From the inside: GVHD and glucose metabolism

In this issue of Blood, Assmann et al show that elevated glycolysis is a traceable metabolic feature of infiltrating T cells, particularly in the early phase of graft-versus-host disease (GVHD).¹  This observation verifies that noninvasive imaging of the metabolic activity of T cells can visualize dynamic inflammatory processes in patients after allogeneic hematopoietic stem cell transplantation (HSCT). Over the past 10 to 20 years, several new approaches, either tested in prospective randomized studies (eg, studies aiming to optimize the conditioning therapy prior transplantation) or evaluated by analyzing large clinical data sets (eg, retrospective studies exploring the impact of better HLA-matching between donor and patients) have led to improved survival outcomes in patients treated with HSCT.^2-4  In addition to these clinical data, experimental research has helped to elucidate the underlying immunological processes. In particular, the better understanding of T-cell biology has recently led to the successful development of novel treatment strategies for the most relevant complication after allogeneic HSCT, GVHD.⁵  Despite these advances, GVHD still affects 30% to 60% of patients after allogeneic HSCT.⁶  Although the diagnosis of GVHD of the skin can be easily established visually in most cases, GVHD of the intestine or liver is still defined by surrogate parameters like stool volume or increased bilirubin, which are nonspecific and have limited value in establishing the diagnosis of GVHD, or prediction of response to treatment or outcome.^7,8 

By visualizing enhanced glycolysis of alloreactive T cells by metabolic magnetic resonance imaging (MRI), Assmann et al were able to detect the onset of GVHD even before clinical manifestation in a chronic GVHD animal model. Further detailed analysis of the experimental GVHD model and data from single-cell sequencing of patient-derived circulating T cells corroborated their hypothesis that increased glycolysis is a feature in the early activation and organ infiltration (liver) of alloreactive T cells causing GVHD. Interestingly, peak metabolic activity (as measured by metabolic MRI using hyperpolarized pyruvate) preceded the clinical symptoms of GVHD. At later time points during the course of GVHD, the differences compared with syngeneic controls were less prominent, probably due to the limited target region analyzed and/or due diminished T-cell activity.

Today, in vivo imaging of metabolic activity using positron emission tomography (PET) with fluorodeoxyglucose (FDG) or other radiopharmaceuticals is a standard method evaluating malignant and nonmalignant diseases. We and others have employed PET for translational studies in GVHD, where enhanced glycolysis of the gut was established as an excellent preclinical and clinical marker for T-cell activity in intestinal GVHD.^9,10  The data presented in this issue of Blood clearly verify these previous observations, using a novel imaging technology (metabolic MRI), and extend these to metabolic T-cell activity in the liver in GVHD.

Although both FDG-PET and pyruvate-lactate metabolic MRI assess the glycolytic metabolic activity in vivo, differences exist with respect to the translational potential and availability. PET, which is similar to hyperpolarized MRI based on the systemic injection of a tracer molecule, allows for whole-body imaging of glycolysis and is not restricted to a single organ such as the liver. This is, so far, not established for metabolic MRI, given the short half-life of the hyperpolarized tracer, the fast metabolic step measured, and the resulting limited scan window.

In PET, coregistration with morphological and functional imaging allows for precise anatomical localization enabling detailed and quantitative assessment of glycolysis in all organs affected by GVHD in mice and humans. To this end, PET-computed tomography (CT) is clinically available in all cancer centers and could therefore be applied directly in translational studies. However, this is not yet established for hyperpolarized MRI.

On the other hand, as correctly mentioned by Assmann et al, PET does come with radiation exposure, whereas metabolic MRI does not, which might limit serial investigations in patients with GVHD. However, radiation doses resulting from PET-CT are diagnostic and rather low, and therefore only minimally add to therapeutic doses, in the frame of allogenic HSCT algorithms. Novel approaches, such as FDG-PET-MRI, further decrease radiation exposure and, at the same time, also provide additional and independent MRI-based evidence of intestinal GVHD (see figure).

The study by Assmann et al provides compelling evidence that in vivo molecular imaging of T-cell–mediated immune reactions, particularly by employing methods detecting metabolic activity of effector cells, is of value to detect the early onset of GVHD. This observation is likely to prove important in the setting of allo-activation of T cells after allogeneic HSCT and could also be useful in assessing immune responses in other cellular therapies. Moreover, noninvasive measurement of GVHD activity might be more informative and reliable compared with the classical clinical assessment of GVHD in evaluating the effectiveness of novel compounds for the treatment of GVHD.