Varicella-zoster virus proteome-wide T-cell screening demonstrates low prevalence of virus-specific CD8 T-cells in latently infected human trigeminal ganglia

Abstract

Background: Trigeminal ganglia (TG) neurons are an important site of lifelong latent varicella-zoster virus (VZV) infection. Although VZV-specific T-cells are considered pivotal to control virus reactivation, their protective role at the site of latency remains uncharacterized.

Methods: Paired blood and TG specimens were obtained from ten latent VZV-infected adults, of which nine were co-infected with herpes simplex virus type 1 (HSV-1). Short-term TG-derived T-cell lines (TG-TCL), generated by mitogenic stimulation of TG-derived T-cells, were probed for HSV-1- and VZV-specific T-cells using flow cytometry. We also performed VZV proteome-wide screening of TG-TCL to determine the fine antigenic specificity of VZV reactive T-cells. Finally, the relationship between T-cells and latent HSV-1 and VZV infections in TG was analyzed by reverse transcription quantitative PCR (RT-qPCR) and in situ analysis for T-cell proteins and latent viral transcripts.

Results: VZV proteome-wide analysis of ten TG-TCL identified two VZV antigens recognized by CD8 T-cells in two separate subjects. The first was an HSV-1/VZV cross-reactive CD8 T-cell epitope, whereas the second TG harbored CD8 T-cells reactive with VZV specifically and not the homologous peptide in HSV-1. In silico analysis showed that HSV-1/VZV cross reactivity of TG-derived CD8 T-cells reactive with ten previously identified HSV-1 epitopes was unlikely, suggesting that HSV-1/VZV cross-reactive T-cells are not a common feature in dually infected TG. Finally, no association was detected between T-cell infiltration and VZV latency transcript abundance in TG by RT-qPCR or in situ analyses.

Conclusions: The low presence of VZV- compared to HSV-1-specific CD8 T-cells in human TG suggests that VZV reactive CD8 T-cells play a limited role in maintaining VZV latency.

Keywords: Herpes simplex virus; Human; Latency; T-cells; Trigeminal ganglion; Varicella-zoster virus.

Fig. 1

HSV-1 and VZV responses of human TG-derived T-cell lines. A Gating strategy applied to determine T-cell reactivity, consisting of selection of viable cells followed by size exclusion and selection of single cells, selection of CD3⁺ T-cells, and finally selection of CD3⁺CD4⁺ and CD3⁺CD8⁺ T-cells. FSC-A, forward scatter—area; SSC-A, side scatter—area; FSC-H, forward scatter—height. B Flow cytometry analysis of IFNγ production by CD4 T-cells derived from TG donors TG02–TG05 following 6 h of co-culture with autologous B-LCL pulsed with lysates derived from mock, HSV-1, or VZV-infected ARPE-19 cells. C Flow cytometry analysis of IFNγ production by CD8 T-cells following 6 h of coculture with autologous mock- or HSV-1-infected B-LCL. Numbers indicate percentages of cells within the respective gate

Fig. 2

VZV antigen specific response of human TG-derived CD8 T-cells. A, B TG-derived T-cell lines of donor TG04 (A) or TG07 (B) were screened for reactivity with artificial antigen presenting cells expressing the indicated subject-specific HLA class I allele(s) and individual VZV open reading frames or fragments thereof. For subject TG04, all 4 HLA alleles were co-transfected simultaneously, while for subject TG07, only HLA-A*11:01 was used. Empty vector (control) and phytohemagglutinin (PHA-P), were used as negative and positive controls, respectively. Levels of secreted IFNγ was determined by ELISA. Data are presented as the individual (circles) and mean (bars) OD₄₅₀ values of two independent replicates. Horizontal dashed lines indicate the threshold of T-cell response set at twice the value of the empty vector control. Viral ORF nomenclature according to VZV reference strain Dumas (Genbank Accession number X04370)

Fig. 3

TG-derived CD8 T-cells recognize an epitope in VZV ORF9 epitope in an HLA-A*11:01-dependent manner. A IFNγ secretion by TG07-derived T-cells following 24 h of incubation with Cos-7 cells co-transfected with a vector encoding HLA-A*11:01 and either an empty control vector or a vector encoding VZV ORF9. B Overview of the peptide composition of each ORF9 peptide pool used in panel C. Reactive pools are shaded. C IFNγ secretion by TG07-derived T-cells co-cultured with Cos-7 cells transfected with HLA-A*11:01 and pulsed with the indicated peptide pools, showing reactivity for two individual pools that overlap at peptide 28. D IFNγ secretion by TG07-derived T-cells following co-culture with HLA-A*11:01-transfected Cos-7 cells transfected and pulsed with ORF9 peptide 28 (amino acids 109–121, EDAVYENPLSVEK) or internal ORF9 peptide 28* (amino acids 111–121, AVYENPLSVEK). E Alignment of the VZV ORF9 epitope (AVYENPLSVEK, underlined) with the homologous regions of HSV-1 and HSV-2 UL49. Amino acids in bold differ between VZV and HSV-1/HSV-2; numbers indicate amino acid positions in the respective proteins. F IFNγ secretion by TG07-derived T-cells following co-culture with Cos-7 cells transfected with HLA-A*11:01 in combination with either an empty control vector or vectors encoding HSV-1 UL49 or HSV-2 UL49. Data are presented as the individual (circles) and mean (bars) OD₄₅₀ values of at two independent replicates

Fig. 4

TG-VZV reactive CD8 T-cells recognize an HSV/VZV cross-reactive epitope in an HLA-A*01:01-dependent manner. A IFNγ secretion by TG04-derived T-cells following incubation with Cos-7 cells transfected with the indicated subject-specific HLA class I allele together with an empty control vector (−) or a vector encoding VZV ORF29 (+). B Schematic of VZV ORF29 fragments used in (C) to map the T-cell epitope. Numbers indicate amino acid positions. C ELISA showing IFNγ secretion by TG04-derived T-cells following incubation with Cos-7 cells transfected with HLA-A*01:01 and an empty control vector (−) or vectors encoding the indicated VZV ORF29 fragments from (B). D IFNγ production by TG04-derived T-cells following coculture with Cos-7 cells transfected without (−) or with (+) HLA-A*01:01 and pulsed with the indicated ORF29 peptides or DMSO as a negative control (−). E Alignment of the VZV ORF29 epitope-containing region of VZV ORF29 with the homologous regions of HSV-1 UL29 and HSV-2 UL29 (underlined). Amino acid in bold differ between VZV, HSV-1, and HSV-2. Numbers indicate amino acid positions in the respective proteins. F IFNγ secretion by TG04-derived T-cells following co-culture with Cos-7 cells transfected to express the indicated HLA class I alleles in combination with either an empty control vector or vectors encoding full-length HSV-1 UL29 or HSV-2 UL29. Data are presented as the individual (circles) and mean (bars) OD₄₅₀ values of 2 (A, C, F) or 3 (D) independent replicates

Fig. 5

No correlation between T-cell infiltrates and VZV latency in human TG. A Consecutive TG sections were stained for CD3 by immunohistochemistry (red; left column) or VLT RNA by in situ hybridization (dark red, right column). Filled and open arrow heads indicate CD3⁺ cells and a VLT⁺ neuron, respectively; the inset shows a VLT⁺ neuron. Data is shown for two different areas (1 and 2) of one donor that are representative of five independent TG donors; scale bar: 50 µm. B, C Abundance of VZV DNA (B) or VZV VLT and VZV VLT-ORF63 RNA (C) as well as human CD3D transcripts in 18 TG specimens (described in [3]) as determined by q(RT)-PCR and presented as relative DNA/transcript levels normalized to single-copy gene HMBS or cellular housekeeping gene GAPDH. Spearman r and p values are indicated