Clinical Impact of a LAG3 Single-Nucleotide Polymorphism in Relapsed, Refractory DLBCL Patients Treated with Glofitamab

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,

Kindly review the concerns outlined in the attached document and address them thoroughly in the revised manuscript.

Please ensure that all points are carefully considered and that appropriate revisions and clarifications are incorporated.

Comments for author File: Comments.pdf

Author Response

Dear Authors,

The submitted manuscript represents the first study evaluating the impact of LAG3 and CTLA4 single-nucleotide polymorphisms in relapsed/refractory DLBCL patients treated with glofitamab. The research addresses a clinically relevant question by exploring predictive biomarkers for bispecific antibody therapy. The study is supported by a clear biological rationale, highlighting the mechanistic link between LAG3-mediated immune regulation and T-cell-engaging therapies. Clinical endpoints, including progression-free survival (PFS) and overall survival (OS), are well defined and appropriately analyzed. A dose-dependent genetic effect across LAG3 genotypes further strengthens the biological plausibility of the findings. Ethical approval and informed consent are clearly documented, ensuring regulatory compliance. However, despite its strengths, the manuscript has several important limitations that substantially weaken the robustness of the conclusions;

1. The very small overall sample size (n = 28), together with extremely small genotype subgroups (some containing only 3-6 patients), severely limits statistical power and increases the risk of spurious associations. The retrospective, single-center design introduces potential selection bias and limits generalizability.
2. The absence of a control cohort prevents differentiation between prognostic and predictive biomarker effects.
3. Imbalances in baseline characteristics, particularly sex distribution and prior CAR-T therapy exposure may confound survival outcomes.
4. The wide confidence intervals observed in multivariable analyses (HRs ~14–16) suggest instability of effect estimates and possible overfitting.

Response: We thank the reviewer for the helpful comments. We have addressed the concerns in the revised discussion. This study had several limitations. The retrospective single-center study design and the small sample size with only a few patients in the genetic subgroups may have introduced selection bias and affected the generalizability. Due to the small sample size the investigation is explorative, including the risk of false positive and false negative results. Furthermore, due to the lack of a control cohort, it was not possible to differentiate between prognostic and predictive value. Certain imbalances in the baseline characteristics, particularly the different prior treatments with CAR T-cell therapy, may also influence the outcome. Only 33% of patients in the I455hom subgroup had received CAR-T cell therapy, whereas more than 85% of patients in the T455hom and the I455het subgroup were treated with CAR-T cell therapy before glofitamab treatment. Also 83% of the I455hom subgroup were men, whereas in the other genetic subgroups men and women were equally represented, indicating an imbalance which may have influenced the results. Furthermore, as most patients were not only diagnosed with R/R DLBCL but were also multi-drug resistant and refractory to CAR-T therapy, the results of this study cannot be generalized to R/R DLBCL. Objective parameters were assessed, and a multivariate analysis was performed to evaluate the potential influence of confounding factors on the observed outcomes. The multivariate analysis identified the I455hom genotype as the most significant predictor of both PFS and OS, suggesting that genetic characteristics of the inhibitory cell-surface receptor may have greater prognostic relevance than established clinical factors such as sex, age or number of prior therapies.

5. The lack of functional or mechanistic validation further weakens biological interpretation of the SNP findings.

Response: Mechanistic studies will be required to analyze the projected differential functions of the LAG3 variant proteins. This sentence was added in the conclusions.

6. Multiple comparisons were performed without adjustment for type I error.

Response: We respectfully note that our cohort of 28 patients represents the first exploratory investigation of Glofitamab response in R/R DLBCL according to LAG3 genotype. Applying formal multiple testing corrections in this setting would further reduce already limited statistical power, increasing the likelihood of failing to detect clinically meaningful differences. Following recommendations from Rothman (Rothman KJ. No adjustments are needed for multiple comparisons. Epidemiology. 1990 Jan;1(1):43-6. PMID: 2081237.) and Althouse (Althouse AD. Adjust for Multiple Comparisons? It's Not That Simple. Ann Thorac Surg. 2016 May;101(5):1644-5. doi: 10.1016/j.athoracsur.2015.11.024. PMID: 27106412.), we have chosen to report unadjusted p-values. However, we have added a sentence into the discussion describing this limitation.”

7. The conclusion appears overstated, as the designation of LAG3 polymorphism as a “prognostic response marker” is premature without independent validation in a larger, prospective cohort.

Response: To validate this hypothesis, further investigations and larger, prospective studies are required. This sentence was added in the conclusions.

Remarks: Kindly address these concerns before it is accepted for publication.

Response: We feel that we have incorporated all the suggestions thereby gaining relevant insights and improving the quality and integrity of the manuscript to a considerable degree.

Reviewer 2 Report

Comments and Suggestions for Authors The manuscript Clinical Impact of a LAG3 Single-Nucleotide Polymorphism in Relapsed, Refractory DLBCL Patients Treated with Glofitamab by Maeva Ullmann et al. assesses the effects of LAG3 rs870849 and CTLA4 rs231775 SNPs on clinical outcomes in R/R DLBCL patients treated with glofitamab, identifying a significant correlation between LAG3 genotypes and patients’ PFS and OS, with no evident clinical impact of CTLA4 genotypes observed.   This retrospective single-center observational study adopts standardized methodologies for treatment regimens, endpoint definition and genetic testing. The covariates included in the multivariate analysis (age, sex, prior CAR-T therapy, transplantation history, etc.) cover the key prognostic factors for R/R DLBCL, ensuring basic scientific rigor.   However, the study is limited by the small sample size of the single-center design, leading to baseline imbalances in the study population and inadequate stratification and adjustment of confounding factors. As noted in the study background, LAG3 and CTLA4 polymorphisms affect CAR-T efficacy, yet 75% of enrolled patients received glofitamab after CAR-T failure. Subgroup analysis shows only 33% of LAG3 I455 homozygous patients had a CAR-T history, significantly lower than the T455 homozygous (86%) and I455T heterozygous (87%) subgroups (p=0.029). CAR-T failure itself impacts subsequent glofitamab outcomes, and its distribution is markedly imbalanced with LAG3 genotypes. While "prior CAR-T therapy" was included as a covariate in the multivariate analysis, no interaction analysis between this factor and LAG3 genotypes was performed, nor stratified analysis to exclude CAR-T history’s interference with the LAG3 genotype-clinical outcome association, leaving the baseline imbalance unresolved. In addition, the LAG3 I455 homozygous subgroup (n=6) and CTLA4 A17 homozygous subgroup (n=3) suffer from extremely low statistical power. Small sample sizes increase the risk of false positive/negative results in log-rank tests and multivariate Cox regression; the latter’s reliability and stability are further compromised by reduced degrees of freedom from multiple covariates. Moreover, the study used chi-square or Fisher’s exact tests for categorical variables but did not report correction for multiple comparisons, raising the risk of Type Ⅰ errors. I have some suggestions for the discussion section and conclusion. 1, Highlight the impacts of small sample size and subgroup imbalance on the reliability of results and conclusions, as well as baseline bias from the imbalanced distribution of CAR-T history and genotypes. Emphasize that most patients were multi-drug resistant and CAR-T-refractory (not simply R/R DLBCL), so the findings should not be extrapolated to all R/R DLBCL populations. 2, Cite basic research literature to elaborate on how LAG3 polymorphisms regulate T cell function (especially in TCE therapy) and thus drive differences in clinical efficacy, enhancing the depth and scientific merit of the discussion.

Author Response

Comments and Suggestions for Authors

The manuscript Clinical Impact of a LAG3 Single-Nucleotide Polymorphism in Relapsed, Refractory DLBCL Patients Treated with Glofitamab by Maeva Ullmann et al. assesses the effects of LAG3 rs870849 and CTLA4 rs231775 SNPs on clinical outcomes in R/R DLBCL patients treated with glofitamab, identifying a significant correlation between LAG3 genotypes and patients’ PFS and OS, with no evident clinical impact of CTLA4 genotypes observed.   This retrospective single-center observational study adopts standardized methodologies for treatment regimens, endpoint definition and genetic testing. The covariates included in the multivariate analysis (age, sex, prior CAR-T therapy, transplantation history, etc.) cover the key prognostic factors for R/R DLBCL, ensuring basic scientific rigor.   However, the study is limited by the small sample size of the single-center design, leading to baseline imbalances in the study population and inadequate stratification and adjustment of confounding factors. As noted in the study background, LAG3 and CTLA4 polymorphisms affect CAR-T efficacy, yet 75% of enrolled patients received glofitamab after CAR-T failure. Subgroup analysis shows only 33% of LAG3 I455 homozygous patients had a CAR-T history, significantly lower than the T455 homozygous (86%) and I455T heterozygous (87%) subgroups (p=0.029). CAR-T failure itself impacts subsequent glofitamab outcomes, and its distribution is markedly imbalanced with LAG3 genotypes. While "prior CAR-T therapy" was included as a covariate in the multivariate analysis, no interaction analysis between this factor and LAG3 genotypes was performed, nor stratified analysis to exclude CAR-T history’s interference with the LAG3 genotype-clinical outcome association, leaving the baseline imbalance unresolved. In addition, the LAG3 I455 homozygous subgroup (n=6) and CTLA4 A17 homozygous subgroup (n=3) suffer from extremely low statistical power. Small sample sizes increase the risk of false positive/negative results in log-rank tests and multivariate Cox regression; the latter’s reliability and stability are further compromised by reduced degrees of freedom from multiple covariates. Moreover, the study used chi-square or Fisher’s exact tests for categorical variables but did not report correction for multiple comparisons, raising the risk of Type Ⅰ errors.

Response: We thank the reviewer for the helpful comments. We have addressed the concerns in the revised discussion.

I have some suggestions for the discussion section and conclusion.

1, Highlight the impacts of small sample size and subgroup imbalance on the reliability of results and conclusions, as well as baseline bias from the imbalanced distribution of CAR-T history and genotypes. Emphasize that most patients were multi-drug resistant and CAR-T-refractory (not simply R/R DLBCL), so the findings should not be extrapolated to all R/R DLBCL populations.

Response: This study had several limitations. The retrospective single-center study design and the small sample size with only a few patients in the genetic subgroups may have introduced selection bias and affected the generalizability. Due to the small sample size the investigation is explorative, including the risk of false positive and false negative results. Furthermore, due to the lack of a control cohort, it was not possible to differentiate between prognostic and predictive value. Certain imbalances in the baseline characteristics, particularly the different prior treatments with CAR T-cell therapy, may also influence the out-come. Only 33% of patients in the I455hom subgroup had received CAR-T cell therapy, whereas more than 85% of patients in the T455hom and the I455het subgroup were treated with CAR-T cell therapy before glofitamab treatment. Also 83% of the I455hom subgroup were men, whereas in the other genetic subgroups men and women were equally represented, indicating an imbalance which may have influenced the results. Furthermore, as most patients were not only diagnosed with R/R DLBCL but were also multi-drug resistant and refractory to CAR-T therapy, the results of this study cannot be generalized to R/R DLBCL. Objective parameters were assessed, and a multivariate analysis was performed to evaluate the potential influence of confounding factors on the observed outcomes. The multivariate analysis identified the I455hom genotype as the most significant predictor of both PFS and OS, suggesting that genetic characteristics of the inhibitory cell-surface receptor may have greater prognostic relevance than established clinical factors such as sex, age or number of prior therapies.

2, Cite basic research literature to elaborate on how LAG3 polymorphisms regulate T cell function (especially in TCE therapy) and thus drive differences in clinical efficacy, enhancing the depth and scientific merit of the discussion.

Response: LAG3 polymorphism may regulate T cell function by altering the inhibitory signaling capacity, thereby modulating T Cell activation, proliferation and differentiation. LAG3 receptor localizes to the immunological synapse and associates with the T cell receptor (TCR)-CD3 complex in CD4+ and CD8+ T cells [24–26]. Variants that reduce LAG3 expression result in lower inhibitory signaling at the immunological synapse, leading to enhanced T-cell activation and increased proliferation [27]. Polymorphisms altering LAG3 expression or function may influence the threshold for T cell activation, given that LAG3 suppresses T cell activation by disrupting TCR-CD3 signaling [24]. In allogenic hematopoietic stem cell transplantation, LAG3 polymorphism in donor cells can enhance T cell reactivity and lead to severe Graft-versus-host-Disease (GVHD) [27,28]. The rs870849 snp causes a substitution of isoleucine to threonine (I455T) within the LAG3 transmembrane domain affecting protein processing and proteolytic cleavage [28]. In the R/R DLBCL LAG3 I455hom carriers glofitamab treatment triggered an abortive T-cell activation accompanied by excessive inflammatory cytokine production but did not induce sustained cytolytic activity necessary for tumor destruction, a combination of low efficacy and high toxicity, characteristic of an abnormal immune response. This biochemical disorder may result from decreased cleavage of the LAG3 molecule at the cell surface, leading to T cell dysfunction.

 

Submission Date 06 February 2026

Date of this review 01 Mar 2026 15:36:03

Date of revision: 9 March 2026

Reviewer 3 Report

Comments and Suggestions for Authors

Thank the Editor and authors for the opportunity to review this well-written and interesting paper, which examines clinical genomic determinants of therapeutic response to Glofitamab in R/R-DLBC and the pivotal role of the Lymphocyte-activation gene 3 (LAG3) rs870849 Single-Nucleotide Polymorphism. Although in theory, a standard immunochemotherapy regimen can cure 60% of DLBCL patients, in practice, treating R/R patients remains a major clinical challenge. However, with the arrival of T-cell-engaging therapies such as CAR-T cell therapy and BsAb therapy, the landscape of R/R DLBCL therapy has changed dramatically. Among the newly introduced therapies, glofitamab is a first-in-class BsAb with a unique 2:1 configuration that allows bivalently binding to CD20 and monovalently binding to CD3 on endogenous T cells. Although glofitamab is a significant therapy in this field, there is a marked difference in clinical response rates among R/R DLBCL patients, suggesting that factors such as the patient's immune system play a crucial role in clinical outcomes. In this direction, immune checkpoint receptors LAG3 and T-lymphocyte-associated protein 4 (CTLA4) are crucial for T cell fitness. In this respect, glofitamab, by binding to two CD20 molecules, has been found to produce a more potent interaction with lymphomas, particularly in cases with low antigen expression. Additionally, monovalent binding to CD3 is strategically designed to avoid non-specific activation of T cells. In addition, LAG3 is a critical immune checkpoint receptor that supports immune homeostasis by suppressing T cell activation. To my knowledge, this study, conducted at the University Hospital of Bern, is the first to investigate the impact of LAG3 and CTLA4 polymorphisms on the clinical outcomes of glofitamab therapy in R/R DLBCL patients. The study included 28 patients with R/R DLBCL, 75% of whom had received CAR-T cell therapy before the study. Patients were divided into three groups depending on the LAG3 polymorphism variant allele, T455hom, I455Thet, and I455hom. The study found that the LAG3 polymorphism, particularly the rs870849 genotype, is a significant predictor of patient survival after receiving glofitamab therapy. Patients who are T455 homozygotes have the best outcomes, while those who are I455 homozygotes have the worst outcomes, where all the patients who are I455hom have progressive disease as the best response, showing that the patients are resistant to glofitamab therapy. On the contrary, patients with T455hom have durable response rates, with a median PFS of 19 months. The intermediate response is due to the accumulation of genetic effects, where each T allele contributes to the T cell engager's capacity to fight the patient's lymphoma. It is noteworthy that the multivariate analysis confirmed that the LAG3 genotype, namely the I455hom genotype, is the most significant prognostic factor for PFS and OS, with HRs>14 and >16, respectively. This shows that the genetic makeup of the host cell-surface receptor on the inhibitory cell may hold the most critical prognostic clue, irrespective of conventional prognostic factors, such as age, sex, and the number of prior therapies. In addition to the LAG3 polymorphism, the glofitamab study also investigated the CTLA4 polymorphism at rs231775, which encodes a threonine-to-alanine substitution at amino acid seventeen of the N-terminal signal peptide of the CTLA4 receptor, affecting the regulation of early T-cell priming in lymph nodes. CTLA4 regulates the first activation of naïve T cells. This process is critical in CAR T cell therapy, in which the transferred T cells are expected to undergo significant expansion and differentiation within the patient's body. On the other hand, bispecific antibodies like glofitamab target endogenous T cells that are already primed in the tumor microenvironment. At this point, T cells are chronically stimulated by their antigen and are, to some extent, exhausted. In this regard, LAG3 is a hallmark of exhaustion, and its overexpression is a key cause of immune resistance in lymphomas. Therefore, the efficacy of glofitamab would be more influenced by LAG3-mediated suppressive signals than CTLA4-mediated priming signals. As is known, the safety profile of glofitamab includes immune-related adverse events, including cytokine release syndrome (CRS) and immune-effector cell-associated neurotoxicity syndrome (ICANS). In the Bern cohort, the incidence of CRS was 32%. Of these, 10% were high-grade CRS (Grade 3-4). When stratified by LAG3 status, the I455hom group had a higher incidence of CRS than the I455Thet and T455hom groups. This is the paradox the study's authors observed: the I455hom group had the worst efficacy and the worst toxicity profile. These results are characteristic of an abnormal immune response and show that in I455hom carriers, glofitamab triggers an abortive T-cell activation accompanied by excessive inflammatory cytokine production (leading to CRS) but does not induce sustained cytolytic activity necessary for tumor destruction. This biochemical disorder may result from decreased cleavage of the LAG3 molecule at the cell surface, leading to T cell dysfunction. Furthermore, analysis of CTLA4 subgroups showed that CRS was most prevalent in the T17hom subgroup (55%), although this was not statistically associated with clinical outcomes. This once again emphasizes the intricacies of modulating immune checkpoints in which a single nucleotide variation may influence drug efficacy and toxicity in different pathways. Thus, in conclusion, the clinical results of the Bern study are encouraging and call for the clinical use of personalized immunotherapy for R/R DLBCL patients. In fact, the use of germline SNPs as risk factors may enable early identification of 'high-risk' R/R DLBCL patients, for example, those with LAG3 I455H homozygous genotype. For these 'high-risk' patients, combination therapy with glofitamab may prevent early disease progression and improve survival outcomes. Thus, although future prospective studies are needed to confirm this genomic marker and prove a standard framework for the use of this and other emerging therapies to maximize the curative potential of patients with aggressive DBCL, this paper, acceptable in its present form, significantly adds to this field.

Author Response

Comments and Suggestions for Authors

Thank the Editor and authors for the opportunity to review this well-written and interesting paper, which examines clinical genomic determinants of therapeutic response to Glofitamab in R/R-DLBC and the pivotal role of the Lymphocyte-activation gene 3 (LAG3) rs870849 Single-Nucleotide Polymorphism. Although in theory, a standard immunochemotherapy regimen can cure 60% of DLBCL patients, in practice, treating R/R patients remains a major clinical challenge. However, with the arrival of T-cell-engaging therapies such as CAR-T cell therapy and BsAb therapy, the landscape of R/R DLBCL therapy has changed dramatically. Among the newly introduced therapies, glofitamab is a first-in-class BsAb with a unique 2:1 configuration that allows bivalently binding to CD20 and monovalently binding to CD3 on endogenous T cells. Although glofitamab is a significant therapy in this field, there is a marked difference in clinical response rates among R/R DLBCL patients, suggesting that factors such as the patient's immune system play a crucial role in clinical outcomes. In this direction, immune checkpoint receptors LAG3 and T-lymphocyte-associated protein 4 (CTLA4) are crucial for T cell fitness. In this respect, glofitamab, by binding to two CD20 molecules, has been found to produce a more potent interaction with lymphomas, particularly in cases with low antigen expression. Additionally, monovalent binding to CD3 is strategically designed to avoid non-specific activation of T cells. In addition, LAG3 is a critical immune checkpoint receptor that supports immune homeostasis by suppressing T cell activation. To my knowledge, this study, conducted at the University Hospital of Bern, is the first to investigate the impact of LAG3 and CTLA4 polymorphisms on the clinical outcomes of glofitamab therapy in R/R DLBCL patients. The study included 28 patients with R/R DLBCL, 75% of whom had received CAR-T cell therapy before the study. Patients were divided into three groups depending on the LAG3 polymorphism variant allele, T455hom, I455Thet, and I455hom. The study found that the LAG3 polymorphism, particularly the rs870849 genotype, is a significant predictor of patient survival after receiving glofitamab therapy. Patients who are T455 homozygotes have the best outcomes, while those who are I455 homozygotes have the worst outcomes, where all the patients who are I455hom have progressive disease as the best response, showing that the patients are resistant to glofitamab therapy. On the contrary, patients with T455hom have durable response rates, with a median PFS of 19 months. The intermediate response is due to the accumulation of genetic effects, where each T allele contributes to the T cell engager's capacity to fight the patient's lymphoma. It is noteworthy that the multivariate analysis confirmed that the LAG3 genotype, namely the I455hom genotype, is the most significant prognostic factor for PFS and OS, with HRs>14 and >16, respectively. This shows that the genetic makeup of the host cell-surface receptor on the inhibitory cell may hold the most critical prognostic clue, irrespective of conventional prognostic factors, such as age, sex, and the number of prior therapies. In addition to the LAG3 polymorphism, the glofitamab study also investigated the CTLA4 polymorphism at rs231775, which encodes a threonine-to-alanine substitution at amino acid seventeen of the N-terminal signal peptide of the CTLA4 receptor, affecting the regulation of early T-cell priming in lymph nodes. CTLA4 regulates the first activation of naïve T cells. This process is critical in CAR T cell therapy, in which the transferred T cells are expected to undergo significant expansion and differentiation within the patient's body. On the other hand, bispecific antibodies like glofitamab target endogenous T cells that are already primed in the tumor microenvironment. At this point, T cells are chronically stimulated by their antigen and are, to some extent, exhausted. In this regard, LAG3 is a hallmark of exhaustion, and its overexpression is a key cause of immune resistance in lymphomas. Therefore, the efficacy of glofitamab would be more influenced by LAG3-mediated suppressive signals than CTLA4-mediated priming signals. As is known, the safety profile of glofitamab includes immune-related adverse events, including cytokine release syndrome (CRS) and immune-effector cell-associated neurotoxicity syndrome (ICANS). In the Bern cohort, the incidence of CRS was 32%. Of these, 10% were high-grade CRS (Grade 3-4). When stratified by LAG3 status, the I455hom group had a higher incidence of CRS than the I455Thet and T455hom groups. This is the paradox the study's authors observed: the I455hom group had the worst efficacy and the worst toxicity profile. These results are characteristic of an abnormal immune response and show that in I455hom carriers, glofitamab triggers an abortive T-cell activation accompanied by excessive inflammatory cytokine production (leading to CRS) but does not induce sustained cytolytic activity necessary for tumor destruction. This biochemical disorder may result from decreased cleavage of the LAG3 molecule at the cell surface, leading to T cell dysfunction. Furthermore, analysis of CTLA4 subgroups showed that CRS was most prevalent in the T17hom subgroup (55%), although this was not statistically associated with clinical outcomes. This once again emphasizes the intricacies of modulating immune checkpoints in which a single nucleotide variation may influence drug efficacy and toxicity in different pathways. Thus, in conclusion, the clinical results of the Bern study are encouraging and call for the clinical use of personalized immunotherapy for R/R DLBCL patients. In fact, the use of germline SNPs as risk factors may enable early identification of 'high-risk' R/R DLBCL patients, for example, those with LAG3 I455H homozygous genotype. For these 'high-risk' patients, combination therapy with glofitamab may prevent early disease progression and improve survival outcomes. Thus, although future prospective studies are needed to confirm this genomic marker and prove a standard framework for the use of this and other emerging therapies to maximize the curative potential of patients with aggressive DBCL, this paper, acceptable in its present form, significantly adds to this field.

Submission Date 06 February 2026

Date of this review 22 Feb 2026 11:16:45

 

Response: We thank the reviewer for the elaborate comments and suggestions.  We feel that we have incorporated all the suggestions thereby gaining relevant insights and improving the quality and integrity of the manuscript to a considerable degree.

Date of revision 9 March 2026.