ER-associated degradation adapter Sel1L is required for CD8+ T cell function and memory formation following acute viral infection

Abstract

The maintenance of antigen-specific CD8⁺ T cells underlies the efficacy of vaccines and immunotherapies. Pathways contributing to CD8⁺ T cell loss are not completely understood. Uncovering the pathways underlying the limited persistence of CD8⁺ T cells would be of significant benefit for developing novel strategies of promoting T cell persistence. Here, we demonstrate that murine CD8⁺ T cells experience endoplasmic reticulum (ER) stress following activation and that the ER-associated degradation (ERAD) adapter Sel1L is induced in activated CD8⁺ T cells. Sel1L loss limits CD8⁺ T cell function and memory formation following acute viral infection. Mechanistically, Sel1L is required for optimal bioenergetics and c-Myc expression. Finally, we demonstrate that human CD8⁺ T cells experience ER stress upon activation and that ER stress is negatively associated with improved T cell functionality in T cell-redirecting therapies. Together, these results demonstrate that ER stress and ERAD are important regulators of T cell function and persistence.

Keywords: CD8 T cell; CP: Immunology; ER stress; ERAD; Myc; T cell memory; immunometabolism; protein homeostasis.

Figure 1.. T cell activation induces ER stress in vitro

(A) Experimental schema. Splenocytes from wild-type (WT) P14 transgenic mice were activated in vitro with LCMV gp33–41 peptide in the presence of IL-2 for 3 days (T_ACT), then incubated with either IL-2 to generate IL-2 “effector” cells (IL-2 T_E) or IL-15 to generate IL-15 “memory” cells (IL-15 T_M). Created with BioRender.com. (B) (Left) Confocal microscopy and (right) associated quantification of calreticulin (CALR) in in vitro naive (D0), T_ACT (D3), IL-2 T_E, and IL-15 T_M P14 cells. Scale bar represents 10 μm. (C) (Left) Representative histograms of PROTEOSTAT in in vitro naive (D0), T_ACT, IL-2 T_E, and IL-15 T_M. (Right) Fold change in median fluorescence intensity (MFI) of PROTEOSTAT normalized to naive. (D) (Left) Representative immunoblot of K48-Ub in serially collected in vitro naive, T_ACT, IL-2 T_E, and IL-15 T_M P14 cells. (Right) Densitometry quantification of the K48-Ub immunoblot bands in indicated conditions. (E) (Left) Representative immunoblot of ATF4 and XBP1 (spliced/unspliced) in serially collected in vitro naive, T_ACT, IL-2 T_E, and IL-15 T_M P14 cells. (Right) Densitometry quantification of the immunoblot bands. (F) (Left) Representative histograms of XBP1s in in vitro naive, T_ACT, IL-2 T_E, and IL-15 T_M. (Right) Fold change in MFI of XBP1s normalized to naive. Data representative of naive n = 276, T_ACT n = 376, IL-2 T_E n = 882, and IL-15 T_M n = 751 (B), n = 4 (C and F), and n = 5 (D and E). All immunoblot data are normalized β-actin, then to naive. *p < 0.05, **p < 0.01, ****p < 0.0001, one-way ANOVA with uncorrected Fisher’s least significant difference (B–F).

Figure 2.. Antigen-specific CD8⁺ T cells experience dynamic ER stress during an acute viral infection in vivo

(A) (Left) Uniform manifold approximation and projection (UMAP) of P14 cells responding to LCMV infection (GEO: GSE131847) determined by scanpy. Each dot corresponds to one individual cell colored by day of infection. (Right) enrichment (adjusted p values) of gene module “hallmark unfolded protein response” determined by Fisher’s exact test (decouplr) and used to color UMAP plots. (B and C) (Left) Representative histograms of PROTEOSTAT (B) or ER Tracker (C) in gp33⁺CD44⁺CD8⁺TCRb⁺ splenocytes isolated on indicated days post infection (p.i.). LCMV compared to naive CD8⁺ T cells from uninfected mice. (Right) Fold change in MFI of PROTEOSTAT or ER Tracker normalized to uninfected naive CD8⁺ T cells. (D and E) (Left) Representative histograms of intracellular XBP1s (D) and intracellular CHOP (E) in gp33⁺ or NP396⁺CD44⁺CD8⁺TCRb⁺ splenocytes. (Right) Fold change in MFI of XBP1s (D) or CHOP (E) in gp33⁺ or NP396⁺ normalized to uninfected naive CD8⁺ T cells. Fluorescence minus one (FMO) depicted as negative control in histograms. Data are representative of n = 4–6 from two independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant (p R 0.5); one-way ANOVA with Dunnett’s multiple comparison test. Data are shown as mean ± SD.

Figure 3.. Sel1L is induced in antigen-experienced cells

(A) (Left) Representative immunoblot of Sel1L and β-actin in serially collected in vitro naive, T_ACT, IL-2 T_E, and IL-15 T_M cells (10 mg lysates). (Right) Densitometry quantification of the immunoblot band normalized to β-actin. (B) (Left) Representative immunoblot of uninfected naive CD8⁺ T cells and gp33⁺ and NP396⁺ CD8⁺ T cells (2 × 10⁵ cells) from mice on day 8 (D8) with LCMV. (Right) Densitometry quantification of the immunoblot band normalized to β-actin. Data are representative of n = 5–6 from two independent experiments. *p < 0.05, **p < 0.01, Friedman test with Dunnett’s multiple comparisons test (A) and Kruskal-Wallis test with Dunnett’s multiple comparisons test (B).

Figure 4.. Sel1L/ERAD is required for optimal CD8⁺ T cell effector function

Peripheral blood lymphocytes from Sel1LcKO P14 (CD45.2) and WT P14 (CD45.1/2) mice were mixed to generate a 1:1 mix of donor P14 cells, which was subsequently transferred into B6.SJL (CD45.1) mice. These mice were infected with LCMV-Armstrong the following day. At day 8 p.i., splenocytes were stimulated ex vivo with gp33 peptide. (A) Representative plots of intracellular TNF-α and IFN-γ expression (left) and frequencies (top right) of indicated cell populations in donor P14 cells, as well as MFI of indicated cytokine among cytokine⁺ populations (bottom right). (B) Cytokine co-expression in ex vivo restimulated WT or KO P14 cells at day 8 p.i. Dashed lines separate columns by number of cytokines produced. Data are representative of n = 8/genotype from two independent experiments. *p < 0.05, **p < 0.01; ns, not significant; paired t test.

Figure 5.. Sel1L/ERAD is required for CD8⁺ T cell survival and memory formation

Sel1LcKO P14 (CD45.2) and WT P14 (CD45.1/2) cells were mixed 1:1 and adoptively co-transferred into B6.SJL (CD45.1) mice, which were infected with LCMV-Armstrong the following day. Flow-cytometric analysis of donor P14 cells was performed on day 8 (D8; A–E) and peripheral blood mononuclear cells (PBMCs) serially collected (F). (A) (Left) Representative flow-cytometric analysis of donor P14 input and at D8 p.i. (center). (Right) Donor frequencies of indicated genotypes in spleen (SP), lymph nodes (LN), and peripheral blood (PB). (B) (Left) Representative flow-cytometric analysis of Ki67 in naive CD8⁺ T cells as negative control and donor P14. (Right) Frequency of Ki67⁺ population in donor P14 cells. (C) Frequency of 7AAD⁺ Annexin V (AV)⁺ populations among donor P14. (D) Representative flow-cytometric analysis of CD127 and KLRG1 expression in donor P14 cells. (E) MFI of TCF-1 in donor P14 populations. (F) Frequency of indicated genotypes among donor P14 cells in PBMCs at indicated time points ± SEM. Data are representative of n = 6/genotype from two independent experiments (A and B); n = 8/genotype from two independent experiments (C); n = 5–9 from 2–3 independent experiments (D); n = 11 from three independent experiments (E); n = 12 from three independent experiments (F). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant (p > 0.05); paired t test.

Figure 6.. Sel1L/ERAD regulates CD8⁺ T cell metabolism

(A) Pathway analysis of pathways significantly altered (false discovery rate <0.05) in Sel1LcKO P14 relative to WT P14 transcriptome at day 8 p.i. (B) Extracellular flux analysis of activated Sel1LcKO P14 and WT P14 cells to assess OXPHOS activity including basal oxygen consumption rate (OCR), maximalOCR, and metabolic fitness as measured by spare respiratory capacity (SRC). (C) Extracellular flux analysis of activated Sel1LcKO P14 and WT P14 quantifying basal extracellular acidification rate (ECAR). (D and E) MFI of TMRM (D) or MitoTracker green (E) in Sel1LcKO P14 and WT P14 cells at day 8 p.i. (F) (Left) Representative immunoblot of Sel1L, c-Myc, and β-actin in WT and KO CD8⁺ T_ACT and (right) densitometry quantification of the c-Myc immunoblot bands in indicated conditions normalized to β-actin, then to WT. Data are representative of n = 3/genotype from one independent experiment (A), n = 3/genotype from three independent experiments (B and C), n = 7/genotype from two independent experiments (D), n = 12/genotype from three independent experiments (E), and n = 5 WT, n = 4 KO from two independent experiments (F). *p < 0.05, **p < 0.01, ****p < 0.0001; ns, not significant (p > 0.05); unpaired t test (B, C, and F) or paired t test (D and E). Data are shown as mean ± SD. (B, C, and F)

Figure 7.. Human CD8⁺ T cells experience of ER stress is associated with terminal differentiation and reduced persistence

(A) (Left) Principal component (PC) analysis of human CD8⁺ T cell transcriptomes from peripheral blood of healthy donors from GSE179613. (Right) PC analysis from left colored by fold enrichment of genetic signature (GSVA) “GOBP response to endoplasmic reticulum stress” relative to naive cells. (B) Log₂ intensity of representative ER stress markers PDIA6, PERK, BiP, and ERO1L in indicated subset and activation state from human CD8⁺ T cell proteomes from PXD004352. (C) Representative flow cytometry histograms of (top) PROTEOSTAT and XBP1s MFI and (bottom) PROTEOSTAT and XBP1s in CD8⁺CD3⁺ cells from healthy human donors 3 days after activation with CD3/CD28 Dynabeads + IL-2 or rested in IL-7. (D) Leading-edge plots of gene set enrichment analysis (GSEA) of human T cells undergoing continuous stimulation from bispecific CD3xCD19 (AMG 562) T cells versus cells that experienced a treatment-free interval (TFI) at day 15 from GEO: GSE196463. (E) Leading-edge plots of GSEA of continuously stimulated GD2 CAR-T cells versus transiently rested cells at day 15 from GEO: GSE164950. n = 7–11/subset (A), n = 4/subset (B), n = 6 PROTEOSTAT, n = 3 XBP1s (C) from 2–3 independent experiments; n = 3/group (D), n = 3/group (E). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant (p > 0.05); unpaired t test (B) or paired t test (C).