The role of vascular function on exercise capacity in health and disease

Abstract

Three sentinel parameters of aerobic performance are the maximal oxygen uptake $\left({\dot{V}}_{{\text{O}}_2\text{max}}\right)$, critical power (CP) and speed of the ${\dot{V}}_{{\text{O}}_2}$ kinetics following exercise onset. Of these, the latter is, perhaps, the cardinal test of integrated function along the O₂ transport pathway from lungs to skeletal muscle mitochondria. Fast ${\dot{V}}_{{\text{O}}_2}$ kinetics demands that the cardiovascular system distributes exercise-induced blood flow elevations among and within those vascular beds subserving the contracting muscle(s). Ideally, this process must occur at least as rapidly as mitochondrial metabolism elevates ${\dot{V}}_{{\text{O}}_2}$. Chronic disease and ageing create an O₂ delivery (i.e. $\text{blood\hspace{0.17em}flow\hspace{0.17em}}\times \text{\hspace{0.17em}arterial\hspace{0.17em}}\left[{\text{O}}_2\right],{\dot{Q}}_{{\text{O}}_2}$) dependency that slows ${\dot{V}}_{{\text{O}}_2}$ kinetics, decreasing CP and ${\dot{V}}_{{\text{O}}_2\text{max}}$, increasing the O₂ deficit and sowing the seeds of exercise intolerance. Exercise training, in contrast, does the opposite. Within the context of these three parameters (see Graphical Abstract), this brief review examines the training-induced plasticity of key elements in the O₂ transport pathway. It asks how structural and functional vascular adaptations accelerate and redistribute muscle ${\dot{Q}}_{{\text{O}}_2}$ and thus defend microvascular O₂ partial pressures and capillary blood-myocyte O₂ diffusion across a ~100-fold range of muscle ${\dot{V}}_{{\text{O}}_2}$ values. Recent discoveries, especially in the muscle microcirculation and ${\dot{Q}}_{{\text{O}}_2}$-to-${\dot{V}}_{{\text{O}}_2}$ heterogeneity, are integrated with the O₂ transport pathway to appreciate how local and systemic vascular control helps defend ${\dot{V}}_{{\text{O}}_2}$ kinetics and determine CP and ${\dot{V}}_{{\text{O}}_2\text{max}}$ in health and how vascular dysfunction in disease predicates exercise intolerance. Finally, the latest evidence that nitrate supplementation improves vascular and therefore aerobic function in health and disease is presented.

Graphical Abstract

Three sentinel parameters of aerobic performance are the O₂ uptake $\left({\dot{V}}_{{\text{O}}_2}\right)$ kinetics following the onset of exercise, critical power (CP) or critical speed (CS) (asymptote of the power/speed–time relation for high intensity exercise) and the maximal O₂ uptake $\left({\dot{V}}_{{\text{O}}_2\text{max}}\right)$. The dependence of each parameter on O₂ delivery is highly subject, exercise mode and context dependent. That said, for upright rhythmic cycling or running exercise the boxes apportion the relative importance of cardiac, vascular and mitochondrial O₂ delivery/utilization to each in the untrained state (pre-) and the participation of each in the training adaptation (post-) for each parameter. This brief review explores that dependency in health and disease utilizing exercise training and other conditions such as nitrate supplementation to unveil how vascular function and dysfunction predicate exercise tolerance and intolerance within the scope of these three parameters of aerobic function.

Scope

This brief review will consider vascular health principally from a functional viewpoint. There is an extensive literature detailing with the impact of acute and chronic exercise on endothelial function and vasomotor control. In contrast herein, we ask what are the net down-stream effects of vasomotor function/dysfunction across the continuum from healthy to diseased or aged with respect to muscle O₂ delivery (i.e. $\text{blood\hspace{0.17em}flow,\hspace{0.17em}}\dot{Q},\times \text{\hspace{0.17em}arterial}\left[{\text{O}}_2\right]={\dot{Q}}_{{\text{O}}_2}$) in relation to local metabolic demands $\left({\dot{V}}_{{\text{O}}_2}\right)$. We argue that the ${\dot{Q}}_{{\text{O}}_2}$-to-${\dot{V}}_{{\text{O}}_2}$ ratio, insofar as it determines the microvascular O₂ partial pressure $\left(P_{{\text{O}}_{2\text{mv}}}\right)$ that drives O₂ across the blood-myocyte interface, is of paramount importance for metabolic control, exercise energetics and exercise tolerance. This schema is placed within key parameters of aerobic performance, ${\dot{V}}_{{\text{O}}_2}$ kinetics, critical power (CP)/critical speed (CS) and maximal oxygen uptake $\left({\dot{V}}_{{\text{O}}_2\text{max}}\right)$, in order to understand how vascular function predicates exercise tolerance in health and how ageing and disease erode that function. Each of these parameters provides its own unique window into determining exercise tolerance, its improvement with exercise training and impairment with organic disease. However, the CP parameter is, perhaps, the most powerful predictor of exercise capacity for high intensity (heavy and severe) exercise and thus performance for the bulk of contested athletic events (or activities) longer than 1–2 min duration. Indeed, the overarching importance of ${\dot{V}}_{{\text{O}}_2}$ kinetics and also ${\dot{V}}_{{\text{O}}_2\text{max}}$ may lie in their capacity to influence CP in health and disease. Finally, the mechanistic bases by which exercise training (Fig. 1) and also nitrate supplementation increases exercise capacity in health and disease are assessed from the perspective of their ability to improve muscle ${\dot{Q}}_{{\text{O}}_2}$-to-${\dot{V}}_{{\text{O}}_2}$ and prevent potentially catastrophic falls of $P_{{\text{O}}_{2\text{mv}}}$ during exercise, especially during non-steady state conditions.

Figure 1. Schematic representation of maximal O₂ uptake $\left({\dot{V}}_{{\text{O}}_2\text{max}}\right)$ plotted as a function of the venous or microvascular $P_{{\text{O}}_2}\left(P_{{\text{O}}_{2\text{mv}}}\right)$.

Conflation of perfusive (${\dot{Q}}_{{\text{O}}_2}$, Fick’s principle, ${\dot{V}}_{{\text{O}}_2}$ = $\dot{Q}$ (arterial-venous O₂ difference), continuous and dashed curved lines) and diffusive [$D_{{\text{O}}_2}$, Fick’s law, ${\dot{V}}_{{\text{O}}_2}$ (microvascular $D_{{\text{O}}_2}$), continuous and dashed straight lines from origin] O₂ conductances permits a mechanistic analysis of how these components resolve to achieve a given ${\dot{V}}_{{\text{O}}_2}$. Exercise training (dashed lines) acts to increase both perfusive (${\dot{Q}}_{{\text{O}}_2}$, black arrow 1) and diffusive ($D_{{\text{O}}_2}$, black arrow 2) O₂ conductances by elevating maximal cardiac output, muscle ${\dot{Q}}_{{\text{O}}_2}$ and also $D_{{\text{O}}_2}$ consequent to capillary neogenesis, which increases the longitudinal recruitment of capillary surface area and also helps to constrain/prevent a reduction in mean capillary RBC transit time after training. Notice that a substantial increase in $D_{{\text{O}}_2}$ is required to produce even a small lowering of venous or microvascular $P_{{\text{O}}_2}$ (i.e. increased arterial–venous O₂ difference, downward pointing arrows on x-axis; black arrow is pre- and white arrow is post-training). From Roca et al. (1992) and Poole et al. (2012). See text for additional details.

Expansion of key points and perspectives

1. For moderate intensity exercise in health, the compelling weight of evidence supports that the speed of the ${\dot{V}}_{{\text{O}}_2}$ kinetics is not limited by bulk O₂ delivery $\left({\dot{Q}}_{{\text{O}}_2}\right)$ and there is evidence that this may also be true for heavy and severe-intensity exercise (reviewed by Poole & Jones, 2012). However, in patient populations, aged or very unfit individuals, and especially within fast-twitch fibres, there may be an influence of O₂ supply on metabolic regulation. Also, in health, although the kinetics of ${\dot{Q}}_{{\text{O}}_2}$ is far faster than that of ${\dot{V}}_{{\text{O}}_2}$, it is possible that a less than ideal ${\dot{Q}}_{{\text{O}}_2}$-to-${\dot{V}}_{{\text{O}}_2}$ matching lowers regional microvascular O₂ and intramyocyte partial pressures ($P_{{\text{O}}_{2\text{mv}}}$ and $P_{{\text{O}}_{2\text{im}}}$, respectively) and thereby impacts metabolic control (reviewed by Koga et al. 2014; Heinonen et al. 2015). Indeed, it is possible that the metabolic behaviour of muscles comprised of fast-twitch fibres is, in part, dependent on the presence of a low $P_{{\text{O}}_{2\text{mv}}}$. See below, ‘Evidence that, in health, muscle ${\dot{Q}}_{{\text{O}}_2}$ kinetics is faster than ${\dot{V}}_{{\text{O}}_2}$ kinetics’. Exercise training promotes proportionally greater vascular than mitochondrial adaptations such that ${\dot{Q}}_{{\text{O}}_2}$ is redistributed preferentially towards more oxidative fibres. Following the transition to exercise, this effect raises $P_{{\text{O}}_{2\text{mv}}}$ and lessens the increase in deoxy[Hb+Mb] even in the face of substantially faster ${\dot{V}}_{{\text{O}}_2}$ kinetics (Murias et al. 2010, 2011; Hirai et al. 2012, 2014). See below, ‘A novel vascular-based perspective on the mechanisms by which exercise training decreases metabolic perturbation’.

2. For a given individual, the highest ${\dot{V}}_{{\text{O}}_2}$ achievable where the ${\dot{V}}_{{\text{O}}_2}$ slow component can be stabilized at a submaximal level, generally expressed as a finite power or speed (CP or CS), constitutes the upper extreme of the heavy intensity domain. All work rates or speeds above CP or CS incur an inexorable rise in ${\dot{V}}_{{\text{O}}_2}$ to ${\dot{V}}_{{\text{O}}_2\text{max}}$ and imminent fatigue in time (t_lim) as predicted from the relation:

   $$t_{\text{lim}}=W^{\prime }/P-\text{CP}$$ (1)

   where W′ constitutes a discrete primarily non-O₂-related energy store within the exercising muscles that cannot be reconstituted during severe intensity exercise (reviewed by Gaesser & Poole, 1996; Poole & Jones, 2012). Exhaustion is manifested when the available W′ is depleted. Slow ${\dot{V}}_{{\text{O}}_2}$ kinetics following the onset of exercise necessitates a greater depletion of W′ and ‘sows the seeds for exercise intolerance’ (Rossiter, 2011). Isotonic exercise training increases CP/CS by elevating the metabolic capacity, especially of slow twitch and the more oxidative fast-twitch fibres, delaying recruitment of their less oxidative counterparts. Together with enhanced ${\dot{Q}}_{{\text{O}}_2}$ distribution (greater vascularity, fine-tuning of ${\dot{Q}}_{{\text{O}}_2}$-to-${\dot{V}}_{{\text{O}}_2}$) to these fibres, the cascade of destabilizing metabolic events, which incurs a continuous drain on W′ and drives the ${\dot{V}}_{{\text{O}}_2}$ slow component to ${\dot{V}}_{{\text{O}}_2\text{max}}$ and the individual to imminent exhaustion, is shifted to a higher power/speed after training. These so-called isotonic exercise training paradigms do not generally increase W′. See Graphical Abstract, centre panel.

3. By definition ${\dot{V}}_{{\text{O}}_2\text{max}}$ is confined to the severe intensity domain. Within this domain the ${\dot{V}}_{{\text{O}}_2}$ profile, be it approximately linear, monoexponential or comprised of two discrete dynamic components (i.e. fast + slow), will be dependent on the forcing function (reviewed by Poole & Jones, 2012). Specifically, an incremental or ramp exercise test often induces a linear ${\dot{V}}_{{\text{O}}_2}$. response with a gain of 9–11 ml O₂/watt/min, sometimes evincing a discrete plateau of ${\dot{V}}_{{\text{O}}_2}$ vs. power, the classic definition of ${\dot{V}}_{{\text{O}}_2\text{max}}$ (see Graphical Abstract, right panel). Whereas any constant-load exercise performed in the severe domain (i.e. >CP) will yield ${\dot{V}}_{{\text{O}}_2\text{max}}$ if continued to exhaustion, there are two distinct profiles. For power outputs in the upper portion of this domain, ${\dot{V}}_{{\text{O}}_2}$ increases rapidly and exponentially to ${\dot{V}}_{{\text{O}}_2\text{max}}$, whereas down closer to CP a secondary slow component of the response emerges, perhaps 100–200 s following exercise onset, from the underlying fast kinetics. For exercise performed just above CP where W′ is depleted so slowly that exercise can continue for 15 min or more, the size of the ${\dot{V}}_{{\text{O}}_2}$ slow component can be extraordinary (i.e. 1–1.5 l min⁻¹; reviewed by Gaesser & Poole, 1996; Poole & Jones, 2012). Because of the close association of the ${\dot{V}}_{{\text{O}}_2}$ slow component and exhaustion, the past three decades has seen intense efforts dedicated to understanding its mechanistic bases (Rossiter, 2011). Interventions that delay or reduce the ${\dot{V}}_{{\text{O}}_2}$ slow component at a given power, such as exercise training or breathing hyperoxic inspirates, invariably enhance exercise tolerance.

Unlike the case for moderate or heavy intensity exercise where ‘aerobic’ exercise training either does not change (moderate) or reduces (heavy) the steady-state ${\dot{V}}_{{\text{O}}_2}$ at a given work rate, because ${\dot{V}}_{{\text{O}}_2}$ projects to ${\dot{V}}_{{\text{O}}_2\text{max}}$ for all severe-intensity work rates, exercise training will invariably increase the ${\dot{V}}_{{\text{O}}_2}$ response. Use of the ‘Wagner diagram’ (Fig. 1) facilitates analysis of the training-induced perfusive and diffusive adaptations that conflate to increase the overall ${\dot{V}}_{{\text{O}}_2\text{max}}$ (Roca et al. 1992; Poole et al. 2018). Perfusive elements enhanced by training include cardiac output, selective distribution of that blood flow $\left(\dot{Q}\right)$ to the exercising muscles, enhanced smooth muscle vasodilatation and potentially improved ${\dot{Q}}_{{\text{O}}_2}$-to-${\dot{V}}_{{\text{O}}_2}$ matching within muscles. Diffusive improvements include greater capillarity (i.e. increased capillary-to-fibre ratio) and capillary volume and constrained reduction in capillary red blood cell (RBC) transit time which would otherwise result from the augmented post-training $\dot{Q}$. There is also the likelihood that there is an enhanced longitudinal recruitment of exchange surface area along the length of capillaries as fractional O₂ extraction increases combined with the potential for greater $\dot{Q}$ values to increase capillary haematocrit and therefore reduce inter-RBC spacing (see below, ‘A novel vascular-based perspective on the mechanisms by which exercise training decreases metabolic perturbation’). The Wagner analysis reveals that the augmented diffusive capacity contributes far more quantitatively than surmised simply from the modest widening of the a–vO₂ difference. One intriguing question that remains, especially in young healthy individuals, is where is the site of ${\dot{V}}_{{\text{O}}_2\text{max}}$ limitation? Certainly the answer is very different for small (e.g. knee extension ~2–3 kg muscle per leg) versus large (e.g. conventional cycling ~15 kg) muscle mass exercise with the latter recruiting sufficient muscle to exceed the pumping capacity of the heart, necessitating enhanced sympathetic vasoconstriction within the exercising muscles during maximal exercise. Thus, knee extension exercise has yielded some of the highest muscle $\dot{Q}$ values and ${\dot{V}}_{{\text{O}}_2}$ values ever measured (3.85 l/kg/min, 640 ml/kg/min) in humans (Richardson et al. 1993), moving the site of limitation within the muscle and towards the mitochondria. In contrast, for subjects performing conventional cycling exercise, with approximately the same ${\dot{V}}_{{\text{O}}_2\text{max}}$ of ~60 ml/kg/min, leg muscle $\dot{Q}$ only averaged 1.5 l/kg/min and ${\dot{V}}_{{\text{O}}_2}$ 150 ml/kg/min (Knight et al. 1992).

These observations provided the first clues, perhaps, that different exercise training paradigms could be explored for their utility to benefit distinct patient populations for whom either a pathologically low cardiac output (i.e. in heart failure; HF) or ventilatory (chronic obstructive pulmonary disease; COPD) ceiling constrains the patient to a very suboptimal metabolic stress during conventional cycling or treadmill paradigms. Indeed, Esposito and colleagues (2010, 2011, 2018) have utilized small muscle mass training to restore muscle and whole body exercise capacity in HF patients.

Parameters of exercise performance

The most recognizable parameter of aerobic exercise performance is arguably the ${\dot{V}}_{{\text{O}}_2\text{max}}$, which defines the maximum pulmonary-cardiovascular-muscle O₂ flux attainable during large muscle mass exercise (Wagner, 1996, 2000; Wagner et al. 1997; Poole & Jones, 2012, 2017). The ${\dot{V}}_{{\text{O}}_2\text{max}}$ has proven extremely valuable also in stratifying cardiovascular and other patient populations with respect to the predations of their disease and recommendations for treatments such as heart transplant (Weber et al. 1982; Mancini et al. 1991). However, because individuals rarely exercise to reach ${\dot{V}}_{{\text{O}}_2\text{max}}$, the speed of their ${\dot{V}}_{{\text{O}}_2}$ kinetics and their CP/CS, which denotes the maximal metabolic rate or ${\dot{V}}_{{\text{O}}_2}$ that can be sustained for a prolonged duration, constitute two sentinel aerobic parameters of at least equal importance (Poole et al. 1988, 2016; Barker et al. 2006; Jones et al. 2010; Poole & Jones, 2012; Jones et al. 2019). This contention is supported in healthy individuals and, maybe even more so, for aged and patient populations (Rossiter, 2011). We shall explore the fundamental mechanisms that control the rate of muscle ${\dot{Q}}_{{\text{O}}_2}$, especially relative to that of muscle ${\dot{V}}_{{\text{O}}_2}$, recognizing that it is the instantaneous ${\dot{Q}}_{{\text{O}}_2}$-to-${\dot{V}}_{{\text{O}}_2}$ ratio in any muscle compartment that regulates the $P_{{\text{O}}_{2\text{mv}}}$ and thus the driving force for blood-myocyte O₂ flux across its final frontier to the mitochondria.

In pursuit of brevity reference will be made to state-of-the art reviews that establish foundational principles such as (1) the role of endothelial function and the arterioles in muscle vascular control and adaptations to training (Clifford & Hellsten, 2004; Clifford & Tschakovsky, 2008; Casey & Joyner, 2011; Laughlin et al. 2012; Joyner and Casey, 2015; Campbell et al. 2019), (2) lack of ${\dot{Q}}_{{\text{O}}_2}$ limitations to ${\dot{V}}_{{\text{O}}_2}$ kinetics in young healthy humans but increasing role of ${\dot{Q}}_{{\text{O}}_2}$ deficits with ageing and disease slowing ${\dot{V}}_{{\text{O}}_2}$ kinetics (Rossiter, 2011; Poole & Jones, 2012; Korzeniewski et al. 2018), (3) presence of ${\dot{Q}}_{{\text{O}}_2}$ limitations to ${\dot{V}}_{{\text{O}}_2\text{max}}$ in health and disease (Wagner, 1996, 2000; Wagner et al. 1997; Piepoli et al. 2010a,b; Poole & Erickson, 2011; Poole et al. 2012, 2018; Hirai et al. 2015) and plasticity of the O₂ transport system in response to exercise training (Saltin & Gollnick, 1983; Poole, 1997; Poole & Erickson, 2011; Laughlin et al. 2012; Hirai et al. 2015; Poole et al. 2018), (4) spatial and temporal heterogeneity of muscle(s) deoxygenation during exercise (Vogiatzis et al. 2015; Richardson et al. 2001; Koga et al. 2007, 2014; Heinonen et al. 2015), (5) importance of the CP concept for partitioning the heavy and severe exercise intensity domains and determining exercise performance in health and disease (Jonesetal. 2010,2019;Poole&Jones, 2012; Poole et al. 2016), and (6) role of dietary nitrate supplementation for improving physiological function in health and disease (Ferguson et al. 2013b; Zamani et al. 2015, 2017; Jones et al. 2016, 2018; McDonagh et al. 2019).

Evidence that, in health, muscle ${\dot{Q}}_{{\text{O}}_2}$ kinetics is faster than ${\dot{V}}_{{\text{O}}_2}$ kinetics

Perhaps as late as the 1990s it was widely considered, and taught, that muscle $\dot{Q}$ (and therefore ${\dot{Q}}_{{\text{O}}_2}$) increased largely in response to the vasodilating effects of muscle metabolites generated by contractions. Such a model of delayed vasodilatation and dependence on feedback regulation would have dire consequences for the cardiovascular system. Specifically, as cardiac output increases substantially within a single heartbeat for large muscle mass exercise, if muscle vascular conductance did not increase in synchrony, arterial blood pressure would spike (Poole & Jones, 2012; Joyner & Casey, 2015). Moreover, because of the extremely small O₂ stores within human muscle(capillarybloodvolume~1–2%;Saltin&Gollnick, 1983) and [myoglobin] ~0.5 mM (Reynafarje, 1962; Terrados, 1990; van Beek-Harmsen, 2004) increasing ${\dot{V}}_{{\text{O}}_2}$ without commensurately elevating ${\dot{Q}}_{{\text{O}}_2}$ would decrease $P_{{\text{O}}_{2\text{mv}}}$ precipitously within the contracting muscle thereby compromising blood–myocyte O₂ flux, slowing ${\dot{V}}_{{\text{O}}_2}$ kinetics, drastically depleting high-energy phosphates, accumulating metabolites (e.g. H⁺) and impairing metabolic control.

Grassi and colleagues (1996) adapted the constant-infusion thermodilution technique developed by Andersen and Saltin (1985), combined with rapid sampling of the venous effluent, to measure ${\dot{Q}}_{{\text{O}}_2}$ and ${\dot{V}}_{{\text{O}}_2}$ kinetics across the leg following the onset of moderate-intensity cycling exercise (Fig. 2B, Grassi et al. 1996). Surprisingly, given the then-conventional wisdom, leg $\dot{Q}$ and ${\dot{Q}}_{{\text{O}}_2}$ increased within the first contraction cycle preceding that of ${\dot{V}}_{{\text{O}}_2}$ such that, for several seconds after exercise onset, fractional O₂ extraction actually decreased transiently, raising venous effluent O₂ content. That ${\dot{Q}}_{{\text{O}}_2}$ kinetics was at least as fast, and maybe more so, than ${\dot{V}}_{{\text{O}}_2}$ kinetics was subsequently demonstrated for heavy (Koga et al. 2005: Fig. 2A) and severe-intensity (Bangsbo et al. 2000) exercise. Collectively, these investigations emphasized the importance of the muscle pumping action on perfusion and also raised awareness that rapid vasodilatory mechanisms including hyperpolarization and the effects of mechanical compression (Naik et al. 1999; Clifford & Hellsten, 2004; Clifford et al. 2006; Clifford & Tschakovsky, 2008) may provide an effective increase of ${\dot{Q}}_{{\text{O}}_2}$ in parallel with muscle contraction(s). These almost instantaneous (typically less than 1 s) dilatory mechanisms (Clifford & Hellsten, 2004; Wray et al. 2005; Clifford et al. 2006; Clifford & Tschakovsky, 2008; Behnke & Delp, 2010; McDaniel et al. 2010; Casey & Joyner, 2011; Casey et al. 2013; Hughes et al. 2016, 2018; Sinkler et al. 2016; Sinkler & Segal, 2017) are then supplemented by a spectrum of vasodilators providing a fine-tuning of ${\dot{Q}}_{{\text{O}}_2}$-to-${\dot{V}}_{{\text{O}}_2}$ matching. That such fine-tuning occurs is supported by fractional O₂ extractions as high as 80–90% across healthy muscles during maximal exercise (Knight et al. 1992; Richardson et al. 1993).

Figure 2. Rate of leg blood flow increase compared with that of VO₂ following the onset of heavy (A) and moderate (B) intensity exercise.

A, temporal profile of leg blood flow $\left(\dot{Q}\right)$ (continuous line, measured by ultrasound of femoral artery) compared with that of alveolar ${\dot{V}}_{{\text{O}}_2}$ (dashed line) across the transient from unloaded to heavy intensity knee extension exercise for one subject. Notice that the $\dot{Q}$ response is far faster than that for ${\dot{V}}_{{\text{O}}_2}$. From Koga et al. (2007) and Poole and Jones (2012). B, temporal responses of leg $\dot{Q}$, ${\dot{V}}_{{\text{O}}_2}$ and arterial-venous O₂ difference (a–vO₂ difference) measured by constant-infusion thermodilution in the femoral vein combined with rapid femoral venous and radial arterial sampling for blood-gas analysis across the transition to moderate intensity cycling exercise. Standard errors are omitted for clarity. Note that leg $\dot{Q}$ increases more rapidly than leg ${\dot{V}}_{{\text{O}}_2}$ such that the O₂ content in the femoral venous effluent increases transiently, decreasing a–vO₂ difference (i.e. reduced fractional O₂ extraction). From Grassi et al. (1996). These data exemplify that, for young healthy subjects performing knee extension and conventional cycling exercise, bulk muscle(s) O₂ delivery is faster than muscle ${\dot{V}}_{{\text{O}}_2}$ across the rest–exercise transition for moderate and heavy intensity exercise. Bangsbo et al. (2000) have demonstrated that this is also true for severe-intensity exercise.

However, it was recognized that measurements made across the exercising leg were remote from events occurring at the blood–myocyte interface within those capillaries adjacent to the contracting muscle fibres. Accordingly, Kindig and colleagues (2002) developed an intravital microscopy model of the rat spinotrapezius muscle, a close analogue of the human quadriceps with respect to fibre-type composition and oxidative capacity (cf. Delp & Duan, (1996) and Leek et al. (2001)). As predicted from Grassi et al. (1996) and Bangsbo et al. (2000), RBC velocity and flux increased rapidly within the first contraction cycles of electrically induced rhythmic muscle contractions (Kindig et al. 2002; Fig. 3, filled circles). Pairing these measurements with phosphorescence quenching determination of $P_{{\text{O}}_{2\text{mv}}}$ (Fig. 3, inverted triangles), Behnke et al. (2001, 2002a) resolved that, at the individual capillary level and collectively across the observable capillary bed of healthy muscle, ${\dot{Q}}_{{\text{O}}_2}$. a ${\dot{V}}_{{\text{O}}_2}$ (Fig. 3, filled and open circles, respectively) increased in concert such that, for multiple contraction cycles lasting 10–20 s, $P_{{\text{O}}_{2\text{mv}}}$ was defended near resting levels before decreasing in a close-to-exponential fashion towards its steady-state value reflecting the overall increased a–vO₂ difference (i.e. elevated fractional O₂ extraction) (Ferreira et al. 2006b; Poole et al. 2007). The capability to follow events in the capillary (i.e. RBC distribution and flux as well as oxygenation) set the imperative for better understanding muscle microvascular dysfunction in disease and also the role of exercise training and other therapeutic interventions to improve muscle function and exercise tolerance.

Figure 3. Increase in rat spinotrapezius red blood cell (RBC) flux (filled circles) and microvascular $P_{{\text{O}}_2}$ ($P_{{\text{O}}_2\text{mv}}$, triangles) conflated to estimate oxygen uptake (${\dot{V}}_{{\text{O}}_2}$, open circles) in response to 1 Hz electrically induced contractions (initiated at time = 0 s) normalized to 100% response amplitude.

Data from Behnke and colleagues (2001) and Kindig et al. (2002). Capillary RBC flux was measured by high-resolution videomicroscopy analysis within individual capillaries. $P_{{\text{O}}_{2\text{mv}}}$ was resolved by phosphorescence quenching techniques (Rumsey et al. 1988; Poole et al. 1995; Behnke et al. 2001, 2002a,c). ${\dot{V}}_{{\text{O}}_2}$ was estimated using the measured RBC flux and arterial O₂ content and microvascular $P_{{\text{O}}_2}$ as an approximation of end-capillary $P_{{\text{O}}_2}$ and O₂ dissociation curves established for the rat (Altman & Dittmer, 1974). Model fits are shown. Both RBC flux and ${\dot{V}}_{{\text{O}}_2}$ (but not $P_{{\text{O}}_{2\text{mv}}}$) were fit by a single exponential without time delay. TD, time delay. τ, time constant. The value for τ is close to that for human muscle and also Phase II pulmonary ${\dot{V}}_{{\text{O}}_2}$ kinetics (e.g. Grassi et al. 1996). Figure from Poole (2019).

Because muscle microvascular units (i.e. terminal arterioles and their dependent capillaries) are not spatially matched with the fibres of discrete motor units, it is an enduring dilemma how fractional O₂ extraction can be so high given that quiescent fibres form a mosaic with their recruited counterparts. Is it possible that O₂ may diffuse effectively from fibre-to-fibre in physiologically meaningful quantities? The uniformly low $P_{{\text{O}}_{2\text{im}}}$ values measured in the human quadriceps by proton magnetic resonance spectroscopy supports this notion (Richardson et al. 1995). The recently developed capacity to measure both $P_{{\text{O}}_{2\text{mv}}}$ and muscle interstitial $P_{{\text{O}}_2}$ during muscle contractions may provide important clues as to the viability of this mechanism (Hirai et al. 2018, 2019a).

Evidence for ${\dot{Q}}_{{\text{O}}_2}$ dependence of ${\dot{V}}_{{\text{O}}_2}$ kinetics in disease

In stark contrast to the healthy muscle capillary bed, where the vast majority of capillaries support at least some plasma and RBC flux, intravital microscopy reveals that in type II diabetes and HF a substantial portion of the capillary bed does not (Kindig et al. 1999; Richardson et al. 2003; Padilla et al. 2006; reviewed by Poole et al. 2018; Poole, 2019). Crucially, these non-RBC flowing capillaries do not start flowing when the muscle contracts rhythmically (Richardson et al. 2003), and even in those capillaries that sustain RBC flux at rest, the increase with contractions is extremely sluggish. Thus, following exercise onset, when the energetic demands of the muscle mitochondria and contractile apparatus are accelerating rapidly, ${\dot{Q}}_{{\text{O}}_2}$ is not (Fig. 4A). In muscles of both HF and type II diabetic rats $P_{{\text{O}}_{2\text{mv}}}$ may fall faster and more precipitously than in healthy controls (Fig. 4B; Diederich et al. 2002; Behnke et al. 2002a; Padilla et al. 2007; Hirai et al. 2015). Furthermore there are additional signs of dysfunction with the lowered $P_{{\text{O}}_{2\text{mv}}}$ appearing jagged and far less stable, which likely results from impaired vasomotor function and a loss of the ability to fine-tune regional ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ (Hirai et al. 2019a). Indeed, with disease (e.g. cancer) and disuse conditions (e.g. diaphragm with mechanical ventilation), oscillations in the $P_{{\text{O}}_{2\text{mv}}}$ profile are associated with impaired resistance vessel function (Davis et al. 2012; Horn et al. 2019; McCullough et al. 2013). Importantly, exercise training in disease states, such as cancer, leads to a stable (i.e. less oscillations around the mean value) $P_{{\text{O}}_{2\text{mv}}}$ profile in non-contracting tissue (McCullough et al. 2013), indicating the efficacy of training to fine tune ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$, even in non-skeletal muscle tissue (e.g. tumours). Rapid falls in $P_{{\text{O}}_{2\text{mv}}}$ can account for the progressive slowing of ${\dot{V}}_{{\text{O}}_2}$ kinetics as the O₂ transport pathway, via both perfusive and diffusive O₂ conductances, becomes progressively more dysfunctional with disease severity (i.e. leftwards movement beyond the ‘tipping point’ in Fig. 5). That near infrared spectroscopy (NIRS) has revealed, following the onset of exercise, the existence of a substantial ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ mismatch even in ‘optimally treated’ HF with reduced ejection fraction (HFrEF) patients suggests that $P_{{\text{O}}_{2\text{mv}}}$ represents a key therapeutic target (Sperandio et al. 2009). This finding, coupled with evidence that muscle structural and functional plasticity are retained in HF, emphasizes the potential for exercise training, at least in HFrEF, to improve convective and diffusive O₂ transport and thereby ˙V_O2max and exercise tolerance (Esposito et al. 2010, 2011, 2018). Moreover, high intensity training may yield even greater benefits than the more traditional endurance training paradigms in HFrEF and also HF with preserved ejection fraction (HFpEF) (Wisloff et al. 2007; Adams et al. 2015; Angadi et al. 2015).

Figure 4. Temporal profiles of capillary hemodynamics (A) and microvascular O₂ partial pressure (B) following the onset of contractions at time 0.

A, capillary red blood cell (RBC) flux (means ± SE) in spinotrapezius muscles from healthy and heart failure (HF) rats. Muscles were electrically stimulated at 1 Hz to produce a rhythmic contraction for 180 s. From Kindig et al. (2002) and Richardson et al. (2003). B, spinotrapezius microvascular O₂ partial pressure $\left(P_{{\text{O}}_{2\text{mv}}}\right)$ in different groups of healthy and HF rats across the same transition. Note the far lower $P_{{\text{O}}_{2\text{mv}}}$ in HF especially across the first 60–80 s when the most rapid increases in mitochondrial ${\dot{V}}_{{\text{O}}_2}$ are expected. From Copp et al. (2010) and Poole et al. (2012).

Figure 5. Idealized depiction of the relationship between the speed of the ${\dot{V}}_{{\text{O}}_2}$ kinetics (given by the time constant, τ) and muscle(s) O₂ delivery pre- and post-exercise training.

Note the presence of O₂-dependent (leftwards) and O₂-independent (rightwards) zones falling either side of the ‘tipping point’ (TP). ${\dot{V}}_{{\text{O}}_2}$ kinetics becomes demonstrably slowed in disease (e.g. heart failure, diabetes and COPD) and also with ageing (red downward arrows). Note that exercise training functions to speed ${\dot{V}}_{{\text{O}}_2}$ kinetics and shifts TP rightwards (horizontal red arrow) to higher levels of O₂ dependency. Fortunately, exercise training-induced vascular adaptations, at least in the short term, exceed the mitochondria-induced speeding of ${\dot{V}}_{{\text{O}}_2}$ kinetics such that the latter may limit ${\dot{V}}_{{\text{O}}_2}$ kinetics irrespective of training. It is likely that the tipping point varies across different muscle fibre types and, possibly, for different exercise paradigms. See text for additional details.

A novel vascular-based perspective on the mechanisms by which exercise training decreases metabolic perturbation

Health.

Across the metabolic transition from rest to exercise, the degree of intracellular metabolic perturbation (i.e. Δ[PCr], [ADP_free], [P_i], [H⁺], [NADH⁺]) and rate of glycolysis and glycogen depletion will be dependent on the ${\dot{V}}_{{\text{O}}_2}$ kinetics, with slower kinetics predicating a greater O₂ deficit and therefore perturbation. Following the development of breath-by-breath gas analysis systems, pulmonary ${\dot{V}}_{{\text{O}}_2}$ kinetics has traditionally been modelled as a time delay (TD) followed by an exponential increase to the steady state such that:

$${\dot{V}}_{{\text{O}}_2\left(\text{t}\right)}={\dot{V}}_{{\text{O}}_2\left(\text{BL}\right)}+\Delta {\dot{V}}_{{\text{O}}_2}\left(1-e^{-t-\text{TD}/\mathcal{T}}\right)$$ (2)

Where ${\dot{V}}_{{\text{O}}_2(t)}$ is the ${\dot{V}}_{{\text{O}}_2}$ at time t following the onset of exercise, $\Delta {\dot{V}}_{{\text{O}}_2}$ is the difference between the baseline ${\dot{V}}_{{\text{O}}_2}\left({\dot{V}}_{{\text{O}}_2(\text{BL})}\right)$ and the exercising steady state ${\dot{V}}_{{\text{O}}_2}$, and the parameter T (i.e. time constant) denotes the time necessary for ${\dot{V}}_{{\text{O}}_2}$ to increase to 63% of the difference between rest and exercise. For heavy and severe-intensity exercise the emergence of a secondary slow component of the ${\dot{V}}_{{\text{O}}_2}$ response becomes manifested requiring addition of a second exponential term (Whipp & Wasserman, 1986; Whipp, 1987; Barstow & Mole, 1991; Gaesser & Poole, 1996; Rossiter, 2011; Poole & Jones, 2012).

For moderate intensity exercise this concept is implicit in the inverse relation between the speed of the ${\dot{V}}_{{\text{O}}_2}$ kinetics and the O₂ deficit:

$${\text{O}}_2\text{\hspace{0.17em}deficit}=\Delta {\dot{V}}_{{\text{O}}_2}\times T$$ (3)

The value of T is typically 20–30 s in young, healthy individuals and increases with age and diseases that impact mitochondrial function (e.g. mitochondrial myopathies; Porcelli et al. 2019) and/or the O₂ transport system (e.g. COPD, HF, diabetes; reviewed by Rossiter, 2011; Poole & Jones, 2012). T may be as short as 10 s in elite athletes such as cyclists and distance runners, specifically marathon World-record holder, Paula Radcliffe, and also thoroughbred racehorses (Jones & Poole, 2009; Poole & Erickson, 2011; Poole & Jones, 2012). Whereas the three parameters of aerobic function considered herein, namely T, CP/CS and ${\dot{V}}_{{\text{O}}_2\text{max}}$, do correlate with one another, especially in cohorts expressing a wide range of exercise capacities, they may be regulated independently (Gaesser & Wilson, 1988; Rossiter, 2011; Poole & Jones, 2012).

Why does exercise training speed ${\dot{V}}_{{\text{O}}_2}$ kinetics (reduce T) and decrease the O₂ deficit? Exercise training increases mitochondrial volume density and this means that each respiratory chain has to produce proportionally less ATP (lower ${\dot{V}}_{{\text{O}}_2}$ per unit mitochondria) and therefore requires less metabolic perturbation to drive the same muscle work and pulmonary ${\dot{V}}_{{\text{O}}_2}$ (Saltin & Gollnick, 1983; Holloszy & Coyle, 1984). Consider the essential mitochondrial respiratory equation that establishes the P:O₂ ratio of ~6:

$$6\text{ADP}+6{\text{P}}_{\text{i}}+2\text{NADH}+2{\text{H}}^{+}+{\text{O}}_2\to 6\text{ATP}+2{\text{NAD}}^{+}+2{\text{H}}_2\text{O}$$ (4)

Whereas the steady state ${\dot{V}}_{{\text{O}}_2}$, certainly for exercise of moderate and potentially higher intensities, may not be limited by O₂ availability, the same may not be said for the degree of metabolic perturbation. Specifically, even above the ‘critical $P_{{\text{O}}_2}$’ where ${\dot{V}}_{{\text{O}}_2}$ is not O₂ delivery-dependent, there is an inverse relationship between $P_{{\text{O}}_2}$ and the other substrates expressed in eqn (4). This was demonstrated by Wilson and colleagues (1977) initially in liver mitochondria and later by Hogan, Arthur and colleagues in contracting canine skeletal muscles (Arthur, 1992; Hogan et al. 1992). The modulation of metabolic controllers at supra-critical $P_{{\text{O}}_2}$ values raises the intriguing possibility that key aspects of metabolic control, once considered solely dependent upon the metabolic characteristics of muscle fibres per se, may be subject to modulation by upstream vascular control as this determines the ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ ratio and thus $P_{{\text{O}}_{2\text{mv}}}$ and $P_{{\text{O}}_{2\text{im}}}$. Specifically, according to Fick’s law:

$${\dot{V}}_{{\text{O}}_2}=D_{{\text{O}}_2\text{\hspace{0.17em}m}}\left(P_{{\text{O}}_{2\text{mv}}}-P_{{\text{O}}_2\text{\hspace{0.17em}im}}\right)$$ (5)

where $D_{{\text{O}}_2\text{m}}$ is effective muscle O₂ diffusing capacity as considered by Roca et al. (1992). From eqn (4), for presiding values of ${\dot{V}}_{{\text{O}}_2}$ and $D_{{\text{O}}_2\text{m}}$, raising $P_{{\text{O}}_{2\text{mv}}}$ by elevating the ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ ratio will increase $P_{{\text{O}}_{2\text{im}}}$ and impact metabolic regulation. Could this phenomenon help explain how exercise training decreases the degree of metabolic perturbation, speeds ${\dot{V}}_{{\text{O}}_2}$ kinetics, and increases exercise tolerance at a given work rate or ${\dot{V}}_{{\text{O}}_2}$?

In addition to exercise training raising $P_{{\text{O}}_{2\text{mv}}}$ (and likely $P_{{\text{O}}_{2\text{im}}}$), growth of additional capillaries via capillary neogenesis increases capillary volume density (Saltin & Gollnick, 1983; Poole & Mathieu-Costello, 1996) and thus the number of RBCs adjacent to contracting muscle fibres, which is believed to be the primary determinant of $D_{{\text{O}}_2\text{m}}$ in eqn (5) (Federspiel & Popel, 1986; Groebe & Thews, 1990). It is also possible that, as capillary haematocrit increases with hyperaemia (Klitzman & Duling, 1979; Kindig et al. 2002; Hirai et al. 2015), and muscle $\dot{Q}$ during maximal exercise is elevated by exercise training, there is a greater number of RBCs per unit length of capillaries in trained muscles. Together, these effects produce a substantial (34%) increase in $D_{{\text{O}}_2\text{m}}$ which is proportionally greater than the increased $\dot{Q}$ (19%). Thus, in Roca et al.’s investigation, compared to increased $\dot{Q}$, the rise of $D_{{\text{O}}_2\text{m}}$ contributed more than twice as much of the training-induced increase in ${\dot{V}}_{{\text{O}}_2\text{max}}$ (Fig. 6; Roca et al. 1992). Interestingly, this magnitude of $D_{{\text{O}}_2\text{m}}$ increase only resulted in a modest ~10% elevation of fractional O₂ extraction. A question that may be asked, in light of the proportionally greater increase in $D_{{\text{O}}_2\text{m}}$ vs. $\dot{Q}$ with training and the incredibly high $\dot{Q}$ capacity of skeletal muscle when cardiac output is not limited (see Richardson et al. 1993) is whether vascular function in skeletal muscle needs to be improved with exercise training. In other words, if $\dot{Q}$ did not increase after exercise training, could increases in $D_{{\text{O}}_2\text{m}}$ elicit the same increase in ${\dot{V}}_{{\text{O}}_2\text{max}}$?

Figure 6. Relationship between leg ${\dot{V}}_{{\text{O}}_2}$ at ${\dot{V}}_{{\text{O}}_2\text{max}}$ and calculated mean capillary $P_{{\text{O}}_2}$ pre- (filled symbols) and post- (open symbols) training in humans breathing 12% (left), 15% (middle) and 21% inspired O₂ (right).

$D_{{\text{O}}_2\text{m}}$ is calculated muscle diffusing capacity, which increased substantially with training. See text for additional details. From Roca et al. (1992).

Surprisingly small changes in vascular function and/or vessel diameter can elevate $\dot{Q}$ substantially. Specifically, based upon Poiseuille’s law, with all other variables (length, perfusion pressure, blood viscosity) held constant, ~18–19% vasodilatation will double $\dot{Q}$ through a vessel. Therefore, to elicit an ~20% elevation of $\dot{Q}$, vessel radius would only have to increase by a mere 5% to account for almost all of the exercise training-induced gains in whole-muscle $\dot{Q}$. Although this is an oversimplification of training-induced changes in $\dot{Q}$, it serves to illustrate that even a minimal enhancement of vasodilatation (through exercise training, nitrate or other pharmacological intervention) can elicit increases in vessel diameter that result in substantial elevations in $\dot{Q}$. Conversely, an elevated constriction or entrenchment of a given blood vessel can greatly diminish $\dot{Q}$.

To address the question of the necessity of training-induced improvements in vascular function and $\dot{Q}$, we should first determine how much $D_{{\text{O}}_2\text{m}}$ would have to increase to reach the same post-training increase in ${\dot{V}}_{{\text{O}}_2\text{max}}$ if $\dot{Q}$ and/or vascular function was not impacted (enhanced) post-training. In short, even if $D_{{\text{O}}_2\text{m}}$ could increase to infinity, which it clearly cannot, at a pre-training fractional O₂ extraction of 0.85 (i.e. 85%), there is not enough available O₂ remaining at the end of the capillaries to allow this to happen as seen from the relationship in eqn (6) (Roca et al. 1992):

$${\text{O}}_2\text{\hspace{0.17em}extraction}={\dot{V}}_{{\text{O}}_2}/{\dot{Q}}_{{\text{O}}_2}={\dot{Q}}_{{\text{O}}_2}\left[1-{\text{exp}}^{\left(D_{{\text{O}}_2\text{\hspace{0.17em}m}}/\beta \dot{Q}\right)}\right]$$ (6)

where β is the slope of the O₂ dissociation curve in the applicable range. Note that any training-induced increase in fractional O₂ extraction means that $D_{{\text{O}}_2\text{m}}$ must increase proportionally more than $\dot{Q}$ otherwise O₂ extraction will fall (this is also implicit in Fig. 1). This relationship also explains why in HF fractional O₂ extraction at a given ${\dot{V}}_{{\text{O}}_2}$ may actually be elevated. Specifically, even though HF reduces $D_{{\text{O}}_2\text{m}}$, $\dot{Q}$ is decreased proportionally more. In addition to enhancing muscle $\dot{Q}$ capacity, at the bulk scale, post-training there are considerable adjustments to the kinetics of conduit and resistance artery vasodilatation and constriction (Shoemaker et al. 1996; Murias et al. 2010, 2011; Hirai et al. 2012, 2014). We propose that these changes (e.g. more rapid and sensitive arteriolar vasodilatation) facilitate the fine-tuning of the cardiovascular system to better match the metabolic demands (i.e. facilitate a tighter metabolic control via elevated ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$), especially across the on-exercise transition.

Mechanistically, measurements of steady-state $\dot{Q}$, typically made across the whole muscle, provide a perfusive snapshot in time. Potentially more valuable insights can be gained by tracking the rate of change in $\dot{Q}$ and/or dilatation following the onset of exercise. As discussed previously, exercise training speeds ${\dot{V}}_{{\text{O}}_2}$ kinetics, and, without a concomitant speeding of $\dot{Q}$ kinetics, a precipitous fall in $P_{{\text{O}}_{2\text{mv}}}$ would occur, sowing the seeds for premature exhaustion (Rossiter, 2011; Poole et al. 2012, 2018; Hirai et al. 2015). Although vasomotor kinetics of resistance and conduit arteries is an under-studied field, several investigations demonstrate an exercise-training induced acceleration in the rate of arterial vasodilatation. One of the first studies to investigate such dilatation kinetics came from the work of Shoemaker and colleagues (1996) who reported faster femoral blood velocity kinetics following exercise onset that was evinced with as little as 10 days of exercise training. Although they did not measure the rate of smooth muscle relaxation per se, as cardiac output kinetics was not significantly different with this short training period, the results suggest a faster vasodilatation occurring within the exercising muscle post-training. More recently, work from Casey and colleagues has found that, in older individuals, exercise training abolished age-related differences in rapid onset vasodilatation (ROV) in conduit arteries (both brachial and femoral), arguably through a speeding of dilatory kinetics (Casey et al. 2013). This is important as old age slows the dilatation of conduit arteries (Hughes et al. 2016, 2018), feed arteries (Park et al. 2016) and resistance arterioles (Behnke & Delp, 2010), as well as disrupts the coordination of up- and down-stream resistance vessel vasomotor responses by attenuating ascending vasodilatation via blunted hyperpolarization conduction (Sinkler et al. 2016). The latter is intricately involved in ROV (Crecelius et al. 2013) and, with old age, greater H₂O availability increases Ca² 2 +-activated K⁺ channel function that impairs endothelium electrical conduction (Socha et al. 2015). Importantly, such blunted dilatation kinetics with age can lead to a ${\dot{Q}}_{{\text{O}}_2}$ limitation to ${\dot{V}}_{{\text{O}}_2}$ kinetics in skeletal muscle (Poole & Jones, 2012; Poole et al. 2012). Therefore, in addition to the well-established increase in $\dot{Q}$ capacity of skeletal muscle post-training, the ability to augment $\dot{Q}$ rapidly is also enhanced and this occurs across muscles of disparate fibre types and oxidative capacities. The kinetics of resistance artery vasodilatation, to several endothelium-dependent agonists (e.g. flow-induced, acetylcholine), varies depending upon the fibre type of their parent muscle (e.g. predominantly fast or slow twitch, high versus low oxidative capacity). For instance, dilatation to a single bolus of acetylcholine (1 × 10⁻⁶ M) is considerably slower in the low-oxidative, mostly type-IIB and IID/X white portion of the gastrocnemius muscle (Delp & Duan, 1996) versus either the highly oxidative, type I and IIA red portion of the same muscle (Fig. 7) or the type I fibre soleus muscle (Fig. 7; Behnke & Delp, 2010). $P_{{\text{O}}_{2\text{mv}}}$ profiles from skeletal muscle support the contention that dilatation dynamics are slower in the low-oxidative fibres (white gastrocnemius) as evidenced by the more rapid fall in $P_{{\text{O}}_{2\text{mv}}}$ following the onset of contractions (Behnke et al. 2003; McDonough et al. 2005). Conversely, the soleus, with its faster resistance vessel dilatation dynamics (Fig. 7; Behnke & Delp, 2010), demonstrates a slower fall in $P_{{\text{O}}_{2\text{mv}}}$ which helps maintain a higher blood–myocyte O₂ driving pressure to support fast ${\dot{V}}_{{\text{O}}_2}$ kinetics (Behnke et al. 2003; McDonough et al. 2005; Hirai et al. 2013, 2014, 2015). Importantly, differences in dilatation kinetics in arterioles from muscles of contrasting fibres were abolished in response to an exogenous NO donor (i.e. sodium nitroprusside; Behnke & Delp, 2010). This identifies the endothelium (e.g. hyperpolarization) and/or the bioavailability of NO as the primary site of limitation in the speed of dilatation in skeletal muscle resistance vessels. With exercise training, there is an increased ‘resilience’ of the endothelium in that dilatation can be maintained when exposed to multiple inhibitors (e.g. no change in dilatation with NOS blockade) (Durand et al. 2015), although whether this is vascular branch specific (i.e. first-order versus terminal) or impacts intramuscular/tissue blood flow distribution is unknown. With respect to the latter, in sedentary subjects NOS blockade diminishes ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ matching in skeletal muscle, consistent with that found in upstream arterioles (Spier et al. 2004). However, at the level of the skeletal muscle, NOS blockade after exercise training reduces, but does not abolish, the improved ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ matching (Hirai et al. 2012), indicating that maintenance of tone in one vessel segment may not necessarily indicate preserved intra-tissue blood flow distribution. A central mechanism for this exercise-induced microcirculatory phenotype resides with the transcriptional coactivator endothelial peroxisome proliferator-activated receptor γ coactivator 1-α (PGC-1α), given its upregulation in human arterial endothelial cells with exercise training (Kim et al. 2014). For example, Kadlec et al. (2018) recently demonstrated that loss of PGC-1α (via siRNA) eliminates vasodilator robustness in arterioles from human athletes. Conversely, forced upregulation of PGC-1α mimics an exercise-trained phenotype in arteries from sedentary subjects and patients with coronary artery disease (Kadlek et al. 2017). Given the well-documented upregulation of endothelial function with exercise training, a speeding of vasodilatory dynamics would be the expected corollary. Accordingly, multiple direct and indirect measures of muscle $\dot{Q}$ dynamics and arterial/arteriolar dilatation evince a more rapid increase in muscle $\dot{Q}$ across the critical rest-to-exercise transition. This more rapid perfusion post-training ensures that the faster post-training ${\dot{V}}_{{\text{O}}_2}$ kinetics does not outstrip the available O₂ supply (see ‘tipping points’ preversus post-training in Fig. 5).

Figure 7. Dynamic vasodilatation response of isolated arterioles from the red and white gastrocnemius and soleus of young rats to direct acetyl choline (1 × 10⁻⁶ M) application (Behnke & Delp, 2010).

Note the significantly faster response for the arterioles from the oxidative muscles (soleus, red gastrocnemius) versus that of the low oxidative white gastrocnemius. τ denotes the time to 63% of the final response; SS is the estimated time to steady state.

${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ matching and heterogeneity

An overarching precept in cardiovascular physiology is that $\dot{Q}$, and thereby ${\dot{Q}}_{{\text{O}}_2}$, is distributed to different organs and tissues in relation to their specific needs, be they substrate delivery, metabolite removal, signalling function or, most often, metabolic rate (i.e. ${\dot{V}}_{{\text{O}}_2}$, reviewed by Koga et al. 2014, 2015; Heinonen et al. 2015). With respect to locomotory muscles during treadmill running in rats, there is a profound $\dot{Q}$ heterogeneity among and within different muscles that is based, in part, upon their fibre type, oxidative capacity and recruitment profiles(Laughlin & Armstrong, 1982; Armstrong & Laughlin, 1984; Poole et al. 2000; Laughlin et al. 2012). Across multiple superficial sites in the human quadriceps, there may be extensive dynamic heterogeneity in the deoxy[Hb+Mb] profile within and across subjects, which undermines confidence in the global interpretation of single-site NIRS measurements. For instance, as seen in Fig. 8, Koga et al. (2007) demonstrated that, even across a cohort of young healthy subjects performing heavy intensity cycling exercise, there was a profound difference in the kinetic profiles of muscle deoxygenation within and among individuals. Important unanswered questions include to what extent is this heterogeneity - or lack thereof - consequent to fibre type differences, recruitment patterns and/or differences in power production among contractions across and within subjects? A crucial component of this puzzle is very likely to encompass differences in duty cycle among and within muscles, muscle groups and across subjects. Broxterman and colleagues (2014) have demonstrated that, for the same external work rate, a 20% versus 50% handgrip duty cycle induces a far higher $\dot{Q}$ and CP. Such manipulation of muscle duty cycle, insofar as it alters the perfusive/diffusive O₂ transport balance and exercise tolerance may prove invaluable to patient exercise rehabilitation and other training programs. Another unanswered question is whether, in any given instance, this heterogeneity reflects an advantageous ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ distribution or, akin to the ventilation-to-perfusion ratio (i.e. ${\dot{V}}_A/\dot{Q}$) in the lung, gas exchange is most effectively subserved when there is a uniform ratio close to 1 (Hammond & Hempleman, 1987). At least with respect to speeding of the primary ${\dot{V}}_{{\text{O}}_2}$ kinetics, reduction of this ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ heterogeneity after priming exercise is ineffectual (Saitoh et al. 2009). However, this intervention does speed the overall kinetics following the onset of heavy exercise by reducing the magnitude of the ${\dot{V}}_{{\text{O}}_2}$ slow component (Saitoh et al. 2009; Koga et al. 2011, 2014; reviewed by Poole & Jones, 2012; Fukuoka et al. 2015; Heinonen et al. 2015).

Figure 8. Relative increases in quadriceps muscle deoxygenation (deoxy[Hb+Mb]) (continuous lines) for 10 sites (measured by near infrared spectroscopy) and pulmonary ${\dot{V}}_{{\text{O}}_2}$ (dashed line, note comparatively slower response) during heavy exercise for 2 different subjects.

Values are normalized to end-exercise increase from resting baseline. Subjects shown had the least (top) and most (bottom) inter-site heterogeneity across the group. Thick line denotes response at the single site most often studied (Koga et al. 2007; Poole & Jones, 2012). See text for further details.

It is pertinent that, in electrically stimulated contracting rat muscles, deeper slow-twitch muscles sustain an appreciably higher ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ ratio versus more superficial fast-twitch muscles (Behnke et al. 2003; McDonough et al. 2005). In agreement with this notion, recent use of positron emission tomography (Heinonen et al. 2007, 2015; Laaksonen et al. 2013) and also NIRS (Koga et al. 2007, 2014, 2015, 2017; Okushima et al. 2015) has demonstrated that, in humans, deeper, more oxidative muscles have higher $\dot{Q}$ values and ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ ratios (lower deoxy[Hb+Mb]) during exercise than their more superficial, less oxidative counterparts. This is supported by the greater sympathetic constraint of $\dot{Q}$ in muscles comprised predominately of fast-twitch, low oxidative fibre types (many superficial muscles) versus muscle of more oxidative, typically deeper (e.g. soleus, red-portion of the gastrocnemius) fibres (Behnke et al. 2011). This scenario predicates a far higher steady-state $P_{{\text{O}}_{2\text{mv}}}$ in slow-twitch muscles with implications for differential metabolic control as implicit in eqns (4) and (5).

Except for exercise in the severe-intensity domain where the presence of a ${\dot{V}}_{{\text{O}}_2}$ slow component drives ${\dot{V}}_{{\text{O}}_2}$ to ${\dot{V}}_{{\text{O}}_2\text{max}}$ (Poole et al. 1988; Gaesser & Poole, 1996; Poole & Jones, 2012, 2016), exercise training does not increase bulk $\dot{Q}$ to skeletal muscles at a given submaximal work rate. However, this intransigence of bulk $\dot{Q}$ belies a training-induced redistribution of ${\dot{Q}}_{{\text{O}}_2}$ from less to more oxidative muscles and muscle parts (Armstrong & Laughlin, 1984; Kent et al. 2016) and also an upregulation of arteriolar vasomotor function (Shoemaker et al. 1996; reviewed by Laughlin et al. 2012) that may speed $\dot{Q}$ kinetics following the onset of exercise. It is pertinent to acknowledge that, as exercise training substantially speeds ${\dot{V}}_{{\text{O}}_2}$ kinetics, if $\dot{Q}$ and ${\dot{Q}}_{{\text{O}}_2}$ kinetics, following exercise onset, did not accelerate with training the ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ ratio and thus $P_{{\text{O}}_{2\text{mv}}}$ would be lower after training, at least across the on-exercise transition. Hirai and colleagues (2012) and also McCullough et al. (2011) showed that this was not the case. Specifically, exercise-trained rats evinced a slower decline of $P_{{\text{O}}_{2\text{mv}}}$ (i.e. higher ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ ratio) following the onset of muscle contractions supporting that the training-induced plasticity of vascular function exceeds that of mitochondrial ${\dot{V}}_{{\text{O}}_2}$ kinetics. Interestingly, only a portion of the $P_{{\text{O}}_{2\text{mv}}}$ training response could be attributed to greater NO bioavailability (Hirai et al. 2012).

In an intriguing series of studies Murias and colleagues have examined the temporal relationships among NIRS-measured muscle deoxygenation (i.e. deoxy[Hb+Mb] - a corollary of $P_{{\text{O}}_{2\text{mv}}}$ that also includes myoglobin) and pulmonary ${\dot{V}}_{{\text{O}}_2}$ kinetics in humans (Murias et al. 2010, 2011; McLay et al. 2017). In the untrained condition there was a temporal mismatch between the ${\dot{V}}_{{\text{O}}_2}$ (slower) and deoxy[Hb+Mb] (faster) such that the deoxy[Hb+Mb]/${\dot{V}}_{{\text{O}}_2}$ ratio of the normalized responses increased above unity. After training, the temporal profiles of the kinetics overlaid one another such that the ratio of the normalized responses did not increase above unity. It was also demonstrated that rat aorta, iliac and femoral vessels dilated more rapidly after high intensity training in diabetic rats (Murias et al. 2013). Whereas these data collectively were interpreted as evidence that ${\dot{Q}}_{{\text{O}}_2}$ kinetics limit ${\dot{V}}_{{\text{O}}_2}$ kinetics, the strength of this conclusion is undermined by (1) the relationship being correlative and based solely upon a ratio of normalized values, (2) the NIRS measurements being restricted to the superficial vastus lateralis, the deoxygenation profile of which is not emblematic of the rest of the quadriceps, especially the deep, most metabolically active portion (Okushima et al. 2015; Koga et al. 2015, 2017), and (3) that across a range of models from the exercising human (MacDonald et al. 1997, 2000; Koga et al. 2019) to the electrically stimulated canine gastrocmemius-plantaris complex (Grassi et al. 1998), increasing either ${\dot{Q}}_{{\text{O}}_2}$ or arterial O₂ content does not speed ${\dot{V}}_{{\text{O}}_2}$ kinetics (reviewed by Grassi, 2000; Rossiter, 2011; Poole & Jones, 2012). Moreover, although microvascular $\dot{Q}$ kinetics may not reflect closely that of bulk muscle(s) $\dot{Q}$ (Harper et al. 2006), it remains to be explained how an inherently faster process $\left(\dot{Q}\right)$ can limit a slower one (i.e. ${\dot{V}}_{{\text{O}}_2}$, Fig. 2).

In summary, there are thus at least five mechanisms by which upstream vascular control may participate, in coordination with increased muscle mitochondrial volume and capacity, to improve exercise tolerance for heavy and severe-intensity exercise after training. The first is speeding overall ${\dot{V}}_{{\text{O}}_2}$ kinetics by accelerating the primary response and reducing the size of the ${\dot{V}}_{{\text{O}}_2}$ slow component and, thus, decreasing the steady-state ${\dot{V}}_{{\text{O}}_2}$ up to several hundred millilitres per minute at any given heavy intensity work rate (Poole et al. 1994, 2016; Poole & Jones, 2012; Gaesser et al. 1994; Womack et al. 1995; Jones et al. 2010). Training also elevates CP/CS increasing the upper extremes of the heavy intensity domain in which ${\dot{V}}_{{\text{O}}_2}$ can be stabilized (Poole & Jones, 2012). Thus, work rates that were in the severe domain pre-training that constitute heavy (i.e. <CP) post-training will no longer drive ${\dot{V}}_{{\text{O}}_2}$ to ${\dot{V}}_{{\text{O}}_2\text{max}}$, and will result in a substantially lower, and submaximal, end-exercise ${\dot{V}}_{{\text{O}}_2}$ that is associated with greatly enhanced exercise tolerance. The second is redistributing $\dot{Q}$ to more oxidative muscle fibres (Armstrong & Laughlin, 1984; Laughlin et al. 2012; Kent et al. 2016). The third is improving the sensitivity and speed of arteriolar relaxation, which raises $P_{{\text{O}}_{2\text{mv}}}$ (McCullough et al. 2011; Hirai et al. 2012, 2014) and likely $P_{{\text{O}}_{2\text{im}}}$, allowing for tighter metabolic control (less intramyocyte perturbation, decreasing glycolysis and the rate of utilization of finite and limiting glycogen stores). The fourth is increasing muscle capillary volume density (Saltin & Gollnick, 1983; Poole & Mathieu-Costello, 1996) and, therefore, the RBCs adjacent to contracting muscle fibres. This, combined with any elevation of capillary haematocrit and better longitudinal capillary recruitment, will raise $D_{{\text{O}}_2\text{m}}$ and potentially elevate fractional O₂ extraction (Federspiel & Popel, 1986; Groebe & Thews, 1990; Roca et al. 1992; Poole et al. 2013; Poole, 2019). The fifth is greater cardiovascular capacity, which enables exercising skeletal muscles to receive more $\dot{Q}$, which conflates with the greater $D_{{\text{O}}_2\text{m}}$ to raise ${\dot{V}}_{{\text{O}}_2\text{max}}$ (Fig. 1). It is notable that, after training, CS, when expressed as ${\dot{V}}_{{\text{O}}_2}$, moves closer to absolute ${\dot{V}}_{{\text{O}}_2\text{max}}$ (i.e. >90% versus 70–80% ${\dot{V}}_{{\text{O}}_2\text{max}}$ in less well-trained individuals; Poole et al. 1988; Jones et al. 2010; Poole et al. 2016). This compression of the severe-intensity exercise domain is epitomized by the extraordinary distance runner Paula Radcliffe, whose women’s marathon record (2 h 15 min 25 s) stood for over 16 years until 2019 (Jones & Poole, 2009).

Ageing and disease.

Whereas exercise training speeds ${\dot{V}}_{{\text{O}}_2}$ kinetics, increases CP/CS and elevates ${\dot{V}}_{{\text{O}}_2\text{max}}$, the predations of ageing and chronic diseases such as HF and diabetes do the obverse. In particular for ${\dot{V}}_{{\text{O}}_2}$ kinetics, this can be appreciated in Fig. 5 where ${\dot{Q}}_{{\text{O}}_2}$ on the x-axis moves leftwards into the O₂ delivery-dependent region where T increases (i.e. slowed ${\dot{V}}_{{\text{O}}_2}$ kinetics) as ${\dot{Q}}_{{\text{O}}_2}$ (and, of course, $P_{{\text{O}}_{2\text{mv}}}$) continue to decrease. Given the pivotal importance of $P_{{\text{O}}_{2\text{mv}}}$, as considered above, it is instructive to consider the impact of ageing and disease and the degree to which skeletal muscle retains its plasticity in response to training or exercise rehabilitation in these conditions.

For any given increase in ${\dot{V}}_{{\text{O}}_2}$ or work rate, reductions in the ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ ratio exacerbate the fall in $P_{{\text{O}}_{2\text{mv}}}$ in animal muscles (HF, Diederich et al. 2002; diabetes, Behnke et al. 2002b; Padilla et al. 2007; ageing, Behnke et al. 2007; McCullough et al. 2011) and its equivalent, increase in deoxy[Hb+Mb], in human muscles (HF, Sperandio et al. 2009, 2012; diabetes, Wilkerson et al. 2011). This phenomenon may also be present, and help account for the diaphragmatic failure that occurs, in patients following prolonged mechanical ventilation, a pressing medical concern (Davis et al. 2012; Horn et al. 2019). As is evident from eqn (5), even if $D_{{\text{O}}_2\text{m}}$ were to remain unchanged in disease, and it is very likely to fall with reductions in the proportion and volume density of flowing capillaries (Kindig et al. 1999; Richardson et al. 2003; Russell et al. 2003; Padilla et al. 2006), maintenance of O₂ flux $\left({\dot{V}}_{{\text{O}}_2}\right)$ will mandate a fall in $P_{{\text{O}}_{2\text{im}}}$ (Richardson et al. 1995). The metabolic consequences of even numerically modest changes in $P_{{\text{O}}_{2\text{im}}}$ are profound as demonstrated by Richardson and colleagues (1998, 1999) using hypoxic and hyperoxic inspirates. Specifically, just reducing $P_{{\text{O}}_{2\text{im}}}$ from 3.1 to 2.3 mmHg by breathing 10% O₂ increased lactate efflux from submaximally exercising muscle several-fold, although, as mentioned by the authors, part of this effect may have been mediated via hypoxia raising circulating catecholamines (Richardson et al. 1998). Conversely, increasing $P_{{\text{O}}_{2\text{im}}}$ from 3.0 to 4.1 mmHg by breathing 100% O₂ was associated with elevating muscle ${\dot{V}}_{{\text{O}}_2\text{max}}$ by 19% (Richardson et al. 1999). It is important to consider that exercise performance is exquisitely sensitive to altered inspired O₂, with hypoxia (10–15% O₂) dramatically compromising and hyperoxia (e.g. 100% O₂) enhancing exercise tolerance (Amann et al. 2006; Katayama et al. 2007). Consequently, CP is reduced by inspiratory hypoxia and increased by hyperoxia (Vanhatalo et al. 2010; Simpson et al. 2015; Goulding et al. 2019). One pertinent caveat here is that, as Peter Wagner has elucidated, systemic O₂ flux is subject to multiple sites of limitation each of which may assume greater, or lesser, importance under different circumstances (Wagner, 1996, 2000; Wagner et al. 1997). Accordingly, carbon monoxide-induced hypoxia may not appreciably alter arterial $P_{{\text{O}}_2}$, $P_{{\text{O}}_{2\text{im}}}$ or muscle ${\dot{V}}_{{\text{O}}_2}$ at submaximal exercise intensities, but ${\dot{V}}_{{\text{O}}_2\text{max}}$ will invariably be decreased (Richardson et al. 2002).

Although there are numerous pathways by which ageing and disease impact vascular function, loss of NO bioavailability, whether through production or clearance, is implicated across most major disease conditions in skeletal muscle. Irrespective of the impact of ageing or disease on the NO synthases, blockade of NO synthase by nitro-l-arginine methyl ester, for example, reveals that NO bioavailability and hence its ability to defend ${\dot{Q}}_{{\text{O}}_2}$ -to- ${\dot{V}}_{{\text{O}}_2}$ matching and $P_{{\text{O}}_{2\text{mv}}}$, especially in HF, is compromised (Hirai et al. 1994, 1995; Ferreira et al. 2006). Combined with a reduction of sympathetic vasoconstriction, lessened humourally mediated vasoconstriction and its anti-inflammatory and antioxidant effects (reviewed by Piepoli, 2010a,b; Poole et al. 2012; Hirai et al. 2015) it is evident that exercise training enhances the muscle ${\dot{Q}}_{{\text{O}}_2}$ -to- ${\dot{V}}_{{\text{O}}_2}$ ratio and thus $P_{{\text{O}}_{2\text{mv}}}$, at least across the rest-contractions transition in HF (Fig. 9) as Murias et al. (2011) (decreased muscle deoxygenation) have evidenced in healthy humans. We consider that this higher $P_{{\text{O}}_{2\text{mv}}}$, in concert with elevated $D_{{\text{O}}_2\text{m}}$, is likely to account, at least in part, for the improved aerobic parameters (i.e. faster ${\dot{V}}_{{\text{O}}_2}$ kinetics, greater CP and ${\dot{V}}_{{\text{O}}_2\text{max}}$) that, in turn, mitigate improved exercise tolerance in health as well as in HF and other patient populations after training.

Figure 9. Effect of exercise training in rats with heart failure on temporal profile of microvascular O₂ partial pressure following the onset of contractions (time 0, A,B) and putative effect on ${\dot{V}}_{{\text{O}}_2}$ kinetics (C).

A, microvascular $P_{{\text{O}}_2}$ in the rat spinotrapezius muscle during electrical stimulation, beginning at time 0, in sedentary (i.e. pre-training, open symbols) and treadmill-trained heart failure (HF, filled symbols) rats with reduced ejection fraction (HFrEF). These training responses are qualitatively similar to those seen also in healthy rats pre- and post-training (Hirai et al. 2012). B, the magnitude of the increased $P_{{\text{O}}_{2\text{mv}}}$ evoked by training at each point in the transition. Note that the largest increase occurs over that interval when mitochondrial energetics and thus ${\dot{V}}_{{\text{O}}_2}$ are expected to be increasing most rapidly. C, theoretical depiction of ${\dot{V}}_{{\text{O}}_2}$ kinetics pre- and post-training (Hirai et al. 2015). Murias and colleagues (2011) have correlated the reduced deoxy[Hb+Mb] post-training with faster ${\dot{V}}_{{\text{O}}_2}$ kinetics.

Alternative nitric oxide pathway therapeutic interventions that improve physiological function, in part, by raising $P_{{\text{O}}_{2\text{mv}}}$

We are not suggesting that elevating $P_{{\text{O}}_{2\text{mv}}}$ in and of itself is responsible for all, or even most, of the improvements in physiological function after training. However, as reductions in muscle oxygenation (whether measured as ${\dot{Q}}_{{\text{O}}_2}$, $P_{{\text{O}}_{2\text{mv}}}$ or $P_{{\text{O}}_{2\text{im}}}$) impair, and increasing muscle oxygenation improves, ${\dot{V}}_{{\text{O}}_2}$ kinetics, CP, ${\dot{V}}_{{\text{O}}_2\text{max}}$ and exercise tolerance, it is useful to evaluate the efficacy of select therapeutic countermeasures to elevate $P_{{\text{O}}_{2\text{mv}}}$. Such interventions may be extremely valuable for individuals who cannot, or will not, engage in exercise rehabilitation/training. For instance, in the USA today it is notable that only a dismally small proportion of patients, for whom cardiac rehabilitation is prescribed, complete their course of treatment. This situation is tragic because improvements in exercise tolerance and ${\dot{V}}_{{\text{O}}_2\text{max}}$ in HF patients substantially elevate life quality, decrease morbidity and mortality, and reduce hospital visits (Fleg et al. 2015).

Reduced nitric oxide (NO) bioavailability in HF and other diseases occurs consequent to NO synthase downregulation and/or uncoupling together with the proliferation of reactive oxygen species that pirate NO and convert it to peroxynitrite (Poole et al. 2012; Hirai et al. 2015; Coggan & Peterson, 2016; Zamani et al. 2015, 2017; Sharma & Kass, 2017). This reduced NO bioavailability to the upstream resistance vasculature is believed to be the primary cause of stopped capillaries and compromised muscle $P_{{\text{O}}_{2\text{mv}}}$ and $D_{{\text{O}}_2\text{m}}$ in HF (Ferreira et al. 2006; Hirai et al. 2015), but whether increasing NO bioavailability exogenously re-establishes RBC flux in stopped capillaries in HF remains to be conclusively resolved. Notwithstanding the above, targeting the NO pathway via either direct sodium nitroprusside or nitrite superfusion (Ferreira et al. 2006; Colburn et al. 2017), nitrite infusion (Glean et al. 2015; Ferguson et al. 2016a) or dietary nitrate supplementation does increase $\dot{Q}$ and $P_{{\text{O}}_{2\text{mv}}}$ in contracting healthy and HF rat muscles (Ferguson et al. 2013a,b, 2015, 2016b; Jones et al. 2016; reviewed by Hirai et al. 2015; Poole, 2019). NO raises $P_{{\text{O}}_{2\text{mv}}}$ by not only elevating ${\dot{Q}}_{{\text{O}}_2}$, as discussed above, but also decreasing ${\dot{V}}_{{\text{O}}_2}$ (Larsen et al. 2007). There is evidence that this latter effect results from a combination of improved mitochondrial efficiency (Larsen et al. 2011) and decreased ATP cost of force production (Bailey et al. 2010). Increased NO bioavailability also enhances muscle contractile function in vitro especially in fast-twitch fibres by improving calcium handling (Hernandez et al. 2012) and speeding ${\dot{V}}_{{\text{O}}_2}$ kinetics in vivo for exercise that specifically recruits more fast twitch fibres (Breese et al. 2013; Bailey et al. 2015). The effects of nitrate and nitrite supplementation are most pronounced in fast twitch muscles (Ferguson et al. 2013b; 2015; Jones et al. 2016), possibly because their lower $P_{{\text{O}}_{2\text{mv}}}$ (Behnke et al. 2003; McDonough et al. 2005) and pH help drive the final reduction of nitrite to NO. Importantly, elegant single-fibre experiments in mouse fast-twitch muscle demonstrate improved sarcoplasmic reticulum calcium handling and myofilament calcium sensitivity at physiologically relevant $P_{{\text{O}}_2}$ values (Bailey et al. 2019). It is pertinent that HF patients are more dependent on recruitment of fast-twitch fibres than their healthy counterparts (Piepoli et al. 2010a,b; Poole et al. 2012; Hirai et al. 2015).

Collectively, these investigations and others in healthy rats and humans and HF rats laid the groundwork for human studies in HF. Consequently, there are now convincing data supporting that nitrate and nitrite supplementation both elevate cardiac output, ${\dot{Q}}_{{\text{O}}_2}$, ${\dot{V}}_{{\text{O}}_2\text{max}}$ and exercise tolerance in HFpEF patients (Zamani et al. 2015, 2017). These patients are inarguably O₂ delivery limited and this may account for nitrate/nitrite supplementation not reducing exercising ${\dot{V}}_{{\text{O}}_2}$ as seen for healthy counterparts. Specifically, greater O₂ availability likely allowed a reduction in the pathologically induced O₂ deficit for these patients. It will be important to demonstrate whether or not nitrate/nitrite supplementation also speeds ${\dot{V}}_{{\text{O}}_2}$ kinetics in these patients as Fig. 5 hypothesizes. Greater NO bioavailability in HFpEF patients does increase exercise efficiency (i.e. lower ${\dot{V}}_{{\text{O}}_2}$/work rate; Zamani et al. 2015, 2017) and this, in and of itself, likely reflects improved contractile function that may be related to improvements in intrafibrillar calcium handling (Bailey et al. 2019), although it is true that nitrate/nitrite this has not been a universal finding in HFpEF (Borlaug et al. 2018) or HFrEF (Hirai et al. 2017) patients. One valid concern is that NO is a reactive O₂ species (ROS) that can form the pernicious ROS peroxynitrite via reaction with superoxide, which itself is expected to be elevated in HF and other diseases. Alternative solutions may involve targeting the NO-vasodilatation pathway down-stream of NO itself. Indeed, in HFrEF patients, Sperandio et al. (2012) found that sildenafil decreases muscle deoxygenation (akin to increased $P_{{\text{O}}_{2\text{mv}}}$), speeds ${\dot{V}}_{{\text{O}}_2}$ kinetics and improves exercise tolerance in the severe intensity domain. A newer and potentially very powerful therapeutic avenue for raising muscle ${\dot{Q}}_{{\text{O}}_2}$ and $P_{{\text{O}}_{2\text{mv}}}$ may be via activating and/or stimulating soluble guanylate cyclase directly using Vericiguat and Riociguat (Adempas^™), the latter of which is currently approved for treatment of pulmonary hypertension (Lian et al. 2017). Whereas this approach would presumably not have the advantage of reducing muscle ${\dot{V}}_{{\text{O}}_2}$ at a given exercise level by directly improving contractile function, calcium handling or mitochondrial efficiency as nitrate/nitrite supplementation do, it would elevate muscle O₂ delivery whilst removing the potential for elevated peroxynitrite formation and other pernicious treatment sequelae.

Conclusions

In young, healthy and reasonably fit humans performing rhythmic exercise such as cycling or running, ${\dot{V}}_{{\text{O}}_2}$ kinetics, unlike CP/CS or ${\dot{V}}_{{\text{O}}_2\text{max}}$, does not appear to be limited by muscle O₂ delivery $\left({\dot{Q}}_{{\text{O}}_2}\right)$. This means that arteriolar vasodilatation and muscle pump-induced $\dot{Q}$ increase at least as rapidly as mitochondrial energetics elevates ${\dot{V}}_{{\text{O}}_2}$. However, ageing and also diseases such as HF and diabetes compromise endothelial function impairing exercise-induced vasodilatation thereby slowing and lowering ${\dot{Q}}_{{\text{O}}_2}$ increases and the ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ ratio and therefore decreasing $P_{{\text{O}}_{2\text{mv}}}$. Together this impaired perfusive O₂ conductance conflates with its reduced diffusive counterpart $\left(D_{{\text{O}}_2\text{m}}\right)$ slowing ${\dot{V}}_{{\text{O}}_2}$ kinetics and lowering CP and ${\dot{V}}_{{\text{O}}_2\text{max}}$. Exercise training improves endothelial function, probably to a greater extent than mitochondrial adaptations, elevating both perfusive and diffusive O₂ delivery. These effects conflate to raise the ${\dot{Q}}_{{\text{O}}_2}$ -to-${\dot{V}}_{{\text{O}}_2}$ ratio and $P_{{\text{O}}_{2\text{mv}}}$ thereby accelerating ${\dot{V}}_{{\text{O}}_2}$ kinetics and raising CP/CS and ${\dot{V}}_{{\text{O}}_2\text{max}}$, all of which, in concert with intramyocyte and neuro-muscular adaptations, constitute the bases for improved exercise tolerance after training for both healthy and patient populations. Accumulating evidence demonstrates that dietary nitrate (or nitrite) supplementation and sildenafil-induced phosphodiesterase inhibition both speed ${\dot{V}}_{{\text{O}}_2}$ kinetics, decreasing muscle deoxygenation (raising $P_{{\text{O}}_{2\text{mv}}}$) for submaximal exercise and enhancing ${\dot{V}}_{{\text{O}}_2\text{max}}$ in healthy and patient populations. Whether these strategies or direct activation/stimulation of soluble guanylate cyclase can improve the efficacy and success of exercise training or patient rehabilitation programmes is an exciting prospect and the focus of intense ongoing scientific scrutiny.

Biography

David C. Poole, Brad J. Behnke and Tim I. Musch are professors in the Departments of Kinesiology and Anatomy and Physiology (D.C.P., T.I.M.) at Kansas State University. Their work utilizes a range of novel approaches to unveiling mechanisms of control within the oxygen transport pathway, especially as these relate to skeletal muscle and exercise function, in health, heart failure, diabetes and cancer.