Cannabinoid use is increasingly common among patients with inflammatory bowel disease (IBD), despite limited evidence that it improves underlying disease activity¹⁴ and evidence from a retrospective study suggesting potential harm¹⁵. We have previously demonstrated that cell-specific CB₂R deletion or inverse agonism attenuates a model of human Crohn’s disease^{11, 16}. This current study further informs how CB₂R signalling may influence chronic intestinal inflammation. CB₂R activation increased uptake of the fluorescent glucose analogue 2-NBDG in Jurkat T cells, whereas receptor blockade had the opposite effect. In the brain, CB₂R activation stimulates glucose uptake in astrocytes and neurons. This is inhibited by CB₂R-specific antagonists, consistent with a role for metabolic regulation¹⁷. Glucose metabolism is initially mediated by glucose uptake, primarily regulated by the transporter Glut1, the primary glucose transporter expressed in immune cells⁵. Glucose uptake via membrane-bound glucose transporters suppresses T cell proliferation, yet increased maturation and memory T cell formation¹⁸. Similarly, loss of Glut1 suppresses effector T cell expansion and protected against piroxicam-induced colitis¹⁹ while blocking GLUT1 in a humanized UC model attenuated disease, in part by suppression of central memory CD4⁺ T cell development²⁰. Single-sample gene-set enrichment analysis describes significantly higher rates of glycolysis in IBD patients’ intestinal tissues relative to controls²¹.

CB₂R activation alters metabolism with increased intracellular glucose metabolised by the pentose phosphate pathway (PPP) at the expense of glycolysis. The apparent shift to PPP metabolism in CB₂R⁺ T cells is consistent with previous studies of neuronal ischemia/reperfusion injury in which cannabinoids stimulated glucose metabolism through the PPP to maintain redox balance and energy conservation²². Memory T cells have lower glycolytic rates and higher levels of oxidative metabolism thereby limiting their growth but promoting their persistence²³. In rheumatoid arthritis, a subset of glucose-6 phosphate (G6P) dehydrogenase G6PD^High cells divert glucose away from energy production toward the PPP, resulting in pathogenic CD4⁺ T cells with increased tissue invasiveness, accelerated conversion from naïve to memory phenotype and worsening inflammation²⁴. Increased expression of TALDO1 alone would have minimal impact on nucleotide synthesis as the supply of G6P becomes rate limiting. GALT is a crucial enzyme in galactose metabolism, where it catalyses the reversible transformation galactose-1-phosphate into glucose-1-phosphate²⁵. However, because CB₂R activation also promotes glucose uptake and GALT upregulation to increase G6P synthesis from galactose, now G6P is no longer rate limiting. According to the principles of metabolic control, the pathway then becomes more highly sensitive to TALDO1 regulation, thereby allowing for a rapid increase in nucleotide synthesis without markedly decreasing concentration of metabolic intermediates in the main pathway (glutamine and aspartate) which would oppose the stimulation of macromolecule synthesis. PPP-metabolised glucose produces NADPH and metabolic intermediates required to generate macromolecules such as nucleotides, lipids and cholesterol for biosynthetic fitness^26–28. NADPH is also important for maintaining a degree of redox balance within a cell, providing reduced glutathione and thioredoxins to neutralise ROS²⁹. In lymphocytes, the NADP/NADPH ratio decreases from 1.4 in resting naïve cells to 0.2 in cells 4 days after stimulation, highlighting the requirement for the PPP and subsequent NADPH production in meeting the biosynthetic demands of clonal expansion³⁰. We found CB₂R signalling to significantly enhance the production of NADPH in cannabinoid-treated T cells, while also decreasing the NADP/NADPH ratios.

Consistent with the observed increases in NADPH following CB₂R activation, receptor activation significantly enhanced TALDO1 concentrations. TALDO1 is a crucial enzyme in the PPP, facilitating the transfer of a dihydroxyacetone group from sedoheptulose 7-phosphate to glyceraldehyde-3-phosphate, forming erythrose 4-phosphate and fructose 6-phosphate³¹. This indicates a rapid shunt towards the PPP upon receptor activation. According to the principles of metabolic control laid out by Newsholme et al. (1985), if the flux through one branch is greatly in excess of the other, which in this case could be GALT in providing G6P, then the sensitivity of the low-flux pathway, namely PPP in this regard, becomes highly sensitive to pathway alterations such as TALDO1 regulation³². Pyruvate derived from the PPP is then preferentially shuttled into the TCA cycle and converted into acetyl CoA to fuel increased levels of OxPhos. CNR2 CRISPRi cells confirmed the requirement for CB₂R in mediating these effects, indicating the direct role for CB₂R signalling in T cell glucose metabolism. These cells are more resistant to apoptosis, less metabolically active, and are smaller in size than effector cells, indicative of a cell that is less reliant on metabolism for clonal expansion²⁶.

CB₂R activation in T cells to induce a significant increase in mitochondrial mass. The increase in mitochondrial biogenesis and mass following receptor activation facilitates excessive OxPhos activity, mediating a metabolic shift away from glycolysis towards OxPhos. This shift is marked by enhanced SRC and increased mitochondrial efficiency, mediated by the expression of UCPs 2 and 3, driving proton leak and maintaining mitochondrial efficiencies through the potential reduction in mitochondrial ROS generation. As OxPhos and FAO are mitochondrial metabolic processes, alterations in mitochondrial mass allows memory T cells to effectively utilise these metabolic pathways to facilitate their long-term survival. These alterations in mitochondrial structure and function mediate distinct differences in the metabolic profile of a cell, and are distinct features distinguishing memory T cells from effector T cells. While effector T cells proliferate via mitochondrial fission, memory T cells undergo mitochondrial fusion, tightly assembling their mitochondria with the endoplasmic reticulum resulting in greater mitochondrial mass³³. Memory T cells also exhibit higher SRC, the extra capacity of mitochondria to produce energy in response to increased stress or work, associated with prolonged cellular survival³³. Memory T cells with their high SRC rates and greater mitochondrial mass more rapidly recall and respond to infections using this bioenergetic advantage³³. In contrast, diminished mitochondrial mass of effector cells makes them heavily dependent on glycolysis due to their decreased ability to generate energy from alternative sources such as fatty acids via OxPhos.

Proton leak results in slow respiration, accounting for up to 25% of the basal metabolic rate^{34, 35}. Proton leak reduces the electrochemical proton gradient, which operates in a cytoprotective manner in times of oxidative stress, in turn reducing superoxide production and lowering ROS levels^{34, 36}. We found both SRC and proton leak mirror each other in cannabinoid-treated T cells, with receptor activation significantly enhancing both and CB₂R inverse agonism reducing both. Generally, proton leak in mitochondria is induced by UCPs, particularly UCP2 and UCP3. These are proton channels expressed on the inner mitochondrial membrane which mediate proton leak to reduce ROS production. UCP2 is primarily expressed in immune cells, including T cells, with its expression tightly regulated and increasing significantly during T cell activation and proliferation, suggesting a role for this protein in the immune response^{37, 38}. Both UCP2 and UCP3 are significantly increased in mRNA expression following CB₂R activation, offering a potential mechanism through which CB₂R signalling drives proton leak in T cells. In future, it would be useful to confirm that this increased proton leak and UCP upregulation translates to elevated mitochondrial ROS generation in T cells specifically following CB₂R activation.

Using LC-MS/MS on cannabinoid-treated lysates provided unbiased screening of the T cell metabolome across multiple metabolite compound classes. We employed the MxP^® Quant500 system to measure concentrations of 500 metabolites across 26 different metabolite classes including glycerophospholipids, ceramides, and amino acids which were those most significantly altered following manipulation of CB₂R signalling. The majority of the significantly altered glycerophospholipids were decreased upon CB₂R activation, consistent with their role as substrates for the enhanced FAO and subsequent OxPhos observed in these T cells. In particular, we see a decrease in ceramides with CB₂R activation, which have been previously linked with enhanced glucose utilisation and metabolism^{39, 40}. Ceramide accumulation has also been linked to decreased mitochondrial respiratory chain activity and OxPhos, increased ROS generation, oxidative stress, and mitochondrial dysfunction. Ceramides directly inhibit mitochondrial complex III, to suppress mitochondrial OxPhos^41–43.

To understand the changes in glycerophospholipid and ceramide profile in these T cells, both metabolite classes were plotted based on both their acyl chain carbon length and number of acyl chain double bonds. While most glycerophospholipids are decreased in T cells following CB₂R activation, SM(OH) C24:1, with one double bond and an acyl chain carbon length of 24, was significantly increased. Its role in T cell biology is not well established and therefore warrants further study. CB₂R activation increased PPP-associated amino acids, histidine, phenylalanine, and tyrosine. Histidine production is increased by catabolism of the PPP intermediate ribose 5-phosphate^{44, 45}. Leucine treatment has also been associated with decreases in glycolysis, even in models where glycolysis substrates are sufficient⁴⁶ and leucine-rich environments have been linked with a metabolic switch from glycolysis towards OxPhos⁴⁷. These findings are consistent with the overall hypothesis that CB₂R activation skews T cell metabolism from glycolysis to PPP and OxPhos to facilitate long-term survival. This shift would support memory T cell expression but this has not been proven unequivocally to date.

Human intestinal tissue-resident memory T cells express a range of markers such as CD45RO, CD69, CD95 and CD103¹³. CD103 or integrin αE forms a heterodimeric pair with integrin β7 which we have previously demonstrated is upregulated by CB₂R activation¹¹. CD103 binding to epithelial-expressed E-cadherin promoting the mucosal retention, provides long-lasting, localised immune surveillance⁴⁸. We found a significant decrease in CD103⁺ CD4⁺ T cells, as well as a trend towards a decrease in CD103⁺ CD69⁺ CD4⁺ T cells with CB₂R inverse agonism. These results suggest that CB₂R inverse agonism may not only reduce the number of T cells trafficking to the gut, but also specifically targeting those that could become intestinal-associated tissue-resident memory T cells. This represent a novel therapeutic target for the management and treatment of IBD, given that proinflammatory tissue-resident memory T cells, marked by CD69 and CD103, are elevated in the intestinal mucosa of IBD patients, contributing to the chronic inflammation observed in IBD^49–51. CB₂R activation also increased the levels of CD45RO⁺ CD69⁺ and CD45RO⁺ CD95⁺ CD4⁺ T cells ex vivo. Increased levels of these cells is associated with disease flares and a more complicated disease course in patients^{52, 53}.

Several limitations of this study must be acknowledged. Much of the metabolic work was performed in Jurkat T cells, and although these provide a tractable system for mechanistic study, they do not fully reflect the heterogeneity of primary human intestinal T cells. In addition, the link between the observed metabolic programme and bona fide memory-cell differentiation remains partly inferential. Future work will validate the mitochondrial and PPP findings at the protein and flux levels, directly measure ROS, and determine whether CB₂R signalling promotes durable memory formation in primary T-cell systems ex vivo and in vivo.

Overall, these data identify CB₂R as a regulator of T-cell metabolic organisation, linking cannabinoid signalling to glucose uptake, mitochondrial function, PPP activity and memory-associated phenotypes. These findings suggest that while CB₂R activation may suppress effector T cell proliferation in the short term, it does so by supporting the shift to long-lived, gut-associated T-cell states that could exacerbate chronic intestinal inflammation. This provides a mechanistic framework for understanding why cannabinoid use may fail to improve IBD despite symptomatic benefit, and supports further investigation of CB₂R blockade as a potential novel strategy to limit pathogenic T-cell persistence in chronic intestinal disease.