In this issue of Blood, Subburaj et al report on results from using metabolomics (the science of studying small molecules known as metabolites) to measure the metabolites in biofluids, cells, and tissues. They longitudinally followed 222 pediatric and adolescent patients from 27 international transplantation centers as part of the prospective Predictive Biomarkers for Pediatric Chronic Graft-Versus-Host Disease (ABLE-cGVHD) biomarker trial.1 Risk-assignment analysis of centrally stored and analyzed plasma samples on days +100 to 114 after allogeneic hematopoietic cell transplantation (HCT) and onset analysis at diagnosis of chronic graft-versus-host disease (cGVHD) showed that plasma α-ketoglutaric acid was consistently elevated both before and at onset of cGVHD in multiregression analysis. Furthermore, metabolomic profiling allowed late acute GVHD (aGVHD) to be distinguished from cGVHD.
cGVHD is still a serious complication after allogeneic HCT and is the leading cause of nonrelapse mortality after transplantation.2 Objective diagnostic and staging tools for early cGVHD are needed because patients who meet the National Institutes of Health (NIH) diagnostic criteria based on clinical symptoms often have significant morbidity and possibly irreversible organ damage at diagnosis. A biomarker that is easy to assess would advance clinical care by speeding up diagnostic evaluation, providing additional certainty of the diagnosis, and improving the quality of clinical data for patients in clinical trials. A good diagnostic biomarker for distinguishing cGVHD from other diseases should reflect cGVHD activity and be noninvasive, accurate, simple, rapid, inexpensive, and standardized. A number of serum and plasma candidate proteins, including CXCL9, CXCL10, and serum BAFF and cells such as CD19+CD21low B-cell, T-cell, and natural killer (NK)-cell subsets have been explored as diagnostic biomarkers of cGVHD. However, all of them lack NIH-defined qualification and validation.3,4 The ABLE consortium recently advanced the field by presenting immune profile differences between cGVHD and late aGVHD by comprehensively analyzing 71 T-cell, B-cell, NK-cell, dendritic-cell, and myeloid immune–cell populations and 12 cGVHD plasma markers in a pediatric patient cohort.5 Subburaj et al continued their excellent efforts by using mass spectrometry–based metabolomic analysis, which allows identification and quantification of up to 150 endogenous metabolites. By using the plasma samples from days +100 to +114 after HCT, the consortium identified a unique metabolic pattern for late aGVHD compared with cGVHD. These data confirm the biological differences between these 2 entities that were originally defined by the NIH consensus group based on clinical features.2
Interestingly, plasma alpha-ketoglutaric acid, a key intermediary metabolite of the mitochondrial tricarboxylic acid cycle and a product of glutaminolysis, emerged as the single most significant elevated metabolite both before and at the onset of cGVHD in children and adolescents. Glutamine has been considered a crucial source of energy and macromolecule production in activated T cells, and glutaminolysis has been shown to play a central role in metabolic reprogramming of T cells during their activation, which could also be the case in the activation of GVHD-inducing alloreactive T cells.6 In addition, alpha-ketoglutaric acid kynurenine, a product of the tryptophan catabolism known to regulate T-cell proliferation and survival, was increased in the cGVHD cohort.7 The tryptophan catabolites kynurenine, 3-hydroxykynurenine, and 3-hydroxyanthranilic acid are potent inhibitors of T-cell activation and induce T-cell apoptosis.8 Thus, a strong activation of the kynurenin pathway in cGVHD could be an indicator of an anti-inflammatory response. It would be of interest to investigate metabolomics longitudinally and to include patients who respond to immunosuppressive treatments.
Although patients with either active aGVHD or cGVHD before day +114 after HCT were excluded, it is tempting to speculate that individuals with significantly elevated plasma alpha-ketoglutaric acid in the risk-assignment analysis performed between days +100 and +114 already had ongoing inflammation and immune dysregulation, which lead to clinical symptoms of cGVHD. Therefore, the question arises of whether plasma alpha-ketoglutaric acid could be an indicator for early cGVHD before clinical symptoms appear. It is unclear whether plasma α-ketoglutaric acid, even when adequately validated, could serve as an NIH-defined prognostic biomarker,4 because the window for therapeutic intervention that would have an impact on patients’ outcome is brief, and other pathways involved in alloreactive proliferation and disturbance of immune reconstitution leading to cGVHD have already been activated.
The study by Subburaj represents the largest prospective, well-characterized pediatric cohort with cGVHD and late aGVHD to date. Additional strengths of their study are a multicenter design and prospective, calendar-driven, serial sample collection with central storage and metabolomic analysis, comparison with age and time-matched controls without cGVHD, and thorough close clinical adjudication of GVHD using the 2005 NIH consensus criteria for cGVHD.2 Despite a relatively high number of patients, however, different subgroups such as de novo and progressive onset types and pulmonary cGVHD remain too small to allow for valid comparisons. Future studies should consider the heterogeneity of cGVHD with individual organ involvement and overall disease phenotypes, including inflammatory, fibrotic, and immune dysfunction features, and they should enroll well-characterized patients accordingly.
Subburaj et al used targeted metabolomic analyses that quantitatively measure the concentrations of a predefined set of metabolites that were previously identified. The data presented here support a prospective, longitudinal validation study in an independent pediatric patient cohort with the goal of combining potential metabolic biomarker panels with immune cellular and cytokine markers to obtain a biomarker algorithm for accurate risk assignment before onset and diagnosis of cGVHD. The authors must be congratulated for embarking on this important and very large project and should consider further collaborations, including adding an adult patient cohort.
Because of the inherent sensitivity of metabolomics, subtle alterations in biological pathways can be detected and can provide novel insights into pathophysiologic mechanisms of diseases. Thus, metabolomics and innovative developments in informatics and analytical technologies should be investigated not only for biomarker discovery but also for increasing our understanding of cGVHD clinical phenotypes and their association with immunoregulatory pathways relevant in studying the development and persistence of cGVHD. This should allow for a more individualized treatment approach in the future, as recently suggested by the NIH consensus group.9 Furthermore, metabolism is an attractive target for therapeutic intervention in GVHD because differentiation, proliferation, and function of innate immune cells are also subjected to metabolism-dependent regulation.6,10 Targeting metabolism for therapy of GVHD, however, will require a thorough understanding of the unique metabolic properties and programs of the multiple cellular components involved in cGVHD.
Conflict-of-interest disclosure: H.T.G. received honoraria from speaker’s bureaus and participated in advisory boards for Amgen, Celgene, Novartis, Sanofi, and Therakos.