P-glycoprotein expression skews mitochondrial dye measurements in T cells

Abstract

Assays to monitor metabolic parameters of immune cells at a single cell level provide efficient means to study immunometabolism. We show here that staining intensity of mitochondria targeting probes in T cells is dramatically influenced by P-glycoprotein/P-gp expression, a xenobiotic efflux pump that extrudes these fluorescent dyes. Discrepancies between MitoTracker Green FM/MTG signals and multiple dye-independent measurements are seen in CD4 T and CD8 T cell subsets and are corrected by P-gp inhibition (PSC833) during MTG staining. We further investigate invariant Natural Killer T (iNKT) cells, which express the highest level of P-glycoprotein among T cells. Using mtDNA abundance, mitochondrial volume, respiration and proteomics, we establish that iNKT cells have higher mitochondrial content and activity than CD4 T cells, opposite to what MTG signals reveal. A similar phenomenon is also seen in human PBMCs, and with TMRE, a dye indicator of mitochondrial membrane potential. Collectively, these data argue that P-glycoprotein expression is a significant confounding factor when analyzing T cells using mitochondrial specific dyes. Complementary methods are necessary to reliably assess mitochondrial features in T cells.

Keywords: P-glycolprotein; T cells; TMRE; mitochondria; mitotracker; oxidative phosphorilation.

Figure 1

P-glycoprotein (P-gp) expression affects mitochondrial dye measurements in T cells. (a, b) MTG signals from splenic naïve (CD44^-CD62L⁺) and memory (CD44⁺CD62L⁺) CD8 T cells stained with or without PSC833 (1uM). (c) Protein expression level of P-glycoprotein based on Variance Stabilization Normalization (VSN) Normalized Intensities from proteomic analysis. (d, e) MTG and (f, g) TMRE signals from thymic CD4 T (CD8a^-CD1d-aGC-tetramer^- CD4⁺TCRβ^hi) and iNKT (CD8a^- CD1d-aGC-tetramer⁺ TCRb^int) cells stained with or without PSC833 as in (a). (h, i) MTG and (j, k) TMRE signals from PBMC CD4 T (CD8a^-CD1d-aGC-tetramer^- CD4⁺TCRb^hi) and iNKT (CD8a^- CD1d-aGC-tetramer⁺ TCRb^int) cells stained with or without PSC833 as in (a). The data shown in (a, b, d-k) are one representative experiment out of three independent experiments. The data shown in c are from samples collected from two sorting experiments and proteomes analyzed together. ns, non-significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, paired Student’s t test.

Figure 2

Mitochondrial proteome supports high oxidative phosphorylation and mitochondrial metabolism in iNKT cells. (a) Gene set enrichment analysis of 107 oxidative phosphorylation related proteins comparing the proteomes of thymic iNKT cells and CD4 T cells. (b, c) Differentially expressed Electron Transport Chain (ETC) complex proteins (b), TCA cycle related and pyruvate dehydrogenase complex proteins (c).

Figure 3

iNKT cells exhibit high mitochondrial content and activity. (a) Quantification of mitochondrial genome encoded genes relative to nuclear DNA (nDNA) in freshly isolated thymic T cells. (b) Ideogram of the 16,299 basepair (bp) mtDNA genes and annotations, and the variant frequencies plotted at each variant locus identified from thymic and splenic CD4 T cells and iNKT cells. (c, d) Confocal imaging of freshly isolated thymic T cells from Dendra2 transgenic mice. Scale bars indicate 2μm. (e) Quantification of mitochondrial volume based on Dendra2 signals as in c and d. (f) Average oxygen consumption rate (OCR) in sorted thymic T cells measured in a Seahorse Mitostress test. (g) Basal OCR, (h) proton leak, (i) maximal OCR, and (j) spare respiratory capacity were calculated based on (f). The data shown in a, c, d and f are one representative experiment out of three independent experiments. The data shown in e and g-j are pooled from three independent experiments. **p < 0.01, ****p < 0.0001, unpaired Student’s t test.