High-intensity interval training modulates retinal microvascular phenotype and DNA methylation of p66^Shc gene: a randomized controlled trial (EXAMIN AGE)

Abstract

Aims

Impairments of retinal vessel diameter are associated with major adverse cardiovascular (CV) events. Promoter DNA methylation is a repressor of the mitochondrial adaptor p66^Shc gene transcription, a key driver of ageing-induced reactive oxygen species. The study aimed to investigate whether high-intensity interval training (HIIT) affects retinal microvascular phenotype as well as p66^Shc expression and oxidative stress in ageing subjects with increased CV risk from the EXAMIN AGE cohort.

Methods and results

Eighty-four sedentary subjects (mean age 59.4 ± 7.0 years) with ≥2 CV risk factors were randomized into either a 12-week HIIT or standard physical activity recommendations. Retinal arteriolar and venular diameters were measured by use of a retinal vessel analyser. As a marker of oxidative stress plasma 3-nitrotyrosine (3-NT) level was determined by ELISA. Gene expression of p66^Shc and DNA methylation were assessed in mononuclear cells by RT-qPCR and methylated-DNA capture (MethylMiner Enrichment Kit) coupled with qPCR, respectively. High-intensity interval training reduced body mass index, fat mass, low-density lipoprotein and increased muscle mass, as well as maximal oxygen uptake (VO₂max). Moreover, HIIT restored microvascular phenotype by inducing retinal arteriolar widening (pre: 175 ± 14 µm vs. post: 181 ± 13 µm, P = 0.001) and venular narrowing (pre: 222 ± 14 µm vs. post: 220 ± 14 µm, P = 0.007). After HIIT, restoration of p66^Shc promoter methylation (P = 0.034) reduced p66^Shc gene expression (P = 0.037) and, in turn, blunted 3-NT plasma levels (P = 0.002).

Conclusion

High-intensity interval training rescues microvascular dysfunction in ageing subjects at increased CV risk. Exercise-induced reprogramming of DNA methylation of p66^Shc gene may represent a putative mechanistic link whereby exercise protects against age-related oxidative stress.

Clinical trial registration

 ClinicalTrials.gov: NCT02796976 (https://clinicaltrials.gov/ct2/show/NCT02796976).

Ageing, Retinal microcirculation, Oxidative stress, p66Shc gene, DNA methylation, Exercise

Introduction

Retinal vessels are part of the cerebrovascular bed and represent an accessible window to investigate microvascular health and subclinical vascular remodelling. The retinal microcirculation has previously been described as a window to the heart.¹ Alterations of the retinal microvascular phenotype are associated with heart failure² and coronary heart disease mortality³ and have been shown to be predictive of long-term cardiovascular (CV) outcome in the general population.⁴ In ageing subjects, narrower retinal arterioles and wider venules have been associated with increased CV events such as stroke,⁵ coronary artery disease,⁶ and higher CV mortality.⁷ Physical inactivity has been associated with worse retinal microvascular phenotype.⁸

A key feature of vascular ageing is the imbalance between NO bioavailability and reactive oxygen species (ROS) leading to endothelial dysfunction, an early step in the pathogenesis of CV events. In this setting, the adaptor protein p66^Shc has emerged as an important redox enzyme implicated in mitochondrial ROS generation. In patients with Type 2 diabetes (T2DM), we previously found that up-regulation of p66^Shc correlates with urinary 8-iso-prostaglandin F2α levels, an in vivo marker of oxidative stress and endothelial dysfunction.⁹ Epigenetic regulation of gene transcription is mediated primarily by DNA methylation and post-translational modifications of histone proteins.¹⁰ While the benefits of physical activity are widely acknowledged, the role of exercise-induced epigenetic regulation of genes implicated in vascular ageing remains poorly understood. Regular exercise can modulate methylation levels which translate into differential gene expression at genome-wide level in healthy men and women.¹¹ High-intensity interval training (HIIT) can improve endothelial function and cardiorespiratory fitness comparable or even superior to moderate continuous training.¹² However, the impact of exercise on the interplay between microvascular phenotype, ROS generation, and p66^Shc gene transcription is unknown. This randomized controlled trial was designed to assess the effects of HIIT on retinal vessel diameters as a microvascular biomarker of CV risk as well as DNA methylation of p66^Shc gene and oxidative stress in ageing subjects with CV risk from the EXAMIN AGE cohort.¹³

Methods

The complete methods section of this randomized controlled trial is available in the Supplementary material online including the power calculation and sample size estimation.

Results

We recruited 452 participants through our Outpatient Prevention Clinic and advertisements in local newspapers in and around the City of Basel. Eighty-four patients (mean age 59 ± 7 years, 42 female) met the inclusion criteria and were randomized into HIIT or standard physical activity recommendations (control group). Forty subjects undergoing HIIT and 34 controls were measured post-intervention (Supplementary material online, Figure S1). No changes in medication during the HIIT-period or HIIT-related adverse events were observed. Distribution of CV risk factors is listed in Supplementary material online, Figure S2. Anthropometric measurements before and after 12 weeks of intervention or standard physical activity are listed in Supplementary material online, Table S1.

Microvascular phenotype

After 12 weeks of HIIT, retinal arteriolar diameters significantly increased (pre: 175 ± 14 µm vs. post: 181 ± 13 µm, P = 0.001) and venular diameters decreased (pre: 222 ± 14 µm vs. post: 220 ± 14 µm, P = 0.007) when compared with the control group (Figure 1A, Supplementary material online, Table S2). Further adjustment for maximal oxygen uptake (VO₂max) showed that arteriolar widening and venular narrowing were dependent on changes in VO₂max. The increased arteriolar-to-venular diameter ratio (AVR) in the intervention group [β (95% confidence interval): 0.03 (0.01–0.05), P = 0.005] was independent of age, change (Δ) in body mass index (BMI), systolic and diastolic blood pressure, CV medication, and ΔVO₂max (Supplementary material online, Table S2). Retinal diameters did not change in the control group. We also performed an intention-to-treat analysis for the primary outcome. The exercise-induced effects on retinal arteriolar diameter and AVR remained significant but not the effect of venular narrowing (Supplementary material online, Table S3).

Figure 1

(A) Retinal vessel diameters in subjects with increased cardiovascular risk before and after intervention. Forty subjects before and after high-intensity interval training and 34 before and after standard physical activity recommendations (control group) were included in the final analysis. (B) Expression of p66^Shc relative to ACTB was measured before and after high-intensity interval training or standard physical activity recommendations in 20 subjects randomly selected from each group. Oxidative stress measured as plasma 3-nitrotyrosine levels in 40 subjects before and after high-intensity interval training and 34 before and after standard physical activity recommendations. (C) p66^Shc gene and CpG islands proximal to p66^Shc promoter (grey lines indicate CpG rich regions amplified with specific primers). Levels of DNA methylation at region 3, 2, and 1 of p66^Shc promoter were measured in the same 20 subjects before and after high-intensity interval training or standard physical activity recommendations. Values are expressed as mean ± standard deviation. ANCOVA P-values corrected for baseline and control group are shown for multiple comparisons. *P < 0.05 for ANCOVA and between group differences; ^†  P < 0.05 for t-test and within group differences. ACTB, actin beta; ANCOVA, analysis of covariance; AVR, arteriolar-to-venular diameter ratio; CRAE, central retinal arteriolar equivalent; CRVE, central retinal venular equivalent; Ct, cycle threshold; HIIT, high-intensity interval training.

Oxidative stress and p66^Shc expression

Twelve weeks of HIIT when compared with control condition significantly reduced mitochondrial adaptor p66^Shc gene expression in peripheral blood mononuclear cells [pre: 6.5 ± 8.4 arbitrary units (AU) vs. post: 1.9 ± 1.5 AU, P = 0.037] and blunted 3-nitrotyrosine (3-NT) plasma levels (pre: 5.6 ± 5.1 µg/mL vs. post: 3.8 ± 2.2 µg/mL, P = 0.002) after adjustment for age, ΔBMI, systolic and diastolic blood pressure, CV medication, and ΔVO₂max (Figure 1B, Supplementary material online, Table S2).

DNA methylation of p66^Shc gene

p66^Shc promoter was analysed for DNA methylation using Methylminer and qPCR. Three different sets of primers were designed to amplify the CPG islands in the p66^Shc promoter (−225/+676 bp of the transcription start site) and to comprehensively examine the methylation status of the CpG islands in the two experimental groups (Figure 1C). A significant restoration of DNA methylation status of p66^Shc promoter was observed in all three regions in the HIIT group (P < 0.05, within group analysis). In contrast, no changes were found in the control group. Following adjustment for confounders DNA methylation levels in region 3 remained significantly up-regulated after HIIT (P = 0.034, Figure 1C).

Discussion

Twelve weeks of HIIT improved retinal microvascular phenotype. Moreover, HIIT-induced reprogramming of DNA methylation of p66^Shc gene was associated with down-regulation of p66^Shc expression in peripheral blood mononuclear cells and subsequent decrease of systemic oxidative stress. Improvements in retinal AVR, p66^Shc gene expression and oxidative stress were independent of age, ΔBMI, systolic and diastolic blood pressure, CV medication, and ΔVO₂max.

Our randomized controlled trial demonstrates that microvascular remodelling in an older population with increased CV risk can be reversed by short-term exercise training. HIIT resulted in arteriolar widening and venular narrowing. In the intention-to-treat analysis, exercise-induced arteriolar widening but not venular narrowing remained significant suggesting predominant effects on arterioles. This retinal microvascular phenotype has been linked with a lower incidence of CV events and CV mortality.^{4  ,  7  ,  14} Reduced oxidative stress with subsequent restoration of NO bioavailability are major drivers of exercise-induced improvement of microvascular function. Indeed, endurance exercise has the capacity to directly increase endothelial NO production. We have previously shown that exercise-induced dilatation of retinal arteriolar diameters in obese individuals was accompanied by a reduction of asymmetrical dimethyl-L-arginine (ADMA), an endogenous inhibitor of the L-arginine/NO pathway.¹⁵ HIIT has been reported to improve antioxidant capacity and vascular reactivity more potently than moderate continuous training.¹² Higher blood flow and shear stress during HIIT increase plasma glutathione peroxidase availability, which may contribute to ROS reduction and, in turn, increased NO bioavailability.¹⁶ In patients with heart failure, exercise-induced attenuation of oxidative stress has been associated with improvement of vascular function.¹² Endothelial homeostasis depends in large part on the balance between oxidant and antioxidant pathways. Higher shear stress during HIIT has been shown to result in more distinct improvement of endothelial function.¹² Indeed, post-exercise AVR was restored independent of potential confounders and improvement of microvascular endothelial function may be a key mechanism involved.

Direct effects of exercise on clinical risk factors may also impact microvascular health. Blood pressure and fasting glucose levels did not change significantly during exercise training. In patients with diabetes, hyperglycaemia is associated with increased p66^Shc gene expression.⁹ In our study, p66^Shc expression was blunted after exercise without any modification in fasting blood glucose, indicating that exercise-induced changes in p66^Shc gene expression are not necessarily linked to blood glucose levels. The observed changes in body composition with a decrease in fat mass as well as the reduction of low-density lipoprotein-cholesterol levels may contribute to the observed improvement of microvascular function. Interestingly, we adjusted our analyses for BMI, CV medications including other potential confounders and the results remained significant. In view of the different beneficial effects elicited by HIIT, the underlying mechanisms responsible for the restoration of microvascular phenotype are likely to be multifactorial. Future studies are warranted to investigate the relative contribution of direct and indirect exercise-induced effects on the vasculature.

This is the first study in ageing subjects with increased CV risk to report that exercise-induced down-regulation of p66^Shc gene expression is associated with a reduction of oxidative stress. In this regard, it is well established that ONOO⁻ formation originating from the reaction of NO and O₂  ⁻ contributes to increased 3-NT plasma levels.^{17  ,  18} This study shows that HIIT, by rescuing methylation of p66^Shc promoter, can reduce p66^Shc gene transcription that, in turn, may contribute to a decrease of 3-NT plasma levels. We have previously demonstrated in experimental models of diabetes and in patients with T2DM that hypomethylation of p66^Shc promoter causes gene overexpression, oxidative stress, and endothelial dysfunction.^{7  ,  19} In addition, genetic deletion of p66^Shc protects against age-induced, ROS-mediated endothelial dysfunction, most likely by restoring NO bioavailability.²⁰

Our intervention trial investigates for the first time the interconnection between exercise, microvascular phenotype, and epigenetics in ageing subjects with increased CV risk. Interestingly, p66^Shc methylation is dynamic showing robust changes in the post-HIIT group. Our study provides a proof of concept that epigenetic regulation of p66^Shc is related to age-induced oxidative stress and microvascular phenotype. Although only part of the epigenetic crosstalk has been elucidated, the present findings shed light on a key epigenetic mark linking exercise-induced reprogramming of p66^Shc expression, decreased ROS generation and improved microvascular health. The conclusion of our work is summarized in the Take home figure. Previous evidence suggest mechanistic links between p66^Shc transcription, oxidative stress, and endothelial function.^{9  ,  20  ,  21} It is therefore plausible to hypothesize that exercise-induced down-regulation of p66^Shc transcription via DNA methylation contributes to rescue microvascular function. Future research has to prove the causative link in this setting. The disentanglement of single molecular pathways with their systemic impact and potential effects on complex organ function in humans remains a major scientific challenge.

Take home figure

Exercise-induced improvement of microvascular phenotype and reprogramming of p66^Shc DNA methylation. High-intensity interval training improves microvascular phenotype. Reprogramming of DNA methylation mark on p66^Shc gene promoter may represent a mechanistic link whereby physical exercise protects against age-related oxidative stress in the microcirculation.

Our study validates a new multidisciplinary research perspective in clinical medicine encompassing integrated physiology and molecular mechanisms. In conclusion, exercise improves microvascular health in subjects with increased CV risk leading to healthier ageing and eventually better CV outcomes. Reprogramming of DNA methylation on p66^Shc gene promoter may represent a putative link whereby exercise protects against age-related oxidative stress. The entire mechanistic landscape remains to be addressed in future studies.

Acknowledgements

We would like to thank all the staff and the students of the Division of Sports and Exercise Medicine at the Department of Sport, Exercise and Health at the University of Basel for their management support during the study visits and the exercise training. Special thanks to Monique Nussbaumer and Denis Infanger for their statistical and analytical support.

Funding

This work was supported by the Swiss National Science Foundation, Swiss National Science Foundation (32003B_159518/1 to H.H.); and the Nora van Meeuwen-Häfliger Stiftung (to H.H.); the Swedish Research Council (VR 2016-02706 to F.C.); Swedish Heart and Lung Foundation (20140360 to F.C.); and Konung Gustav: Vs Och Drottning Victorias Frimurarestiftelse (to F.C.); Diabetes Wellness Sverige Foundation (to A.W.K.); and European Society of Cardiology Research Grant (to R.S.).

Conflict of interest: none declared.

Author notes

Lukas Streese and Abdul Waheed Khan contributed equally to this work.