Stored RBC metabolism as a function of caffeine levels

Abstract

BACKGROUND:

Coffee consumption is extremely common in the United States. Coffee is rich with caffeine, a psychoactive, purinergic antagonist of adenosine receptors, which regulate red blood cell energy and redox metabolism. Since red blood cell (purine) metabolism is a critical component to the red cell storage lesion, here we set out to investigate whether caffeine levels correlated with alterations of energy and redox metabolism in stored red blood cells.

STUDY DESIGN AND METHODS:

We measured the levels of caffeine and its main metabolites in 599 samples from the REDS-III RBC-Omics (Recipient Epidemiology Donor Evaluation Study III Red Blood Cell-Omics) study via ultra-high-pressure-liquid chromatography coupled to high-resolution mass spectrometry and correlated them to global metabolomic and lipidomic analyses of RBCs stored for 10, 23, and 42 days.

RESULTS:

Caffeine levels positively correlated with increased levels of the main red cell antioxidant, glutathione, and its metabolic intermediates in glutathione-dependent detoxification pathways of oxidized lipids and sugar aldehydes. Caffeine levels were positively correlated with transamination products and substrates, tryptophan, and indole metabolites. Expectedly, since caffeine and its metabolites belong to the family of xanthine purines, all xanthine metabolites were significantly increased in the subjects with the highest levels of caffeine. However, high-energy phosphate compounds ATP and DPG were not affected by caffeine levels, despite decreases in glucose oxidation products—both via glycolysis and the pentose phosphate pathway.

CONCLUSION:

Though preliminary, this study is suggestive of a beneficial correlation between the caffeine levels and improved antioxidant capacity of stored red cells.

Sixty four percent of adult Americans (18 years old or older) consume at least a cup of coffee a day, according to a 2018 survey by Reuters and National Coffee Association.¹ The Food and Drug Administration recommends keeping the daily caffeine intake below 400 mg, i.e. ~4 cups of coffee per day. The average American consumes ~3.1 cups of coffee per day—making the United States one of the world leading consumers of coffee. While caffeine intake can be sustained through many caffeinated beverages, ~75% of caffeine intake derives from consumption of coffee. Altogether, these numbers explain the economic impact of coffee consumption, a market size of ~80.96 billion dollars in 2019 in the US alone.² It is difficult to understand the popularity of this beverage—and the relevance to the study we are presenting here—without considering the science-backed health benefits of coffee.

Studies have shown how coffee—and caffeine, the psychoactive purine of the methyl-xanthine class it contains—exerts a stimulatory effect and thus increases alertness and performance during the day and night.^3,4 Caffeine also boosts thermogenesis and fat burning,⁵ decreases the risk of type 2 diabetes⁶ and cirrhosis,⁷ lowers the risk of some neurodegenerative diseases like Alzheimer’s⁸ and Parkinson’s⁹ (beneficial effects that are not conferred upon consumption of decaffeinated beverages), counteracts depression¹⁰ and even extends the life expectancy.¹¹ Other than caffeine, coffee contains several antioxidants—including carotenoids—and is de facto the largest dietary source of antioxidants in western diets.¹²

Several mechanisms underlie the above-mentioned benefits associated with coffee consumption. The best established one involves the antagonistic effect of caffeine on adenosine receptors,¹³ which occurs at caffeine concentrations consistent with those achieved during normal human consumption of caffeinated beverages, including coffee. In the central nervous system, the antagonistic effect on adenosine receptors (especially A2A) promotes a stimulatory effect on dopaminergic neurons, which enhances neuronal firing and protects from neurological diseases involving defects of the dopaminergic system, such as Parkinson’s disease.¹⁴

Caffeine absorption takes place in the small intestine and it needs around 45 minutes to become fully (99%) bioavailable, with no significant first pass effect of caffeine. Caffeine absorption rate constant K₀₁ is around 0.33 minute⁻¹ and its half-life is around 4 hours.¹⁵ Caffeine is indeed metabolized in vivo by the liver cytochrome P450 1A2 - CYP1A2, whose polymorphisms influence the distribution of caffeine metabolites that are generated by means of phase I oxidation reactions: paraxanthine (dimethyl-xanthine – 81.5%), theobromine (10.8%), and theophylline (5.4%).¹⁶ In circulation, caffeine interacts with circulating cells, including red blood cells (RBCs), the most abundant cell in the human body (~83% of the total human cells).¹⁷

RBCs also express adenosine receptors^18,19—specifically A2B²⁰—which are subject to the antagonistic effect of caffeine. Recently, we and others have demonstrated the relevance of adenosine signaling through adenosine receptor A2B in promoting RBC glycolysis and oxygen off-loading capacity following physiological exposure to high-altitude hypoxia or pathological hypoxia, such as in the case of hemoglobinopathies (e.g., sickle cell disease).^21–24 Adenosine signaling in mature RBCs activates mechanisms involving the adenosine receptor A2B, a G-protein coupled receptor that interfaces with an adenylate cyclase subunit and thus promotes the synthesis of cyclic AMP. In turn, kinases activate downstream to the adenosine receptor, including Protein Kinase A (PKA) and AMP-dependent protein kinase (AMPK), their targets sphingosine kinase 1, extracellular nucleotide transporter 1, and biphosphoglycerate mutase; these events contribute to the synthesis of 2,3-diphosphoglycerate (DPG) and other molecular cascade that promote a metabolic adaptation to high-altitude hypoxia.^21,25

Owing to the role of caffeine in vasomodulation,²⁶ blood donors are usually recommended to avoid drinking coffee prior to donation because of the potential vasoconstrictive effects that could result in prolonging the time required to finalize the donation. However, coffee—the main source of dietary caffeine intake, above all caffeinated beverages—is also enriched with a series of antioxidant compounds¹² that could benefit the RBC and improve its capacity to cope with the oxidant stress arising during storage in the blood bank. Moreover, we recently demonstrated a beneficial effect of purinergic agonist and antagonist in the modulation of RBC metabolic function, with respect to energy and redox metabolism in response to high-altitude hypoxia or pathological oxidant stresses (e.g., sickle cell disease).^21–25 As such, to further understand the impact of caffeine levels on RBC storage quality, in the present study we leveraged the biobank established within the framework of the REDS-III RBC Omics study to investigate whether the RBC levels of caffeine and its main metabolites (especially paraxanthine) correlated with any specific improvements in stored RBC metabolism.

MATERIALS AND METHODS

REDS-III RBC-Omics study participants and samples

This study is part of the REDS-III RBC-Omics study. As such, details related to donor selection and recruitment have been extensively described.^27–29 Briefly, 13,403 healthy blood donors were recruited at four different blood centers in the United States of America to donate a unit of leukocyte-filtered erythrocyte concentrates. These units were stored until the end of their shelf-life (storage day 42), when they were tested for the propensity of stored RBCs to hemolyze, either spontaneously, or following osmotic, oxidative, or mechanical insults, as previously detailed.^27,28 Donors were thus ranked based on these measurements of hemolysis, and those donors showing the lowest (5th percentile) and highest (95th) hemolytic propensity for any of the hemolytic parameters were contacted again and asked to donate a second unit of blood. Sterile sampling of these units was performed at storage day 10, 23, and 42, as described.³⁰ All technical aspects of this phase of the study, including blood collection, sample processing, and other aspects of the screening and recall phases of the RBC-Omics Study are detailed in prior publications.^28,31

Sample processing and metabolite extraction

An isotopically labeled internal standard mixture including a mix of ¹³C¹⁵N-labeled amino acid standards (2.5 μM) was prepared in methanol. A volume of 100 μL of frozen RBC aliquots was mixed with water and the internal standards (1:1:1, v/v/v). The samples were extracted with methanol (final concentration of 80% methanol). After incubation at −20°C for 1 hour, the supernatants were separated by centrifugation and stored at −80°C until analysis.³² Samples were vortexed and insoluble material pelleted as described.^33,34

Ultra-high-pressure liquid chromatography-mass spectrometry metabolomics

Analyses were performed using a Vanquish UHPLC coupled online to a Q Exactive mass spectrometer (Thermo Fisher). Samples were analyzed using a 3-minute isocratic condition³⁵ or a 5-, 9-, and 17-minute gradient as described.^34,36–38 For the analysis of caffeine, paraxanthine/theobromine/theophylline, methyl-xanthine levels were determined, samples were analyzed using a 5-minute C18 gradient and MS positive ion mode acquisition as described.^34,36–38 Solvents were supplemented with 0.1% formic acid for positive mode runs and 1 mM ammonium acetate for negative mode runs. MS acquisition, data analysis, and elaboration were performed as described.^34,35 Additional analyses, including untargeted analyses were performed with Compound Discoverer 2.0 (Thermo Fisher). Graphs and statistical analyses (repeated measures ANOVA and correlation analyses) were prepared with GraphPad Prism 8.0 (GraphPad Software, Inc). Pathway analyses and other statistical analyses were performed with MetaboAnalyst 4.0³⁹ and the KEGG pathway-based interactome was generated with the OmicsNet tool.⁴⁰

RESULTS

Caffeine and caffeine metabolites across the REDS-III RBC-Omics recalled donor population

A summary of the experimental design of the REDS-III RBC-Omics study is provided in Fig. 1A, as described.²⁷ In the present study, we directly measured caffeine and caffeine metabolites paraxanthine, theobromine, theophylline, and methyl-xanthine in the REDS-III RBC-Omics recalled donor population were performed via UHPLC–MS. Measurements were validated against chromatographic retention times and high-resolution intact mass for commercially available standards, an analytical measurement that also affords the determination of raw chemical formulae upon deconvolution of isotopic patterns (Fig. 1B). Data are extensively reported in tabulated form in Table S1, available as supporting information in the online version of this paper. Caffeine levels did not change during storage in this study (Fig. 1C), while minor, albeit not significant increases were noted in the levels of caffeine metabolites (paraxanthine in Fig. 1D). Caffeine and paraxanthine (and isobaric metabolites) levels were not normally distributed across the whole population, with the majority of healthy donors displaying caffeine (260 samples out of 599 tested in this study) and paraxanthine (198 out of 599) levels lower than the median levels in the whole recalled donor population at any given time point (Fig. 1E).

Fig. 1.

UHPLC–MS measurements of caffeine and caffeine metabolites in the REDS-III RBC-Omics recalled donor population. Within the framework of the REDS-III RBC-Omics study, a large cohort (13,400) of healthy volunteers donated a unit of blood (A). Units were stored until the end of their shelf-life (42 days) and then tested for the propensity to hemolyze (either spontaneously or following oxidant or osmotic insults). The lowest (5th) and highest (95th) percentile donors with respect to these hemolysis parameters were asked to donate a second unit of blood. The second unit was sampled at storage days 10, 23, and 42 for metabolomics analyses. These analyses included measurements of caffeine based on chromatographic retention times and high-resolution intact mass spectra via UHPLC–MS (B). Caffeine levels did not change during storage in this study (C), while minor, not significant increases were noted in the levels of the main caffeine metabolite paraxanthine (D). Caffeine and paraxanthine levels were not normally distributed across the whole population (E), with the majority of healthy donors displaying caffeine (260 samples out of 599 tested in this study) and paraxanthine (198 out of 599) levels >3 standard deviations lower than the median levels in the whole recalled donor population at any given time point. [Color figure can be viewed at wileyonlinelibrary.com]

Caffeine levels are associated with distinct metabolic phenotypes

After determining the inter-donor heterogeneity in blood caffeine levels, we set out to investigate whether low or elevated caffeine consumption prior to blood donation is associated with distinct metabolic phenotypes. As such, we leveraged the measurements of caffeine and its metabolites to identify 15 subjects with the lowest levels of both caffeine and paraxanthine, theobromine, and theophylline (the most abundant caffeine metabolites in the cohort investigated this study) and 15 subjects with the highest levels of these metabolites at storage day 10 (earliest time point available in this study—Table S1, available as supporting information in the online version of this paper, which also includes the demographics of this population). Metabolic phenotypes were assessed in the two sub-groups at storage days 10, 23, and 42, and processed with multivariate analyses, including partial least square-discriminant analysis (PLS-DA – Fig. 2A), hierarchical clustering analysis (HCA-Fig. 2B) and metabolic pathway enrichment analysis (Fig. 2C). PLS-DA separated samples on the basis of the metabolic phenotypes across the 90 samples tested in this part of the study (i.e., n = 15 for low or high-caffeine; storage days 10, 23, and 42). Storage time emerged as one of the top loading variables (vector aligned with PC2, explaining 11.1% of total variance – Fig. 2A). Caffeine and paraxanthine levels mostly discriminated samples across principal component 3 (PC3: 9.7% of the total variance—red vs. blue—Fig. 2A). Significant metabolites were thus determined by repeated measure ANOVA (two factors: time and caffeine/metabolite levels— Table S1, available as supporting information in the online version of this paper) and used to filter metabolites in the unsupervised hierarchical clustering reported in the heat map in Fig. 2B (map provided in scalable format in Fig. S1, available as supporting information in the online version of this paper.). Pathway analysis of significant metabolites as a function of caffeine levels and storage duration suggested an impact of these variables on energy metabolism (glycolysis), tryptophan, arginine, methionine, and amino acid metabolism at large, gltuathione metabolism, fatty acid, and purine oxidation (Fig. 2C). In the paragraphs that follow we provide a detailed overview of key metabolites from these pathways and how they are affected by storage duration as a function of caffeine levels.

Fig. 2.

Multivariate analysis of metabolomics data from the 15 subjects showing the highest and lowest caffeine and caffeine metabolite levels at storage days 10, 23, and 42. In (A), partial least square discriminant analysis (PLS-DA) separates samples on the basis of the metabolic phenotypes and clusters them as a function of storage time (vector aligned with PC2, explaining 11.1% of total variance in this subset of 90 samples) and caffeine levels (PC3: 9.7% of the total variance—red vs. blue). In (B), hierarchical clustering of significant metabolites by repeated measure ANOVA (two factors: time and caffeine/metabolite levels) for the 15 subjects showing extremes in the levels of caffeine and its metabolites (a vectorial version of this panel is provided in Fig. S1, available as supporting information in the online version of this paper.). In (C), OmicsNet-based pathway analysis of significant metabolites by ANOVA. [Color figure can be viewed at wileyonlinelibrary.com]

High levels of caffeine are associated with lower glucose oxidation via either glycolysis and the pentose phosphate pathway

High levels of caffeine and its metabolites were associated with significantly lower levels of glucose and its oxidation products via glycolysis (glucose 6-phosphate, fructose bisphosphate, glyceraldehyde 3-phosphate, pyruvate, and lactate) and the pentose phosphate pathway (PPP – ribose phosphate) (Fig. 3). Specifically, significantly lower levels of glucose, glucose 6-phosphate, fructose bisphosphate, and lactate at all the storage days tested in this study (Fig. 3A). However, no significant differences were noted in the levels of the main high-energy phosphate compounds adenosine triphosphate (ATP) and 2,3-diphosphoglycerate (DPG) at any storage time tested in this study (Fig. 3). These subjects were also characterized by significantly higher levels of reduced glutathione and its precursors, glutamate, and cysteine through the whole storage period (Fig. 4A). Vice versa, lower levels of oxidized glutathione (GSSG) and glutathionyl-cysteine (a glutathione adduct to cysteine that forms under prooxidant conditions) were detected in high caffeine subjects at storage day 10, which were instead characterized by higher levels of the glutathione turn-over product, 5-oxoproline especially by the end of the storage period (Fig. 4A). Higher availability of reduced glutathione seemed to fuel the glyoxylate pathway to counteract triose sugar aldehyde formation, overall resulting in decreases in the levels of lactaldehyde and increases in the levels of methylglyoxal and lactoyl-glutathione in those subjects with higher levels of caffeine through the whole storage period (Fig. 4A). Similarly, glutathione availability resulted in increases in the levels of glutathionylated forms of lipid peroxidation products, such as 4-hydroxynonenal detoxification catabolites glutathionyl-hydroxynonenal (GSHNE) and glutathionyl-hydroxynonanoic acid (GSHNA), (Fig. 4A).⁴¹ These pathways are dependent on NADPH-dependent aldose reductases in mammalian erythrocytes,⁴¹ suggestive of a potential cross-talk between caffeine levels and sugar metabolism in human erythrocytes. Interestingly, higher caffeine levels were not only associated with lower levels of glucose, but also of ascorbate and dehydroascorbate—which compete with glucose for the same transporter (GLUT1) for their uptake in red cells.⁴² The glutamate precursor, glutamine was also increased in RBCs from subjects with high levels of caffeine (Fig. 4A), as it was the glutamate transamination products^35,43 alanine and aspartate.

Fig. 3.

Glycolysis (A) and the pentose phosphate pathway (B) in RBCs at storage days 10, 23, and 42 in 15 subjects with low (blue) or high (red) levels of caffeine and its metabolites in the REDS-III RBC-Omics recalled donor population. Asterisks indicate significant results (ANOVA, Tukey multiple column comparison: * p < 0.05; ** p < 0.01 *** p < 0.001). [Color figure can be viewed at wileyonlinelibrary.com]

Fig. 4.

Glutathione, ascorbate, and glyoxylate metabolism (A) and purine oxidation (B) in RBCs at storage days 10, 23, and 42 in 15 subjects with low (blue) or high (red) levels of caffeine and its metabolites in the REDS-III RBC-Omics recalled donor population. Asterisks indicate significant results (ANOVA, Tukey multiple column comparison: * p < 0.05; ** p < 0.01 *** p < 0.001). [Color figure can be viewed at wileyonlinelibrary.com]

Since caffeine and its metabolites are purines, we thus focused on purine metabolism—which is critical to RBC storage quality and correlates with the stored RBC capacity to circulate upon transfusion in mice and humans.⁴⁴ Notably, all purine metabolites in the pathway appeared to be increased in the subjects with high levels of caffeine, including hypoxanthine (only at storage day 42), xanthine, urate, and 5-hydroxy-isourate (Fig. 4B). While storage-dependent increases in the levels of these metabolites are a hallmark of oxidant stress to the stored erythrocyte,^44–46 increases in the levels of aspartate and decreases in the levels of fumarate and malate were here noted in the subjects with high caffeine levels (Fig. 4B and Table S1, available as supporting information in the online version of this paper). These observations are consistent with a decreased activation of aspartate-dependent purine salvage reactions, active—though minimally—in the mature erythrocyte.⁴⁴ This consideration is further supported by the lack of differences in ATP (Fig. 3) and the levels of its breakdown products, AMP (Fig. 4B) and adenosine (Fig. 5A) between the low and high caffeine groups.

Fig. 5.

Methionine (A), fatty acid and oxylipin (B), arginine (C), and tryptophan metabolism (D) in RBCs at storage days 10, 23, and 42 in 15 subjects with low (blue) or high (red) levels of caffeine and its metabolites in the REDS-III RBC-Omics recalled donor population. Asterisks indicate significant results (ANOVA, Tukey multiple column comparison: * p < 0.05; ** p < 0.01 *** p < 0.001). [Color figure can be viewed at wileyonlinelibrary.com]

Subjects with elevated levels of caffeine were characterized by higher methionine levels and decreased consumption of methionine by 1) direct oxidation (methionine sulfoxide) or 2) utilization as a methyl-group donor to form the S-Adenosylmethionine (Fig. 5A), which is necessary to fuel methyl-transferases involved in repair of protein damage to aspartyl groups elicited by oxidant stress^47,48 Finally, no significant changes were observed in the levels of oxylipins or free fatty acids as a function of caffeine levels, though lower (albeit not significant) levels of 5-HETE, 15-HETE, and 16-HDoHE were detected in RBCs from high caffeine donors at storage day 42 (Fig. 5B). High caffeine levels were also positively associated with increases in arginine metabolites citrulline and ornithine (Fig. 5C), as well as tryptophan and several of its metabolites, including kynurenine and picolinic acid (Fig. 5D, E).

Metabolic correlates to caffeine in stored RBCs

Correlation analyses (Spearman) were performed between caffeine levels and metabolic measurements in the whole population (Fig. 6A) or just the 15 subjects with extreme levels of caffeine and its metabolites (Fig. 6B). Values are detailed in ranked order (top negative to positive correlates) in Table S1, available as supporting information in the online version of this paper. Expectedly, caffeine and paraxanthine levels correlated significantly in most of the samples tested in this study (Fig. 6C). A sub-population of donors showed a lack of correlation between caffeine and paraxanthine levels and a positive correlation between caffeine and methyl-xanthine (Fig. 6D), reflecting well-established heterogeneity in caffeine metabolism in humans as a function of polymorphisms in cytochrome P450 1A2.¹⁶ Levels of caffeine and its metabolites were overall positively associated, albeit not significantly, with resistance to spontaneous and oxidative hemolysis through the whole storage period and osmotic hemolysis at least at day 10 (Fig. 7). No significant association was noted with markers of smoking (cotinine), though subjects with higher levels of caffeine also showed higher levels of ethyl-glucoronide (Fig. 7), a stable marker of moderate alcohol consumption.⁴⁹ Despite the positive correlation of alcohol consumption with oxidant stress (D’Alessandro et al. under review), subjects with higher caffeine levels were instead characterized by overall decreases in markers of oxidant stress, such as methionine sulfoxide, S-adenosylmethionine (SAM) to S-adenosyl-homocysteine (SAM/SAH) ratios, lactaldehyde (Fig. 6E) and increases in markers of tryptophan and indole metabolism (tryptophan, kynurenine, picolinic acid, indole acetaldehyde—Fig. 6A and E). Similarly, other aromatic amino acids were positively correlated to caffeine levels, such as tyrosine and its metabolites, dopamine, and adrenaline (Fig. 7).

Fig. 6.

Correlation analysis of caffeine levels and metabolomics data in the whole REDS-III RBC-Omics recalled donor population (A), in the subjects showing the highest and lowest caffeine levels at any storage day (B) and highlighted correlations (C-E). [Color figure can be viewed at wileyonlinelibrary.com]

Fig. 7.

Caffeine levels had a limited impact on oxidative osmotic and spontaneous hemolysis (y axis indicates % hemolysis in A). Caffeine drinking was not associated with increased nicotine exposure, but it was associated with higher levels of ethyl-glucuronide, a marker of moderate alcohol exposure (B). In (C), tyrosine metabolism coffee high (red) versus low (blue) caffeine levels. [Color figure can be viewed at wileyonlinelibrary.com]

DISCUSSION

In recent years, our understanding of the RBC metabolic storage lesion has significantly improved.^50,51 Studies have extensively shown a linkage between altered RBC energy metabolism and increased susceptibility to oxidant stress.⁵² Ultimately, the storage lesion is thought to influence transfusion outcomes such as intravascular lysis or extravascular clearance in the recipient upon transfusion.^53,54 While the progression of the storage lesion is inevitable, the rate and extent of this phenomenon is impacted by factors such as donor biology (sex, ethnicity, age²⁷), genetic makeup (e.g., glucose 6-phosphate dehydrogenase deficiency^55,56), processing strategies (e.g., additive solutions^30,57–59) and, preliminary data seems to suggest, habits that impact RBC biochemistry—such as exercise⁶⁰ or smoking.^61,75 While the role of purinergic signaling in boosting RBC energy metabolism is well established,²⁰ little is known about the impact of caffeine levels on stored RBC metabolism. To fill this gap in knowledge, in the present study we leveraged samples collected within the framework of the recall phase of the Recipient Epidemiology and Donor Evaluation Study—REDS-III RBC-Omics.^27,30,31,62 One caveat of this study is that this cohort is biased toward those subjects whose end of storage RBCs showed extreme (5th or 95th percentile—i.e., lowest or highest) propensity to hemolyze spontaneously or following oxidant or osmotic insults out of the original cohort of 13,403 donors. For this reason, this cohort offers the opportunity to determine whether caffeine levels in stored RBC units are associated with extremes in hemolytic propensity. Through advanced mass spectrometry-based metabolomics approaches here we correlated RBC levels of caffeine and its primary metabolites to specific metabolic and hemolytic phenotypes as a function of storage duration. However, it is important to note that self-reported questionnaires on coffee consumption were not available for this study. As such, in the present study any inference of coffee intake from caffeine levels would be speculative, in that caffeine levels can be influenced by a combination of factors such as the actual caffeine intake (e.g., through coffee or caffeinated energy drinks—see Bertolone et al.),⁷⁶ the inter-donor heterogeneity in caffeine metabolism¹⁶ and the time that has elapsed between coffee consumption and blood donation. Having acknowledged these limitations, here we noted a significant positive correlation between those subjects with the highest levels of caffeine and the activation of a series of antioxidant pathways. Above all, we reported a positive correlation between the levels of caffeine and its metabolites and glutathione homeostasis, glutathione-dependent detoxification of oxidized triose sugar aldehydes (glyoxylate pathway) and oxidized lipids (hydroxynonenal metabolism) and other sulfur-metabolites like methionine, critical in antioxidant metabolism and repair mechanisms of isoaspartyl protein damage by oxidant stress.⁴⁸ On the other hand, we saw no decreases in the levels of oxidized purines, which however appeared to be fueled by the breakdown of caffeine and its metabolites (xanthine purines) —rather than by ongoing purine oxidation. Indeed, while the levels of hypoxanthine and its metabolites had been previously reported as critical markers of the metabolic age of stored RBCs^45,46 and a predictor of post-transfusion performances in vivo,^44,63,64 previous studies had reported that increases in hypoxanthine were tied to breakdown of ATP—the main energy currency of the RBC and a positive correlate to RBC lifespan and post-transfusion recovery in vivo.⁶⁵ In the present study, increases in the levels of xanthine purines was not accompanied by decreases in ATP, nor by increases in its breakdown products AMP and adenosine or the products (e.g., the carboxylic acid fumarate) of salvage reactions pathways that can be activated to partially compensate for purine deamination. Vice versa, increases in the levels of aspartate (amine group donor in such reactions) and other transamination products such as alanine and glutamate are suggestive of a more active amino acid metabolism in RBCs from donors with higher levels of caffeine. Interestingly, this also affects tyrosine and its catabolites, dopamine, and adrenaline, consistent with the extensive literature on the stimulatory role of caffeine and its interaction with the dopaminergic system.^9,14 Here we also noted decreases in glucose and glycolytic metabolites and byproducts (such as lactate), as well as ribose phosphate—an intermediate of the PPP. These results could be in part explained by decreases in the activity of by glucose transporters—such as GLUT1,⁴² resulting in decreased uptake of glucose or other metabolic substrates of this transporter (e.g., dehydroascorbate). In the absence of flux data, these steady state observations are suggestive of either 1) decreased fluxes through the PPP, 2) decreased glucose uptake overall, or 3) decreased oxidant stress in RBCs from subjects with high levels of caffeine. The latter hypothesis is supported by prior reports in the literature⁶⁶: caffeine had been previously suggested to activate antioxidant pathways in RBCs by stabilizing the relaxed state of hemoglobin and its interaction with N-terminus of Band 3.⁶⁶ Alterations of such mechanism have been reported to correlate with the severity of the storage lesion, and interventional studies to leverage this mechanism have relied on hypoxic storage to modulate the oxygen saturation of hemoglobin.⁴⁸ These mechanisms would be chronically triggered through the shelf-life of the unit by caffeine, owing to the apparent lack of caffeine metabolism in the mature erythrocyte. Indeed, no significant changes in caffeine and its metabolite levels were observed over storage in this study. Interventional studies will need to be designed to disentangle the independent impact of such variables (caffeine, taurine⁶⁷ or dietary antioxidants present in coffee or other caffeinated beverages) on the metabolic phenotypes of stored RBCs. In this view, it is interesting to note a positive correlation between caffeine levels and indole/tryptophan metabolites. This observation is relevant in that indole metabolites of bacterial origin have been previously associated to altered synthesis of nicotinamide adenine dinucleotide (NAD) in aging⁶⁸ and Down syndrome.⁶⁹ Positive correlation between caffeine and the tryptophan catabolite picolinic acid is also relevant in that this metabolite has been reported to inhibit the neurotoxicant, but not the neuroexcitant effects of other tryptophan catabolites like quinolinic acid.⁷⁰ Finally, alterations of RBC tryptophan metabolism have been found to link erythrocyte metabolism to the microbiome of transfusion recipients, especially those receiving end of storage units and a bolus of iron resulting from their extravascular hemolysis.⁷¹

Though preliminary, the present study expands on the growing appreciation of the impact of donor-dependent variables on blood storability. While prior work had suggested a negative impact of smoking on the accumulation of markers of oxidant stress in stored RBCs such as carboxyhemoglobin,⁶¹ in independent study we have expanded on these observations suggesting a negative correlation between smoking and even moderate alcohol consumption (as gleaned by direct measurements of ethyl-glucoronide) on stored RBC metabolism.⁷⁵ Previous studies had shown an impact of sex on blood storability.²⁷ However, caffeine levels did not significantly correlate with sex, consistent with national statistics showing similar figures for coffee consumption in male and female (66% vs. 62%¹). On the other hand, here we comment on the positive correlation between caffeine levels and the antioxidant metabolism of stored RBCs, despite the observation of a positive correlation between alcohol consumption and caffeine levels. However, this apparent improvement in RBC energy and redox metabolic markers did not result in significant (only trending) improvements in spontaneous, oxidative, or osmotic hemolysis. In this view, it is interesting to note that the impact of some of these variables on stored RBC metabolism may be negligible and the impact of one variable can end up masking the other, further complicating mixed mode analyses in the absence of a clear indicator of transfusion efficacy beyond the highly debated and often-time challenged parameters of hemolysis, post-transfusion recovery⁷² or mortality.⁷³ In this view, it is worth noting that polymorphisms of cytochrome p450 1A2 which promote faster caffeine clearance have been previously correlated with higher post-transfusion recovery in pilot studies on only three subjects.¹⁶ More recent and extensive investigations on metabolic correlates to post-transfusion recovery under normoxic and hypoxic storage conditions showed a link between hemoglobin oxygenation state, caffeine levels, and post-transfusion recovery⁷⁷—consistent with a potential role of caffeine in the purinergic signaling component of the so-called oxygen-dependent metabolic modulation.⁶⁶ Studies linking donor-recipient databases,⁷⁴ whereas accompanied by metabolic testing of the transfused unit at the time of administration, could potentially help bridging this gap in knowledge between the molecular characterization of the storage lesion and its clinical impact.

Supplementary Material

S Fig 1

Supplementary Figure S1. A detailed heat map showing significant metabolites by ANVOA.

S Table 1

Supplementary Table S1. Metabolomics report of caffeine levels and stored RBC metabolism.