All human procedures were performed in accordance with the Guidelines of the Human Research Ethics Committee of the Shandong Second Provincial General Hospital Affiliated to Shandong University (Jinan, China). The Human Research Ethics Committee of the Shandong Second Provincial General Hospital Affiliated to Shandong University (Jinan, China) approved the experimental protocols (approval no. XYK20181224). All participants agreed to use their samples for scientific research, and informed consent was obtained. In accordance with the Institutional Review Board under Food and Drug Administration regulations (24), venous blood was collected via puncturing the brachiocephalic vein from 10 healthy male donors aged 21 to 32 years old, after they had provided consent, and the blood was collected in heparinized tubes on ice. Adults with diabetes, hyperlipidemia or other diseases that affect blood lipid levels were excluded. Lipoproteins were isolated from plasma using sequential ultracentrifugation with a Beckman TL-100 tabletop ultracentrifuge (Beckman Coulter, Inc.), which was extracted from blood supernatant by centrifuging at 1,500 x g for 20 min at 4˚C (25). The lipoproteins were treated with 0.3 mM EDTA in 1X PBS (pH 7.4) overnight at 4˚C and subsequently sterilized using a 0.22-µM filter (MilliporeSigma). The Folin Lowry method was used to calculate the protein concentration in the lipoproteins. After dialysis using 5 µM copper sulphate at 4˚C overnight, ox-LDL was sampled from the native LDL immediately, as previously described (26). Thiobarbituric acid reactive substances (TBARS; Sigma-Aldrich; Merck KGaA) were used to monitor the degree of LDL oxidation and to ensure ox-LDL quality and reproducibility using a microplate reader at a wavelength of 532 nm (BioTek Instruments, Inc.) (27). Specifically, the TBARS value was maintained at 40-50 nmol malondialdehyde/mg protein. There were no detectable TBARS in the native LDL. All product was then stored at 4˚C and used within 1 month of preparation.

All animal procedures were performed in accordance with the Guidelines of the Animal Care Committee of the Shandong Second Provincial General Hospital Affiliated to Shandong University (Jinan, China). The Animal Care Committee of Shandong Second Provincial General Hospital Affiliated to Shandong University approved the experimental protocols (approval no. XYK20181225). All mice were maintained at room temperature with 40-60% humidity and a 12 h light/dark cycle, with ad libitum access to food and water.

A total of 10 randomized, age-matched wild-type (WT) male and female C57BL/6 mice (weight, 20±3 g; age, 4-6 weeks; Jackson Laboratory) were administered 50 µg prepared ox-LDL daily via tail vein injections for 3 days, as described previously (7). A total of 10 25±5 g LDLR^-/- C57BL/6 male and female mice (age, 4-6 weeks) were also obtained from Jackson Laboratory. The genotyping for WT and LDLR^-/- mice were further confirmed by Southern blots. All mice were fed a normal diet (ND) until 8 weeks of age, after which they were fed a high-fat diet (HFD; 17% anhydrous milk fat and 0.2% cholesterol; Harlan Laboratories, Inc.) for 6 months to induce hyperlipidemia. Age-matched male and female WT C57BL/6 mice on an HFD or ND, and LDLR^-/- male and female mice fed with ND were used as the controls.

After 6 months of HFD treatment, isoflurane was used to induce (3%) and maintain (1.5%) anesthesia in mice for blood collection (300-500 µl) via cardiac puncture. Animals were then immediately euthanized using CO₂ (50-70% of the chamber volume per min) and death was confirmed by ascertaining cardiac and respiratory arrest or by observing fixed and dilated pupils. Aorta and BM were collected after confirming death of animals.

After 6 months of HFD treatment, blood plasma samples from all mice were collected for lipid profile testing. Plasma (40 µl) was tested using the Cholestech LDX lipid profile cassette (Alere 10-989; Central Infusion Alliance, Inc.) for each test coupled with the Alere Cholestech LDX system (Alere Cholestech). Total cholesterol (TC), triglyceride (TRG), LDL, high density lipoprotein (HDL), non-HDL and the TC/HDL ratio were measured. Mouse aortas were also isolated for the atherosclerotic plaque formation test. Red oil (MilliporeSigma) was used to stain the atherosclerotic plaque at room temperature for 5 min, where the plaque area against the total inner surface of aorta was calculated as previously described (28).

BM and blood cells were harvested to observe the effects of ox-LDL and hyperlipidemia on the population of blood and BM EPCs in male and female mice. After eliminating red blood cells (RBCs) with RBC lysis buffer (Thermo Fisher Scientific, Inc.), a BD™ LSRII system (BD Biosciences) was used to perform multicolor analysis for BM and blood EPCs.

An endothelial cell marker combined with a stem cell marker, including CD34⁺/fetal liver kinase-1 (Flk-1)⁺, Stem cell antigen-1 (Sca-1)⁺/Flk-1, c-Kit⁺/CD31⁺ and CD34⁺/CD133⁺, were used to identify EPCs as previously described (29). Functional EPCs express the endothelial markers Flk-1 and CD31(29). BM and blood EPCs with a total of 50,000 cells in each sample were carefully analyzed and described (Fig. S1). All cell populations were carefully compensated (each cell population percentile was confirmed further using single antibody staining) and determined using flow cytometry, as previously described (30-36). Flk-1 APC-Cy™7 (cat. no. 561252) antibody was obtained from BD Biosciences and CD34 FITC (cat. no. 11-0341-82) from eBioscience (Thermo Fisher Scientific, Inc.). Sca-1 AF700 (cat. no. 108142), c-Kit-APC (cat. no. 105812), CD31-PE-Cy7 (cat. no. 102418) and CD133-PE (cat. no. 141204) were purchased from BioLegend, Inc. All antibodies were diluted 1:100.

Mouse BM and blood were harvested following intravenous injection of 50 µg ox-LDL into each mouse for 3 days, as described previously (7). For LDLR^-/- mice, BM and blood were harvested after 6 months of HFD feeding. RBC lysis buffer was used to remove all RBCs (37). A total of four groups of BM and circulating EPCs were selected for intracellular ROS detection. The mean of the four groups of ROS levels in EPCs were statistically analyzed.

Intracellular ROS generation was measured using FITC conjugated ROS Detection Reagent (cat. no. D399; Invitrogen; Thermo Fisher Scientific, Inc.) as previously described (38). A total of 1x10⁶ were incubated at 37˚C for 10 min with 5 µg/ml reagent. All labeled cells were washed with PBS twice before suspending in warm PBS. Flow cytometry was used for analysis. BD™ LSRII (BD Biosciences) at a wavelength of 525 nm was used to calculate the positively fluorescent cells, as previously described (39).

Mouse blood samples were harvested after 6 months of HFD or ND treatment. The plasma was obtained from the blood samples after centrifugation at 300 x g for 20 min at 4˚C. The plasma levels of the proinflammatory cytokines IL-1β (cat. no. 432601) and TNF-α (cat. no. 430904) were evaluated using ELISA kits from BioLegend, Inc. according to the manufacturer's protocols.

Data are presented as the mean ± standard deviation, and analyzed using an unpaired Student's t-test (two-sided) for comparisons between two groups of data, or a two-way ANOVA followed by a Bonferroni post hoc test for comparing the subgroups of data between male and female groups to minimize type I errors as appropriate in GraphPad Prism version 4 (GraphPad Software, Inc.). A two-tailed P<0.05 was considered to indicate a statistically significant difference.

To study the effects of differences in sex on hyperlipidemia, male and female LDLR^-/- mice were fed a HFD for 6 months. The TC, TRG, LDL and non-HDL lipoprotein levels, and the TC/HDL ratio were markedly increased in male and female hyperlipidemic LDLR^-/- mice fed a HFD compared with their respective control groups, confirming that the hyperlipidemic mouse model was successfully established (Table I). Of note, the lipid levels in female hyperlipidemic mice was notably lower compared with that in male hyperlipidemic mice (Table I). In addition, a number of atherosclerotic plaques were present in the aorta of the male hyperlipidemic mice, whereas plaque severity was significantly decreased in the female mouse group (P<0.01; Fig. 1).

Persistent endothelial cell dysfunction or injury promotes the progression of atherosclerosis and coronary heart disease (4). Therefore, the EPC profiles were examined in the BM of male and female LDLR^-/- mice. Hyperlipidemia did not change the BM Sca-1⁺/Flk-1⁺, c-Kit⁺/CD31⁺ and CD34⁺/CD133 levels in male mice (Fig. 2B-D), which only exhibited significantly decreased CD34⁺/Flk-1⁺ levels (P<0.05; Fig. 2A). By contrast, these BM EPC cell populations, except for those expressing c-Kit⁺/CD31⁺, were found to be significantly increased in female LDLR^-/- mice fed a HFD compared with female LDLR^-/- mice fed a ND (P<0.05) and those in male LDLR^-/- mice fed a HFD (P<0.05; Fig. 2A, B and D). The c-Kit⁺/CD31⁺ cell population in female hyperlipidemic mice was lower compared with that in the male mice, which may be due to the low basal numbers of this cell population in female mice (Fig. 2C).

The circulating EPC numbers were also measured in both LDLR^-/- male and female mice. As shown in Fig. 2E-H, the numbers of circulating EPCs were significantly reduced in both male and female hyperlipidemic mice (P<0.05) compared with those in their control groups fed a ND. The numbers of blood EPCs, including those expressing CD34⁺/Flk-1⁺, Sca-1⁺/Flk-1⁺, c-Kit⁺/CD31⁺ and CD34⁺/CD133⁺, were significantly higher (P<0.05) in the hyperlipidemic female mice compared with those in the respective male counterparts (Fig. 2E-H).

To investigate the cause of the differences in EPC numbers between male and female mice with hyperlipidemia, the BM and blood EPC intracellular ROS levels were determined. Although there were no intracellular ROS changes in BM EPCs in both male and female mice, a significantly increased ROS level was observed in the blood EPCs of male hyperlipidemic mice compared with that in their corresponding control group fed a ND (P<0.01; Fig. 3A and B). The blood EPC intracellular ROS levels were significantly decreased in female LDLR^-/- mice fed with HFD compared with that in the female LDLR^-/- mice fed with ND (P<0.05), and was also significantly lower compared with that in the male LDLR^-/- mice fed a HFD (P<0.05; Fig. 3B).

The plasma inflammatory factor TNF-α and IL-6 levels were next measured in both male and female mice. After feeding the mice with a HFD for 6 months, except for those in the female WT mice, the TNF-α and IL-6 levels were found to be significantly increased in both female and male LDLR^-/- mice compared with those in the LDLR^-/- mice fed a ND (P<0.05; Fig. 4A and B). Of note, the TNF-α and IL-6 levels in all HFD-fed female LDLR^-/- mice were significantly lower compared with those in the male LDLR^-/- mice fed a HFD (P<0.05; Fig. 4A and B).

Ox-LDL treatment was used as the primary hyperlipidemic mediator to treat both male and female WT mice for 3 days prior to measuring their EPC numbers and intracellular ROS levels. In ox-LDL-treated male mice, the BM CD34⁺/Flk-1⁺, c-Kit⁺/CD31⁺ (Fig. 5A and C) and circulating Sca-1⁺/Flk-1⁺ and c-Kit⁺/CD31⁺ (Fig. 5F and G) cell populations were significantly decreased (P<0.05) compared with mice without ox-LDL treatment. However, BM and circulating CD34⁺/CD133⁺ (Fig. 5D and H) and circulating CD34⁺/Flk-1⁺ cell numbers (Fig. 5E) were significantly increased (P<0.05). There was little to no change in the entire EPC population (Fig. 5A, B and D-H), except for the fact that the BM c-Kit⁺/CD31⁺ population (Fig. 5C) was significantly increased (P<0.05) in female mice following ox-LDL treatment compared with female mice without ox-LDL treatment (Fig. 5). The female BM and circulating CD34⁺/Flk-1⁺ (Fig. 5A and E), CD34⁺/CD133⁺ (Fig. 5D and H) and BM c-Kit⁺/CD31⁺(Fig. 5C) populations were significantly increased (P<0.05) compared with those in male mice treated with ox-LDL.

Subsequently, the BM and blood EPC intracellular ROS levels were measured. Similar to that in the hyperlipidemic mice, there were no changes in ROS levels in the BM in both male and female mice with or without ox-LDL treatment (Fig. 6A). However, the intracellular blood ROS levels were significantly elevated in ox-LDL-treated male mice (P<0.01) compared with that in the ox-LDL-treated female mice (Fig. 6B).