation of LAG-3 expression stemming from decreased transcription and persisting after T cell stimulation. These findings further support the potential clinical benefits of a LAG-3 agonist in the treatment of human autoimmunity.

This article is featured in Top Reads, p.

Autoimmune diseases are chronic and debilitating and affect more than 20 million people in the United States (1). Although immunotherapies have greatly improved treatment, we are still unable to prevent or cure autoimmune disease. An understanding of the molecular and cellular mechanisms underlying the loss of tolerance and progression to autoimmune disease is crucial to move this field forward. In this study, we investigate the role of the inhibitory receptor (IR) lymphocyte activation gene-3 (LAG-3) in two organ-specific autoimmune diseases: relapsing-remitting multiple sclerosis (RRMS) and type 1 diabetes (T1D). IRs counterbalance costimulatory signals and prevent excessive effector T cell activation contributing to autoimmunity (2). Consequently, the expression of IRs is necessary to promote appropriate self-tolerance, but overexpression can lead to an inability of T cells to mount effective immune responses against tumors or pathogens (3). This critical role in regulating immune responses has made IRs an attractive target for immunotherapy, especially in the setting of cancer, in which targeting CTLA-4 and PD-1 have been highly successful. More recently, clinical trial results from combination LAG-3/PD-1 immune checkpoint inhibition in patients with melanoma doubled the progression-free survival time compared with PD-1 blockade alone, underscoring the value of targeting LAG-3 in the clinic (4). Despite the clinical success of immune checkpoint inhibition in the treatment of cancer, these therapies are considered dangerous for the treatment of patients with underlying autoimmunity (5, 6) and have been associated with both disease relapse and death in patients with multiple sclerosis (MS) (7).

To date, almost all of the work investigating the role of LAG-3 in autoimmunity has been done in murine models. Experimental autoimmune encephalomyelitis (EAE) is a commonly used murine model for the inflammatory demyelinating disease MS and is considered a prototype for T cell–mediated autoimmune disease. In studies of the myelin oligodendrocyte glycoprotein–specific TCR transgenic mouse EAE model, the upregulation of LAG-3 on myelin-specific CD4+ gut-induced intraepithelial lymphocytes that migrate to the CNS is required to reduce disease severity (8). The role of LAG-3 shedding in the context of EAE has also been explored, as the inhibitory function of T cells is enhanced in noncleavable LAG-3 (LAG-3NC) (9). Less severe EAE is seen in LAG-3NCCD4+Foxp3 cells compared with LAG-3NC CD8 T cells or LAG-3 cleavable controls. Furthermore, LAG-3NC on CD4+Foxp3 T cells reduced proliferation and increased cleaved caspase 3 (cCasp3), suggesting a selective role for LAG-3NC mediating T cell functionality and proliferation in EAE (10). In addition, studies in which Ag-specific LAG-3+Foxp3+ regulatory T cells are expanded through vaccination demonstrate enhanced suppression of Ag-specific autoreactive effector T cells and bystander immunosuppression (11). Thus, the LAG-3 inhibitory pathway may function to selectively immunosuppress activated T cells through a variety of pathways that may be clinically effective in the treatment of autoimmunity. Although these studies looking at the impacts of Lag3 deletion or overexpression provide a foundation for hypothesizing a role for LAG-3 in autoimmunity, the source and extent of LAG-3 expression in human autoimmunity have not been investigated.

In this study, we address this gap in knowledge by investigating regulation of LAG-3 expression in primary T cells from patients with two organ-specific autoimmune diseases: RRMS and T1D. We explore the source of altered LAG-3 expression in these diseases and the functional implications of this altered expression.

All samples used in this study were from the Benaroya Research Institute Registry and Repository. The study was approved by the Benaroya Research Institute’s Institutional Review Board (protocol no. IRB07109), and all subjects gave written informed consent. Two patient cohorts and healthy control subjects (HCs) were selected for these studies; both cohorts were heterogeneous with respect to disease duration, disease activity, and therapy for RRMS and T1D. The HCs were selected based on the absence of autoimmune disease or any family history of autoimmunity and were age- and sex-matched to the subjects with RRMS and T1D. Individuals were diagnosed with RRMS based on the Revised McDonald diagnostic criteria for MS (12). All experiments were performed in a blinded manner. Characteristics of study participants are listed in Supplemental Table I.

Thawed PBMCs were rested in serum-free X-VIVO 15 Medium (Lonza Bioscience) for 1 h and washed with PBS. Total CD3 T cells, CD4 T cells, or CD8 T cells were purified by negative selection using MACS technology (Miltenyi Biotec). T cell stimulation was carried out in 96-well flat-bottom plates (Thermo Fisher Scientific) at 2 × 106 cells/ml in 200 μl of RPMI-1640 medium (supplemented with 10% human serum, 2 mM glutamine, 100 IU/ml penicillin, and 0.1 mg/ml streptomycin) with plate-bound anti-CD3 (OKT3; 1 μg/ml) and soluble anti-CD28 (CD28.2; 2 μg/ml) for 24 h.

The relevant fluorochrome-conjugated anti-human surface Abs used were specific for LAG-3 (polyclonal goat IgG; R&D Systems), ADAM10 (clone 163003; R&D Systems), ADAM17 (clone 111633; R&D Systems), CD25 (clone BC96; BioLegend), and CD95/Fas (clone DX2; BioLegend). Dead cells were discriminated by staining with the LIVE/DEAD Fixable Blue Dead Cell Stain Kit (Invitrogen). General T cell gating strategy, gating on LAG-3+ T cells, as well as ADAM10 and ADAM17 expression are shown in Supplemental Fig. 1A–D. For intracellular staining, cells were stained with surface markers (in this study, a LAG-3 mAb; clone 3DS223H; eBioscience), fixed in Fix/Perm buffer (eBioscience) for 30 min, washed in permeabilization buffer (eBioscience) twice, and stained for intracellular factors Ki67 (clone B56; BD Biosciences) and LAG-3 (clone 874512; R&D Systems) in permeabilization buffer for 30 min on ice. To stain for cCasp3 (clone C92-605; BD Biosciences), cells were fixed and permeabilized using Fix Buffer I and Perm Buffer III (BD Biosciences), respectively. All cells were acquired on an LSR Fortessa (BD Biosciences), and data were analyzed using FlowJo version v10.6.2 (Tree Star).

Magnetically enriched CD3+ T cells were distributed to 96-well flat-bottom plates (Corning) at 2 × 106 cells/ml in 200 μl supplemented RPMI-1640 medium (10% human serum) with plate-bound anti-CD3 (OKT3; 1 μg/ml), soluble anti-CD28 (CD28.2; 2 μg/ml), and either the pan-metalloproteinase inhibitor TNF-α protease inhibitor-1 (TAPI-1) or the ADAM10 inhibitor GI254023X (20 μM; Selleck Chemicals). After 24 h, supernatants were collected, and soluble LAG-3 (sLAG3) concentrations were determined using the LAG-3 Human ELISA Kit (Invitrogen).

Magnetically enriched CD4 or CD8 T cells were either rested or stimulated with plate-bound anti-CD3 (OKT3; 1 μg/ml), soluble anti-CD28 (CD28.2; 2 μg/ml) at 1 × 106 cells/ml in 200 μl RPMI-1640/well for 48 h. RNA was extracted from 6 × 105 cells using the RNAqueous-Micro Total RNA Isolation Kit (Invitrogen) with on-column DNA digestion (Qiagen). SuperScript III (Life Technologies) was used to generate cDNA, and gene expression was measured by multiplex real-time PCR performed on an ABI 7500 Fast Real-Time PCR System. TaqMan expression assays for LAG3 (Hs00158563_m1), PDCD1 (Hs01550088_m1), TIGIT (Hs00545087_m1), and HAVCR2 (Hs00958618_m1) were used in combination with RPL36AL (Hs00733231_m1) for normalization (Thermo Fisher Scientific). Quantitative detection of LAG3 mRNA levels was determined by 2−ΔΔCt calculations and reported relative to a Jurkat cell line (clone E-61; ATCC TIB-152).

PBMCs were plated at 3.75 × 106 cells/ml in RPMI-1640 medium (supplemented with 10% human serum, 2 mM glutamine, 100 IU/ml penicillin, and 0.1 mg/ml streptomycin), and a LAG-3 antagonist Ab (anti–LAG-3, 10 μg/ml; 17B4; Abcam) was added to select wells for 30 min before the addition of PepTivator CMV pp65 and Adv5 hexon (175 ng/ml; Miltenyi Biotec) and incubated at 37°C for 6 d.

Statistical analysis was performed using GraphPad Prism 9 and JMP software from SAS. To assess statistical significance, a one-way ANOVA with Holm-Sidak multiple-comparison test with single paired variance was carried out to correct for multiple testing. Results were expressed as medians, and differences were considered statistically significant at p < 0.05. Multiple variable and simple linear regression were performed, and the coefficient of determination (r2) was reported stratified by patient group. Bivariate and multivariate (partial) Pearson correlation analyses were also performed. Outliers were removed from (Fig. 1 using the ROUT method (coefficient Q = 0.1%) (13).

FIGURE 1.

Frequency of T cell LAG-3 surface expression. LAG-3+ frequency in: CD4 T cells (A), regulatory CD4 T cells (Treg) (CD25+, CD127int) (B), conventional (Tconv; non-Treg) (C), CD8 T cells (D), naive CD8 T cells (CD45RA+CD45RO) (E), and memory CD8 T cells (Tmem; CD45RACD45RO+) (F) from HCs (n = 69) and subjects with T1D (n = 104) and RRMS (n = 121). Lines shown are median, and p values are shown after a one-way ANOVA. Results are shown from 25 independent experiments.

FIGURE 1.

Frequency of T cell LAG-3 surface expression. LAG-3+ frequency in: CD4 T cells (A), regulatory CD4 T cells (Treg) (CD25+, CD127int) (B), conventional (Tconv; non-Treg) (C), CD8 T cells (D), naive CD8 T cells (CD45RA+CD45RO) (E), and memory CD8 T cells (Tmem; CD45RACD45RO+) (F) from HCs (n = 69) and subjects with T1D (n = 104) and RRMS (n = 121). Lines shown are median, and p values are shown after a one-way ANOVA. Results are shown from 25 independent experiments.

Close modal

We measured LAG-3 cell surface expression on T cells from patients with RRMS (n = 121) and T1D (n = 104) compared with age- and sex-matched HCs (n = 69) (Supplemental Table I, cohort 1). Subjects with T1D were not on disease-modifying therapies (DMT), whereas 42% of subjects with RRMS were treated with at least one DMT (natalizumab, glatiramer acetate, dimethyl fumarate, and IFN-β1a and 1b). We observed significantly fewer LAG-3+ CD4 T cells in both regulatory (CD4+CD25hiCD127low) and conventional CD4 T cell subsets (Fig. 1A–C). Further gating of the CD4 population based on naive and memory markers CD45RA and CD45RO showed fewer LAG-3+ CD4 memory cells in RRMS as compared with HCs, but not in T1D (Supplemental Fig. 1E, 1F). Altered LAG-3 expression was not limited to the CD4 T cell compartment; there were also fewer LAG-3+ CD8 T cells in both RRMS and T1D compared with HCs (Fig. 1D). Although there is generally more LAG-3 expression on memory CD8 T cells than naive, we observed fewer LAG-3+ cells in both memory and naive CD8 T cells isolated from subjects with RRMS and T1D (Fig. 1E, 1F). In subjects with RRMS, the frequency of LAG-3 expression does not appear to be influenced by treatment with DMT (Supplemental Fig. 1G, 1H), disease flare (Supplemental Fig. 1I, 1J), age, or sex (not shown). Together, these data suggest that T cell LAG-3 surface expression is globally dysregulated in subjects with RRMS and T1D.

Cell surface LAG-3 is regulated by proteolytic cleavage that results in the shedding of sLAG3 (14). This cleavage is mediated by the metalloproteinases ADAM10 and ADAM17 and necessary for optimal T cell function (9). Because low LAG-3 cell surface expression on CD4 and CD8 T cells could be the result of enhanced proteolytic cleavage, we measured shedding of LAG-3 in T cell supernatants after activation. For these experiments, we selected a subset of subjects from cohort 1 based on their resting surface LAG-3 expression. We selected HCs with a LAG-3+ T cell frequency greater than the third quartile and subjects with RRMS and T1D with LAG-3+ frequency less than the first quartile. We found a significantly lower concentration of sLAG3 in supernatant taken from the RRMS and T1D cultures as compared with HCs (Fig. 2A). In the same assay, we measured metalloproteinase expression at rest and upon activation. We observed significantly lower ADAM10 on both CD4 and CD8 resting T cells from subjects with RRMS and T1D compared with HCs at rest, whereas ADAM10 was lower on T1D CD8 cells (Fig. 2B, Supplemental Fig. 2A). Although ADAM10 expression increased upon activation, subjects with RRMS and T1D had significantly lower ADAM10 in CD8 T cells (Fig. 2C), and subjects with RRMS also had less ADAM10 expression on CD4 T cells after activation compared with HCs (Supplemental Fig. 2B). ADAM17 was not differentially expressed in resting or stimulated T cells between cohorts (Supplemental Fig. 2C–F). To further address whether the altered frequency of LAG-3+ T cells could be influenced by shedding, we used metalloproteinase inhibitors to determine if blockade of ADAM10 and ADAM17 impacted cleavage to a greater degree in subjects with RRMS and T1D. Both the pan-metalloproteinase inhibitor TAPI-1 and the ADAM10 inhibitor GI254023X decreased the concentration of sLAG3 in conjunction with TCR stimulation (Fig. 2D). In cultures from subjects with RRMS and T1D, the change in sLAG3 concentration after ADAM10 inhibition was comparable to that in HCs (Fig. 2E). Together, these results indicate that although ADAM10 has reduced expression in RRMS and T1D, it is fully functional.

FIGURE 2.

The role of surface LAG-3 expression and ADAM10 expression in the generation of sLAG3. (A) The concentration of sLAG3 after 48-h TCR stimulation of CD3+ T cells. The gMFI of ADAM10 in unstimulated CD8 T cells (B) and after a 48-h TCR stimulation (C). (D) Concentration of sLAG3 after 48-h TCR stimulation with or without the addition of the pan-metalloproteinase inhibitor TAPI-1 or the ADAM10 inhibitor GI254023X. (E) Concentration of sLAG3 after TCR stimulation and with ADAM10 inhibition in paired subjects. Results from one experiment: HCs, n = 5; T1D, n = 4; RRMS, n = 4. (F) The frequency of CD8 T cell surface LAG-3 versus the concentration of sLAG3 after TCR stimulation in HCs and subjects with RRMS and T1D (p < 0.0001, r2 = 0.6390 for all subjects). (G) The gMFI of ADAM10 on CD8 T cells versus the concentration of sLAG3 after TCR stimulation (p < 0.0001, r2 = 0.3450 for all subjects). Results from two independent experiments: HCs (green), n = 17; RRMS (red), n = 15; and T1D (orange), n = 13. Lines shown are median, and p values are shown after a one-way ANOVA (A–C), paired t test (D and E), and Spearman correlation test (F and G).

FIGURE 2.

The role of surface LAG-3 expression and ADAM10 expression in the generation of sLAG3. (A) The concentration of sLAG3 after 48-h TCR stimulation of CD3+ T cells. The gMFI of ADAM10 in unstimulated CD8 T cells (B) and after a 48-h TCR stimulation (C). (D) Concentration of sLAG3 after 48-h TCR stimulation with or without the addition of the pan-metalloproteinase inhibitor TAPI-1 or the ADAM10 inhibitor GI254023X. (E) Concentration of sLAG3 after TCR stimulation and with ADAM10 inhibition in paired subjects. Results from one experiment: HCs, n = 5; T1D, n = 4; RRMS, n = 4. (F) The frequency of CD8 T cell surface LAG-3 versus the concentration of sLAG3 after TCR stimulation in HCs and subjects with RRMS and T1D (p < 0.0001, r2 = 0.6390 for all subjects). (G) The gMFI of ADAM10 on CD8 T cells versus the concentration of sLAG3 after TCR stimulation (p < 0.0001, r2 = 0.3450 for all subjects). Results from two independent experiments: HCs (green), n = 17; RRMS (red), n = 15; and T1D (orange), n = 13. Lines shown are median, and p values are shown after a one-way ANOVA (A–C), paired t test (D and E), and Spearman correlation test (F and G).

Close modal

After observing both decreased ADAM10 expression and fewer LAG-3+ T cells in subjects with RRMS, we addressed which of these two factors most contributed to the concentration of sLAG3. We observed a positive correlation between the frequency of sLAG3 and surface LAG-3 expression on CD8 T cells when pooling all subjects together (p < 0.0001; r2 = 0.6390) (Fig. 2F) and similarly in CD4 T cells (p < 0.0001; r2 = 0.4150) (Supplemental Fig. 2G). In contrast to the correlation with LAG-3 frequency, the relationship between ADAM10 and sLAG3 was much weaker in CD8 and CD4 T cells (Fig. 2G, Supplemental Fig. 2H). In CD8 T cells from subjects with RRMS, the weak relationship between ADAM10 expression and sLAG3 concentration suggests that altered expression of ADAM10 was not the primary cause of altered sLAG3. From this, we conclude that alterations in LAG-3 cleavage are not the primary driver of altered LAG-3 expression on T cells in T1D and RRMS.

A linear regression analysis taking into account the surface expression of LAG-3 on CD8 T cells and cohort (HC, RRMS, and T1D) found that the frequency of surface LAG-3 explains ∼66% of the variability in the concentration of sLAG3, whereas ADAM10 expression and cohort type only explains 29%. Together, the frequency of surface LAG-3+ CD8 T cells and ADAM10 geometric mean fluorescence intensity (gMFI) explain ∼69% of the variability in the concentration of sLAG3, after controlling for cohort. The effect of adding cohort to this model had minimal impact (4%) on the total model r2 value. Moreover, a multivariable model investigating interaction effects between surface LAG-3 and ADAM10 expression with cohort demonstrated that the relationship between these two factors and the concentration of sLAG3 did not differ between cohorts. This analysis implies that the expression of ADAM10 only modestly contributes to the cleavage of cell surface LAG-3, regardless of cohort.

LAG-3 is stored in lysosomal compartments to facilitate rapid translocation to the cell surface following TCR stimulation; in the absence of stimulation, intracellular LAG-3 is degraded (15, 16). To determine whether intracellular LAG-3 is low after TCR stimulation, we used the same subset of subjects from cohort 1 who were previously used for sLAG3 studies. To measure both surface and intracellular LAG-3, instead of using a polyclonal Ab, we used two LAG-3 mAbs on separate colors (874512, R&D Systems; and 3DS223H, eBioscience). Representative flow cytometry plots show LAG-3 expression in resting (blue) and stimulated (red) CD8 T cells from HCs and subjects with RRMS (Fig. 3A). Similar to our initial observations in resting T cells, following stimulation, surface LAG-3 in CD4 and CD8 T cells from subjects with RRMS was significantly lower than in T cells from HCs (Fig. 3B, 3C). Intracellular LAG-3 was also significantly lower in CD4 T cells and trended lower in CD8 T cells from subjects with RRMS (Fig. 3D, 3E). Although we did not observe statistically significant differences in surface LAG-3 expression on CD4 and CD8 T cells from subjects with T1D, intracellular LAG-3 in CD8 T cells trended lower in these subjects. These data indicate that low LAG-3 protein production in RRMS contributes to the frequency of LAG-3+ T cells in RRMS, whereas in T1D, there may be less intracellular LAG-3 available for cell surface expression.

FIGURE 3.

Frequency of intracellular and surface LAG-3 on T cells after stimulation. (A) Representative flow cytometry plots of surface and intracellular LAG-3 staining before (blue) and after (red) 48-h TCR stimulation in CD8 T cells from three independent experiments. After 48-h TCR stimulation of purified CD3+ T cells, the surface LAG-3 expression was measured in CD4 (B) and CD8 T cells (C) as well as the intracellular LAG-3 expression (D and E) from HCs and subjects with RRMS and T1D. HCs, n = 17; RRMS, n = 15; T1D, n = 12. Lines shown are median, and p values are shown after a one-way ANOVA.

FIGURE 3.

Frequency of intracellular and surface LAG-3 on T cells after stimulation. (A) Representative flow cytometry plots of surface and intracellular LAG-3 staining before (blue) and after (red) 48-h TCR stimulation in CD8 T cells from three independent experiments. After 48-h TCR stimulation of purified CD3+ T cells, the surface LAG-3 expression was measured in CD4 (B) and CD8 T cells (C) as well as the intracellular LAG-3 expression (D and E) from HCs and subjects with RRMS and T1D. HCs, n = 17; RRMS, n = 15; T1D, n = 12. Lines shown are median, and p values are shown after a one-way ANOVA.

Close modal

To determine whether low LAG-3 protein production and, ultimately, fewer LAG-3+ T cells from subjects with RRMS and T1D were due to reduced transcription, we measured LAG3 mRNA levels using real-time quantitative RT-PCR. CD4 and CD8 T cells were magnetically enriched from frozen PBMCs isolated from a smaller cohort of subjects with RRMS (n = 20) and T1D (n = 15) and HCs (n = 19) (Supplemental Table I, cohort 2). In T cells from subjects with RRMS, LAG3 mRNA expression was less than in resting CD4 (median 2.58-fold) and CD8 T cells (2.16-fold) from HCs, whereas no difference was observed in T1D CD4 or CD8 T cells (Fig. 4A, 4B). We next examined the expression of LAG3 after TCR stimulation and observed an increase in LAG3 transcript in all subjects. There were no differences in expression after stimulation within CD4 T cells (Fig. 4A); however, in CD8 T cells, both subjects with RRMS and T1D expressed significantly less LAG3 after stimulation (Fig. 4B). The expression of LAG3 was not significantly different between subjects with RRMS on and off DMT (Supplemental Fig. 3A). These findings suggest that low LAG3 mRNA levels contribute to less LAG-3 protein expression in subjects with RRMS and may contribute in part to fewer LAG-3+ CD8 T cells in T1D.

FIGURE 4.

Transcriptional expression of LAG3 in T cells. LAG3 mRNA expression in resting CD4 (A) and CD8 T cells (B). (C) PDCD1 mRNA levels correlated with LAG3 in HCs but not subjects with RRMS and T1D in resting CD4 T cells. Quantitative RT-PCR for LAG3 and PDCD1 was normalized to the ribosomal gene RPL36A and reported relative to a Jurkat cell line. HCs, n = 19; T1D, n = 15; RRMS, n = 20; results shown are from three independent experiments. Lines shown are median, and p values are from a one-way ANOVA. Linear regressions are shown with Pearson correlation coefficient. Stim, stimulation.

FIGURE 4.

Transcriptional expression of LAG3 in T cells. LAG3 mRNA expression in resting CD4 (A) and CD8 T cells (B). (C) PDCD1 mRNA levels correlated with LAG3 in HCs but not subjects with RRMS and T1D in resting CD4 T cells. Quantitative RT-PCR for LAG3 and PDCD1 was normalized to the ribosomal gene RPL36A and reported relative to a Jurkat cell line. HCs, n = 19; T1D, n = 15; RRMS, n = 20; results shown are from three independent experiments. Lines shown are median, and p values are from a one-way ANOVA. Linear regressions are shown with Pearson correlation coefficient. Stim, stimulation.

Close modal

Knowing that IRs can function as a gene module in response to stimulation (3), we asked if altered LAG3 mRNA expression was part of a broader dysregulation of IRs in subjects with T1D and RRMS. Unlike our findings with LAG3, we found no differences in the mRNA expression of PDCD1 (PD-1), HAVCR2 (TIM3), and TIGIT in resting CD4 and CD8 T cells from individuals with RRMS and T1D when compared with HCs (Supplemental Fig. 3B). Further, the positive correlation between LAG3 and PDCD1 transcript seen in HCs was not present in resting CD4 T cells from subjects with RRMS and T1D (Fig. 4C). After TCR stimulation, LAG3 mRNA levels were tightly correlated with TIGIT, PDCD1, and HAVCR2, indicating the relationship between LAG3 and other coinhibitory molecules is retained upon activation, but not maintained in the resting state (Supplemental Fig. 3C).

To address the functional implications of fewer LAG-3+ T cells in RRMS, we first investigated the effect of LAG-3 inhibition on T cell proliferation in samples from three HCs and three off-therapy subjects with RRMS selected from cohort 1 on the basis of high and low surface LAG-3 expression, respectively. We measured proliferation after a 6-day peptide stimulation in the presence or absence of a LAG-3 antagonistic Ab (anti–LAG-3) and observed more proliferation of LAG-3+ T cells when LAG-3 was blocked in both HCs and RRMS T cells (Fig. 5A, 5B). Proliferation was comparable in activated CD4 T cells from HCs and those with RRMS (Fig. 5A); CD8 T cells from subjects with RRMS were significantly less proliferative than those from HCs (Fig. 5B). Although T cell proliferation was higher after LAG-3 inhibition in both HCs and RRMS, the CD8 T cells from subjects with RRMS remained hypoproliferative relative to those from HCs (Fig. 5B). Although these data are gated on activated CD25+ T cells (the bulk of which are CD45RACCR7 effector memory), similar differences between groups were seen in total CD4 and CD8 T cells. Additionally, the frequency of CD25 was comparable between HCs and subjects with RRMS on both CD4 and CD8 T cells (Fig. 5C, 5D). LAG-3 inhibition comparably enhanced proliferation of T cells from subjects with RRMS and HCs, suggesting T cell LAG-3 is functional in subjects with RRMS.

FIGURE 5.

T cell proliferation after low-dose Ag stimulation and LAG-3 inhibition. Frequency of Ki67 after PBMC peptide stimulation (CMV/Adv5, 175 ng/ml, 6 d) in activated CD4 (A) and CD8 T cells (B) after subtracting the frequency of Ki67 in unstimulated T cells. Frequency of CD25 in CD4 (C) and CD8 T cells (D). Unstimulated in white; peptide-stimulated in black; peptide-stimulated with anti-LAG3 in gray. Circles, HCs; triangles, RRMS. The p values are paired or unpaired t tests. Results are from a single experiment: three HCs and three subjects with RRMS. aLAG3, anti-LAG3; Stim, stimulation.

FIGURE 5.

T cell proliferation after low-dose Ag stimulation and LAG-3 inhibition. Frequency of Ki67 after PBMC peptide stimulation (CMV/Adv5, 175 ng/ml, 6 d) in activated CD4 (A) and CD8 T cells (B) after subtracting the frequency of Ki67 in unstimulated T cells. Frequency of CD25 in CD4 (C) and CD8 T cells (D). Unstimulated in white; peptide-stimulated in black; peptide-stimulated with anti-LAG3 in gray. Circles, HCs; triangles, RRMS. The p values are paired or unpaired t tests. Results are from a single experiment: three HCs and three subjects with RRMS. aLAG3, anti-LAG3; Stim, stimulation.

Close modal

Earlier studies in Lag3-deficient mice and noncleavable LAG-3 mutants indicate a role for LAG-3 in promoting cell death. Furthermore, a resistance to activation-induced cell death has been implicated in the pathogenesis of MS, and DMT have been shown to induce apoptosis (1719). To determine the association of LAG-3 with markers of apoptosis, we used surface LAG-3 expression to select the same subset of patients from cohort 1 as studied in (Fig. 2 (Supplemental Table I) and then measured expression of Fas and caspase 3, both of which play a central role in apoptosis. Fas, once bound to its ligand, activates caspase 3 via caspase 8, leading to apoptosis (20). In this study, we observed a positive correlation in the expression of LAG-3 and Fas in resting CD4 and CD8 T cells from HCs and those with RRMS and T1D (Fig. 6A, 6B). LAG-3 inhibition decreased the frequency of Fas on CD4 and CD8 T cells from six HCs and subjects with RRMS after a 6-day peptide stimulation (Fig. 6C, 6D).

FIGURE 6.

Fas expression and response to LAG-3 inhibition. Frequency of Fas positively correlates with LAG3 in CD4 (A) and CD8 T cells (B) in a pooled cohort of HCs and subjects with RRMS and T1D. Results are from two independent experiments: HCs, n = 14; RRMS, n = 15; T1D, n = 14. Frequency of Fas after peptide stimulation (CMV/Adv5, 175 ng/ml, 6 d; black) and anti–LAG-3 (aLAG3) (10 μg/ml; gray) in CD4 (C) and CD8 T cells (D). Results are from a single experiment, three HCs and three subjects with RRMS (C and D). The p values are shown after a Spearman correlation with 95% confidence interval shown (A and B) and a paired t test (C and D). aLAG3, anti-LAG3; Stim, stimulation.

FIGURE 6.

Fas expression and response to LAG-3 inhibition. Frequency of Fas positively correlates with LAG3 in CD4 (A) and CD8 T cells (B) in a pooled cohort of HCs and subjects with RRMS and T1D. Results are from two independent experiments: HCs, n = 14; RRMS, n = 15; T1D, n = 14. Frequency of Fas after peptide stimulation (CMV/Adv5, 175 ng/ml, 6 d; black) and anti–LAG-3 (aLAG3) (10 μg/ml; gray) in CD4 (C) and CD8 T cells (D). Results are from a single experiment, three HCs and three subjects with RRMS (C and D). The p values are shown after a Spearman correlation with 95% confidence interval shown (A and B) and a paired t test (C and D). aLAG3, anti-LAG3; Stim, stimulation.

Close modal

In addition, subjects with RRMS had fewer cCasp3+ CD4 and CD8 T cells as well as lower expression (gMFI) of cCasp3 on those cells; T cells from subjects with T1D showed similar trends (Fig. 7A–D). Within cCasp3+ cells, there was a positive correlation between the expression of LAG-3 and cCasp3 in CD8 T cells, but not in CD4 T cells from pooled HCs and subjects with RRMS and T1D (Fig. 7E, 7F). Blocking LAG-3 decreased the frequency of cCasp3+ CD4 T cells (p = 0.014) and CD8 T cells (p = 0.039) in all subjects (Fig. 7G, 7H). As we observed in resting T cells, there were fewer cCasp3+ T cells from subjects with RRMS after stimulation and LAG-3 inhibition decreased cCasp3 to a lesser degree. LAG-3 expression is associated with the expression of both Fas and cCasp3, suggesting that low expression of LAG-3 may contribute to resistance to cell death.

FIGURE 7.

cCasp-3 expression and response to LAG-3 inhibition. Frequency of cCasp3 in resting CD4 (A) and CD8 T cells (B) from HCs and subjects with RRMS and T1D. gMFI of cCasp3 within cCasp3+ CD4 (C) and CD8 T cells (D). Correlation of gMFI of LAG-3 versus cCasp3 in cCasp3+ CD4 (E) and CD8 T cells (F). Results are from two independent experiments: HCs, n = 10; RRMS, n = 10; T1D, n = 10. Frequency of cCasp3 after peptide stimulation (Stim) in CD4 (G) and CD8 T cells (H). Results are from a single experiment: three HCs and three subjects with RRMS. Lines shown are median, and p values are shown after a one-way ANOVA (A–D), a Spearman correlation with 95% confidence interval shown (E and F), and an unpaired t test (G and H). aLAG3, anti-LAG3; Stim, stimulation.

FIGURE 7.

cCasp-3 expression and response to LAG-3 inhibition. Frequency of cCasp3 in resting CD4 (A) and CD8 T cells (B) from HCs and subjects with RRMS and T1D. gMFI of cCasp3 within cCasp3+ CD4 (C) and CD8 T cells (D). Correlation of gMFI of LAG-3 versus cCasp3 in cCasp3+ CD4 (E) and CD8 T cells (F). Results are from two independent experiments: HCs, n = 10; RRMS, n = 10; T1D, n = 10. Frequency of cCasp3 after peptide stimulation (Stim) in CD4 (G) and CD8 T cells (H). Results are from a single experiment: three HCs and three subjects with RRMS. Lines shown are median, and p values are shown after a one-way ANOVA (A–D), a Spearman correlation with 95% confidence interval shown (E and F), and an unpaired t test (G and H). aLAG3, anti-LAG3; Stim, stimulation.

Close modal

In this study, we investigated the source and extent of LAG-3 expression in two tissue-specific autoimmune diseases: RRMS and T1D. We found significantly lower protein expression of LAG-3 primarily due to alterations in mRNA expression and not cell surface cleavage in cells from subjects with RRMS and T1D. Our functional studies of the cleavage of LAG-3 in human autoimmunity corroborate murine findings on the generation of sLAG3 primarily by the ADAM10 sheddase and demonstrate the strong association of sLAG3 with LAG-3 surface expression.

We found fewer LAG-3+ T cells in individuals with RRMS and T1D compared with HCs. Notably, this lower expression of LAG-3 was found in all major CD4 and CD8 T cell subsets. In T cells from RRMS, lower expression of LAG-3 was predominantly at the level of transcription and not through posttranscriptional mechanisms, such as extracellular transport or metalloprotease cleavage. In T1D, a single mechanism was not clearly identified, suggesting a more complex set of factors impacting LAG-3 expression in T cells from these subjects. Our findings are consistent with the literature in which low LAG-3 expression has been linked to murine models of autoimmunity (21). Although relevant patient-based studies are largely missing from the literature, fewer LAG-3+ CD4 T cells were recently reported in patients with active psoriatic arthritis (22). In a NOD mouse model, genetic deletion and LAG-3 inhibition have both been shown to accelerate the development of autoimmune diabetes (23). In the setting of EAE, upregulation of LAG-3 on myelin-specific CD4+ gut-induced intraepithelial lymphocytes limited disease severity (8). Prevention of LAG-3 cleavage on CD4 T cells reduced both disease severity and the number of pathogenic IL-17+IFN-γ+GM-CSF+ cells isolated from the brain (10).

We identified transcriptional dysregulation as the source of altered LAG-3 expression in T cells from subjects with RRMS. This is supported by prior studies in MS in which lower LAG3 mRNA expression in PBMCs is associated with a more severe MS outcome and increased likelihood of progression to secondary progressive MS, whereas high LAG3 expression correlates with persisting RRMS and lower disability score up to 10 y later (24). In T1D, the role of transcription in altered LAG-3 surface expression was less clear, as LAG3 transcript was only significantly decreased in CD8 T cells after activation. Altered expression of LAG3 in T cells has been shown to be the result of epigenetic dysregulation in some patients with cancer, suggesting a possible explanation for the lower frequency of LAG-3+ T cells (2527). In addition, although LAG3 is part of an IR gene module that is coregulated under certain conditions (3), we did not observe altered PDCD1, HAVCR2, or TIGIT mRNA expression in resting T cells, suggesting that LAG3 is uniquely dysregulated in subjects with T1D and RRMS. Although we did not observe a decrease in the mRNA expression of HAVCR2 in subjects with MS, a functional defect in TIM-3 immunoregulation has previously been reported to be reversed by treatment with glatiramer acetate or IFN-β (28, 29). It is possible that our findings are influenced by the makeup of our cohort in which 80% of the subjects were taking DMT at the time of sample collection. Moreover, the altered relationship between PDCD1 and LAG3 in CD4 T cells may be indicative of epigenetic modifications, causing a targeted change in LAG3 regulation.

LAG-3 cleavage is mediated by ADAM10 and ADAM17 and is associated with the inhibitory function of T cells (9). We examined the possibility that altered LAG-3 cleavage may contribute to the decrease in LAG-3+ T cells from subjects with RRMS and T1D. Differences in expression of these sheddases were limited to ADAM10, which was lower in T cells from subjects with RRMS and T1D. Furthermore, T cells from subjects with RRMS and T1D shed significantly less sLAG3 after activation, which appeared to be driven by the frequency of LAG-3+ T cells and not by altered cleavage. We concluded that ADAM10 is functional in its ability to cleave surface LAG-3 in these subjects, as inhibition of ADAM10 and ADAM17 decreased the concentration of sLAG3 similarly in all subjects. Interestingly, inhibiting ADAM10 reduced the concentration of sLAG3 significantly more than the pan-metalloproteinase inhibitor, which is less effective at blocking ADAM10; this suggests ADAM10 is primarily responsible for the cleavage of sLAG3 from T cells in response to TCR stimulation. This observation corroborates recent findings that LAG-3 surface expression on CD4+Foxp3 T cells was modulated by ADAM10-mediated cell surface shedding (10). We were surprised to observe decreased but functional ADAM10 in T cells from subjects with low LAG-3 in whom cleavage was not perturbed; future studies may explore the possibility of coregulation between these two surface molecules.

Our functional studies implicate LAG-3 in the regulation of T cell proliferation and apoptosis in autoimmune disease. T cells from individuals with T1D and RRMS have been shown to have altered responses to activation and regulation. There is broad consensus that LAG-3 plays a role in regulating the expansion of activated T cells (3032). Early research posited that LAG-3 was a negative regulator of T cell activation and function with the observation that blocking LAG-3 on CD4 T cell clones enhanced both proliferation and Th1 cytokine production (33). In assessing whether T cell LAG-3 was functional in subjects with RRMS and how lower LAG-3 may alter the response to activation, we demonstrated that LAG-3 blockade resulted in enhanced proliferation in a similar respect in HCs and those with RRMS. Although we used only a small cohort for these experiments, it is interesting that Ag-stimulated CD8 T cells from subjects with RRMS not prescribed DMT proliferated less than those from HCs, both before and after LAG-3 inhibition. This may be due to underlying or acquired defects from chronic in vivo stimulation and independent of LAG-3 expression. Future studies may explore if dysregulated LAG-3 expression is also present in autoreactive cells that recognize CNS Ag. Our functional assays explored apoptosis in addition to proliferation as they relate to LAG-3 expression and inhibition. Subjects with RRMS exhibited both fewer cCasp3+ T cells and lower expression of cCasp3 on those cCasp3+ T cells that correlated with LAG-3; cCasp3 was further decreased after LAG-3 blockade. Taken together, these data suggest low LAG-3 expression may be linked to the decrease in cell death observed in T cells from subjects with RRMS and possibly also subjects with T1D. Certain DMTs for RRMS have been shown to induce activation-induced cell death (18, 34), and the resistance of T cells to Fas-mediated apoptosis has been implicated in both the pathogenesis and prognosis of MS (17, 35, 36). We observed a modest correlation between Fas and LAG-3 expression in resting T cells. Blocking LAG-3 decreased the expression of Fas on T cells from both healthy and off-therapy subjects with RRMS, further underscoring the association between LAG-3 expression and Fas-induced cell death. Our studies were generally not powered to examine the role of therapy in subjects with RRMS and include subjects both on and off therapy. Future studies could explore the role of LAG-3 expression in the context of immunosuppressive therapies, particularly in light of a recent preprint reporting the upregulation of LAG-3 in the presence of IFN-β, which is also a first-line treatment in MS (T. S. Sumida, S. Dulberg, J. Schupp, H. A. Stillwell, P.-P. Axisa, M. Comi, M. Lincoln, A. Unterman, N. Kaminski, A. Madi, et al., manuscript posted on bioRxiv, DOI: 10.1101/2020.10.30.362947).

In human autoimmunity, targeting LAG-3 therapeutically could function to selectively downmodulate activated T cells (32). Yet the efficacy of such approaches will depend in part on the function and expression of LAG-3 on autoreactive T cells being targeted by these therapies. Our findings indicate that LAG-3 is dysregulated in RRMS and T1D. This LAG-3 dysregulation may contribute to the escape of autoreactive T cells in autoimmunity by evading apoptosis. We find that the dominant source of this dysregulation is a decrease in transcription rather than cleavage of the expressed protein. Thus, targeting LAG-3 therapeutically in RRMS and T1D should take into account the source of dysregulation. This study further endorses the need to study human autoimmune samples to develop an in depth understanding of the regulation and expression of therapeutic targets in the context of disease.

We thank the Benaroya Research Institute Clinical Core Laboratory, in particular T. Nguyen and acknowledge the efforts of research assistants David Kook, Jenna Snavely, and Kim Varner in the Translational Research Program and the Diabetes Clinical Research Program in addition to Sylvia Posso for sample selection. We also thank A. Hocking and V. Green for assistance in editing and all participants in the Benaroya Research Institute Registry and Repository.

This work was supported by the National Institutes of Health (National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant R01 AI132774 to J.H.B.).

B.E.J. and J.H.B. conceived the project and designed the experiments. B.E.J. performed the experiments, analyzed the data, prepared the figures, and wrote the manuscript. M.D.M. assisted in conducting experiments and reviewed the article. H.T.B. was the consulting statistician. A.S. developed the technique for intracellular LAG-3 staining by flow cytometry, provided guidance on experimental design, and reviewed the article. L.H.M. and C.S. are the clinical principal investigators responsible for Neurology and Diabetes sample repositories. J.H.B. oversaw the project and reviewed the article.

The online version of this article contains supplemental material.

Abbreviations used in this article

cCasp3

cleaved caspase 3

DMT

disease-modifying therapy

EAE

experimental autoimmune encephalomyelitis

gMFI

geometric mean fluorescence intensity

HC

healthy control subject

IR

inhibitory receptor

LAG-3

lymphocyte activation gene-3

LAG-3NC

noncleavable lymphocyte activation gene-3

MS

multiple sclerosis

RRMS

relapsing-remitting multiple sclerosis

sLAG3

soluble LAG-3

TAPI-1

TNF-α protease inhibitor-1

T1D

type 1 diabetes

1.
National Institute of Environmental Health Sciences
.
2020
.
NIEHS Autoimmune Diseases Fact Sheet: Autoimmune Diseases and Your Environment.
.
2.
Joller
N.
,
V. K.
Kuchroo
.
2017
.
Tim-3, Lag-3, and TIGIT.
Curr. Top. Microbiol. Immunol.
410
:
127
156
.
3.
Chihara
N.
,
A.
Madi
,
T.
Kondo
,
H.
Zhang
,
N.
Acharya
,
M.
Singer
,
J.
Nyman
,
N. D.
Marjanovic
,
M. S.
Kowalczyk
,
C.
Wang
, et al
2018
.
Induction and transcriptional regulation of the co-inhibitory gene module in T cells.
Nature
558
:
454
459
.
4.
Bristol Myers Squibb
.
2021
.
Bristol Myers Squibb Announces LAG-3-Blocking Antibody Relatlimab and Nivolumab Fixed-Dose Combination Significantly Improves Progression-Free Survival vs. Opdivo (nivolumab) in Patients with Previously Untreated Metastatic or Unresectable Melanoma.
.
5.
Huang
C.
,
H.-X.
Zhu
,
Y.
Yao
,
Z.-H.
Bian
,
Y.-J.
Zheng
,
L.
Li
,
H. M.
Moutsopoulos
,
M. E.
Gershwin
,
Z.-X.
Lian
.
2019
.
Immune checkpoint molecules. Possible future therapeutic implications in autoimmune diseases.
J. Autoimmun.
104
:
102333
.
6.
Dougan
M.
,
M.
Pietropaolo
.
2020
.
Time to dissect the autoimmune etiology of cancer antibody immunotherapy.
J. Clin. Invest.
130
:
51
61
.
7.
Garcia
C. R.
,
R.
Jayswal
,
V.
Adams
,
L. B.
Anthony
,
J. L.
Villano
.
2019
.
Multiple sclerosis outcomes after cancer immunotherapy.
Clin. Transl. Oncol.
21
:
1336
1342
.
8.
Kadowaki
A.
,
S.
Miyake
,
R.
Saga
,
A.
Chiba
,
H.
Mochizuki
,
T.
Yamamura
.
2016
.
Gut environment-induced intraepithelial autoreactive CD4(+) T cells suppress central nervous system autoimmunity via LAG-3.
Nat. Commun.
7
:
11639
.
9.
Li
N.
,
Y.
Wang
,
K.
Forbes
,
K. M.
Vignali
,
B. S.
Heale
,
P.
Saftig
,
D.
Hartmann
,
R. A.
Black
,
J. J.
Rossi
,
C. P.
Blobel
, et al
2007
.
Metalloproteases regulate T-cell proliferation and effector function via LAG-3.
EMBO J.
26
:
494
504
.
10.
Andrews
L. P.
,
A.
Somasundaram
,
J. M.
Moskovitz
,
A. L.
Szymczak-Workman
,
C.
Liu
,
A. R.
Cillo
,
H.
Lin
,
D. P.
Normolle
,
K. D.
Moynihan
,
I.
Taniuchi
, et al
2020
.
Resistance to PD1 blockade in the absence of metalloprotease-mediated LAG3 shedding.
Sci. Immunol.
5
:
eabc2728
.
11.
Krienke
C.
,
L.
Kolb
,
E.
Diken
,
M.
Streuber
,
S.
Kirchhoff
,
T.
Bukur
,
Ö.
Akilli-Öztürk
,
L. M.
Kranz
,
H.
Berger
,
J.
Petschenka
, et al
2021
.
A noninflammatory mRNA vaccine for treatment of experimental autoimmune encephalomyelitis.
Science
371
:
145
153
.
12.
Polman
C. H.
,
S. C.
Reingold
,
G.
Edan
,
M.
Filippi
,
H. P.
Hartung
,
L.
Kappos
,
F. D.
Lublin
,
L. M.
Metz
,
H. F.
McFarland
,
P. W.
O’Connor
, et al
2005
.
Diagnostic criteria for multiple sclerosis: 2005 revisions to the “McDonald Criteria”.
Ann. Neurol.
58
:
840
846
.
13.
Motulsky
H. J.
,
R. E.
Brown
.
2006
.
Detecting outliers when fitting data with nonlinear regression - a new method based on robust nonlinear regression and the false discovery rate.
BMC Bioinformatics
7
:
123
.
14.
Li
N.
,
C. J.
Workman
,
S. M.
Martin
,
D. A.
Vignali
.
2004
.
Biochemical analysis of the regulatory T cell protein lymphocyte activation gene-3 (LAG-3; CD223).
J. Immunol.
173
:
6806
6812
.
15.
Woo
S. R.
,
N.
Li
,
T. C.
Bruno
,
K.
Forbes
,
S.
Brown
,
C.
Workman
,
C. G.
Drake
,
D. A.
Vignali
.
2010
.
Differential subcellular localization of the regulatory T-cell protein LAG-3 and the coreceptor CD4.
Eur. J. Immunol.
40
:
1768
1777
.
16.
Bae
J.
,
S. J.
Lee
,
C.-G.
Park
,
Y. S.
Lee
,
T.
Chun
.
2014
.
Trafficking of LAG-3 to the surface on activated T cells via its cytoplasmic domain and protein kinase C signaling.
J. Immunol.
193
:
3101
3112
.
17.
Durelli
L.
,
L.
Conti
,
M.
Clerico
,
D.
Boselli
,
G.
Contessa
,
P.
Ripellino
,
B.
Ferrero
,
P.
Eid
,
F.
Novelli
.
2009
.
T-helper 17 cells expand in multiple sclerosis and are inhibited by interferon-β.
Ann. Neurol.
65
:
499
509
.
18.
Boziki
M.
,
R.
Lagoudaki
,
P.
Melo
,
F.
Kanidou
,
C.
Bakirtzis
,
I.
Nikolaidis
,
E.
Grigoriadou
,
T.
Afrantou
,
T.
Tatsi
,
S.
Matsi
,
N.
Grigoriadis
.
2019
.
Induction of apoptosis in CD4(+) T-cells is linked with optimal treatment response in patients with relapsing-remitting multiple sclerosis treated with Glatiramer acetate.
J. Neurol. Sci.
401
:
43
50
.
19.
Petelin
Z.
,
V.
Brinar
,
D.
Petravic
,
N.
Zurak
,
K.
Dubravcic
,
D.
Batinic
.
2004
.
CD95/Fas expression on peripheral blood T lymphocytes in patients with multiple sclerosis: effect of high-dose methylprednisolone therapy.
Clin. Neurol. Neurosurg.
106
:
259
262
.
20.
Elmore
S.
2007
.
Apoptosis: a review of programmed cell death.
Toxicol. Pathol.
35
:
495
516
.
21.
Grebinoski
S.
,
D. A.
Vignali
.
2020
.
Inhibitory receptor agonists: the future of autoimmune disease therapeutics?
Curr. Opin. Immunol.
67
:
1
9
.
22.
Gertel
S.
,
A.
Polachek
,
V.
Furer
,
D.
Levartovsky
,
O.
Elkayam
.
2021
.
CD4+ LAG-3+ T cells are decreased in active psoriatic arthritis patients and their restoration in vitro is mediated by TNF inhibitors.
Clin. Exp. Immunol.
206
:
173
183
.
23.
Bettini
M.
,
A. L.
Szymczak-Workman
,
K.
Forbes
,
A. H.
Castellaw
,
M.
Selby
,
X.
Pan
,
C. G.
Drake
,
A. J.
Korman
,
D. A.
Vignali
.
2011
.
Cutting edge: accelerated autoimmune diabetes in the absence of LAG-3.
J. Immunol.
187
:
3493
3498
.
24.
Lavon
I.
,
C.
Heli
,
L.
Brill
,
H.
Charbit
,
A.
Vaknin-Dembinsky
.
2019
.
Blood levels of co-inhibitory-receptors: a biomarker of disease prognosis in multiple sclerosis.
Front. Immunol.
10
:
835
.
25.
Klümper
N.
,
D. J.
Ralser
,
E. G.
Bawden
,
J.
Landsberg
,
R.
Zarbl
,
G.
Kristiansen
,
M.
Toma
,
M.
Ritter
,
M.
Hölzel
,
J.
Ellinger
,
D.
Dietrich
.
2020
.
LAG3 (LAG-3, CD223) DNA methylation correlates with LAG3 expression by tumor and immune cells, immune cell infiltration, and overall survival in clear cell renal cell carcinoma.
J. Immunother. Cancer
8
:
e000552
.
26.
Sasidharan Nair
V.
,
H.
El Salhat
,
R. Z.
Taha
,
A.
John
,
B. R.
Ali
,
E.
Elkord
.
2018
.
DNA methylation and repressive H3K9 and H3K27 trimethylation in the promoter regions of PD-1, CTLA-4, TIM-3, LAG-3, TIGIT, and PD-L1 genes in human primary breast cancer.
Clin. Epigenetics
10
:
78
.
27.
Fröhlich
A.
,
J.
Sirokay
,
S.
Fietz
,
T. J.
Vogt
,
J.
Dietrich
,
R.
Zarbl
,
M.
Florin
,
P.
Kuster
,
G.
Saavedra
,
S. R.
Valladolid
, et al
2020
.
Molecular, clinicopathological, and immune correlates of LAG3 promoter DNA methylation in melanoma.
EBioMedicine
59
:
102962
.
28.
Koguchi
K.
,
D. E.
Anderson
,
L.
Yang
,
K. C.
O’Connor
,
V. K.
Kuchroo
,
D. A.
Hafler
.
2006
.
Dysregulated T cell expression of TIM3 in multiple sclerosis.
J. Exp. Med.
203
:
1413
1418
.
29.
Yang
L.
,
D. E.
Anderson
,
J.
Kuchroo
,
D. A.
Hafler
.
2008
.
Lack of TIM-3 immunoregulation in multiple sclerosis.
J. Immunol.
180
:
4409
4414
.
30.
Maçon-Lemaître
L.
,
F.
Triebel
.
2005
.
The negative regulatory function of the lymphocyte-activation gene-3 co-receptor (CD223) on human T cells.
Immunology
115
:
170
178
.
31.
Lichtenegger
F. S.
,
M.
Rothe
,
F. M.
Schnorfeil
,
K.
Deiser
,
C.
Krupka
,
C.
Augsberger
,
M.
Schlüter
,
J.
Neitz
,
M.
Subklewe
.
2018
.
Targeting LAG-3 and PD-1 to enhance T cell activation by antigen-presenting cells.
Front. Immunol.
9
:
385
.
32.
Angin
M.
,
C.
Brignone
,
F.
Triebel
.
2020
.
A LAG-3-specific agonist antibody for the treatment of T cell-induced autoimmune diseases.
J. Immunol.
204
:
810
818
.
33.
Huard
B.
,
M.
Tournier
,
T.
Hercend
,
F.
Triebel
,
F.
Faure
.
1994
.
Lymphocyte-activation gene 3/major histocompatibility complex class II interaction modulates the antigenic response of CD4+ T lymphocytes.
Eur. J. Immunol.
24
:
3216
3221
.
34.
Lopatinskaya
L.
,
J.
Zwemmer
,
B.
Uitdehaag
,
K.
Lucas
,
C.
Polman
,
L.
Nagelkerken
.
2006
.
Mediators of apoptosis Fas and FasL predict disability progression in multiple sclerosis over a period of 10 years.
Mult. Scler.
12
:
704
709
.
35.
Comi
C.
,
M.
Leone
,
S.
Bonissoni
,
S.
DeFranco
,
F.
Bottarel
,
C.
Mezzatesta
,
A.
Chiocchetti
,
F.
Perla
,
F.
Monaco
,
U.
Dianzani
.
2000
.
Defective T cell fas function in patients with multiple sclerosis.
Neurology
55
:
921
927
.
36.
Okuda
Y.
,
B. R.
Apatoff
,
D. N.
Posnett
.
2006
.
Apoptosis of T cells in peripheral blood and cerebrospinal fluid is associated with disease activity of multiple sclerosis.
J. Neuroimmunol.
171
:
163
170
.

The authors have no financial conflicts of interest.

Supplementary data