Sex-specific alterations in blood-borne factors in physically inactive individuals are detrimental to endothelial cell functions

Abstract

Mechanisms underlying the protective effects of both habitual endurance exercise and the female sex on vascular function are incompletely understood. Blood-borne circulating factors, such as circulating microRNAs (ci-miRs), may partially explain these effects. Blood samples were obtained from young, healthy men and women who either habitually performed endurance exercise (endurance trained) or were relatively inactive (sedentary). Women were tested during the early follicular phase of the menstrual cycle or the placebo pill phase of oral contraceptive to control for estrogen. Cultured human umbilical vein endothelial cells (HUVECs) were exposed to participants’ serum in migration, proliferation, and reactive oxygen species (ROS) assays. Real-time quantitative polymerase chain reaction was used to quantify an initial array of 84 cardiovascular disease (CVD)-related ci-miRs, followed by validation of 10 ci-miRs. All participants were devoid of traditional CVD risk factors, and circulating estradiol concentration was not different between groups. Serum of endurance-trained women induced greater HUVEC migration compared with serum of sedentary women. HUVEC ROS production was greater in response to serum of sedentary men compared with serum of endurance-trained men and sedentary women. There were sex effects on the levels of nine ci-miRs, with greater levels in men, while ci-miRs-140–5p and 145–5p were also higher in sedentary compared with endurance-trained men and/or women. In a sex-specific manner, habitual endurance exercise was associated with beneficial effects of serum on HUVECs. Thus, alterations in circulating factors may contribute to the protective effects of habitual endurance exercise on vascular health. Additionally, sex had a greater impact than habitual activity level on the levels of vascular-related ci-miRs.

NEW & NOTEWORTHY Serum from sedentary women caused impaired endothelial migration, whereas serum from sedentary men elicited increased endothelial reactive oxygen species production as compared with serum from their endurance-trained counterparts. Select CVD-related circulating microRNAs (ci-miRs) were higher in men than women, while ci-miRs-140–5p and 145–5p were also higher in sedentary versus trained men and/or women. Our data suggest that alterations in circulating factors may contribute to the protective effects of habitual exercise and sex on vascular health.

INTRODUCTION

Physical inactivity is a major independent risk factor for cardiovascular disease (CVD), which remains the primary global cause of death (8, 71). The endothelium is a monolayer of cells lining the lumen of all blood vessels that is critical in maintaining proper vascular function and preventing atherogenesis (13). The effects of chronic inactivity manifest as impaired endothelial-dependent dilation and adverse structural remodeling of the arteries (10, 79, 87). Conversely, regular physical activity improves endothelial function and protects against CVD, with the most active individuals showing up to a 40% reduction in CVD risk (24, 38, 79, 82). The primary roles of both reductions in traditional risk factors and changes in hemodynamics in mediating the positive effects of regular exercise, and negative effects of inactivity, on vascular health have been well described (29, 38, 68, 89). However, the training-induced decrease in CVD risk and improvements in vascular function cannot be fully attributed to the attenuation of traditional risk factors (30, 32, 39) and are not likely fully explained by changes in blood flow (69). Research suggests that circulating factors in the blood may be partially responsible for the vascular protection conferred by regular exercise, though the association between habitual exercise and the effects of the circulating milieu on endothelial cells has not been determined (29, 30, 59, 69).

Additionally, premenopausal women exhibit lower rates of CVD in comparison with age-matched men (8). These sex-based disparities are hypothesized to be largely dependent on sex hormones, primarily estrogen, which exerts numerous protective effects on the endothelium in women (9, 37, 64, 84). However, a recent study including thousands of participants found sex differences in 61 of 71 investigated circulating proteins involved in CVD development, with the greatest differences apparent between premenopausal women and age-matched men (55). Thus, other circulating factors likely contribute to disparities in CVD risk between men and women, and it may be useful to determine sex-based differences in the effects of all blood-borne factors on the endothelium, excluding differences in estrogen.

MicroRNAs (miRs) are a relatively novel class of blood-borne molecules proposed to mediate the effects of exercise on endothelial health via posttranscriptional regulation (5, 77). Exercise training alters cardiovascular-related circulating miRs (ci-miRs) (reviewed in 75, 77), but given the difficulty in performing long-term exercise training studies, long-term effects (>6 mo) of habitual exercise on the ci-miR profile are not known. Additionally, sex-based differences in the expression of select miRs have been identified and may underlie disparities in CVD development (3, 12, 21, 72, 73), but it is unclear whether habitual exercise impacts ci-miRs differently in men and women. Therefore, we used a cross-sectional study comparing young, healthy men and women who either habitually performed endurance exercise (endurance trained) or were relatively inactive (sedentary).

We sought to determine the associations of habitual activity level and sex with the effects of the circulating milieu (in serum) on endothelial cell function and the levels of cardiovascular-related ci-miRs. By comparing the effects of serum derived from young individuals devoid of traditional CVD risk factors and differing only in chronic exercise habits, the aim was to identify alterations in the circulating milieu that likely play a role in the early development of CVD in physically inactive individuals. To exclude acute effects of circulating estrogen, women were tested during the menstrual/early follicular phase (first 3 days) of the menstrual cycle or the placebo phase of oral contraceptive, when estradiol is at its lowest concentration and is similar to that in men (81, 83). It was hypothesized that serum from endurance-trained individuals would confer beneficial effects on endothelial cell functions as compared with serum of sedentary individuals, and that serum of women would have beneficial effects compared with that of men in each respective group. Likewise, we hypothesized habitual endurance exercise- and sex-based differences in the levels of cardiovascular-related ci-miRs, such that ci-miRs that promote inflammation and oxidative stress would be greater in sedentary than trained individuals, and in men than women.

METHODS

Ethical approval.

All procedures and documents conformed to the Declaration of Helsinki and were approved by the University of Maryland Institutional Review Board before participant recruitment. Before beginning any testing, all participants were provided both a verbal and written explanation of the study, and participants gave their written informed consent.

Participants.

Healthy men and women of any race/ethnicity between the ages of 18 and 39 yr who reported being either endurance trained or sedentary were recruited for this study. Physical activity status was determined first by a questionnaire regarding habitual physical activity level over at least the past 5 years. Self-reported habitual activity level was then corroborated by a maximal oxygen consumption (V̇o_2max) test. Endurance-trained participants reported structured aerobic exercise (including running, cycling, swimming, and/or rowing) for >30 min/day on 4 or more days per week and/or >4 h/wk. On average, the endurance-trained participants had been training continuously for the past 11.5 ± 6.4 yr. A majority of the endurance-trained individuals also reported that they regularly trained for and competed in endurance races. While some participants reported that they occasionally (i.e., not on a regular weekly basis) participated in other types of exercises (e.g., resistance training, yoga), endurance exercise was by far the major exercise type for each individual. Sedentary individuals did not engage in regular scheduled exercise and reported ≤20 min/day on ≤2 days per week of structured physical activity.

To confirm health status, participants completed a health history questionnaire and further testing. Measurements included height and weight, seated blood pressure and heart rate (HR), and body fat measurement via the seven-site skinfold method (49). A resting, fasted blood sample was also used to assess blood chemistry for CVD risk factors. Exclusion criteria included any one of the following: prior CVD or metabolic diseases, current or past smoker, use of any potential study-confounding medications such as nonsteroidal anti-inflammatory drugs (NSAIDs), body weight of <110 pounds, body mass index (BMI) of >29 kg/m², resting HR of >100 bpm, systolic blood pressure of >140 mmHg or diastolic blood pressure of >90 mmHg, fasting total cholesterol of >200 mg/dL, low-density lipoprotein (LDL) cholesterol of >130 mg/dL, high-density lipoprotein (HDL) cholesterol of <40 mg/dL, or blood glucose of >100 mg/dL.

Blood sampling.

Participants arrived to the laboratory in the morning after an overnight fast of ≥8 h. Participants were also asked to refrain from caffeine overnight; alcohol, NSAIDs, or other medications for 24 h; and exercise for at least 16 h before the visit. All women were tested during the first 3 days of their menstrual cycle or during the placebo pill phase of oral contraceptive, which was self-reported by the participants. After seated rest of ≥5 min, 10 mL of blood was drawn from an antecubital vein into a serum separator tube and was allowed to clot at room temperature for 45 min, after which it was centrifuged at 1,500 g for 15 min at 4°C. Serum was aliquoted and stored at −80°C until use. An additional 8–10 mL of blood was also obtained and sent for analysis of blood chemistry for screening (Quest Diagnostics, Baltimore, MD). Serum samples from some participants were used in other studies and were therefore limited. Serum from all the participants was used in the endothelial migration and proliferation assays, while the number of samples used in the estradiol enzyme-linked immunoassay (ELISA), endothelial reactive oxygen species (ROS) assay, and real-time quantitative polymerase chain reaction (RT-qPCR) for each group was as follows: endurance-trained men (n = 12), endurance-trained women (n = 12), sedentary men (n = 11), and sedentary women (n = 9).

Maximal oxygen consumption test.

The participants had the option to perform the V̇o_2max test on the same visit as the fasting blood draw, or on a separate, nonfasting visit. Considering this, the menstrual cycle phase/oral contraceptive phase in which V̇o_2max was assessed in women varied. Previous research suggests no effects of either menstrual cycle phase or oral contraceptive use on V̇o_2max (46, 50, 61). Following ~5 min of warm-up, the participants completed either a treadmill or cycle ergometer exercise test during which they had a mask covering their nose and mouth to measure their oxygen consumption via indirect calorimetry. HR was monitored during the exercise with a chest strap monitor (POLAR T31). Participants ran at a self-selected speed or cycled at ≥80 rpm for the entire test, during which the treadmill grade increased by 2%–3% or cycling intensity increased by 25–50 W every 2 min until volitional exhaustion. A true maximum was confirmed by achievement of at least three of the following criteria: a plateau in V̇o₂ (increase in V̇o₂ of <250 mL/min despite increased workload), HR_max ± 10 bpm within age-predicted max, respiratory exchange ratio of ≥1.15, rating of perceived exertion of ≥17, or postexercise blood lactate level of ≥8 mmol/L.

Estradiol ELISA.

The concentration of estradiol in serum was quantified using an ELISA according to the manufacturer’s protocol (Eagle Biosciences, Nashua, NH). Samples were assayed in duplicate according to the manufacturer’s instructions. A spectrophotometer (Synergy H1 Hybrid Reader; BioTek, Winooski, VT) was used to measure absorbance at 450 nm immediately after stopping the reaction.

Endothelial migration assay.

Pooled donor human umbilical vein endothelial cells (HUVECs) were purchased (Lonza, Basel, Switzerland), grown out to passage 2 (P2), and cryopreserved until further use. Experiments were performed on HUVECs (P3–P5) grown in endothelial growth medium (EGM-2, Lonza) supplemented with 2% fetal bovine serum (FBS) at 37°C and 5% CO₂. HUVECs were harvested using trypsin plus EDTA and seeded onto the Radius 96-well cell migration plate (Cell Biolabs Inc., San Diego, CA) at 25,000 cells/well overnight. Once HUVECs reached confluency around the central circular gel, the media was removed and replaced with a gel removal solution in EGM-2 without FBS for 30 min, followed by multiple washes using EGM-2 without FBS. EGM-2 without FBS plus 10% human serum was then added using serum from each participant. In a dose–response experiment (data not shown), 10% human serum was chosen as the optimal concentration to induce endothelial migration and to detect differences between samples. Serum at higher concentrations resulted in impaired and incomplete migration for the majority of samples tested. Participants’ serum was assessed in triplicate. Control wells in each 96-well plate were cultured with media containing neither FBS nor human serum, which induced minimal migration over 24 h (negative control) (76). Pictures were taken manually using a microscope at ×10 magnification immediately upon addition of samples (0 h time point). Migration was then tracked by taking pictures every 4 h for 24 h. The outer edge of migrating cells was traced using ImageJ (NIH), and the area was quantified at each time point. Quantification was performed separately by two blinded investigators. Migration rate at each time point was calculated as percentage of area closure over time [100 – (size of area at time point/size of initial area × 100)]. Area under the curve (AUC) was then calculated using migration at time points 0, 4, 8, 12, 16, and 24 h.

Endothelial proliferation assay.

HUVECs were seeded into a white-walled, clear-bottom 96-well plate at a concentration of 4,000 cells/well and allowed to attach overnight. The media was then removed, cells were washed, and EGM-2 without FBS containing 20% serum from each participant was added based on previous literature and dose–response experiments (59). Serum from each participant was added to wells in triplicate. Control wells were cultured with the same conditions used in the migration assay. After 36 h, proliferation was assessed using a fluorometric cell proliferation assay kit (BioVision, Inc., Milpitas, CA). Wells were washed with endothelial basal media to remove any dead cells or debris, after which a cell lysis buffer and a nuclear cell dye were added and the 96-well plate was gently shaken for 15 min. Fluorescence was read on a fluorescence microplate reader (BioTek FLx800TBIE) at Ex/Em = 485/528 nm.

Reactive oxygen species assay.

The effects of serum on ROS production in HUVECs were determined using the 2′,7′-dichlorofluorescin diacetate (DCFDA) Cellular ROS Detection Assay Kit (Abcam, Cambridge, MA). DCFDA diffuses into cells and is converted to the fluorescent molecule 2′,7′-dichlorofluorescein (DCF) by ROS. Cells were seeded into a black, clear-bottom 96-well plate at 15,000 cells/well in EGM-2 + 2% FBS without phenol red and allowed to attach overnight. HUVECs were then washed with PBS and stained with 25 μM DCFDA for 45 min. After removing the DCFDA and washing again, duplicate wells of cells were exposed to 20% human serum from the participants in EGM-2 without FBS or phenol red for 6 h. Following the incubation, fluorescence was read at Ex/Em = 485/535 nm.

Circulating microRNA quantification.

Ci-miR concentrations were assessed in serum using RT-qPCR. Although some studies have found differences in ci-miR concentrations between sample types (serum versus plasma) potentially due to the coagulation process, ci-miR concentrations between serum and plasma are generally well correlated (36, 62, 90). We chose to use serum as a sample type, as it was used in the culture-based endothelial cell assays. Total RNA was first isolated from 200 μL of serum using the miRNeasy serum/plasma kit (Qiagen, Germantown, MD). Prior to addition of chloroform during the isolation process, a synthetic spike-in control [Caenorhabditis elegans miR-39 (cel-miR-39)] was added for normalization of RT-qPCR data. RNA (2 μL) from each sample was reverse transcribed using the miScript II RT kit (Qiagen). RT-qPCR was performed using 2.5 µL of input cDNA with the miScript SYBR Green PCR Kit (Qiagen, Germantown, MD). The PCR arrays were run on an ABI 7300 Real-Time PCR System (Applied Biosystems), whereas an Agilent Mx3005P qPCR system was used for individual ci-miR quantification.

PCR array.

The Human CVD miRNA PCR Array (SABiosciences) was first used to investigate the expression of 84 miRs previously implicated in CVD pathologies (supplemental information available at https://doi.org/10.6084/m9.figshare.12133155). Samples from three participants of each activity level were chosen to be pooled and investigated in the miR array based on their performance in the migration assay. Three participants from the endurance-trained group whose serum induced expedited and complete (100%) HUVEC migration in the migration assay and three participants from the sedentary group whose serum induced impaired migration were included. This consisted of two women and one man of each activity level. RNA from these participants was pooled for each group and 2 μL was reverse transcribed into cDNA using the miScript II RT kit (Qiagen). Analysis of the array data was performed using a Web-based tool (Qiagen, GeneGlobe Data Analysis Center). This tool first calibrated the cycle threshold (CT) values based on recovery of spiked-in cel-miR-39 and then normalized to the average CT values of the five most invariant miRs between the groups for the ΔΔCT calculations. The 2^-ΔΔCT method of relative quantification was used to compare groups. Fold difference was calculated as the normalized 2^-ΔΔCT value for the trained sample over that of the sedentary sample. Fold difference values <1, indicating lower expression in the trained sample, were transformed to negative values by taking the inverse (i.e., 1/fold difference value).

Individual miR.

For quantification of individual miR targets chosen both based on the results of the PCR array and selected a priori, specific miScript primer assays (Qiagen) were used and each reaction was performed in duplicate. Ci-miRs chosen for analysis based on results of the array included miRs 23b-3p, 31–5p, 140–5p, 145–5p, and 199a-5p, whereas those chosen for analysis a priori included miRs 126–3p, 146a-5p, 181b-5p, 221–3p, and 222–3p. MiR expression levels were determined using the 2^-ΔΔCT method of relative quantification. For each miR within each sample, ΔCT = CT of miR – CT of spike-in control miR; ΔΔCT = ΔCT for individual sample – average ΔCT of all endurance-trained individuals. Ci-miRs quantified in all samples in <35 cycles (CT < 35) were determined as sufficiently expressed.

Statistics.

All statistical analyses were performed using GraphPad Prism 8. Migration (AUC values), proliferation, ROS, and ci-miR data were analyzed using two-way ANOVA (activity level × sex). The effects of oral contraceptive use on serum-induced endothelial responses were also determined in separate comparisons of women by two-way ANOVA (oral contraceptive use × activity level). When a significant or borderline significant interaction was identified, planned Fisher’s least significant difference post hoc comparisons were performed to compare activity levels within the sexes and sexes within the activity levels. Pearson’s correlation coefficients were calculated to determine associations between ci-miR concentrations and HUVEC responses to serum in the assays. All tests were two-sided, with P < 0.05 considered statistically significant and 0.05 <P < 0.09 considered borderline (i.e., trending toward) significant.

RESULTS

Participant characteristics.

Participants were all healthy and groups were generally well matched. There were no differences between groups in age or BMI (Table 1). On average, the relative V̇o_2max of the endurance-trained men (P < 0.0001) and women (P < 0.0001) was greater than those of their sedentary counterparts by ~50%–57% (~20 ml·kg⁻¹·min⁻¹), while men in both groups also had significantly greater V̇o_2max than women of the same activity level (trained: P = 0.0002; sedentary: P < 0.0001). Similar statistical differences were seen between groups in body fat percentage, with the sedentary groups and women having greater body fat levels than the trained groups and men, respectively (all P ≤ 0.0003). Men had greater systolic blood pressure than women of the same activity level (trained P = 0.0001; sedentary P = 0.004), though there was only a significant difference in mean arterial pressure between endurance-trained men and women (P = 0.01). Sedentary men (P = 0.002) and trained women (P = 0.02) both had significantly higher glucose than sedentary women, while trained women also had significantly higher HDL cholesterol than trained men (P = 0.003). Seven women in the endurance-trained group (39%) and seven women in the sedentary group (50%) reported being on oral contraceptives. There were no significant differences between groups in circulating estradiol concentration (P = 0.40–0.99).

Table 1.

Participant characteristics

	Men		Women
	Trained	Sedentary	Trained	Sedentary
	(n = 16)	(n = 16)	(n = 18)	(n = 14)
Age, y	23 ± 5	24 ± 5	25 ± 6	25 ± 4
V̇o2max, L/min	4.35 ± 0.62*†	3.05 ± 0.36†	3.06 ± 0.49*	1.94 ± 0.36
V̇o2max, mL·kg−1·min−1	62.6 ± 7.9*†	41.6 ± 4.2†	52.4 ± 6.0*	33.4 ± 4.4
BMI, kg/m2	22 ± 2	23 ± 2	22 ± 2	22 ± 2
Body fat, %	8.5 ± 2.7*†	16.1 ± 5.7†	17.6 ± 4.2*	23.6 ± 3.3
SBP, mmHg	123 ± 5†	123 ± 10†	113 ± 7	111 ± 9
DBP, mmHg	73 ± 9	76 ± 9	69 ± 7	74 ± 9
MAP, mmHg	89 ± 6†	92 ± 9	81 ± 10	86 ± 8
Glucose, mg/dL	89 ± 7	90 ± 6†	88 ± 6*	81 ± 9
Total cholesterol, mg/dL	151 ± 25	152 ± 27	164 ± 23	166 ± 31
HDL-C, mg/dL	56 ± 12†	55 ± 9	71 ± 15	63 ± 17
LDL-C, mg/dL	81 ± 20	83 ± 15	79 ± 19	88 ± 21
VLDL-C, mg/dL	14 ± 4	14 ± 6	14 ± 5	14 ± 5
Triglycerides, mg/dL	72 ± 22	70 ± 28	70 ± 21	70 ± 23
Estradiol, pg/mL	101 ± 70	100 ± 59	156 ± 219	144 ± 218

Mean ± SD. Participant demographics were compared between groups using independent t tests. BMI, body mass index; DBP, diastolic blood pressure; HDL, high-density lipoprotein; LDL, low-density lipoprotein; MAP, mean arterial pressure; SBP, systolic blood pressure; VLDL, very low-density lipoprotein.

^*P < 0.05 vs. sedentary of the same sex; and

^†P < 0.05 vs. women with same activity status.

Endothelial cell assays.

Two-way ANOVA revealed a borderline significant interaction effect (P = 0.053) of serum on endothelial cell migration AUC over 24 h (Fig. 1). There was a significant effect of habitual activity level (P = 0.03), with no effect of sex (P = 0.67). Post hoc tests indicated that serum from endurance-trained women induced significantly greater migration compared with serum from sedentary women (P = 0.004) and also approached statistical significance compared with trained men (P = 0.08). There were no significant interaction (P = 0.15), activity level (P = 0.12), or sex (P = 0.87) effects on endothelial cell proliferation (Fig. 2). There was a borderline significant interaction effect (P = 0.08) on ROS production of HUVECs in response to serum (Fig. 3). There was a significant sex effect (P = 0.004) on ROS levels, with no effect of activity level (P = 0.29). Specifically, serum from sedentary men caused significantly greater ROS production compared with serum from both their endurance-trained male (P = 0.04) and sedentary female (P = 0.002) counterparts. There was no significant difference in endothelial ROS production in response to serum of trained compared with sedentary women (P = 0.61).

Fig. 1.

A: endothelial cell migration over 24 h. B: area under the curve in response to 10% serum from endurance-trained men (n = 16) and women (n = 18) and sedentary men (n = 16) and women (n = 14). Means ± SE; *P < 0.05.

Fig. 2.

Endothelial cell proliferation in response to 20% serum from endurance-trained men (n = 16) and women (n = 18) and sedentary men (n = 16) and women (n = 14). Fluorescence is reported relative to the endothelial growth medium no serum negative control (set as 1). Means ± SE.

Fig. 3.

Endothelial cell reactive oxygen species production in response to 20% serum from endurance-trained men (n = 12) and women (n = 12) and sedentary men (n = 11) and women (n = 9). Means ± SE; *P < 0.05.

Comparisons of serum from women based on the use of oral contraceptives revealed no significant interaction effects or effects of oral contraceptive use on endothelial migration (P = 0.13, P = 0.66), proliferation (P = 0.30, P = 0.40), or ROS production (P = 0.90, P = 0.25) (Fig. 4). There remained a significant effect of activity level on serum-induced endothelial migration (P = 0.002). Additionally, in these comparisons that only included women, there was a borderline significant effect of activity level on serum-induced endothelial proliferation (P = 0.066), though there was still no effect on ROS production (P = 0.40).

Fig. 4.

The effects of serum from trained and sedentary women using oral contraceptives (OC) compared with those not using OC (no OC) on endothelial migration (A), proliferation (B), and reactive oxygen species production (C). For migration and proliferation: no OC trained (n = 11), OC trained (n = 7), no OC sedentary (n = 7), and OC sedentary (n = 7). For ROS production: no OC trained (n = 6), OC trained (n = 6), no OC sedentary (n = 5), and OC sedentary (n = 4). Means ± SE; *P < 0.05.

Human cardiovascular disease miR PCR array.

Of the 84 miRs investigated, 70 showed appropriate abundance for analysis in samples from both the endurance-trained and sedentary groups (CT ≤ 35). Controls for reverse transcription and PCR indicated acceptable performance for analysis. CT values were first calibrated based on the recovery of spiked-in cel-miR-39, and data were further normalized using the five most invariant ci-miRs between groups in the array (miRs 210–3p, 122–5p, let-7c-5p, 424–5p, and 100–5p). Ci-miRs showing greater than fourfold difference between groups were said to be differentially expressed. One ci-miR was more highly expressed in the trained sample, while seven were higher in the sedentary sample. Those ci-miRs and their respective fold differences (endurance-trained group/sedentary group) were miR-31–5p (317), miR-23b-3p (−5.0), miR-30c-5p (−4.2), miR-93–5p (−5.7), miR-125a-5p (−4.5), miR-140–5p (−19), miR-145–5p (−5.3), and miR-199a-5p (−7.0).

Specific ci-miR assays.

Individual ci-miR quantification was performed on a set of five chosen from the array and on a set of five selected a priori based on previous literature (Fig. 5). Ci-miR-31–5p was not reliably detected (CT > 35) in all samples and was therefore excluded from further analysis. There were no significant interaction effects on any ci-miR, though there was a borderline significant interaction effect (P = 0.07) on ci-miR-145–5p. All ci-miRs exhibited significant sex effects, while there were also significant effects of activity level on ci-miRs-140–5p (P = 0.02) and 145–5p (P = 0.02), with greater levels in men. Specifically, endurance-trained (P = 0.07) and sedentary men (P = 0.0002) had greater levels of ci-miR-145–5p compared with their female counterparts. Compared with trained men, ci-miR-145–5p was also higher in sedentary men (P = 0.004).

Fig. 5.

Circulating microRNA (ci-miR) expression in serum from endurance-trained men (n = 12) and women (n = 12) and sedentary men (n = 11) and women (n = 9). Means ± SE; *P < 0.05.

Correlations.

The correlations between serum-induced endothelial cell functions and ci-miRs with cardiovascular and metabolic factors for both men and women separately are shown in Table 2. In men, serum-induced endothelial migration correlated positively with age and LDL concentration, whereas in women, it correlated positively with relative V̇o_2max and negatively with resting HR. Serum-induced endothelial proliferation was related to endothelial ROS production and body fat percentage in men and endothelial migration, diastolic blood pressure, and circulating glucose concentration in women. Circulating glucose concentration was also associated with endothelial ROS production in women only. In men, ci-miR-181b-5p correlated with endothelial migration and ci-miRs 126–3p, 146a-5p, and 199a-5p correlated with systolic blood pressure. In women, there were select associations between ci-miRs and endothelial proliferation and ROS production, circulating triglycerides, and glucose. Circulating estradiol correlated only with serum-induced endothelial ROS production in men.

Table 2.

Pearson’s correlations (r values) between serum-induced endothelial cell functions and circulating microRNAs (miR) with demographic and cardiometabolic factors in men and women

Men	Migration	Prolif.	ROS	Age	BMI	V̇o2max	SBP	DBP	HR	%Body fat	HDL	LDL	Triglyc.	Glucose	Estradiol
Migration	x	0.33	0.35	0.35*	−0.02	0.26	0.11	−0.19	−0.17	−0.04	−0.12	0.38*	0.10	0.23	−0.16
Prolif.	0.33	x	0.58*	0.02	0.30	−0.10	0.17	0.15	−0.02	0.39*	−0.11	0.10	−0.03	0.30	0.33
ROS	0.35	0.58*	x	0.03	0.17	−0.17	0.27	0.40	0.01	0.28	0.02	−0.14	0.01	0.20	0.43*
miR-23b	0.21	−0.08	0.07	0.25	−0.01	0.09	0.37	−0.12	0.22	−0.08	−0.20	−0.33	−0.02	−0.31	−0.23
miR-126	0.21	−0.18	0.18	−0.05	−0.26	0.01	0.46*	0.19	0.16	−0.16	−0.06	−0.26	−0.18	−0.06	−0.04
miR-140	0.33	0.19	0.20	0.099	0.07	−0.23	0.16	−0.04	0.16	0.05	−0.00	−0.13	−0.12	−0.06	−0.14
miR-145	0.19	−0.11	0.22	0.058	−0.18	−0.19	0.25	0.05	0.34	−0.08	−0.01	−0.36	−0.19	−0.12	−0.11
miR-146a	0.24	−0.18	0.21	−0.04	−0.31	−0.01	0.42*	0.22	0.19	−0.15	−0.06	−0.27	−0.20	−0.00	−0.04
miR-181b	0.44*	0.25	0.33	0.15	0.08	0.05	0.30	−0.01	0.09	−0.05	−0.15	−0.25	−0.07	−0.12	−0.04
miR-199a	0.22	−0.21	0.17	−0.10	−0.34	0.00	0.45*	0.30	0.16	−0.19	0.02	−0.24	−0.21	0.06	−0.04
miR-221	0.23	−0.17	0.13	−0.06	−0.33	−0.00	0.34	0.15	0.14	−0.16	−0.08	−0.30	−0.15	0.02	−0.08
miR-222	0.30	−0.08	0.18	0.03	−0.23	−0.03	0.32	0.03	0.10	−0.11	−0.15	−0.34	−0.11	−0.04	−0.12
Women
Migration	x	0.42*	0.15	0.00	0.07	0.45*	0.20	−0.29	-0.55*	−0.21	−0.07	−0.31	0.02	0.34	−0.01
Prolif.	0.42*	x	0.06	−0.15	−0.04	0.31	−0.18	-0.43*	−0.18	0.02	0.04	−0.23	0.07	0.53*	−0.03
ROS	0.15	0.06	x	0.33	−0.03	0.20	0.38	0.17	−0.30	−0.14	0.30	0.17	0.25	0.57*	0.25
miR-23b	−0.04	−0.14	0.45*	−0.15	0.07	−0.04	0.13	−0.07	0.05	0.11	−0.22	0.29	0.55*	0.18	−0.24
miR-126	−0.30	−0.41	0.48*	−0.16	−0.04	−0.10	−0.04	0.07	0.37	0.07	−0.12	−0.07	0.74*	0.41	−0.07
miR-140	0.01	−0.43	0.07	−0.15	−0.12	−0.19	0.11	0.20	0.20	0.14	−0.08	−0.04	0.36	0.02	−0.26
miR-145	−0.03	-0.50*	0.14	−0.34	−0.06	0.00	−0.13	0.32	0.31	0.15	−0.14	0.02	0.29	0.34	−0.13
miR-146a	−0.22	−0.42	0.44*	−0.26	−0.14	−0.11	−0.04	−0.00	0.29	0.02	−0.21	−0.09	0.78*	0.26	−0.12
miR-181b	−0.09	−0.42	0.31	−0.39	−0.09	0.04	0.05	−0.01	0.16	0.01	−0.36	0.05	0.63*	0.15	−0.26
miR-199a	−0.17	-0.45*	0.44*	−0.21	−0.08	0.07	−0.06	−0.03	0.22	−0.08	−0.15	−0.04	0.70*	0.46*	−0.16
miR-221	−0.19	-0.44*	0.44*	−0.25	−0.04	0.02	−0.03	0.03	0.20	−0.03	−0.25	0.03	0.69*	0.36	−0.18
miR-222	−0.19	-0.46*	0.42	−0.26	0.02	−0.01	0.00	0.05	0.22	0.01	−0.27	−0.01	0.68*	0.32	−0.11

BMI, body mass index; DBP, diastolic blood pressure (n = 59); Glucose (n = 63); HDL, high-density lipoprotein (n = 63); HR, resting heart rate (n = 43); LDL, low-density lipoprotein (n = 63); Prolif., proliferation; ROS, reactive oxygen species; SBP, systolic blood pressure (n = 59); Triglyc., triglycerides (n = 63); V̇o_2max, maximal oxygen consumption (n = 61); % Body fat (n = 63). Sample sizes are specified for those variables missing data for some participants not already specified in the text.

^*P < 0.05.

Significant correlations are in bold.

DISCUSSION

As evidenced by the effects of serum on cultured endothelial cells, we found that the circulating milieu differs in association with habitual activity level in a sex-specific manner. There was a significant effect of activity level on serum-induced endothelial cell migration and a significant effect of sex on serum-induced endothelial ROS production. Compared with serum from sedentary individuals of the same sex, serum from endurance-trained women induced greater HUVEC migration, whereas serum of trained men caused lower ROS production. There were also sex-based and some activity level-based differences in the concentrations of ci-miRs, suggesting they may contribute to known disparities in endothelial function and CVD development. Importantly, all participants were matched by age and BMI, as well as being devoid of any CVD risk factors. Owing to our study design, circulating estradiol concentration was also not different between men and women, thus ruling out a major contributor to sex-based differences in endothelial function.

We sought to determine endothelial cell migration and proliferation in response to serum as a method to identify differences in the cumulative effects of blood-borne factors on endothelial cell functions. The migration and proliferation of endothelial cells are important cellular functions in endothelial repair and angiogenesis (20, 22, 47, 52, 56). These processes may be modulated by a variety of factors, with faster migration and proliferation indicating superior function (25, 52, 56). In a previous report, 6 wk of lower-body resistance exercise training in young men did not alter serum-induced HUVEC proliferation (6). Similarly, we did not observe a difference in endothelial migration or proliferation in response to serum of endurance-trained compared with sedentary men. However, we did observe a significant effect of activity level, with serum of endurance-trained women exerting a beneficial effect on endothelial migration in comparison with the serum of their sedentary counterparts, indicating favorable alterations in circulating factors due to habitual exercise.

To our knowledge, we are the first to examine the effects of serum from young, healthy women on endothelial cell functions. In a study of postmenopausal women, 13 wk of walking-based training improved serum’s capacity to induce chemoattractive migration and capillary-like tube formation of HUVECs (48). This is interesting in light of the common phenomenon that endothelial function of postmenopausal women measured at the vascular level commonly does not respond to exercise training, purportedly due to the loss of estrogen’s permissive effect on training adaptations (64, 65). Indeed, the beneficial effects of estrogen on endothelial function are well recognized (9, 84). Estrogen exerts both genomic and nongenomic effects that are anti-inflammatory and antioxidant, also promoting vasodilation, endothelial repair, and angiogenesis (9, 11, 14, 45). To exclude the acute effects of estrogen in our study, women were sampled during the early follicular phase of the menstrual cycle or the placebo phase of oral contraceptive use, and circulating estradiol concentrations were similar to those found in men. Our results show that in young women, sedentary behavior is associated with a profile of circulating factors responsible for acute detrimental effects on endothelial cell function. Further, this may represent a physiological adaptation to chronic endurance exercise/inactivity that is specific to women. Identification of the responsible circulating factors could benefit vascular therapies for postmenopausal women.

The effects of oral contraceptives on the endothelium are currently unresolved (2, 34, 42, 57, 67, 81, 88). Although we controlled for circulating estradiol, additional hormones and other circulating factors may be affected by the use of oral contraceptives (26, 28, 33). Thus, we performed exploratory comparisons including the endurance-trained and sedentary groups of women in our study and found no significant effects of oral contraceptive use on serum-induced endothelial cell migration, proliferation, or ROS production. Future targeted studies should determine the effects of the diverse types of oral contraceptives, length of oral contraceptive use, and the different phases of oral contraceptives and the menstrual cycle on circulating factors and their impacts on the endothelium.

Traditional and nontraditional CVD risk factors are detrimental to endothelial cell health/function largely via the induction of chronic oxidative stress, a state of elevated ROS production that exceeds the rate of removal by antioxidants and nitric oxide (NO). Excessive ROS reduces NO bioavailability and leads to increased vascular inflammation and cellular damage (44). Thus, a chronic state of oxidative stress is a major mechanism underlying the development of endothelial dysfunction and CVD (15, 44). In young, healthy men and women, longitudinal exercise training and cross-sectional studies comparing athletes and sedentary controls suggest that the antioxidant capacity of the blood may increase with training (reviewed in 35). In our study, we observed a significant effect of sex on serum-induced endothelial ROS production. The interaction effect additionally approached statistical significance, and endothelial ROS production was specifically found to be higher in response to serum of sedentary men compared with that of endurance-trained men, as well their sedentary female counterparts. Habitual activity level was not associated with a difference in serum-induced endothelial ROS production in women, suggesting the observed differences in endothelial cell migration were due to other mechanisms. Thus, alterations in blood-borne factors that increase endothelial ROS production may represent an early mechanism underlying CVD development due to inactivity in otherwise healthy men.

A previous study found that compared with serum of anaerobic male athletes (soccer players and sprinters), serum of male triathletes had the highest NO bioavailability, induced the lowest ROS production (measured by thiobarbituric acid reactive substances), and caused the greatest survival and proliferation of cultured endothelial cells (19). Those results suggest aerobic-type exercise training may be the most beneficial in terms of endothelial cell health due to alterations in the circulating milieu. Although that study included only men, our results suggest that exercise training does not influence the effects of serum from young women on endothelial ROS production. It is possible that sedentary, yet otherwise healthy, young women are protected from the endothelial ROS production that was apparent in men in our study, due to the residual effects of estrogen on other antioxidant molecules in circulation (7).

Classically, endurance exercise training improves the circulating cardiometabolic risk factor profile, including circulating glucose, triglycerides, and cholesterol levels (29, 63). In an effort to exclude these factors as the cause of differences in HUVEC functions, all the participants included in our study had healthy levels of circulating glucose, cholesterol, and triglycerides. Indeed, our study groups were very well matched in regard to their metabolic profile, despite large differences in V̇o_2max. Within sexes, the only significant difference in circulating metabolic profile was a higher glucose concentration in endurance-trained compared with sedentary women. Surprisingly, circulating glucose concentration exhibited a positive correlation with serum-induced endothelial proliferation and ROS production in women, but not men. Both of these endothelial measures were also associated with the levels of several ci-miRs exclusively in women. Conversely, only a negative correlation between endothelial proliferation and body fat percentage was observed in men. Adipose tissue is known to secrete several circulating factors (i.e., adipokines) that regulate cardiovascular health (27, 30). These findings support the conclusion that different circulating factors may contribute to endothelial health in men and women. Interestingly, cardiorespiratory fitness, as measured by V̇o_2max and resting HR, correlated with endothelial migration only in women.

Exercise training is also proposed to protect against atherogenesis at least partially via alterations in the circulating concentrations of cytokines (29, 30). Previously however, our laboratory found no differences in the concentrations of a number of growth factors or inflammatory proteins, including vascular endothelial growth factor (VEGF), interleukin-6 (IL-6), and tumor necrosis factor alpha (TNF-α), between young, healthy endurance-trained and sedentary men (53). To our knowledge, similar studies in women have not been performed. A recent large-scale study found differences in many circulating CVD biomarkers between healthy, premenopausal women and age-matched men; unfortunately, they did not focus on fitness or activity levels (55). Thus, studies determining sex- and habitual activity-based differences in blood-borne circulating factors important in early CVD development are needed. Here, we determined differences in the levels of ci-miRs, as they are novel proposed mediators of both exercise training adaptations and CVD development.

Endurance exercise training has been shown to influence the circulating levels of miRs important to endothelial health, though the effects of chronic endurance exercise are not well known and are understudied in women (75, 77). Recently, using serum samples from the HERITAGE Family Study, the effects of a 20-wk endurance exercise training program on ci-miRs related to cardiovascular health were determined in previously sedentary, but healthy, men and women (aged 43.7 ± 12.8 yr) (4). By using the same PCR array of ci-miRs related to cardiovascular health used in our study, the investigators identified 14 ci-miRs with either upregulated or downregulated expression after training, including ci-miRs 126–3p, 146a–5p, and 221–3p (4). In our study, after validation, only ci-miRs 140–5p and 145–5p were different in association with habitual activity level. Both exhibited higher levels in sedentary individuals, with ci-miR-145–5p higher specifically in sedentary men. MiR-140–5p increases ROS production by directly targeting nuclear factor erythroid 2–related factor 2 (NRF2) and sirtuin 2 (SIRT2), whereas miR-145–5p decreases NO production by targeting solute carrier family 7 member 1 (SLC7A1), an l-arginine transporter protein (58, 91). Additionally, both inhibit angiogenesis, with miRs 140–5p and 145–5p directly targeting VEGF and integrin β8, respectively (18, 85). Thus, a reduction in these ci-miRs with training could contribute to increased capillarization and protection against oxidative stress.

Owing to major differences in the limited studies performed to date on the association between habitual physical activity and ci-miR expression in healthy individuals, it is difficult to make conclusions on the associations between fitness level and habitual exercise with ci-miRs (12, 23, 92). In our study, sex was a more important factor than habitual activity level in determining ci-miR concentrations. We found significant sex effects on all nine vascular-related ci-miRs that were detected during validation, with higher levels in men compared with women. Although we cannot infer the directionality of the ci-miR correlations identified, it appears that ci-miRs may also be related to different cardiometabolic factors in young men and women. These results add to the literature showing sex differences in ci-miR concentrations in a variety of age-groups and health conditions (1, 3, 21, 31, 80). In this context, estrogen regulates proteins in the miR biogenesis pathway within endothelial cells, as well as the expression of specific miRs within vascular cells and the circulation (31, 40, 70). The levels of ci-miRs have been found to stay consistent over the course of the menstrual cycle and, therefore, do not appear to be altered due to acute changes in circulating hormones in women (74). Thus, sex differences in ci-miRs likely persisted in our study due to these long-term genomic mechanisms of regulation in women.

Specifically, the anti-inflammatory and antioxidant effects of estrogen are likely partially accomplished via its influence on the expression of several miRs within endothelial cells (31, 70). The ci-miRs assessed in our study regulate targets to primarily induce anti-inflammatory effects, although they have varying effects on NO and ROS. For example, miRs 23b–3p, 145–5p, 146a-5p, and 181b-5p all either directly or indirectly downregulate NFkB expression or activity (16, 43, 60, 86). MiR-126–3p may upregulate endothelial nitric oxide synthase activity at least partially via the phosphatidylinositol 3-kinase/Akt pathway, whereas miR-199–5p and 221–3p may act to downregulate this pathway (17, 51, 66). Endothelial ROS production may be increased by miRs 181b-5p, 199a-5p, and 221–3p/222–3p downregulating sirtuin 1 (SIRT1), superoxide dismutase (SOD1), and peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α), respectively, and decreased by miR-146a-5p-induced NADPH oxidase 4 (NOX4) suppression (41, 51, 93). The cumulative effects of the higher levels of these ci-miRs in men compared with women are unclear, but these may contribute to the elevated serum-induced endothelial ROS production that we observed.

There are some limitations to our study that should be acknowledged. Our study was cross-sectional in nature, so follow-up longitudinal exercise training studies should be performed to investigate changes in circulating factors in sedentary individuals. While we focused on functional assays of endothelial cell function, future studies should also investigate the effects of exercise training-induced changes in circulating factors on intracellular pathways related to endothelial function. Although the sample sizes used in our study were comparable to those used in similar previous studies performed by us and others (19, 54), they may have limited our ability to detect statistically significant differences, considering some of our comparisons approached significance. Although estradiol is the major sex hormone promoting vascular protection in women, there are other sex hormones with roles in endothelial function that we did not measure or control for, such as progesterone and other estrogens (84). Still, circulating progesterone is low in the early follicular/menstrual phase and was likely similar to concentrations in men, considering the estradiol concentrations were not different (81, 83). We also did not control for the effects of androgens in men, such as testosterone, which may exert either positive or negative cardiovascular effects depending on concentration (9, 84). Owing to financial restraints, we pooled samples including men and women for the initial PCR arrays. The results of the array likely would have been different if we had performed arrays separately for samples from men and women, considering we found sex differences in ci-miR levels. Finally, although we found differences in the concentrations of specific ci-miRs and determined correlations, we did not experimentally determine their mechanistic roles in the endothelial cell assays in our study.

In conclusion, we have found that the blood-borne circulating milieu in young, sedentary, but otherwise healthy, individuals exerts detrimental effects on cultured endothelial cells as compared with that of their endurance-trained counterparts. These effects are sex specific, as are the cardiometabolic factors that they correlate with. Identification of the major responsible circulating factors may be useful for the development of early cardiovascular interventions in individuals unable to exercise, as well as therapies for postmenopausal women who commonly do not respond to exercise training. Additionally, in our population, ci-miRs related to vascular health are influenced by sex more so than habitual activity level. Mechanisms underlying the regulation of these miRs, such as estrogen, and the long-term effects of these alterations on cardiovascular health are of interest, as they could contribute to sex disparities in CVD risk. Finally, ci-miRs 140–5p and 145–5p were upregulated in inactive individuals. Their roles in vascular training adaptations and the development of vascular diseases due to sedentary behavior require further investigation.

GRANTS

This study was supported by the National Heart, Lung, and Blood Institute Grant R21-HL98810 (to J. M. Hagberg); National Institutes of Health Predoctoral Institutional Training Grant T32AG000268 (to J. M. Hagberg); University of Maryland Summer Research Fellowship (to R. Q. Landers-Ramos); and University of Maryland Department of Kinesiology graduate research initiative projects (to R. M. Sapp and R. Q. Landers-Ramos).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

R.M.S., R.Q.L.-R., and J.M.H. conceived and designed research; R.M.S., R.Q.L.-R., and D.D.S. collected the data; R.M.S. and C.B.S. analyzed data; R.M.S., R.Q.L.-R., D.D.S., C.B.S., and J.M.H. interpreted results of experiments; R.M.S. prepared figures and drafted manuscript; R.M.S., R.Q.L.-R., D.D.S., C.B.S., and J.M.H. edited and revised manuscript and approved final version of manuscript.