Venous blood samples, taken from the forearm vein of healthy adult volunteers, were treated with the anticoagulant sodium citrate (15 IU/mL). The collected blood was centrifuged at 800 × g for 3 min to separate the plasma and red blood cells. Then, the red blood cells were washed with phosphate-buffered saline (PBS), the supernatant was completely removed with a pipette, and the pellet was centrifuged 3 times for 2 min each time. Finally, 0.45 L/L hematocrit was retained, to which citrate phosphate dextrose solution was added. This solution can provide nutrients to the red blood cells and maintain the cellular osmotic pressure (main ingredients: citric acid, sodium citrate, glucose, and phosphate). The resulting mixture was a suspension of red blood cells.

In a subsequent experiment, we established six experimental groups: the healthy red blood cells group (control), the hypoxia group (hypoxia), the healthy red blood cells add resveratrol group (C-Res), the healthy red blood cells add salidroside group (C-SAL), the hypoxic red blood cells add resveratrol group (H-Res), and the hypoxic red blood cells add salidroside group (H-SAL), with the salidroside group serving as the active control and the hypoxia group serving as the negative control. The working concentration of resveratrol and salidroside was 70 μM (purchased from Beijing Solarbio Science & Technology Co., Ltd.).

Blood is a non-Newtonian fluid, and the viscosity of blood at high shear rates is related to erythrocyte deformability, whereas at low shear rates, blood viscosity is related to erythrocyte aggregation (Shiga et al., 1990; Beris et al., 2021). Isolated plasma was added to washed red blood cells from each experimental group to simulate whole blood for whole blood viscosity testing (0.45 L/L hematocrit). The viscosity of each sample suspension was measured by an automatic hemorheology rapid detector (FASCO-5010DX, Chongqing Weiduo Technology Co., Ltd., China) to determine the deformability and aggregation of red blood cells. The hemorheometer was heated to and maintained at 37°C, after which the channel was cleaned. Samples from each group were then placed into tubes for whole blood viscosity determination. At the end of the test, the channel was cleaned again.

Atomic force microscopy (AFM; BRUKER, Dimension ICON010803, United States) was used to measure the Young’s modulus and surface roughness of the cells. Pretreated erythrocytes were resuspended in PBS and added to polylysine-treated glass slides. Peak force quantitative nanomechanical mapping (PFQNM) mode was selected for AFM calibration. Measurements were started after a single red blood cell was observed under the microscope using a soft-cone silicon nitrided cantilever with an elastic constant of 0.92 N/m (F), a probe angle of 18°, and a probe radius of 23.1 nm. In PFQNM mode, three kinds of images can be obtained: a high sensor, peak force error and DMT modulus. The Young’s modulus of the cells was determined in the DMT Modulus channel using Nanoscale analysis software. In addition, the 5 μm × 5 μm region of the cell was selected in the DMT Modulus channel to determine the surface roughness.

For each experiment, 30 cells from each group were randomly examined under a microscope.

AFM was performed based on the Sneddon model (Sneddon, 1965) formula: F = π 2 E 1 − v 2 tan   α δ 2 where E is the elastic or Young’s modulus, v is the Poisson ratio (assumed to be 0.5), α is the opening angle and δ is the indentation depth. Here, F = 0.92 N/m.

The surface roughness was calculated with the following formula (https://www.bruker.com/): R q = ∑ Z i 2 N where R_q is the mean square root roughness, N is the number of sampling points, and Z is the height of the ith sampling point.

For the detection of free hemoglobin (FHb) in the supernatants from different treatment groups, the absorbance at 510 nm was measured after different color development reactions. First, the relationship between the optical density and the concentration of the standard substance was determined. Then, 0.15 mL of supernatant was collected from each group, and 2.5 mL of chromogenic agent was added. In the control group, 0.15 mL of double distilled water and 2.5 mL of chromogenic agent were added. Each mixture was heated in a water bath at 37°C for 20 min, after which the optical density was measured at a wavelength of 510 nm. The free hemoglobin concentration was calculated according to the standard curve. Then FHb concentration was experimentally calculated according to: FHb (mg/L) = 126.03*(OD_treatment-OD_blank) + 3.5041, 126.03, and 3.5041 were based on a standard curve.

The method for detecting MetHb and MDA were consistent with that described in Section 2.2.2.

All the parameters and calculation methods used were measured using kits provided by Nanjing Jiancheng Bioengineering Institute (Njjcbio.com/product.asp) according to the manufacturer’s instructions.

Differently treated erythrocytes were centrifuged at 800 × g for 5 min, after which the supernatant was discarded and the cells were collected and resuspended in PBS. Then, 1–5 × 10⁵ cells were taken and centrifuged again at 800 × g for 5 min, and the supernatant was discarded. Cell survival was detected using an Annexin V-FITC apoptosis detection kit (China, http://www.solar.bio.com). Briefly, 500 μL of binding solution was added to the cellular precipitate and mixed gently. Then, 5 μL of Annexin V-FITC was added, and the mixture was incubated for 10 min at 25°C in the dark. The CytoFLEX assay was performed immediately (Beckman Coulter, United States).

Both the reduced glutathione (GSH) molecules in the erythrocytes and the sulfhydryl groups in the erythrocyte membrane proteins can react with 5,5'-dithiol-di-nitrobenzoic acid (DTNB), and the resulting compounds have a maximum absorption peak at 412 nm. The contents of GSH and sulfhydryl groups in membrane proteins can be measured via visible spectrophotometry (Aksenov and Markesbery, 2001).

Cells were fixed by mixing 4% paraformaldehyde and 0.01% glutaraldehyde at a 1:1 ratio and then washed with PBS at 1200 rpm 3 times for 2 min each time. Then, 0.01% Triton X-100 was used to permeabilize the red blood cell membranes, after which the cells were washed with PBS at 1200 rpm 3 times for 2 min each time. Bovine serum albumin (BSA) was used to block the other sites on the membrane so that subsequent antibodies could bind specifically. After the corresponding antibody was selected, the cells were incubated with mouse monoclonal spectrin antibody and rabbit monoclonal band 3 antibody (Abcam, UK, No. ab2808; Abcam, UK, No. ab108414) diluted 1:200 in 5% BSA for 2 h at room temperature. The cells were washed with PBS three times (two minutes each time). Then, the cells were incubated with secondary antibodies (Alexa Fluor 488 goat anti-mouse, No. A0428 and Alexa Fluor 488 goat anti-rabbit, No. A0423; Beyotime, China), diluted with 5% BSA at a 1:400 dilution, and washed three times with PBS. Finally, the cells were imaged to observe fluorescence using an Olympus IX71 microscope equipped with a 63/1.25 oil immersion objective and a CCD camera (Olympus, Tokyo, Japan).

Rinsed erythrocytes were placed in a cold lysis buffer mixture containing 1 mM MgCl₂, 1 mM ethylenediaminetetraacetic acid sodium, 10 mM Tris-HCl, and 0.1 mM phenylmethylsulfonyl fluoride (pH 8.0) at 4°C for hemolysis. Erythrocyte membrane proteins were collected after treatment with lysis buffer and centrifugation at 12,000 rpm for 10 min, and protein concentrations were quantified by the Bradford method. The proteins were solubilized in SDS‒PAGE loading buffer with or without dithiothreitol (DTT) at a volume ratio of 4:1, and each mixture was incubated at 100°C for 10 min.

Clear bands were obtained by electrophoresis on a preconfigured polyacrylamide separation gel and a concentrated gel. The proteins were subsequently transferred from the gel to a polyvinylidene fluoride membrane and immunostained with primary antibodies (mouse monoclonal to spectrin, No. S3396, Sigma, Germany, diluted 1:500; rabbit monoclonal to band 3, No. ab108414, Abcam, UK, diluted 1:5000; rabbit monoclonal to β-actin, No. 380624, ZENBIO, China, diluted 1:5000; and rabbit monoclonal to band 3 phospho Y359 and phospho Y21, Abcam, UK, diluted 1:20,000 and 1:10,000. No. ab77236 and No. ab125070). The sections were incubated with anti-mouse (No. 701051) or anti-rabbit (No. 511203) secondary antibodies (ZENBIO, China) diluted 1:10,000, and the proteins were detected via chemiluminescence (Thermo Scientific, United States). Quantification of the immunocomplexes was conducted by lengthwise scanning densitometry using a gel analyzer image processing program (Azure Biosystem, United States).

The experimental results are expressed as the mean ± standard deviation. One-way ANOVA was used to compare more than three groups.

Origin 9.0 software was used for statistical analysis. Quantity One was used for gray analysis. FlowJo was used for cell eryptosis analysis.

When a red blood cell suspension was placed in a hypoxic environment, echinocytes (spiculated cells) appeared. To determine an appropriate treatment concentration of resveratrol, normal red blood cells were incubated with different concentrations of resveratrol (Figure 1A), and then, some of the cells were subjected to hypoxic conditions (Figure 1B) and observed under a light microscope to determine the resveratrol concentration that had a protective effect on red blood cell morphology.

As shown in Figure 1C, when the resveratrol concentration was increased to 100 μM, the number of echinocytes increased to more than 20%, and at 150 μM resveratrol, 40% were echinocytes. Compared with the normal group, high concentrations of resveratrol (100 μM and 150 μM) were not conducive to protecting cell morphology (p < 0.01, p < 0.001). Compared with the hypoxia group, 70 μM resveratrol protected cell morphology (p < 0.01). Moreover, there were no significant differences between the low (10 μM and 40 μM) and high (100 μM and 150 μM) resveratrol concentration groups and the hypoxia group.

To determine the optimal working concentration of resveratrol, the levels of CAT, SOD, MDA, and MetHb in normal and hypoxic erythrocytes incubated with resveratrol were measured.

Compared with the control group, the levels of CAT (Figure 2A) and SOD (Figure 2B) were significantly lower after hypoxia treatment (p < 0.001; p < 0.01), but the MDA (Figure 2C) and MetHb (Figure 2D) levels were higher (p < 0.01; p < 0.001). After incubating normal erythrocytes with different concentrations of resveratrol, the CAT content gradually decreased with increasing concentration (100 μM and 150 μM) and was significantly different from that in the control group (p < 0.05); however, the SOD content did not change significantly with increasing resveratrol concentration. With respect to the MDA content, when the concentration of resveratrol was increased to 150 μM, the MDA content increased and was significantly different from that in the control group (p < 0.05). When the concentration of resveratrol was 100 μM or 150 μM, the MetHb content was significantly higher than that in the control group (p < 0.05; p < 0.01).

After hypoxia treatment after incubation with 70 µM resveratrol, the levels of CAT and MetHb were significantly different from those in the hypoxia group (p < 0.01; p < 0.001). The levels of SOD and MDA after treatment with 70 μM and 100 μM resveratrol were significantly different from those in the hypoxia group (p < 0.01; p < 0.01; p < 0.05; p < 0.01).

Considering the changes in red blood cell morphology and CAT, SOD, MDA, and MetHb levels, we selected 70 μM resveratrol as the optimal concentration for use subsequent experiments.

Whole-blood viscosity can reflect the aggregation and deformability of red blood cells. According to the results in Figure 3A, we found that the simulated whole-blood viscosity of the hypoxia group was significantly greater than that of the normal group under both high (p < 0.05), medium (p < 0.001), and low shear rate (p < 0.05) conditions. When resveratrol and salidroside were added, the whole-blood viscosity did not significantly change compared with that of the control group, and resveratrol achieved the same effect as salidroside.

To investigate the cellular mechanical properties of hypoxic erythrocytes after resveratrol treatment, 30 single cells from each sample were selected for AFM measurements, which were repeated three times. The whole experiment consisted of six independent replicate tests, and the resulting data collation is shown in Figure 3B. Neither the presence of resveratrol nor the presence of salidroside affected the mechanical properties of red blood cells. After hypoxia treatment, the Young’s modulus of the red blood cells decreased significantly (p < 0.001). However, after resveratrol or salidroside were added, the Young’s modulus increased (p < 0.01), and there was no obvious difference between them. The surface of the normal red blood cells was smooth. After hypoxia treatment, the surface roughness of the red blood cells increased, and after resveratrol and salidroside treatment, the surfaces of the red blood cells were smoother than those of the hypoxia group, as shown in Figure 3C.

Figure 3D presents the DMT modulus diagram, which shows representative images of single cells selected from each group.

The levels of FHb, MetHb, and MDA were measured to investigate the effect of resveratrol on hypoxic injury in erythrocytes. As shown in Figure 4, the contents of FHb (Figure 4A), MetHb (Figure 4B) and MDA (Figure 4C) did not change significantly in control red blood cells after the addition of resveratrol or salidroside. After hypoxia, the FHb, MetHb and MDA levels increased (p < 0.01; p < 0.001; p < 0.001). In the presence of resveratrol and salidroside, the FHb, MetHb, and MDA contents were lower than those in hypoxia group.

Flow cytometry was used to analyze whether resveratrol could protect red blood cells from eryptosis. The results (Figure 4D) show that when normal red blood cells were incubated with resveratrol and salidroside, there was no significant change in the number of eryptotic cells. After hypoxia, the percentage of cells binding Annexin V increased from 6.12% to 63.01%. However, under hypoxic conditions, resveratrol and salidroside significantly reduced the percentage of cells binding Annexin V from 63.01% to 24.44% and 41.86%, respectively.

Spectrin is the main skeletal component of the erythrocyte membrane that plays an important role in maintaining erythrocyte deformability. Immunofluorescence localization analysis (Figure 5A) showed that resveratrol and salidroside had no effect on the expression of the protein spectrin, although hypoxia decreased the average fluorescence intensity of the spectrin protein (Figure 5B; p < 0.01). After incubation with resveratrol, the fluorescence intensity of the cells was significantly greater than that in the hypoxia group (p < 0.05).

Western blotting confirmed the effects of hypoxia and resveratrol supplementation on the spectrin protein level (Figure 5C). Hypoxia led to a decrease in the spectrin protein content (Figure 5D; p < 0.001), but after the addition of resveratrol and salidroside to the hypoxia treatment group, the protein content of spectrin was significantly greater than that in the hypoxia group (Figure 5D; p < 0.01; p < 0.01), indicating that resveratrol could resist the damage to spectrin caused by hypoxic conditions.

Compared with those in the control group, the intracellular GSH and membrane-SH contents in the hypoxia group were significantly lower (Figure 6A, p < 0.01; Figure 6B, p < 0.001), but the GSH and -SH contents were significantly greater in the presence of resveratrol (p < 0.01; p < 0.01) and salidroside (p < 0.01; p < 0.01). After hypoxia treatment, there was no significant difference in the GSH or -SH content between the groups incubated with resveratrol and salidroside.

Resveratrol may regulate the deformability of hypoxic erythrocytes under hypoxic conditions by regulating the cross linking of band 3. To explore this possibility, we examined the status of band 3 after incubation under hypoxic conditions or with resveratrol in the presence or absence of DTT. Figure 6C–F shows that the band 3 protein content in the hypoxia group decreased significantly in the absence of DTT (p < 0.05, p < 0.001). Additionally, this protein was partially oxidized to a high polymer (p < 0.05, p < 0.05) but reduced back to a band between 90 and 100 kDa in erythrocytes after incubation with DTT. After incubation with resveratrol in the absence of DTT, the band 3 protein content was significantly greater than that in the hypoxia group (p < 0.05; p < 0.001) the polymers were reduced or absent with no significant difference compared with the crosslinked band 3 in the control group (Figures 6C–F). Hypoxia-associated oxidative stress leads to damage caused by the free sulfhydryl groups in membrane proteins, thereby inducing the oxidation of band 3 proteins through the formation of reducible intermolecular and/or intramolecular disulfide bonds.

Immunofluorescence localization analysis of the membrane proteins was also performed (Figures 6G, H). Compared with those in the control group, the relative fluorescence intensity of the band 3 protein in normal erythrocytes incubated with resveratrol and salidroside did not significantly change. After hypoxia, the relative fluorescence intensity of the band 3 protein on the erythrocyte membrane was significantly lower than that in the control group (p < 0.05; p < 0.01). After the addition of resveratrol, the relative fluorescence intensity of the band 3 protein was significantly greater than that in the hypoxia group (p < 0.05; p < 0.05). After the addition of salidroside, the relative fluorescence intensities of band 3 and the contractile protein were not significantly different from those in the hypoxia group. There was no significant difference between the resveratrol group and the salidroside group after hypoxia treatment.

Western blot analysis of tyrosine phosphorylation at positions 359 and 21 in the band 3 protein (Figure 7A) showed that band 3 phosphorylation was induced after erythrocytes were subjected to hypoxia (Figures 7B, C; p < 0.001; p < 0.001). After resveratrol or salidroside were added, compared with that in the hypoxia group, the phosphorylation of band 3 at position 21 was decreased (Figure 7B; p < 0.001, p < 0.001). However, after the addition of resveratrol or salidroside, there was no significant difference in phosphorylation at position 359 of the band 3 protein compared with that in the hypoxia group (Figure 7C).