Costimulatory signals via B7/CD28 family molecules (signal 2) are critical for effective adaptive CD8+ T cell immune responses. In addition to costimulatory signals, B7/CD28 family coinhibitory receptor/ligands that modulate immune responses have been identified. In acute hepatitis C virus (HCV) infection, programmed death receptor 1, an inhibitory receptor in the CD28 family, is highly expressed on virus-specific CD8+ T cells, yet vigorous immune responses often develop. We hypothesized that other costimulatory signals present during the acute phase of HCV infection would be important to counter this negative signaling. In this study, we found that CD86 was highly expressed on HCV-specific CD8+ T cells early in acute HCV infection and was lost on transition to chronic HCV infection; the expression of CD86 was different from other activation markers, because expression was delayed after in vitro TCR stimulation and required sufficient IL-2 signaling; and HCV-specific CD8+ T cells in the liver of patients with chronic HCV infection were highly activated (CD69, CD38, and HLA-DR expression), but only a minority expressed CD86 or showed evidence of recent IL-2 signaling (low basal phosphorylated STAT5), despite persistent viremia. Our study identified B7 ligand expression on HCV-specific CD8+ T cells as a distinct marker of effective T cell stimulation with IL-2 signaling in acute HCV infection. Expression of costimulatory molecules, such as CD86, early in HCV infection may be essential in overcoming inhibitory signals from the high level of programmed death receptor 1 expression also seen at this phase of infection.
A majority of patients infected with hepatitis C virus (HCV) do not spontaneously clear the virus and become chronically infected. It is hypothesized that in those individuals developing persistent infection, an effective adaptive T cell response fails to develop during the acute phase of HCV infection or is lost on progression to the chronic phase of infection (see review in Ref. 1). In the acute phase of HCV infection, a clinical hepatitis is often observed that is marked by a significant elevation in liver transaminase levels. In contrast, during chronic infection, a clinical hepatitis is often absent; only a mild elevation in liver transaminases is noted. Over many years, this mild inflammation can lead to liver fibrosis and the eventual development of cirrhosis. Evaluation of HCV-specific CD8+ T cells from the peripheral blood of patients with acute infection showed that a majority are highly activated, expressing the markers CD69, CD38, and HLA-DR (2–7). Likewise, during the chronic phase of infection, HCV-specific CD8+ T cells in the liver are highly activated and prevalent (7–9); however, they induce only mild liver injury, as measured by serum liver transaminase levels. The loss of functionality of HCV-specific CD8+ T cells, despite the continued high level of activation that is seen in chronic infection (at the site of infection), is not completely understood.
Recently, it was reported that the expression of programmed death receptor 1 (PD-1), an inhibitory receptor in the B7/CD28 family, during chronic viral infection was associated with a loss of T cell function and a state of exhaustion that could be reversed by PD-1 blockade (10). In chronic HCV infection, a high level of PD-1 was expressed on HCV-specific CD8+ T cells infiltrating the liver of patients with chronic HCV infection (11–13), and expression of PD-1 on these cells contributed to their decreased function, despite the high level of activation (12, 13). Somewhat surprisingly, high levels of PD-1 expression have also been noted on HCV-specific CD8+ T cells in the acute phase of infection, yet these cells still induce significant liver injury (7, 13–15). We hypothesized that other stimulatory signals are important to counteract this high PD-1 expression to enable vigorous T cell responses in acute infection.
B7 ligands, such as CD80 and CD86, are typically described as being expressed on APCs. However, these molecules can also be expressed on in vitro activated T cells (16–23), and the expression of these ligands on T cells is hypothesized to be important in T–T cell interactions (17). For example, after 10 d of stimulation with anti-CD3 and IL-2, >80% of human CD4+ and CD8+ T cells expressed CD80 (16), and stimulation of human PBMCs with anti-CD3 and IL-2 led to maximal CD86 expression (60% of T cells) 3 wk after stimulation (18). Expression of the B7 ligands CD80, CD86, and PD-L1 also was identified directly ex vivo on CD3+ T cells from patients with HIV infection (24, 25) and autoimmune diseases (26, 27), and the expression of these ligands was hypothesized to be a marker of disease progression (24–27).
Despite these studies, the expression of B7 family ligands on HCV-specific CD8+ T cells isolated from the peripheral blood or liver of patients with HCV infection has not been investigated. Because CD80 and CD86 expression on T cells can provide a costimulatory signal to other T cells (16, 18), the study of the expression of these ligands on HCV-specific CD8+ T cells in acute and chronic infection is important for a more complete understanding of the signals driving a functional adaptive T cell response to this virus. In the current study, we evaluated B7 ligand expression in four patients with acute HCV infection and detectable HCV-specific tetramer responses. We report that HCV-specific CD8+ T cells expressed high levels of CD86 in all of the patients in the acute phase of HCV infection at a time when significant liver injury was clinically apparent. We found that CD86 expression on CD8+ T cells was linked to effective TCR stimulation with sufficient IL-2 signaling. This differed from the expression of the activation markers CD69, CD38, HLA-DR, and CD25 that occurred rapidly after TCR stimulation alone, without the addition of IL-2 to in vitro cultures of PBMCs. Unlike in acute HCV infection, in chronic infection, despite the high-level expression of activation markers, the majority of HCV-specific CD8+ T cells did not express CD80 or CD86, even at the site of infection, and there was no evidence of recent common γ-chain cytokine signaling. This study improves our understanding of the deficits in T cell stimulation at the site of HCV infection: HCV-specific CD8+ T cells are activated partially (expression of CD69, CD38, and HLA-DR) but not effectively [low IL-2 signaling, low CD86 expression, and low proliferation (7)]. Furthermore, this study highlights the early loss of supportive cytokine stimulation of HCV-specific CD8+ T cells during the waning response to HCV infection and identifies B7 ligand expression (CD80 and CD86) on T cells as an indicator of effective TCR stimulation with supportive IL-2 signaling.
Materials and Methods
Four patients with acute HCV infection, as evidenced by HCV Ab seroconversion in the presence of a clinical syndrome of acute hepatitis, and 28 patients with chronic HCV infection (HCV Ab and HCV PCR positive) were enrolled in the study from the clinic or hospital of Emory University, Atlanta Veterans Affairs Medical Center, Grady Memorial Hospital, Atlanta, or the Montreal Acute Hep C Cohort at Centre de Recherche du Centre Hospitalier de l’Universite de Montreal (CRCHUM), Hôpital St-Luc, as previously described (28, 29). The patient characteristics are summarized in Table I. Acutely infected patients are denoted by an “a” preceding the patient number, and chronically infected patients are denoted by a “c” preceding the patient number. Nine of the chronically infected patients were enrolled in the Emory Liver Transplant Program, and liver specimens were procured at the time of hepatectomy for liver transplantation (explant liver). An “E” following the patient number denotes these patients. Patients a802, a240, a808, a4915, c113E, and c671 were HLA-A2 positive by FACS analysis, which enabled analysis with HLA-A2–restricted tetramers. None of the patients with acute HCV infection spontaneously cleared viremia. a802 began therapy with pegylated IFN and ribavirin at day 80 of the study; a802 was also HIV positive and had an undetectable HIV load on highly active antiretroviral therapy (HAART) for ∼9 mo prior to our analysis. There were no changes in HIV load (remained undetectable throughout the study). a240, a808, and a4915 were not treated for HCV infection during the course of this study. The patients each provided informed consent, and the protocol was approved by the local ethics committees of Emory University, the Atlanta Veterans Affairs Medical Center, Grady Health Systems, and CRCHUM.
HCV Ab testing, viral load determination, and genotyping
HCV Ab testing by ELISA was performed at the Emory Immunology Laboratory using a kit, per the manufacturer’s instructions (Abbott Laboratories, Abbott Park, IL); at the Atlanta Veterans Affairs Immunology Laboratory (Bio-Rad, Hercules, CA); and at CRCHUM, by standard Anti-HCV EIA1 and EIA2 tests (Abbott Laboratories). HCV load quantification was performed at the Emory Molecular Laboratory and Atlanta Veterans Affairs Medical Center, using a real-time RT-PCR assay (Roche Molecular Systems, Alameda CA), and at CRCHUM, by automated COBAS AmpliPrep/COBAS Amplicor HCV test, version 2.0 (sensitivity 50 IU/ml) (Roche Molecular Systems, Branchburg, NJ). HCV genotyping was performed at the Emory Molecular Laboratory, using a commercially available assay (Siemens Medical Solutions Diagnostics, Berkeley, CA); at the Atlanta Veterans Affairs Medical Center, using a line probe assay (Bayer Diagnostics, Research Triangle Park, NC); and at CRCHUM, using standard sequencing for the NS5B region by the Laboratoire de Santé Publique du Québec (Ste-Anne-de-Bellevue, Québec, Canada).
EDTA and heparin anticoagulated blood was collected from each patient and used directly for FACS staining or for PBMC isolation. PBMCs were isolated using Ficoll-Paque PLUS density gradient (Amersham, Oslo, Norway), washed twice in PBS, and analyzed immediately or cryopreserved in media containing 90% FCS (Hyclone, Logan, UT) and 10% DMSO (Sigma-Aldrich, St. Louis, MO).
Liver tissue obtained by ultrasound-guided needle biopsy, via transjugular fluoroscopic technique, or from explant liver was immediately put into RPMI 1640 medium (Life Technologies, Carlsbad, CA) containing 10% FCS (Hyclone) for immunological assays. Another fragment was fixed in formalin for histological examination. Classification of the histological changes was performed, using the Scheurer scoring system, at the Emory University and Atlanta Veterans Affairs Medical Center pathology laboratories.
Intrahepatic T cell isolation
The liver biopsy sample obtained by ultrasound-guided needle biopsy or via transjugular fluoroscopic technique was washed three times with RPMI 1640 medium containing 10% FCS to remove cell debris and RBCs. Isolation of liver-infiltrating lymphocytes was performed using an automated, mechanical disaggregation system (Medimachine, BD Biosciences, San Jose, CA). The sample was inserted into a 50-mm Medicon, inserted into the Medimachine, and run for 15 s. Disaggregated cells were removed using a syringe in the syringe port. The Medicon was rinsed twice with RPMI medium containing 10% FCS to ensure maximum cell recovery. Cells were used immediately for FACS staining. For explant liver samples, the liver section was perfused with HBSS, sectioned using sterile scissors, resuspended in 0.5% collagenase in F12 media, and disaggregated in a Stomacher 400 Circulator (Seward, Bohemia, NY) set at 200 rpm for 20 min. The suspension was washed in R10 and filtered using sterile autoclaved cheesecloth. To isolate liver-infiltrating mononuclear cells (LIMCs), the supernatant was resuspended in 40% Percoll, layered over 70% Percoll, and centrifuged at 2000 rpm for 10 min on low break. LIMCs were harvested from the interface layer and washed in R10. Cells were used immediately for FACS, phosflow, or culture.
Abs, HLA-A2 tetramers, and flow cytometry
Cells were stained with FITC-, PE-, PerCP-, APC-, and Pacific Blue-labeled mAbs or tetramers according to the manufacturers’ instructions, and flow cytometry was performed using FACSCalibur or LSRII. FACS data were analyzed with FlowJo software (Tree Star, Ashland, OR). The following mAbs were purchased from BD Pharmingen (San Diego, CA): anti-CD25 FITC, CD28 FITC, CD38 FITC, Ki67 FITC, IgG1k FITC, HLA-DR PE, CD86 PE, PD-L1 PE, PD-L2 PE, IgG1k PE, IgG2Ak PE, Granzyme B PE, Perforin FITC, Bcl-2 PE, CD8 PerCP, and CD3 Pacific Blue. CD80 PE, CD69 PE, CD127 PE, and CD8 APC were obtained from Beckman Coulter (Fullerton, CA). Anti–PD-1 PE-conjugated Ab (clone EH12) was obtained from BioLegend (San Diego, CA). Dead cells were excluded by staining with 7-aminoactinomycin D (7AAD; BD Biosciences) or Alexa Fluor 430 (Invitrogen). HLA-A2 tetramers were specific for the following CD8+ T cell epitopes: HCV/1073: CINGVCWTV; HCV/ 1406: KLVALGINAV; CMV/NLV: NLVPMVATV. The tetramers were generated at the National Tetramer Core Facility, Emory University School of Medicine.
Analysis of STAT phosphorylation in T cells
The following mAbs were purchased from BD Pharmingen: anti-phosphorylated STAT (pSTAT)1 A488 (clone 4a), pSTAT5 A488 (clone 47), pSTAT6 A488 (clone J71-773.58.11), IgG1k A488 (MOPC-21), CD86 PE (clone FUN-1), IgG1k PE (MOPC-21), and CD3 Pacific Blue (SP34-2). CD8 PECy5 (B9.11) was obtained from Beckman Coulter. For pSTAT analysis in HCV-specific CD8+ T cells, whole blood or LIMCs were stained with tetramer for 10 min at room temperature; incubated with no stimulation, IL-2 (Roche), IL-7 (R&D Systems, Minneapolis, MN), or IL-15 (R&D Systems) for 10 min (with tetramer) at 37°C; then immediately fixed in Lyse/Fix buffer I (BD Biosciences); washed with PBS; permeabilized with Perm III buffer (BD Biosciences) at 4°C; washed in FACS buffer (PBS containing 0.5% BSA and 0.1% NaN3); and stained for 30 min with anti-pSTAT and other Abs. For pSTAT analysis on cultured cells, cells were first washed in FACS buffer and then fixed with Fix Buffer I (BD Biosciences) prior to permeabilization. Flow cytometry was performed using FACS Calibur or LSRII cytometers. FACS data were analyzed with FlowJo software.
A total of 1 × 106 PBMCs or LIMCs in 1 ml RPMI 1640/10% FCS were cultured for 5–7 d. For specific Ag stimulation, HCV NS3 1073-1081 peptide (CINGVCWTV; 1 μg/ml), HCV NS3 1406-15 (KLVALGINAV; 1 μg/ml), or CMV NLV peptide (NLVPMVATV; 1 μg/ml) was used (Genemed Synthesis, San Antonio, TX). For nonspecific TCR stimulation, cells were incubated with anti-CD3 (Beckman Coulter, Somerset, NJ) at a concentration of 0.1 μg/ml. IL-2 (Roche), IL-7 (R&D Systems), or IL-15 (R&D Systems) was added to the culture media at the time of TCR stimulation at concentrations indicated in the Results section. For some experiments, CD8+ T cells were negatively selected using a CD8 negative-selection kit (Invitrogen) and then used for in vitro stimulation. Purity of CD8+ T cell selection was >95%. In some conditions, blocking Ab to CD86 (anti-CD86, 10 μg/ml; BD Biosciences) was added at the time of TCR stimulation. After stimulation, PBMCs or isolated CD8+ T cells were washed in FACS buffer (PBS containing 0.5% BSA and 0.1% NaN3), surface stained for 20 min at room temperature, washed in FACS buffer, and fixed with 1% paraformaldehyde if not analyzed immediately on a flow cytometer. In some experiments, after stimulation, isolated CD8+ T cells were washed twice in FACS buffer and lysed with RLT Lysing Buffer (Qiagen, Valencia, CA) for use in RT-PCR experiments. For the FACS analyses, dead cells were excluded with 7AAD (BD Biosciences) (unfixed cells), per the manufacturer’s instructions, or by using Alexa Fluor 430 staining in PBS at the time of surface staining (fixed cells). For intracellular Ab staining, the cells were washed twice with FACS buffer after surface staining, permeabilized for 10 min at room temperature with 500 μl Perm II Buffer (BD Biosciences), washed with FACS buffer, stained with the indicated intracellular Abs (BD Biosciences) for 20 min at room temperature, washed, and fixed with 1% paraformaldehyde. Flow cytometry was performed using FACS Calibur or LSRII cytometers. FACS data were analyzed with FlowJo software.
RNA preparation and cDNA synthesis
To determine mRNA expression profiling of selected genes, the relative quantitative real-time PCR was performed. Cells were harvested and lysed immediately with 350 μl RLT lysis buffer from an RNeasy kit (Qiagen) and then frozen at −80°C. Total RNA was extracted from collected samples using the RNeasy kit, according to the protocol. Total RNA was eluted in 50 μl water, and OD measurements were taken immediately. All total RNA was reverse-transcribed using a High-Capacity cDNA Archive Kit protocol (Applied Biosystems, Foster City, CA).
Quantitative real-time PCR analysis
The reaction was carried out on a 384-well optical plate (Applied Biosystems) in a 10-μl reaction volume with TaqMan Gene Expression Master Mix (Applied Biosystems). All sequences were amplified using the Applied Biosystems 7900HT Sequence Detection System with the following PCR profile: 50°C for 2 min and 95°C for 10 min, followed by 45 cycles at 95°C for 15 s and 60°C for 1 min. Samples were tested in duplicate, in parallel with the housekeeping gene 18S. Real-Time StatMiner software (Integromics, Granada, Spain) was used to perform a quality control for all runs, and relative quantification δ-δ Ct analysis was used to calculate the fold differences between samples. For CD86 expression, primer and probe were designed and obtained from Applied Biosystems (part number 4329514T).
Results were graphed and analyzed using GraphPad Prism (v4) (GraphPad, San Diego, CA). Comparisons between the expression of CD86 by FACS after anti-CD3 stimulation alone versus anti-CD3 and IL-2 stimulation were performed using the Mann–Whitney U test. Comparison of fold changes in mRNA of CD86 was performed using paired t tests.
Early and transient expression of CD86 on HCV-specific CD8+ T cells during acute HCV infection
The characteristics of four patients with acute HCV infection (a802, a240, a808, and a4915) are shown in Table I. Acute HCV infection was identified by HCV Ab seroconversion in the presence of a clinical syndrome of acute hepatitis. At the time of blood sampling, all of the patients had significantly elevated liver transaminases. None of the patients spontaneously cleared HCV. Patient a802 was also HIV positive and had undetectable HIV viremia with HAART for ∼9 mo prior to acquiring HCV. The other three patients were HIV negative. These patients had detectable HCV-specific CD8+ T cell responses by tetramer analysis. Blood sampling from each of these patients was obtained at the earliest time points after identification of acute HCV infection. Day 0 corresponds to the first time point sampled for this study, rather than day 0 of infection, because the exact time point of HCV acquisition was not known.
B7 ligand expression on HCV-specific CD8+ T cells from the peripheral blood of each patient was measured at the earliest time points available for evaluation (day 0 of sampling for patients a802 and a240, day 25 for patient a808, and day 33 for patient a4915; Fig. 1A ). Approximately 30–60% of HCV-specific CD8+ T cells expressed CD86 at these early time points in HCV infection for all four patients (Fig. 1A ). For a240, two HCV-specific class I tetramer responses could be identified for two epitopes (HCV/1073 and HCV/1406), and HCV-specific CD8+ T cells directed at both epitopes highly expressed CD86 (Fig. 1A ). CD80 expression was not detected on HCV-specific CD8+ T cells from a802, a240, or a4915 at the earliest time points (Fig. 1A ), nor at any time point postinfection (data not shown). In contrast, HCV-specific CD8+ T cells from a808 expressed CD80 and CD86 at the earliest time point studied (Fig. 1A ). Approximately 30% of HCV-specific CD8+ T cells expressed CD86 and 40% expressed CD80 at the earliest time point sampled for a808 (Fig. 1A ). Neither CMV- nor Flu-specific CD8+ T cells expressed CD80 or CD86 during acute HCV infection (Fig. 1A ). In addition, HCV-specific CD8+ T cells did not express the other B7 ligands, PD-L1, or PD-L2 at the earliest time points sampled or at any time point of infection (data not shown).
Frequent blood sampling of a802 and a240 enabled a precise longitudinal assessment of CD86 expression on HCV-specific CD8+ T cells. For both patients, the high-level expression of CD86 on HCV-specific CD8+ T cells was transient (Fig. 1B ). After the first month of evaluation, <10% of HCV-specific CD8+ T cells continued to express CD86 (Fig. 1B ). For a802 and a240, an early and transient increase in CD86 expression on bulk CD8+ T cells was also noted (Fig. 1C ). For a240, who was HIV negative, a peak ∼5% of bulk CD8+ T cells expressed CD86 at the earliest time points of HCV infection, which decreased to <1% of CD8+ T cells within the first month of follow-up (Fig. 1C ). For a802, who was HIV positive, ∼14% of CD8+ T cells expressed CD86 at the earliest time point, which decreased to ∼11% after the first month (Fig. 1C ). Because elevation of CD86 on CD3+ T cells was reported in patients with HIV infection (24, 25), we hypothesize that continued elevation of CD86 on bulk CD8+ T cells from a802 was related to HIV infection. In contrast to these findings in acute HCV infection, <1% of bulk CD8+ T cells typically expressed CD86 in healthy donors (Fig. 1D ).
Lack of frequent blood sampling for a808 precluded a complete prospective evaluation of CD80 and CD86 expression for this patient; however, peak CD80 expression was noted on day 25 (40% of HCV-specific CD8+ T cells), which decreased to <5% by day 49 (Fig. 1E ). CD86 expression was noted on ∼30% of HCV-specific CD8+ T cells on day 25 and 25% of HCV-specific CD8+ T cells at day 49 (Fig. 1E ); at the next available blood sampling (day 167), CD86 expression was <5% (data not shown). For a4915, CD86 continued to be expressed on HCV-specific CD8+ T cells 17 d after the initial time point (day 50; data not shown); however, no CD86 expression was found at the next available time point (day 130; data not shown).
The loss of detectable HCV-specific CD8+ T cells expressing CD86 coincided with the loss of other activation markers, such as CD38 and HLA-DR, during the acute phase of infection (Fig. 1F , 1G ); however, the decrease of CD86 expression to <10% of HCV-specific CD8+ T cells preceded the loss of detectable HCV-specific CD8+ T cells expressing HLA-DR and CD38 to <10%. In this regard, CD86 most resembled the early activation marker CD69, for which a significant detectible population of HCV-specific CD8+ T cells expressing CD69 was also lost very early during infection (Fig. 1F , 1G ).
More specifically, for a802 (Fig. 1F ), HLA-DR was expressed on >75% of HCV-specific CD8+ T cells until antiviral treatment initiation on day 80 of follow-up. The frequency of HCV-specific CD8+ T cells expressing CD38 remained >50% until treatment initiation, although a substantial early decrease in CD38 expression from nearly 100% of HCV-specific CD8+ T cells to ∼60% coincided with the decrease in CD86 expression (Fig. 1F ). The early activation marker CD69 was expressed on only ∼25% of HCV-specific CD8+ T cells at the earliest time point sampled for a802, and it remained <10% after the first month (Fig. 1F ).
For a240 (Fig. 1G ), by day 25 only ∼5% of HCV-specific CD8+ T cells expressed CD86, whereas ∼30% continued to express CD38, and 70% expressed HLA-DR. As with a802, the level of CD69 expression correlated most closely with the level of CD86 expression. For a802 and a240, the costimulatory molecule CD28 remained highly expressed (>75%) on HCV-specific CD8+ T cells throughout infection (Fig. 1F , 1G ).
Addition of IL-2 to culture media induces high-level expression of CD80 and CD86 on CD8+ T cells after TCR stimulation but is not required for CD69, CD38, HLA-DR, or CD25 expression after brief in vitro culture of PBMCs
We next investigated the signals driving CD80 and CD86 expression on CD8+ T cells and compared this with the expression of other activation markers in response to the same stimulation (Fig. 2A ). Although TCR stimulation via brief in vitro culture of PBMCs with anti-CD3 mAb led to a significant (>25%) expression of CD69, CD38, HLA-DR, and CD25 within 1–2 d, it led to only a low-level expression of CD80 and CD86 (<25% of bulk CD8+ T cells) (Fig. 2A ). In particular, anti-CD3 stimulation alone led to very high levels of CD69 expression (>50% of CD8+ T cells) within 1–2 d (Fig. 2A ). Significant expression of CD86 on bulk CD8+ T cells after brief in vitro culture (7 d) with anti-CD3 could be achieved by the addition of IL-2 to the culture media (Fig. 2B ). Furthermore, increasing amounts of IL-2 led to increasing CD86 expression (Fig. 2B ). We further assessed the importance of IL-2 for CD86 expression after brief in vitro culture on sorted CD8+ T cells from 11 patients with chronic HCV infection (c128, c144, c148, c152, c157, c161, c163, c167, c176, c177, and c181) using a negative bead-selection protocol (Invitrogen). After TCR stimulation without IL-2 added to the culture media, minimal CD86 expression was detected on the sorted CD8+ T cells (Fig. 2C ). Addition of IL-2 led to consistent and significant CD86 expression on the CD8+ T cells after TCR stimulation with anti-CD3 (Fig. 2C ). Although the expression of CD69 could be enhanced by coculture with IL-2 among the sorted CD8+ T cells, TCR stimulation alone with anti-CD3 led to the significant expression of CD69, in distinct contrast with B7 ligand upregulation (data not shown). Hence, although CD86 expression during acute HCV infection (Fig. 1) most resembled CD69 expression in terms of the timing of expression (both seen transiently at the earliest phase of acute infection only), the signal driving the expression of CD86 was unique in its dependence on additional cytokine signaling and in its delayed expression in vitro.
We also investigated the signals driving CD80 and CD86 expression on virus-specific CD8+ T cells from the peripheral blood (Fig. 2D ). Fresh PBMCs from a808 (day 0) were used in 5-d culture to assess the ability of TCR stimulation plus IL-2 to increase CD86 expression on HCV-specific CD8+ T cells (Fig. 2D ). Few detectible HCV-specific CD8+ T cells could be obtained after 5-d culture for the no-stimulation condition (Fig. 2D ). This was in concordance with our previous work, indicating that a large fraction of these HCV-specific CD8+ T cells at this early time point were highly susceptible to cytokine-withdrawal apoptosis (7). Five-day stimulation with HCV peptide plus IL-2 led to high expression of CD86 (Fig. 2D ) and CD80 (data not shown) to nearly 100% of HCV-specific CD8+ T cells.
Expression of CD86 on virus-specific CD8+ T cells after in vitro culture with IL-2 was not unique to HCV infection. After 5-d culture of fresh PBMCs from a healthy donor (HD209), culture with CMV peptide alone led to only low-level CD86 expression on CMV-specific CD8+ T cells (Fig. 2E ). Culture with CMV peptide plus IL-2 (50 U/ml) led to high expression of CD86 (Fig. 2E ). Culture with CMV peptide plus the common γ-chain cytokines, IL7 (5 ng/ml) and IL-15 (5 ng/ml), also led to high-level expression of CD86 on CMV-specific CD8+ T cells (Fig. 2E ).
De novo production of CD86 on CD8+ T cells
Two mechanisms exist by which CD86 (and other B7 ligands) could be expressed on CD8+ T cells in general. The first is via de novo generation in highly stimulated cells (16, 30); the second is by a relatively recently described mechanism termed trogocytosis, whereby T cells capture surface molecules through the immunological synapse from APCs (31–33). Both mechanisms were reported to occur in T cells: trogocytosis occurring after stimulation for ≤24 h and endogenous production occurring after 3 d of stimulation (30). In support of de novo production of CD86 on CD8+ T cells, we demonstrated that sorted CD8+ T cells, without the presence of CD86-expressing APCs from which to steal CD86, expressed CD86 after TCR stimulation with IL-2 supplementation in 7-d culture (Fig. 2C ). In addition, sorted CD8+ T cells from these patients with chronic HCV infection (c128, c134, c144, c148, c152, c157, c161, c163, c167, c176, c177, and c181) increased the mRNA expression of CD86 after in vitro culture with anti-CD3 and IL-2 compared with no stimulation (Fig. 3A ) or stimulation with anti-CD3 alone (Fig. 3B ). For the sorting experiments for c134, the quantity of sorted CD8+ T cells only allowed mRNA analysis and not flow cytometry analysis. In further support of the ability of HCV-specific CD8+ T cells to synthesize B7 ligands de novo after stimulation, we generated an HCV-specific CD8+ T cell line that has been propagated via anti-CD3+ IL-2 stimulation without APCs. These cells highly expressed CD86 (Fig. 3C ).
Early loss of pSTAT5 of HCV-specific CD8+ T cells in a patient with acute HCV infection
pSTAT5 in T cells is critical for signal transduction after common γ-chain cytokine binding (34). No study has evaluated pSTAT5 directly ex vivo or in vitro on Ag-specific CD8+ T cells using tetramer analysis. Because we noted a dependence on common γ-chain cytokine signaling for maximal CD80 and CD86 expression of HCV-specific CD8+ T cells in vitro, we assessed the level of pSTAT5 in T cells directly ex vivo from fresh blood of a808 during acute infection (Fig. 4). Only frozen PBMCs were available for the other three acutely infected patients, precluding an ex vivo, basal pSTAT analysis for these patients. We found an increase in the basal level of pSTAT5 in HCV-specific CD8+ T cells compared with pSTAT1 or pSTAT6 for a808 (Fig. 4A ). This increase in pSTAT5 was also detected in CD3+ T cells compared with CD3+ T cells from a patient with chronic HCV infection (Fig. 4B ). Prospective evaluation of a808 showed that the increased level of detectable pSTAT5 in HCV-specific CD8+ T cells, as well as in CD3+ T cells, was lost early in the acute phase of infection, because by day 49 of blood sampling we were unable to detect increased pSTAT5 (Fig. 4C ). These findings in HCV infection contrast with the recent report in HIV infection, in which increased basal pSTAT5 was noted in bulk CD4+ and CD8+ T cells, even in chronic HIV infection (35), which likely indicated the persistent high level of immune activation that often characterizes HIV infection.
We also assessed the ability of HCV-specific CD8+ T cells during early HCV infection to further signal through common γ cytokines and found that brief culture with IL-15, in particular, increased pSTAT5 (Fig. 4D , red box). Although the concentrations of IL-2 (100 U/ml), IL-7 (5 ng/ml), and IL-15 (5 ng/ml) were adequate to cause phosphorylation of STAT5 on bulk CD3+ T cells after very brief culture (10 min) (note the shift in the nontetramer CD3+ T cells in Fig. 4D ), there was a lack of increase in pSTAT5 via IL-2 and -7 stimulation of HCV-specific CD8+ T cells. This deficit in pSTAT5 in response to IL-7 and -2 stimulation in HCV-specific CD8+ T cells may be related to the low level of IL-7R (CD127) and IL-2R (CD25) characterizing the HCV-specific CD8+ T cells from a808 at that time point (day 25) (Fig. 4E ). CD127 expression, in particular, was also reported to be low on HCV-specific CD8+ T cells during acute infection (7, 28, 36–38). Our finding of relatively low pSTAT5 in HCV-specific CD8+ T cells after IL-2 stimulation in vitro in acute HCV infection corresponds with deficits in signaling of CD4+ and CD8+ T cells seen in chronic HIV infection, in which increased basal levels of pSTAT5 correlated with poor pSTAT5 signaling in response to further IL-2 stimulation (35).
Low-level CD80, CD86, and basal pSTAT5 expression in HCV-specific CD8+ T cells in the liver of patients with chronic HCV infection despite high-level expression of activation markers CD69, CD38, and HLA-DR
To fully understand the deficits in the adaptive T cell response to HCV infection, it is important to study immune cells at the site of infection. Studying bulk CD8+ T cells from the liver of five patients with chronic HCV infection (c159, c280, c147E, c671, and c684), we found that the majority of CD8+ T cells expressed the activation markers CD69, CD38, and HLA-DR (Fig. 5A ). Despite high-level expression of these activation markers in the liver of patients with chronic HCV infection, only a minority of CD8+ T cells expressed the B7 ligands CD80 and CD86. We identified two patients with chronic HCV infection and a detectable HCV-specific CD8+ T cell response by tetramer analysis in the liver (c671 and c113e). In previous work, we (7) and other investigators (8, 9) found high-level expression of the activation markers CD69, CD38, and HLA-DR on HCV-specific CD8+ T cells from the liver of patients with chronic HCV. Nearly 100% of HCV-specific CD8+ cells in the liver of patients with chronic HCV infection expressed these activation markers (7). Despite expressing high levels of these activation markers (7), the frequency of liver-infiltrating HCV-specific CD8+ T cells from c671 and c113e that expressed CD86 (∼14%) or CD80 (<1%) was low, despite persistent HCV viremia (Fig. 5B ). Similarly, in blood, no detectible CD80 or CD86 expression was found on HCV-specific CD8+ T cells from c671 (Fig. 5B ). C113e did not have a detectible HCV-specific CD8+ T cell response in the peripheral blood.
When evaluating the basal level of pSTAT5 in bulk CD3+ T cells and in HCV-specific CD8+ T cells for c113E, no basal elevation in recent common γ-chain cytokine signaling could be detected (Fig. 5C , upper plots), nor could we detect a difference in pSTAT5 in the liver versus blood (Fig. 5C , lower plot). We analyzed ex vivo, basal pSTAT5 in the blood and liver of five additional patients with chronic HCV infection (c128E, c135E, c140E, c142E, and c270) and did not detect an increase in the basal level of pSTAT5 in CD8+ T cells in blood or liver of these patients (data not shown).
We further evaluated for the presence of any underlying impairment in IL-2, -7, or -15 signaling of liver-infiltrating CD3+ T cells that could explain the low expression of CD80 and CD86, despite persistent activation (Fig. 5D ). We found no deficit in pSTAT5 in response to IL-2 or -15 (Fig. 5D ). In comparison, we found a relatively lower level of pSTAT5 for the IL-7 condition for the liver CD3+ T cells (Fig. 5D ). We also investigated the ability of HCV-specific CD8+ T cells in the liver to signal after exposure to IL-2, -7, and -15 (Fig. 5E ). Similar to c166 (Fig. 5D ), for patient c113E, liver-derived CD3+CD8+ T cells increased pSTAT5 after exposure to IL-2 and -15 efficiently (Fig. 5E , upper row). However, again, there was lower pSTAT5 for the IL-7 condition (Fig. 5E , upper row). Mirroring these liver CD3+CD8+ T cells, liver HCV-specific CD8+ T cells also efficiently increased pSTAT5 after brief exposure to IL-2 and -15; however, we noted lower pSTAT5 after IL-7 exposure (Fig. 5E , lower row). We hypothesize that this deficit in IL-7 signaling in the liver that is seen in these patients may be related to the decreased frequency of CD127 expression seen on bulk T cells and HCV-specific CD8+ T cells in the liver of patients with chronic HCV infection in general (11). This contrasts with the deficit in pSTAT5 expression in acute HCV, in which HCV-specific CD8+ T cells showed deficiency in pSTAT5 in response to IL-2 and -7 exposure (Fig. 4D ). Future studies with larger cohorts of acute and chronically infected patients will be important to characterize the noted deficits in pSTAT5 signaling. The defect in pSTAT5 phosphorylation of HCV-specific CD8+ T cells in the liver in response to IL-7 that we detected could contribute to the lack of CD80 and CD86 expression seen in the majority of CD8+ T cells in the liver of patients with chronic HCV infection. However, overall, we conclude that there is not an overriding, underlying deficit in the ability of HCV-specific CD8+ T cells from the liver to respond to cytokine (particularly IL-2), which might explain the lack of B7 ligand expression on HCV-specific CD8+ T cells in the liver. Rather, based on these findings, we hypothesize that a dearth of common γ-chain cytokines at the site of chronic HCV infection explains the low frequency of CD80 or CD86 expression on HCV-specific CD8+ T cells in the liver.
In further support of a lack of an underlying deficit in the ability of liver-infiltrating T cells to signal after exposure to cytokines, 5-d in vitro culture of liver CD8+ T cells with anti-CD3 and IL-2 led to high expression of CD86 (Fig. 6A ). High-level CD80 expression was also seen on CD8+ T cells after culture with anti-CD3 and IL-2 (Fig. 6B ). Comparing blood and liver for c101E, 5-d culture with IL-2 alone led to a higher expression of CD80 and CD86 for liver CD8+ T cells (∼30%) compared with blood (9.1% for CD80 and 2.8% for CD86) (Fig. 6B ). In contrast, anti-CD3 culture alone led to greater expression of CD86 on CD8+ T cells in blood compared with liver. We hypothesize that these findings indicate that liver-infiltrating T cells of patients with chronic HCV infection recently received stimulation via their TCR without supporting cytokine. As such, they respond poorly to further TCR stimulation alone (via anti-CD3) but do respond rapidly to IL-2 alone (or the combination of anti-CD3 + IL-2) by increased expression of CD80 and CD86.
Expression of CD86 is linked to STAT5 signaling
Our studies identified the importance of IL-2 signaling for the expression of CD86 on T cells. In fact, without IL-2 in the culture media, minimal CD86 expression was observed on sorted CD8+ T cells (Fig. 2C ). However, we noted low-level CD86 expression on CD8+ T cells after in vitro stimulation of whole PBMCs from some patients with anti-CD3 alone (Figs. 2A , 6B ) in contrast to a lack of CD86 expression after anti-CD3 stimulation alone of sorted CD8+ T cells. Thus, we investigated the level of pSTAT5 expression in relation to CD86 expression in cell culture after TCR stimulation with anti-CD3 alone (Fig. 7A ) and the level of pSTAT5 expression after anti-CD3 stimulation alone in PBMCs versus sorted CD8+ T cells (Fig. 7B , 7C ). The anti-CD3 and IL-2 conditions are shown as positive controls. Importantly, we found that anti-CD3 stimulation alone of PBMCs also led to transient levels of STAT5 phosphorylation in CD8+ T cells (25% noted at day 1) and to transient, low-level expression of CD86 (Fig. 7A, upper row). Addition of IL-2 to anti-CD3 stimulation led to a greater frequency of CD8+ cells expressing pSTAT5 and to prolonged pSTAT5 expression, as expected (Fig. 7, lower row). Evaluating CD8+ T cells in culture over time showed that all CD8+ T cells eventually expressing CD86 also expressed pSTAT5, whether stimulated with anti-CD3 alone or with anti-CD3 and IL-2 (Fig. 7A ). Because we did not observe the expression of CD86 on sorted CD8+ T cells in culture after anti-CD3 stimulation alone (Fig. 2C ), we repeated the pSTAT5 experiment on sorted CD8+ T cells (Fig. 7B , 7C ). In concordance with the importance of IL-2 signaling for CD86 expression, there was no pSTAT5 expression on sorted CD8+ T cells after anti-CD3 stimulation alone, in contrast with the findings in PBMCs (Fig. 7B ). With the anti-CD3 stimulation alone of PBMCs, ∼20–30% of CD8+ T cells expressed pSTAT5 at day 1 (Fig. 7A , 7C ). Consistently, we found a lack of pSTAT5 expression after anti-CD3 stimulation alone of sorted CD8+ T cells, in contrast to anti-CD3 stimulation of PBMCs (data not shown). We hypothesize that IL-2 released by CD4+ T cells in cultured PBMCs after anti-CD3 stimulation alone contributed to the expression of CD86 on CD8+ T cells in this condition and that the lack of CD4+ T cells in sorted CD8+ T cells led to a lack of pSTAT5 expression (Fig. 7) and a lack of CD86 expression (Fig. 2A ) after anti-CD3 stimulation alone. Based on these studies, we conclude that pSTAT5 expression in CD8+ T cells is critical for CD86 expression after TCR stimulation of bulk PBMCs and sorted CD8+ T cells.
B7/CD28 family molecules play a central role in the generation and modulation of the adaptive T cell immune response. A balance of costimulatory and coinhibitory signaling governs the activation and function of the responding T cells (see review in Ref. 21). Classically, it is the B7 ligand expressed on the APC signaling to the CD28 family receptor on the T cell that directs the T cell response. However, a number of studies also identified B7 ligand expression on T cells (16–23) and linked its expression to an enhancement of the T cell response (16). Expression of CD80 on a human CD4+ T cell clone enhanced an MLR to resting peripheral blood responder T cells (16), and expression of CD86 on anti-CD3 stimulated and paraformaldehyde-fixed human T cells enhanced IFN-γ production and proliferation of naive CD4+ T cells responding to suboptimal concentrations of anti-CD3 (18). Furthermore, fixed CD86+, but not CD86−, T cells induced an MLR response that was partly decreased by neutralizing anti-CD86 mAbs (18). Despite these studies, the importance of B7 ligand expression on virus-specific CD8+ T cells is not known.
In the acute phase of infection, the inhibitory receptor PD-1 is often highly expressed on HCV-specific CD8+ T cells (7, 15), so we hypothesized that other costimulatory signals are important at this early phase of infection to enable an effective immune response. In the current study, we demonstrated that the B7 ligand CD86 is highly expressed on HCV-specific CD8+ T cells during the early acute phase of infection and not during the later phase of acute infection or during chronic infection, even at the site of infection in the liver. Significant CD86 expression on HCV-specific CD8+ T cells was not detected at any later time points for any of the acutely infected patients developing chronic infection nor for any other chronically infected patients that we evaluated. For some patients in the acute phase of HCV infection, HCV-specific CD8+ T cells also expressed CD80, although expression was not seen in other patients with acute infection. It is not well understood why CD80 is expressed on HCV-specific CD8+ T cells from some patients with acute infection but not others. We hypothesize that this may be related to differing kinetics of CD86 and CD80 expression or to different levels of signaling required for the expression of CD86 and CD80.
In this study, we investigated the significance of B7 ligand expression on T cells and found that high-level expression was delayed after brief in vitro culture (5–7 d) and was linked with recent common γ-chain cytokine signaling. This was in contrast with other activation markers, such as CD69, CD38, HLA-DR, or CD25, whose expression could rapidly (1–3 d) be induced by TCR stimulation alone after brief in vitro culture of PBMCs. Hence, our findings support the hypothesis that B7 ligand expression on T cells is a unique marker that identifies recent stimulation via TCRs in the presence of sufficient supportive cytokine. The lack of B7 ligand expression on liver-infiltrating HCV-specific CD8+ T cells in chronic infection, despite the high-level expression of activation markers, highlights a critical deficit in supportive cytokine signaling that contributes to the waning immune response to HCV. Recent studies in mice demonstrated that IL-7 can be produced by hepatocytes themselves and that IL-7 is important in regulating the expansion of T cells in response to LPS (39). Although hepatocyte IL-7 was not found to be important in pathogen-specific CD8+ T cell proliferation in this study (39), improved understanding of the cytokine milieu in the liver of patients with hepatotropic viral infection is clearly important.
IL-2, in particular, was shown to be important in the generation of effective immune responses to HCV infection, and secretion of IL-2 by CD4+ T cells during the acute phase of infection is critical for sustained and effective adaptive CD8+ T cell responses (38, 40, 41). CD4+ T cells from patients with self-limited evolution of infection produced considerably more IL-2 in response to HCV recombinant proteins compared with patients with chronically evolving disease (38, 40, 41). In the chimpanzee model of HCV infection, depletion of CD4+ T cells prior to infection led to an inability to clear viremia (42). A preferential loss of IL-2–secreting CD4+ T cells has been noted on progression to chronic HCV infection (43), and HCV-specific CD8+ T cells from the peripheral blood of patients with chronic infection have an impaired ability to proliferate that can be rescued in vitro by exposure to IL-2 (44). Our study further supports the critical loss of IL-2 during progression to chronic infection and identifies ex vivo pSTAT5 signaling and expression of B7 ligands (CD86 and CD80) as important markers of recent effective signaling. Although our study highlights the role of common γ-chain cytokines, such as IL-2, in the expression of CD86 on T cells, future studies will need to investigate the role of other inflammatory cytokines in the expression of CD86 or CD80 on T cells during HCV infection. Furthermore, determining whether the level of CD86 expression or the timing of CD86 expression is a determinant of viral clearance versus persistence will require longitudinal studies with larger numbers of patients with acute disease.
Studies of liver-infiltrating HCV-specific CD8+ T cells are critical to understand the failure of the immune response seen in most patients with HCV infection. Previous studies demonstrated the high activation state of these cells in the liver but poor functionality in the chronic phase of infection (7–9). A number of factors likely contributes to the waning immune response and include high-level expression of PD-1 (11–13), infiltration by regulatory T cells (Tregs) [(45) and reviewed in Ref. 46], and a loss of CD4+ T cell help (43). We hypothesize that a central feature of each of these mechanisms is the loss of IL-2 signaling on HCV-specific CD8+ T cells. Recent studies on the mechanism of action of PD-1 signaling indicate the possibility that PD-1 signaling might directly prevent STAT5 phosphorylation via activation of the Src homology 2-containing tyrosine phosphatase (45, 47). In addition, one of the proposed mechanisms of action of Tregs is to act as an IL-2 sink and deplete the immunological milieu of the supportive cytokine (48). Our study is the first to characterize pSTAT5 on virus-specific CD8+ T cells using tetramers. Our findings indicate an early loss of pSTAT5 that occurs in the acute phase of infection and a lack of high-level pSTAT5 in liver-infiltrating HCV-specific CD8+ T cells, despite persistent infection and persistent activation. Future studies will need to determine the relative contribution of PD-1 signaling, Treg infiltration, and CD4+ Th cell loss in the reduction of pSTAT5 in HCV-specific CD8+ T cells.
Clearly, differences in the acute versus chronic immune response are evident in HCV infection, and this can be seen in the differing clinical responses and degree of liver injury as measured by alanine aminotransferase (ALT) in the patients infected with HCV. Based on our study, we hypothesize that other B7 molecules, and in particular, CD86, which are expressed at this early phase of infection, provide costimulatory signals via T–T interactions that enhance the immune response at this early stage of infection. In this study, we assessed the ability of HCV-specific CD8+ T cells expressing CD86 to function as APCs in a T–T-dependent manner by sorting on CD3+CD8+ T cells from fresh PBMCs from a808 (day 0) and culturing in the presence of HCV peptide, HCV peptide plus IL-2, and HCV peptide plus anti-CD86 with/without IL-2 (data not shown). We were unable to demonstrate an ability of HCV-specific CD8+ T cells to present Ag in this manner, indicating that these T cells functioned poorly in presenting Ag, despite the expression of B7 ligands (data not shown). Although we did not see evidence of direct Ag presentation by HCV-specific CD8+ T cells to other T cells, we hypothesize that the expression of CD86 on HCV-specific CD8+ T cells is important for other T–T interactions, by providing costimulation to neighboring T cells interacting with an APC or via an cell-autonomous costimulatory signal. Unfortunately, given the difficulty in separating the effect of CD86 expression on APCs from CD86 expression on T cells during in vitro assays, we were unable to demonstrate this effect directly.
In this study, we found that CD86 expression was lost early during HCV infection (Fig. 1), despite persistent high-level PD-1 expression on HCV-specific CD8+ T cells in acute infection (7, 15). This loss of CD86 expression also coincided with decreased liver inflammation, as measured by ALT levels. Thus, we hypothesize that as HCV progresses to chronic infection, the persistent negative signals via receptors, such as PD-1, and the loss of positive signals via CD86 on T cells tip the balance in favor of a waning response. Net negative costimulatory/coinhibitory signals to T cells at this phase of infection may be adaptive for a host that is unable to clear a virus, and waning CD86 expression may be a mechanism to prevent further high-level liver damage. If CD86 expression on HCV-specific CD8+ T cells is shown to provide direct costimulation to other HCV-specific CD8+ T cells in acute infection, prolonging or modulating CD86 expression on these T cells may also be a mechanism that can be used to enhance future therapies for patients with chronic HCV infection.
We thank Francie Lasseter, Beverly Weaver, Melissa K. Osborn, Ana Howells-Ferrerira, and the Emory Transplant clinical group for patient cohort coordination. We also thank Benton Lawson and Natalia Kozyr for assistance with RT-PCR experiments, Enrique Martinez and Naasha Talati for patient referrals, and the patients who agreed to participate.
Disclosures The authors have no financial conflicts of interest.
This work was supported by the Bill and Melinda Gates Foundation Grand Challenges in Global Health (AG, GC#12 to Rafi Ahmed); Center For AIDS Research Immunology Core P30 AI050409 (to H.R., C.I., and A.G.); Cancer Research Institute Investigator Award (to A.G.); the Yerkes Research Center Base Grant RR-00165; the Canadian Institutes for Health Research (MOP-74524 [to N.H.S. and J.B.]); Fonds de la Recherche en Santé du Quebec AIDS and Infectious Disease Network (to N.H.S. and J.B.); Fonds de la Recherche en Santé du Quebec Senior clinical research award (to J.B.); Canadian Institutes for Health Research New Investigator Award (to N.H.S.); and the Public Health Service (K08 AI072191 [to H.R.] and AI070101 [to A.G.]).
Abbreviations used in this paper:
- 7-aminoactinomycin D
- alanine aminotransferase
- Centre de Recherche du Centre Hospitalier de l’Universite de Montreal
- highly active antiretroviral therapy
- hepatitis C virus
- isotope control Ab
- liver infiltrating mononuclear cell
- no stimulation
- programmed death receptor 1
- phosphorylated STAT
- regulatory T cell.
- Received September 21, 2009.
- Accepted December 21, 2009.
- Copyright © 2010 by The American Association of Immunologists, Inc.