A High Throughput Whole Blood Assay for Analysis of Multiple Antigen-Specific T Cell Responses in Human Mycobacterium tuberculosis Infection

Abstract

Antigen-specific CD4 and CD8 T cells are important components of the immune response to Mycobacterium tuberculosis, yet little information is currently known regarding how the breadth, specificity, phenotype, and function of M. tuberculosis-specific T cells correlate with M. tuberculosis infection outcome in humans. To facilitate evaluation of human M. tuberculosis-specific T cell responses targeting multiple different Ags, we sought to develop a high throughput and reproducible T cell response spectrum assay requiring low blood sample volumes. We describe here the optimization and standardization of a microtiter plate-based, diluted whole blood stimulation assay utilizing overlapping peptide pools corresponding to a functionally diverse panel of 60 M. tuberculosis Ags. Using IFN-γ production as a readout of Ag specificity, the assay can be conducted using 50 μl of blood per test condition and can be expanded to accommodate additional Ags. We evaluated the intra- and interassay variability, and implemented testing of the assay in diverse cohorts of M. tuberculosis-unexposed healthy adults, foreign-born adults with latent M. tuberculosis infection residing in the United States, and tuberculosis household contacts with latent M. tuberculosis infection in a tuberculosis-endemic setting in Kenya. The M. tuberculosis-specific T cell response spectrum assay further enhances the immunological toolkit available for evaluating M. tuberculosis-specific T cell responses across different states of M. tuberculosis infection, and can be readily implemented in resource-limited settings. Moreover, application of the assay to longitudinal cohorts will facilitate evaluation of treatment- or vaccine-induced changes in the breadth and specificity of Ag-specific T cell responses, as well as identification of M. tuberculosis-specific T cell responses associated with M. tuberculosis infection outcomes.

Figure 1. Schematic summary of the T cell RSA

Fresh, heparinized whole blood is diluted 1:2 with RPMI-1640; 100µl of 1:2 diluted blood is added to each well containing 100µl of RPMI-1640 containing Mtb peptide pools (blue wells), negative controls containing media alone with no antigen (red wells) and positive controls containing PHA (green wells). The final dilution of blood in each well is 1:4. The plates are incubated at 37°C in a 5% CO₂ incubator for 7 days. Plates are then centrifuged and plasma supernatants are removed for measurement of IFN-γ production by ELISA.

Figure 2. Whole blood dilutions and kinetics of IFN-γ production following stimulation with Mtb peptide pools

Fresh whole blood from two QFT⁺ individuals was used either undiluted (red circles), or diluted 2-fold (green triangles), 5-fold (light blue triangles), and 10-fold (dark blue squares) with RPMI-1640. CFP-10, ESAT-6, and TB10.4 peptide pools were added to each dilution in 96-well plates, and the plates incubated at 37°C for the indicated number of days. On days 1, 3, 5, and 7, plates were centrifuged and plasma was removed for measurement of IFN-γ in supernatants by ELISA. Data from the first donor is shown in the top row; data from the second donor is shown in the bottom row.

Figure 3. Optimization of antigen concentration in the whole blood stimulation assay

Fresh whole blood was collected from 15 QFT⁺ and 13 QFT⁻ individuals. Blood was diluted 2-fold with RPMI-1640 and incubated with the indicated concentrations of CFP-10 (A), ESAT-6 (B), and TB10.4 (C) peptide pools, and with PHA (D), for 7 days. Plasma supernatants were harvested on day 7 for measurement of IFN-γ in supernatants by ELISA. The median and IQR are shown in panels A – C. Differences in IFN-γ production between antigen concentrations were determined using the Wilcoxon matched-pairs signed rank test.

Figure 4. Comparison of IFN-γ production in 1:2 and 1:4 diluted whole blood multiple antigen RSAs

Fresh whole blood was diluted 2-fold and 4-fold with RPMI-1640. Diluted blood from 15 QFT⁺ individuals was incubated with 20 different Mtb peptide pools and PHA for 7 days. Each peptide pool was measured individually in triplicate wells. Plasma supernatants were harvested on day 7 for measurement of IFN-γ in supernatants by ELISA. (A) Comparison of IFN-γ production in 1:2 and 1:4 diluted blood in the negative control wells (no antigen) and positive control wells (5µg/ml PHA). Box plots represent the median and IQR; outliers are shown as individual points. (B) Pearson’s correlation between IFN-γ production in 2-fold and 4-fold diluted to 20 Mtb peptide pools. Each peptide pool was tested in triplicate, and the average IFN-γ production of the triplicate wells is shown for 2-fold and 4-fold diluted blood from the same individuals. The Kendall rank correlation coefficient was used to determine the p-value. (C) Coefficient of variation (CV) of IFN-γ production to 20 different Mtb peptide pools across triplicate wells of each peptide pool (intra-assay variation, grey box plots), and across individual donors (inter-individual variation, green box plots). The average IFN-γ production in the triplicate wells for each peptide pool in each individual was used to calculate the inter-individual CV of IFN-γ production.

Figure 5. Comparison of IFN-γ production in frozen and freshly prepared RSA Ag plates

96-well plates containing Mtb peptide pools (CFP-10 and TB10.4) and PHA in RPMI-1640 were prepared and frozen at −80°C. On the day of blood draw, an identical set of 96-well plates were freshly prepared with Mtb peptide pools and PHA in RPMI-1640. Fresh blood was diluted 1:4 in RPMI-1640 and added to the 96-well plates that had been previously prepared (‘Frozen Ag plate’, open circles; plates were thawed at room temperature on the day of blood collection), as well 96-well plates that had been prepared with Ags just prior to blood collection (‘Fresh Ag plate’, filled circles). Plasma supernatants were harvested on day 7 for measurement of IFN-γ in supernatants by ELISA. Comparisons of IFN-γ production in the fresh and frozen antigen plates were done using the Wilcoxon matched-pairs signed rank test.

Figure 6. Reproducibility of the T cell RSA across three study visits

Blood was collected from 15 QFT⁺ individuals at three time points, spaced at weekly intervals (Visit 1, Visit 2, and Visit 3). Blood was diluted 1:4 in RPMI-1640 and stimulated for 7 days with 20 Mtb peptide pools and PHA. Background IFN-γ production in the negative control wells was subtracted from the Ag-stimulated wells. The mean and 95% confidence level of IFN-γ production are shown for each peptide pool at each study visit for 15 QFT⁺ individuals.

Figure 7. Application of the T cell RSA to Mtb-unexposed and LTBI cohorts across international study sites

The whole blood multiple antigen response spectrum assay was conducted with 1:4 diluted blood from 6 Mtb-unexposed healthy donors enrolled in Atlanta, GA (A), 15 QFT⁺ individuals enrolled in a refugee cohort in Atlanta, GA (B), and 15 QFT⁺ individuals enrolled in Kisumu, Kenya (C). None of the healthy donors have been vaccinated with BCG. Of the 15 QFT⁺ individuals enrolled in Atlanta, 10 have been vaccinated with BCG, 3 have not been vaccinated, and 2 are unsure if they have received the BCG vaccine. All QFT⁺ individuals enrolled in Kenya have been vaccinated with BCG. Diluted blood was stimulated with 60 Mtb peptide pools and PHA. Results are shown after subtraction of background IFN-γ production in the negative control wells. Horizontal bars represent the median; vertical bars represent the interquartile range.

Figure 8. Application of the T cell RSA to longitudinal cohorts of individuals with LTBI in a TB-endemic setting in Kisumu, Kenya

The whole blood multiple antigen RSA was conducted with 4-fold diluted blood from 10 QFT⁺ individuals with LTBI enrolled in Kisumu, Kenya, at three time points: enrollment (Baseline, top row), month 6 (middle row) and month 12 (bottom row). All individuals have been vaccinated with BCG. Diluted blood was stimulated with 60 Mtb peptide pools and PHA; the numbers in each cell represent the concentration of IFN-γ after subtraction of the background IFN-γ production in the negative control wells. Data are shown as a heat map, with low IFN-γ levels shown in light blue, and increasing concentrations of IFN-γ indicated by darker blue cells. The 60 Mtb peptide pools are shown in the columns; the same 10 individuals are shown consecutively in the same order in each row for the three time points.