Activated PD-1⁺ CD4⁺ T cells represent a short-lived part of the viral reservoir and predict poor immunologic recovery upon initiation of ART

Abstract

Objective: 

Activated (CD38⁺HLA-DR⁺) PD-1⁺ CD4⁺ T cells are strongly associated with virus replication and disease progression in untreated HIV-1 infection, and viral persistence in individuals on ART. Few studies have examined cell-associated viral load (CAVL) in different activated CD4⁺ T-cell populations to measure relative contributions to viral reservoirs.

Design: 

Longitudinal assessment of HIV-1 chronically infected Ugandans initiating ART, to investigate activated CD4⁺ T-cell populations and their contribution to viral reservoirs.

Methods: 

We followed 32 HIV-1 chronically infected individuals from Kampala, Uganda, and determined their CD4⁺ T-cell counts and viral load at baseline, 6, and 12 months after the initiation of ART. T-cell populations were sorted based on activation profiles and gag DNA was measured to determine CAVL within these populations. Soluble factors associated with inflammation were measured in plasma using a multiplexed platform.

Results: 

Concomitant with viral load decline and CD4⁺ T-cell count rebound, the activated PD-1⁺ CD4⁺ T-cell population contracted upon initiation of ART. Baseline levels of activated PD-1⁺ CD4⁺ T cells correlated with plasma levels of IP-10 and TNFRII. Interestingly, a higher baseline level of activated PD-1⁺ CD4⁺ T cells was associated with poorer CD4⁺ T-cell recovery after 12 months of ART. This population contributed significantly to the cell-associated HIV DNA load at baseline, whereas their contribution declined on ART, indicating high turnover.

Conclusion: 

Activated PD-1⁺ CD4⁺ T cells are predictors of poor immunologic recovery on ART and may represent a short-lived component of HIV-1 reservoirs.

Introduction

HIV-1-associated immune activation perpetuates viral replication, drives disease progression, and persists despite successful antiretroviral therapy (ART) ^[1,2]. Levels of activated CD4⁺ T cells co-expressing CD38, HLA-DR, and PD-1 in circulation are predictive of disease progression in untreated HIV-1 infection ^[3]. Similarly, activated and exhausted CD4⁺ T cells are associated with more severe disease in HIV-2 infection ^[4]. Approaches leveraging bioinformatics independently substantiated the significance of this population in predicting disease outcome in untreated infection ^[5,6]. In treated infection, PD-1 expression on activated CD4⁺ T cells has been linked with viral persistence after effective ART ^[7]. However, few studies have examined the cell-associated viral load (CAVL) in phenotypically distinct CD4⁺ T-cell populations based on activation phenotype to define their contribution to the HIV reservoir.

Here, 32 chronically HIV-1 infected individuals from Kampala were followed to explore the immunologic and virologic response to ART initiation. Soluble and cellular biomarkers of immune activation were measured, and HIV-1 gag DNA quantified to determine CAVL in T-cell populations based on activation profiles. Of note, poorer immunologic recovery was associated with higher baseline levels of activated PD-1⁺ CD4⁺ T cells after 12 months of ART. This population showed detectable levels of HIV-1 gag DNA before treatment; however, compared with baseline, CAVL within the activated PD-1⁺ CD4⁺ T-cell population was significantly lower after 12 months of ART. Therefore, the activated PD-1⁺ CD4⁺ T-cell population represents a short-lived component of the viral reservoir and elevated levels of these cells at baseline predict poor immunologic recovery on ART.

Methods

Study participants

Study participants aged 29–38 years were from the Couples Observation Study (COS) in Kampala, Uganda, as previously described ^[8] (Supplementary Table 1). Samples were collected; CD4⁺ T-cell counts determined and viral load assessments made at baseline, and 6 and 12 months after initiation of ART. The study was approved by the institutional review boards of Uganda's National Council for Science and Technology (UNCST) and the National AIDS Research Committee, the University of Washington, as well as the Division of Human Subjects Protection at the Walter Reed Army Institute of Research. Participants provided written informed consent.

Flow cytometry and soluble inflammatory marker measurement

Peripheral blood mononuclear cells (PBMC), isolated and cryopreserved as previously described ^[9], were thawed, washed, and stained with the LIVE/DEAD Fixable Aqua Dead Cell Stain Kit (Molecular Probes, Eugene, Oregon, USA), and monoclonal antibodies for CD3(SK7) PerCP-Cy5.5, CD8(SK1) PE-Cy7, CD38(HB7) APC, HLA-DR(L243) FITC, PD-1(EH12.1) PE, CD14(MΦP9) APC-H7, and CD16(3G8) Pacific Blue (all from BD Biosciences, San Jose, California, USA), as well as CD4⁺ (SFCI12T4D11) ECD (Beckman Coulter, Indianapolis, Indiana, USA). Fixed cells were run on a BD FACSARIA II SORP (BD Biosciences) where populations were sorted, and dry pellets stored. A multiplex array was used to quantify plasma levels of: IFNγ, IL-1α, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-15, IL-17, IP-10, MCP-1, TNF, and TNFRII (Quansys Biosciences, Logan, Utah, USA). ELISAs were used to measure neopterin (GenWay Biotech, San Diego, California, USA), as well as IFNα, I-FABP, and sCD14 (R&D Systems, Minneapolis, Minnesota, USA). Plasma was tested in triplicate and mean values used for data analysis.

Cell-associated DNA

HIV-1 quantification in cell populations was performed similar to a previous report ^[10]. Dry cell pellets were thawed, lysed, and kept on ice. The PCR reaction included 10× PCR buffer, dNTP, MgCl₂, primers (forward primer, Gag762V2, 5′-TGA CTA GCG GAG GCT AGA A-3′, reverse primer, GagRev909, 5′- CTC YCT GCT TGCCCA TA-3′), Platinum Taq DNA Polymerase, and cell lysates. The reaction mixture was subjected to preamplification to increase the sensitivity of the assay using an MJ Research PTC-225 thermal cycler (Bio-Rad Laboratories, Hercules, California, USA). Preamplification product included a reaction mixture containing PCR buffer, dNTP, MgCl₂, primers, Platinum Taq, ROX reference dye and probe (5′-[6FAM] AAA ATT CGG TTA AGS CCA GGG GGA AAG AA [BHQ1]-3′). DNA copies were quantified against dilutions of the 8E5 cell line engineered to express 1 HIV copy per cell.

Statistical methods

For this study, 36 COS study participants originally qualified for inclusion, but two did not have pre-ART samples available and two only received prevention of mother-to-child transmission (PMTCT) prophylaxis, but not regular ART treatment, and were excluded. Cellular and soluble factor data was log₁₀ transformed for analysis. Immunologic and virologic changes between baseline, month 6, and month 12 visits were evaluated using a paired t-test. A t-test was used to calculate differences between groups. The relationship between biomarkers and clinical indices were tested using linear regression. P values less than 0.05 were considered statistically significant.

Results

Levels of activated PD-1⁺ CD4⁺ T cells at baseline associate with CD4⁺ T-cell recovery after effective antiretroviral therapy

Activated (CD38⁺HLA-DR⁺) PD-1⁺ CD4⁺ T cells were identified by flow cytometry in 32 Ugandans with chronic untreated HIV-1 infection (Fig. 1a). The PD-1⁺ activated CD4⁺ T cells correlated positively with viral load (R² = 0.238, P = 0.005) and inversely with CD4⁺ count (R² = 0.261, P = 0.003), similar to previous observations ^[3] (Fig. 1b). In addition, levels of these cells correlated directly with IP-10 (R² = 0.362, P < 0.001) and soluble TNFRII (R² = 0.237, P = 0.006) in plasma (Fig. 1b). Samples were also examined at 6 and 12 months after ART initiation, and response to therapy assessed. As viral load was suppressed and CD4⁺ counts improved, the frequency of the activated PD-1⁺ CD4⁺ T-cell population declined (Fig. 1c). Baseline pretreatment viral load was not predictive of CD4⁺ T-cell count recovery by 12 months of therapy and neither was baseline CD4⁺ counts, although a nonsignificant trend towards a correlation was observed for the latter (R² = 0.168, P = 0.07) (Fig. 1d). In contrast, pretreatment levels of bulk activated CD4⁺ T cells and activated PD-1⁺ CD4⁺ T cells showed inverse relationships to CD4⁺ T-cell recovery on ART (R² = 0.250, P = 0.04 and R² = 0.240, P = 0.04, respectively, Supplementary Table 2, https://links.lww.com/QAD/B582). No other cell population or soluble plasma cytokine measured at baseline correlated significantly with CD4⁺ T-cell recovery (Supplementary Table 2, https://links.lww.com/QAD/B582).

Fig. 1: Activated PD-1⁺ CD4⁺ T cells are associated with disease progression and immune recovery on antiretroviral therapy.

Rapid decline of the cell-associated viral load in activated PD-1⁺ CD4⁺ T cells after effective antiretroviral therapy

We hypothesized that the association between the activated PD-1⁺ CD4⁺ T-cell population and HIV immunopathogenesis may indicate that this cell population makes up a significant part of, or reflects the size of, HIV-1 reservoirs. To investigate this, we sorted resting (CD38^-HLA-DR^-PD-1^-), CD38⁺HLA-DR^-PD-1^-, and activated PD-1⁺ (CD38⁺HLA-DR⁺PD-1⁺) CD4⁺ T-cell populations from the 32 donors at baseline and at 6 and 12 months on ART (Fig. 2a) (Supplementary Table 3). The CAVL at baseline was variable across the three sorted CD4⁺ T-cell populations with the highest levels in the CD38⁺HLA-DR^-PD-1^- CD4⁺ T cells (mean 62 025 HIV-1 gag DNA copies/million sorted cells). Activated PD-1⁺ CD4⁺ T cells harbored a mean of 17 394 HIV-1 gag DNA copies/million sorted cells. Upon initiation of ART, the activated PD-1⁺ CD4⁺ T-cell CAVL declined rapidly and was undetectable in 91% of donors by month 12 (Fig. 2b). The other two CD4⁺ T-cell populations also saw declining CAVL but not to the same extent, indicating that the activated PD-1⁺ CD4⁺ T-cell population containing virus has a short half-life.

Fig. 2: Cell-associated viral load measurements in patients initiating antiretroviral therapy.

Finally, we investigated associations between the activated PD-1⁺ CD4⁺ T-cell population and the viral reservoir as measured by the CAVL, in the three sorted cell populations at baseline. Interestingly, the activated PD-1⁺ CD4⁺ T-cell levels were positively correlated with CAVL in all three CD4⁺ T-cell populations (Fig. 2c). HIV-1 plasma viral load, IP-10, and TNFRII levels at baseline also correlated with the CAVL in all sorted populations (Supplementary Table 4, https://links.lww.com/QAD/B582). Interestingly, no associations were observed between any of the cellular or soluble biomarkers and CAVL at month 6 or month 12 after initiation of ART.

Discussion

In this study, we analyzed the level of activated (CD38⁺ and HLA-DR⁺) CD4⁺ T-cells expressing PD-1 in relation to CD4⁺ T-cell recovery and viral reservoir size in a cohort of Ugandan patients initiating ART. The activated PD-1⁺ CD4⁺ T-cell population is associated with more severe disease outcome in untreated infection ^[3–6]. In this study, the activated PD-1⁺ CD4⁺ T-cell population declined after initiation of ART. High pretreatment levels of these cells were predictive of slow immunologic recovery and associated with CAVL. Additionally, activated PD-1⁺ CD4⁺ T cells were found to contain HIV-1 DNA at baseline. However, the relative contribution to the total CAVL diminished on ART, indicating higher turnover in this population.

A number of studies have examined immune activation (CD38 and HLA-DR) and exhaustion (PD-1) in relation to immunologic recovery and the size of the HIV-1 reservoirs after ART ^[7,11–14]. These studies were cross sectional and correlated immune activation or exhaustion to outcome after treatment. Our findings extend these observations by showing that the combined phenotype of activated PD-1⁺ CD4⁺ T cells at baseline, prior to treatment is associated with longitudinal immunologic recovery after ART initiation. It is interesting to note that the bulk activated (CD38⁺HLA-DR⁺) CD4⁺ T-cell population predicted CD4⁺ T-cell recovery, irrespective of PD-1 expression (Supplementary Table 2, https://links.lww.com/QAD/B582), as previously reported ^[15–17]. However, the frequency of activated CD4⁺ T cells lacking PD-1 expression does not significantly associate with CD4⁺ T-cell recovery in this study (R² = 0.077, P = 0.238) (Supplementary Table 2, https://links.lww.com/QAD/B582), thereby suggesting that the activated PD-1⁺ CD4⁺ T-cell population is driving the association with immunologic outcome.

In both treated and untreated infection, the importance of central, effector, and transitional memory CD4⁺ T-cell populations have been clearly demonstrated ^[10,18]. The shift from activated to resting memory state in the CD4⁺ T-cell population may leave them highly susceptible to latency ^[19]. Moreover, in treated HIV-1 infection, distinct populations within the memory pool display different decay rates whereby less differentiated populations may persist at higher rates over time ^[20]. The activated PD-1⁺ CD4⁺ T-cell population primarily consists of an effector memory (T_EM) phenotype ^[3]. We observed measurable CAVL at baseline and a decline after 6 months of therapy, supporting a model where the activated PD-1⁺ CD4⁺ T_EM cell population might be more differentiated with high rates of turnover. Another possible explanation for the reduction in frequency and CAVL of the activated PD-1⁺ CD4⁺ T-cell population is that the phenotype may change after ART initiation. Although PD-1 is epigenetically imprinted and expression is thought to persist ^[21], CD38 and HLA-DR may be subject to downregulation after ART.

The inclusion of the immune checkpoint receptor PD-1 was based on previous observations showing significant associations with disease ^[3]. It should be noted that the CD38⁺HLA-DR⁺ population marks only a fraction of the total PD-1⁺ CD4⁺ T-cell population; median (range) 10% (3–18%) at baseline. Future studies of HIV-1 reservoirs should evaluate the role of PD-1 and the heterogeneity of the CD4⁺ T-cell compartment in HIV-1 infected individuals. Interestingly, PD-1 expression in combination with activation markers at baseline correlated with CAVL in all subsets analyzed. This association may not represent a direct mechanistic link between these cell populations but rather reflect the extent of disease. This reinforces previous findings from us and others showing an HIV-associated increase in frequency of PD-1⁺ activated CD4⁺ T cells that is highly predictive of disease ^[3,4], which our current observations extend to CAVL and viral reservoirs. Assays that quantify HIV-1 reservoirs are of paramount importance, although many assays are not specific for replication-competent latent provirus ^[22]. We analyzed HIV-1 DNA to quantify the total pool of infected cells and to maximize sensitivity; however, despite correlating well with the number of infected cells, quantified HIV-1 DNA includes defective or unintegrated forms of the virus ^[23]. Although activated PD-1⁺ CD4⁺ T cells may not represent a long-lasting component of the reservoir, the associations with disease and immunologic recovery after initiation of ART highlights the importance of these cells in HIV-1 immunopathogenesis and response to therapy.

Acknowledgements

The authors would like to thank the volunteers for participating in this study and allowing samples to be used for this research. PBMC samples were processed and cryopreserved by the Makerere University Walter Reed Project (MUWRP) in Kampala, Uganda. The views expressed are those of the authors and should not be construed to represent the positions of the US Army or the Department of Defense.

Author contributions: The manuscript was written by M.A.E., T.H., S.T. and J.K.S. M.A.E., J.K.S., J.M.B., S.T. S.R.K., and S.J.K. contributed to study design. M.C., M.N., B.S., and M.A.E. performed experimentation and collection of data. M.C., M.N., M.A.E., and T.H. performed data and statistical analysis. M.A.E., S.T., J.K.S., D.L.B., and J.M.B. contributed to interpretation of the data. S.R.K., M.L.R., and N.L.M. supported the funding and M.J.E., E.T.K., and J.M.B. contributed to the design, execution, collection and analysis of the primary cohort data. All authors provided review of the final manuscript.

Funding: This work was supported by cooperative agreements (W81XWH-11-2-0174 and W81XWH-18-2-0040) between the Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., and the U.S. Army Medical Research and Materiel Command. Additional funding was provided by the Division of Intramural Research, NIAID, NIH, and the Swedish Research Council. Funding for the Couples Observational Cohort was supported by the Bill & Melinda Gates Foundation (grant 41185). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Disclaimer: The views expressed are those of the authors and should not be construed to represent the positions of the US Army or the Department of Defense.

Conflicts of interest

There are no conflicts of interest.