Read on Twitter
Excited to share our latest work on the metabolism of T cell exhaustion in @NatImmunol. We identify mitochondrial stress, induced via combinations of signals common in tumor microenvironments, as a key driver of T cell exhaustion. A (lengthy) thread: https://www.nature.com/articles/s41590-020-00834-9 1/
Back in 2016, we reported in @ImmunityCP that T cells infiltrating tumors had a striking metabolic phenotype: repressed glucose uptake (also seen by @Tcellogic/ @HolabUnil) and mitochondrial atrophy. The latter was directly correlated to the severity of T cell exhaustion. 2/
However, what remained unclear to us was whether the mitochondrial phenotype observed was a consequence of T cell exhaustion, or whether metabolic stress actually caused exhaustion. So @scharpingne and I embarked on a 4 year journey to address this question. 3/
We set out to design an in vitro system, based on in vivo readouts, that would generated exhausted T cells: cells with low metabolic sufficiency and self-renewal capacity, high co-inhibitory molecule expression, and reduced polyfunctionality. 4/
Hypoxia is a common driver of mitochondrial dysfunction, so we used pimonidazole to mark hypoxic cells in tumors: turns out CD8+ TIL do not uniformly experience hypoxia. Terminally exhausted T cells (PD1hiTim3+) experience much more hypoxia compared to progenitor-like cells. 5/
But, we reasoned hypoxia alone was not the whole answer to driving exhaustion. As persistent antigen plays a major role in exhaustion, Nicole designed an in vitro system in which T cells experience continuous stimulation, hypoxia, or both. 6/
T cells get activated overnight using stimulatory beads, then passed into 1 of 4 conditions: a ‘typical’ expansion in IL-2 or continuous coculture with stimulatory beads. Either of these conditions occurred either at ambient normoxia or in a hypoxia chamber set to 1.5% O2. 7/
Remarkably, T cells could 'withstand' either continuous stimulation or hypoxic conditions (in fact, hypoxic culture alone made T cells exceptional polyfunctional effectors). But experiencing the combination of continuous stimulation AND hypoxia generated dysfunctional T cells. 8/
Sequencing the 4 conditions showed cont stimulation + hypoxia promoted distinct transcriptional programs consistent w/ exhaustion; not just combo of ‘continuous stim’ and ‘hypoxia’ signatures: thus continuous TCR may upregulate a program altering how cells respond to hypoxia. 9/
Thanks to major help from @PoholekLab, we determined that continuous stim upregulated the transcriptional repressor Blimp1, which suppressed a host of genes associated with metabolic reprogramming, including one of our favorites, PGC1a, driving mitochondrial biogenesis. 10/
Blimp1-deficient T cells could not be rendered exhausted in our in vitro system, and using a temporal deletion strategy, we showed that removal of Blimp from pre-existing exhausted T cells in tumors could restore their metabolism and effector function. 11/
But what was important about mitochondria for the avoidance of exhaustion? We started culturing T cells (after 24h activation) in mitochondrial inhibitors. Intriguingly, culture in a typical ‘Seahorse mix’ of inhibitors (rotenone and antimycin A) didn’t induce dysfunction. 12/
But when Nicole did the controls (rotenone or antimycin A alone) she found something super cool: culture of antimycin A alone DID induce exhaustion. Antimycin A targets complex III, generating mitochondrial reactive oxygen species at complex I. This also happens under low O2! 13/
ROS was directly inducing an exhausted-like state: T cells cultured in antimycin A (alone), cont. stim + hypoxia, or hypoxic T cells in tumors showed elevated mROS (and a second wave of cytoplasmic ROS), and antioxidants (n-ac-cysteine) could reverse dysfunction. 14/
Thus, it turns out we had it wrong: it wasn’t that functional mitochondria were preventing exhaustion, rather, dysfunctional mitochondria (and consequent ROS production) were driving exhaustion. 15/
So what was ROS doing? I reckon that a lot of people are addressing this question in T cells, and I’m sure ROS may be inducing DNA damage, altering metabolic reprogramming, and inhibiting histone/DNA demethylases. But ROS (as peroxide) can also act as a phosphatase inhibitor. 16/
Indeed, just inducing ROS in T cells (using antimycin A) elevated total phospho-tyrosine levels sufficient to drive NFAT into the nucleus, trigger Nur77-GFP reporting, and elevate Blimp1 and Tox. 17/
We thought this was pretty neat: continuous TCR stimulation altered metabolism, but also altered metabolism (via ROS) fed forward and mimicked TCR signaling. We reason that this may be part of what continually drives this pathologic differentiation to exhaustion in disease. 18/
Finally, how do we overcome these problems? As removal of continuous TCR stimulation would be a complex therapeutic effort, we instead focused on mitigating ROS/hypoxia. Reducing ROS via overexpression of antioxidant proteins in therapeutic T cells prevented exhaustion. 19/
We could also lower hypoxia by altering the tumor microenvironment using anti-angiogenics. Treatment with low-dose axitinib reduced intratumoral hypoxia, prevented intratumoral T cell exhaustion, and synergized with immunotherapy in B16. 20/
While axitinib + anti-PD1 is approved in RCC, we are currently running this trial with @YanaNajjarMD in PD-1 refractory melanoma ( https://clinicaltrials.gov/ct2/show/NCT04493203). 21/
Our data, along with fantastic complementary findings from @SantoshVardhana and @HolabUnil (also at @NatImmunol), place the mitochondria critical determinants of T cell fate, especially in terms of exhaustion. 22/
This was a heruclean effort by @scharpingne to cap off her PhD before she headed off to postdoc with @GoldrathLab. Across the country during a pandemic she stewarded this paper through review, along with incredible efforts from Dayana, Ashley, and Paolo in the lab. 23/
And of course, we could not do any of this work without immense support of @UPMCHillmanCC, @PittTweet, collaboration with @poholeklab and @rhodesf0rd, and funding from @TheMarkFdn, @CancerResearch, @SU2Cscience, @ACGTFoundation and NIH. Full text at: https://rdcu.be/cc2CG 24/24
