Association of circulating Notch1 and VEGF with flow-mediated dilation and aerobic fitness in healthy adults

Experimental design

Participants reported to the lab abstained from alcohol, caffeine, and vigorous exercise 12 h prior to testing, as well as a water-only overnight fast. On the day of the experiment, a venous catheter was inserted into a superficial antecubital vein. A pneumatic cuff was secured around the belly of the forearm on the same arm as the catheter for the FMD. Blood samples were taken prior to cuff inflation (baseline) and 1-min post-cuff deflation (to time align with peak vessel dilation). The FMD followed procedures as outlined in [18] and is outlined below. Following the FMD protocol, participants rested for 15 min and then completed a V̇O₂ peak ramp test on a cycle ergometer. Figure 2 provides an overview of the experimental timeline.

Brachial flow-mediated dilation

All procedures were performed in the supine position and followed the guidelines described in Thijssen et al. [18]. Briefly, a narrow pneumatic cuff was positioned around the belly of the forearm. Following 5 min of quiet rest, the brachial artery was imaged using a duplex ultrasound (T3200, Terason, Burlington, MA, USA) for 1 min of baseline measures. Thereafter, the cuff was rapidly inflated to 220 mmHg for 5 min. The cuff was released after 5 min, and ultrasound recordings continued for an additional 3 min. Ultrasound recordings were analyzed offline using custom-designed edge detection software [19]. Shear rate was calculated as 4x the ratio between mean blood velocity (in cm/s) and artery diameter (in cm) [Shear Rate=4 × (mean velocity/diameter)]. To account for differences in baseline brachial diameter, FMD was scaled allometrically [20], and the resulting data was used in all comparative analyses. Allometric scaling was conducted by first determining the natural logarithm of both the baseline and peak diameters (mm). The values obtained were utilized to conduct a regression analysis to find the slope (i.e., β), which was applied to the following formula: Allometrically scaled FMD=peak diameter/(baseline diameter ˆ β) [20].

Peak exercise test

Aerobic fitness was assessed by using an incremental cycle exercise test on a cycle ergometer (Ergoselect 100, Ergoline). Open-circuit spirometry was used to measure respiratory gases (Quark CPET, COSMED, Chicago, IL, USA). Participants completed either a 20 W, 25 W, 30 W, or 35 W protocol based on their aerobic fitness as determined by the aerobic activity history questionnaire. There was a 2-min warm-up followed by increments of the determined wattage every minute until exhaustion while maintaining 60–80 rotations per minute (RPM). This was followed by a 2-min cool-down (RPM<50). The average test duration was 10.95 ± 1.79 min. Peak exertion was determined when participants met the following criteria: 1) had a respiratory exchange ratio of ≥1.15, 2) failed to maintain ≥60 RPM, and 3) had a plateau in heart rate and V̇O₂ despite an increased workload.

Quantification of NECD and VEGF

Whole blood was collected in sodium citrate tubes (Vacutainer, cat. no. 363080, BD Biosciences, San Jose, CA, USA) and centrifuged for 10 min at 1500 g. Plasma was extracted and immediately stored at −80 °C. The plasma samples were then used to determine the concentration of Notch1 ECD using commercially available immunosorbent assays (RayBio, cat. no. ELH-NOTCH1, RayBiotech Life, Peachtree Corners, GA, USA), with an intra- and inter-assay coefficient of variation of <10 % and <12 %, respectively. The detection limit of the assay is 20 pg/mL to 7,000 pg/mL. Samples were read at 450 nm on a microplate reader (BioTek Synergy HT, BioTek Instruments, Winooski, VT, USA) using Gen5 software (Version 1.11, BioTek Instruments). VEGF was quantified using commercially available immunosorbent assays (Bio-Techne, cat. No. DVE00, USA R&D Systems Inc., Minneapolis, MN, USA), with an intra- and inter-assay coefficient of variation of <7 % and <9 %, respectively. The minimal detectable dose is <9.0 pg/mL. Samples were read at 450 nm with a wavelength correction of 540 nm on a microplate reader (BioTek Synergy HT, BioTek Instruments, Winooski, VT, USA) using Gen5 software (Version 1.11, BioTek Instruments).

Statistical analysis

Data are presented as means ± SD with 95 % confidence intervals (CI), or standard deviation (±SD) for descriptive measures. Tests for significant changes in NECD and VEGF through the FMD protocol were determined by one-tailed paired Student’s t-tests (baseline vs. peak dilation). Correlations were determined using one-tailed Pearson’s correlation coefficient (r). Tests for normality were confirmed by Shapiro-Wilk tests. Statistical outliers were identified as values exceeding twice the standard deviation from the mean and were excluded from the analyses where applicable. Following the removal of outliers (either n=1 or n=2, depending on the variable), all values were normally distributed. The removal of outliers did not influence the interpretation of the results; however, the association between V̇O₂ peak and allometrically scaled FMD no longer reached significance. All statistical analyses were performed in PRISM (GraphPad, Boston, MA, USA). Tests to evaluate sex differences were not performed due to insufficient statistical power, therefore males and females were combined for analysis.

Absolute NECD and VEGF concentrations throughout the FMD

The absolute concentrations of NECD and VEGF throughout the FMD protocol are depicted in Figure 3, expressed as pg/mL. A summary of baseline and peak hemodynamic measures for the FMD test, including shear rate and arterial diameter, is presented in Table 1. Compared to baseline, concentrations of NECD were not statistically different following cuff release (peak dilation) (p=0.458). The baseline concentration of NECD was 267.85 ± 202.21 pg/mL with lower and upper 95 % CI of 171.72–363.97 pg/mL. The concentration remained similar following cuff release at 272.08 ± 215.79 pg/mL with lower and upper 95 % CI of 169.51–374.66 pg/mL. Similarly to NECD, baseline VEGF concentrations were unchanged following cuff release (p=0.069). The baseline concentration of VEGF was 145.48 ± 58.72 pg/mL with lower and upper 95 % CI of 117.56–173.39 pg/mL. Following cuff release, the VEGF concentration was 158.79 ± 47.56 pg/mL with lower and upper 95 % CI of 136.19–181.40 pg/mL.

Figure 3:

Individual values for NECD (pg/mL) (n=15) (left) and VEGF (pg/mL) (n=14) (right) at baseline and 1 min after cuff release (peak dilation).

Associations of NECD and VEGF with FMD

To investigate the involvement of Notch1 signalling in the FMD response, the percent change of NECD and VEGF concentrations were compared to allometrically scaled FMD, as shown in Figure 4. The percent change of NECD across the FMD was not correlated with allometrically scaled FMD (p=0.336, r= −0.118). Additionally, the change in VEGF across the FMD was also unrelated to allometrically scaled FMD (p=0.342, r= −0.114).

Figure 4:

Relationships between allometrically scaled FMD (x-axis) with NECD and VEGF (y-axis) concentrations expressed as percent change across the FMD (n=15). r denotes Pearson’s correlation coefficient (95 % confidence limits).

Relationship between NECD and VEGF

To determine the relationship between Notch1 and VEGF activity during the FMD protocol, the percent change in NECD across the FMD was compared to the percent change in VEGF across the FMD, as shown in Figure 5. Contrary to our second hypothesis, the change in NECD was not correlated with the change in VEGF across the FMD (p=0.331, r=0.127). However, NECD and VEGF at baseline were moderately (r= −0.475) and significantly (p=0.042) correlated.

Figure 5:

Relationship between the change (%Δ) in NECD across the FMD (x-axis) and the change (%Δ) in VEGF across the FMD (y-axis) (n=14). r denotes Pearson’s correlation coefficient (95 % confidence limits).

NECD and VEGF with aerobic fitness

To determine the relationship between Notch1 and VEGF activity with aerobic fitness, the percent change in NECD and VEGF across the FMD were compared to individual V̇O₂ peak values, as shown in Figure 6. The change in NECD across the FMD was moderately (r=0.515) and statistically significantly (p=0.024) correlated with V̇O₂ peak. However, the change in VEGF across the FMD was not correlated to V̇O₂ peak (p=0.287, r= −0.157). Additionally, there were no significant associations between aerobic fitness and baseline NECD (p=0.287, r= −0.163) or VEGF (p=0.089, r= −0.367). The allometrically scaled FMD was not statistically associated with V̇O₂ peak (p=0.087. r=0.357).

Figure 6:

Relationships between V̇O₂ peak (mL/kg/min) (x-axis) with NECD (n=15) and VEGF (n=15) activity expressed as percent change across the FMD (y-axis). r denotes Pearson’s correlation coefficient (95 % confidence limits).

No change in NECD and VEGF across the FMD

Despite our previous report demonstrating increases in circulating NECD following elevated antegrade shear stress [14], there was no change in NECD from baseline to 1-min post-cuff release (Figure 3). In other words, NECD did not change from 5 min of reduced antegrade shear (cuff inflation) and 1 min of increased antegrade shear (cuff deflation). It was surprising that NECD was not elevated during the FMD protocol, considering the interactions between Notch1 and the potent vasodilator, nitric oxide [21], which is largely responsible for the FMD response [22]. Moreover, we have previously reported an elevation in Notch1 with an acute increase in shear stress [14]. However, compared to Badour et al., the lack of NECD change in the present study is likely explained by the differing experimental protocol [14]. That is, in Badour et al. [14], the ischemic condition lasted for 20 min (as opposed to 5 min), and an upper cuff was also placed over the arm (just below the axilla), which was inflated to 40 mmHg to ‘trap’ circulating proteins released from the shear stimulus. In the current study, a proximal pressure cuff to trap circulating NECD was not employed, as it would have interfered with the standardized FMD protocol. A large concentration of NECD released from the FMD may have, therefore, been washed away in the circulation, rendering it relatively undetectable. The 20-min vs. 5-min ischemia may also explain the differences with Badour et al., whereby more prolonged tissue hypoxia may have potentiated NECD release [14].

Unaltered concentrations of VEGF across the FMD protocol may also be due to the duration of blood flow occlusion, as hypoxia is a chief regulator of VEGF release [23] and, therefore, may also require prolonged occlusion to stimulate an upregulation. Moreover, VEGF has been shown to exhibit a delayed release [24], 25], suggesting that a subsequent blood draw at 12–24 h may be necessary for the detection of any alterations in VEGF.

Interaction between NECD and VEGF

It was hypothesized that the FMD process would stimulate angiogenic activity, followed by downstream Dll4-Notch1 signalling [26], demonstrated by increased VEGF and NECD activity. Nevertheless, despite no significant increases in NECD or VEGF across the FMD and the aforementioned delayed release of VEGF, potential acute relations between NECD and VEGF were still explored to possibly shed light on the in vitro data demonstrating mechanistic interactions [13], 23], 26]. Indeed, VEGF and Notch1 function in a competitive manner to regulate angiogenesis, wherein increased levels of VEGF lead to an upregulation of Dll4, subsequently enhancing Notch1 signalling, which in turn inhibits the expression of VEGFR [9], 12], 27], 28]. This was demonstrated by the inverse correlation between NECD and VEGF at baseline (p=0.042), whereby increased NECD corresponds to reduced concentrations of VEGF. However, the individual change in NECD was not related to the change in VEGF across the FMD (Figure 5). These data indicate there are no interactions between the changes in VEGF and NECD acutely following 5 min of blood flow occlusion, at least that can be gleaned from the circulation, or that this stimulus was not robust or long enough to elicit any measurable effects.

Positive association between V̇O₂ peak and change in NECD, but not VEGF, across the FMD

Despite the lack of a consistent directional change in NECD and VEGF across the FMD, there was a significant correlation between the V̇O₂ peak and the change in NECD across the FMD (Figure 6; p=0.024, r=0.515). That is, those with the greatest aerobic fitness, on average, had the greatest release of NECD during the FMD. The correlation between NECD and V̇O₂ peak indicates that the vascular benefits of aerobic fitness [16], 29] may be partly attributed to Notch1 signalling, whereby Notch1, in non-tumorigenic states, is linked to vascular endothelial protection [6]. Specifically, the Notch1 signalling pathway has been shown to enhance the endothelial barrier by improving junctional integrity, consequently reducing vascular permeability [9], 27], 30], 31]. Indeed, diminished Notch1 expression has deleterious repercussions for the endothelium, including increasing pro-atherogenic signalling, enhanced endothelial adhesion of inflammatory molecules, increased endothelial cell proliferation, and reduced endothelial junctional integrity [6], 8].

Allometrically scaled FMD values were not associated with V̇O₂ peak (Figure 7), which is consistent with previous studies [32], 33], albeit not a universal finding [34], 35]. However, it is interesting that the change in Notch1 across the FMD was associated with V̇O₂ peak. These data may point to other vascular protective effects of Notch1 that may be upregulated with improved aerobic fitness, i.e., atheroprotection [10], that are, at least acutely, unrelated to vascular dilation in response to shear stress. Future investigations relating to the relationship between NECD responses to shear stress and endothelial plaque (e.g., across aging and fitness) may provide insight into the observations from this study.

Figure 7:

Relationship between V̇O₂ peak (mL/kg/min) (x-axis) and allometrically scaled FMD (y-axis) (n=16). r denotes Pearson’s correlation coefficient (95 % confidence limits).

Considerations

In this initial pilot study (n=16) investigating Notch1 with vascular function and aerobic fitness, we examined NECD and VEGF in the venous circulation while assessing vascular function in the brachial artery. It is also important to note that NECD is an indirect measure of intracellular Notch1 activity. During canonical Notch1 signalling, NECD is cleaved via γ-secretase, resulting in its subsequent release into the bloodstream, enabling convenient quantification of plasma NECD [10]. NICD is then cleaved and translocated to the nucleus, where it regulates atheroprotective gene transcription [13]. As such, NICD would be a more accurate indicator of Notch1 signalling. However, quantification of NICD is not feasible in vivo in humans, and NECD as a proxy for Notch1 activity is the best alternative. Additionally, it is important to note that while the typical peak dilation occurs around 60 s, this can vary among participants [22]. Therefore, the ‘peak dilation’ blood draw may not align perfectly with the individual peak dilation times, and the venous vs. arterial compartments must be considered.

Secondly, we did not control for menstrual phases in female participants or contraceptive use, and we were underpowered to perform a sex difference analysis. The different phases of the menstrual cycle may have contributed to a larger variability in the data and should be considered in future studies. Lastly, the participants were young, healthy adults, and therefore, further research is needed to elucidate the implications of Notch1 and aerobic fitness in aging vasculature.