Preclinical characterization of MTX-101: a novel bispecific CD8 Treg modulator that restores CD8 Treg functions to suppress pathogenic T cells in autoimmune diseases

Abstract

Introduction: Regulatory CD8 T cells (CD8 Treg) are responsible for the selective killing of self-reactive and pathogenic CD4 T cells. In autoimmune disease, CD8 Treg may accumulate in the peripheral blood but fail to control the expansion of pathogenic CD4 T cells that subsequently cause tissue destruction. This CD8 Treg dysfunction is due in part to the expression of inhibitory killer immunoglobulin-like receptors (KIR; KIR2DL isoforms [KIR2DL1, KIR2DL2, and KIR2DL3]); these molecules serve as autoimmune checkpoints and limit CD8 Treg activation.

Methods: Here we describe the pre-clinical characterization of MTX-101, a bispecific antibody targeting inhibitory KIR and CD8. Using human peripheral blood mononuculear cells (PBMC) derived from healthy donors and autoimmune patients, humanized mouse models, and human derived tissue organoids, we evaluated the molecular mechanisms and functional effects of MTX-101.

Results: By binding to KIR, MTX-101 inhibited KIR signaling that can restore CD8 Treg ability to eliminate pathogenic CD4 T cells. MTX-101 bound and activated CD8 Treg in human peripheral blood mononuclear cells (PBMC), resulting in increased CD8 Treg cytolytic capacity, activation, and prevalence. Enhancing CD8 Treg function with MTX-101 reduced pathogenic CD4 T cell expansion and inflammation, without increasing pro-inflammatory cytokines or activating immune cells that express either target alone. MTX-101 reduced antigen induced epithelial cell death in disease affected tissues, including in tissue biopsies from individuals with autoimmune disease (i.e., celiac disease, Crohn's disease). The effects of MTX-101 were specific to autoreactive CD4 T cells and did not suppress responses to viral and bacterial antigens. In a human PBMC engrafted Graft versus Host Disease (GvHD) mouse model of acute inflammation, MTX-101 bound CD8 Treg and delayed onset of disease. MTX-101 induced dose dependent binding, increased prevalence and cytolytic capacity of CD8 Treg, as well as increased CD4 T cell death. MTX-101 selectively bound CD8 Treg without unwanted immune cell activation or increase of pro-inflammatory serum cytokines and exhibited an antibody-like half-life in pharmacokinetic and exploratory tolerability studies performed using IL-15 transgenic humanized mice with engrafted human lymphocytes, including CD8 Treg at physiologic ratios.

Conclusion: Collectively, these data support the development of MTX-101 for the treatment of autoimmune diseases.

Keywords: CD8 regulatory T cells; autoimmune disease affected tissue organoids; autoreactive CD4 T cells; immunomodulatory bispecific antibodies; killer immunoglobulin like receptor (KIR).

Figure 1

KIR CD8 regulatory T cells (Tregs) express the surface markers NKG2C, Helios, and KLRG1 and, when enriched from celiac donor peripheral blood mononuclear cells (PBMCs), specifically eliminate a gliadin T-cell receptor (TCR)-transduced SKW line in a dose-dependent manner. (A, B) Healthy donor PBMCs were plated overnight and then stained for KIRs, NKG2C, Helios, and KLRG1 (A) and the cytolytic marker, Granzyme B (B). p-Values were determined with a paired t-test. (B) The percentage positive of each of these markers is shown for the KIR+ CD8 Treg and KIR− CD8 non-CD8 Treg populations from seven healthy donors. (C) KIR+ CD8 Tregs were isolated from celiac PBMCs following expansion with IL-7, IL-15, and gliadin peptides. Sorted CD8 Tregs were then cultured with gliadin-expanded CD4 T cells and autologous antigen-presenting cells (APCs) with or without (unstimulated) additional gliadin peptide stimulation for 3 days and evaluated for CD8 Treg number. Results are representative of two independent experiments across three different celiac donors. p-Values were determined with an unpaired t-test. (D) CD8 Tregs were enriched from celiac donor PBMCs following expansion in IL-7 and IL-15 for 7 days and sorting for CD5+NKp46-CD4-CCR7-CD28- T cells. CD8 Tregs were quantified, and increasing ranges of CD8 Tregs were cultured with 20,000 GFP+ gliadin-responsive TCR-transduced SKW target cells. Percent change in GFP+ objects (as determined by Incucyte software) from baseline (t = 8 hours) is shown for target cells in the absence of CD8 Tregs and with increasing CD8 Treg number. Results are representative of three independent experiments. (E) CD8 Treg-enriched cells were cultured with either green fluorescent protein (GFP)-expressing parental SKW or gliadin TCR-transduced SKW target cells for 48 hours and the percent change in GFP signal relative to baseline (t = 8 hours, n = 2 independent experiments). (A–E) ns: p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. (D, E) p-Values were determined through a one-way ANOVA of calculated areas under the curves followed by Šídák’s multiple-comparisons test.

Figure 2

MTX-101 specifically binds to target surface receptors on CD8 regulatory T cells (Tregs). (A) Table shows the binding affinities of MTX-101 for human KIR2DL1/2/3 and CD8α based on association and dissociation rate constants measured for each binding interaction, and K_D binding affinity calculated from the measured rate constants. (B) Dose–response curves of MTX-101 binding to stably expressed KIR2DL1 and CD8α SKW cell lines. EC₅₀ values were calculated for each cell line using Prism GraphPad software with non-linear regression [agonist vs. response curve (three parameters)]. Binding curves are representative of binding to KIR2DL1 expressing (n = 14 independent experiments) and CD8a expressing (n = 10) cell lines. (C) Simultaneous co-binding of antigens to MTX-101 as determined by Bio-Layer Interferometry. MTX-101 was captured with either KIR2DL or CD8α, and then additional antigen binding was detected with CD8α or KIR2DL, respectively. (D) Dose-dependent MTX-101 binding to immune cell populations within total human peripheral blood mononuclear cells (PBMCs) is shown as detected with an anti-human Fc secondary antibody. Data are representative of seven independent experiments with different donors. (E) Quantification of CD8 and KIR binding sites per CD8 Treg target cell as determined by receptor mean fluorescence intensity (MFI) and quantitative counting beads is shown in the figure. p-Values were determined through a one-way ANOVA of calculated areas under the curves followed by Šídák’s multiple-comparisons test. ns: p > 0.05. (F) Detection of unbound KIR sites with a saturating dose of fluorescently labeled single-arm KIR-Fc antibody on CD8 Tregs in total PBMCs following incubation with increasing doses of MTX-101. IC50 values were calculated for KIR-Fc using GraphPad Prism software and non-linear fit log (inhibitor) vs. response with variable slope and four parameters. The graph is representative of binding data across three independent healthy donors.

Figure 3

MTX-101 activates CD8 regulatory T cells (Tregs) in celiac and healthy peripheral blood mononuclear cells (PBMCs) and induces lysis of gliadin-specific target cells in celiac T-cell co-culture assay. (A) Percentage of CD8 Tregs detected in healthy (n = 2) and celiac (n = 3) PBMCs. (B) MTX-101 dose-dependent CD8 Tregs binding across the same donors as in panel (A) in total PBMCs. p-Values were determined with an unpaired t-test. ns: p > 0.05, *p < 0.05, **p < 0.01. (C) MTX-101 dose-dependent activation of CD8 Tregs in celiac PBMCs (n = 3 donors) following 48 hours of incubation. (D, E) Celiac KIR+ CD8 Tregs were expanded in IL-7+IL-15+ gliadin peptides. Sorted CD8 Tregs were then cultured with autologous gliadin peptide-expanded CD4 T cells and peptide-pulsed antigen-presenting cells (APCs) +/− 100 nM MTX-101 for 7 days. (D) Percentage of CD8 Treg and Granzyme B and CD107a expression in CD8 following MTX-101 treatment. (E) CD4 readouts in co-cultures including number of total CD4 T cells and IFN-γ-producing CD4 T cells following 72-hour incubation with MTX-101. (F) Meso Scale Discovery (MSD) quantification of secreted IFN-γ, TNF-α, and granulocyte-macrophage colony-stimulating factor (GM-CSF) in the supernatant of co-cultures stimulated with gliadin peptides in the presence of MTX-101 shown relative to untreated controls. (D–F) Results are representative of two independent experiments (n = 3 donors). p-Values were determined with an unpaired t-test. ns: p > 0.05, *p < 0.05, **p < 0.01. (G, H) Activation of CD8 T cells (G) or NK (H) cells as determined percentage CD25 positive is shown for two healthy donor PBMCs following 48-hour incubation in the shown concentrations of anti-KIR single-arm antibody, anti-CD8 single-arm antibody, or MTX-101 in either untreated (resting) or anti-CD3-treated (0.1 µg/mL, activated) PBMCs.

Figure 4

MTX-101 restores CD8 regulatory T-cell (Treg) functions and reduces antigen-induced proinflammatory responses in Crohn’s donor peripheral blood mononuclear cells (PBMCs) and organoids. (A) Healthy or Crohn’s donors (n = 6 each) CD8 Tregs were stained for Granzyme B and Helios. CD4 T cells were stained for co-expression of CD161+CXCR3+CD39+ in PBMCs from eight healthy (n = 8) or Crohn’s donors (n = 11). (B, C) Crohn’s PBMCs were incubated for 7 days with a mixture of IL-7+IL-15, bacterial flagellin, and OmpC peptide in the presence or absence of MTX-101 at 100 nM. (B) The percentage increase in concentration of Granzyme B in the supernatant of antigen-responsive or non-responsive Crohn’s donors (t = 5 days). (C) Carboxyfluorescein succinimidyl ester (CFSE) dilution of antigen-responsive CD4 T cells (t = day 7) antigen-expanded CD4 T cells as determined by diluted CFSE is also shown on day 7 for both responders and non-responders relative to untreated controls. Secreted IFN-γ and TNF-α (t = day 7) in MTX-101-treated donors relative to untreated. (D) Helios expression on CD4 T cells in PBMCs from Crohn’s donors stimulated with anti-CD3 (1 μg/mL) overnight in the presence and absence of MTX-101 (100 nM). (A–D) p-Values were determined with an unpaired t-test. ns: p > 0.05, *p < 0.05. (E) Change in antigen-induced epithelial cell death in both celiac and Crohn’s donor-derived organoids +/− 100 nM MTX-101. Results represent celiac (n = 3) and Crohn’s donors (n = 2). (F) Healthy or celiac PBMCs (n = 3 each) were restimulated with CEFT, Influenza HA, SARS-CoV-2, or AVV5/6/8 peptide (0, 0.1, or 1 µg/mL), in the presence and absence of MTX-101. IFN-γ detected in PBMC culture supernatant on day 5. (G) IFN-γ concentrations in supernatants following 5-day bacterial stimuli of tetanus toxoid (5 µg/mL) or SEB (100 ng/mL). Each point represents the mean of two technical replicates. Values untreated and with MTX-101 were compared using unpaired t-tests and two-way ANOVA followed by Šídák’s multiple-comparisons test.

Figure 5

MTX-101 binding, efficacy, mechanism of action, and dose titration in acute graft-versus-host disease (GvHD) human PBMC-engrafted NSG mouse model. (A) Survival study experimental design. NSG mice were irradiated (0.75 Gy) and injected intravenously with 1 × 10⁷ human PBMC healthy donor 3578 (haplotype HLA-DQ2.5). MTX-101 (2 mg/kg) or saline was injected every 7 days until day 28. Low-dose IL-2 (25,000 IU) was injected every other day from study day (SD) 0 to 10, n = 20/cohort. (B) MTX-101 binding to CD4, CD8, and CD8 regulatory T-cell (Treg) populations as detected by mean fluorescence intensity (MFI) of anti-human Fc antibody in the blood on SD20, and spleens were collected on SD15 (MTX-101 and saline, n = 3 mice; IL-2, n = 2 mice) and at study termination (MTX-101 and saline, n = 6 mice). SD20 timepoint is representative of mice remaining in the MTX-101 (n = 8) and IL-2 treatments (n = 6). (C) CD25 and ICOS-positive CD8 non-Tregs and CD8 Tregs in peripheral blood as detected on SD9 and 15 (n = 10/cohort). CD8 Tregs and CD8 non-Treg Granzyme B MFI on SD20. nd: not determined. (D) Absolute counts of CD25+CD4 T cells/μL of blood on SD15 (MTX-101 and saline, n = 3 mice; IL-2, n = 2 mice). Ki67 MFI in CD4 T cells in peripheral blood on SD15. (A–D) p-Values were determined with an unpaired t-test. ns: p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. (E) Survival curve in the acute GvHD model (saline, MTX-101, and low-dose IL-2, n = 18/cohort). Survival curves were compared using the log-rank (Mantel–Cox) test. p=0.0043. (F–H) Human NSG mice engrafted with 1 × 10⁷ human peripheral blood mononuclear cells (PBMCs) following irradiation (0.5 Gy) from donor LZ0007 and dosed with increasing concentrations of MTX-101 intravenously every 7 days on SD0–SD21. (F) Binding of MTX-101 to CD4, CD8, and CD8 Tregs (KLRG1+ CD8 T cells) in the blood on SD14. Granzyme B (MFI and percentage) on non-Treg CD8 and CD8 Treg populations in the blood and spleens from mice that were terminated on study day 11 (n = 3 mice). (G) CD25-positive CD8 Tregs and Helios-positive CD8 T cells in peripheral blood on SD14 (n = 8) and all surviving mice on SD42. (H) Percentage of CD25+ and Annexin V+ CD4 T cells on SD42 in peripheral blood and splenocytes. (F–H) p-Values were determined by an unpaired t-test. ns: p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 6

MTX-101 treatment does not activate CD8 or NK cells in vitro or in humanized CD34+ NSG-Tg(Hu-IL-15) mice despite antibody-like half-life in the serum. (A) Binding of MTX-101, single-arm anti-KIR, single-arm anti-CD8, or anti-KIR bivalent monoclonal antibody to CD8 regulatory T cells (Tregs) (left), non-Treg CD8 T cells (center), and NK cells (right) after a single dose of 5 mg/kg. Data are representative of two samples of pooled blood from three mice that received CD34 cells from two different donors at baseline, 3 hours, 24 hours, and 72 hours and all six mice at 168 hours. (B) Activation marker CD69+/CD25+ expression on CD8 Tregs (left), non-Treg CD8 T cells (center), and NK cells (right) after one 5 mg/kg dose of MTX-101, single-arm anti-KIR, anti-CD8 antibody, bivalent anti-KIR monoclonal antibody, or anti-CD3 (OKT3) control antibody at 0.5 mg/kg. (C) In vivo study experimental design. Humanized CD34+ NSG-Tg(Hu-IL-15) were injected with 1 and 10 mg/kg MTX-101, and blood was collected at specified timepoints in the diagram. Exposure of MTX-101 in CD34+ NSG-Tg(Hu-IL-15) was quantified from at start to 672 hours. Table shows the half-life of MTX-101, clearance rate, and volume distribution.