The study was conducted between October 2023 and May 2025 at seven hospitals in northern China and included neonates born at 35 weeks’ gestation or later during their first 28 days after birth. Cases were defined using an hour-specific bilirubin nomogram, with a total serum bilirubin concentration above the 95th percentile for postnatal age. Their elevated bilirubin was predominantly unconjugated, and routine etiological evaluation had not identified a common clinical cause. The investigators excluded infants with hemolytic disease, infection, polycythemia, clinically significant inadequate intake or delayed meconium passage, as well as those with previous allogeneic blood transfusion, congenital malformations, a known family history of hereditary disease or a clinically suspected or confirmed genetic disorder. The results therefore concern a specifically defined group with unexplained hyperbilirubinemia rather than neonatal jaundice of every cause.

Control neonates were enrolled at birth from six of the participating hospitals during the same period. They had not been diagnosed with hyperbilirubinemia on the day of enrollment, had normal Apgar scores and no history of asphyxia, and were subject to the same general exclusions concerning transfusion, congenital malformations, family history and suspected genetic disorders. Importantly, controls were not required to remain free of hyperbilirubinemia throughout the neonatal period. A telephone follow-up was conducted only for controls with detected target variants, and those observations were analyzed descriptively without reclassifying control status. The primary comparison was thus between hospitalized cases and newborns without a diagnosis at enrollment, rather than between cases and a uniformly verified unaffected group.

DNA analysis targeted 12 predefined hotspot loci in five genes involved in bilirubin production, transport or metabolism: UGT1A1, G6PD, SLC25A13, SLC10A1 and ABCC2. The variants were screened simultaneously by multicolor melting curve analysis. Genotypes were assigned by comparing each sample’s melting curves with known wild-type, heterozygous and homozygous controls, and approximately 10% of samples were randomly selected for validation by Sanger sequencing. The genotyping results showed 100% concordance with sequencing, and all enrolled neonates had successful calls for the 12 loci, with no assay failures reported. Because the panel covered only selected hotspots rather than the complete genetic landscape, finding no targeted variant did not establish that an infant lacked other genetic factors. Detection also did not imply that an allele was pathogenic or likely pathogenic.

UGT1A1 variants were detected in 60.3% of cases and 40.8% of controls, yielding an odds ratio of 2.20, a 95% confidence interval of 1.45–3.33 and a P value below 0.001. The UGT1A1 association remained significant after the study’s multiple-testing correction, with a false-discovery-rate-adjusted P value of 0.00063. UGT1A1 encodes the enzyme that catalyzes conjugation of unconjugated bilirubin with glucuronic acid, forming water-soluble conjugated bilirubin. Within this gene, c.211G>A, also known as UGT1A1*6, was the most frequently observed targeted locus. At least one copy of the variant was present in 53.3% of cases, compared with 37.5% of controls; the odds ratio was 1.90, with a 95% confidence interval of 1.25–2.89 and a P value of 0.002. This carrier association also remained significant after false-discovery-rate correction.

Genotype-specific comparisons produced higher odds ratios among homozygotes, although the confidence intervals showed substantial uncertainty. The heterozygous UGT1A1 c.211G>A genotype occurred in 40.7% of cases and 32.5% of controls, giving an odds ratio of 1.67 and a 95% confidence interval of 1.08–2.60. The homozygous genotype occurred in 12.6% of cases and 5.0% of controls, with an odds ratio of 3.37 and a wider 95% confidence interval of 1.39–8.17. Both genotype-specific comparisons met the study’s false-discovery-rate-adjusted significance threshold, with adjusted P values of 0.0413 and 0.0193, respectively. The pattern is compatible with a possible dose-related influence at this locus, but the case-control design cannot establish that c.211G>A directly causes hyperbilirubinemia, exclude effects from other linked genetic factors or predict an individual infant’s outcome.

Variants in the other four tested genes were less common and did not differ significantly between cases and controls. SLC10A1 variants were detected in 3.9% of cases and 5.0% of controls, with an odds ratio of 0.77 and a 95% confidence interval of 0.29–2.00. Seven cases carried G6PD variants, while no controls did, but the difference was not statistically significant. One case carried the tested SLC25A13 deletion, and no ABCC2 variant was detected in either group. The cohort was characterized predominantly by unconjugated hyperbilirubinemia, whereas SLC25A13 deficiency-associated neonatal intrahepatic cholestasis and ABCC2-associated Dubin-Johnson syndrome typically involve intrahepatic cholestasis dominated by conjugated bilirubin. That biochemical pattern helps explain why variants in these pathways were uncommon in the selected cohort, but the study cannot determine their broader prevalence among northern Chinese newborns.

Several infants carried more than one targeted variant, but the multi-site detection rate did not differ significantly between groups: 8.2% of cases and 4.2% of controls had variants at multiple sites. All 25 neonates with multiple variants within one gene had combinations involving UGT1A1. The most frequent case-group combination was UGT1A1 c.211G>A with c.1091C>T, detected in 14 cases. Other combinations included c.211G>A with SLC10A1 c.800C>T, UGT1A1 c.1456T>G or G6PD c.1388G>A. Fourteen neonates carried variants in more than one gene. Three male cases with hemizygous G6PD c.1388G>A also carried heterozygous UGT1A1 c.211G>A, while seven cases carried both UGT1A1 c.211G>A and SLC10A1 c.800C>T. These findings document coexistence among variants affecting glucuronidation, hepatic uptake and red blood cell enzyme function, but they do not establish additive effects, biological interactions or a combined clinical risk.

Exploratory follow-up added evidence that some initially classified controls later developed hyperbilirubinemia. Among the 54 controls with detected target variants, 32, or 59.3%, subsequently developed the condition. This included 21 of 39 with a single heterozygous UGT1A1 c.211G>A variant and four of six with the homozygous variant. Other observed outcomes included two of five with heterozygous UGT1A1 c.1091C>T, one of two with heterozygous UGT1A1 c.1456T>G, none of one with heterozygous UGT1A1 c.686C>A and two of six with heterozygous SLC10A1 c.800C>T. Three of the five controls with multi-site variants also developed hyperbilirubinemia. These follow-up data were descriptive, were available only for variant-positive controls and were not used to make inferential comparisons, so they require cautious interpretation.

Sensitivity analyses restricted the sample to the six hospitals that enrolled both cases and controls. Multivariable logistic regression models adjusted for sex and recruitment center yielded effect estimates similar to those from the primary analyses for overall target-variant detection, UGT1A1 variants and carriage of UGT1A1 c.211G>A. Nevertheless, the study has important limitations. Detailed quantitative clinical phenotypes were not collected consistently across centers, limiting genotype–phenotype analysis, while the absence of family samples and haplotype analysis prevented the investigators from determining the phase of multiple UGT1A1 variants or evaluating possible compound heterozygous effects. The relatively small control group, targeted assay, uneven follow-up and possible selection bias also limit interpretation. Population-structure effects and residual confounding cannot be completely excluded, and extrapolation to other populations should be cautious. Overall, the study provides a regional description of selected bilirubin-metabolism variants among hospitalized northern Chinese neonates with unexplained hyperbilirubinemia, with the clearest statistical association at UGT1A1 c.211G>A rather than evidence that variants across all five genes are equally implicated.

Subject of Research: Targeted genetic variants associated with unexplained neonatal hyperbilirubinemia in northern Chinese neonates

Article Title: Bilirubin metabolism-related genetic variants in neonatal hyperbilirubinemia: a multicenter case-control study in northern China

Article References: Sheng, S., Li, Y., Ma, L., Chang, H., Guan, Y., Lei, Y., Tan, C., Arigonggaowa, Wen, H., Kou, C., & Wang, Y. (2026). Bilirubin metabolism-related genetic variants in neonatal hyperbilirubinemia: a multicenter case-control study in northern China. Frontiers in Pediatrics, 14, Article 1945311. https://doi.org/10.3389/fped.2026.1945311

Image Credits: AI Generated

DOI: 10.3389/fped.2026.1945311

Keywords: neonatal hyperbilirubinemia, UGT1A1, genetic variants, bilirubin metabolism, northern China, newborn jaundice, case-control study, G6PD, SLC10A1, multicenter research, Bilirubin, metabolism-related

Glucose-6-phosphate dehydrogenase, commonly abbreviated G6PD, is an enzyme that protects red blood cells from oxidative damage. When the enzyme is deficient, the cells become vulnerable to destruction under oxidative stress, predisposing affected individuals to hemolytic crises. The condition is the most common enzymatic disorder of red blood cells worldwide, affecting more than 400 million people, with particularly high prevalence in the Mediterranean basin, the Middle East, sub-Saharan Africa, and parts of Asia. In newborns, G6PD deficiency is consistently recognized as one of the leading causes of severe hyperbilirubinemia and kernicterus, a devastating form of bilirubin-induced brain injury that can cause sensorineural hearing loss and long-term neurodevelopmental impairment.

What makes neonatal jaundice in G6PD-deficient infants particularly insidious is that it does not always arise from acute hemolysis. Although oxidative triggers can abruptly increase bilirubin production, many affected newborns develop jaundice through a more gradual imbalance between bilirubin production and conjugation, with low-grade hemolysis compounded by impaired bilirubin clearance. This means that clinicians cannot rely on classic laboratory signs of hemolysis to identify at-risk babies, and the jaundice may not appear until bilirubin levels have already climbed to dangerous territory.

Current clinical guidelines reflect a selective approach to testing. The 2022 American Academy of Pediatrics guideline recommends G6PD testing in cases of unexplained jaundice or hemolysis, particularly in high-risk populations. The World Health Organization similarly recommends screening only in at-risk populations where more than 3 to 5 percent of males are affected. But these targeted strategies, based on reported ethnicity, family history, or clinical suspicion, may have limited sensitivity. Prior studies from diverse populations have indicated that a substantial proportion of G6PD-deficient neonates are missed by targeted screening protocols, and there is particular concern that female heterozygotes with intermediate enzyme activity may be overlooked despite being at risk for hemolysis and severe jaundice.

Israel presents a uniquely challenging context for ethnicity-based screening. The population includes Ashkenazi and Sephardic Jews, Arab and Druze communities, and increasing rates of interethnic marriage. Earlier studies demonstrated a highly variable prevalence of G6PD deficiency across specific populations, ranging from less than 1 percent in Ashkenazi Jews up to 58 percent in Sephardic Jews of Kurdish descent. As ethnic admixture has increased, ancestry-based screening strategies have become progressively less reliable. Recognizing this, the Israeli Ministry of Health recommended in 2021 that universal G6PD deficiency screening be added to the newborn screening program, though this recommendation has not yet been fully implemented across all Israeli hospitals.

At Hadassah Medical Center, universal G6PD screening has been part of routine care for more than five years, but its clinical impact had never been formally evaluated. The research team, led by Aviad Schnapp and Ohad Neufeld, conducted a retrospective cohort study with a nested case-control analysis, reviewing the medical records of all neonates born between January 1 and June 30, 2024. The cohort included 5,413 newborns with gestational ages ranging from 32 to 42 weeks. Of these, 4,867 infants, or 89.9 percent, had a valid enzyme activity result. Screening was routinely performed on umbilical cord blood collected at birth, using a quantitative Motulsky enzymatic assay. Results were reported as international units per gram of hemoglobin within 24 hours of laboratory analysis.

The findings revealed that 81 neonates had severe G6PD deficiency, defined as enzyme activity below 2 IU/g Hb, representing 1.66 percent of all screened newborns and 2.9 percent of screened males. An additional 108 neonates had intermediate G6PD activity, between 2 and 10 IU/g Hb, corresponding to 2.2 percent of all screened newborns and 4.4 percent of screened females. Because the number of normal-screen infants substantially exceeded those with deficient or intermediate activity, the researchers used a nested case-control sampling strategy, randomly selecting 567 controls with normal enzyme activity in a 3:1 ratio to form a detailed comparison group of 756 newborns.

The clinical outcomes were striking. Phototherapy was required in 29.6 percent of neonates with severe deficiency, compared with 13.0 percent of those with intermediate activity and 10.1 percent with normal activity. After adjusting for known risk factors including prematurity, mode of delivery, presence of hematoma, ABO or Rh alloimmunization, and family history of conditions that increase hyperbilirubinemia risk, severe G6PD deficiency remained an independent predictor of phototherapy, with an adjusted odds ratio of 4.78. Notably, the time to reach the phototherapy threshold during the birth hospitalization did not differ between infants with and without G6PD deficiency, challenging the common clinical assumption that jaundice in G6PD deficiency evolves later than jaundice due to blood group alloimmunization.

Perhaps the most consequential finding relates to what targeted screening would have missed. Among neonates with G6PD deficiency, a substantial proportion of parents reported no family history of the condition, with 43.2 percent of those with complete deficiency and 53.7 percent of those with partial deficiency reporting no known family history. These infants would not have undergone enzyme testing under a targeted approach. Furthermore, many were not prescribed enhanced bilirubin surveillance during the first 24 hours of life. Six of 94 neonates who required phototherapy, or 6.4 percent, were identified through universal screening as having G6PD deficiency despite the absence of any reported family history. Identification enabled enhanced monitoring and earlier initiation of phototherapy in accordance with the lower treatment thresholds applied to G6PD-deficient neonates.

The biological basis of jaundice in G6PD-deficient newborns is complex. Prior work has shown that clinically important hyperbilirubinemia may occur with limited classic laboratory evidence of hemolysis, while reduced bilirubin conjugation, particularly when G6PD deficiency coexists with UGT1A1 variants associated with Gilbert syndrome, may increase risk. The findings are consistent with the current view that neonatal G6PD-associated jaundice reflects an imbalance between bilirubin production and conjugation rather than a single mechanism. The intermediate-activity group warrants careful interpretation: partial G6PD deficiency was not associated with the excess phototherapy use seen in infants with severe deficiency, though the study was not powered to establish equivalence between the intermediate and normal groups. Because genotyping was not performed, the phenotype-based activity bands cannot be directly aligned with the revised WHO classification, which combines former class II and III variants into a single category and recognizes that heterozygous females may still confer risk. The researchers emphasize that their findings should not be interpreted as excluding risk among infants with intermediate activity or heterozygosity, and that future genotype-informed studies are needed to refine risk stratification. Nationwide neonatal screening for G6PD deficiency is currently implemented in only a limited number of countries worldwide, and this study provides additional evidence from a distinct population and healthcare setting that universal screening identifies infants who would otherwise be missed by targeted approaches, potentially enabling earlier risk-informed management, follow-up planning, and parental education about avoiding oxidative triggers later in life.

Subject of Research: Universal newborn screening for G6PD deficiency and severe hyperbilirubinemia at a tertiary center in Jerusalem

Article Title: Universal newborn screening for G6PD deficiency and severe hyperbilirubinemia at a tertiary center in Jerusalem

Article References: Schnapp, A., Neufeld, O., Abu-Leil, S., Abu Omar, R., Azzam, H., Ofek Shlomai, N., & Eventov Friedman, S. (2026). Universal newborn screening for G6PD deficiency and severe hyperbilirubinemia at a tertiary center in Jerusalem. Journal of Perinatology. https://doi.org/10.1038/s41372-026-02901-9

Image Credits: AI Generated

DOI: 10.1038/s41372-026-02901-9

Keywords: Universal, newborn, screening, G6PD, deficiency, severe, hyperbilirubinemia, tertiary, center, Jerusalem, scientific research

The study, published in Cancer Immunology Research, a journal of the American Association for Cancer Research, tackles a question that immunologists have rarely been able to address directly in human tissue: does the way a T cell generates and uses energy determine whether it mounts an effective attack or slides into exhaustion? Exhaustion is a state in which a cell, activated too long and too continuously, stops responding to the threats it was designed to destroy. Checkpoint immunotherapy, such as antibodies against the PD-1 receptor, releases one of the brakes holding those cells back. For some patients the effect is dramatic; for many others it fails, and the reasons have remained stubbornly unclear.

Metabolism has long seemed a plausible place to look for answers. An immune cell’s capacity to act — to proliferate, to produce inflammatory signals, to kill tumor cells — is constrained by how it generates energy and building blocks. But probing that relationship in human tumors has been technically formidable. The standard laboratory method for measuring cellular metabolism requires a large, pure population of a single cell type, something that is nearly impossible to obtain from a biopsy. As first author Sujeetha A. Rajakumar, Ph.D., a former research associate in the Department of Otolaryngology–Head and Neck Surgery at Johns Hopkins Medicine, explains, what arrives from the clinic is a mixture, there is not much of it, and the cells researchers most want to study are often the rarest. Most of what the field knows about immune cell metabolism has therefore come from mouse tissue and cells grown in dishes, systems that do not always faithfully reflect what happens inside a patient.

The Johns Hopkins team circumvented this bottleneck by measuring each cell’s metabolic machinery alongside the surface markers that identify it. This approach allowed them to read three things from a single cell simultaneously: which cell type it is, what functional state it is in, and what metabolic profile it uses. They applied the method to several populations of tumor-infiltrating immune cells, including mucosal-associated invariant T (MAIT) cells, conventional CD8+ T cells, and innate lymphoid cells (ILCs). Because the technique works on scarce and mixed samples, it opened a window onto metabolic states that had previously been invisible in human tumor tissue.

The tumor samples came from patients enrolled in two phase II clinical trials at Johns Hopkins Medicine in which immunotherapy was administered before surgery. Participants received either the anti-PD-1 antibody nivolumab alone or nivolumab combined with an antibody against interleukin-8, a signaling protein that recruits immune cells to sites of inflammation and infection. Tissue was collected before immunotherapy began and again four weeks later at the time of surgery, giving the investigators a rare opportunity to observe how the immune and metabolic states of individual cells changed within the same patients over the course of treatment.

Three findings stood out. First, the team identified a population of T cells within these tumors that proved to be metabolically fitter and less exhausted than their neighbors, suggesting that not all tumor-infiltrating lymphocytes are equally depleted and that metabolic state tracks closely with function. Second, the more exhausted cells showed elevated levels of the enzyme glucose-6-phosphate dehydrogenase, or G6PD, the rate-limiting enzyme of a metabolic pathway those cells appear to rely on — making it a possible point of intervention. Third, a subset of innate lymphoid cells carried a metabolic profile associated with immune suppression, raising the possibility that these cells may actively pose an obstacle to checkpoint therapy rather than merely failing to help it.

To test whether the exhausted cells truly depend on the G6PD pathway, the researchers turned to laboratory cultures of tumor-infiltrating CD8+ T cells from a separate group of untreated patients. They sorted the cells according to their expression of CD39, a protein that can contribute to immunosuppression by limiting the activation of T cells. The cultures were then treated with a G6PD inhibitor, with an anti-PD-1 antibody, with both, or with neither. The combination treatment produced the largest gene expression changes in the most exhausted cells and increased their secretion of CD27, a marker of T-cell activation. In other words, the cells that seemed least reachable by conventional immunotherapy were the ones that changed the most when the metabolic pathway was blocked alongside checkpoint blockade.

Senior author Martin Alphonse, Ph.D., an assistant professor of dermatology at the Johns Hopkins University School of Medicine, frames the work around a deceptively simple question: take a piece of head and neck tumor tissue and you will find immune cells inside it whose sole purpose is to recognize something abnormal and destroy it — so what has happened to them? The answer emerging from this study is that part of what has happened is metabolic. Exhausted cells appear to have rewired their energy production in a way that can be mapped, and at least one node of that rewiring can be targeted in the laboratory to partially restore function. Rajakumar notes that when the pathway was blocked alongside checkpoint inhibitor treatment, the cells expected to be least reachable changed the most and became reactivated — evidence, she says, that the pathway is worth pursuing.

The clinical context makes the work consequential. Head and neck squamous cell carcinoma, which arises from the cells lining the mouth, throat and voicebox, is the seventh most common cancer worldwide, with roughly 890,000 new cases each year. With checkpoint inhibitors helping fewer than a quarter of these patients, any strategy that could expand the fraction of responders would address a major unmet need. Growing evidence has suggested that metabolic reprogramming influences whether a T cell mounts an effective response or becomes exhausted, but this study is among the relatively few to examine that link directly in human tumor tissue rather than in animal models or cell lines, and to do so across multiple immune cell types at single-cell resolution.

The researchers and their collaborators are careful to stress what the findings are and what they are not. Blocking G6PD in laboratory cultures reactivated exhausted cells, but no patient in the trials received a G6PD inhibitor, and the approach should not be considered a treatment itself. Alphonse emphasizes that these findings need confirmation in larger patient groups before anything can be built on them. The next steps, he says, are to reproduce the results in independent cohorts and to test whether interfering with this metabolic step has any effect in a living system, not just in a dish. Additional co-authors are Namya Nanda, Chloe Kim, Dustin Dikeman, Tanguy Y. Seiwert, Zubair Khan and Sewon Kang, all affiliated with Johns Hopkins Medicine. Study co-author Carole Fakhry, M.D., M.P.H., senior associate dean for clinical affairs at Johns Hopkins, underscores that such interdisciplinary, collaborative studies are critical to understanding how to make therapies more effective in the future, leveraging the translational capabilities of clinical trials to generate insights that laboratory work alone cannot provide. The work was supported by the Johns Hopkins Department of Dermatology, a Career Development Award from the Dermatology Foundation, the Department of Otolaryngology–Head and Neck Surgery, and the Bloomberg~Kimmel Institute for Cancer Immunotherapy. Alphonse and Rajakumar are inventors on two provisional patent applications related to the work, assigned to The Johns Hopkins University, which manages these relationships under its conflict-of-interest policies.

Subject of Research: Immunometabolic profiling of exhausted tumor-infiltrating T cells in head and neck squamous cell carcinoma

Article Title: Johns Hopkins investigators identify a metabolic weak point in exhausted immune cells in head and neck cancer

Article References: Johns Hopkins investigators identify a metabolic weak point in exhausted immune cells in head and neck cancer. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: head and neck cancer, T cell exhaustion, immunotherapy, checkpoint inhibitors, G6PD, tumor metabolism, CD8+ T cells, innate lymphoid cells, MAIT cells, nivolumab, single-cell profiling, Cancer Immunology Research

The distinction between healthy and harmful heart enlargement sits at the center of the work. Cardiac hypertrophy, the thickening of heart muscle, comes in two physiologically opposite flavors. Physiological hypertrophy, triggered by exercise or pregnancy, enlarges the heart while preserving or improving its contractile function, and it is accompanied by efficient energy metabolism, robust blood vessel growth, and molecular profiles dominated by fatty acid oxidation. Pathological hypertrophy, driven by chronic hypertension, valve disease, or heart attack, produces similar gross enlargement but with stiffening walls, impaired pumping, fibrosis, metabolic inflexibility, and a march toward heart failure. Decades of research have sought the molecular switch that separates these two trajectories, and the new findings place cytosolic NADPH firmly on the protective side of that divide.

Technically, the study hinged on the ability to see and manipulate NADPH in living systems, a long-standing challenge because NADPH and its oxidized counterpart NADP+ are structurally almost identical and interconvert rapidly. Wu and colleagues combined genetic models targeting glucose-6-phosphate dehydrogenase, or G6PD, the rate-limiting enzyme of the oxidative pentose phosphate pathway and the principal cytosolic source of NADPH, with readouts of cardiac structure, function, and metabolism. When exercise-induced increases in G6PD activity were blunted, the beneficial cardiac remodeling that normally follows training was lost, demonstrating that the pathway is not merely a byproduct of exercise but a necessary component of its cardiac benefits.

The protective payoff became most apparent in models of ischemia/reperfusion injury, the cellular catastrophe that occurs when blood supply to the heart is briefly cut off and then restored, as happens during a heart attack. Animals with elevated cytosolic NADPH, whether achieved through exercise or through the spermidine-derived compound, showed improved cardiac function and reduced injury after the insult. The team traced this protection to NADPH’s established biochemical roles: the molecule is the electron donor for thioredoxin and glutathione antioxidant systems that neutralize reactive oxygen species, the corrosive byproducts of reoxygenation that kill cardiomyocytes in the minutes and hours after blood flow returns. By keeping antioxidant capacity high, elevated NADPH effectively raises the heart’s threshold for reperfusion damage.

What makes the result especially compelling is the pharmacological angle. Spermidine itself has been linked in prior studies to cardioprotection and extended healthspan, prompting interest in its derivatives as drug candidates. Wu and colleagues identified a candidate spermidine derivative that raises cytosolic NADPH and reproduces key features of the exercise phenotype: physiological growth of the heart, enhanced antioxidant buffering, and resilience to ischemic injury, without the maladaptive gene expression patterns that characterize pathological hypertrophy. In effect, the compound simulates a metabolic signature of the trained heart in sedentary animals.

The concept of an exercise-mimicking pill has long hovered at the edge of cardiovascular research, often disappointing in translation because broad interventions that mimic one aspect of exercise tend to disrupt others. The NADPH-centered approach is narrower and, its proponents argue, more principled. Rather than attempting to replicate the whole-body storm of hormones, neural signals, and mechanical loading that exercise produces, the strategy targets a single downstream metabolite that sits at a convergence point for the cardiac benefits of training. NADPH is consumed in the synthesis of reduced glutathione, regenerated by G6PD and by malic enzyme and isocitrate dehydrogenase reactions, and shuttled across compartments by dedicated transporters, making its cytosolic pool a well-positioned control point for redox homeostasis.

The study also builds on a growing appreciation that NADPH metabolism is not uniform across cellular compartments. Mitochondrial NADPH, generated by transhydrogenase and other enzymes, has distinct roles in antioxidant defense and biosynthesis, while nuclear and cytosolic pools feed separate redox circuits. By focusing specifically on the cytosolic compartment and linking it to G6PD flux, the work offers a more precise target than earlier, cruder attempts to boost cellular reduction potential. Fluorescent biosensors that distinguish NADPH from NADH, developed in recent years, made the compartment-specific measurements feasible and are likely to accelerate follow-up studies in other tissues where NADPH balance governs immune function, fat metabolism, and aging.

Important caveats remain before the findings can inform human medicine. The work was conducted in animal models, and the dose, safety, and long-term consequences of chronically elevating cytosolic NADPH are unknown. Excessive NADPH generation has been implicated in other contexts in fueling proliferative signaling and in providing reducing equivalents to NADPH oxidases, which produce the very reactive oxygen species the heart must defend against. The authors and commentators, including Bryce J. Carpenter and Pieterjan Dierickx of the Max Planck Institute for Heart and Lung Research, who wrote an accompanying News and Views analysis, emphasize that the therapeutic window will need careful definition: the goal is the physiological range achieved by exercise, not an unbounded increase. Questions also remain about whether NADPH elevation alone can recapitulate the vascular and neuronal adaptations of exercise or whether it must be paired with other signals.

Even with those qualifications, the study reframes a central question in cardiovascular medicine. If the protective effects of exercise on the heart converge on a measurable metabolite, then patients confined to hospital beds, older adults with frailty, and the vast population that fails to meet physical activity guidelines may one day have access to a therapy that borrows the heart’s own training program. A candidate spermidine derivative that lifts cytosolic NADPH is an early but concrete step along that path, and it signals that the metabolic underpinnings of exercise’s benefits are finally becoming druggable.

Subject of Research: The role of cytosolic NADPH produced by the pentose phosphate pathway in mediating exercise-induced physiological cardiac hypertrophy and protection against ischemia/reperfusion injury.

Article Title: Increasing cytosolic NADPH to mimic exercise

Article References: Carpenter, B. J., & Dierickx, P. (2026). Increasing cytosolic NADPH to mimic exercise. Nature Metabolism, 8(9), 1827-1828. https://doi.org/10.1038/s42255-026-01580-2

Image Credits: AI Generated

DOI: 10.1038/s42255-026-01580-2

Keywords: NADPH, cardiac hypertrophy, exercise mimicry, pentose phosphate pathway, G6PD, spermidine, ischemia/reperfusion injury, heart failure, cardiac metabolism, antioxidant defense, Nature Metabolism, cardioprotection