TREX1 as a Novel Immunotherapeutic Target

Abstract

Mutations in the TREX1 3' → 5' exonuclease are associated with a spectrum of autoimmune disease phenotypes in humans and mice. Failure to degrade DNA activates the cGAS-STING DNA-sensing pathway signaling a type-I interferon (IFN) response that ultimately drives immune system activation. TREX1 and the cGAS-STING DNA-sensing pathway have also been implicated in the tumor microenvironment, where TREX1 is proposed to degrade tumor-derived DNA that would otherwise activate cGAS-STING. If tumor-derived DNA were not degraded, the cGAS-STING pathway would be activated to promote IFN-dependent antitumor immunity. Thus, we hypothesize TREX1 exonuclease inhibition as a novel immunotherapeutic strategy. We present data demonstrating antitumor immunity in the TREX1 D18N mouse model and discuss theory surrounding the best strategy for TREX1 inhibition. Potential complications of TREX1 inhibition as a therapeutic strategy are also discussed.

Keywords: cancer; exonuclease; immunotherapy; inhibition; small-molecule.

Figure 1

Crystal Structure of the Dimeric Exonuclease mTREX1(1-242). Structure includes only the TREX1 catalytic domain (1-242). Protomers are distinguished by green and cyan, ssDNA by blue sticks, and calcium ions by magenta coloring. Crystal structure was visualized in PyMOL using the PDB structure 2OA8 from ref (6).

Figure 2

TREX1 is a Member of the DEDD Family of Exonucleases. Structure includes only the TREX1 catalytic domain (1-242). Protomers are distinguished by green and cyan cartoons, and D18-E20-D130-D200 motif residues are shown as red sticks with black labels. Crystal structure was visualized in PyMOL using the PDB structure 3MXJ from ref (16).

Figure 3

TREX1^D18N Mice Display T-cell Dependent Antitumor Immunity. (A, B) 5x10⁶ H31m1 tumor cells were injected subcutaneously into WT and D18N mice, and survival (A) and tumor volume (B) tracked daily (see Methods). Mice were treated with αCD4, αCD8, or the respective isotype-control antibodies to test the effects of T-cell depletion (see Methods). Isotype controls are presented together. Tumor volumes are average and standard deviation. Background of mice and tumor cells was 129S1/SvImJ, and each group represents 8-16 mice across 2-4 independent experiments. Data originally submitted for ASBMB 2020 conference (93). Graphs generated with Prism 7.0 (GraphPad).

Figure 4

Similar T-Cell Number but Decreased PD-1 Expression in TREX1^D18N Mice During Tumor Progression. (A–C) WT or D18N mice were challenged with 5x10⁶ H31m1 cells, cells were isolated from the indicated tissue on Day 8, and (A) activated/memory CD4⁺ and CD8α⁺CD44^high T-cells were measured by flow cytometry (see Methods). Numbers of indicated (B) CD4⁺ or (C) CD8+ T cells were determined. ‘SPL’ = spleen, ‘CLN’ = contralateral lymph nodes, ‘TDLN’ = tumor draining lymph node, and ‘TIL’ = tumor infiltrating lymphocytes. (C) PD-1 M.F.I. on activated/memory CD8⁺CD44^high T-cells in the tumor were determined, and the fold change compared to naïve T-cells in the spleen was calculated (see Methods). Individual mice (6-9 total, 3 independent experiments) are plotted, with averages represented by horizontal bars. *p-value < 0.05 via two-tailed independent Student’s t-test. All graphs prepared in Prism 9.0 (GraphPad).

Figure 5

Structural Comparison of TREX1 and TREX2. Graphic (A) shows structural alignment of mTREX1(1-242) and hTREX2 in cyan and green, respectively, and graphic (B) is the same alignment with discrepant residues colored red. Alignment and graphics were generated in PyMOL using the PDB structures 3MXJ and 1Y97 from refs (16, 108).

Figure 6

Small Molecule Inhibitor with High Specificity for hTREX1. (A–C) Standard time-course reactions were prepared in 150 μL volumes containing vehicle or indicated concentrations of inhibitor, and hTREX1 (A), mTREX1 (B), or hTREX2 (C). Reactions were incubated 1-hr at room temperature, and 20 μL samples of each reaction taken at time-points of 0, 5, 10, 20, 30, 45, & 60 minutes and quenched in 20 μL of 15X SYBR Green. Fluorescence was measured, and fluorescence vs time plots were normalized to maximal initial fluorescence and background fluorescence (see Methods). Plots were fit with nonlinear regression. Plots were generated in Prism (GraphPad) and combined in PowerPoint.

Figure 7

Active Sites are Accessible by Solvent Channel in TREX1 Apoenzyme Crystals. Structural representation of crystal lattice for mTREX1(1-242) apoenzyme crystal. Functional unit of interest is colored cyan with DEDD active site residues for one protomer shown as red sticks; other functional units are colored green. Graphic (A) is a slice through the crystal lattice where the active site is visibly facing the solvent channel, and graphic (B) looks through the solvent channel into the crystal lattice. Alignment and graphic were generated in PyMOL using the PDB structure 3MXJ from ref (16).