MEK inhibition prevents human skin graft rejection by promoting CD8+TCF1+ over CD8 effector T cells

Abstract

Pharmacological MEK inhibition might be an innovative approach to complete the immunosuppressive regimen that enables solid organ transplantation. While MEK inhibitors like trametinib are approved in oncology, their immunomodulatory properties remain poorly investigated in the context of organ transplantation, especially in human context. Using a human skin transplantation model in NSG mice reconstituted with third-party human PBMCs, we evaluated the effects of trametinib on graft survival and the human allogeneic immune response. MEK inhibition significantly prolonged graft survival without reducing graft infiltrate, while preserving the human epidermal tissue. Single-cell RNA sequencing of splenic cells revealed that MEK inhibition impaired CD8⁺ T cell differentiation into effector phenotypes, favoring an accumulation of CD8⁺ TCF1⁺ stem-like cells. Additionally, MEK inhibition supported CD4⁺ T cell homeostasis by maintaining IL-7R expression. These findings suggest that MEK inhibition may simultaneously control the alloimmune response and support immune recovery, highlighting its potential in solid organ transplantation.

Keywords: Cell biology; Immune response; Transcriptomics.

Graphical abstract

Figure 1

MEK inhibition delays allogenic human skin graft rejection without affecting immune reconstitution (A) NSG mice received a 1 cm² split-thickness human skin graft followed by 5–10 × 10⁶ hPBMCs i.v (allogeneic to the skin) 7 weeks later. Trametinib (0.9 mg/kg) or vehicle was given daily by oral gavage. (B) Bar plot showing human chimerism assessed in the peripheral blood of mice at day 14. Vehicle treated group n = 15, trametinib treated group, n = 19. (C–E) Flow cytometry analysis of a replicated experiments on spleen harvested from animals at day 14 after hPBMCs injection, vehicle-treated group n = 9, trametinib-treated group n = 8. (C) Bar plots showing the human chimerism, the human CD4/CD8 ratio and the total number of CD4 and CD8 T cells. (D and E) Representative dot plot of flow cytometry analysis showing the T cell populations of naive T cells, (CD45RA⁺CD27⁺), central memory (CD45RA⁻CD27⁺), effector memory (CD45RA⁻CD27⁻) and TemRA (CD45RA⁺CD27⁻) T cells, along with bar plots showing the frequencies of each population and total spleen cell number for hCD4 (D) and hCD8 (E) T cell population in vehicle-treated mice (n = 10) and trametinib-treated mice (n = 9). (F and G) (F) Kaplan-Meier skin graft survival curve. Trametinib treatment delays allogeneic human skin rejection compared with vehicle-treated controls (30.7 ± 4.3 days, n = 17 vs. 18.4 ± 3.3 days, n = 15, log rank test ∗∗∗∗p < 0.0001. Pooled data of n = 3 experiments with skin from 3 different HD and hPBMC from 3 other different HD (G) Bar plot showing skin graft rejection scores at day 14 among vehicle-treated mice (n = 15) and trametinib-treated mice (n = 19). (H) Photographs representative of human skin grafts at day 0 or day 14 after allogeneic hPBMCs injection in vehicle and trametinib-treated animals. (B) and (D) are pooled data of n = 3 experiments with skin from 3 different HD and hPBMC from 3 other different HD. (C–E), and (G) are pooled data of n = 4 experiments with skin from 2 different HD and allogeneic hPBMC from 4 other different HD. (B–E), and (G) are shown as means ± SD, Mann-Whitney test. ns: non-significant, ∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 2

MEK inhibition limits the onset of human skin epithelial cell injury without affecting CD4 and CD8 allogeneic T cell infiltration Human skin grafts were harvested from transplanted NSG mice 14 days after the injection of allogeneic hPBMCs. Mice were daily treated with trametinib or vehicle from days 0–14. (A) Bar plot showing skin graft rejection blinded Banff score, using a validated grading scale, in vehicle- (n = 10) and trametinib-treated mice (n = 9), Fisher’s exact test ∗p < 0.05. (B) Representative H&E staining illustrating the difference between grade 3A (only single cell keratinocyte necrosis) and 3B (focal and full thickness necrosis). (C) Representative images of immunohistochemistry staining Ki67 (purple line delineates dermal/epidermal junction). (D) Bar plot showing the means ± SD of proliferation index of Ki67. (E) Representative images showing fluorescence patterns of hCD4 (green) and hCD8 (red) in human skin grafts of vehicle- and trametinib-treated mice. (F and G) (F) Bar plot showing the means ± SD of CD4 and CD8 human T cells (top) and the CD4/CD8 ratio (bottom) in the skin grafts (G) Representative images showing fluorescence patterns of CD4 (green) and FoxP3 (red) in human skin grafts. (H) Bar plot showing the means ± SD of human CD4 FoxP3 positive T cells (top) and their proportion among human CD4 T cells (bottom) in human skin grafts from vehicle and trametinib-treated mice. (D), (F), and (H) are pooled data of 4 experiments, vehicle-treated group n = 10, trametinib-treated group n = 8. Significance was calculated by non-parametric Mann-Whitney test. ∗∗p < 0.01.

Figure 3

scRNA-seq cluster analysis of human CD45-positive cells isolated from the spleens of human skin grafted humanized mice (A) Mice treated with the vehicle or trametinib (two in each group) were sacrificed on day 14 post-reconstitution, living (negative DAPI staining) hCD45 positive spleen cells were sorted, labeled with oligo-tagged antibodies, pooled and subjected to 3′ single-cell RNA-seq on the 10× platform with the V(D)J sequencing. (B) UMAP projections based on reduction dimension using PCA applied to the entire dataset (after quality control) and colored by gene expression of B cell (CD79A) and T cell (CD3E, IL7R, CD8A) marker genes. (C) UMAP projections based on the reduction dimension after subsetting the dataset to the T cell subset, colored according to the result of clustering with the Louvain algorithm. (D) UMAP projections based on reduction dimension using PCA applied to the T cell subset and colored by gene expression of CD8 (CD8A, GZMK), CD4 (CD4, IL7R) and CD4 Treg (FOXP3, IL2RA) T cells marker genes. (E) Stacked barplot representing the proportion of each T cell subset (CD8, CD4 non-Treg, CD4 Treg) according to the treatment exposure (vehicle vs. trametinib).

Figure 4

The CD8 compartment reveals a differentiation pathway from CD8⁺TCF1⁺ to effector CD8 cells impaired by trametinib (A) Dot plot depicting the expression profiles (average normalized expression and percentage) of genes identified as markers base on differential expression analysis, across the three CD8 clusters. (B) UMAP projection restricted to the CD8 subset, colored by cluster and splited by treatment (vehicle vs. trametinib treated). (C) Stacked barplot representing the proportion of each CD8 cluster cell subset (TCF1⁺, glycolytic, effector). (D) Trajectory analysis using Monocle 3, on the left UMAP showing cells colored by cluster assignment, on the right: UMAP underlining the trajectory - starting at the bottom left and ending at the top right - which defines the pseudotime coordinate of each cell. (E) Relative expression of main marker genes for each cell along pseudotime and colored by cluster. (F) Gene signatures from a set a gene up regulated in CD8+TCF1+ cells as reported in viral and tumoral models of persistent antigen, see Figure S8 for detail. (G–K) Analysis of T cell clonotype, (G) percentage of cells into each cluster with a unique clonotype, (H) distribution of CDR3 region length per cluster. (I) Stacked barplot representing for each cluster the proportion of cells belonging to the different clonotype groups ranging from rare to hyperexpended (J), Morosita’s overlap index for T cell receptor between CD8⁺ T cell subsets. (K) Plot representing the dominant CD8 clones across the three clusters.

Figure 5

Both flow cytometry analysis of spleens and immunofluorescence of skin grafts at Day 14 post PBMC infusion show enrichment in CD8⁺TCF1⁺ in trametinib-treated animals Human skin grafts and spleens were harvested from transplanted NSG mice 14 days after injection of allogeneic hPBMCs. Mice were treated with trametinib or vehicle from days 0–14. (A) Representative contour plots of flow cytometry analysis showing the expression of TCF1 or Granzyme A in CD8 T cells splenocytes of vehicle- or trametinib-treated animals. (B) Bar plots showing total cell number and percentage of CD8⁺TCF1⁺ and CD8⁺GZA⁺ in splenocytes of vehicle- or trametinib-treated animals. Data were from four distinct experiments for CD8⁺TCF1⁺ (vehicle-treated group n = 10, trametinib-treated group n = 8) and from two distinct experiments for CD8⁺GZA⁺ (vehicle-treated group n = 5, trametinib-treated group n = 4). (C) Representative histogram of PD1 expression in CD8⁺TCF1⁺ T splenocytes of vehicle of trametinib treated animals. Bar plot showing percentage of PD1⁺ cells in CD8⁺TCF1⁺ T splenocytes. Data were from four independent experiments (vehicle-treated group n = 10, trametinib-treated group n = 8). (D) Representative flow cytometry plots showing TCF1 and granzyme A expression in splenic CD8⁺ T cells from vehicle- or trametinib-treated mice., Bar plot showing percentage of CD8⁺TCF1⁺GZA⁻ and CD8⁺TCF1⁻GZA⁺ in splenocytes of vehicle- or trametinib-treated animals. Data were from two independent experiments (vehicle-treated group n = 5, trametinib-treated group n = 4). (E) Representative images showing fluorescence patterns of CD8 (green, membranous) and TCF1 (red, intranuclear) in human skin grafts with different scales, in vehicle- or trametinib-treated animals. White arrow highlights double-positive CD8⁺TCF1⁺ cells. (F) Bar graph showing quantification of double-positive CD8⁺TCF1⁺ in vehicle or trametinib-treated animals. Pooled data of 4 experiments, vehicle treated group n = 9, trametinib treated group n = 8). (B, C, D, and F) Data are shown as means ± SD, Mann-Whitney test. ∗p < 0.05.

Figure 6

An IL7R-high subset within the CD4 compartment is promoted by trametinib treatment (A) Dot plot depicting the expression profiles (average normalized expression and percentage) of genes identified as markers based on differential expression analysis, across the three CD4 clusters. (B) UMAP projection restricted to the CD4 subset, colored by cluster and split by treatment (vehicle vs. trametinib). (C) Stacked bar plot showing the proportion of each CD4 cluster (IL7R-high, Tfh-like, TH1-like). (D) Violin plot showing the relative expression of IL7R in each CD4 cluster according to their exposure to trametinib or vehicle. (E and I) Spleens were harvested from transplanted NSG mice at day 14 after the injection of allogeneic hPBMCs. Mice were treated with trametinib or vehicle from day 0 to day 14. (E and F) For IL7R staining, results are pooled data from 2 to 4 experiments, vehicle-treated group n = 5 to 9, trametinib-treated group n = 4 to 8. (E) Representative contour plots showing expression of IL7R in splenocytes CD4 T cells in vehicle or trametinib treated animals. (F) Bar plots showing the percentage and the total cell number of CD4⁺IL7R⁺ in the splenocytes of vehicle- or trametinib-treated animals. (G and H) Splenocytes were cultured in medium alone or were stimulated with PMA-ionomycin in the presence of Brefeldine A for 4 h and stained for detection of IFNg. (G) Representative contour plots showing IFNg expression in CD4 T cells. (H) Bar plots showing the percentages of IFNg expressing CD4 T cells and the mean fluorescence intensity of IFNg expression in vehicle- or trametinib-treated animal. Results are pooled data from 4 experiments, vehicle-treated group n = 10, trametinib-treated group n = 9. (I) Representative flow cytometry plots of the gating strategy of CXCR5⁺ PD1⁺ Tfh cells in splenocytes CD4 T cells from vehicle- or trametinib-treated mice. Histogram overlays showed expression of BCL6, ICOS, and TCF1 in Tfh cells from vehicle- or trametinib-treated mice. (J) Bar plots showing total cell number and percentage of Tfh cells in splenocytes of vehicle- or trametinib-treated animals. Results are pooled data from 4 experiments, vehicle-treated group n = 10, trametinib-treated group n = 9. (F, H, and J) Data are shown as means ± SD, non-parametric Mann-Whitney test. ∗p < 0.05, ∗∗∗p < 0.001.