To investigate the cell-intrinsic role of PSGL-1 in T cells over the course of chronic viral infection, we co-injected small numbers (1-2 x10³) of TCR transgenic (Tg) CD4⁺ and CD8⁺ T cells specific for LCMV into WT mice. Naïve WT (CD45.1⁺) and Selplg ^-/- (Thy1.1⁺) P14⁺ CD8⁺ and WT (CD45.1⁺) and Selplg ^-/- (Thy1.1⁺) SMARTA⁺ CD4⁺ T cells were adoptively transferred in WT (CD45.2⁺Thy1.2⁺) hosts at 1:1 equal ratio ( Figure 1A ). After one day, mice were infected with LCMV Cl13 and T cell responses were analyzed at 9, 15, 21, and 30 days post-infection (dpi) ( Figure 1A ). We compared the frequencies of WT and Selplg ^-/- P14⁺ CD8⁺ T cells in the spleen and observed a higher frequency and ratio of Selplg^-/- P14⁺ T cells to WT at 9dpi, however, Selplg^-/- P14⁺ T cell frequencies decreased by 15dpi and remained lower than WT at 30dpi ( Figures 1B, C ). We also observed lower frequencies of Selplg ^-/- P14⁺ T cells compared to WT T cells in lymph nodes at 30dpi ( Figure 1D ). We next examined the ratio and frequencies of CD4⁺ T cells and observed an initial increase in Selplg ^-/- SMARTA⁺ CD4⁺ T cells at 9dpi, which then significantly decreased throughout infection compared to WT T cells ( Figures 1E, F ). These findings showed the cell-intrinsic PSGL-1 inhibitory function in virus-specific T cells at the peak of anti-viral T cell expansion and revealed that virus-specific CD4⁺ and CD8⁺ T cells required PSGL-1 expression for their long-term maintenance during chronic viral infection.

We next characterized the exhaustion phenotype in T cells during chronic viral infection. We first assessed ex vivo cytokine production in the spleen of infected mice and observed similar frequencies of IFN-γ⁺ WT and Selplg^-/- P14⁺ T cells at 9 and 30dpi ( Figure 2A ). Exhausted T cells are characterized by their loss of polyfunctionality, and we confirmed a decrease in IFN-γ⁺TNF-α⁺ production by both WT and Selplg^-/- T cells, however Selplg^-/- P14⁺ T cells had a slight increase in these cytokine⁺ cells at 9 and 30dpi ( Figure 2A ). We further characterized CD8⁺ T cell effector functions and observed high expression of IFN-γ⁺CD107⁺ and GranzymeB⁺ T cells, with no differences between WT or Selplg^-/- P14⁺ T cells at 9 or 30dpi ( Figure 2B ). We evaluated proliferation by Ki67 staining and observed up to ~90% Ki67⁺ P14⁺ T cells at 9dpi and ~20% Ki67⁺ P14⁺ T cells at 30dpi, noting no differences between WT and Selplg^-/- P14⁺ T cells ( Figure 2C ). We next evaluated functional changes in CD4⁺ T cells and observed similar IFN-γ production in both WT and Selplg^-/- SMARTA⁺ T cells at 9dpi ( Figure 2D ). Although IFN-γ⁺ cells increased by 30dpi, there were no differences between WT and Selplg^-/- T cells ( Figure 2D ). We observed a slight increase in IFN-γ⁺TNF-α⁺ Selplg^-/- SMARTA⁺ T cells at 9dpi, however these levels were like WT T cells at 30dpi ( Figure 2D ). Selplg^-/- SMARTA⁺ T cells had increased Ki67⁺ cells at 9dpi, but no differences were observed between WT and Selplg^-/- SMARTA⁺ T cells at 30dpi ( Figure 2E ). We next evaluated PD-1 expression on both P14⁺ and SMARTA⁺ T cells and observed high PD-1 expression at 9dpi, with no differences between WT and Selplg^-/- T cells ( Figures 3A, B ). Although PD-1 expression decreased by 30dpi compared to 9dpi, expression levels remained high, with no differences between WT and Selplg^-/- T cells ( Figures 3A, B ). These findings showed that P14⁺ and SMARTA⁺ Selplg^-/- T cells are phenotypically and functionally exhausted to similar levels as WT T cells by 30dpi.

Since we observed a decrease in the frequencies of Selplg ^-/- P14⁺ cells by 30dpi ( Figure 1 ), we next determined whether this decrease was due to changes in Tpex or Tex cells. Using the same co-transfer approach, we analyzed the frequencies of Tpex (SLAMF6^hiTIM3^lo) and Tex (SLAMF6^loTIM3^hi) cells in WT and Selplg ^-/- P14⁺ T cells. We observed a decrease in Tpex frequencies and a slight increase in Tex frequencies in Selplg^-/- P14⁺ T cells at 21dpi ( Figures 4A, B ). We quantified the absolute number of Tpex and Tex cells in the spleens of infected mice and found decreased numbers of Tpex and Tex in Selplg^-/- P14⁺ T cells compared to WT T cells at 21 and 30dpi ( Figures 4C, D ). We observed decreased frequencies of PD-1⁺TIM3⁺, TOX⁺, and TCF-1⁺ Selplg^-/- P14⁺ T cells at 21dpi, which are also markers of Tpex cells ( Figures 4E–G ). These findings showed that in Selplg^-/- P14⁺ T cells, frequencies of Tpex cells were decreased, which may result in the decreased maintenance of the exhausted Selplg^-/- T cells over the course of chronic infection.

We next evaluated whether blocking the PD-1 pathway reinvigorated the response of exhausted Selplg^-/- P14⁺ T cells. We co-transferred WT and Selplg^-/- P14⁺ T cells in WT mice, infected with Cl13, and injected either IgG or anti-PD-L1 starting at 30dpi and every 3 days to 42dpi ( Figure 5A ). We evaluated T cell frequencies and phenotype in blood before (30dpi) and during antibody treatment ( Figure 5A ). We observed a significantly higher frequency of WT P14⁺ T cells (~3.5%) than Selplg^-/- P14⁺ cells (~1%) at 30dpi before antibody treatment ( Figure 5B ). However, at 35dpi, the frequency of Selplg^-/- P14⁺ T cells increased to similar levels as WT P14⁺ cells and remained higher to 45dpi ( Figure 5B ). Since WT and Selplg^-/- P14⁺ T cells were initially injected at a 1:1 ratio, we determined ratio and fold change throughout the course of anti-PD-L1 treatment. At 30dpi, WT P14⁺ T cells represented ~80% of the transferred population while Selplg^-/- P14⁺ T cells were ~20% ( Figure 5C ). However, by 35dpi and onwards, Selplg^-/- P14⁺ T cells increased to ~50% of the transferred cells, and they remained increased at 45dpi ( Figure 5C ). We determined the fold change of P14⁺ T cells before and after treatment and observed that Selplg^-/- P14⁺ T cells had a greater fold change from pre-treatment compared to WT P14⁺ T cells ( Figure 5C ). Since we observed an increase in Selplg^-/- P14⁺ T cells after anti-PD-L1 treatment, we determined whether there was a difference in their proliferation compared to WT T cells. Although Selplg^-/- P14⁺ trended for higher Ki67⁺ cells, there was no statistical significance compared to WT T cells pre-treatment (30dpi) ( Figure 5D ). However, after anti-PD-L1 treatment, Ki67⁺ WT P14⁺ T cells increased from ~18% pre-treatment to ~46% by 35dpi and decreased thereafter ( Figure 5D ). Ki67⁺ Selplg^-/- P14⁺ T cells increased from ~34% pre-treatment to ~82% by 35dpi and were significantly higher than WT T cells at 45dpi ( Figure 5D ). We also observed a significant increase in the frequencies of IFN-γ⁺, IFN-γ⁺CD107⁺, and IFN-γ⁺TNF-α⁺ Selplg^-/- P14⁺ T cells during anti-PD-L1 treatment ( Figures 5E, F ). These findings showed that anti-PD-L1 treatment reinvigorated the proliferation and effector function of exhausted Selplg^-/- P14⁺ T cells more than WT P14⁺ T cells.

We next determined whether anti-PD-L1 treatment changed Selplg^-/- SMARTA⁺ T cell responses. WT and Selplg^-/- SMARTA⁺ T cells were co-transferred into WT mice, which were then infected with Cl13 ( Figure 5A ). T cells were analyzed in the blood at 30dpi, before IgG and anti-PD-L1 injections, and during antibody treatment ( Figure 5A ). We observed significantly lower frequencies of Selplg^-/- SMARTA⁺ T cells than WT T cells at 30dpi, before anti-PD-L1 treatment ( Figure 5G ). However, after anti-PD-L1 treatment, Selplg^-/- SMARTA⁺ T cells increased to similar frequencies as WT T cells ( Figure 5G ). WT and Selplg^-/- SMARTA⁺ T cells were co-transferred at a 1:1 ratio, but this ratio changed to 4:1 by 30dpi ( Figure 5H ). Selplg^-/- SMARTA⁺ T cells increased after anti-PD-L1 treatment and approached a 1:1 ratio with WT T cells at 45dpi ( Figure 5H ). In addition, Selplg^-/- SMARTA⁺ T cells had a higher fold change from pre-treatment than WT SMARTA⁺ T cells ( Figure 5H ). We next assessed proliferation and observed a significant increase in Ki67⁺ Selplg^-/- SMARTA⁺ T cells at 35dpi with no differences at later timepoints ( Figure 5I ). We observed a significant increase in IFN-γ⁺ and IFN-γ⁺TNF-α⁺ in Selplg^-/- SMARTA⁺ T cells which peaked at 40dpi ( Figure 5J ). These findings showed that anti-PD-L1 treatment increased the accumulation, proliferation, and effector function of exhausted Selplg^-/- SMARTA⁺ T cells more than WT T cells.

We next assessed whether PSGL-1 expression impacted the CD8⁺ T cell response in melanoma tumors. We co-transferred WT and Selplg^-/- P14⁺ T cells at a 1:1 ratio in WT mice and one day later, subcutaneously injected B16-GP₃₃ melanoma cells ( Figure 6A ). We observed increased frequencies of Selplg ^-/- P14⁺ T cells in tumors compared to WT T cells 15 days after tumor injection ( Figure 6B ). Consistent with the increased frequencies, we also observed increased numbers of Selplg^-/- P14⁺ cells per gram of tumor ( Figure 6C ). We noted that the 1:1 pre-injection T cell ratio changed to 1:4 in tumors, with the increase of Selplg^-/- P14⁺ T cells ( Figure 6C ). We observed no differences in proliferation of WT or Selplg^-/- P14⁺ T cells in tumors as measured by Ki67⁺ cells ( Figure 6D ). We also noted that both WT and Selplg^-/- P14⁺ T cells in tumors had similar frequencies of IFN-γ⁺ and IFN-γ⁺CD107a⁺ T cells ( Figure 6E ). We next evaluated donor T cell frequencies in pooled tumor draining lymph nodes (TdLN) and observed similar frequencies of WT and Selplg ^-/- P14⁺ T cells ( Figure 6F ). Furthermore, there were no differences in cell numbers or the ratio of WT to Selplg^-/- P14⁺ T cells, in TdLN ( Figure 6G ). We did note a significant increase in Ki67⁺ cells in Selplg^-/- P14⁺ T cells in the TdLN ( Figure 6H ). We next evaluated whether anti-PD-L1 treatment changed the frequencies of WT and Selplg^-/- P14⁺ T cells and found similar ratios between IgG and anti-PD-L1 treated mice ( Figure 6I ). We observed that compared to IgG treated mice, anti-PD-L1 treatment decreased the frequencies of Selplg^-/- SLAMF6^hiTIM3^lo (Tpex) cells ( Figure 6J ). We observed increased GranzymeB⁺ Selplg^-/- P14⁺ T cells after anti-PD-L1 treatment ( Figure 6K ), and no significant differences in tumor mass between IgG and anti-PD-L1 treated mice ( Figure 6L ). These findings showed that Selplg^-/- P14⁺ T cells were increased in B16-GP₃₃ melanoma tumors.

We next determined the relevance of PSGL-1 (SELPLG) expression in CD8⁺ T cells from patients with Hepatitis C infection and analyzed RNA sequencing data from the Hensel et al. study (31). In short, low-input RNA sequencing was performed on HCV-specific CD8⁺ T cells from blood of patients infected with chronic Hepatitis C virus (cHCV) and HCV patients who were spontaneous resolvers (SPR) that cleared the viral infection. We observed significantly increased SELPLG expression in CD8⁺ T cells from cHCV patients compared to SPR patients ( Figure 7 ). These findings showed that SELPLG was highly expressed in CD8⁺ T cells during chronic Hepatitis C viral infection and was decreased when the virus was eliminated from the host.

C57BL/6J were purchased from Jackson Laboratory and then bred in specific-pathogen-free (SPF) facilities and maintained in biosafety level 2 (BSL-2) facilities after infection in the vivarium at UC Irvine. P14 and SMARTA TCR transgenic mice were obtained from The Scripps Research Institute (originally from Dr. Charles D. Surh). These mice were bred to Ly5.1 (B6.SJL-Ptprc^a Pepc^b/BoyJ) mice and to Thy1.1 (B6.PL-Thy 1^a/CyJ). Selplg^-/- mice were purchased from Jackson Laboratory and bred with the TCR transgenic mice in house. Both female and mice were used and greater than 6 weeks of age. All experiments were approved by the animal care and use committees at UC Irvine (AUP-18-148).

LCMV Clone 13 (Cl13) strain was propagated in baby-hamster kidney cells and titrated on Vero African-green-monkey kidney cells. Frozen stocks were diluted in Vero cell media and mice were infected by intravenous injection of 2x10⁶ plaque-forming units (PFUs) of LCMV Cl13.

Naïve WT or Selplg^-/- P14 and SMARTA T cells were isolated from mouse spleens by magnetic sorting (Stem Cell Technologies, using negative selection) in accordance with the manufacturer’s protocol. WT and Selplg^-/- P14 and SMARTA T cells were transferred in equal numbers (1x10³ P14 and 2x10³ SMARTA) in 200µL PBS into naïve C57BL/6 mice by intravenous (i.v) injection. One day later, WT mice were infected with 2x10⁶ PFU LCMV Cl13 by i.v. injection. For the melanoma tumor model, naïve WT or Selplg^-/- P14 T cells were isolated from mouse spleens by magnetic sorting (Stem Cell Technologies, using negative selection) according to the manufacturer’s protocol. WT and Selplg^-/- P14 T cells were transferred in equal numbers (1x10⁶) in 200µL of PBS into naïve C57BL/6 mice by intravenous injection (i.v). One day later, mice were injected with 200µL of 1x10⁶ B16-GP₃₃ tumor cells by subcutaneous (s.c) injection.

Cells from the spleens, lymph nodes (inguinal, axial, brachial), and tumors were dissociated in HBSS. For cell surface staining, 2x10⁶ cells were incubated in staining buffer (PBS, 2% fetal bovine serum (FBS) and 0.01% NaN3) for 20 minutes at 4°C with antibodies for expression of surface proteins at a 1:200 dilution. For intranuclear transcription factor staining, cells were fixed and permeabilized using a Foxp3/transcription factor fixation/permeabilization kit (Fisher) according to manufacturer’s protocol. For functional assays, cells from infected mice were cultured for 4 hours at 37°C with 2ug/mL of GP_33-41 or GP_61-80 peptides (AnaSpec) in the presence of brefeldin A (1µg/ml; Sigma-Aldrich), and IL2 (50u/ml). The cells were then stained with antibodies for expression of surface proteins, fixed and permeabilized using a Cytofix/Cytoperm kit (BD Bioscience), and stained with antibodies for intracellular cytokine detection at a 1:100 dilution. To evaluate cell degranulation, splenocytes were incubated with anti-CD107α. The culture media used for the 4-hour incubation was RPMI-1640 containing 10 mM HEPES, 1% non-essential amino acids and L-glutamine, 1mM sodium pyruvate, 10% heat-inactivated FBS, and penicillin/streptomycin antibiotics. The following antibodies were used in this study: Biolegend CD8 (53-6.7), CD4 (RM4-5), Vα2 (B20.1), CD90.1(OX-7), CD45.1 (A20), CD107α (1D4B), IFNγ (XMG1.2), TNFα (MP6-XT22), PD-1 (RMP1-30), TIM-3 (RMT3-23). Cell Signaling Technology TCF-1 (C63D9), Miltenyi Biotec TOX (REA473). BD Biosciences Vb8.1.2 (MR5-2), Ly-108 (13G3), Ki67 (RUO). Invitrogen GranzymeB (GB12).

200μg (200µL PBS) of rat anti-mouse PD-L1 (10F.9G2) was administered by intra peritoneal (i.p) injection five separate times in WT mice. Antibody injections started at 30 days post Cl13-infection and continued every three days until 42 days post infection (30-, 33-, 36-, 39-, and 42-dpi). In vivo mAb were purchased from BioXcell (New Hampshire, USA). For the tumor study, WT mice were co-injected with WT and Selplg^-/- P14⁺ T cells (1x10⁶ cells of each) i.v. and then injected with 1x10⁶ B16-GP₃₃ melanoma cells s.c. IgG (200µg) or anti-PD-L1(200µg) (10F.9G2) was injected i.p. at day 8, 10, 12 post melanoma cell injection and mice euthanized at 15 days post melanoma cell injection.

WT and Selplg^-/- P14⁺ T cells were transferred in equal numbers (1x10⁶) in 200µl into naïve C57BL/6 mice by intravenous injection (i.v). One day later, mice were injected with 200µl of 1x10⁶ B16-GP₃₃ tumor cells by subcutaneous injection (s.c). At day 15 after B16-GP₃₃ injection, tumors were excised, minced, and digested in gentleMACS C Tubes for 40 minutes at 37°C using a gentleMACS Dissociator (Miltenyi Biotec). Tumor digestions were then passed through a 70-μm cell strainer to obtain a single-cell suspension. Cells were plated at 2x10⁶ in 96 well plates and stained for flow cytometry.

Low-input RNA sequencing of HCV-specific CD8⁺ T cells was performed by Hensel et al., 2021 (31) [GSE150345]. Briefly, HCV-specific CD8⁺ T cells transcriptome analysis was done using an Illumina NextSeq 500 platform. For our analysis, RNA reads were aligned to the human reference genome and post-processing of the low-input sequencing data was performed at the Institute for Genomics and Bioinformatics (UCI IGB) at UC Irvine. Gene expression levels were compared in chronically infected HCV patients and HCV spontaneous resolved (SPR) HCV patients.

Flow cytometry data was analyzed with FlowJo software. Graphs were made using GraphPad Prism software. GraphPad Prism was used for statistical analysis to compare outcomes using a two-tailed unpaired t-test. Significance was set to p <0.05 and represented as *<0.05, **<0.001, ***<0.005, and ****<0.0001. Error bars show standard error of the mean (s.e.m.).