CD160 isoforms and regulation of CD4 and CD8 T-cell responses

Abstract

Background: Coexpression of CD160 and PD-1 on HIV-specific CD8+ T-cells defines a highly exhausted T-cell subset. CD160 binds to Herpes Virus Entry Mediator (HVEM) and blocking this interaction with HVEM antibodies reverses T-cell exhaustion. As HVEM binds both inhibitory and activatory receptors, our aim in the current study was to assess the impact of CD160-specific antibodies on the enhancement of T-cell activation.

Methods: Expression of the two CD160 isoforms; glycosylphosphatidylinositol-anchored (CD160-GPI) and the transmembrane isoforms (CD160-TM) was assessed in CD4 and CD8 primary T-cells by quantitative RT-PCR and Flow-cytometry. Binding of these isoforms to HVEM ligand and the differential capacities of CD160 and HVEM specific antibodies to inhibit this binding were further evaluated using a Time-Resolved Fluorescence assay (TRF). The impact of both CD160 and HVEM specific antibodies on enhancing T-cell functionality upon antigenic stimulation was performed in comparative ex vivo studies using primary cells from HIV-infected subjects stimulated with HIV antigens in the presence or absence of blocking antibodies to the key inhibitory receptor PD-1.

Results: We first show that both CD160 isoforms, CD160-GPI and CD160-TM, were expressed in human primary CD4+ and CD8+ T-cells. The two isoforms were also recognized by the HVEM ligand, although this binding was less pronounced with the CD160-TM isoform. Mechanistic studies revealed that although HVEM specific antibodies blocked its binding to CD160-GPI, surprisingly, these antibodies enhanced HVEM binding to CD160-TM, suggesting that potential antibody-mediated HVEM multimerization and/or induced conformational changes may be required for optimal CD160-TM binding. Triggering of CD160-GPI over-expressed on Jurkat cells with either bead-bound HVEM-Fc or anti-CD160 monoclonal antibodies enhanced cell activation, consistent with a positive co-stimulatory role for CD160-GPI. However, CD160-TM did not respond to this stimulation, likely due to the lack of optimal HVEM binding. Finally, ex vivo assays using PBMCs from HIV viremic subjects showed that the use of CD160-GPI-specific antibodies combined with blockade of PD-1 synergistically enhanced the proliferation of HIV-1 specific CD8+ T-cells upon antigenic stimulation.

Conclusions: Antibodies targeting CD160-GPI complement the blockade of PD-1 to enhance HIV-specific T-cell responses and warrant further investigation in the development of novel immunotherapeutic approaches.

Figure 1

Expression of CD160 isoforms in primary CD4 ⁺ T-cells and binding to HVEM. A) Left panel: Representative FACS analysis of CD160 on primary CD4⁺ T-cells isolated from total PBMCs of a healthy donor (ex vivo at baseline), gated on CD3⁺CD4⁺CD8⁻ cells. Middle panels: CD160 surface expression following 48 h of resting (non-stimulated, NS) or TCR activation (plate-bound anti-CD3 and soluble anti-CD28). Right panel: overlapping histograms showing CD160 surface expression from TCR-stimulated CD4⁺ T-cells (dotted empty histogram) in comparison to 48 h rested CD4 (filled grey histogram) and freshly isolated CD4 cells (filled black histograms) all from the same individual donor. B) Frequency of CD160⁺CD4⁺ double positive population following 48 h of resting or TCR stimulation compared to freshly isolated (ex vivo) cells (n = 3). C) Kinetics of CD160-GPI and CD160-TM isoform expression at the mRNA level by quantitative RT-PCR in primary CD4⁺ T-cells (cells from n = 3 independent healthy donors) stimulated through TCR for 4 days. Values are relative to the house-keeping GAPDH gene transcripts (n = 3). HeLa cells were used as a negative control for CD160 TM transcription. D) Binding of HVEM to the two isoforms CD160. Left panel: Schematic representation for the TRF binding assay between CD160 (over-expressed by CHO-K1 cells) and the soluble ligand HVEM containing the human Fc1 (detection with anti-human Fc1). Right panel: Measuring the signal/background (S/B) for HVEM binding to both CD160-GPI and CD160-TM cells by the TRF assay under decreasing concentrations of HVEM-Fc.

Figure 2

Differential inhibition of HVEM/CD160 binding with benchmark tool antibodies. TRF assay measuring the potency of different antibodies to inhibit the binding of recombinant human HVEM-Fc chimera to CD160⁺ CHO-K1 cells. A) CD160 monoclonal antibodies inhibit binding of HVEM-Fc to both CD160-GPI and CD160-TM isoforms. B) Polyclonal HVEM (left panel) and monoclonal HVEM (right panel) antibodies both enhance binding of HVEM-Fc to CD160-TM isoform. The polyclonal anti-HVEM inhibits HVEM-Fc binding to CD160-GPI (left panel). Antibody concentrations are plotted on the X axis whereas, the calculated percentage of inhibition of binding is plotted on the Y axis. Matched isotype control antibody for each individual antibody candidate was also used in the assay (empty circles and squares). CTL = control, mAb = monoclonal antibody, pAb = polyclonal antibody.

Figure 3

Triggering of CD160-GPI is consistent with a positive co-stimulation role. A) Triggering of primary CD4⁺ T-cells with either plate-bound anti-CD3 (1 μg/ml) and anti-CD28 (0.5 μg/ml) or anti-CD3, anti-CD28 and HVEM-Fc (0.2 μg/ml) in the presence or absence of either anti-HVEM (left panel) or anti-CD160 clone CL1-R2 (right panel). IL-2 was measured in the supernatant by ELISA at 24 h post stimulation. Iso-IgG represents the matched isotype control antibody. P values were determined by two-tailed paired t test (data from three independent healthy donors). B) Left panels: Surface expression of CD160-GPI and CD160-TM on Jurkat-NFAT-Luc cells stably-transfected with CD160 plasmids. Mock-transfected cells (light grey histograms in middle and right panels) were used to set the positive and negative gates for FACS. CD160-TM is weakly detected with CD160-GPI antibodies (BY55 clone). Right panels: Quantitative RT-PCR for CD160-GPI and CD160-TM isoforms in Jurkat cells over-expressing either CD160-GPI or CD160-TM, values are relative to the house-keeping GAPDH gene transcripts (One representative experiment, n = 2). Non-transfected Jurkat (control cells) and HeLa cells were used as additional negative controls for CD160 expression. The left graph represents results with a set of Taqman probes that were not isoform selective and hybrdize both CD160-GPI and CD160-TM to demonstrate similar RNA expression levels, whereas the right graph used a set of probes that were CD160-TM specific to confirm the exclusive expression of the different CD160 isoforms in the two cell lines. C) Simultaneous triggering of TCR and CD160 using magnetic Dynal beads coated with anti-CD3, anti-CD28 and either HVEM-Fc (left panels), CD160 monoclonal antibodies (right panels) or their matched IgGs. Cell activation was monitored by measuring the absolute luciferase counts. Control cells are original Jurkat-NFAT-Luc cells non-transfected with either of the CD160 isoforms. NS: non-stimulated. P values were calculated by non-parametric two-tail t test (Mann–Whitney).

Figure 4

CD160 and HVEM antibodies specifically enhance IFNγ production by CD4 T-cells in response to the Tetanus toxoid recall antigen. A) IFNγ production by total PBMCs (1.5 × 10⁶ cells/ml) stimulated with 2.5 μg/ml of Tetanus toxoid in the presence or absence of anti-CD160 clone CL1-R2 mAb (left panel), HVEM pAb (right panel) or their matched isotype control Abs (n = 10). IFNγ was measured by ELISA from the supernatant following 5 days of stimulation. B) Left panel: Surface staining of cells using anti-CD3, anti-CD4, anti-CD8, anti-CD25 and anti-CD134 (OX40) analyzed by FACS (gating on CD3⁺ lymphocytes followed by gating on either CD4⁺ or CD8⁺ T-cells). Right Panel: analysis of the frequency of CD25⁺CD134⁺ double positive CD4⁺ and CD8⁺ T-cell populations (n = 5). P values were determined using the nonparametric Kruskal-Wallis and Dunn’s post-test.

Figure 5

Enhanced CD8 ⁺ T-cell proliferation to antibody-mediated blockade of PD-1 in combination with either CD160 or HVEM antibodies. A) Histogram summarizing the phenotypic analysis showing the frequencies of CD160⁺PD-1⁺ double positive population on total CD8 (left panel: each dot represents an independent staining from the same subject) and HIV-specific (right panel) T-cells from the four recruited study subjects. L933 represents the HIV-uninfected donor used as a control. HIV pentamers from each subject is annotated above each bar (right panel). Gating was done on CD3⁺ lymphocytes followed by gating on either total CD8⁺ T-cells or pentamer HIV-specific CD8⁺ T-cells. Note that we were unable to fold the peptides identified in the ELISPOT assay with HLA-restricted multimers from cART-treated subject. B) CFSE lymphoproliferation assays on total PBMCs from the viremic subjects NF-1042 and KBC-1035 gated on CD3⁺CD4⁻CD8⁺ T-cells. PBMCs stimulated or not with Gag7876 (restricted by HLA-B*1501 for KBC-1035, and HLA-A*0301 for NF-1042) in the absence or presence of blocking antibodies (4 replicates for each condition). C) PBMCs from ST-1041 and RJP-1038 stimulated or not with Pol5683 (restricted by HLA-A*11, A*03, A*68) and Gag7937, restricted by HLA-B*2705, respectively (4 replicates for each condition). P values were determined using the nonparametric Kruskal-Wallis and Dunn’s post-test.