Single-cell RNAseq identifies clonally expanded antigen-specific T-cells following intradermal injection of gold nanoparticles loaded with diabetes autoantigen in humans

Abstract

Gold nanoparticles (GNPs) have been used in the development of novel therapies as a way of delivery of both stimulatory and tolerogenic peptide cargoes. Here we report that intradermal injection of GNPs loaded with the proinsulin peptide C19-A3, in patients with type 1 diabetes, results in recruitment and retention of immune cells in the skin. These include large numbers of clonally expanded T-cells sharing the same paired T-cell receptors (TCRs) with activated phenotypes, half of which, when the TCRs were re-expressed in a cell-based system, were confirmed to be specific for either GNP or proinsulin. All the identified gold-specific clones were CD8⁺, whilst proinsulin-specific clones were both CD8⁺ and CD4⁺. Proinsulin-specific CD8⁺ clones had a distinctive cytotoxic phenotype with overexpression of granulysin (GNLY) and KIR receptors. Clonally expanded antigen-specific T cells remained in situ for months to years, with a spectrum of tissue resident memory and effector memory phenotypes. As the T-cell response is divided between targeting the gold core and the antigenic cargo, this offers a route to improving resident memory T-cells formation in response to vaccines. In addition, our scRNAseq data indicate that focusing on clonally expanded skin infiltrating T-cells recruited to intradermally injected antigen is a highly efficient method to enrich and identify antigen-specific cells. This approach has the potential to be used to monitor the intradermal delivery of antigens and nanoparticles for immune modulation in humans.

Keywords: immunomodulation; nanoparticles; proinsulin peptide; scRNAseq; type 1 diabetes.

Figure 1

Macroscopic and microscopic appearance of skin changes following intradermal injection of C19-A3 GNP. (A) Images show representative injection sites in deltoid region immediately after injection (left), several months after injection (centre) with the enlarged image showing dark centre and surrounding erythematous induration (right); (B, C) Sections from punch skin biopsies of participants EEASI-A and EEASI-B were stained for CD20, CD4, CD8 and Ki-67 and imaged with a fluorescence confocal microscope. Six fields of view were analysed. CD8⁺ T Lymphocytes predominated in all parts of the immune clusters: total overall counts/1.54mm² area: CD8⁺=6537, CD4+=1526, CD20⁺=1363/1.54mm²; (D) number of Ki67⁺ cells (expressed as mean+/-SEM) was higher amongst CD8⁺ cells (29.83+/-4.65) vs CD20⁺ (5.58+/-2.33) and CD4⁺ (29.83+/-4.65) cells. Punch biopsies were performed 7 and 3 months after injection in participants EEASI-A and EEASI-B, respectively.

Figure 2

scRNAseq reveals memory T-cell infiltration of injection site of C19-A3 GNP and clonal expansions (A) UMAP plot of pooled samples of cells from the blister fluid of 3 donors (B) UMAP plots of the blister fluid cells split by donor; (C) UMAP plot of all clonally expanded T-cells in the blisters (clones share an identical TCRα and TCRβ CDR3 at nucleotide level, and V and J gene usage, or where one chain is not detected, the remaining chain is identical); (D) UMAP plot of clonally expanded T-cells in the blister split by donor. Suction blisters were performed in three participants EEASI-A, EEASI-B and EEASI-C, 28, 10 and 10 months after injection, respectively. T_EM -effector memory T-cell, T_RM – tissue resident T cells, T_TR/EM – true tissue resident T-cell with some characteristic of effector memory, T_PM – peripheral memory T-cell, Treg – T regulatory T-cell, DC – dendritic cell.

Figure 3

Antigen specificity of re-expressed TCRs. TCRs were re-expressed in 5KC cell lines. Each graph represents an individual 5KC cell line with a single TCR. (A) GNP reactivity was measured by fluorescent reporter downstream of the NFAT promoter. CD3 - αCD3/CD28 positive control; EBV - patient-autologous EBV transformed cells used to present antigen-negative control; GNP: EBV transformed cells used to present GNP; C19-A3 GNP: EBV transformed cells used to present C19-A3 GNP; EBV transformed cells used to present PI - 50µg proinsulin. Data are pooled from three independent experiments. (B) Proinsulin reactivity was assessed by mIL-2 production measured by ELISA, with background mIL-2 from 5KC+ EBV cells alone subtracted. GNP: 25µL GNP, PI: proinsulin concentration in µg/ml. Data are pooled from at least three independent experiments. Values indicate results of unpaired T-tests, *p<0.05.

Figure 4

Specificities and T-cell phenotypes of clonally expanded cells in blister fluid. (A) combined UMAP plot of PI-specific clones; (B) combined UMAP plot of GNP-reactive clones; (C) combined UMAP plot of clones of unknown specificity; (D) UMAP plots of individual T-cell clones. T_EM -effector memory T-cell, T_RM – tissue resident T cells, T_TR/EM – true tissue resident T-cell with some characteristic of effector memory, Treg – T regulatory T-cell.

Figure 5

Top 5 DEG between PI-specific clones, gold-specific clones and clones of unknown specificity (designated non). (A) expression of KIR2 DL1; (B) expression of KIR3 DL1; (C) expression of granulysin (GNLY); expression of KIR3 DL1; (D) expression of CD300A; (E) expression of TYROBP; (F) KLRC2; (G) KLRD1.