Comparative analysis of activation induced marker (AIM) assays for sensitive identification of antigen-specific CD4 T cells

Abstract

The identification and study of antigen-specific CD4 T cells, both in peripheral blood and in tissues, is key for a broad range of immunological research, including vaccine responses and infectious diseases. Detection of these cells is hampered by both their rarity and their heterogeneity, in particular with regards to cytokine secretion profiles. These factors prevent the identification of the total pool of antigen-specific CD4 T cells by classical methods. We have developed assays for the highly sensitive detection of such cells by measuring the upregulation of surface activation induced markers (AIM). Here, we compare two such assays based on concurrent expression of CD69 plus CD40L (CD154) or expression of OX40 plus CD25, and we develop additional AIM assays based on OX40 plus PD-L1 or 4-1BB. We compare the relative sensitivity of these assays for detection of vaccine and natural infection-induced CD4 T cell responses and show that these assays identify distinct, but overlapping populations of antigen-specific CD4 T cells, a subpopulation of which can also be detected on the basis of cytokine synthesis. Bystander activation had minimal effect on AIM markers. However, some T regulatory cells upregulate CD25 upon antigen stimulation. We therefore validated AIM assays designed to exclude most T regulatory cells, for both human and non-human primate (NHP, Macaca mulatta) studies. Overall, through head-to-head comparisons and methodological improvements, we show that AIM assays represent a sensitive and valuable method for the detection of antigen-specific CD4 T cells.

Fig 1. Comparing the OX40/CD25 and CD40L/CD69 AIM assays at their optimal time points.

(A) A timeline of the assays’ setup and stimulation periods. (B) Example plots of OX40⁺CD25⁺ (red) and CD40L⁺CD69⁺ (blue) expression from human CD4 T cells when PBMC samples were unstimulated (UN), stimulated with peptides from HIV-Env, HIV-Gag, or hCMV, or stimulated with SEB as a positive control. Stimulation period was 9 hrs for the CD40L/CD69 assay and 18 hrs for the OX40/CD25 assay. Flow cytometry was gated as per S1 Fig (C,D) Quantitation of signal response to each antigen after 9 hrs or 18 hrs of stimulation for the CD40L/CD69 and OX40/CD25 assays, respectively. n = 8 for Env and Gag groups, n = 10 for NS and hCMV groups. P values shown comparing signal in the Gag and hCMV stimulations to the unstimulated (UN) condition for both assays (Wilcoxon test). (E,F) Quantitation of response measured by fold-increase (antigen-stimulation signal divided by the signal in the unstimulated control) for each antigen after 9 or 18 hrs of stimulation for the CD40L/CD69 and OX40/CD25 assays, respectively. Dotted line is shown at a fold-increase of 2, indicating positive signal threshold. (G,H) Correlation between the signals detected by the two assays after computing the fold increase and background subtracted values. The different antigen stimulations are shown in orange (Env), purple (Gag), or dark blue (hCMV). R^s = Spearman's Rank correlation coefficient with associated p values. For (E-H) n = 8 for Env and Gag groups, n = 10 for hCMV.

Fig 2. AIM markers identify overlapping populations of antigen-specific CD4 T cells.

(A) Experiment protocol overview. Briefly, PBMCs were rested, then stimulated with antigen for 9-18hrs and stained for CD69, CD40L, OX40 and CD25 surface upregulation. (B) Example flow cytometry plots following 18-hr stimulation. (C) Quantification of data in (B) following a 9 hr (left) and 18 hr (right) antigen stimulation. CD69⁺CD40L⁺ is shown in blue and OX40⁺CD25⁺ in red. (D) Correlations between the two AIM assays following a 9 hr (left) and 18 hr (right) antigen stimulation. Data shown are background subtracted. Different antigens are indicated by color: HIV-1 Gag (purple), HIV-1 Env (orange), hCMV (dark blue) and HBV (green). R^s = Spearman's Rank correlation co-efficient with associated p values. (E,F) Example flow cytometry plots demonstrating the co-expression of different AIM at 18 hr post-antigen stimulation for (E) OX40/CD25 expression on CD69⁺CD40L⁺ CD4 T cells and (F) CD69/CD40L expression on OX40⁺CD25⁺ CD4 T cells. Numbers shown in bottom row indicate percentage from parent gate. (G) Quantification of data in (E,F). Venn diagrams illustrate overlap between the two AIM marker⁺ populations. All numbers represent mean percentages of AIM marker⁺ cells from total CD4 T cells. For (B-G) n = 3–5 HIV-infected, hCMV-positive donors per antigen. (H,I) Example flow cytometry plot showing OX40/CD25 expression on cytokine-producing cells (ICS CD40L⁺IFNγ⁺, top panels), and cytokine production from OX40⁺CD25⁺ cells (bottom panels), (M. mulatta PBMC). (I) Quantification of (H). (J) Detection of antigen-specific CD4 T cells by AIM vs. cytokine assays (M. mulatta PBMC). Cells from vaccinated animals were stimulated with BG505 antigen for a total of 21 hr with Brefeldin A added in the final 4 hrs. For H-J, n = 4 animals.

Fig 3. Bystander activation by contaminants or non-TCR stimulation has limited impact on upregulation of AIM markers.

(A,B) OX40/CD25 expression of PBMCs cultured in the presence of increasing amounts of LPS (25–2500 EU/ml), with an unstimulated (UN) and SEB condition as negative and positive control. (A) Example plots from a representative donor cultured with or without LPS. (B) Data shown as fold-increase (LPS/UN); n = 3 donors. (C) Experimental setup for transwell plate experiment, where antigen-exposed PBMCs were placed within the top well of a transwell plate, and antigen-naïve PBMCs were placed within the bottom well, with the two wells separated by a 0.4 μm pore-size permeable membrane. The cultures were then left unstimulated (UN) or stimulated with antigen (Ag) for 24 hours at 37°C. (D) Representative flow cytometry plot of the antigen-experienced PBMCs in the top well of the transwell plate (left column), the antigen-naïve PBMC in the bottom well of the transwell plate (middle column), and the antigen-naïve PBMC control in a separate plate (right column) after 24 hours in culture; with the unstimulated condition (UN) on the top row and antigen stimulated (Ag) on the bottom row. Gated on total CD4 T cells. (E) Quantification of the OX40⁺CD25⁺ signal for the three culture conditions, unstimulated (UN) or stimulated with antigen (Ag). n = 3 antigen-naïve donors. (F) Diagram of the CD4 T cell line bystander activation experiment from panels G-K. Briefly, cell tracer-labeled (CellVue⁺) PBMCs were cocultured with an autologous CD4 T cell line at different ratios and stimulated with peptide antigens for 18 hr. (G,H) Example flow plots show the frequency of antigen-specific CD4 T cells CellVue⁺ PBMCs in the top row compared to the total frequency of antigen-specific cells within the coculture (CD4 T cell line + CellVue⁺ PBMCs) in the bottom row, at indicated ratios of CD4 T cell line and CellVue⁺ PBMCs. Antigen-specific cells are identified as CD69⁺CD40L⁺ (G) or OX40⁺CD25⁺ (H) respectively. (I,J) Quantification of data shown in (G,H) for CD69/CD40L (I) and OX40/CD25 expression (J) respectively. The total frequency of antigen-specific cells in the culture (CD4 T cell line + CellVue⁺ PBMCs) is indicated by grey bars; the frequency of CellVue⁺ autologous CD4 T cells from PBMCs is indicated by the colored dots/lines. (K) Data shown as a relative change in the percentage of antigen-specific CD4 T cells from the CellVue⁺ PBMCs at each ratio, compared to the PBMC alone condition. n = 3 independent experiments and donors, with HIV-Gag or hCMV pp65 peptide stimulation. Bars represent mean +/- SEM, dotted line at 100% (i.e. no change).

Fig 4. Treg cells in human PBMC.

(A) Graph showing the frequency of Foxp3⁺ Treg cells within the AIM⁺ total CD4 population after 18 hrs of no stimulation (UN) or stimulation with tetanus peptide pool. (B) Increasing amounts of OX40⁺CD25⁺ nTreg cells (Foxp3⁺Helios⁺) within the total CD4 T cell population in human PBMC in response to tetanus peptide pool stimulation. (C) Example plot of CD39 and Foxp3 expression of total CD4 T cells in an unstimulated human PBMC sample after 18 hrs of incubation at 37°C. (D) Venn diagram showing overlap between Foxp3 and CD39 expression on total CD4 T cells in unstimulated human PBMC (18 hrs at 37°C). Numbers shown are mean percentages; n = 12. (E) Example plots of signal detected using different combinations of AIM markers: OX40/CD25 in red, PD-L1/CD25 in orange, and PD-L1/OX40 in yellow. (F) Quantitation of AIM response via different marker combinations after stimulation with tetanus peptide pool. (G) Fold increase in signal for the different marker combinations (tetanus peptide pool stimulation response divided by the unstimulated condition). (H) Graph showing the percent of Foxp3⁺ Treg cells within the AIM⁺ total CD4 population in tetanus stimulated human PBMC where the AIM⁺ population is defined by different combinations of markers. For A, B, E, F, G, and H, n = 12 donors. (I-K) PBMCs from 5 HIV-infected donors stimulated for 18 hrs with various HIV peptide antigens (Gag, Pol, and Nef). Antigen-specific CD4 T cells were identified by the upregulation of AIM markers and analyzed for the contributions of Helios⁺Foxp3⁺ Tregs. n = 4–5 donors. (I) Example flow plots showing gating of AIM⁺ populations, and subsequent identification of AIM⁺ Foxp3⁺Helios⁺ Tregs, following Gag stimulation (all plots from same donor). (J) Quantification of AIM⁺ signal from the various marker combinations shown in the top panel of (I). (K) Quantification of Helios⁺Foxp3⁺ Tregs within AIM⁺ gates. Treg data shown only where antigen-specific responses were 2-fold over background (2–5 responses per antigen).

Fig 5. Treg cells in M. mulatta PBMC and lymph nodes.

(A) Example gating scheme showing Foxp3 and Helios expression in OX40⁺CD25⁺ GC Tfh (PD-1^hiCXCR5^hi) from M. mulatta lymph node. (B) Quantitation of the percent of Treg cells within the OX40⁺CD25⁺ GC Tfh population after stimulation with antigen (BG505) (M. mulatta LN, n = 8 animals). (C) Quantitation of Treg cells within the OX40⁺CD25⁺ total CD4 T cell population after stimulation with antigen (BG505) (M. mulatta PBMC, n = 9 animals). (D) Frequencies of OX40⁺CD25⁺ nTreg cells (Foxp3⁺Helios⁺) within the total CD4 T cell population in M. mulatta PBMC compared between the unstimulated condition “UN” and in response to BG505. n = 7. (E) Example plot of CD39 and Foxp3 expression of unstimulated total CD4 T cells in M. mulatta PBMC after 18 hrs of incubation at 37°C. (F) Venn diagram demonstrating overlap between Foxp3 and CD39 expression on total CD4 T cells in M. mulatta PBMC after 18 hrs of incubation at 37°C without stimulation. Numbers shown are mean; n = 7. (G) Example gating scheme for detection of OX40 and 4-1BB (purple) or OX40 and CD25 (red) AIM⁺ CD4 T cells in M. mulatta LN. (H) Fold-increase (signal from the antigen condition divided by the unstimulated condition) for the different marker combinations: OX40⁺CD25⁺ (red) or OX40⁺4-1BB⁺ (purple) (M. mulatta PBMC and LN, n = 15). (I) Comparison of Treg cell frequencies within the AIM⁺ CD4 T cell population for the different marker combinations: OX40⁺CD25⁺ (red) or OX40⁺4-1BB⁺ (purple) (PBMC, incubated at 37°C for 18 hrs. n = 3).