Pathogen-Specific T Cell Polyfunctionality Is a Correlate of T Cell Efficacy and Immune Protection

Abstract

Introduction: Understanding the factors that delineate the efficacy of T cell responses towards pathogens is crucial for our ability to develop potent therapies against infectious diseases. Multidimensional evaluation of T cell functionality at the single-cell level enables exhaustive analysis of combinatorial functional properties, hence polyfunctionality. We have recently invented an algorithm that quantifies polyfunctionality, the Polyfunctionality Index (Larsen et al. PLoS One 2012). Here we demonstrate that quantitative assessment of T cell polyfunctionality correlates with T cell efficacy measured as the capacity to kill target cells in vitro and control infection in vivo.

Methods: We employed the polyfunctionality index on two datasets selected for their unique ability to evaluate the polyfunctional imprint on T cell efficacy. 1) HIV-specific CD8+ T cells and 2) Leishmania major-specific CD4+ T cells were analysed for their capacity to secrete multiple effector molecules, kill target cells and control infection. Briefly, employing the Polyfunctionality Index algorithm we determined the parameter estimates resulting in optimal correlation between T cell polyfunctionality and T cell efficacy.

Results: T cell polyfunctionality is correlated with T cell efficacy measured as 1) target killing (r=0.807, P<0.0001) and 2) lesion size upon challenge with Leishmania major (r=-0.50, P=0.004). Contrary to an approach relying on the Polyfunctionality Index algorithm, quantitative evaluation of T cell polyfunctionality traditionally ignores the gradual contribution of more or less polyfunctional T cells. Indeed, comparing both approaches we show that optimal description of T cell efficacy is obtained when gradually integrating all levels of polyfunctionality in accordance with the Polyfunctionality Index.

Conclusions: Our study presents a generalizable methodology to objectively evaluate the impact of polyfunctionality on T cell efficacy. We show that T cell polyfunctionality is a superior correlate of T cell efficacy both in vitro and in vivo as compared with response size. Therefore, future immunotherapies should aim to increase T cell polyfunctionality.

Fig 1. The principles of single-cell polyfunctionality analysis for modelling associated variables, such as T cell efficacy.

3 (n) bimodial effector molecules (A, B and C) were measured at the single-cell level identifying 2³ = 8 distinct combinatorial cell subsets, which can be stratified according to the number of simultaneous functions at the single-cell level (i). A) “Classical” polyfunctionality analysis generally assess only cells positive for a defined minimal number of simultaneous effector molecules (i.e. 3 simultaneous functions). B) Contrarily, the polyfunctionality index considers all 8 functional subsets and ingeniously parameterizes the influence of individual (φ_A>φ_B>φ_C) as well as combined (q) functionalities. C) Polyfunctionality assessed with the polyfunctionality index algorithm is therefore modulable, contrary to “classic” polyfunctional analysis, enabling a more optimal fit (green line) of model variables. D) Using regression analysis it is feasible to obtain proper parameter estimates (φ_A, φ_B, φ_C and q), which have biological significance. Indeed, interpretation of such parameters enables an objective evaluation of the influence of individual as well as combinatorial functions on the predictive capacity of polyfunctionality with regards to a desired model variable, such as T cell efficacy.

Fig 2. Polyfunctionality of HIV-specific CD8⁺ T cells is associated with T cell mediated target killing.

In vitro cultured human HIV-specific (KK10) CD8⁺ T cell clones were incubated with KK10-loaded target cells (HLA-B*2705⁺ lymphoblastoid cell lines). To measure T cell killing a standard chromium release assay was employed using ⁵¹Cr charged target cells loaded with a serial dilution of KK10-antigen [10^-6M—10^-13M]. Non-⁵¹Cr charged target cells with an identical serial dilution of KK10-antigen were employed to analyse T cell polyfunctionality by multiparametric flow cytometry (CD107a, IFN-γ, TNF-α, IL-2, MIP-1β). A) Scatter plots showing the association between T cell killing (⁵¹Cr release / max ⁵¹Cr release) and T cell polyfunctionality quantified as the polyfunctionality index (q = 1). B) Non-linear regression for all T cell clones identifies the optimal q-value as 1.24.

Fig 3. Polyfunctionality of Leishmania major-specific CD4⁺ T cells is associated with the degree of protection induced by vaccination.

Mice were vaccinated with Leishmania major antigens employing different antigen preparations (recombinant leishmanial polyprotein (MML) with adjuvant (CpG) and replication-defective adenovirus expressing MML), different doses and different routes of injection. Mice were challenged 28 days post vaccination by intradermal ear injections of Leishmania major carrying parasites. Leishmania major protection levels were measured as peak lesion size. Spleen derived lymphocytes harvested 28 days post vaccination were stimulated with anti-CD28 and MML for 6 hours and analysed for the production of IFN-γ, TNF-α and IL-2 by multiparametric flow cytometry. Experiments were repeated in four independent experiments A) Scatter plots show the association between normalized lesion size and the T cell polyfunctionality quantified as the polyfunctionality index (q = 1). B) Non-linear regression analysis over the 4 experiments identified the optimal q-value as 1.33. Statistical analysis was conducted with parametric aggregate correlation statistics.

Fig 4. Redundancy of effector molecules with regards to the association between HIV-specific CD8⁺ T cell polyfunctionality and T cell mediated target killing.

In vitro cultured human HIV-specific (KK10) CD8⁺ T cell clones were incubated with a serial dilution of KK10-antigen [10^-6M—10^-13M] loaded target cells. T cell polyfunctionality was analysed by multiparametric flow cytometry (CD107a, IFN-γ, TNF-α, IL-2, MIP-1β). A) Stacked bar diagram indicates the probability distribution of T cells expressing 5, 4, 3, 2 or 1 simultaneous effector molecules, given that they express one particular effector molecule. B) Variance Inflation Factor (VIF) analysis of the 5 effector molecules shows the lowest VIF for each molecule after iterative retraction of the effector molecule with the highest VIF>5 (cf. Table 3).