You are here

  • Home
  • Archive
  • Volume 12, Issue 8
  • Targeting TNFRSF25 by agonistic antibodies and multimeric TL1A proteins co-stimulated CD8+ T cells and inhibited tumor growth

Article Text

Article menu
Basic tumor immunology
Original research
Targeting TNFRSF25 by agonistic antibodies and multimeric TL1A proteins co-stimulated CD8+ T cells and inhibited tumor growth
  1. Xueyuan Lyu1,2,
  2. Linlin Zhao1,2,
  3. Sijia Chen2,
  4. Yulu Li2,
  5. Yajing Yang2,
  6. Huisi Liu2,
  7. Fang Yang2,
  8. Wenhui Li2,3 and
  9. Jianhua Sui2,3
  1. 1Peking University-Tsinghua University-National Institute of Biological Sciences Joint Graduate Program, School of Life Sciences, Tsinghua University, Beijing, China
  2. 2National Institute of Biological Sciences, Beijing, China
  3. 3Tsinghua Institute of Multidisciplinary Biomedical Research, Tsinghua University, Beijing, China
  1. Correspondence to Dr Jianhua Sui; suijianhua{at}nibs.ac.cn

Abstract

Background Tumor necrosis factor receptor superfamily 25 (TNFRSF25) is a T-cell co-stimulatory receptor. Expression of its ligand, TNF-like cytokine 1A (TL1A), on mouse tumor cells has been shown to promote tumor regression. This study aimed to develop TNFRSF25 agonists (both antibodies (Abs) and TL1A proteins) and to investigate their potential antitumor effects.

Methods Anti-mouse TNFRSF25 (mTNFRSF25) Abs and multimeric TL1A proteins were generated as TNFRSF25 agonists. Their agonism was assessed in luciferase reporter and T-cell co-stimulation assays, and their antitumor effects were evaluated in syngeneic mouse tumor models. TNFRSF25 expression within the tumor microenvironment and the effects of an anti-mTNFRSF25 agonistic Ab on tumor-infiltrating T cells were evaluated by flow cytometry. Cell depletion assays were used to identify the immune cell types that contribute to the antitumor effect of the anti-mTNFRSF25 Ab. The Fc gamma receptor (FcγR) dependence of TNFRSF25 agonists was assessed in an in vivo T-cell expansion model and a mouse tumor model using Fc variants and FcγR-deficient mice.

Results TNFRSF25 agonists exhibited antitumor effects in syngeneic mouse tumor models without causing observed side effects. We identified an anti-mTNFRSF25 agonistic Ab, 1A6-m1, which exhibited greater antitumor activity than a higher affinity anti-TNFRSF25 Ab which engages an overlapping epitope with 1A6-m1. 1A6-m1 activated CD8+ T cells and antigen-specific T cells, leading to tumor regression; it also induced long-term antitumor immune memory. Although activating TNFRSF25 by 1A6-m1 expanded splenic regulatory T (Treg) cells, it did not influence intratumoral Treg cells. Moreover, 1A6-m1’s antitumor effects required the engagement of both inhibitory FcγRIIB and activating FcγRIII. Replacing 1A6-m1’s CH1-hinge region with that of human IgG2 (h2) conferred enhanced antitumor effects. Finally, we also generated multimeric human and mouse TL1A fusion proteins as TNFRSF25 agonists, and they co-stimulated CD8+ T cells and reduced tumor growth, even in the absence of Fc-FcγR interactions.

Conclusion Our data demonstrates the potential of activating TNFRSF25 by Abs and multimeric TL1A proteins for cancer immunotherapy and provides insights into their development as

therapeutics.

  • Antibody
  • co-stimulatory molecules
  • Immunotherapy

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/jitc-2024-008810

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Activating tumor necrosis factor receptor superfamily 25 (TNFRSF25) signaling co-stimulates T cells, and expression of TNF-like cytokine 1A (TL1A) on mouse tumor cells has been shown to promote tumor regression.

    WHAT THIS STUDY ADDS

    • Targeting TNFRSF25 by agonistic antibodies (Abs) and multimeric TL1A proteins co-stimulates antigen-specific T cells and exerts antitumor effects with a good safety profile. The antitumor effects of the agonistic Abs are influenced by Ab affinity, CH1-hinge flexibility, and Fc-Fc gamma receptor (FcγR) interactions. Multimeric TL1A proteins exert potent antitumor effects, even in the absence of Fc-FcγR interactions.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • Findings from this study provide a rationale for further investigation of anti-TNFRSF25 agonistic Abs and multimeric TL1A proteins as therapeutics for cancer immunotherapy.

    Background

    Antagonistic antibodies (Abs) against inhibitory immune checkpoints, such as programmed cell death protein 1 (PD-1)/programmed cell death ligand and cytotoxic T lymphocyte antigen 4, elicit antitumor immune responses and have shown remarkable clinical benefits against multiple types of cancer.1 Additionally, the antitumor immune responses of T cells are also regulated by co-stimulatory receptors. Agonistic targeting of co-stimulatory receptors, such as 4-1BB, glucocorticoid-induced tumor necrosis factor receptor (TNFR)-related protein (glucocorticoid-induced TNFR-related protein (GITR)), OX40, CD27, CD40, and TNFR2 of the TNFR superfamily (TNFRSF), has been considered as a potential approach for driving effective antitumor immunity.2

    Tumor necrosis factor receptor superfamily 25 (TNFRSF25) (death receptor 3), a member of the TNFRSF, is a T-cell co-stimulatory receptor.3 4 Its ligand, TNF-like cytokine 1A (TL1A), is a trimeric type II transmembrane protein of the TNFSF. Previous studies of patients with rheumatoid arthritis, psoriasis, asthma, or inflammatory bowel disease have reported that both TL1A and TNFRSF25 expression are upregulated in affected areas and correlate with disease severity.5–8 Studies using mouse models of these diseases have found that activating TNFRSF25 signaling by agonists (Abs and TL1A proteins) aggravates disease progression; conversely, blocking it by anti-TL1A Abs or knockout of TL1A or TNFRSF25 attenuates established inflammation.3 9–13 These findings demonstrate a pro-inflammatory role of TL1A/TNFRSF25 signaling. Moreover, phase II clinical trials have revealed the efficacy and safety of anti-TL1A Abs in treating ulcerative colitis.14

    TNFRSF25 is known to be expressed on CD4+ T cells, CD8+ T cells, and regulatory T (Treg) cells.10 15 16 Previous studies have confirmed that TNFRSF25 agonists can expand Treg cells and inhibit allograft rejection in mouse models of organ transplantation and hematopoietic cell transplantation (HCT).17 18 Activating TNFRSF25 has also been shown to promote CD4+ and CD8+ T-cell activation.3 19 Moreover, expression of TL1A on mouse tumor cells has been found to inhibit tumor growth in a CD8+ T cell-dependent manner,19 suggesting TNFRSF25 as a potential target for cancer treatment.

    In this study, we developed an anti-mTNFRSF25 agonistic Ab (1A6-m1) and found it inhibited tumor growth in syngeneic mouse tumor models. Additionally, 1A6-m1 induced stronger CD8+ T-cell activation and conferred greater antitumor effects than a higher affinity anti-TNFRSF25 Ab that recognizes an overlapping epitope with 1A6-m1. In tumor-bearing mice, 1A6-m1 co-stimulated antigen-specific CD8+ T cells, promoted the activation of intratumoral CD8+ T cells (but not the proliferation of intratumoral Treg cells), and elicited durable immune memory. Experiments using the syngeneic melanoma model revealed that 1A6-m1’s antitumor activity required the engagement of both inhibitory FcγRIIB and activating FcγRIII, and replacing 1A6-m1’s CH1-hinge region with that of human IgG2 (h2) led to improved antitumor effects. Finally, we developed multimeric human and mouse TL1A fusion proteins as agonists and found that these proteins co-stimulated CD8+ T cells and inhibited tumor growth. Thus, our study demonstrates the potential of using Abs or multimeric ligands as TNFRSF25 agonists for cancer immunotherapy. It also illustrates that TNFRSF25 agonistic Abs can exert antitumor effects by co-stimulating antigen-specific CD8+ T cells and engaging FcγRs, including both inhibitory FcγRIIB and activating FcγRIII.

    Methods

    Cell lines

    Chinese hamster ovary (CHO) cells were sourced from the Cell Bank of Type Culture Collection (Chinese Academy of Sciences). A20, B16F10, EL4, E.G7-OVA, and Jurkat cells were sourced from the American Type Culture Collection. FreeStyle 293-F cells were from Life Technologies. B16F10-OVA cells were generously provided by Dr Yu Zhang (National Institute of Biological Sciences, Beijing (NIBS)). MC38 cells were generously provided by BeiGene, China. All cells were cultured in accordance with the recommended conditions or following the instructions provided by the respective suppliers.

    Mice

    All animals were housed in a specific pathogen-free facility in compliance with the National Guidelines for Housing and Care of Laboratory Animals in China. The animal experiments were conducted following the approved protocols of the Institutional Animal Care and Use Committee of NIBS (NIBS2018M0039). C57BL/6JNifdc and BALB/c mice were procured from Beijing Vital River Laboratory. OTI mice were provided by Dr Liang Chen (NIBS). Fcgr2b−/− mice were obtained from the Jackson Laboratory, and Fcer1g−/− mice were purchased from Shanghai Model Organisms. FcγR-null mice (Fcer1g−/−Fcgr2b−/−) were generated through the breeding of Fcer1g−/− mice and Fcgr2b−/− mice.

    Expression and purification of proteins

    The extracellular domain (ECD) of human TNFRSF25 (hTNFRSF25) and mTNFRSF25 were fused to a C-terminal His (× 6)-Avi-Tag. The hFc-hTL1A fusion proteins were generated by adding an N-terminal human IgG1 (h1) Fc (hFc) tag to human TL1A (hTL1A) ECD (L72-L251). The hFc-TNC-TL1A fusion proteins were generated by inserting a chicken tenascin C (TNC) tag (amino acids 110–139: ACGCAAAPDIKDLLSRLEELEGLVSSLREQ) between the hFc tag and TL1A ECD. The CH1-hinge-engineered h2/m1 isotype was generated by replacing the CH1-hinge region of mouse IgG1 (m1) with that of h2. The variable heavy chain (VH) and variable light chain (VL) gene sequences of IgG Abs were subcloned into VH and VL expression vectors, respectively.20 The fusion proteins and Abs were produced by transient transfection of FreeStyle 293-F cells and purified through affinity chromatography. Purified proteins were analyzed by size exclusion chromatography (SEC) on a Superose 6 Increase 10/300 GL column (Cytiva, GE29-0915-96).

    Generation of anti-TNFRSF25 Abs

    For anti-mTNFRSF25 Ab panning, the ECD of mTNFRSF25 was biotinylated by BirA ligase and captured on Dynabeads M-280 Streptavidin beads (Life Technologies, 11205D), and then used for panning against a human naïve phage display Ab library.21 Following two rounds of selection, approximately 600 single clones were randomly picked and screened for binding to mTNFRSF25. Subsequently, the chosen clones were converted into full-length Abs in the h1 isotype for analysis. The VH and VL gene sequences of the anti-mTNFRSF25 Ab 4C12 were synthesized according to the published amino acid sequence.10 The anti-hTNFRSF25 Ab 4F4 was produced in-house, using conventional hybridoma technology.

    Surface plasmon resonance analysis

    Kinetic analysis of Abs binding to mTNFRSF25 and FcγRs was carried out using a Biacore T200 instrument (GE HealthCare) at 25°C. Anti-human IgG Ab, anti-mouse IgG Ab, or protein A/G was covalently immobilized on a CM5 sensor chip (Cytiva, BR100399) using an Amine Coupling Kit (Cytiva, BR100050). All Abs analyzed were captured at similar levels on the chip, and the analytes were then injected at twofold serially diluted concentrations. The binding kinetic parameters were determined by fitting the sensorgrams to a 1:1 binding model using Biacore T200 evaluation software.

    ELISA

    For ELISA-based binding assays, round-bottom immuno 96-well plates (Thermo Fisher Scientific, 449824) were coated with 2 µg/mL of streptavidin (Sigma-Aldrich, SA101) at 4°C overnight. Biotin-labeled TNFRSF25 ECD protein at a concentration of 2 µg/mL was captured onto the coated plates by incubating at 30°C for 1 hour. Abs serially diluted in phosphate-buffered saline (PBS) containing 2% non-fat milk were added to each well. The binding was detected by HRP-conjugated anti-human IgG Fc Ab (Invitrogen, 31413) at 30°C for 1 hour. Afterward, the signal was developed using a TMB substrate (Sigma-Aldrich, 87748) for 5–10 min at room temperature and stopped by adding 2 M H2SO4. The colorimetric absorbance was measured at 450 nm using an Imark Microplate Absorbance Reader (Bio-Rad).

    For ELISA-based competition assays, streptavidin and mTNFRSF25 ECD protein were coated as previously described. Different Abs at serially diluted concentrations mixed with 0.4 nM of Abs or protein were added to the ELISA plates to compete for binding with the captured mTNFRSF25 ECD protein.

    For ELISA-based detection of mouse interferon (IFN)-γ, 2 µg/mL anti-IFN-γ Ab R4-6A2 (Bio X Cell, BE0054) was coated at 4°C overnight. The culture medium of cultured T cells was diluted in 1% bovine serum albumin/PBS and incubated at room temperature for 1 hour. Biotin-labeled anti-IFN-γ Ab XMG1.2 (BioLegend, 505803) was added as a detection Ab and incubated at room temperature for 1 hour. The binding was subsequently detected using HRP-Streptavidin (Sigma-Aldrich, RABHRP3) at room temperature for 1 hour.

    For detection of the mouse IgGs against the h2’s CH1-hinge region, a control Ab with the h2 isotype (5 µg/mL) was immobilized on a plate as an antigen. Serum samples were twofold serially diluted starting from an initial 1:250 dilution and added to the immobilized ELISA plate. The binding was subsequently detected using HRP-anti-mouse IgG (Invitrogen, 31439).

    NF-κB reporter assay

    Jurkat-nuclear factor kappa B (NF-κB) luciferase reporter stable cell line was constructed using single-cell cloning and antibiotics Blasticidin (Invivogen, ant-bl-05) selection. We transfected the chimeric mTNFRSF25 construct, expressing the hTNFRSF25 intracellular signaling domain, and full-length hTNFRSF25 into Jurkat-NF-κB luciferase cells. This process resulted in the creation of Jurkat-mTNFRSF25/NF-κB luciferase reporter cells and Jurkat-hTNFRSF25/NF-κB luciferase reporter cells, respectively. To assess NF-κB activation, reporter cells were incubated with Abs at 37°C for 8 hours. Subsequently, the level of NF-κB activation was quantified by measuring luciferase activity using Bright-Glo Luciferase Assay reagents (Promega, E2620).

    In vitro T-cell co-stimulation assay

    Mouse CD8+ T cells were isolated from splenocytes using a MojoSort Mouse CD8 T Cell Isolation Kit (BioLegend, 480008). We stimulated purified cells with 2 µg/mL of plate-bound anti-CD3 Ab 145–2 C11 and the tested proteins for 3 days before analysis. Unpurified splenocytes from OTI mice were stimulated with 10 pM OVA257-264 peptide (Sigma-Aldrich, S7951) and tested proteins for 3 days. The culture medium for mouse T cells consisted of Roswell Park Memorial Institute 1640 supplemented with 10% heat-inactivated fetal bovine serum, 0.05 mM 2-ME, 2 mM L-glutamine, and 100 U/mL penicillin and 100 mg/mL streptomycin.

    Human CD3+ T cells were isolated from human peripheral blood mononuclear cells (PBMCs) (Shanghai Aoneng Biotechnology, PB3003F-C, China) using an EasySep Human T Cell Enrichment Kit (STEMCELL, 17951). Purified cells were stained with carboxyfluorescein succinimidyl ester (CFSE) (Invitrogen, C34554), stimulated with 2 µg/mL of plate-bound anti-CD3 Ab OKT-3 and the tested proteins, and cultured in X-VIVO 15 Serum-free Hematopoietic Cell Medium (Lonza, 02-053Q). The medium was supplemented with 5% heat-inactivated fetal bovine serum, 2 mM L-glutamine, and 100 U/mL penicillin and 100 mg/mL streptomycin.22

    Adoptive T-cell transfer assay

    For OTI cell expansion, splenocytes from OTI mice were activated with 10 nM OVA257-264 peptides and 50 IU/mL IL-2 (3SBio, Inleusin) for 48 hours. The cells were subcultured with 50 IU/mL interleukin (IL)-2 for another 48 hours following a previously described protocol.23 To assess the ability of Abs to induce OTI cell expansion, 1×106 OTI cells were intravenously transfused into mice 1 day before the intraperitoneal (i.p.) injection of 100 µg OVA protein and 100 µg Abs. Mice were bled at indicated time points and the proportion of OTI cells was assessed by flow cytometry.

    Tumor challenge and treatment

    In syngeneic mouse tumor models, 6–10-week-old female mice were inoculated subcutaneous with A20, MC38, B16F10-OVA, or E.G7-OVA cells on the right flank. OTI cells were intravenously transfused by tail vein as indicated in each figure. Mice were grouped based on age and tumor volume before receiving i.p. injection of Abs 1 day after OTI cell injection. Tumor dimensions were measured with an electronic caliper, and tumor volume was calculated using the formula (length×width2/2. Mice were euthanized when tumor volumes reached a maximum of 2,000 mm3 or if animal health was compromised.

    For tumor rechallenge studies, gender-matched and age-matched naive mice and cured mice were inoculated with 2×105 tumor cells in the left flank without any other treatment.

    For the B16F10-OVA lung metastasis model, mice were intravenously infused with 3×105 B16F10-OVA cells via the tail vein on day 0, followed by an intravenous infusion of 4×105 OTI cells on day 3. The lungs were harvested on day 14 and analyzed for the presence of metastatic foci using a dissecting microscope.

    Hematotoxicity and hepatotoxicity measurements

    B16F10-OVA tumor-bearing mice were treated with Abs or protein as indicated in a germane figure. Hematological analyses and serum alanine aminotransferase (ALT) levels were carried out at Vital River Labs (Beijing, China). Liver pathology was assessed by H&E staining by Beijing BrightShines (Beijing, China).

    Flow cytometry

    In the binding assay, EL4 cells stably transduced with TNFRSF25 (EL4-TNFRSF25) were incubated with Abs at 4°C for 30 min. Subsequently, they were stained with PE-labeled goat anti-human IgG Fc (eBioscience, 12-4998-82) as a secondary Ab. To detect TNFRSF25 expression, human PBMCs underwent blocking with 10% human serum at 4°C for 15 min, followed by incubation with a biotinylated anti-hTNFRSF25 Ab 4F4-hDANG (the h1 isotype with D265A and N297G mutations to abolish binding to all FcγRs24) and staining with PE streptavidin (BioLegend, 405203). Single-cell suspensions of mouse spleens and tumors were pretreated with Fc blockade (anti-CD16/32 Ab, 2.4G2) at 4°C for 15 min, stained with an anti-mTNFRSF25 Ab (4C12-hDANG), and incubated with PE-labeled goat anti-human IgG Fc as a secondary Ab. For epitope mapping, CHO cells were transiently transduced with TNFRSF25 variants with a green fluorescent protein (GFP) tag. Positive cells expressing GFP were identified, and the binding of Abs to TNFRSF25 variants was detected using PE-labeled goat anti-human IgG Fc.

    For intratumoral immune cell analysis, tumors were excised and digested with 200 IU/mL collagenase type I (Gibco, 17018029) and 60 mg/mL DNaseI (Roche, 4536282001) at 37°C for 20 min following a previously described protocol.25 Cells were then passed through a 70 µm cell strainer and used for assays. The detection of OTI cells involved the use of anti-mouse T-cell receptor (TCR) Vα2 (B20.1, eBioscience) and anti-mouse TCR Vβ5.1/5.2 (MR9-4, BioLegend), as previously described.26 For intracellular cytokine staining, cells were activated for 4 hours with Cell Activation Cocktail (with Brefeldin A) (BioLegend, 423303) and treated with a Cytofix/Cytoperm Fixation/Permeabilization Kit (BD Biosciences, 554714) according to the manufacturer’s instructions. For nuclear transcription factor staining, cells were treated with Foxp3 Transcription Factor Staining Buffer Set (eBioscience, 00-5523-00).

    The reagents used for flow cytometry were as follows: anti-mouse CD45 (30-F11), anti-mouse CD3 (17A2), anti-mouse CD8 (53–6.7), anti-mouse CD4 (GK1.5), anti-mouse NK1.1 (PK136), anti-mouse CD19 (6D5), anti-mouse/human CD11b (M1/70), anti-mouse CD11c (N418), anti-mouse Ly6G (1A8), anti-mouse Ly6C (HK1.4), anti-mouse F4/80 (BM8), anti-mouse MHC-II (M5/114.15.2), anti-mouse IL-17A (TC11-18H10.1), anti-mouse FcγRIIB Ab (S17012B), anti-mouse FcγRIII Ab (S17014E), anti-mouse perforin (S16009A), anti-mouse/human CD44 (IM7), anti-mouse CD62L (MEL-14), anti-mouse IFN-γ (XMG1.2), anti-human/mouse granzyme B (GzmB) (QA18A28), anti-human CD8 (HIT8a), anti-human CD56 (5.1H11), anti-human CD19 (HIB19), anti-human CD3 (HIT3a), and anti-human Foxp3 (259D) from BioLegend; anti-human CD4 (OKT-4), anti-Foxp3 (FJK-16s), and Fixable Near-IR Dead Cell Stain Kit (L10119) from Thermo Fisher Scientific; DAPI (D9542) from Sigma-Aldrich.

    All samples were analyzed on BD FACSAria III, BD FACSAria Fusion, or BD FACS LSRFortessa cytometers (BD Biosciences). Data were analyzed with BD FACSDiva (BD Biosciences) and FlowJo software (TreeStar).

    Statistical analysis

    Graphs and statistical analyses were conducted using Prism V.9 software (GraphPad Software). The Student’s t-test (unpaired, two-tailed, 95% CI) was employed to compare two groups with one variable. Statistical comparison between multiple groups was determined by either one-way analysis of variance (ANOVA) (one variable) or two-way ANOVA (two variables). For tumor growth analyses, data were analyzed by two-way ANOVA followed by Tukey’s multiple comparisons test. For mice survival analyses, data were analyzed by log-rank (Mantel-Cox) tests. Significance levels were denoted as follows: *p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001.

    Results

    Anti-TNFRSF25 agonistic Abs in combination with adoptive T-cell transfer inhibit tumor growth in mice

    Given that expression of TL1A on tumor cells has been shown to inhibit tumor growth in syngeneic mouse tumor models,19 27 28 we explored the direct targeting of TNFRSF25 by agonistic Abs in vivo. Two anti-mTNFRSF25 Abs were assessed for their binding activities, epitopes, and agonistic activities: one is 1A6, we selected from a naïve phage display Ab library, and the other one is 4C12, an Ab reported in previous studies that was generated by immunizing Armenian hamsters.10

    We used flow cytometry and ELISA assays to measure the binding activity of these two Abs (with the human IgG1 isotype, h1) to mTNFRSF25 and found that 4C12-h1 had a higher binding activity than 1A6-h1 (figure 1A and online supplemental figure 1A). Subsequently, surface plasmon resonance (SPR) analysis revealed that 4C12-h1 (KD=5.60 nM) had about sevenfold higher binding affinity than 1A6-h1 (KD=39.73 nM) and that the dissociation rate (Kd) of 4C12-h1 (Kd=7.07 × 10−4 s−1) was about ninefold slower than 1A6-h1 (Kd=6.12 × 10−3 s−1) (figure 1B).

    Supplemental material

    Figure 1
    Figure 1

    Anti-TNFRSF25 agonistic Abs in combination with adoptive T-cell transfer inhibit tumor growth in mice. (A) Flow cytometry determined the binding of two Abs to EL4-mTNFRSF25 cells. (B) Surface plasmon resonance sensorgrams depicted the binding of two Abs to mTNFRSF25 ECD protein. The table shows calculated affinity constants (ka, kd, and KD). (C) 1A6 competed with 4C12 for mTNFRSF25 binding. mTNFRSF25 ECD protein was incubated with 4C12 (0.4 nM) in the presence of 1A6 at serially diluted concentrations. The binding of 4C12 to mTNFRSF25 was assessed by ELISA. The competition activity of 1A6 was expressed as the percentage of inhibition for 4C12 binding to mTNFRSF25. (D) The NF-κB signaling induced by Abs. Jurkat-mTNFRSF25/NF-κB luciferase reporter cells, incubated with each Ab for 8 hours, and the relative luminescence units (RLU) were measured and are presented as fold changes relative to controls without Ab incubation. (E) The IFN-γ secretion of CD8+ T cells induced by Abs. Purified mouse CD8+ T cells were activated with 2 µg/mL plate-bound anti-CD3 in the presence of each Ab for 3 days. The concentration of IFN-γ in the culture medium was measured by ELISA. (F) Schematic illustration of tumor cell inoculation and treatment in a B16F10-OVA mouse tumor model. (G) Tumor growth and survival curves of mice treated as illustrated in the panel F. n=5–7/group. Data are presented as means±SEM. The data shown are representative of two or three independent experiments. Ab, antibody; ECD, extracellular domain; IFN, interferon; i.p., intraperitoneal; MFI, mean fluorescence intensity; mTNFRSF25, mouse tumor necrosis factor receptor super family 25; NF-κB, nuclear factor kappa B; RU, response units; s.c., subcutaneous; TNFRSF25, tumor necrosis factor receptor superfamily 25.

    The ECD of TNFRSF25 comprises four cysteine-rich domains (CRDs), among which CRD1 is divided into the A and B modules.29 To map the binding epitopes of these two Abs, we constructed several truncated TNFRSF25 variants lacking CRDs and human/mouse TNFRSF25 chimeras with CRD1 or the A module replaced by its human/mouse TNFRSF25 counterpart (online supplemental figure 1B). Flow cytometry analysis revealed that, similar to a mouse TL1A (mTL1A) fusion protein (hFc-TNC-mTL1A, see below), both 1A6-h1 and 4C12-h1 recognized epitopes in CRD1. Distinct from 4C12-h1, which bound to both hTNFRSF25 and mTNFRSF25, 1A6-h1 only bound to mTNFRSF25, engaging an epitope spanning both the A and B modules of CRD1 (online supplemental figure 1B). A competitive ELISA assay showed that both 1A6-h1 and 4C12-h1 blocked mTL1A binding with mTNFRSF25 (online supplemental figure 1C) and that these two Abs cross-competed with each other for mTNFRSF25 binding (figure 1C and online supplemental figure 1D), findings indicating that they recognize an overlapping epitope in CRD1.

    Previous studies reported that activating TNFRSF25 results in the activation of the transcription factor NF-κB,30 so we established Jurkat-mTNFRSF25/NF-κB luciferase reporter cells and examined NF-κB activation on exposure to the Abs. 1A6-h1 and 4C12-h1 each triggered NF-κB signaling, doing so in a dose-dependent manner. Interestingly, 1A6-h1 with relatively lower affinity induced higher levels of NF-κB signaling than 4C12-h1 (figure 1D). A CD8+ T-cell co-stimulation assay in which mouse naïve CD8+ T cells were stimulated with anti-CD3 Ab and treated with each Ab followed by ELISA-based quantification of IFN-γ production by CD8+ T cells showed that both 1A6-h1 and 4C12-h1 induced IFN-γ production in a dose-dependent manner and that 1A6-h1 induced a greater extent of IFN-γ production than 4C12-h1 (figure 1E). These results suggest that 1A6 and 4C12 activate TNFRSF25 signaling and co-stimulate CD8+ T cells and that the relatively lower affinity Ab 1A6 induces greater agonism than 4C12.

    To investigate the agonism of 1A6 and 4C12 in vivo, they were isotype-switched from h1 to m1 to enable efficient engagement with mouse FcγRs. Consistent with the aforementioned data using the h1 isotype Abs, we found that 1A6-m1 had a lower affinity but greater NF-κB signaling activity and T-cell co-stimulatory activity than 4C12-m1 (online supplemental figure 1E–H). To assess their antitumor activities, we used a syngeneic B16F10-OVA (a derivative of B16F10 melanoma that expresses OVA) tumor model. Despite administrating 1A6-m1 twice a week and initiating treatment on day 4 after tumor cell inoculation to ensure a sustained high concentration of antibodies in vivo and early intervention, 1A6-m1 did not confer antitumor effects in this model (online supplemental figure 1I). B16F10-OVA tumor-bearing mice that received adoptive transfer of OVA257-264-specific OTI cells and i.p. injection of a control Ab showed rapid tumor growth; whereas combining OTI cells with 1A6-m1 or 4C12-m1 both significantly inhibited tumor growth and significantly prolonged the mouse survival (figure 1F,G). It was worth noting that while 1A6-m1 conferred superior antitumor effects compared with 4C12-m1, no significant survival benefit was observed (figure 1G). Taken together, these results demonstrate that anti-mTNFRSF25 agonistic Abs 1A6-m1 and 4C12-m1 in combination with adoptive T-cell transfer can inhibit tumor growth and that 1A6-m1 has greater agonism and antitumor effects than 4C12-m1. Thus, we selected 1A6-m1 for further investigation.

    1A6-m1 inhibits tumor growth by enhancing intratumoral CD8+ T-cell function

    To identify the immune cell types that contribute to 1A6-m1’s antitumor effects, we analyzed the expression pattern of TNFRSF25. We first analyzed PBMCs and mouse spleen cells using flow cytometry and found that CD3+ T cells had the highest TNFRSF25 expression among the examined immune cells (online supplemental figure 2A,B). We then examined TNFRSF25 expression in tumor tissues. Mice received OTI cells on day 4 following B16F10-OVA cell inoculation, and single-cell suspensions of tumors and spleens were analyzed when the tumors reached a diameter of 5–8 mm. TNFRSF25 was detected on both intratumoral and splenic CD4+ (including CD4+Foxp3+ and CD4+Foxp3−) T cells, CD8+ T cells, and a minority of natural killer (NK) cells (5–10%); and among T cells, its expression on intratumoral CD8+ T cells and Treg cells was relatively lower compared with its expression on splenic CD8+ T cells and Treg cells, respectively (figure 2A). In addition, TNFRSF25 is not expressed on B16F10-OVA cells cultured in vitro, consistently, it was not detected on B16F10-OVA cells isolated from the implanted tumors (online supplemental figure 2C). Taken together, these results indicate that T cells are the likely contributors to 1A6-m1’s antitumor effects.

    Figure 2
    Figure 2

    1A6-m1 inhibits tumor growth by enhancing intratumoral CD8+ T-cell function. (A) Expression of mTNFRSF25 in the spleens and tumors. Mice (n=3/group) were s.c. inoculated with 5×105 B16F10-OVA cells on day 0, intravenously transfused with 1×106 OTI cells on day 4, and analyzed on day 22. Left: representative histograms, right: median fluorescence intensity of mTNFRSF25 expression. All T cells were gated within the CD45+ CD3+ population. 5 µg/mL anti-mTNFRSF25 Ab (4C12-hDANG) was used as the primary Ab, followed by 0.5 µg/mL PE-labeled goat anti-human IgG Fc as the secondary Ab. (B) The proportion of OTI cells in the blood and tumors. Mice were s.c. inoculated with 5×105 B16F10-OVA cells on day 0, intravenously transfused with 1×106 OTI cells on day 4, i.p. injected with 5 mg/kg Abs on days 5 and 9, and analyzed when the tumors reached 5–8 mm in diameter. Anti-mouse CD8α (clone: 53–6.7), anti-mouse TCR vα2 (clone: B20.1), and anti-mouse TCR Vβ5.1/5.2 (clone: MR9-4) were used to stain OTI cells. Left: representative dot plots, right: the proportion of OTI cells among CD8+ T cells. (C) A20 Tumor growth after 1A6-m1 treatment. BALB/c mice (n=5–6/group) were inoculated s.c. with 7×105 A20 cells on day 0 and i.p. injected with 5 mg/kg of Abs twice a week from day 5 as indicated. (D) Cytokine production of intratumoral CD8+ T cells in B16F10-OVA tumor-bearing mice. Cells were preactivated with Cell Activation Cocktail (with Brefeldin A) for 4 hours and assessed by intracellular cytokine staining (n=3–6/group). Mice were treated as illustrated in the panel B. (E) Cytokine production of intratumoral CD8+ T cells in A20 tumor-bearing mice. BALB/c mice (n=4–5/group) were inoculated s.c. with 1×106 A20 cells on day 0, i.p. injected with 5 mg/kg Abs on days 7 and 12, and analyzed for cytokine production on day 15. (F) A20 tumor growth after 1A6-m1 treatment with or without CD8+ depletion and IFN-γ blockade. 5 mg/kg Abs were i.p. injected twice a week from day 5 following 6×105 A20 cells inoculation (n=5–6/group). 200 µg anti-mouse CD8α Ab (clone: 2.43) or anti-IFN-γ Ab (clone: XMG1.2) was i.p. injected on days 5 and 12. (G) The proportion and number of Treg cells. Mice (n=5–6/group) were treated as illustrated in the panel B. Data are presented as means±SEM. The data shown are representative of two or three independent experiments. Abs, antibodies; IFN, interferon; i.p., intraperitoneal; s.c., subcutaneous; TNFRSF25, tumor necrosis factor receptor superfamily 25; TCR, T-cell receptor; Treg, regulatory T cells.

    To assess whether 1A6-m1 co-stimulates OTI cells, we used a co-stimulation assay in which splenocytes from OTI mice were activated with OVA256-264 peptide and Abs in vitro.19 1A6-m1 augmented IFN-γ secretion by OTI cells, doing so in a dose-dependent manner (online supplemental figure 2D). In B16F10-OVA tumor-bearing mice, 1A6-m1 treatment following OTI cell infusion induced a significant OTI cell expansion in the blood and an apparent increase in the proportion of OTI cells among CD8+ T cells in the tumors compared with the control Ab (figure 2B), suggesting that 1A6-m1 co-stimulates antigen-specific CD8+ T cells, leading to the inhibition of tumor growth.

    To assess 1A6-m1’s antitumor effects as a monotherapy in additional tumor models, three syngeneic tumor models: A20 (B-cell lymphoma), E.G7-OVA (a derivative of EL4 thymoma that expresses OVA), and MC38 (colon adenocarcinoma) were tested. 1A6-m1 did not inhibit tumor growth in the E.G7-OVA and MC38 tumor models (online supplemental figure 3E,F), whereas in the A20 model, 1A6-m1 significantly inhibited tumor growth (figure 2C). Similar to B16F10-OVA cells, A20 cells were also found to lack expression of TNFRSF25 (online supplemental figure 2C). Within the A20 tumors, like the B16F10-OVA model, we observed TNFSF25 expression on intratumoral T-cell subsets, including CD4+Foxp3+, CD4+Foxp3−, and CD8+ T cells (online supplemental figure 2G). Intracellular cytokine staining revealed that 1A6-m1 induced significantly elevated production of intracellular IFN-γ, GzmB, and perforin in intratumoral CD8+ T cells compared with the control Ab in both the B16F10-OVA and A20 tumor models (figure 2D,E). Depletion of CD8+ T cells using an anti-CD8α Ab (clone: 2.43) abolished 1A6-m1’s antitumor effects, as did blocking IFN-γ signaling with an anti-IFN-γ Ab (clone: XMG1.2) (figure 2F); depletion of NK cells using the anti-Asialo-GM1 Ab (clone: Poly21460) did not affect 1A6-m1’s antitumor effects (online supplemental figure 2H). These results collectively demonstrate that 1A6-m1 inhibits tumor growth in a CD8+ T cell-dependent and IFN-γ-dependent manner.

    Considering TNFRSF25’s role in promoting Treg cell and Th17 cell proliferation,11 31 we investigated 1A6-m1’s effects on these cells. We analyzed peripheral and intratumoral Treg cells in B16F10-OVA tumor-bearing mice. Compared with the control Ab, although 1A6-m1 treatment induced a significant increase in the proportion of Treg cells among CD4+ T cells in the spleens, there was no significant change in either the proportion or the number of Treg cells within the tumors (figure 2G), indicating that 1A6-m1 has the activity to expand peripheral Treg cells but does not expand intratumoral Treg cells. We also analyzed Th17 cells in B16F10-OVA tumor-bearing mice, 1A6-m1 significantly decreased the proportion of intratumoral Th17 cells among CD4+ T cells compared with the control Ab, and it did not affect the proportion of Th17 cells among CD4+ T cells in mesenteric lymph nodes (online supplemental figure 2I). Collectively, these results demonstrate that 1A6-m1 co-stimulates intratumoral CD8+ T cells and antigen-specific CD8+ T cells, leading to the inhibition of tumor growth.

    1A6-m1 promotes the development of immune memory

    To investigate whether 1A6-m1 treatment results in long-term immune memory, mice that had achieved complete regression of their B16F10-OVA tumors following the combined treatment of 1A6-m1 and OTI cells were rechallenged with a second inoculation of B16F10-OVA cells about 90 days after the initial tumor challenge (and in the absence of any treatment in the interim) (figure 3A). Compared with age-matched naïve mice, tumor growth in the rechallenged mice was significantly slower, and 80% of the rechallenged mice completely rejected B16F10-OVA cells, whereas all naive mice died from progressive tumor growth (figure 3B). Analysis of the immune cells of peripheral blood in these mice at 4 days after tumor inoculation showed that the rechallenged mice exhibited significantly higher proportions of central memory (CD44+ CD62L+) and effector memory (CD44+ CD62L−) CD4+ and CD8+ T cells than the naïve mice (figure 3C). These findings demonstrate that the survival mice develop immune memory.

    Figure 3
    Figure 3

    1A6-m1 promotes the development of immune memory. (A) Schematic of tumor rechallenge. Cured mice underwent rechallenge with 2×105 B16F10-OVA or B16F10 cells on the left flank approximately 90 days after the initial tumor cell inoculation, and subsequently received eight doses of 1A6-m1 administration (twice a week). Age-matched naïve mice served as controls. (B) Tumor growth and survival curves of B16F10-OVA rechallenged mice (n=5/group) treated as illustrated in the panel A. (C) Analysis of memory T cells after B16F10-OVA tumor rechallenge in the mice from panel B. The percentages of CD44+CD62L− and CD44+CD62L+ among CD4+ (left) and CD8+ T cells (right) in the blood were measured 4 days after B16F10-OVA tumor rechallenge (n=5/group). (D) Tumor growth and survival curves of B16F10 rechallenged mice (n=6–7/group) treated as illustrated in the panel A. Data are presented as means±SEM. The data shown are representative of two or three independent experiments. Abs, antibodies; i.v., intravenous; s.c., subcutaneous.

    We next evaluated whether 1A6-m1 treatment establishes immune memory against tumor antigens in addition to the OVA antigen by challenging cured mice with parental B16F10 tumor cells (which do not express OVA). The rechallenged mice showed significantly slower tumor growth of B16F10 tumors and significantly prolonged survival compared with the naive mice (figure 3D), indicating that 1A6-m1 treatment leads to immune memory against B16F10 tumor antigens beyond the OVA antigen. Taken together, these results demonstrate that 1A6-m1 promotes the development of immune memory responses, leading to the establishment of long-lasting antitumor immunity.

    Both FcγRIIB and FcγRIII are required for 1A6-m1’s antitumor effects

    Previous studies have shown that the activity of Abs targeting the co-stimulatory receptors in the TNFRSF relies on inhibitory FcγRIIB-mediated crosslinking.32 33 To investigate the contribution of Fc-Fc gamma receptor (FcγR) interactions to the in vivo activity of 1A6-m1, we used the OTI cell expansion model. Specifically, mice were adoptively transferred with OTI cells followed by the administration of OVA protein and Abs (figure 4A). Compared with the control Ab, 1A6-m1 significantly expanded OTI cells in WT mice, while it failed to expand OTI cells in FcγR-null mice (lacking all FcγRs) (figure 4B), indicating 1A6-m1’s in vivo agonism requires Fc-FcγR interactions. Consistently, we also found that a 1A6-m1 variant bearing a D265A mutation (1A6-m1-D265A) in its Fc region, which lacked all FcγR binding activity as measured by SPR (online supplemental figure 3A), did not expand OTI cells in WT mice (online supplemental figure 3B). These results demonstrate that FcγR engagement is required for A6-m1’s in vivo agonism.

    Figure 4
    Figure 4

    Both FcγRIIB and FcγRIII are required for 1A6-m1’s antitumor effects. (A) Schematic of OTI cell expansion model. Mice were intravenously transferred with 1×106 OTI cells and i.p. injected with 100 µg OVA protein and 100 µg Abs 1 day later. (B) Analysis of OTI cell expansion induced by 1A6-m1. The percentage of OTI cells among CD8+ T cells in mice (n=3–4/group) treated as illustrated in the panel A was measured by flow cytometry. (C) Schematic illustration of tumor cell inoculation and treatment in a B16F10-OVA mouse tumor model, involving WT, FcγR-null, Fcer1g−/−, or Fcgr2b−/− mice. (D–G) Tumor growth and survival curves in the B16F10-OVA mouse model. Mice were treated as illustrated in the panel C. WT mice (n=5/group) (D) FcγR-null mice (n=6–7/group) (E) Fcer1g−/− mice (n=10–11/group) (F) Fcgr2b−/− mice (n=6–7/group) (G). Data are presented as means±SEM. The data shown are representative of two or three independent experiments. Abs, antibodies; FcγR, Fc gamma receptor; i.p, intraperitoneal; s.c., subcutaneous.

    There are two types of FcγRs in mice, activating FcγRs (FcγRI, FcγRIII, and FcγRIV) and inhibitory FcγRIIB, and it is known that the m1 isotype Abs only bind to FcγRIIB and FcγRIII.34 Thus, we investigated the contribution of FcγRIIB and FcγRIII to the observed in vivo activity of 1A6-m1 by using Fcer1g−/− and Fcgr2b−/− mice. Fcer1g−/− mice lack the common Fc receptor γ chain, which is required for the assembly and surface expression of activating FcγRI, FcγRIII, and FcγRIV32; Fcgr2b−/− mice lack expression of inhibitory FcγRIIB. In Fcer1g−/− mice, the extent of 1A6-m1-mediated OTI expansion was significantly reduced compared with WT mice; in Fcgr2b−/− mice, 1A6-m1 failed to expand OTI cells (figure 4B). These results indicate that the engagement of both FcγRIIB and FcγRIII is required for 1A6-m1’s in vivo activity.

    We next evaluated whether Fc-FcγR interactions are required for 1A6-m1’s antitumor activity. Consistent with the observations in the aforementioned OTI expansion model, the antitumor activity of 1A6-m1 was abolished in FcγR-null mice bearing B16F10-OVA tumors (figure 4C–E). Such FcγR-dependent antitumor activity was confirmed by 1A6-m1-D265A which exhibited no antitumor activity in WT mice (online supplemental figure 3C). In Fcgr2b−/− and Fcer1g−/− mice bearing B16F10-OVA tumors, 1A6-m1’s antitumor activity was also abolished (figure 4F,G). These results demonstrate that both FcγRIIB and FcγRIII engagement are required for 1A6-m1’s antitumor effects. Additionally, we evaluated the expression of FcγRIIB and FcγRIII on tumor-infiltrating lymphocytes and found that both receptors were expressed on monocytic myeloid-derived suppressor cells, macrophages, dendritic cells, and NK cells; while B cells expressed only FcγRIIB (online supplemental figure 3D). These cells likely provide the Fc-FcγR interactions necessary for 1A6-m1’s antitumor effects. Taken together, these results demonstrate that 1A6-m1’s in vivo agonism and antitumor effects require both FcγRIIB and FcγRIII engagement and suggest that, in addition to FcγRIIB, activating FcγRIII can also provide FcγR-mediated crosslinking to agonistic Abs.

    CH1-hinge-engineered 1A6-h2/m1 confers enhanced antitumor effects

    It has been reported that anti-CD40, 4-1BB, and OX40 Abs in the human IgG2 (h2) isotype exhibit higher agonism than other isotypes,35–37 which has been attributed to the unique arrangement of disulfide bonds that leads to a compact and less flexible shape in the h2’s CH1-hinge region.38 39 To determine whether the h2 isotype may improve 1A6’s agonism, we replaced 1A6-m1’s CH1-hinge region with that of the h2 isotype (1A6-h2/m1). SPR analysis revealed that 1A6-m1 and 1A6-h2/m1 possessed similar binding affinity to FcγRIIB, FcγRIII, and mTNFRSF25 (online supplemental figure 4A), demonstrating that changing the CH1-hinge region of 1A6-m1 does not alter its binding profiles for mTNFRSF25 and FcγRs.

    We next compared 1A6-m1’s and 1A6-h2/m1’s agonism using the mTNFRSF25/NF-κB luciferase reporter system and found that 1A6-h2/m1 induced NF-κB signaling to a greater extent compared with 1A6-m1 (figure 5A). Moreover, the aforementioned in vitro OTI cell co-stimulation assay showed that 1A6-h2/m1 outperformed 1A6-m1 in promoting significantly greater IFN-γ secretion by OTI cells (figure 5B). To compare their antitumor effects, OTI cells were adoptively transferred on day 4 after 5×105 B16F10-OVA tumor cell inoculation, followed by injection of 1 mg/kg Abs. We observed that 1A6-h2/m1 conferred a significantly greater therapeutic benefit compared with 1A6-m1 (figure 5C,D), despite that a high titer of mouse anti-h2’s CH1-hinge antibody (1:4,000) was detected in the B16F10 mouse model following 1A6-h2/m1 treatment (online supplemental figure 4B). Additionally, in the B16F10-OVA lung metastasis model, 1A6-h2/m1 in combination with OTI cells significantly reduced the number of lung metastases compared with the control Ab (online supplemental figure 4C); in the E.G7-OVA tumor model (lacking TNFRSF25 expression, online supplemental figure 2C), 1A6-h2/m1 in combination with OTI cells also significantly inhibited the tumor growth (figure 5E,F). These results collectively demonstrate that CH1-hinge-engineered 1A6-h2/m1 induces stronger agonism than 1A6-m1 and that the h2’s CH1-hinge region confers greater antitumor activity to 1A6.

    Figure 5
    Figure 5

    CH1-hinge-engineered 1A6-h2/m1 confers enhanced antitumor effects. (A) The NF-κB signaling induced by Abs. The experiment was performed as described in the figure 1D. (B) The IFN-γ secretion of CD8+ T cells induced by Abs. Purified mouse CD8+ T cells were activated with 2 µg/mL plate-bound anti-CD3 in the presence of Abs for 3 days. (C) Schematic illustration of tumor cell inoculation and treatment in a B16F10-OVA mouse tumor model. (D) Tumor growth and survival curves of mice treated as illustrated in the panel C, n=5/group. (E) Schematic illustration of tumor cell inoculation and treatment in the E.G7-OVA mouse tumor model with OTI cell transfer. (F) Tumor growth of mice treated as illustrated in the panel E, n=6–7/group. Data are presented as means±SEM. The data shown are representative of two or three independent experiments. Abs, antibodies; IFN, interferon; i.p., intraperitoneal; i.v., intravenous; NF-κB, nuclear factor kappa B; RLU, relative luminescence units; s.c., subcutaneous.

    Multimeric TL1A fusion proteins inhibit tumor growth

    Given the essential role of receptor superclustering in mediating effective signaling induction of the TNFRSF receptors and the trimeric structure of soluble TL1A, we explored the potential of TL1A in activating TNFRSF25 signaling and conferring antitumor effects.3 40 41 We generated a fusion protein hFc-hTL1A comprising an N-terminal hFc tag and hTL1A ECD (L72-L251). SEC analysis revealed that hFc-hTL1A was eluted as two peaks (figure 6A), indicating its unstable structure. To stabilize hTL1A’s trimeric structure, we introduced an additional TNC trimerization domain,42 yielding hFc-TNC-hTL1A. SEC analysis showed that hFc-TNC-hTL1A was eluted as a single symmetric peak corresponding to a size of ~550 kDa (figure 6A), indicating that the TNC domain facilitates the self-assembly of a stable hTL1A protein. Moreover, compared with hFc-hTL1A, hFc-TNC-hTL1A had higher binding activity to hTNFRSF25 and mTNFRSF25 and induced a greater level of NF-κB signaling in hTNFRSF25/NF-κB luciferase reporter cells (figure 6B,C), indicating that stabilizing the trimeric structure of hTL1A enhances its activity.

    Figure 6
    Figure 6

    A multimeric hTL1A fusion protein inhibits tumor growth. (A) Analysis of hTL1A fusion proteins through SEC. (B) Flow cytometry analysis of hTL1A fusion proteins binding to hTNFRSF25 and mTNFRSF25. (C) The NF-κB signaling induced by hTL1A fusion proteins. The experiment was performed as described in the figure 1D. (D) The proliferation of CD4+ and CD8+ T cells induced by hFc-TNC-hTL1A. Purified CD3+ T cells from peripheral blood mononuclear cells were stained with CFSE to evaluate cell proliferation and activated with 2 µg/mL plate-bound anti-CD3 and 1 µg/mL control or hFc-TNC-hTL1A for 3 days. Left: representative histograms of CFSE dilution, right: the proportions of proliferated CD4+ (upper) or CD8+ (lower) T cells. (E) The IFN-γ secretion of CD8+ T cells induced by hFc-TNC-hTL1A. Purified CD8+ T cells from mouse spleen were treated with 2 µg/mL plate-bound anti-CD3 in the presence of 1 µg/mL proteins for 3 days. (F) The IFN-γ secretion of OTI cells induced by hFc-TNC-hTL1A. Splenocytes from OTI mice were treated with 10 pM OVA257-264 peptides and 1 µg/mL proteins for 3 days. (G) Schematic illustration of tumor cell inoculation and treatment in the B16F10-OVA mouse tumor model. (H and J) Tumor growth and survival curves in WT mice (n=5/group) (H) and FcγR-null mice (n=8–9/group) (J) treated as illustrated in the panel G. (I)A20 tumor growth after hFc-TNC-hTL1A treatment. BALB/c mice (n=5–7/group) were inoculated s.c. with 7×105 A20 cells on day 0 and i.p. injected with 5 mg/kg proteins every other day from day 6 as indicated. Data are presented as means±SEM. The data shown are representative of two or three independent experiments. Abs, antibodies; CFSE, carboxyfluorescein succinimidyl ester; hTL1A, human TNF-like cytokine 1A; hTNFRSF25, human tumor necrosis factor receptor super family 25; IFN, interferon; i.p., intraperitoneal; i.v., intravenous; mTNFRSF25, mouse tumor necrosis factor receptor super family 25; RLU, relative luminescence units; s.c., subcutaneous.

    To examine whether hFc-TNC-hTL1A has co-stimulatory activity in human and mouse T cells, we conducted T-cell co-stimulation assays in which CD3+ T cells isolated from human PBMCs were labeled with CFSE and stimulated with anti-CD3 Ab and different proteins. Flow cytometry analysis revealed that hFc-TNC-hTL1A treatment significantly increased the proliferation of CD4+ and CD8+ T cells compared with the control protein (figure 6D). Additionally, hFc-TNC-hTL1A also significantly enhanced IFN-γ secretion of mouse CD8+ T cells and OTI cells compared with the control protein (figure 6E,F). These results demonstrate that hFc-TNC-hTL1A co-stimulates both human and mouse T cells. We also generated a multimeric hFc-TNC-mTL1A fusion protein containing the hFc tag, the TNC domain tag, and mTL1A ECD (K87-L270) (online supplemental figure 5A), and hFc-TNC-mTL1A can co-stimulate mouse CD8+ T cells and OTI cells (online supplemental figure 5B,C).

    We examined whether these two multimeric TL1A proteins confer therapeutic benefits in the B16F10-OVA tumor model (figure 6G). Compared with the control protein, hFc-TNC-hTL1A significantly inhibited tumor growth and significantly prolonged survival time (figure 6H); similar results were observed for hFc-TNC-mTL1A (online supplemental figure 5D,E). In the A20 model, hFc-TNC-hTL1A significantly inhibited the tumor growth (figure 6I). We next evaluated whether multimeric TL1A proteins’ in vivo activity requires FcγR engagement. We tested an mTL1A variant (DANG-TNC-mTL1A) with D265A and N297G mutations in the Fc tag (known to abolish the binding to all FcγRs24) in WT mice bearing B16F10-OVA tumors, and it exhibited comparable antitumor effects to hFc-TNC-mTL1A (online supplemental figure 5E), indicating that FcγR engagement is not required for the antitumor activity of hFc-TNC-mTL1A. We further tested hFc-TNC-hTL1A and hFc-TNC-mTL1A in FcγR-null mice bearing B16F10-OVA tumors and found that both of these two proteins retained antitumor activity in the absence of FcγRs (figure 6J and online supplemental figure 5F). Collectively, these findings indicate that the multimeric TL1A proteins induce antitumor effects by potently activating TNRFSF25 signaling in T cells, suggesting a mechanism, that is, likely not reliant on Fc-FcγR interactions.

    To compare the antitumor effects of TNFRSF25 agonists, we used the B16F10-OVA tumor model, in which mice were treated with 5 mg/kg 1A6-m1, 1A6-h2/m1, or hFc-TNC-hTL1A 1 day after OTI cell transfer (online supplemental figure 6A). We found that 1A6-h2/m1 conferred significantly greater antitumor effects than 1A6-m1. It also appeared to have greater antitumor effects than hFc-TNC-hTL1A, though the difference was not statistically significant (online supplemental figure 6B). To assess the potential side effects induced by TNFRSF25 agonists, we measured the body weight, hematoxicity, and hepatotoxicity of B16F10-OVA tumor-bearing mice. The body weight of mice treated with 1A6-m1 and hFc-TNC-hTL1A did not show significant differences compared with the control group, and the decreased weight observed in 1A6-h2/m1-treated mice compared with the control group was likely attributed to the reduced tumor burden (online supplemental figure 6C). Three weeks after TNFRSF25 agonist treatment, no significant changes were detected in counts of white blood cells, lymphocytes, neutrophils, monocytes, or platelets (online supplemental figure 6D,E); nor was there a change in serum ALT levels (online supplemental figure 6F). Histological evaluation revealed no signs of immune cell infiltration or hepatic toxicity in mice treated with TNFRSF25 agonists (online supplemental figure 6G). Together, these results indicate that TNFRSF25 agonists exhibit good antitumor effects and a favorable safety profile.

    Discussion

    In the present study, we demonstrated that activating TNFRSF25 by agonistic Abs and multimeric TL1A fusion proteins co-stimulates CD8+ T cells and reduces tumor growth in syngeneic mouse tumor models with a good safety profile. The anti-mTNFRSF25 agonistic Ab 1A6-m1’s antitumor effects rely on both inhibitory FcγRIIB and activating FcγRIII, and replacing its CH1-hinge region with that of h2 confers enhanced antitumor activity, findings highlighting that both the CH1-hinge region and the Fc region contribute to the Ab’s antitumor activity. Moreover, multimeric TL1A fusion proteins (hFc-TNC-hTL1A and hFc-TNC-mTL1A) confer potent antitumor effects, even in the absence of Fc-FcγR interactions, providing an alternative approach for activating TNFRSF25 signaling.

    Several studies have reported that agonistic Abs targeting Fas, CD40, 4-1BB, and PD-1 exhibit an inverse correlation between affinity and agonism, where an increase in affinity leads to a decrease in agonism.43 44 A slow dissociation rate has been proposed to explain this decreased agonism: one fragment of antigen binding (Fab) of an Ab dissociates to recruit more receptor monomers to cluster, and Abs with slow dissociation rates “lock” two receptor monomers into a non-signaling complex.43 44 Consistent with this notion, we found that 1A6-m1 with a relatively faster dissociation rate and lower affinity exhibited greater in vitro activity and antitumor activity than 4C12-m1, which recognizes an overlapping epitope with 1A6-m1.

    Previous studies of patients with colon cancer or breast cancer have reported a reduction in TNFRSF25 expression within cancer tissues,45 46 and it has been observed that decreased expression of TNFRSF25 is positively correlated with shorter long-term survival rates.46 47 In B16F10-OVA tumor-bearing mice, we observed that TNFRSF25 expression on intratumoral CD3+ T cells was relatively low compared with peripheral CD3+ T cells, particularly on Treg cells and CD8+ T cells, indicating that the tumor microenvironment reduces TNFRSF25 expression on T cells and that stimulation of intratumoral T cells through TNFRSF25 can lead to improved antitumor effects.

    We found that TNFRSF25 agonists exhibited antitumor effects in a CD8+ T cell-dependent manner. Notably, in the A20 tumor model, 1A6-m1 as a single agent enhanced T-cell activation and inhibited tumor growth. However, in the B16F10-OVA and E.G7-OVA models, TNFRSF25 agonists as a single agent did not confer antitumor effects unless combined with antigen-specific OTI cells. Considering that the A20 tumor model likely has more T-cell infiltration compared with the B16F10-OVA and E.G7-OVA models,48 the different responses observed among the three models might be attributed to a lower proportion of intratumoral CD8+ T cells and/or the presence of immune-suppressive signaling that hinders the activation of CD8+ T cells within the B16F10-OVA and E.G7-OVA tumor microenvironments. Therefore, exploring the combination of TNFRSF25 agonists with other immune-enhancing agents, such as T-cell engagers, chimeric antigen receptor T cells, or T-cell receptor T cells, to enhance the function of activated T cells, holds promise for future development.

    In addition to enhancing activated T-cell function, activating TNFRSF25 also promotes Treg cell proliferation in naïve mice. A previous study reported that activating TNFRSF25 signaling before rather than after allogeneic HCT results in differential impacts on T cells: prophylactic activation (before HCT) expands Treg cells and attenuates graft-versus-host disease, whereas activating TNFRSF25 after HCT co-stimulates T cells but not Treg cells, leading to severe mortality in mice.18 We found that 1A6-m1 expanded Treg cells in the spleens of tumor-bearing mice; this appears similar to activating TNFRSF25 signaling before HCT, as most T cells are not activated by tumor antigens in the spleen. In the tumor microenvironment, where most tumor-infiltrating T cells are activated by tumor antigens, analogous to T cells being activated by alloantigen after HCT, activating TNFRSF25 signaling enhances CD8+ T-cell function but does not obviously alter Treg cells. These results suggest that TNFRSF25-induced Treg cell expansion is influenced by activated antigen-specific T cells.

    TNFRSF receptors have been shown to require superclustering to mediate activation, and bivalent Abs can induce receptor superclustering by concurrently engaging with FcγRs.2 49 Previous studies have demonstrated that anti-CD40 and anti-4-1BB Abs in m1 isotype rely on inhibitory FcγRIIB-mediated crosslinking (but not activating FcγRIII-mediated crosslinking) to induce strong agonistic responses.32 33 50 51 Similarly, we found that FcγRIIB engagement was required for 1A6-m1’s agonism. However, we also found that 1A6-m1’s antitumor activity was abolished when lacking the Fc-FcγRIII interaction. The essential role of activating FcγRIII in 1A6-m1’s activity suggests that activating FcγRIII-mediated crosslinking can also contribute to receptor superclustering in vivo and has implications for the design of agonistic Abs targeting the TNFRSF receptors. It has been shown that anti-TNFRSF receptor Abs in the h2 isotype induce greater agonism than other isotypes.35–38 The CH1-hinge’s biophysical flexibility inversely correlates with Abs’ agonism, and using the compact h2’s CH1-hinge confers better agonism.36 Changing the disulfide bonds in the hinge (VL: C214S and VH: C233S) to create a compact structure where Ab’s Fab regions covalently link to the hinge leads to greater agonism.38 39 Consistent with these previous findings, we found that changing the CH1-hinge region to the h2 isotype enhanced 1A6’s agonism. Thus, both the CH1-hinge region and the Fc region contribute to anti-TNFRSF25 Abs’ agonism.

    Another strategy used for activating TNFRSF receptors is using TNFSF ligand fusion proteins. Previous studies have shown that oligomerization of TNFSF ligands with secondary Abs can enhance their agonism and that multimeric fusion proteins can overcome the requirement for oligomerization to gain high activity in vitro.42 52 A hexavalent GITR ligand fusion protein has been reported to induce potent agonism independent of FcγR-mediated crosslinking.53 Similarly, we demonstrated that multimeric TL1A fusion proteins (hFc-TNC-hTL1A and hFc-TNC-mTL1A) inhibit tumor growth independently of FcγR engagement, indicating that they effectively activate downstream signaling without requiring additional FcγR-mediated crosslinking and have strong potential for clinical development as TNFRSF25 agonists.

    Supplemental material

    Data availability statement

    All data relevant to the study are included in the article or uploaded as supplementary information.

    Ethics statements

    Patient consent for publication

    Ethics approval

    Human peripheral blood mononuclear cells were purchased from Shanghai Aoneng Biotechnology. The use of these samples was approved by the Institutional Review Board of the National Institute of Biological Sciences, Beijing (NIBS). All animal experiments were conducted following the approved protocols of the Institutional Animal Care and Use Committee of NIBS (NIBS2018M0039).

    Acknowledgments

    We thank Dr Wanli Liu and Dr Hai Qi (Tsinghua University) for providing Fcgr2b−/− mice, Dr Gelin Wang (Tsinghua University) for critical discussions related to this project, and Zhizhong Wei and Zhiheng He for technical assistance with experiments. We also thank the NIBS Animal Facility for helping in handling and caring for the mice.

    References

    1. ↵
      1. Ribas A,
      2. Wolchok JD
      . Cancer immunotherapy using checkpoint blockade. Science 2018;359:1350–5. doi:10.1126/science.aar4060
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Dadas O,
      2. Ertay A,
      3. Cragg MS
      . Delivering co-stimulatory tumor necrosis factor receptor agonism for cancer immunotherapy: past, current and future perspectives. Front Immunol 2023;14:1147467. doi:10.3389/fimmu.2023.1147467
    3. ↵
      1. Meylan F,
      2. Davidson TS,
      3. Kahle E, et al
      . The TNF-family receptor DR3 is essential for diverse T cell-mediated inflammatory diseases. Immunity 2008;29:79–89. doi:10.1016/j.immuni.2008.04.021
      OpenUrlCrossRefPubMed
    4. ↵
      1. Migone TS,
      2. Zhang J,
      3. Luo X, et al
      . TL1A is a TNF-like ligand for DR3 and TR6/DcR3 and functions as a T cell costimulator. Immunity 2002;16:479–92. doi:10.1016/s1074-7613(02)00283-2
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Bamias G,
      2. Martin C,
      3. Marini M, et al
      . Expression, localization, and functional activity of TL1A, a novel Th1-polarizing cytokine in inflammatory bowel disease. J Immunol 2003;171:4868–74. doi:10.4049/jimmunol.171.9.4868
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. Cassatella MA,
      2. Pereira-da-Silva G,
      3. Tinazzi I, et al
      . Soluble TNF-like cytokine (TL1A) production by immune complexes stimulated monocytes in rheumatoid arthritis. J Immunol 2007;178:7325–33. doi:10.4049/jimmunol.178.11.7325
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. Bamias G,
      2. Evangelou K,
      3. Vergou T, et al
      . Upregulation and nuclear localization of TNF-like cytokine 1A (TL1A) and its receptors DR3 and DcR3 in psoriatic skin lesions. Exp Dermatol 2011;20:725–31. doi:10.1111/j.1600-0625.2011.01304.x
      OpenUrlCrossRefPubMed
    8. ↵
      1. Machida K,
      2. Aw M,
      3. Salter BMA, et al
      . The Role of the TL1A/DR3 Axis in the Activation of Group 2 Innate Lymphoid Cells in Subjects with Eosinophilic Asthma. Am J Respir Crit Care Med 2020;202:1105–14. doi:10.1164/rccm.201909-1722OC
      OpenUrl
    9. ↵
      1. Bull MJ,
      2. Williams AS,
      3. Mecklenburgh Z, et al
      . The Death Receptor 3-TNF-like protein 1A pathway drives adverse bone pathology in inflammatory arthritis. J Exp Med 2008;205:2457–64. doi:10.1084/jem.20072378
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Fang L,
      2. Adkins B,
      3. Deyev V, et al
      . Essential role of TNF receptor superfamily 25 (TNFRSF25) in the development of allergic lung inflammation. J Exp Med 2008;205:1037–48. doi:10.1084/jem.20072528
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Pappu BP,
      2. Borodovsky A,
      3. Zheng TS, et al
      . TL1A-DR3 interaction regulates Th17 cell function and Th17-mediated autoimmune disease. J Exp Med 2008;205:1049–62. doi:10.1084/jem.20071364
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Wolf D,
      2. Barreras H,
      3. Bader CS, et al
      . Marked in Vivo Donor Regulatory T Cell Expansion via Interleukin-2 and TL1A-Ig Stimulation Ameliorates Graft-versus-Host Disease but Preserves Graft-versus-Leukemia in Recipients after Hematopoietic Stem Cell Transplantation. Biol Blood Marrow Transplant 2017;23:757–66. doi:10.1016/j.bbmt.2017.02.013
      OpenUrl
    13. ↵
      1. Copsel S,
      2. Wolf D,
      3. Kale B, et al
      . Very Low Numbers of CD4+ FoxP3+ Tregs Expanded in Donors via TL1A-Ig and Low-Dose IL-2 Exhibit a Distinct Activation/Functional Profile and Suppress GVHD in a Preclinical Model. Biol Blood Marrow Transplant 2018;24:1788–94. doi:10.1016/j.bbmt.2018.04.026
      OpenUrl
    14. ↵
      1. Danese S,
      2. Klopocka M,
      3. Scherl EJ, et al
      . Anti-TL1A Antibody PF-06480605 Safety and Efficacy for Ulcerative Colitis: A Phase 2a Single-Arm Study. Clin Gastroenterol Hepatol 2021;19:2324–32. doi:10.1016/j.cgh.2021.06.011
      OpenUrl
    15. ↵
      1. Taraban VY,
      2. Ferdinand JR,
      3. Al-Shamkhani A
      . Expression of TNFRSF25 on conventional T cells and Tregs. J Clin Invest 2011;121:463–4. doi:10.1172/JCI45832
      OpenUrlCrossRefPubMed
    16. ↵
      1. Li L,
      2. Lu Y,
      3. Fu L, et al
      . Expression of death receptor 3 (DR3) on peripheral blood mononuclear cells of patients with psoriasis vulgaris. Postgrad Med J 2018;94:551–5. doi:10.1136/postgradmedj-2018-136040
      OpenUrlAbstract/FREE Full Text
    17. ↵
      1. Wolf D,
      2. Schreiber TH,
      3. Tryphonopoulos P, et al
      . Tregs expanded in vivo by TNFRSF25 agonists promote cardiac allograft survival. Transplantation 2012;94:569–74. doi:10.1097/TP.0b013e318264d3ef
      OpenUrlPubMed
    18. ↵
      1. Nishikii H,
      2. Kim B-S,
      3. Yokoyama Y, et al
      . DR3 signaling modulates the function of Foxp3+ regulatory T cells and the severity of acute graft-versus-host disease. Blood 2016;128:2846–58. doi:10.1182/blood-2016-06-723783
      OpenUrlAbstract/FREE Full Text
    19. ↵
      1. Slebioda TJ,
      2. Rowley TF,
      3. Ferdinand JR, et al
      . Triggering of TNFRSF25 promotes CD8. Eur J Immunol 2011;41:2606–11. doi:10.1002/eji.201141477
      OpenUrlCrossRefPubMed
    20. ↵
      1. Liu H,
      2. Zheng S,
      3. Hou X, et al
      . Novel Abs targeting the N-terminus of fibroblast growth factor19 inhibit hepatocellular carcinoma growth without bile-acid-related side-effects. Cancer Sci 2020;111:1750–60. doi:10.1111/cas.14353
      OpenUrl
    21. ↵
      1. Li D,
      2. He W,
      3. Liu X, et al
      . A potent human neutralizing antibody Fc-dependently reduces established HBV infections. Elife 2017;6:e26738. doi:10.7554/eLife.26738
    22. ↵
      1. Zhang X,
      2. Zhang C,
      3. Qiao M, et al
      . Depletion of BATF in CAR-T cells enhances antitumor activity by inducing resistance against exhaustion and formation of central memory cells. Cancer Cell 2022;40:1407–22. doi:10.1016/j.ccell.2022.09.013
      OpenUrlCrossRef
    23. ↵
      1. Li Y,
      2. Liu J,
      3. Chen W, et al
      . A pH-dependent anti-CD47 antibody that selectively targets solid tumors and improves therapeutic efficacy and safety. J Hematol Oncol 2023;16:2. doi:10.1186/s13045-023-01399-4
    24. ↵
      1. Lo M,
      2. Kim HS,
      3. Tong RK, et al
      . Effector-attenuating Substitutions That Maintain Antibody Stability and Reduce Toxicity in Mice. J Biol Chem 2017;292:3900–8. doi:10.1074/jbc.M116.767749
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. Yang F,
      2. Zhao L,
      3. Wei Z, et al
      . A Cross-Species Reactive TIGIT-Blocking Antibody Fc Dependently Confers Potent Antitumor Effects. J Immunol 2020;205:2156–68. doi:10.4049/jimmunol.1901413
      OpenUrlAbstract/FREE Full Text
    26. ↵
      1. Schreiber TH,
      2. Wolf D,
      3. Bodero M, et al
      . T cell costimulation by TNFR superfamily (TNFRSF)4 and TNFRSF25 in the context of vaccination. J Immunol 2012;189:3311–8. doi:10.4049/jimmunol.1200597
      OpenUrlAbstract/FREE Full Text
    27. ↵
      1. Zhao C-C,
      2. Han Q-J,
      3. Ying H-Y, et al
      . TNFSF15 facilitates differentiation and polarization of macrophages toward M1 phenotype to inhibit tumor growth. Oncoimmunology 2022;11:2032918. doi:10.1080/2162402X.2022.2032918
    28. ↵
      1. Zhai Y,
      2. Ni J,
      3. Jiang GW, et al
      . VEGI, a novel cytokine of the tumor necrosis factor family, is an angiogenesis inhibitor that suppresses the growth of colon carcinomas in vivo. FASEB J 1999;13:181–9. doi:10.1096/fasebj.13.1.181
      OpenUrlPubMedWeb of Science
    29. ↵
      1. Bodmer JL,
      2. Schneider P,
      3. Tschopp J
      . The molecular architecture of the TNF superfamily. Trends Biochem Sci 2002;27:19–26. doi:10.1016/s0968-0004(01)01995-8
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Marsters SA,
      2. Sheridan JP,
      3. Donahue CJ, et al
      . Apo-3, a new member of the tumor necrosis factor receptor family, contains a death domain and activates apoptosis and NF-kappa B. Curr Biol 1996;6:1669–76. doi:10.1016/s0960-9822(02)70791-4
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Schreiber TH,
      2. Wolf D,
      3. Tsai MS, et al
      . Therapeutic Treg expansion in mice by TNFRSF25 prevents allergic lung inflammation. J Clin Invest 2010;120:3629–40:42933. doi:10.1172/JCI42933
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. Li F,
      2. Ravetch JV
      . Inhibitory Fcγ receptor engagement drives adjuvant and anti-tumor activities of agonistic CD40 antibodies. Science 2011;333:1030–4. doi:10.1126/science.1206954
      OpenUrlAbstract/FREE Full Text
    33. ↵
      1. White AL,
      2. Chan HTC,
      3. Roghanian A, et al
      . Interaction with FcγRIIB is critical for the agonistic activity of anti-CD40 monoclonal antibody. J Immunol 2011;187:1754–63. doi:10.4049/jimmunol.1101135
      OpenUrlAbstract/FREE Full Text
    34. ↵
      1. Nimmerjahn F,
      2. Ravetch JV
      . Fcgamma receptors as regulators of immune responses. Nat Rev Immunol 2008;8:34–47. doi:10.1038/nri2206
      OpenUrlCrossRefPubMedWeb of Science
    35. ↵
      1. Yu X,
      2. Chan HTC,
      3. Fisher H, et al
      . Isotype Switching Converts Anti-CD40 Antagonism to Agonism to Elicit Potent Antitumor Activity. Cancer Cell 2020;37:850–66. doi:10.1016/j.ccell.2020.04.013
      OpenUrlCrossRefPubMed
    36. ↵
      1. Liu X,
      2. Zhao Y,
      3. Shi H, et al
      . Human immunoglobulin G hinge regulates agonistic anti-CD40 immunostimulatory and antitumour activities through biophysical flexibility. Nat Commun 2019;10:4206. doi:10.1038/s41467-019-12097-6
      OpenUrlCrossRef
    37. ↵
      1. Yu X,
      2. James S,
      3. Felce JH, et al
      . TNF receptor agonists induce distinct receptor clusters to mediate differential agonistic activity. Commun Biol 2021;4:772. doi:10.1038/s42003-021-02309-5
    38. ↵
      1. White AL,
      2. Chan HTC,
      3. French RR, et al
      . Conformation of the human immunoglobulin G2 hinge imparts superagonistic properties to immunostimulatory anticancer antibodies. Cancer Cell 2015;27:138–48. doi:10.1016/j.ccell.2014.11.001
      OpenUrlCrossRefPubMed
    39. ↵
      1. Orr CM,
      2. Fisher H,
      3. Yu X, et al
      . Hinge disulfides in human IgG2 CD40 antibodies modulate receptor signaling by regulation of conformation and flexibility. Sci Immunol 2022;7:eabm3723. doi:10.1126/sciimmunol.abm3723
    40. ↵
      1. Wajant H
      . Principles of antibody-mediated TNF receptor activation. Cell Death Differ 2015;22:1727–41. doi:10.1038/cdd.2015.109
      OpenUrlCrossRefPubMed
    41. ↵
      1. Jin T,
      2. Kim S,
      3. Guo F, et al
      . Purification and crystallization of recombinant human TNF-like ligand TL1A. Cytokine 2007;40:115–22. doi:10.1016/j.cyto.2007.07.193
      OpenUrlCrossRefPubMed
    42. ↵
      1. Wyzgol A,
      2. Müller N,
      3. Fick A, et al
      . Trimer stabilization, oligomerization, and antibody-mediated cell surface immobilization improve the activity of soluble trimers of CD27L, CD40L, 41BBL, and glucocorticoid-induced TNF receptor ligand. J Immunol 2009;183:1851–61. doi:10.4049/jimmunol.0802597
      OpenUrlAbstract/FREE Full Text
    43. ↵
      1. Chodorge M,
      2. Züger S,
      3. Stirnimann C, et al
      . A series of Fas receptor agonist antibodies that demonstrate an inverse correlation between affinity and potency. Cell Death Differ 2012;19:1187–95. doi:10.1038/cdd.2011.208
      OpenUrlCrossRefPubMedWeb of Science
    44. ↵
      1. Yu X,
      2. Orr CM,
      3. Chan HTC, et al
      . Reducing affinity as a strategy to boost immunomodulatory antibody agonism. Nature New Biol 2023;614:539–47. doi:10.1038/s41586-022-05673-2
      OpenUrl
    45. ↵
      1. Ślebioda TJ,
      2. Stanisławowski M,
      3. Cyman M, et al
      . Distinct Expression Patterns of Two Tumor Necrosis Factor Superfamily Member 15 Gene Isoforms in Human Colon Cancer. Dig Dis Sci 2019;64:1857–67. doi:10.1007/s10620-019-05507-8
      OpenUrl
    46. ↵
      1. Ge Z,
      2. Sanders AJ,
      3. Ye L, et al
      . Expression of death receptor-3 in human breast cancer and its functional effects on breast cancer cells in vitro. Oncol Rep 2013;29:1356–64. doi:10.3892/or.2013.2259
      OpenUrlPubMed
    47. ↵
      1. Qi C,
      2. Wang X,
      3. Shen Z, et al
      . Anti-mitotic chemotherapeutics promote apoptosis through TL1A-activated death receptor 3 in cancer cells. Cell Res 2018;28:544–55. doi:10.1038/s41422-018-0018-6
      OpenUrl
    48. ↵
      1. Zhong W,
      2. Myers JS,
      3. Wang F, et al
      . Comparison of the molecular and cellular phenotypes of common mouse syngeneic models with human tumors. BMC Genomics 2020;21:2. doi:10.1186/s12864-019-6344-3
    49. ↵
      1. Zhou Z,
      2. Song X,
      3. Berezov A, et al
      . Human glucocorticoid-induced TNF receptor ligand regulates its signaling activity through multiple oligomerization states. Proc Natl Acad Sci U S A 2008;105:5465–70. doi:10.1073/pnas.0711350105
      OpenUrlAbstract/FREE Full Text
    50. ↵
      1. Li F,
      2. Ravetch JV
      . Antitumor activities of agonistic anti-TNFR antibodies require differential FcγRIIB coengagement in vivo. Proc Natl Acad Sci USA 2013;110:19501–6. doi:10.1073/pnas.1319502110
      OpenUrlAbstract/FREE Full Text
    51. ↵
      1. Qi X,
      2. Li F,
      3. Wu Y, et al
      . Optimization of 4-1BB antibody for cancer immunotherapy by balancing agonistic strength with FcγR affinity. Nat Commun 2019;10:2141. doi:10.1038/s41467-019-10088-1
    52. ↵
      1. Müller N,
      2. Wyzgol A,
      3. Münkel S, et al
      . Activity of soluble OX40 ligand is enhanced by oligomerization and cell surface immobilization. FEBS J 2008;275:2296–304. doi:10.1111/j.1742-4658.2008.06382.x
      OpenUrlCrossRefPubMed
    53. ↵
      1. Richards DM,
      2. Marschall V,
      3. Billian-Frey K, et al
      . HERA-GITRL activates T cells and promotes anti-tumor efficacy independent of FcγR-binding functionality. J Immunother Cancer 2019;7:191. doi:10.1186/s40425-019-0671-4

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Contributors Conceptualization: XL and JS. Methodology: XL. Investigation: XL, LZ, SC, YL, YY, HL, and FY. Visualization: XL. Writing—Original Draft: XL and JS. Writing—Review and Editing: XL and JS. Funding Acquisition: JS. Supervision: WL and JS. JS is the guarantor.

    • Funding This work was supported by grants from the Beijing Municipal Science and Technology Commission and the Beijing Key Laboratory of Pathogen Invasion and Immune Defense (Z171100002217064 to JS).

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.