top of page

HIF-1α-mediated mitochondrial-glycolytic reprogramming controls the transition of precursor to terminally exhausted T cells

Wu et al. (BioRxiv) DOI: 10.1101/2023.08.31.555662

HIF-1α-mediated mitochondrial-glycolytic reprogramming controls the transition of precursor to terminally exhausted T cells


  • T cell exhaustion

  • Immunometabolism

  • CAR T cells

  • mitochondria

Main Findings

CD8+ T cell exhaustion is a phenomenon often observed in individuals living with chronic illnesses such as chronic viral infections and cancer. Due to sustained antigen exposure the cells reduce expression and secretion of cytolytic proteins and cytokines and upregulate inhibitory receptors. Exhaustion follows a trajectory pathway in which progenitor exhausted T cells (TPEX) give rise to a larger number of terminally exhausted T cells (TEX). Clinically, there is an increased interest in understanding how TPEX can be maintained, since these are the primary targets of immune checkpoint inhibition therapy and can still give rise to a substantial burst of functional CD8+ T cells.

The study by Wu et al. dissects the metabolic requirements that drive CD8+ T cell exhaustion. By and large the authors find that the metabolism of TEX is driven by glycolytic pathways, while TPEX seem to more strongly rely on mitochondrial respiration. Using a KO model of the mitochondrial inorganic phosphate transporter Slc25a3 (also known as mPiC) in T cells, the authors find that the loss of this transporter can indeed reduce basal mitochondrial respiration and increase basal glycolysis. This seems to be sufficient to drive increased expression of exhaustion-associated markers (e.g. TIM-3), but also somewhat reduce their ability to produce multiple cytokines. This goes along with an increased drive to apoptosis of Slc25a3-KO cells upon activation.

Importantly, the authors find that under both chronic viral infection (using LCMV cl13) or even acute viral infection (using LCMV Arm), the KO show increased signs of exhaustion. Conversely, overexpression of Slc25a3 in antigen-specific T cells, thereby boosting mitochondrial respiration, shifts these cells towards an increased presence of TPEX.

Metabolomic analysis of TPEX in Slc25a3-KO cells associates with reduced presence of metabolites of the pentose phosphate pathway and a reduced availability of NADPH. Since NADPH is also an important antioxidant, this finding appears to fit to the increased mitochondrial stress (i.e. ROS) in Slc25a3-KO T cells. Treatment with the antioxidant NAC reduced exhaustion marker expression suggesting that mitochondrial respiration is upstream of CD8+ T cell exhaustion. The authors then identify HIF1a stabilization as a main driver of the transcriptional control of many exhaustion associated marker genes.

To validate the role of Hif1a in CD8+ T cell exhaustion, the authors then deleted Hif1a from T cells and found that it (expected) reduced glycolysis and also shifted their phenotype towards TPEX characteristics. Thus implicating Hif1a activity in the differentiation of terminally exhausted CD8+ T cells. Lastly, the authors also apply their findings in the context of pharmacological treatment of CD8+ T cells, by pre-treating antigen-specific CD8+ T cells with the glucose analogue 2-DG (to prevent glycolysis). This pre-treatment seemed to prevent expression of a TEX-associated phenotype in the chronic LCMV cl13 infection model. Similarly, pre-treatment of CAR T cells with 2-DG seemed to slightly improve anti-tumour responses of these cells.

Overall, this thorough study further connects reduced mitochondrial fitness with the differentiation of terminally exhausted T cells. It adds into a series of previous reports trying to identify the molecular and metabolic drivers of CD8+ T cell exhaustion. While the study will certainly add to the field, some few limitations of the study should be observed for the moment.


  • A central aspect to ICI therapy is to push TPEX towards a proliferative burst to give rise to functional CD8+ T cell responses. Unfortunately, the authors did not further investigate how ICI therapy reshapes the metabolic response of TPEX and which metabolic requirements would push T cells themselves down this path. Meaning: What happens during reinvigoration of T cells on a metabolic level? What are the metabolic requirements for reinvigoration of T cells? What happens to T cell metabolism during ICI therapy?

  • While the final pieces of data are interesting (using 2-DG treatment to “prime” T cells towards a particular phenotype), the study would really benefit from an ATACseq analysis of the cells after 2-DG treatment and after having been transferred in the host to check whether 2-DG treatment can manifest some epigenetic changes that promotes their TPEX phenotype.

  • Additionally, the authors state in the abstract that HIF1a is proteasomally controlled however no data is shown about disturbance of the proteasome and its impact on HIF1a. This comes in addition, since mitochondrial health is often also influenced by mitophagy (which has been evidenced by Yu et al. 2020 Nat Immuol).

  • Several studies have already identified the significance impact of mTOR in controlling terminal differentiation of T cells (Ando et al 2020 JCI, Gabriel et al. 2021 Immunity, Bengsch et al. 2016 Immunity) and it is well known that mTOR can control HIF1a signalling. Thus, it would be good if the authors could implicate mTOR in their pathway more specifically.

  • Considering the increased apoptosis of Slc25a3 KO cells, one wonders if the T cell model describes a true exhaustion phenotype or whether it rather just ablates the cells eventually. Since the cells in some transfer experiments seem to completely disappear, it sounds more like that the cells enter an exhaustion-associated apoptosis pathway.

  • Likely outside of the scope of this preprint, but it would have been nice to get more detailed information of the Slc25a3 KO and its impact on acute infections and how it may control normal memory T cell differentiation.

  • Additionally, we would advise the authors to carefully use the language, some sections felt to these readers a little overstated.


The preprint really highlights the impact that mitochondrial disruption can have on the terminal exhaustion of CD8+ T cells. It adds some causative data into a field that has long been associating mitochondrial impairment with CD8+ T cell exhaustion. While perhaps the concept is not super novel, the data presented show a fairly definitive case for its direct impact on transcription and expression of exhaustion-associated genes and (dys)function.

The study can be broadly influential in considering the metabolic tweaking of ICI therapy and CAR T cells, although the data presented may not yet show the optimized cases (i.e. the relatively slight improvement of CAR-T cell therapy on survival and tumor growth upon 2-DG treatment), but certain allows for more research into this avenue to improve cancer therapy. As such, the study is of interest for scientists working in the immunology, immunometabolism and cancer/CAR T cell research.


Reviewed by Felix Clemens Richter and Mariia Saliutina as part of a cross-institutional journal club between the Vanderbilt University (VUMC), the Max-Delbrück Center Berlin, the Charité Berlin, the Medical University of Vienna and other life science institutes in Vienna.

The author declares no conflict of interests in relation to their involvement in the review.

bottom of page