Glycolytic metabolism of pathogenic T cells enables early detection of GVHD by 13C-MRI

Abstract

Graft-versus-host disease (GVHD) is a prominent barrier to allogeneic hematopoietic stem cell transplantation (AHSCT). Definitive diagnosis of GVHD is invasive, and biopsies of involved tissues pose a high risk of bleeding and infection. T cells are central to GVHD pathogenesis, and our previous studies in a chronic GVHD mouse model showed that alloreactive CD4+ T cells traffic to the target organs ahead of overt symptoms. Because increased glycolysis is an early feature of T-cell activation, we hypothesized that in vivo metabolic imaging of glycolysis would allow noninvasive detection of liver GVHD as activated CD4+ T cells traffic into the organ. Indeed, hyperpolarized 13C-pyruvate magnetic resonance imaging detected high rates of conversion of pyruvate to lactate in the liver ahead of animals becoming symptomatic, but not during subsequent overt chronic GVHD. Concomitantly, CD4+ T effector memory cells, the predominant pathogenic CD4+ T-cell subset, were confirmed to be highly glycolytic by transcriptomic, protein, metabolite, and ex vivo metabolic activity analyses. Preliminary data from single-cell sequencing of circulating T cells in patients undergoing AHSCT also suggested that increased glycolysis may be a feature of incipient acute GVHD. Metabolic imaging is being increasingly used in the clinic and may be useful in the post-AHSCT setting for noninvasive early detection of GVHD.

Trial registration: ClinicalTrials.gov NCT00520130.

Figures

Graphical abstract

Figure 1.

Mass spectrometry screen of intracellular metabolites in CD4+ T-cell subsets early in cGVHD reveals increased aerobic glycolysis in allogeneic Tem cells. Representative example of the clinical score (A) and body weight (B) changes over time in the B10.D2 into BALB/c cGVHD model, n = 5 (syngeneic [syn]), n = 16 (allogeneic [allo]). (C) Schematic indicating the potential fate of pyruvate by being converted to either lactate (anaerobic/aerobic glycolysis), acetyl-CoA (TCA cycle), or alanine (transamination). (D-G) Intracellular concentrations of metabolites: lactate (D), citric acid (E), malic acid (F), and alanine (G). Single-cell suspensions were generated from pooled, freshly harvested spleens of syn (n = 19) and allo (n = 22) HSCT recipients on day 14. The cells underwent positive selection using CD4 microbeads. FACS-purified T-cell subsets were collected, and the enzyme activity was quenched with methanol. All samples were supplemented with an internal standard solution, and relative quantification was carried out using capillary electrophoresis time-of-flight mass spectrometry. Cells were pooled from 2 independent HSCTs. Each sample for MS analysis contained ∼3 million cells; n = 1 for syn/allo Tn samples, n = 2 for syngeneic Tem, and n = 3 for allogeneic Tem. Data are represented as mean + SEM, as appropriate. PDH, pyruvate dehydrogenase.

Figure 2.

CD4+ Tem cells exhibit a higher ECAR compared with CD4+ Tn or syngeneic CD4+ Tem in cGVHD. (A) Single-cell suspensions were generated from pooled harvested spleens of syngeneic (n = 5 mice) and allogeneic (n = 9 mice) HSCT recipients on day 14. The cells underwent positive selection using CD4 magnetic microbeads, followed by FACS. FACS-purified T cells were plated using CellTak at a seeding density of 100 000 live sorted cells per well and treated with oligomycin (1 µM), carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP; 0.5 µM), and antimycin A (AA)/Rotenone (1 µM) at indicated time points. The number of technical replicates equals n = 8 to 32 wells per subtype; data are representative of 2 independent experiments. (B) Basal glycolysis vs basal respiration. (C) Basal glycolysis based on the mean of the 2 time points before oligomycin injection. (D) Glycolytic capacity measured after FCCP addition (mean of 4 time points). (E) Basal respiration rate based on the mean of the 3 time points before oligomycin injection. All data are represented as mean + SEM, statistical testing using a 2-way ANOVA, *P < .05, **P < .01, ***P < .001. OCR, oxygen consumption rate.

Figure 3.

RNAseq of allogeneic CD4+ Tn and Tem cells indicates an overall upregulation of glycolytic enzymes in alloreactive Tem cells. (A) FACS-purified allogeneic Tem and Tn cells from the spleen (>0.5 million cells per sample) were pooled from multiple study animals on day 14. RNA was extracted from pooled samples of each cell type (n = 1 for Tn, n = 3 for Tem), and the library was generated using the HyperPrep RNA-Seq Kit. All samples were sequenced on a HiSeq4000 (Illumina), and the reads were trimmed, mapped to the reference genome, and normalized to the library size as counts per million. The shaded squares indicate an increased (red) or decreased (blue) expression in Tem cells over Tn cells, and log2 fold-change values are indicated in panel B. (C) CD4+ T cells were isolated from the spleen of syngeneic and allogeneic animals on day 14 posttransplant and analyzed via flow cytometry to analyze the protein expression of GLUT1, HK2, and GAPDH in naive CD4+ and Tem cells. The ratio of the median fluorescence intensity of Tem to Tn cells is displayed as mean ± SEM; n = 7 (syn)/10 (allo) animals; Welch’s ANOVA test with Dunnett’s T3 multiple comparison test. (D) Gene expression module score of glycolysis genes and (E) TCA cycle genes in a single-cell sequencing data set of CD4+ cells isolated from the liver of syngeneic and allogeneic mice on day 14 posttransplant quantifying the up- or downregulation of a predefined set of genes (see also supplemental Table 3). Data were generated from 3 separately processed biological replicates for each condition that were pooled for the bioinformatic analysis. Two-way ANOVA with Sidak’s multiple comparison test; bar graphs represent mean ± SEM. (F) Quantification of the protein expression of GLUT1, HK2, and GAPDH in T-cell subsets isolated from the liver on day 14 as described in panel C. Aco2, aconitase 2; Aldoa, fructose-bisphosphate aldolase A; Cs, citrate synthase; Eno1, enolase 1; Fh, fumarate hydratase; Gpi1, glucose phosphate isomerase 1; Gpt, glutamate pyruvate transaminase; Hk1, hexokinase 1; Idh, isocitrate dehydrogenase 1; Ldha, lactose dehydrogenase A; Mdh, malate dehydrogenase; MFI, mean fluorescence intensity; Ogdh, oxoglutarate dehydrogenase; Pck2, phosphoenolpyruvate carboxykinase 2; Pdk1, pyruvate dehydrogenase kinase 1; Pdp1, pyruvate dehydrogenase phosphatase 1; Pfkp, phosphofructokinase; Pgk1, phosphoglycerate kinase 1; Pgm1, phosphoglucomutase1; Slc2a1, solute carrier family 2 member 1 (GLUT1); Slc2a3, solute carrier family 2 member 3 (GLUT3); Slc16a1, solute carrier family member 16 (MCT1); Slc16a3, solute carrier family 2 member 1 (MCT4); Sucla, succinyl-CoA ligase; Sudh, succinate dehydrogenase. *P < .05, **P < .01, ***P < .001.

Figure 4.

Hyperpolarized 13C-pyruvate in vivo MRI performed over the postsyngeneic and allogeneic HSCT course. [1-13C]Pyruvic acid (30 μL), containing 15 mM OX063 and 2.5 mM gadolinium chelate, was hyperpolarized using the Hypersense DNP Polarizer (Oxford Instruments). After hyperpolarization was achieved, the sample was dissolved in 4.5 mL of heated alkaline buffer, ie, 40 mM 4-(2-hydroxyethyl)-1 piperazineethanesulfonic acid, 30 mM of NaCl, and 100 mg/L of EDTA. The hyperpolarized [1-13C]pyruvate solution (96 mM) was then administered through an IV tail vein catheter. 13C MRI studies were performed on a 3-T MR Solutions MRI using a custom 13C-1H saddle coil for spectroscopic and anatomical imaging. (A) Schematic representation of the pyruvate-to-lactate conversion with the hyperpolarized 13C indicated in red. (B) Representative coronal anatomical MRI images overlaid with the 13C MRSI spectra for each voxel and an exemplary 13C spectrum indicating the lactate, alanine, and pyruvate peaks shown for a syngeneic (top) and allogeneic (bottom) animal at day 14. (C) MRSI images of syn (top) and allo (bottom) at day 14, indicating the signal intensity for lactate, pyruvate, as well as the combined lactate/pyruvate ratio. A region of interest was drawn around the liver based on the anatomical MRI, which was then used to quantify the signal intensity for each peak. (D) Lactate/pyruvate signal intensity in the liver over time after HSCT in syngeneic and allogeneic animals, n = 3 to 6 animals per group and time point, **P < .01 in a mixed-effect analysis with Sidak’s multiple comparison test. (E) Mouse plasma was collected via mandibular bleed at indicated time points in sodium fluoride/EDTA–coated microtubes. After centrifugation, an internal standard was added, and the lactate levels were quantified using ultra performance liquid chromatography–mass spectrometry, n = 4 to 17 per group and time point, **P < .01 in a mixed-effect analysis with Sidak’s multiple comparison test. All data are shown as mean ± SEM. ALT, alanine aminotransferase.

Figure 5.

Single-cell RNAseq of AHSCT patient-derived PBMCs. (A) Diagram depicting sample selection. PBMCs of 2 patients, one of which developed GVHD shortly after sample collection (pre-GVHD) and one who did not (no GVHD), were thawed; red blood cells were lysed, and cell viability was assessed. Single-cell preparation was performed using the Chromium Next GEM Single Cell 5′ Library & Gel Bead Kit (10× Genomics). Sequenced reads were aligned to the human GRCh38 reference sequence provided by 10× Genomics. Clustering and visualization were performed in R using the Seurat package (v3.1.1) with integrated datasets. (B) Volcano plot of the differential gene expression within 1 cluster comparing the no GVHD vs pre-GVHD sample for the CD4 cluster (left) and T_proliferating cluster (right). Genes with a fold change >1.5 and adjusted P value <.01 are highlighted in red. (C) Dot plot visualizing the scaled gene expression level of key glycolysis enzymes for the CD4, CD8, and T_proliferating cluster. (D) Comparison of the module scores for Glycolysis (left) and TCA (right) between both patients for CD4, CD8, and proliferating T cells. The analyzed cell numbers for each cluster are indicated in parentheses. Data displayed as mean ± SEM. FC, fold change; GI, gastrointestinal; MUD, matched unrelated donor; NS, not significant; PBHSCT, peripheral blood hematopoietic stem cell transplantation.