PD-1 signaling affects cristae morphology and leads to mitochondrial dysfunction in human CD8+ T lymphocytes

Abstract

Background Binding of the programmed death-1 (PD-1) receptor to its ligands (PD-L1/2) transduces inhibitory signals that promote exhaustion of activated T cells. Blockade of the PD-1 pathway is widely used for cancer treatment, yet the inhibitory signals transduced by PD-1 in T cells remain elusive. Methods Expression profiles of human CD8 + T cells in resting, activated (CD3 + CD28) and PD-1-stimulated cells (CD3 + CD28 + PD-L1-Fc) conditions were evaluated by RNA-seq. Bioinformatic analyses were used to identify signaling pathways differentially regulated in PD-1-stimulated cells. Metabolic analyses were performed with SeaHorse technology, and mitochondrial ultrastructure was determined by transmission electron microscopy. PD-1-regulated mitochondrial genes were silenced using short-hairpin RNA in primary cells. Blue native gel electrophoresis was used to determine respiratory supercomplex assembly. Results PD-1 engagement in human CD8 + T cells triggers a specific, progressive genetic program different from that found in resting cells. Gene ontology identified metabolic processes, including glycolysis and oxidative phosphorylation (OXPHOS), as the main pathways targeted by PD-1. We observed severe functional and structural alterations in the mitochondria of PD-1-stimulated cells, including a reduction in the number and length of mitochondrial cristae. These cristae alterations were associated with reduced expression of CHCHD3 and CHCHD10, two proteins that form part of the mitochondrial contact site and cristae organizing system (MICOS). Although PD-1-stimulated cells showed severe cristae alterations, assembly of respiratory supercomplexes was unexpectedly greater in these cells than in activated T cells. CHCHD3 silencing in primary CD8 + T cells recapitulated some effects induced by PD-1 stimulation, including reduced mitochondrial polarization and interferon-γ production following T cell activation with anti-CD3 and -CD28 activating antibodies. Conclusions Our results suggest that mitochondria are the main targets of PD-1 inhibitory activity. PD-1 reprograms CD8 + T cell metabolism for efficient use of fatty acid oxidation; this mitochondrial phenotype might explain the long-lived phenotype of PD-1-engaged T cells. Electronic supplementary material The online version of this article (10.1186/s40425-019-0628-7) contains supplementary material, which is available to authorized users.

Immunometabolism: From basic mechanisms to translation

“…83 Critically, treatments such as checkpoint blockade therapies act in part through relieving this metabolic inhibition. [15][16]84 Siska et al focus on the "Warburg effect" which is a key feature of tumor cells and leads to acidification and the accumulation of lactate in the tumor microenvironment. 85 As Both systemic and local influences can also affect metabolism and anti-tumor immunity.…”

PD-1-induced T cell exhaustion is controlled by a Drp1-dependent mechanism

“…Second, engagement of PD-1 signaling downregulates both glycolysis and OXPHOS in stimulated T cells (41). However, glycolysis is downregulated to a higher extent than OXPHOS, this forcing PD1-engaged T cells to rely more on an OXPHOS-based metabolism (22,41).…”

“…Second, engagement of PD-1 signaling downregulates both glycolysis and OXPHOS in stimulated T cells (41). However, glycolysis is downregulated to a higher extent than OXPHOS, this forcing PD1-engaged T cells to rely more on an OXPHOS-based metabolism (22,41). Although this mechanism may allow tumor-infiltrating T cells to survive in a tumor microenvironment characterized by low glucose availability (22), the underlying mechanism favoring such metabolic adaptation is not known yet.…”

“…Of note, Ogando et al recently analysed the effect of PD-1 signaling on the assembly of mitochondrial cristae (41). Although they reported no variations in the rate of mitochondrial fragmentation in T cells stimulated in the presence or absence of PD-1 engagement, they also reported no differences in the rate of mitochondrial fragmentation between unstimulated and stimulated T cells (41). However, fragmentation of the mitochondrial network upon T cell stimulation (independently of PD-1 signaling) has been reported extensively by us and others (15,17,23).…”

“…Indeed, we believe that such a parameter may not be suitable to compare mitochondria morphology in stimulated and unstimulated naïve T cells, since the former cells grow and enlarge significantly compared to the latter ones upon in vitro culture. In addition, Ogando et al only analysed total levels of Drp1, without focusing on its specific phosphorylated residues (41), which are more reliable indicators of Drp1 activation compared to the total protein amount.…”