Pyruvate kinase deficiency (PKD) is the most frequent congenital enzymatic defect of glycolysis, and one of the most common causes of hereditary non spherocytic hemolytic anemia (Zanella et al., 2007). The disease has diverse clinical presentations, with levels of hemolysis and the resultant anemia varying from patient to patient (Svidnicki et al., 2018; Enegela and Anjum, 2021). As such, clinical presentations range from mild and fully compensated anemia to more severe forms requiring repeated transfusion (Zanella et al., 2005). PKD is traditionally considered an autosomal recessive condition (Zanella et al., 2007). Due to the founder effect, PKD is more prevalent in specific groups including the Pennsylvania Amish and the Romani people (Rider et al., 2011).

A well-established hypothesis posits that common metabolic enzymopathies in humans, like PKD or glucose 6-phosphate dehydrogenase deficiency (Francis et al., 2020), may have arisen and were selected for as a result of evolutive pressure by endemic diseases – such as malaria (Ayi et al., 2008). By altering enzymatic activity at rate-limiting steps erythrocyte energy and redox metabolism, these mutations are deemed to confer resistance to parasitic infection at the erythrocytic stage (Ayi et al., 2008). Comprising over 80% of cells in the human body, red blood cells (RBCs) are involved in a transport system that brings oxygen and nutrients to the tissues (Nemkov et al., 2018a). Because of their ubiquity and movement throughout the body, they offer a unique lens into whole system metabolism (Nemkov et al., 2018a). Lacking a nucleus and organelles, red cells depend almost entirely on glycolysis for energy production. By constraining the rate of one of the two adenosine triphosphate (ATP)-generating steps in glycolysis, PKD dramatically alters RBC function, metabolism, and lifespan (Chapman and Schaumburg, 1967; Nathan et al., 1968; Delivoria-Papadopoulos et al., 1969). Specifically, pyruvate kinase is responsible for catalyzing the ATP producing conversion of phosphoenolpyruvate to pyruvate, a process responsible for the generation of 50% of all ATP in RBCs (Grace et al., 2015). Anemia in PKD patients is thought to occur as a result of reduced levels of ATP which are critical in maintaining processes required for cell function (Nathan et al., 1968), including fueling ion pumps, preserving membrane structural homeostasis and synthesizing key reducing equivalents such as glutathione (GSH; D’Alessandro et al., 2019). As ATP levels decrease, the energy-dependent sodium potassium ATPases slow, and potassium exits the cell membrane – a phenomenon that is particularly relevant in the context of common iatrogenic interventions, such as blood storage and transfusion (Yoshida et al., 2019). Alteration of sodium/potassium equilibria in the inner/extracellular compartment concomitantly impairs the function of ATP/sodium-dependent calcium pumps (Lew and Tiffert, 2017), triggering processes of erythrocyte-specific suicidal death (eryptosis; Nemkov et al., 2020a). Loss of ion homeostasis promotes water leaking out of the cell owing to the hypotonicity of the intracellular compartment, leading to dehydration, dysfunction, and removal of RBCs from the bloodstream (Van Wijk and Van Solinge, 2005). Vice versa, boosting PK activity in the context of hemoglobinopathies that alter RBC redox and energy metabolism (e.g., sickle cell disease) has been proposed as a viable strategy to prevent RBC lysis and untimely removal from the bloodstream (Rab et al., 2021).

Though the understanding of PKD has expanded in recent years, its full effects on RBC metabolism have not been fully described. Moreover, PKD is underdiagnosed in the general population, due in part to diagnostic gaps that arise from the disease’s heterogeneity (Bianchi et al., 2019). An understanding of the effects of this disorder on the metabolome may allow for better diagnosis, management, and treatment of this condition.

While previous reports focused on dried blood spot analysis of whole blood from PKD patients (Dooijeweert et al., 2021), in this study, we attempted to characterize the metabolome of RBCs from PKD patients. We hypothesized that PKD samples would show reprogramming of energy metabolism, with increases in early glycolytic metabolites and decreases in late glycolytic metabolites. We also predicted that PKD samples would display metabolic changes indicative of the systemic hypoxia that occurs in other anemias, but with less severe presentations due to the compensated nature of the disease.

In the present study, we performed a pilot metabolomics analysis of RBCs from individuals with PKD. Despite the pilot nature of the study, given the limited sample size of the cohort enrolled in the test population, we care to stress the sociomedical relevance of investigation aimed at serving underrepresented minorities, such as subjects with polymorphisms of PKD genes in the Amish population (four out of five PKD subjects enrolled in this study; Rider et al., 2011). As a result, we confirm that PKD mutation R479H – the only variant investigated here – is associated with the dysregulation of late RBC glycolysis, with severe depletion of pyruvate and lactate downstream to PKD, and accumulation of metabolites upstream. Accumulation of early glycolytic intermediates and depletion of reducing equivalents may be sufficient to fuel the PPP by law of mass action, as observed for example in the context of glycolytic enzyme inhibitions by oxidant stress (e.g., oxidation of cysteine 152, 156 and histidine 179 of glyceraldehyde 3-phosphate dehydrogenase; Reisz et al., 2016) or binding to the N-terminus cytosolic domain of band 3 (Issaian et al., 2021). Indeed, steady state levels of PPP metabolites were elevated in PKD patients. This is interesting in that adaptations in RBC metabolism of PKD patients seem to be opposite to those reported in G6PD deficient subjects (Tzounakas et al., 2016; Francis et al., 2020). This observation is relevant when considering that both adaptations are deemed to have evolved to counteract the spread of malaria in regions in which this parasitic infection is endemic (Ayi et al., 2008; Mohandas and Gallagher, 2008). Though this shift can reflect increases in oxidative stress and a greater demand for GSH, increased ATP levels and comparable levels (and ratios) of GSH/GSSG in PKD samples when compared to controls suggest otherwise. Instead, the increase in the PPP may reflect a compensatory response to the decrease in late glycolytic metabolites, as the PPP produces metabolites that can reenter late glycolysis. Alternatively, this may reflect an accumulation of early glycolytic products that increase PPP activity through mass-action effects.

Despite altered glycolysis, ATP levels were surprisingly found to be higher in PKD patients in this cohort. The data is inconsistent with standard ATP measurements in the literature and previous studies on whole blood from dried blood spots (Dooijeweert et al., 2021), despite the lack of iatrogenic interventions (i.e., recent transfusion events at the time of blood draws) in any of the subjects investigated here. The high degree of reticulocytosis that accompanies of these patients may have impacted measurements of carboxylic acids (also elevated in PKD samples from earlier dried blood spot studies; Dooijeweert et al., 2021) and ATP production through residual mitochondrial activity, despite the lack of significant differences in the levels of most carboxylates between the two groups with the exception of succinate. Accumulation of succinate in PKD patients is suggestive of a potential disruption in the activity of succinate dehydrogenase (SDH) at complex II in mitochondria in other cells than RBCs, as a function of PKD-induced anemia/hypoxia, similarly to what observed in the face of pathological hemorrhagic (D’Alessandro et al., 2017) or ischemic (Chouchani et al., 2014) hypoxia in previous studies. Previous studies have hypothesized that increased glutaminolysis in PKD reticulocytes may provide an alternative form of anaplerotic carbon, compensating in the tricarboxylic acid cycle (TCA) for decreases in pyruvate from glycolysis (Van Vuren et al., 2020), and leading to reduced levels of glutamine and GSH. Though our data describes both reticulocytes and mature red cells, it revealed elevated glutamine levels and, elevated levels of its catabolites through glutaminolysis, indeed implicating an upregulation of this pathway. Levels of GSH and glutamine catabolites of other pathways were not elevated when compared to controls, further supporting this theory.

Because catabolized branched chain amino acids (BCAAs) can enter the TCA as succinyl CoA, it is not surprising that trends in BCAA levels match those of early TCA metabolites in PKD patients. C3 and C5 carnitines are considered direct products of BCAA catabolism, and C3 carnitine appeared to follow the BCAA trends (Newgard et al., 2009; Lake et al., 2015). The increase in TCA metabolites may reflect a higher proportion of reticulocytes in PKD patients, especially if some of those patients underwent splenectomy.

In the previous paragraphs, we hypothesized that increased levels of PPP metabolites in PKD patients could be at least in part explained by increases in oxidant stress in these subjects. Of note, purine oxidation products and hypoxic markers (Nemkov et al., 2018b) hypoxanthine, xanthine, and allantoate were all increased in RBCs from PKD subjects. Interestingly, previous studies on RBC storage have suggested that lower levels of these metabolites indicated higher oxidative stress (Nemkov et al., 2018b), and a compromised capacity of the transfused RBC to circulate upon transfusion (Nemkov et al., 2018b; D’Alessandro et al., 2020b).

Since purine metabolism is intertwined with methionine oxidation, purine salvage, and polyamine synthesis (Reisz et al., 2018), it is interesting to note a significant increase in several polyamines in RBCs from PKD subjects. The polyamines spermine, putrescine, and spermidine have been shown to stabilize erythrocyte membranes (Ballas et al., 1983). As such, increases in polyamines may be compensating for the decreased membrane stability characteristic of PKD associated hemolytic anemia. Polyamines also play a role in direct scavenging of reactive oxygen species, in like fashion to sulfur-containing compounds methionine (D’Alessandro et al., 2020a) and taurine (Bertolone et al., 2020), both increasing significantly in PKD subjects.

Red blood cells lack the Golgi and endoplasmic reticulum (ER) and thus cannot synthesize phospholipids de novo (Wu et al., 2016). Instead, they make use of the Land’s cycle to remodel oxidatively membranes, which has been shown to be upregulated under physiological (e.g., exercise stress; San-Millan et al., 2020; Nemkov et al., 2021) or pathological conditions that increase susceptibility to osmotic stress (e.g., testosterone-replacement therapy; Alexander et al., 2021; storage of RBCs from obese blood donors; Hazegh et al., 2021). Studies in RBC storage have shown the accumulation of bioactive fatty acids including HETEs, polyunsaturated fatty acids (as a function of redox-dependent fatty acid desaturase activity; Thomas et al., 2021; and oxylipins; Howie et al., 2019) – potential indicators of lipid peroxidation and increased oxidative stress (D’Alessandro et al., 2021). The accumulation of catabolites of membrane phospholipids may testify to an increased metabolism of these molecules.

Other studies have noted high levels of 2,3-BPG in PKD patients, thought to be responsible for their high tolerance of PKD associated anemia (Delivoria-Papadopoulos et al., 1969; Grace et al., 2015). In a previous study of RBCs in chronic kidney disease, S1P was demonstrated to promote 2,3-BPG production, a mechanism which appears to be active in PKD samples (Xie et al., 2020). Thus, increased levels of S1P in PKD in response to anemia increase the production of 2,3-BPG which in turn, increases oxygen offloading capabilities of RBCs. Both the potential activation of HIF1A through increased levels of succinate and the increase of S1P and subsequently 2,3-BPG levels suggest that PKD RBCs are reacting to hypoxic conditions – consistent with the observations above on increased purine breakdown and oxidation (e.g., hypoxanthine as a marker of hypoxia in cells with mitochondria).

In conclusion, in the present study, we provide the first overview of RBC metabolism in patients with PKD, expanding on previous reports on whole blood from dried blood spots (Dooijeweert et al., 2021). Though limited in scope (i.e., small cohort with a single genetic mutation R479H), the study addresses the need for basic science to investigate pathologies targeting underrepresented minorities (four out of five PKD subjects were from Amish families), with the ultimate goal to target treatments to health disparities (Hunter et al., 2021). Indeed, since these subjects suffer from anemia and therapeutic options are limited (e.g., recurring transfusions), this study could pave the way to targeted personalized investigations to drive interventions, such as supplementation of antioxidants (taurine; Bertolone et al., 2020; methionine, and choline for isoaspartyl protein damage-repair; D’Alessandro et al., 2020a); carnitine supplementation (Nikolaos et al., 2000) and dietary interventions to sustain membrane lipid remodeling in response to increased oxidant stress.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

The studies involving human participants were reviewed and approved by IRB Pyruvate Kinase Deficiency Natural History Study Protocol No. 2014-12. The patients/participants provided their written informed consent to participate in this study.

HM and AD’A designed the study. GO and HM enrolled patients and collected the samples. MR, FG, FC, and AD’A performed metabolomics studies. AD’A prepared figures. MR and AD’A wrote the first version of the manuscript that was critically reviewed and approved by MR, FC, GO, FG, HM, and AD’A. All authors contributed to the article and approved the submitted version.

This research was supported by funds from the National Institute of General and Medical Sciences, RM1GM131968 (AD’A), and R01HL146442 (AD’A), R01HL149714 (AD’A), R01HL148151 (AD’A), and R21HL150032 (AD’A), from the National Heart, Lung, and Blood Institute.

AD’A is a founder of Omix Technologies Inc. and Altis Biosciences LLC. AD’A is a consultant for Rubius Therapeutics and an advisory board member for Hemanext Inc. and FORMA Therapeutics Inc.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.