Salvianolic acid B alleviates red blood cell storage lesion with GPX4 preservation and reduced lipid peroxidation

Abstract

Background

Red blood cell (RBC) storage lesion is a cascade of biochemical and physical alterations that impair transfusion efficacy and safety. This study explored the role and underlying mechanism of salvianolic acid B (Sal B) against RBC storage lesion, with a specific focus on a ferroptosis-like process.

Materials and methods

RBC from healthy volunteers were stored in MAP additive solution for 35 days, and samples were analyzed at weekly intervals. Ferroptosis was determined by the levels of cytosolic reactive oxygen species (ROS), antioxidant enzymes superoxide dismutase and glutathione peroxidase, 4-hydroxy-2-nonenal (4-HNE), malondialdehyde (MDA), and glutathione peroxidase 4 (GPX4). Storage lesion was evaluated by erythrocyte morphology, hemolysis rate, microvesicle formation, phosphatidylserine exposure, lactate dehydrogenase release, osmotic fragility, and Band 3. Ferrostatin-1 was used to verify the involvement of ferroptosis.

Results

The results demonstrated a time-dependent escalation of both storage lesion and ferroptosis, evidenced by biochemical changes such as increased hemolysis, phosphatidylserine exposure, ROS accumulation, 4-HNE and MDA, as well as a reduction in antioxidant capacity and protein level of GPX4. Ferrostatin-1 treatment effectively mitigated pathological changes, suggesting that ferroptosis-associated pathways may contribute to storage lesion. Notably, Sal B supplementation correlated with reduced ferroptosis markers and mitigation of storage lesion, as evidenced by preserved GPX4 levels and reduced lipid peroxidation.

Discussion

This study demonstrates that a ferroptosis-like process of lipid peroxidation contributes to RBC storage lesion, with GPX4 regulating erythrocyte stability. Sal B alleviates storage lesion in parallel with preservation of GPX4 and reduced lipid peroxidation, providing a potential strategy to improve the quality of stored blood.

INTRODUCTION

Red blood cell (RBC) transfusion is a commonly used medical method in clinical treatment and emergency rescue, and the advent of anticoagulants and preservative solutions in the early 20^th century was a crucial breakthrough to meet blood storage and transfusion demands. Although blood transfusion can be life-saving in numerous clinical contexts, patients are still at risk of developing serious posttransfusion complications^1,2. The U.S. Food and Drug Administration and the European Council stipulate that the storage life of RBC units is limited to 42 days, although certain preservation methods, such as cryopreservation with glycerol at extremely low temperatures, can extend this period to up to 10 years. During storage, RBC continuously accumulate reactive oxygen species (ROS)³, causing oxidative damage to lipids and proteins^4–6. This progressive and irreversible process is termed “storage lesion”. This kind of damage promotes the formation of vesicles from compromised components⁷, leading to morphologically irregular and smaller RBC, which are predominantly cleared by the recipient’s reticuloendothelial system upon transfusion⁸.

The mechanisms contributing to RBC damage primarily include oxidative damage to proteins, lipids, and hemoglobin and depletion of intracellular metabolites such as ATP and 2,3-DPG⁶. In-depth investigation into the mechanisms and clinical effects of the RBC storage lesion is crucial for identifying interventions to reduce risks for transfusion recipients. Recent research has explored adding various components, such as antioxidants or active ingredients from traditional Chinese medicine to RBC preservation solutions to mitigate cell damage and extend blood storage time. Ethanol⁹, cerium oxide nanoparticles¹⁰, vitamin C¹¹, deuterated linoleic acid¹², and naringin¹³ have been shown to effectively reduce RBC storage lesion. However, their underlying mechanisms of action warrant further investigation. Given that natural products such as naringin have demonstrated positive effects in mitigating RBC storage lesion, this study further focuses on salvianolic acid B (Sal B), aiming to evaluate its potential role in improving the quality of stored RBC.

Ferroptosis is a recently identified mode of programmed cell death that results from the iron-driven buildup of lethal lipid peroxides¹⁴. Previous studies have indicated that ferroptosis is related to multiple diseases including aortic dissection¹⁵, heart failure¹⁶, renal ischemia-reperfusion injury¹⁷, stroke¹⁸, Parkinson’s disease¹⁹, and arrhythmia²⁰. Glutathione peroxidase 4 (GPX4) is the main enzyme that counteracts lipid peroxidation and is a pivotal modulator of ferroptosis²¹. In the context of RBC biology, Stolwijk et al. demonstrated that GPX4 is present as an active enzyme in mature erythrocytes and that its abundance inversely correlates with hemolysis during blood bank storage²². Using erythroid-specific Gpx4 knockout mice and genetic analyses of 13,091 human blood donors, Stephenson et al. demonstrated that GPX4 protects stored RBC from ferroptosis by limiting lipid peroxidation; conversely, GPX4 deficiency or missense variants (e.g., rs73507255) exacerbate lipid peroxidation and impair posttransfusion recovery²³. Page et al. identified GPX4 as a genome-wide significant locus associated with oxidative hemolysis in a large multi-ancestry genome-wide association study of 12,353 blood donors, providing human genetic evidence linking GPX4 to the hemolytic propensity of RBC²⁴. D’Alessandro et al. identified that STEAP3 and ferroptosis-related genes (FADS1/2, EPHX2, LPCAT3, GPX4) regulate a ferroptosis-like process of lipid peroxidation in stored RBC, affecting posttransfusion recovery and hemoglobin increments⁴. These landmark studies have firmly established GPX4 as a central regulator of RBC storage quality and established a ferroptosis-like process as a relevant pathway in storage lesion.

Sal B is the predominant water-soluble constituent extracted from Salvia miltiorrhiza, known for its broad pharmacological effects, including modulation of oxidative stress²⁵, fibrosis²⁶, inflammatory responses²⁷, aging²⁸, and cancer²⁹. Multiple studies have demonstrated that Sal B inhibits platelet activation and aggregation³⁰, regulates glucolipid metabolism³¹, improves hemorheology³² and alleviates liver fibrosis by inhibiting hepatocyte ferroptosis²⁶. Given the mechanistic link between oxidative stress, lipid peroxidation, and ferroptosis, we hypothesized that Sal B might alleviate storage lesion in association with preserved GPX4 function and reduced ferroptosis-associated lipid peroxidation.

In this study, we aimed to examine the involvement of ferroptosis in RBC storage and to evaluate the effect of Sal B supplementation on the storage lesion. We hypothesized that the addition of Sal B to the RBC preservation solution at the beginning of storage would mitigate RBC storage lesion, an effect that correlates with preserved GPX4 levels and reduced ferroptosis markers.

MATERIALS AND METHODS

Ethics

Blood collection from healthy donors and preclinical study in our research complies with all relevant ethical guidelines. Human blood was collected in accordance with standard operating procedures. Informed consent was obtained from each donor. All blood donors were 18 years of age or older (age range, 25–44 years).

Whole blood collection and preparation, treatment and storage of the red blood cells

A single unit of whole blood (volume: 200 mL ± 10%) was collected into a polyvinyl chloride bag containing citrate-phosphate-dextrose-adenine (CPD-A) as the anticoagulant (Weigao, Shandong, China) according to an approved institutional standard operating procedure. The whole blood was centrifuged at 1,000 g for 15 minutes to separate the plasma which was discarded. Mannitol-adeninephosphate (MAP) was added to the packed RBC to obtain suspended RBC. Prior to storage, ferrostatin-1 (Fer-1) was added to the suspended RBC at final concentrations of 100 nM and 500 nM to assay for ferroptosis. Sal B was added to the suspended RBC at a final concentration of 1 μM before storage to elucidate its effects. The suspended RBC were then stored under standard blood banking conditions at 2–6°C and were analyzed weekly after storage. Fer-1 was obtained from MedChem Express (Monmouth Junction, NJ, USA). Sal B was obtained from Solarbio (Solarbio, Beijing, China).

Evaluation of red blood cell storage lesion

To evaluate storage lesion of RBC, microvesicle release, rate of hemolysis, lactate dehydrogenase release, abundance of erythrocyte membrane integrity protein Band 3, level of phosphatidylserine (PS) exposure, and osmotic fragility were determined.

Suspended RBC were centrifuged at 800 g for 10 minutes to harvest the supernatant, then the supernatant was centrifuged at 10,000 g for 10 minutes. The final supernatant was collected and stained with annexin V to evaluate microvesicles using flow cytometery (BD Falcon, San Jose, CA, USA). Flow cytometry data were analyzed following established guidelines³³. RBC were gated by forward and side scatter to exclude debris and platelets. Doublets were excluded by side scatter pulse width. The PS positivity threshold was set using an FMO control and unstained RBC as a negative control; 100,000 events were acquired per sample. Data were analyzed using FlowJo (v10; BD Biosciences, Franklin Lakes, NJ, USA) with identical gating across all groups.

The supernatant from the 10,000 g spin was further utilized to determine the concentration of cell-free hemoglobin by colorimetric assay (Beijing Real Tech, Beijing, China). The hematocrit and total hemoglobin levels were measured using a Sysmex XS-800i hematology analyzer (Sysmex, Kobe, Japan). The rate of RBC hemolysis was calculated according to the following formula:

$$(\mathit{1}-hematocrit)\times cell-free\hspace{0.17em}hemoglobin\hspace{0.17em}(g/dL)/total\hspace{0.17em}hemoglobin\hspace{0.17em}(g/dL)$$

The final supernatant was also used to test lactate dehydrogenase (LDH) level with a lactate dehydrogenase kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

PS exposure on the erythrocyte surface was quantified using flow cytometry via staining with phycoerythrin-labeled annexin V (BD Biosciences). Band 3 abundance on erythrocyte membrane was determined by western blotting.

Osmotic fragility

RBC were aliquoted into 11 tubes containing a saline concentration gradient ranging from 0% to 0.9%, with 0% represented by distilled water. The tubes were mixed and incubated for 1 hour at room temperature; the samples were then centrifuged at 800 g for 10 minutes. Absorbance of the supernatant was measured via a spectrophotometer at a wavelength of 540 nm. The rate of hemolysis was reported as a percentage. Full hemolysis in distilled water was set at 100% reference and treatment with normal saline serving as the 0% reference.

Ghost preparation, spectrin extraction, and western blotting

The supernatant was discarded after centrifuging suspended RBC (2 mL) at 1,000 g for 10 minutes. Isotonic phosphate-buffered saline (pH 7.4) was added at a 1:3 ratio, followed by centrifugation at 1,000 g for 10 minutes and removal of the supernatant. This washing step was repeated three times. Then, a pre-cold Tris-HCl solution (5 mM, pH 7.4) containing 1% protease inhibitor was added at a 1:40 ratio, mixed thoroughly, and incubated overnight at 4°C to ensure complete hemolysis. The pellet was washed three times with 5 mM Tris-HCl solution after centrifugation at 21,100 g for 10 minutes. The resulting milky-white membrane pellet was collected as erythrocyte ghosts.

Erythrocyte ghosts were lysed in 1× radio-immunoprecipitation assay buffer supplemented with protease inhibitors, and gently stirred for 1 hour at 4°C. Debris was discarded after centrifugation at 4°C (12,000 g for 1 hour).

Protein concentration was quantified with the Bradford assay (Thermo Fisher Scientific, Waltham, MA, USA). Equal amounts of protein were separated by 12% sodium dodecylsulfate-polyacrylamide gel electrophoresis. Subsequently, the proteins were transferred onto polyvinylidene difluoride membranes (EMD Millipore, Billerica, MA, USA), blocked with 5% skim milk powder, and incubated with primary antibodies including anti-Band 3 (1:1000, MCE HY-P82074) and anti-GPX4 (1:2,000, ab125066) at 4°C overnight. Membranes were incubated with horseradish peroxidase-conjugated secondary antibody (1:10,000, Santa Cruz Biotechnology, Inc., Dallas, TX, USA) for 1 hour after washing with TBS/Tween. Enhanced chemiluminescence (EMD Millipore) was used to detect the protein bands.

Conventional loading controls (e.g., tubulin, β-actin, and GAPDH) demonstrated limited reliability in western blot analysis of erythrocytes, so we used whole-gel Coomassie staining to verify protein loading uniformity. The gel was immersed in Coomassie Brilliant Blue R-250 Staining Solution (Thermo Fisher Scientific) for 2 hours, followed by destaining in a solution containing 10% acetic acid, 50% methanol, and 40% water. The resulting Coomassie-stained bands were used to confirm the positions of residual proteins.

Morphology

RBC morphology was assessed by examining RBC smears under an optical microscope following Wright staining (Beyotime, Shanghai, China). Meanwhile, flow cytometry was used to analyze cell size and internal complexity. The specific procedure was as follows: the RBC suspension was resuspended in 1× phosphate-buffered saline for detection. The median fluorescent intensity of forward scatter reflected cell size, and the median fluorescent intensity of side scatter was used to estimate cellular complexity.

Detection of reactive oxygen species

Suspended RBC were washed with normal saline three times, resuspended in normal saline, and incubated with 10 μM 5-(and-6)-carboxy-2′,7′-dichlorodihydro fluorescein diacetate (carboxy-H2DCFDA; Sigma-Aldrich, Burlington, MA, USA) at 37°C under light-protected conditions for 30 minutes. After incubation, the cells were centrifuged to remove the supernatant, the cell pellet was washed with normal saline three times, and then analyzed by flow cytometry.

Measurement of malondialdehyde and superoxide dismutase activity

The suspended RBC (50 μL) were washed three times. The pellet was resuspended in 200 μL of distilled water and vortexed for 1 minute to lyse the cells. Subsequently, absolute ethanol (100 μL) was added and vortexed for 30 seconds. Next, 100 μL of trichloromethane was added, and vortexed for 1 minute. Centrifugation at 3,000 g for 8 minutes resulted in separation into three layers. The top layer was collected for the determination of malondialdehyde (MDA) and superoxide dismutase (SOD). MDA concentration and SOD enzyme activity were assessed with a lipid peroxidation MDA assay kit (Beyotime, Shanghai, China) and an SOD assay kit (Nanjing Jiancheng Bioengineering Institute), respectively, according to the manufacturers’ instructions.

Assay of glutathione peroxidase activity and 4-hydroxy-2-nonenal level

A clear hemolysate was prepared by diluting RBC with distilled water in order to assay glutathione peroxidase (GSH-Px) activity and 4-hydroxy-2-nonenal (4-HNE) level. GSH-Px activity and 4-HNE level were assessed with a GSH-Px activity assay kit (Nanjing Jiancheng Bioengineering Institute) and 4-HNE assay kit (Abbkine, Wuhan, China) according to the manufacturers’ instructions.

Statistical analysis

Statistical analysis was performed using GraphPad Prism (v7.0, GraphPad Software Inc., La Jolla, CA, USA). Data are presented as mean ± standard deviation. A total of six independent healthy donors (No.=6 biological replicates per group) were used for all experiments.

For within-group comparisons vs day 0 (indicated by * in figures), a paired Student’s t-test was used (same donor at different timepoints). For between-group comparisons across multiple timepoints (indicated by # in figures), two-way analysis of variance with treatment and time as factors, followed by Sidak’s post-hoc test were used to compare treatment groups at each timepoint.

A p-value <0.05 was considered statistically significant.

RESULTS

Ferroptosis-associated changes are observed in red blood cells during storage

All experiments were performed using RBC from six independent healthy donors (No.=6 biological replicates). Storage lesion in stored RBC was determined by assessing rate of hemolysis, PS exposure and Band 3 abundance. Hemolysis, the release of hemoglobin from RBC into the surrounding fluid, indicates compromised structural integrity of these cells. PS is normally located on the inner leaf let of the RBC membrane, and exposed externally as the cells age, serving as a signal for cellular clearance or eryptosis. Band 3, a key anion exchanger for CO₂ transport and deformability in the RBC membrane, declines with age and stimulates senescence. The results showed a time-dependent increase in both hemolysis and PS exposure, alongside a decrease in the level of Band 3 protein (Figure 1A–C). To clarify the involvement of ferroptosis in RBC storage lesion, we tested ferroptosis-related indicators. The data indicated that with prolonged storage time, cytoplasmic ROS levels and the accumulation of lipid peroxidation products (4-HNE and MDA) increased gradually, while the activities of SOD and GSH-Px enzymes, as well as the abundance of the ferroptosis protective factor, GPX4 protein, progressively decreased (Figure 1D–I). These data indicate that RBC display features associated with ferroptosis during storage.

Figure 1. Ferroptosis-associated changes are observed during the preservation of red blood cells.

Suspended RBC were stored at 2–6°C and collected on days 0, 7, 14, 21, 28, and 35 of storage (No.=6 biological replicates, each from an independent donor). (A) The rate of hemolysis of suspended RBC was measured at different storage timepoints. (B) Phosphatidylserine exposure was detected by flow cytometry using annexin V staining (left panel), and statistical analysis of annexin V-positive populations of RBC was performed (right panel). (C) Band 3 protein was detected by western blot and statistically analyzed. Coomassie staining served as a qualitative control. (D) Level of cytosolic ROS in RBC. The left and right panels display the flow cytometry histogram and the statistical analysis of ROS-positive RBC, respectively. (E, F) Lipid peroxidation in RBC was determined by assaying 4-HNE and MDA. (G) SOD enzyme activity in RBC was measured. (H) GSH-Px enzyme activity in RBC was measured. (I) Western blot and statistical analysis of GPX4. A paired Student’s t-test was used to compare each timepoint to day 0 (same donor). NS: not significant, *p<0.05, **p<0.01, ***p<0.001. 4-HNE: 4-hydroxy-2-nonenal; MDA: malondialdehyde; RBC: red blood cells; ROS: reactive oxygen series; SOD: superoxide dismutase; GPX4: glutathione peroxidase 4; GSH-Px: glutathione peroxidase.

Ferrostatin-1 reduces ferroptosis-associated changes in erythrocytes

To further explore the involvement of ferroptosis in RBC storage lesion, the ferroptosis inhibitor Fer-1 was added to the MAP solution. Carboxy-H2DCFDA was used to label intracellular ROS. MDA and 4-HNE levels were measured to quantify the end products of lipid peroxidation. The results revealed decreased cytosolic ROS (Figure 2A) and lipid peroxidation (Figure 2B, C) in Fer-1-pretreated RBC compared with the levels in the control group. SOD reduces lipid peroxidation by scavenging excessive ROS, thereby indirectly modulating ferroptosis. GPX4, which requires glutathione (GSH) as a cofactor, is a primary defense against ferroptosis. Enzyme activity of SOD and GSH-Px showed a progressive decline over the storage period, while pretreatment with Fer-1 preserved the activity of SOD and GSH-Px (Figure 2D, E). GPX4 is the primary endogenous defense against ferroptosis. GPX4 abundance decreased from day 14 onward, which was reversed by the pretreatment with Fer-1 (Figure 2F). These data demonstrate that Fer-1 attenuates ferroptosis-associated changes in stored RBC.

Figure 2. Ferrostatin-1 reduces ferroptosis-associated changes in red blood cells.

Suspended RBC were randomly divided into a control group, a group treated with a low concentration of Fer-1 (100 nM), and a group treated with a high concentration of Fer-1 (500 nM). The RBC were pretreated with Fer-1 at the beginning of storage. Samples were collected on days 0, 7, 14, 21, 28, and 35 of storage (No.=6 biological replicates per group). (A) The level of reactive oxygen species in RBC was assayed. (B, C) Accumulation of lipid peroxides in RBC was measured by 4-HNE and MDA assays. (D) SOD enzyme activity in RBC was measured. (E) GSH-Px enzyme activity in RBC was measured. (F) GPX4 protein on the RBC membrane was detected by western blot and statistically analyzed. *p<0.05, **p<0.01, ***p<0.001 vs day 0 within the same group (paired Student’s t-test); ^#p<0.05, ^##p<0.01 vs control at the corresponding timepoint (two-way analysis of variance). 4-HNE: 4-hydroxy-2-nonenal; Fer-1: ferrostatin-1; GPX4: glutathione peroxidase 4; GSH-Px: glutathione peroxidase; MDA: malondialdehyde; ROS: reactive oxygen series; SOD: superoxide dismutase. See also Online Supplementary Figure S1.

Ferrostatin-1 treatment protects red blood cells from storage lesion

We next detected RBC storage lesion-related indicators to determine whether the inhibition of ferroptosis further attenuated cell damage. The morphology of erythrocytes was observed under a light microscope. As shown in Figure 3A, compared with the control group, Fer-1 pretreatment reduced the proportion of RBC with a bnormal morphology, including echinocytes, sphero-echinocytes, and spherocytes. To evaluate cell damage of RBC, hemolysis, LDH release, osmotic fragility and protein level of Band 3 on the membrane were determined. As expected, Fer-1 reduced hemolysis on days 28 and 35 of storage (Figure 3B). The release of LDH from RBC is a key indicator of impaired membrane integrity and hemolysis. Fer-1 treatment reduced LDH release on days 14, 21, 28, and 35 (Figure 3C). Osmotic fragility directly reflects membrane vulnerability. At day 35 of storage, cells treated with Fer-1 displayed greater resistance to osmotic stress than the vehicle control group (Figure 3D). Meanwhile, higher levels of Band 3 protein on the erythrocyte membrane were observed in the Fer-1-treated group (Figure 3E). Overall, these data demonstrate that inhibiting ferroptosis with Fer-1 is associated with preserved RBC morphology, reduced hemolysis, and maintained membrane integrity.

Figure 3. Ferrostatin-1 treatment protects red blood cells from storage lesion.

Samples were collected as described in Figure 2 (No.=6). (A) RBC were stained with Wright’s stain. Scale bar, 10 μm. (B) The rate of hemolysis of suspended RBC was measured. (C) Supernatant of suspended RBC was collected for assay of lactate dehydrogenase. (D) Osmotic fragility at day 35 of storage. (E) Band 3 protein level was determined by western blot and statistically analyzed. NS: not significant, *p<0.05, **p<0.01, ***p<0.001 vs day 0 within the same group (paired Student’s t-test); ^#p<0.05, ^##p<0.01 vs control at the corresponding time point (two-way analysis of variance). Ctrl: control; Fer-1: ferrostatin-1; LDH: lactate dehydrogenase; NaCl: sodium chloride; RBC: red blood cells.

Salvianolic acid B treatment correlates with reduced ferroptosis-associated changes in red blood cells

Sal B regulates the progression of various diseases through its antioxidant properties. To explore the effects of Sal B on ferroptosis during storage, Sal B was added directly into the MAP additive solution. Ferroptosis indicators were measured after cell collection. The results showed that Sal B treatment significantly reduced intracellular ROS accumulation, starting from day 14 (Figure 4A). Concurrently, Sal B potentiated the activities of key antioxidant enzymes, GSH-Px and SOD, and increased the abundance of GPX4 (Figure 4B–D). Furthermore, a decrease in the accumulation of lipid peroxidation end products (4-HNE and MDA) was observed in the Sal B-treated group (Figure 4E, F). Collectively, these results demonstrate that Sal B confers a robust protective effect against ferroptosis-associated changes in stored RBC, an effect that correlates with preserved GPX4 level and strengthened cellular antioxidant defense.

Figure 4. Salvianolic acid B treatment correlates with reduced ferroptosis-associated changes in red blood cells.

The RBC were pretreated with 1 μM Sal B at the beginning of storage. Samples were collected as above (No.=6). (A) Flow cytometry histograms of ROS assays (left panel) and statistic analysis (right panel). (B, C). SOD and GSH-Px enzyme activity in RBC was measured. (D) GPX4 protein level was detected by western blot and statistically analyzed. (E, F) Accumulation of lipid peroxides in RBC was measured by 4-HNE and MDA assays. *p<0.05, **p<0.01, ***p<0.001 vs day 0 within the same group (paired Student’s t-test); ^#p<0.05, ^##p<0.01 vs control at the corresponding timepoint (two-way analysis of variance). 4-HNE: 4-hydroxy-2-nonenal; Ctrl: control; Fer-1: ferrostatin-1; GPX4: glutathione peroxidase 4; GSH-Px: glutathione peroxidase; MDA: malondialdehyde; RBC: red blood cells; ROS: reactive oxygen series; Sal B: salvianolic acid B; SOD: superoxide dismutase.

Salvianolic acid B treatment correlates with reduced storage lesion

Building on these findings, we propose that Sal B exerts protective effects against storage lesion. To validate this hypothesis, we examined the damage to RBC treated with Sal B. Wright’s staining of RBC smears revealed that, compared with the control RBC, the Sal B-treated RBC demonstrated fewer schistocytes and acanthocytes (Figure 5A). Flow cytometry was used to determine both the size and complexity of erythrocytes. Analysis of cell size and granularity uniformity, combined with Wright’s staining results, showed that RBC volume gradually decreased with prolonged storage time, which was associated with a relative loss of the biconcave disc structure. In contrast, cells treated with Sal B maintained their biconcave disc morphology better compared to the blank control group (Figure 5A–C). Subsequently, the rate of RBC hemolysis was measured, and it was found that Sal B treatment significantly reduced the hemolysis rate on days 28 and 35 (Figure 5D). Meanwhile, to clarify the extent of loss of RBC membrane integrity, LDH release was assessed and the data showed that Sal B could inhibit LDH release (Figure 5E). Following this, indicators of RBC storage lesion were examined. The results indicated that on day 35, the addition of Sal B showed higher resistance to osmotic stress, reduced microvesicle accumulation from days 21 to 35, and decreased PS exposure and Band 3 protein levels from days 14 to 35 (Figure 5F–I). Together, these results show that the protective effect of Sal B on RBC is associated with reduced markers of ferroptosis and storage lesion.

Figure 5. Salvianolic acid B protects red blood cells against storage lesion.

Samples were collected as above (No.=6). (A) RBC were performed Wright’s stain. Scale bar, 10 μm. (B, C) Statistical evaluation of cell size (forward scatter) and granularity (side scatter) of RBC detected by flow cytometry. (D) Rate of hemolysis of suspended RBC was measured at different storage timepoints. (E) Lactate dehydrogenase in the supernatant of RBC was determined. (F) Osmotic fragility at day 35 of storage. (G) Concentration of microvesicles in the supernatant of RBC was detected using flow cytometry, and statistic analysis was performed. (H) FACS histogram (left panel) of RBC stained by annexin V-phycoerythrin and statistical analysis (right panel). (I) Band 3 protein was detected by western blot and statistically analyzed. *p<0.05, **p<0.01, ***p<0.001 vs day 0 within the same group (paired Student’s t-test); ^#p<0.05, ^##p<0.01 vs control at the corresponding time point (two-way analysis of variance). RBC: red blood cells.

DISCUSSION

Currently, improving preservation solutions is an efficient strategy for promoting the quality and extending the storage duration of RBC^34,35. Here, our data demonstrate that Sal B, a natural polyphenolic compound, mitigates storage-associated oxidative and membrane injury in parallel with GPX4 preservation and reduced lipid peroxidation.

Our findings should be interpreted in the context of recent landmark studies on ferroptosis in RBC storage biology. D’Alessandro et al. demonstrated that genetic variations in ferroptosis-related genes are strongly associated with a ferroptosis-like process of lipid peroxidation and hemolysis in stored RBC, establishing that storage lesion shares mechanistic features with ferroptosis⁴. Our results are consistent with their findings, as we observed time-dependent accumulation of lipid peroxidation products (MDA and 4-HNE) and decreased GPX4 protein level during storage. Page et al. identified GPX4 as a genome-wide significant locus associated with oxidative hemolysis, providing human genetic evidence that complements our pharmacological approach²⁴. Stephenson et al. provided causal evidence, using erythroid-specific Gpx4 knockout mice, that GPX4 deficiency exacerbates lipid peroxidation and impairs posttransfusion recovery, and also identified a human GPX4 missense variant (rs73507255) associated with lower hemoglobin increments in transfusion recipients²³. Our findings align with those of recent landmark studies demonstrating that a ferroptosis-like process of lipid peroxidation contributes to RBC storage lesion. While these studies focused on genetic determinants, our study provides complementary pharmacological evidence that preserving GPX4 with Sal B intervenes in this process.

The importance of oxidative lipid remodeling in RBC injury has been further highlighted by Nemkov et al., who demonstrated that ultra-endurance running triggers inflammation-associated protein, lipid, and purine oxidation in RBC, leading to their impaired deformability and extravascular clearance, features that parallel the molecular signatures of the storage lesion³⁶. This convergence suggests that common oxidative pathways contribute to RBC injury across diverse physiological and ex vivo stressors.

GPX4 is a core protein that protects against ferroptosis, responsible for reducing toxic lipid peroxides²¹. Our data show that Sal B treatment correlates with preserved GPX4 abundance, enhanced antioxidant enzyme activities, and reduced lipid peroxidation, effects comparable to those of the ferroptosis inhibitor Fer-1³⁷. The present study reveals that Sal B alleviates storage lesion in conjunction with GPX4 preservation. Nevertheless, these results do not provide evidence that GPX4 is the specific mediator underlying the protective effects of Sal B. Future pharmacological interventions targeting GPX4 and loss-of-function studies are warranted. Consistent with reports that GPX4 could be degraded via autophagy^17,38, GPX4 decreased during RBC storage in this study. However, as mature RBC lack autophagic organelles, the precise mechanism of GPX4 loss in stored erythrocytes remains to be elucidated.

The PS exposure rate in our control group at day 35 (approximately 3.5%) is comparable to the values reported by She et al. (approximately 4.0%) using the same MAP additive solution¹³. RBC additive solution also affects PS exposure levels. For example, Lu et al. reported 4.5% PS exposure in SAGM at day 42³⁹, while Dinkla et al. reported less than 1% in SAGM⁴⁰. Differences in detection methods³⁹, gating strategies³³, and donor variability⁴⁰ may also contribute to inter-study differences. Despite differences in absolute values, the time-dependent increase in PS exposure during storage is consistent across studies. Thus, the relative reduction in PS exposure by Sal B was consistent, supporting its protective effect regardless of baseline variability. The decline of Band 3 abundance is consistent with reported storage-induced degradation, with western blot showing a progressive decrease in full-length Band 3⁴¹. The changes in other parameters of RBC storage lesion and ferroptosis (hemolysis⁴², ROS¹³, 4-HNE¹², MDA, SOD and GSH-Px⁴³, GPX4²²) are all consistent with the well-established oxidative stress trends during RBC storage.

The present data reveal a correlation between ferroptosis-associated markers and RBC injury readouts, and further show that Fer-1 mitigates these indicators of RBC damage. However, without genetic loss-of-function studies, we cannot conclude that ferroptosis is the definitive mechanism. We therefore interpret our findings as evidence that ferroptosis-associated pathways contribute to storage lesion.

Previous studies have attempted to add components such as ethanol, cerium oxide, and naringin to preservation solutions to mitigate the storage lesion^9,10,13. In this study, Sal B treatment was associated with reduced ferroptosis markers, suggesting that it may act upstream of oxidative damage. Further studies are warranted to decipher the key signaling pathways involved by which Sal B regulates ferroptosis to protect RBC.

From a translational perspective, Sal B may warrant further evaluation as a potential additive for blood preservation solutions. To assess its biocompatibility, we further evaluated the safety of Sal B on vascular endothelial cells. Supernatant from RBC treated with 1 μM Sal B for 35 days did not affect the viability of human umbilical vein endothelial cells (data not shown), supporting the safety of residual Sal B in RBC units. Notably, Sal B is a component of clinically approved intravenous preparations in China (e.g., Danhong injection) with an established safety profile^44,45. Preclinical studies have demonstrated that Sal B exhibits dose-proportional pharmacokinetics in rats at intravenous doses ranging from 10 to 62.5 mg/kg⁴⁶ and effectively attenuates myocardial ischemia-reperfusion injury without significant toxicity in mice at 60 mg/kg⁴⁷, with these doses being substantially higher than the residual concentration expected upon transfusion. However, compatibility with blood banking workflows, posttransfusion recovery, in vivo survival, and product-scale stability have not been assessed.

While this study primarily focuses on the GPX4 pathway, ferroptosis may involve other regulatory factors (e.g., ACSL4, FSP1)^48,49. Metabolomic studies have shown that donor genetics regulate kynurenine metabolism, impacting RBC hemolysis both in vivo and in vitro⁵⁰. The functional network of Sal B warrants further investigation using -omics approaches. Moreover, functional validation in animal transfusion models is needed to assess the in vivo survival and safety of Sal B-treated RBC.

CONCLUSIONS

In summary, our findings are consistent with the hypothesis that ferroptosis-associated lipid peroxidation may contribute to RBC storage lesion, and that Sal B mitigates storage lesion in association with GPX4 preservation and reduced lipid peroxidation. However, further studies are needed to establish causality. These results suggest that adding Sal B to preservation solutions may represent a candidate approach for future evaluation.