eUREkA! T cells answer nature’s call

Nitrogen is critical for cell growth, proliferation, and function. Amino acid (AA) catabolism is a major source of intracellular nitrogen for biosynthesis but also generates the nitrogen-containing molecule ammonia (NH₃⁺), which can accumulate to toxic levels in a cell if not exported to the circulation and detoxified by the urea cycle in the liver. In a recent paper in Nature Immunology, Tang et al. [1] discovered a functional urea cycle in memory T cells, which metabolizes ammonia, prevents its toxic accumulation, and consequently promotes the maintenance of this long-lived cell type critical for recall responses to infection and cancer.

Nitrogen is essential for the synthesis of many biomolecules, including AAs, nucleotides, polyamines, hexosamines, nitric oxide (NO), and glutathione (GSH) [2]. AA catabolism generates glutamate via transamination, which undergoes deamination to α-ketoglutarate (α-KG) and ammonia. Some amino acids, e.g. glutamine, can be directly deaminated, also releasing ammonia. Most cells export ammonia to the circulation, where it is acquired by hepatocytes and metabolized to nontoxic urea, which is exported from the cell and excreted. This cycle was thought to operate only in the liver, but Tang et al. demonstrated urea formation via the same enzymes in CD8⁺ memory T (T_mem) cells (Fig. 1) as a means of detoxifying ammonia in these long-lived cells [1]. They investigated ammonia metabolism in effector (T_eff, Day 7 post-infection) and memory (Day 30) T cell responses to infection with Listeria monocytogenes-expressing ovalbumin (Lm-OVA). The levels of carbamoyl phosphate (CP), urea, and other urea cycle intermediates (ornithine, citrulline, argininosuccinate, and arginine) were elevated in T_mem cells but not T_eff cells after infection, and were higher in in vitro-generated T_mem cells than T_eff cells. Tracing nitrogen into these intermediates using heavy-labeled ammonium (¹⁵N-NH₄Cl) or the endogenous ammonia donor ¹⁵N-glutamine revealed increased labeling of urea cycle intermediates in T_mem cells. Collectively, these results demonstrate an operational urea cycle outside hepatocytes, in T_mem cells.

Fig. 1.

T_mem cells engage the urea cycle to detoxify ammonia. Intracellular ammonia (NH₃⁺) levels can increase due to amino acid metabolism or as a consequence of ammonia accumulation in the microenvironment. Tang et al. identified a urea cycle in T_mem cells, in which ammonia is first metabolized in the mitochondria to carbamoyl phosphate (CP) via CP synthetase-1 (CPS1), the urea cycle’s rate-limiting enzyme. CP combines with ornithine to form citrulline, which is exported to the cytosol, where it reacts with aspartate to form argininosuccinate. Finally, argininosuccinate is metabolized to arginine and fumarate, and arginine is transported into the mitochondria for hydrolysis by ARG2 to ornithine and urea. This pathway may be important for ammonia detoxification and T_mem cell maintenance in contexts such as the tumor microenvironment, in which ammonia can accumulate to toxic levels. The figure was created using BioRender.com

Pharmacological or genetic inhibition of CP synthetase-1 (CPS1) elevated intracellular ammonia levels in T_mem cells, showing that these cells do indeed metabolize ammonia. Restoring ammonia levels decreased T_mem cell numbers, suggesting that CPS1 has a role in T_mem cell formation or persistence. Furthermore, the transfer of CPS1-deficient OT-I CD8⁺ T cells into mice prior to Lm-OVA infection resulted in the generation of fewer T_mem cells in vivo and increased cell death compared to transfer of control cells. As CPS1 was needed to protect T_mem cells from excess ammonia-induced cell death in vitro, it will be important to measure ammonia in the extracellular fluid in vivo, to investigate what levels of ammonia are sufficient to cause death. The urea cycle may therefore maintain T_mem cell populations both by enhancing their formation and by improving their survival.

In hepatocytes, CPS1 generates mitochondrial citrulline, which is exported to the cytosol and is metabolized via ornithine transcarbamoylase (OTC), argininosuccinate synthase (ASS1), argininosuccinate lyase (ASL), and ARG1 to urea. While OTC, ASS1, and ASL levels were increased in T_mem cells vs. T_eff and naïve T cells, ARG1 was not detected, implying a divergence of the T_mem cell urea cycle from the hepatic cycle. The authors hypothesized that mitochondrial arginase, ARG2, may instead catalyze the last step of this cycle. Supporting this hypothesis, ARG2 levels were increased in T_mem cells, and ARG2 inhibition increased the levels of ammonia, CP, citrulline, argininosuccinate, and arginine, but decreased the levels of ornithine and urea (the products of arginase). The use of mitochondrial ARG2 rather than cytosolic ARG1 means that this step is differentially localized in T_mem cells compared to hepatocytes and raises the question of why T_mem cells transport arginine into the mitochondria. ARG2 generates mitochondrial ornithine, and it may be advantageous to have this CPS1 substrate in close proximity to CPS1 to rapidly initiate ammonia detoxification.

Beyond being a source of urea, arginine is metabolized by the citrulline cycle, in which nitric oxide synthase (NOS) uses arginine to regenerate citrulline and form NO [3]. This cycle also operates in T_mem cells, which have higher levels of NOS enzymes (iNOS, eNOS, and nNOS) and release more NO than effector or naïve T cells. iNOS or eNOS inhibition increased T_mem cell ammonia levels as well as ¹⁵N-NH₄Cl labeling of arginine, indicating a blockade of NOS-mediated arginine metabolism. The investigators traced ¹⁵N-arginine into the urea and citrulline cycles to determine their relative contributions to ammonia disposal, but did not investigate other fates of arginine, e.g. protein or polyamine synthesis. Together, these results suggest that T_mem cells are well equipped to metabolize ammonia, as they catabolize ammonia-derived arginine via multiple routes.

The authors then aimed to promote the urea cycle in T_mem cells in vivo by CPS1 overexpression (CPS1-OE). This was sufficient to drive the T_mem urea cycle, as CPS1-OE T cells had less ammonia and more urea cycle intermediates. CPS1-OE in OT-I T cells increased T_mem cell frequency in Lm-OVA-infected mice, perhaps due to enhanced survival of the T_mem cells. While effector molecule expression by these cells was unaltered after acute (4 h) Lm-OVA rechallenge, there were more T_eff cells in infected mice 5 days post-rechallenge, suggesting that CPS1 affects T_eff cell quantity after reinfection. It is unclear whether T_mem cells lose urea cycle enzyme expression as they reactivate, and whether ammonia detoxification is more important for long-term persistence or reactivation. It would also be intriguing to know whether CPS1-OE cells better control infection. The authors additionally reveal how CPS1 expression is modulated in T_mem cells, which may be important for therapeutically targeting CPS1. T_mem cell ketogenesis results in the production of the metabolite β-hydroxybutyrate (BHB) [4], which epigenetically modifies histone H3K9 on the Cps1 locus to drive its expression. Inhibiting BHB production, or H3K9 hydroxybutyration via P300, decreased CPS1 levels, showing how T_mem cell metabolism regulates gene expression to induce the urea cycle, ensuring survival.

The T_mem cell urea cycle may be particularly useful for tumor-infiltrating T cells, as ammonia accumulates in the tumor microenvironment (TME) [5]. Cancer cells recycle ammonia into glutamate, which is then used for AA synthesis to support growth and proliferation, as demonstrated for breast cancer cells [6]. While supporting cancer cell activity, such increased TME ammonia levels may concomitantly antagonize antitumor immune responses. In advanced colorectal cancer, ammonia accumulation in the TME and intracellularly in T cells (but not in colon cancer cells) causes T cell exhaustion and attenuates T cell responses to checkpoint blockade therapy [5]. This study did not specifically examine T_mem cells, which Tang et al. have now shown have the capacity to detoxify ammonia. To this end, Tang et al. transferred CPS1-OE T_mem cells derived from Lm-OVA-challenged mice into naïve mice prior to B16-OVA melanoma inoculation, leading to increased tumor clearance and survival relative to mice receiving control cells. It would be intriguing to further investigate the mechanism of this protection and whether it is due to increased antitumor T cell functionality or numbers.

This work suggests that engineering a functional urea cycle and intracellular ammonia detoxification in other tumor-infiltrating cells may be of therapeutic benefit, e.g. for CAR T cells. While the authors clearly described the mechanisms underlying T_mem CPS1 expression, factors controlling other urea cycle enzymes remain unknown. Ketogenesis-derived BHB may play a role, as it can also modify loci such as Foxo1 and Ppargc1a, which themselves influence T_mem cell formation [4]. HNF4α, a master regulator of hepatic urea cycle factors [7], is another candidate for further investigation. An additional outstanding question is why T_mem cells metabolize ammonia rather than export it for hepatic detoxification. Urea cycle products may support T_mem cells. For example, arginine improves T cell survival and may help maintain central T_mem cells; arginine promotes oxidative phosphorylation (OXPHOS) while inhibiting glycolysis in human CD4⁺ T cells [8], possibly explaining the differential usage of the urea cycle in T_mem (rely on OXPHOS) vs. T_eff cells (more glycolytic). NO can also influence cellular metabolism and modify proteins to alter their activity, e.g. via S-nitrosylation of thiol groups, with consequences for cellular function [9]. Polyamine metabolism also warrants further research. Does compartmentalization of ornithine in the mitochondria restrict its use for cytosolic polyamine generation and eIF5a-dependent translation [10]? Finally, ammonia can inhibit the transsulfuration of methionine to cysteine, inducing oxidative stress [5]. Does ammonia detoxification relieve this inhibition and allow T_mem cells to better clear reactive oxygen species and maintain cell survival? Overall, this study reveals a novel method of controlling nitrogen and amino acid metabolism in T_mem cells, with the potential to exploit these pathways for therapeutic benefit.

Competing interests

The authors declare no competing interests.