Exercise-induced responses in matrix metalloproteinases and osteopontin are not moderated by exercise format in males with overweight or obesity

Obesity and its determinants have now been identified as risk factors for three of the four leading causes of non-communicable diseases worldwide (Swinburn et al. 2019). With an estimated two billion people worldwide living with obesity in 2015 (Swinburn et al. 2019), there is a need to better understand the early developmental stages of obesity and obesity-related conditions. Obesity is characterised by adipose tissue expansion with infiltration of macrophages expressing pro-inflammatory cytokines (Nomiyama et al. 2007; Asghar and Sheikh 2017; Wensveen et al. 2015). These pro-inflammatory cytokines can impair insulin signalling (Lazar 2005) leading to chronically elevated blood glucose concentrations, and contribute to microvascular and macrovascular complications in diabetes mellitus (Bagi et al. 2014; Knight and Imig 2007; Van Gaal et al. 2006; Wellen and Hotamisligil 2003). Microvascular and macrovascular complications largely account for the high rates of all-cause mortality in individuals with diabetes mellitus (Fox et al. 2007; Meshkani and Vakili 2016).

The events initiating the shift to the predominantly pro-inflammatory state in obesity remain unclear (Pirola and Ferraz 2017; Wang et al. 2017), although several candidates have emerged as potential contributors. A class of enzymes known as matrix metalloproteinases (MMPs) play important roles in the early stages of obesity-related disease progression (Bouloumie et al. 2001; Visse and Nagase 2003). Specifically, MMP-2 is released from adipose tissue during adipocyte differentiation, and is understood to regulate the vascular extracellular matrix remodelling during adipose tissue formation (Visse and Nagase 2003; Bouloumie et al. 2001). MMP-3, on the other hand, counteracts the role of MMP-2 by inhibiting adipocyte differentiation and halting hyperplastic adipose tissue expansion (Wu et al. 2017).

Osteopontin (OPN), a multifunctional protein shown to promote macrophage migration (Lund et al. 2013) is another potential candidate. OPN was shown to increase levels of the classically activated macrophages in adipose tissue, which are the primary source of circulating pro-inflammatory cytokines (Kahles et al. 2014; Nomiyama et al. 2007). OPN is a substrate for MMP-2 and MMP-3 (Lindsey et al. 2015), with resultant OPN peptides maintaining distinct biological functions.

Concentrations of serum MMP-2 (Derosa et al. 2008) and OPN are increased in patients with obesity and type 2 diabetes mellitus (T2D) (Daniele et al. 2014), while OPN has been implicated in mediating the relationship between inflammation, obesity and diabetes (Kahles et al. 2014). In contrast, the concentrations of MMP-3 (Goncalves et al. 2015) and the adipose tissue derived anti-inflammatory cytokine adiponectin (APN), which can increase the expression of MMP-3 (Tong et al. 2011), are reduced in patients with T2D. The increased concentrations of MMP-2 and reduced concentrations of MMP-3 in T2D are likely associated with hyperglycaemia (Death et al. 2003).

Exercise training plays an important role in decreasing visceral adipose tissue (Verheggen et al. 2016), and improving both acute and chronic glycaemia (Umpierre et al. 2011; Thomas et al. 2006). The response of MMPs to exercise have only been assessed in a few studies with humans, while the extent to which circulating MMPs respond to direct modifications in exercise variables (e.g., intensity, frequency, duration, type) remain largely unexplored in humans (Jaoude and Koh 2016). Further, the response of MMPs to exercise, remains equivocal in humans (Jaoude and Koh 2016). The inconsistent findings may be due to differences in exercise variables (e.g., resistance exercise versus aerobic exercise) and study-populations (non-obese, obese or T2D) (Jaoude and Koh 2016). Given the purported benefits of high-intensity intermittent exercise (HIIE) in regulating glucose concentrations and adiposity (Savikj and Zierath 2020), this study sought to determine i) the acute responses in MMP-2, MMP-3, OPN and APN; ii) the response in 48-h glucose profile, to a HIIE bout compared with an energy-matched moderate-intensity exercise bout. Due to the purported role of MMPs in adipose tissue remodelling and the importance of exercise in decreasing adipose tissue, this study recruited individuals with overweight or obesity. We hypothesised that exercise would reduce 48 h glucose concentrations, acutely increase OPN and MMP-3, and acutely decrease MMP-2, and this would be more pronounced following HIIE.

The participants in the study were 27.4 ± 6.3 (mean ± SD) years old; classified as either overweight (n = 12) or obese (n = 2) by BMI (29.0 ± 3.2 kg/m²); with total body fat percent of 27.2 ± 4.2% and visceral fat of 613.2 ± 122.7 cm³. Participants were normoglycemic (fasting glucose: 5.2 ± 0.6 mmol/L) and had a VO₂ peak of 35.7 ± 5.1 ml/kg/min. All participants completed the two experimental phases. One participant completed only five sprints during the HIIE bout, whilst one other participant completed all six sprints in the HIIE bout but did not complete the final 4 min (low-intensity portion of the protocol) of the bout. HIIE elicited a significantly higher mean heart rate and RPE (151 ± 33 bpm and 16 ± 2, respectively) than CME (137 ± 25 bpm and 14 ± 2, respectively).

All subjects provided valid wear time data on each of the days during the experimental phases (Table 1). Energy expenditure (kCal), total time in sedentary activity (Sed Time) and the vector magnitude (VM) were no different between days (p > 0.178 for all variables) or conditions (p > 0.196 for all variables).

This study sought to determine the responses in circulating APN, MMP-2, MMP-3 and OPN—key proteins involved in the remodelling of adipose tissue (Wu et al. 2017; Goncalves et al. 2015; Kahles et al. 2014; Sheikh et al. 2022; Unamuno et al. 2021; Akra et al. 2020)—following two different exercise formats in individuals with overweight or obesity. The main findings were (i) MMP-3 was acutely increased in response to exercise, while APN was not significantly different; (ii) MMP-2 did not demonstrate significant changes in response to exercise; (iii) OPN was acutely decreased following exercise, returning to baseline levels by 25 h. In addition, we sought to determine the interstitial glucose response over a 48 h period following the exercise session. The main findings were (i) exercise did not improve diurnal interstitial glucose profiles in this cohort; (ii) the glucose profile immediately following the OGTT supported the diurnal pattern, with the glucose-AUC performed 1 h after exercise (Day 2) and 25 h after exercise (Day 3) being significantly higher than the glucose-AUC on Day 1. In contrast to our hypothesis, the effect of HIIE did not appear to be greater than an energy-matched continuous exercise bout, on any measured marker in this study.

Elevated postprandial glucose is an independent risk factor for cardiovascular events in individuals with or without type 2 diabetes (Bonora and Muggeo 2001). Increased physical activity, however, has been shown to ameliorate the levels of hyperglycaemia (Mikus et al. 2012; Van Dijk et al. 2012). Despite some previous studies showing improvements in systemic glucose following acute exercise (Manders et al. 2010; Mikus et al. 2012; Van Dijk et al. 2012; Oberlin et al. 2014), neither a single bout of CME or HIIE in the present study was sufficient in reducing glucose concentrations under the current experimental conditions. A number of experimental differences may explain this discrepancy between studies. Specifically, the study of Oberlin et al., Manders et al. and Van Dijk et al., were performed in individuals with T2D (Van Dijk et al. 2012; Oberlin et al. 2014; Manders et al. 2010), while participants with overweight or obesity, but no glucose disturbances, were recruited in the current study. The study performed by Mikus et al. modified daily step count as the physical activity intervention (i.e., > 10,000 steps per day versus < 5000 steps per day for three successive days) rather than performing a single exercise session (Mikus et al. 2012). The timing of exercise and the energy intake may also have affected glucose responses (Teo et al. 2018, 2020), given the OGTT was conducted in a fasted state 1 h following exercise in the current study, rather than 2 h following exercise (Oberlin et al. 2014) or consuming breakfast (test meal) prior to exercise (Van Dijk et al. 2012).

There appears to be an important interplay between APN, MMPs and OPN during development of obesity and insulin resistance (Daniele et al. 2014). Among these proteins, the role of APN in the development of obesity and metabolic disease is the most well established, with the concentration of total APN decreasing with increasing adiposity and metabolic dysfunction (Nigro et al. 2014). Accordingly, the concentration of APN in the current study (~ 5 µg/ml) is at the lower end of the expected range (3–30 µg/ml) for human plasma. The acute effect of exercise on APN is equivocal with some reporting increases following acute moderate-to-high intensity (55–70% \(\mathop {\text{V}}\limits^{.}\)O_2max) aerobic exercise (Kriketos et al. 2004; Numao et al. 2011), and others reporting no changes following moderate-intensity (60–65% \(\mathop {\text{V}}\limits^{.}\)O_2max) aerobic exercise (Jamurtas et al. 2006; Magkos et al. 2010). In the current study, neither CME nor HIIE altered the concentration of circulating APN.

To our knowledge, this is the first study to have investigated the effect of exercise intensity on MMP-2 and MMP-3 in individuals with overweight or obesity. The transient increase in MMP-3 immediately post-exercise in the current study was similar to that observed in sedentary men (Urso et al. 2009), as well as endurance-trained and sedentary individuals (Schild et al. 2016). Although the response in MMP-3 to exercise in the study of Schild et al. was significantly more pronounced in endurance-trained individuals than sedentary individuals (Schild et al. 2016). In the current study, MMP-2 did not change in response to acute exercise, which is consistent with previous research reporting an acute increase in plasma MMP-2 concentration only in endurance-trained individuals, not in sedentary individuals (Schild et al. 2016).

The activation of MMPs is complex, requiring an initial cleavage from pro-MMP which can then be activated through oxidative stressors (Jaoude and Koh 2016) such as those expected during an acute, high-intensity exercise session. While MMPs play important roles in tissue remodelling, they may also inactivate cytokines and chemokines including interleukin-1β. It has been postulated that the acute increase in MMPs post-exercise, might function to diminish an excessive post-exercise pro-inflammatory response (Schild et al. 2016). While a main effect of condition was observed in the MMP-2 response, this was due to a higher baseline value where the interaction effect was not significant. Previous research using a cell culture system (Tong et al. 2011) found that APN may also induce the expression of MMP-3; however, in the present study, the transient increase in MMP-3 occurred without a significant change in APN concentration.

OPN appears to play an important initiating role in adipose-tissue expansion, inducing the expression of MMP-2 via pro-MMP-2, and inflammation (Daniele et al. 2014; Philip et al. 2001), as well as directly serving as a substrate for MMP-2 and MMP-3 (Lindsey et al. 2015). The acute decrease in OPN concentration immediately post-exercise coincided with the increase in MMP-3 concentration. This temporal association may explain the acute decrease in OPN immediately post-exercise, but further research is required.

Although previous findings of reduced OPN in mice was beneficial in reducing adipose tissue inflammation and improving glucose tolerance (Kiefer et al. 2010; Nomiyama et al. 2007), the reduction in OPN in the current study was only transient and appeared not to have an effect on glucose tolerance. Indeed, while the respective response of these markers to exercise are consistent with potentially longer-term health benefits; in the current study, these changes were not associated with changes in diurnal glucose profiles as evidenced by a lack of significance over time and the higher glucose responses 1 h and 25 h post-exercise (Figs. 2 and 3).

This study had a number of strengths including (i) duration- and workload-matched CME and HIIE conditions; (ii) recruiting a clinically relevant population, namely individuals who were inactive and overweight or obese- which has, to our knowledge not previously been conducted in a study of this kind; (iii) adopting a randomized cross-over design; (iv) each trial included a specific baseline compromising of 24 h continuous glucose monitoring which were used to compare 24 h differences following exercise for 2  days post-exercise. However, there were also some limitations associated with this study. Despite the measurement of MMP-2 and MMP-3 in this study, we did not measure activity of tissue inhibitors of MMPs which would have accounted for the regulatory activity of these MMPs which may have provided further insight into the acute effect of MMPs following exercise. Additionally, the measurement of MMPs and other proteins were limited to plasma or serum only. Simultaneous measurement of these proteins in the circulation, as well as multiple tissues including skeletal muscle and adipose tissue would provide further mechanistic insights regarding their exercise responses. Implementing a similar study in a population of individuals spanning the spectrum from normoglycemia to individuals with obesity and type 2 diabetes would be a strength for future studies and may allow assessment across clinical and pre-clinical stages.

Although the strict timing on completion of the OGTT allowed for a level of consistency over each trial day, providing all participants with standardized meal over the three days would have allowed for greater control of dietary intake. Similarly, controlling the levels of physical activity on each day (beyond the exercise bout), would be a strength for future research studies; although, based on our physical activity monitoring with a tri-axial activity monitor, energy expenditure, sedentary time and total vector counts appeared to be no different between exercise conditions or days. Finally, conducting multiple regression analysis to assess the effect of respective markers on glucose profile would provide more substantial and concrete answers; however, a greater number of participants would be required to run such an analysis. Due to the lack of significant change in the glucose profile following exercise, we were not able to determine the impact of the changes in these inflammatory markers on the glucose response.

Important roles for MMP-2, MMP-3 and OPN in the remodeling of adipose tissue and development obesity-related diseases are now emerging (Wu et al. 2017; Goncalves et al. 2015; Kahles et al. 2014; Sheikh et al. 2022; Unamuno et al. 2021; Akra et al. 2020). The present study demonstrated that a single bout of moderate- or high-intensity intermittent exercise in individuals with overweight or obesity resulted in an acute and transient increase in MMP-3 and decrease in OPN. While these changes align with expected responses to acute exercise (MMP-3) and improvements in metabolic outcomes (OPN), continuous glucose concentrations measured for 48 h post-exercise under free-living conditions, were not different to the 24 h pre-exercise glucose concentrations under the current experimental conditions. Further research is warranted to assess the response of OPN and MMPs in response to exercise training, particularly in a cohort of individuals with T2D, along with a deeper characterization of the interaction of the MMPs and the various OPN isoforms.