Capillary-Mitochondrial Oxygen Transport in Muscle: Paradigm Shifts

Abstract When exercising humans increase their oxygen uptake (V̇O2) 20-fold above rest the numbers are staggering: Each minute the O2 transport system - lungs, cardiovascular, active muscles – transports and utilizes 161 sextillion (10 21) O2 molecules. Leg extension exercise increases the quadriceps muscles’ blood flow 100-times; transporting 17 sextillion O2 molecules per kilogram per minute from microcirculation (capillaries) to mitochondria powering their cellular energetics. Within these muscles, the capillary network constitutes a prodigious blood-tissue interface essential to exchange O2 and carbon dioxide requisite for muscle function. In disease, microcirculatory dysfunction underlies the pathophysiology of heart failure, diabetes, hypertension, pulmonary disease, sepsis, stroke and senile dementia. Effective therapeutic countermeasure design demands knowledge of microvascular/capillary function in health to recognize and combat pathological dysfunction. Dated concepts of skeletal muscle capillary (from the Latin capillus meaning ‘hair’) function prevail despite rigorous data-supported contemporary models; hindering progress in the field for future and current students, researchers and clinicians. Following closely the 100th anniversary of August Krogh’s 1920 Nobel Prize for capillary function this Evidence Review presents an anatomical and physiological development of this dynamic field: Constructing a scientifically defensible platform for our current understanding of microcirculatory physiological function in supporting blood-mitochondrial O2 transport. New developments include: 1. Putative roles of red blood cell aquaporin and rhesus channels in determining tissue O2 diffusion. 2. Recent discoveries regarding intramyocyte O2 transport. 3. Developing a comprehensive capillary functional model for muscle O2 delivery-to-V̇O2 matching. 4. Use of kinetics analysis to discriminate control mechanisms from collateral or pathological phenomena.


Introduction: What Must Capillaries Do?
Breaking World athletics records and achieving monumental physical milestones (4 min mile, 2 h marathon) captivates public attention whilst extending the proven boundaries of physical performance. Development of novel methods (e.g., near infrared spectroscopy, NIRS; nuclear magnetic resonance, NMR) combined with high fidelity measurements (spatial and temporal) mathematical and kinetics modeling approaches have better resolved the physiology of elite performance in health and also our understanding of the exercise intolerance that erodes life quality and predicates morbidity and mortality in disease.
With these developments, iconic perspectives of skeletal muscle structure and functional control have been overturned, especially as regards O 2 delivery (QO 2 ) and its instantaneous matching toVO 2 . We are far closer to understanding what the key functional parameters of sustained muscular performance are-VO 2 kinetics, critical power (CP or speed, CS), and the max-imumVO 2 (VO 2 max) ( Figure 1). A compelling weight of evidence  supports that the speed ofVO 2 kinetics in healthy young individuals is controlled by mitochondrial energetics rather thaṅ QO 2 rev. 1 and, in part, by constraining the O 2 deficit fastVO 2 kinetics facilitate a greater CP. 2 CP, which is sensitive to mus-cleQO 2 3-6 designates the boundary between the heavy and severe exercise intensity domains and, therefore, sustainable from nonsustainable exercise. 7,8 That experimentally elevating muscleQO 2 , by raising arterial hematocrit rev. 9 or inspired O 2 fraction, 10 increasesVO 2 max demonstrates thatQO 2 limitṡ VO 2 max. Crucially, any model of muscle microcirculatory control and function must cohere with and support these known facets of oxidative performance. Specifically, resting muscle preparations cannot have aVO 2 that isQO 2 dependent (see [10][11][12] ; c.f. 13 ), during exerciseVO 2 and blood flow (Q) should increase in a ratio of ∼1:5 or 6, [14][15][16] across a range of metabolic rates venous O 2 content should fall (and fractional extraction i.e., arterialvenous O 2 content, rise) as a hyperbolic function of increasinġ VO 2 . 14, 16 Moreover, when measured within the skeletal muscle, VO 2 should increase with little delay in an exponential fashion with a time constant (τ ) of some 10 to 40 s in young, healthy individuals. 17-20; rev. 1 The most pervasive capillary control models over the past century are based upon August Krogh's Nobel prizewinning "capillary motor" schema. [21][22][23][24][25] Therein the vast majority of capillaries are, via a contractile endothelium (now debunked) or pericyte constriction (see Role of Pericytes in Controlling Skeletal Muscle Capillary Hemodynamics below), closed in resting muscle but open up and support red blood cell (RBC) flux during muscle contractions. In this review we will demonstrate, across a range of physiological stimuli, that this schema cannot explain the observed behavior, either temporally (i.e., speed oḟ VO 2 kinetics) or quantitatively (i.e., CP,VO 2 max), and is unnecessary to support the dynamic control of microcirculatoryQO 2 andVO 2 kinetics.
Moreover, beyond the Kroghian binary concept of capillaries (i.e., open with a fixed O 2 delivery or closed), our understanding of the transport of O 2 from RBC to mitochondria has undergone substantive revision based upon recent discoveries. For instance, the diffusivity of O 2 across its "final frontier" is likely dependent, in part, upon: (1) Newly discovered channels in the RBC membrane. (2) Recruitment of capillary exchange surface along the capillary length (longitudinal recruitment). (3) Network interdependence of O 2 exchange among adjacent microvessels. (4) More homogeneous O 2 pressures in the interstitial space surrounding myocytes. (5) O 2 tracking along sarcolemmal and intracellular membranes, including interconnected mitochondria. (6) Interplay between deoxygenated myoglobin (Mb) and nitric oxide (NO) release to control cytochrome oxidase activity and better spatially coordinate mitochondrial ATP production and O 2 availability. These considerations are important for construction of a contemporary model of capillary function and O 2 transport: One that has far less reliance on control at the individual capillary level and, at least in health, minimizes the importance of capillary-mitochondrial diffusion distances per se.

A Brief History of Capillaries
When William Harvey's De Motu Cordis was published in 1628 26 there was no concept of capillaries. Rather, blood was supposed to flow from the right to left heart through the pores of Galen in the interventricular septum. Some years later, after boiling organs, including kidneys, liver, lungs, and spleen and examining them closely, Harvey described "capillamenta" which must have been small arteries and veins or arterioles/venules as capillaries cannot be resolved by the naked eye. 27 He thus concluded that, in the periphery, blood flows: ". . . from the arteries into the veins either directly by means of anastomosis or indirectly through the porosities of the flesh. . . " 27, p. 88 Had Harvey been cognizant, at that time, of the eminent Islamic scholar known as Ibn al-Nafis (1213 to 1288, full name: Ala al-Din Abu al-Hassan Ali Ibn Abi-Hazm al-Qarshi al-Dimashq), whose Commentary on Anatomy in Avicenna's Canon became known to the Western world only in the early 20th century, he may have reconsidered his "porosities of the flesh" concept. As translated from "Commentaries. . . " by Meyerhof 28 " ". . . there is no passage between these two cavities [right and left ventricles]; for the substance of the heart is solid. . . " and ". . . .the penetration of the blood into the left ventricle is from the lung. . . " leading Ibn al-Nafis to conclude that there must be small communications between the pulmonary artery and vein. See West, 29 for a reproduction of the original Arabic text opposing the presence of septal connections between the ventricles from Meyerhof. 28 Thus, at least as early as the 13th century in the Arab world the pores of Galen in the interventricular septum had been falsified. In Europe, Michael Servetus (1511 to 1553), who, West 29 speculates, may have been aware of Ibn al-Nafis's work-though historians consider this unlikely-wrote that the blood passed from the pulmonary artery through the lung to the left ventricle and became "reddish-yellow" in the process. 30; see also 29 As a mark of Harvey's intuitive brilliance, a few years before his death in 1657, he experimented with the pulmonary circulation of a throttled man. After ligating certain vessels he demonstrated that water flowed easily from the pulmonary artery through the lungs and into the left ventricle. 31; rev. 29 A finding that, for the emerging generations of microscopists, identified the lungs as an organ where visualization of the smallest vessels connecting the arterial and venous circulation might be possible in vivo.
Enter Marcello Malpighi who was born serendipitously in 1628, the year Harvey published De Motu, in Crevalcore near Bologna. Malpighi used both single magnifying lenses and a compound double-convex lens microscope and observed capillaries in the alveolar walls of the frog lung. 32 He described the capillaries as "tortuous" and clarified conclusively that blood "is always passed through tubules" remaining within the vessels and not "poured out into spaces" (translation from 33 ). Whereas, Jan Swammerdam (1637 to 1680), had described RBCs or corpuscles three years earlier; 34, see also 35 Malpighi mentioned these only as "particles" within the blood, probably because he could not determine their color. 36 Stephen Hales (1677 to 1761), minister of Teddington parish near London, may have been the first person to coin the term "capillary." 37,38 Hales estimated the pulmonary capillary diameter to be around 17 μm (i.e., ∼2to 3-fold the average diameter of muscle capillaries) with pulsatile flow and a RBC transit time just over 2 s. Capillaries were distinguished from other blood vessels by Marshall Hall (1790 to 1857) as being of ". . . an intermediate station. . . " interposed between the ". . . last branches of the arterial system and the first roots of the venous. . . " 39, p.18 He noted that, within capillaries compared with upstream vessels, blood velocity was halved and intuited that ". . . a more diffused and slower circulation is required for administering to the nutrient. . . functions" and considered capillaries to be of "uniform character and dimensions." 39, pp. 29-30 Johannes Müller (1801 to 1858), in 1843 reinforced the notion that capillaries were "the same diameter throughout" 40 but the field was obscured with confusing terms such as "capillary arteries," "capillary veins", and "minute vessels." 41 Improvements in light microscopy toward the middle of the 19th century revealed the cell as the primary structural base of tissues. 42,43; rev. 44 This realization and Schwann's "cell theory" were focal to the recognition of the capillary wall or endothelium.
Reminiscent of Harvey, J.W. Earle in 1835 portrayed capillaries as membrane-less channels such that: ". . . blood in the finest capillaries. . . flows. . . in simple furrows, or canals, whose walls are formed by the surrounding cellular tissue." 45, p. 8 At that time it was considered impossible for "the processes of nutrition and absorption being carried on through the coats of vessels." 42 Counter to this notion was that capillaries in the ears of reptiles and birds when injected with dyes were distinct from adjacent tissues-evidencing "parieties" or walls. 46 It was left to Theodor Schwann (1810 to 1882), in 1839 to identify what would later be called the endothelium in the tails of tadpoles. 43 However, adherents to each side of the argument, walls versus no walls, engaged in strident debate based upon assumed properties of such a barrier and the presumption that it would prevent the known blood-tissue transit of, for example, leukocytes (see ref. 44 for a stimulating account of the controversy). This despite Augustus Waller's (1814 to 1870) demonstration, in real time, of leukocytes "protruding" from vessels at the site of inflammation in his frog's tongue intravital microscopy preparation. 47 A process called diapedesis.
At the close of the 19th century Julius Cohnheim (1839 to 1884), a student of Rudolf Virchow (1821 to 1902), had demonstrated that the capillary wall was a living organ rather than simply an inert membrane. 48 This set the stage for addressing how capillaries subserve their cardinal responsibility for facilitating blood-tissue exchange. Ernest Henry Starling (1866 to 1927) would take up the challenge for substrate and fluid exchange, whereas Schack August Steenberg Krogh's (1874 to 1949) focus was O 2 delivery, primarily in skeletal muscle.
Since Krogh's 1920 Nobel prize, the dominant concepts of capillary function, especially within skeletal muscle, have been driven by Krogh's work and theories. [21][22][23][24][25] This despite a more recent compelling weight of opposing evidence garnered with advanced technologies under rigorously controlled cardiovascular conditions. Major Kroghian concepts include: (A) Most capillaries being "closed" in resting muscle and opened or "recruited" at the onset of contractions. (B) Each capillary representing an independent fixed unit of O 2 delivery with intracapillary diffusion distances determining the efficacy of mitochondrial O 2 delivery. Despite these notions being resoundingly falsified, they remain at the forefront of undergraduate and medical school teaching and exert a persistent influence over the microcirculatory control field today.
In many ways, the microcirculation today is treated as a classic-"everybody talks about it but nobody reads it" to paraphrase the great Mark Twain. 49 For instance, one of the most memorable and enduring "facts" regarding the human circulation is that the total length of microvessels in skeletal muscle (or the human body) is 100 000 km or more; sufficient to encircle the Earth at the equator about 2.5 times. This figure comes initially from Krogh's book, 25, p. 10 where capillarity values from guinea pigs are inappropriately applied to human muscles, and is grossly in error. Our best estimates from contemporary data yield a total length of some 9 000 to 19 000 km; less memorable, perhaps, but more accurate and supporting a very different regulation of capillary structure and function in health and disease. Specifically, where the blood volume necessary to fully perfuse the muscle capillary bed is only 2% to 3% of total blood volume versus over 30% (Krogh's estimate). In the latter instance, if correct, poorly regulated (muscle) capillary opening and perfusion could compromise cardiovascular regulation and homeostasis. Figure 2 portrays some of the essential elements from Krogh, many of which are still retained in the microcirculation lexicon today.
The following section (Pervasive Myths and The Microcirculation) presents some of the major time-honored beliefs originating from Krogh; 21-25 (see Figure 2) and Chambers and Zweifach, as regards so-called precapillary sphincters 50 and, in opposition, the more recent evidence that informs and shapes a contemporary understanding of capillary function and bloodmitochondrial O 2 transport matching structure and function to physiology (Figures 3 and 4).

Evidence Falsifying Myth 1
When circulation-intact resting muscles are visualized by intravital microscopy under physiological conditions (i.e., normotensive, nonhyperoxic, or pharmacologically vasoconstricted, not over stretched or surgically damaged) the vast majority (i.e., >80%) of capillaries support RBC flux (see Figure 3 #1 and Table 1, 51-106 right side . Moreover, that minority of capillaries not supporting RBC flux are "open" and not collapsed. Thus, following the onset of contractions increased perfusive (RBC flux) and diffusive (DO 2 , hematocrit) O 2 transport principally occurs via elevated RBC flux, velocity and hematocrit in already-flowing capillaries ( Figure 4, #s 1 and 2). As capillary hematocrit at rest is only ∼15% to 20%, on average, increases toward systemic values (i.e., ∼45%) can substantially increase capillary RBC content and therefore DO 2 . 107,108,109 As RBC velocity increases in concert with elevated fractional O 2 extraction-from ∼25% at rest to as much as ∼90% during contractions-capillary exchange surface area is recruited along the length of individual capillaries, a process known as "longitudinal recruitment," which contributes to the elevated DO 2 (Figure 4, #3). 110 An important scientific goal will be to apportion the substantial increase in DO 2 from rest-contractions among microvascular (e.g, RBC aquaporin + rhesus channels, hematocrit, RBC distribution [relative to fiber(s)VO 2 ], longitudinal recruitment, network effects), and myocyte-associated processes (interstitial O 2 diffusion pathways, membrane O 2 transport [sarcolemma, mitochondrial membranes], Mb, simple diffusion). These latter processes are considered below (see the section "Intramyocyte O 2 Transport: Sarcolemma-Mitochondria").
Using a combination of intravital microscopy and phosphorescence quenching in the rat spinotrapezius muscle, rapid, and physiologicalVO 2 kinetics (time constant ∼30 s) have been demonstrated in the absence of any increase in the numbers of RBC-perfused capillaries. 20,89 This profile matches closely that measured for pulmonary 17 A valid concern reflects that, because direct observation of the skeletal muscle capillary bed requires that the animal be anesthetized, the presence of RBC flux in most capillaries may be an artifact of anesthesia or surgery. However, several lines of evidence support that neither anesthesia nor surgery perturb the underlying physiology in this regard. Specifically: (A) Surgical exteriorization does not alter muscleQ 113 nor impact the fundamental proportionality between musclė VO 2 andQ. 16 (B) Using endothelial dye perfusions essentially all capillaries in muscles of anesthetized and conscious rats at rest evidence flow (at least plasma) nullifying the opportunity for de novo recruitment. 79,80 (C) Broadly accepted values for resting muscle blood flows are compatible with 80% to 90% of total capillaries sustaining flow with >10 RBCs/s. 89,114-119 (D) Total quadriceps (haemoglobin + myoglobin), measured using near-infrared spectroscopy (NIRS), increases <30% from restexercise. 120-122, rev. 119, 123-125 This observation is inconsistent with the substantive recruitment of previously "closed" capillaries.
One theoretical argument advanced to support that capillary recruitment must occur from rest-exercise is that this is necessary to limit the fall in capillary RBC transit times. However, this notion is specious because, irrespective of how you get there-recruitment or simply increased flux in already flowing capillaries-RBC transit time during maximal exercise will be exactly the same as it is determined by the ratio of capillary volume/blood flow. Moreover, the estimations of Richardson and colleagues, 126,127 based upon some of the highest muscle blood flows measured in human kneeextensors, support that capillary RBC mean transit times of only ∼0.1 s can facilitate fractional O 2 extractions as high as 80%.

Evidence Falsifying Myth 2
At rest and during contractions there is a pronounced heterogeneity of RBC flux, velocity, and hematocrit among capillaries. This is true even for adjacent capillaries served by the same terminal arteriole. If there is little or no de novo recruitment of flowing capillaries from rest to contractions then the perfusive (QO 2 ) and diffusive (DO 2 ) capacity of individual capillaries (see 89 and Table 1, right side for references) must increase many fold to support as much as 100-fold elevation of muscleVO 2 . 126 Conceptually it is important to recognize that some capillaries at rest have such a low RBC flux and/or hematocrit that they do not contribute substantially to support metabolism. During contractions, however, their increased RBC flux and hematocrit facilitate a substantial contribution of these capillaries to muscleVO 2 (Figure 4 #1).

Myth 3:
Capillaries are straight, unbranched vessels that supply O 2 independent of all others to a cylinder of surrounding tissue.

Evidence Falsifying Myth 3
Capillaries evidence substantial tortuosity, especially at short muscle sarcomere lengths, 128,129 and are highly interconnected by branches (Figure 4 #5). 128 This geometry serves to increase muscle capillary volume and surface area and the proportion of myocyte surface immediately adjacent to RBCs within capillaries. Thus, the Hill cylindrical model of O 2 diffusion is likely more appropriate than the "pinpoint" O 2 source that Krogh considered (cf. Specifically, PmvO 2 is regulated as a function of muscle fiber type, being significantly higher at rest and during contractions in slow than fast twitch muscles. 138,139 PO 2 falls steeply across the capillary endothelium being some 8 to 12 mmHg lower for PisO 2 than PmvO 2 . 133 It has not been ruled out that O 2 consumption by the frequency domain phosphorescence method might contribute to some of this gradient. However, the relative constancy of the gradient from rest to contractions, and the very low potential for this method to consume appreciable O 2 , at least in comparison to the extantVO 2 of the tissue, support that this microvascular-interstitial gradient is not an artefact of the technique. Thus, because PimO 2 during contractions decreases to as low as 1 to 5 mmHg (data from dog gracilis 134,135 ) and human quadriceps [140][141][142][143] ) there exists an appreciable (i.e., several mmHg) trans-sarcolemmal (i.e., PisO 2 to PimO 2 ) PO 2 gradient during contractions. Although possibly a limitation of the cryomicrospectroscopy and proton NMRS measurement techniques neither radial nor longitudinal intracellular PimO 2 gradients or anoxic loci (i.e., 0 mmHg) have been detected. If intramyocyte PO 2 gradients do exist, because of the overall low PimO 2 , they must be small. These data support the notion that intramyocyte O 2 transport is extremely effective and therefore that diffusion distances from sarcolemma to even the most distant part of the mitochondrial reticulum are not limiting (see also Skeletal muscle Mitochondria: Structure-Function and Relevance. . . .below).

Evidence Falsifying Myth 5
Krogh's "capillary motor regulating mechanism" 21-23, 144 that won him the Nobel prize was central to the capillary recruitment hypothesis and emplaced capillaries themselves in control of muscleQ (Figure 2 #5). Krogh initially considered that the capillary endothelium was itself contractile-as conveyed in his Nobel lecture drawing of a "capillary" that becomes vastly dilated (>50 μm diameter) when mechanically irritated. 24,25; rev. 125 However, subsequent experiments across tissues from different species led his focus to Rouget cells 145, 146 -now called pericytes-as the source of unitary capillary constriction and  1) Most capillaries support RBC flux at rest and so there is not much scope for de novo capillary recruitment. (2) Evidence of either contractile endothelium or pericytes closing capillary lumen and opening during contractions with requisite speed lacking. (3) Hill model incorporating capillary geometry and, possibly, interstitial space, allows cooperative O2 delivery among capillaries (and other vessels) and reduces supplied tissue volume with distance from capillary. (4) Most of the partial pressure (PO2) drop between red blood cell in the microvasculature (mv), across the interstitial space (is) and mitochondria occurs very close to the capillary such that intramyocyte (im) PO2 during contractions is extremely low. (5) It is now recognized that the majority of vascular resistance and thus blood flow control resides at the arteriolar level. See text (Myths #1 to 8) for more details.
dilation (see Schmidt-Nielsen, 1984 for her fascinating account of these experiments). 147 Despite intense investigation, initially by Bjovulf J. Vimtrup 148,149 over the ensuing decades, there was not incontrovertible evidence that pericytes could reversibly constrict (and close shut, Figure 2 #2) the capillary lumen (see 125 for a detailed account). Throughout the 1930s capillary contractility was repeatedly challenged with advocates 150, 151 and opponents [152][153][154][155][156] who regarded the capillary wall and associated structures as noncontractile.
In the absence of satisfactory evidence that either the capillary endothelium or pericytes could occlude the capillary lumen Benjamin Zweifach's attention shifted to the beginning of the capillary, where the capillary branch departs from the arteriole. At that specific location he detected a narrowing which he termed a precapillary sphincter. 50 Here the plot thickens. In the frog mesentery in 1937 Zweifach finds no precapillary sphincters, 155 but in 1939 he found such in both mouse and rabbit mesentery. 156 In 1942 Chambers and Zweifach produced motion pictures evincing "sphincter like functioning of the precapillaries at their junction with the arteriole." 157 ; see also Sakai and Hosoyamada. 158 Despite Zweifach's observations of precapillary sphincters solely in the mesentery and not in all species, within a decade, they were adopted as the presumptive site of capillary hemodynamic control within ALL microcirculations in physiological and medical textbooks. 159-165; rev. 158 Also disturbing, and a clear case of confirmation bias, was the propensity for researchers, from that point on, to attribute changes in vascular resistance to "precapillary sphincter tone" without any evidence for such. [166][167][168] The extensive search for precapillary sphincters, anatomically and physiologically, detailed by Sakai and Hosoyamada, 158 did not identify these structures in any tissue.
Given that absence of evidence is not necessarily evidence of absence with respect to structure it is also telling that presence of precapillary sphincters has not been demonstrated physiologically. For instance, hyperoxia, K ATP channel blockade with glibenclamide and other conditions that decrease blood flow to skeletal muscles do so via arteriolar vasoconstriction and there is no evidence that RBC flux can be halted by active luminal constriction specifically in individual capillaries. 52, 169; rev. 101,170 Myth 6: Following the onset of contractions it is necessary for muscles to deplete their O 2 stores and produce vasoactive metabolites which then vasodilate arterioles and "open" up previously closed capillaries.

Evidence Falsifying Myth 6
Advances in intravital microscopy, phosphorescence quenching measurement of PO 2 , ultrasound, NMR, and mathematical modeling have permitted resolution of rapid muscle metabolic and hemodynamic transient responses following exercise onset. These techniques have helped unveil the mechanistic bases for vascular control dynamics.
Logically, if the cardiovascular system is to subserve its primary function of blood pressure control, there must be a tight coordination between central cardiovascular (cardiac output) and muscleQ dynamics. Skeletal muscle has such a great capacity for vasodilation that it has been described as the "sleeping giant" in that uncontrolled dilation would crash blood pressure (1) At exercise onset, capillaries are not recruited de novo but red blood cell (RBC) velocity and flux increase and raise the proportion of capillaries important for O2 delivery. If mean muscle blood flow increases by 3-fold for moderate/light intensity exercise and 30fold for heavy/severe intensity exercise, assuming a normal distribution of capillary RBC fluxes at rest and during contractions, the percentage of capillaries having an RBC flux below 10 cells/s (resting data 114 ; contracting 89 ) would decrease from ∼20%-30% to ∼10% and to ∼1-2%, respectively. (2) Elevated capillary hematocrit (∼30%; rat intravital microscopy 89 ; human NIRS 243 ) increases O2 diffusing capacity commensurately. (3) Increased functional surface area is recruited along the length of capillaries (so-called longitudinal recruitment). (4) There may be a modification of the endothelial surface layer (glycocalyx) that facilitates the higher hematocrit. This represents an exciting current field of investigation. (5) The 3D geometry of the microvascular bed facilitates between-vessel O2 exchange by means of the interstitial space. (6) The Hill model better represents capillary-myocyte O2 transport than the Krogh model (pinpoint O2 delivery). (7) Despite that microvascular PO2 falls during exercise (reflecting the greater fractional O2 extraction) the precipitous drop in intramyocyte PO2 means that the overall PO2 gradient between RBC and intramyocyte space actually increases from rest to exercise. (8) Upstream signaling from capillaries (and also potentially venules) vasodilates the terminal arterioles specific to individual capillary modules. This mechanism and potentially microvascular network effects are likely key to facilitating O2 delivery-O2 utilization matching across the exercising muscle. The propensity of papers demonstrating RBC and/or plasma flow in most capillaries in resting muscle (right side) are either direct observations of muscle microcirculation or studies using endothelial staining of capillaries. Unlike those on the right side, many of those on the left side use indirect methods such as 1-MX (1-methyl xanthine) or contrast-enhanced ultrasound (CEU). Updated from Poole el al. 107 See text for more details.
catastrophically. 171 Specifically, if muscle vasodilation occurred prior to cardiac output increasing, blood pressure would plummet. By the same token delayed muscle vasodilation would produce a potentially dangerous blood pressure spike. In young healthy subjects, neither extreme response is seen. Rather, following the onset of rhythmic cycling exercise mean arterial pressure rises modestly with response kinetics that are far slower than that of muscleQ (i.e., time constant, τ , MAP, 89 s, muscleQ, 9 s): 172 Supporting a close dynamic matching between the time course of both processes. As early as 1895 the eminent Swedish physiologist Erik Johan Johansson (1862 to 1938) demonstrated that electrical stimulation of the legs of rabbits increased heart rate within 0.5 s. 173 Later Krogh and Lindhard in 1913, 174 measured correspondingly fast heart rate kinetics at the onset of cycle exercise in humans and Anrep and von Saalfeld in 1935 showed that muscle hyperemia was initiated simultaneously with the onset of contractions. 175 Despite this evidence, as late as the close of the 20th Century, it was widely considered that, following the onset of exercise, "muscle" PO 2 plummeted toward zero and subsequent vasodilation was dependent upon production of metabolites and sympathetic vasodilator activity. 176,177 Today it is recognized that mechanisms initiating the increase of muscleQ may be very different from those that sustain its exercising steady state. 178; rev. 179 Moreover, the obligatory role of the muscle pump 180,181 in initiating the rapid hyperemia has been challenged by experiments supporting that contraction-induced vascular deformation induces an almost immediate vasodilation 182,183 possibly involving integrins 184 and mechanosensitive ion channels. 185 Importantly, Behnke and Delp 186 showed that isolated rat skeletal muscle arterioles do evince the requisite fast relaxation dynamics in response to elevated flow, acetyl choline and NO. However, Clifford and Tschakovsky, 179 argued that a system which elevates muscle 19,[181][182][183]187,188 and capillary 89Q within 1 s after a contraction is unlikely to depend on diffusion of some soluble vasoactive mediator. This conclusion would also be relevant to the extent to which sympatholysis 189 could participate in the initialQ response. 190 Notwithstanding its precise mechanistic bases, following the onset of muscle contractions, in both exercising human muscle 19,111,112,191 and capillary 20,89,192Q increases as fast or faster than mitochondrialVO 2 such that, for the first 10 to 20 s of contractions, muscle PmvO 2 and venous O 2 content do not fall and may even rise ( Figure 6). 19,20,89,192; rev. 124 Subsequently, as mitochondrialVO 2 continues to increase and despite a continued increase ofQ to the steady-state or peak level, PmvO 2 and venous O 2 content decrease in a close-to-exponential fashion reflecting the rising overallVO 2 :QO 2 ratio. 16 It is quite possible that, beyond the initial transient, further vasodilation and increasedQ reflects regional specific metabolic control that acts to enhanceVO 2 -to-QO 2 matching and promotes greater fractional O 2 extraction (see the section Modeling Capillary-Myocyte O 2 Delivery below).

Myth 7:
Muscle O 2 diffusing capacity (DO 2 ) can only increase from rest to exercise by recruiting more capillaries and is determined principally by capillary density.

Evidence Falsifying Myth 7
As seen above (see Myth #1), in healthy skeletal muscle, most capillaries support RBC flux and are, therefore, not available to be recruited. What can, and does, change from rest to contractions is the RBC flux and hematocrit within individual capillaries. 89 This is crucial because O 2 delivery models 108,109 and experimental data 193 support that the primary determinant of DO 2 is the number of RBCs in the capillary bed, adjacent to the muscle fibers, within RBC-flowing capillaries. Accordingly, diseases such Type II diabetes 92 and HF 90 that decrease the proportion (and absolute number) of capillaries that support RBC flux, at least in animal models, are characterized by low DO 2 . 194; rev. 195 As discussed below in "Oxygen transport from capillary to mitochondria" the dependence of DO 2 on capillary hematocrit may depend upon recently discovered aquaporin and rhesus channels in the RBC membrane.
The importance of knowing capillary RBC hemodynamics and not just capillary density per se was effectively demonstrated by Hepple and colleagues in the canine gastrocnemiusplantaris complex contracting atVO 2 max. 196 Specifically, using exercise training and limb immobilization to alter capillarity, DO 2 was completely dissociated from capillary density.

Evidence Falsifying Myth 8
Viewed most commonly histologically in transverse muscle slices, intramyocyte mitochondria appear to be independent "bean-shaped" organelles reminiscent of their bacterial origins. In marked contrast, when viewed 3D intramyocyte mitochondria, in all principal fibre types, constitute an interconnected reticulum reaching from the subsarcolemmal space to the most capillary-distant intermyofibrillar mitochondria (see Figure 7). [197][198][199][200][201][202][203] Moreover, acutely increased mitochondrial interconnectivity 204 may potentially help explain the dramatically elevated intramyocyte O 2 transport and DO 2 found during exercise. That greater [cytochrome C oxidase complex IV] is found in subsarcolemmal versus intermyofibrillar mitochondria but [ATP synthase complex V] is not 202; rev. 205 might enable intramyocyte H + -electrochemical (and O 2 ) gradient tracking. Much like a power grid, charge created at one location can power ATP production at a spatially remote site. Such a system would defend local ATP production in the face of very low PO 2 and small PO 2 gradients. 205 In combination with oxy/deoxy-Mb removing/producing NO to regulate cytochrome C oxidase activity such an interconnected power grid would optimize ATP production in the face of a very low intramyocyte PO 2 (Figure 8).
This schema would: (1) Abolish dependence on long intramyocyte capillary-mitochondria O 2 diffusion pathways and creation of Krogh's anoxic "lethal corners," and (2) Negate the necessity for precise capillary positioning around the fibers.

Role of Pericytes in Controlling Skeletal Muscle Capillary Hemodynamics
Reminiscent of Krogh's search for a mechanism by which individual capillaries might control their own hemodynamics, it has been claimed that pericytes exert a "canonical" role in regulating skeletal muscle vessel diameter. 206 However, some of the strongest evidence that pericytes can constrict/dilate capillaries comes from other tissues, most notably brain 207-210 and heart muscle. 211,212; rev. 213 In particular, beautiful electrophysiological studies by Zhao and colleagues 211 in mouse papillary muscle demonstrate that pericytes are electrically connected via gap junctions to ventricular myocytes. Thus, they propose that, when intramyocyte [ATP] ([ATP]im) falls, K ATP channel opening hyperpolarizes myocytes, capillary endothelial cells, pericytes, and vascular smooth muscle cells; decreasing pericyte and vascular smooth muscle [Ca 2+ ] and relaxing any extant constriction. This elegant potential mechanism would be exquisitely sensitive toQO 2 -VO 2 mismatch ele-vatingQO 2 rapidly and specifically to ventricular myocytes with the greatest needs. That said, under physiological conditions, neither cardiac nor skeletal myocytes undergo any appreciable fall in [ATP]im even across substantial-orders of magnitude-increases of metabolic demand. 214,215 Indeed, as demonstrated by Cain and Davies, 214 it is necessary to completely inhibit creatine phosphoryltransferase with FDNB (1fluoro-2,4-dinitrobenzene), a markedly nonphysiologic condition, to decrease [ATP] substantively during muscle contractions. Such a system might be better triggered, at least in skeletal muscle, by sensitivity to the phosphorylation potential (i.e., [ATP]/[ADP + Pi]). 216 Moreover, even under highly nonphysiological conditions (e.g., capillary PO 2 140 to 700 + mmHg) the extent of the pericyteassociated luminal constriction was extremely small (i.e., ∼20%, 1.2 μm; 210 ∼13%, 211,212 ) and thus, whereas it might increase individual capillary resistance, it would not be expected to prevent RBC passage. In response to whisker stimulation in mice, somatosensory cortex capillaries dilated ∼5% to 7% of their 4.4 μm baseline within ∼4 to 10s 208 -faster than some arterioles and considered, by the authors, to be responsible for 84% of the regional blood flow increase. These kinetics are far faster than measured for bulk brain blood flow 217,218 where the mean response time is ∼80 s suggesting that pericyte action may need to redistribute the existing blood flow in response to elevated regional demands. Interestingly, in brain, ischemia induces pericyte constriction of capillaries which may be irreversible. 208 Capillary perfusion heterogeneity may not be controlled in muscle 219 as it potentially is in brain. 125,220,221 It is also pertinent that, in heart, pericyte coverage is less than in brain with not all capillaries having a pericyte. 211,212 Specifically, endothelial cells are ∼10 × 30 μm in size 222 , whereas brain and heart have an endothelial cell-to-pericyte ratio between 1:1 and 1:3 223,224 Figure 5. The Hill solid cylinder model provides a far more efficient O2 delivery to the myocyte (black circle) as the O2 path is not restricted to the small physical space apposed to the red blood cell (RBC). Thus, for the Hill model the O2 flux density-and therefore the apparent resistance to O2 diffusion into the myocyte-is far less. 130 Figure 6. Temporal changes in capillary red blood cell (RBC) flux, oxygen uptake (VO2), and microvascular O2 partial pressure (PmvO2) following the onset of contractions in the rat spinotrapezius muscle (redrawn from 20 , 89 , 192 ). See text for more details.

Figure 7.
Over the past half-century it has become appreciated that mitochondria in skeletal muscles of humans and animals are not isolated bean-shaped organelles as depicted at left but, rather, form a catenated, interconnected reticulum that may transduce electrical charge (and potentially O2) across and along the myocyte (see right). This realization challenges traditional notions of O2 diffusion distances.

Figure 8. Myoglobin-mediated Regulation of O2 Diffusing Capacity (DO2) and Oxidative
Metabolism. (Top) Mb molecules are saturated with O2 in resting muscle forming a functionally O2-carrier depleted region (FCDR) with low DO2. With exercise, Mb becomes deoxygenated as O2 utilization increases and ATP production accelerates; reducing the FCDR and elevating DO2. (Bottom) Mb molecules function as O2 sensors, spatially limiting, or facilitating cytochrome c oxidase activity via nitric oxide [NO] removal or production. Thus, locales with higher O2 partial pressure [PO2], will have NO scavenged by oxymyoglobin permitting uninhibited cytochrome c oxidase function whereas those with very low PO2 and deoxymyoglobin will reduce nitrite [NO2-] to NO, which will inhibit cyctochrome c oxidase activity and maintain PO2. This mechanism may be important for preventing formation of anoxic loci during maximal exercise (based upon 205 ). in skeletal muscle it is 100:1 225; rev. 226 decreasing the opportunity for many capillaries to even have a pericyte. Also, whereas some authors claim that pericytes play a deterministic role in muscle perfusion, the strongest evidence for such, at present, is restricted to pathological states such as arterial stenosis. 227 It is also pertinent that, of the three tissues considered above, skeletal muscle has by far the greatest range of metabolic rates demanding over a 100-fold increase in blood flow from rest to maximal exercise. 126 Thus, the potential for, but perhaps not the consequences of,QO 2 -VO 2 mismatch is far greater in skeletal muscle versus heart or brain.
In conclusion, for pericyte relaxation to account for the rapid (∼1 s or less) skeletal muscle capillary hyperemia following the onset of contractions (see Myth #6 above) it would be necessary to demonstrate several phenomena; including: (1) That, in resting muscle, pericytes are constricting capillaries and occluding RBC flux. As evidence falsifying Myth #1 above (see also Table 1, right column) substantiates, this does not appear to be the case. (2) That pericytes can relax almost simultaneously across most of the capillary bed in synchrony with the first contraction-relaxation cycle following the onset of exercise.
Perhaps, there might be a role for pericytes improvinġ QO 2 -VO 2 matching in the secondary phase of the muscle capillary hyperemic response wherein PmvO 2 falls, and a-vO 2 difference rises, toward steady-state values. 89 To date, however, direct evidence is lacking. Vasoaction, at least in response to exercise (onset and offset, 89,178 ), chemical stimuli, 228,229 and the K ATP -channel blocker glibenclamide, 101 as well as the pathological derangements (decreased proportion of capillaries supporting RBC flux, impaired capillary hemodynamics, in HF, 90 Type II diabetes, 92 ) appear to be mediated exclusively at the arteriolar level. Specifically, capillary hemodynamics are accelerated in hyperemic states and decelerated in low flow conditions without any discernible narrowing of the capillary lumens whether caused by pericytes or other means. 229 It is also pertinent that, in resting or contracting muscle at sarcomere lengths below ∼2.7 μm (i.e., physiological, nonstretched), capillary RBC velocity and flux through individual capillaries exhibit enormous heterogeneity and do not correlate with capillary luminal diameter. [85][86][87][88][89]230 Oxygen Flux At the Blood-Myocyte Boundary: O 2 Diffusing Capacity (DO 2 ) For many years it was considered that acute and chronic (e.g., exercise training, aging, disease) changes inVO 2 max were driven almost exclusively by changes in muscle(s) perfusive O 2 flux (i.e.,Q × arterial [O 2 ] orQO 2 ) with little attention paid to muscle O 2 diffusing capacity (DO 2 ). However, some 3 decades ago Peter D. Wagner graphically (Figure 9) conflated the Fick principleV with Fick's Law : which demonstrated effectively that the reason skeletal muscle could not remove all the O 2 from the venous effluent, atVO 2 max when the mitochondrial were O 2 -supply limited, was the presence of a finite DO 2 . 231,232; rev. 233 As demonstrated in Figure 9, right panel, measured venous (or calculated mean capillary) PO 2 or O 2 content may be little impacted by exercise training or HF, for example, but conceal substantial changes in DO 2 . For instance, despite only a small increase in a-vO 2 difference, in response to exercise training ∼70% of the increaseḋ VO 2 max is attributed to elevated DO 2 rather thanQO 2. 232 For HF patients a reduced DO 2 contributes substantially to their loweredVO 2 max even though muscle venous effluent PO 2 may be very low. 234; rev. 195 Thus, fractional O 2 extraction is synonymous with a-vO 2 difference and determined by the ratio of where β is the slope of the O 2 dissociation curve in the relevant physiological range. That the number of RBCs immediately adjacent to the muscle fibers within flowing capillaries is a primary determinant of DO 2 108,109,193 coheres with the capillary neogenesis stimulated by exercise training increasing, rev. 235,236 and the capillary involution and high proportion of capillaries that do not support RBC flux in HF decreasing, 90,194 DO 2 (Figure 9, right panel).
As O 2 moves down its pressure gradient from capillary RBCs to mitochondria in skeletal muscle to power oxidative energetics it must (in sequence): Dissociate from hemoglobin, traverse the RBC membrane, plasma, capillary endothelial surface layer (glycocalyx), endothelial cell, interstitial space, myocyte sarcolemma and intervening cytoplasm, cross the mitochondrial membranes, and react with reduced cytochrome c oxidase and protons ( Figure 10). rev. 107 Whereas each of these steps will provide some finite resistance, measurements of PmvO 2 and interstitial PO 2 (PisO 2 ) by phosphorescence quenching and intramyocyte PO 2 (PimO 2 ) by proton NMR and cryomicrospectrophotometry support that the bulk of the apparent resistance to O 2 flux lies in close proximity to the capillary. Thus the majority of the PO 2 decrease between RBC and mitochondria occurs from capillary to interstitial space and from there into the myocyte (i.e., ∼1 μm). [131][132][133] During contractions the PimO 2 is low (i.e., 1 to 5 mmHg in human vastus lateralis) 140 and absent of marked transverse or longitudinal gradients (canine gracilis). 135 Let's consider individual steps in O 2 's pathway from RBC to mitochondria.

Red Blood Cell
Hemoglobin O 2 dissociation has a 4 ms half-time with a velocity constant ∼3 000 mM/s under physiological conditions. [237][238][239] AtVO 2 max, PO 2 decreases <4 mmHg from Hb to the RBC membrane. 240 Exciting recent data suggest that two channels (aquaporin-1, AQP-1, rhesus, RhAg) facilitate transmembrane O 2 flux 241 and may account for over half of the RBC O 2 diffusing capacity. 241,242 It remains to be determined how knocking out these channels impairs skeletal muscle DO 2 and exercise capacity in vivo.

Plasma, Glycolcalyx, Endothelial Cell, And Interstitial Space
Oxygen's insolubility in plasma dictates that only RBCs themselves are quantitatively important O 2 sources. Thus the proportion of the capillary endothelium relevant for O 2 flux is determined by how many capillaries sustain RBC flux as well as their length and crucially their hematocrit. 108,109 Acting like an O 2 bottleneck the RBC-endothelial region concentrates the O 2 flux density forcing a large fall in PO 2 across the microvascular-to-interstitial space. [131][132][133] This bottleneck is widened during exercise as capillary hematocrit increases from 10% to 20% at rest toward systemic values (∼45%). rev. 119 It ) and/or diseases such as heart failure or diabetes (red arrow, i.e., blood flow distribution, ↓capillaries flowing, O2 extraction, ↓DO2, PO2), and aging. 107 , 123 , 250 remains to be determined exactly how high capillary hematocrit rises during heavy or severe intensity exercise. Measurements using Time Resolved Near Infrared Spectroscopy (TRS-NIRS) in human muscles during severe intensity exercise indicate that muscle hematocrit may only increase up to 30% above resting 121,122,[243][244][245] contributing only fractionally to the DO 2 elevation necessary for the 100-fold increase in blood-muscle O 2 flux.
With respect to PmvO 2 and PisO 2 there is a marked fibre type heterogeneity. Specifically,QO 2 at rest and during exercise is maintained far higher, with respect toVO 2 (i.e., ↑QO 2 -to-VO 2 ratio) for slow-versus fast-twitch muscles, such that their PmvO 2 is higher. 133,138,139 Quantifying the PO 2 profile across the microvascular and interstitial compartments at rest and during contractions within contrasting fibre-type muscles helps partition the apparent O 2 flux resistance into myocytes. The 1-2 μm region from the RBC membrane to the myocyte subsarcolemmal space is considered an O 2 carrier-free region (CFR) encompassing the plasma and endothelial surface layers, endothelial cell, interstitial space and sarcolemma ( Figure 10). Interestingly, the substantial trans-capillary PO 2 gradient is similar in both slow-and fast-twitch muscles at rest and during contractions (1 Hz, ∼moderate intensity exercise). Thus, despite the contractions-induced increase of capillary hematocrit, the plasma-glycocalyx-endothelial cell interface presents an appreciable apparent resistance to blood-myocyte O 2 flux amounting to ∼60% (soleus) to 85% (white gastrocnemius) of the capillarymyocyte PO 2 drop: which is consistent with Roy and Popel's 246 theoretical modeling in athletic and nonathletic animal muscles.
In slow-versus fast-twitch muscles the higher PmvO 2 (rest and contracting) and greater PisO 2 promotes a higher transsarcolemmal O 2 flux and may serve to sustain a higher intramyocyte PO 2 (PimO 2 ). This higher PimO 2 will reduce metabolic perturbations (i.e., less increase of [ADP] free , [Pi], [H + ], and glycolysis) and improve muscle contractile performance. 141,[247][248][249] That upstream vascular regulation can dictate muscle metabolic control across fiber types has been heretofore marginalized by attention almost solely to fiber enzymatic properties. Greater focus on this aspect of vascular control may provide novel insights into exercise intolerance in health and especially major diseases such as HF and diabetes. rev. 1, 2 123, 124,195,250

Intramyocyte O 2 Transport: Sarcolemma-Mitochondria
The theoretical model of Krogh and Erlang hypothesized that muscle PO 2 decreased proportionally with distance "R" from the capillary (Figure 2 #4). [21][22][23] The foundations for that model included: Muscles have a finite "diffusivity" that is increased from rest-exercise solely by recruitment of more capillaries each with a unitary perfusive and diffusive capacity supplying a fixed volume of tissue that increases with distance from the capillary (see Myth #4 above). Muscle regions distant from capillaries exist in an anoxic state (PO 2 = 0 mmHg see also 251 ) at rest and become oxygenated during contractions. The role of Mb in intramyocyte O 2 transport was not appreciated.
Decades later development of the PO 2 microelectrode facilitated construction of tissue PO 2 histograms revealing mean PO 2 values between 16 and 39 mmHg in resting skeletal muscles. 252; rev. 253 Later Honig and colleagues' 135 cryomicrospectrophotometry resolved submyocyte myoglobin O 2 saturations, which permitted calculation of local intramyocyte PO 2 's in rapidly frozen resting and contracting dog gracilis muscles. Despite technical shortcomings, that precluded spatial resolution much below 50 μm, 254 their data showing low intramyocyte PimO 2 s during contractions-and therefore absence of appreciable intramyocyte PO 2 gradients-was confirmed by proton-NMRS in human muscles. Specifically, in human muscles PimO 2 was reduced from 20 to 30 mmHg at rest to between 2 and 5 mmHg during heavy-maximal exercise (Figure 3 #4). [140][141][142][143]255 Within human muscles in particular there is an interdigitating mosaic of contrasting fibre types and a dissociation anatomically between capillary modules and muscle fibers within an individual motor unit. Accordingly, the same capillaries may supply both quiescent and active muscle fibres 256; rev. 102 contributing to disparateQO 2 -to-VO 2 and thus variant extraction ratios across fiber types 16 and microregions. Given this behavior it is remarkable that the femoral arterialvenous O 2 difference for human cycling exercise increases (and Figure 10. Key Features of Red Blood Cell-to-Intramyocyte Oxygen Transport. Muscle O2 uptake (VO2) increases rapidly and exponentially following the onset of exercise. Haemoglobin-bound O2 in red blood cells (RBCs) moves down its pressure gradient at the RBC-capillary interface (PmvO2, microvascular partial pressure for O2) crossing plasma, the endothelial surface layer/glycocalyx, endothelial cells, and on into the interstitial space. In addition to simple diffusion, movement of O2 across the RBC plasma membrane occurs via aquaporin (AQP-1) and rhesus (RhAg) channels. 241 , 242 O2 carriers being absent between the RBC and sarcolemma (carrier free region), differences in functional surface area available for O2 flux produce a higher RBC-interstitial O2 flux density than that at the interstitial-intramyocyte barrier: Requiring a substantial pressure differential (i.e., [PmvO2 > PisO2], interstitial space PO2) to transcend the apparent O2 flux resistance between compartments. As seen in Figure 7 subsarcolemmal interconnect with interfibrillar mitochondria 202 , 203 and are surrounded by cytoplasmic Mb. When O2 saturated in muscle at rest, oxymyoglobin forms a functionally depleted O2-carrier region (FCDR) that desaturates during exercise increasing O2 diffusing capacity and O2 transport to the more "distant" mitochondrial reticulum (see text and Figure 8 for further details, rev. 135 ). femoral venous O 2 concentration decreases) hyperbolically 1 according to: whereVO 2 is in l/min. Any model of capillary function and O 2 movement from capillary-mitochondria must be consistent with the physiological observations characterized in equation 4. 16,119,125 It is argued below that a contemporary understanding of O 2 movement across/within myocytes is not consistent with the necessity for de novo capillary recruitment from rest-exercise nor the importance of capillary-mitochondrial diffusion distances with respect to O 2 supply.

Modeling Capillary-Myocyte O 2 Delivery
Abundant evidence demonstrates that, in skeletal muscles, the locus of capillary RBC flux control resides in the arterioles with some modulation from vasodilation signals arising from within individual capillaries. [257][258][259][260][261][262] These signals are conducted via gap junctions rev. 259 and also a pannexin/purinergic-dependent pathway 262 and contribute to the rapid hyperemia (i.e., <1 s). 89,183,259 There may also be fine-tuning of capillary RBC flux across capillary modules resulting from venular vasomotor control. 102 As discussed above, (see the section Role of Pericytes in Controlling Skeletal Muscle Capillary Hemodynamics) evidence for unitary capillary RBC flux modulation by pericytes (Rouget cells) in skeletal muscle-at least in health and with kinetics that can account for the increased capillary RBC flux measuredis wholly absent. Moreover, the presence/participation of precapillary sphincters 21-23,25,50; rev. 158 or a contractile endothelium are not supported by either anatomical or physiological evidence.
In Myth #3, previously, we considered that the capillary 3D geometry, tortuosity, and branching, particularly in passively shortened or contracting muscles, 128 supports the Hill model of O 2 delivery. Critically, for the Hill model the dependent tissue volume decreases with distance from a given capillary and/or the myocyte sarcolemma ( Figure 5, right panel) 130 with all capillaries adjacent to a particular fibre cooperatively supplying O 2 in proportion to their convective and diffusive O 2 conductances and the fibre's metabolic demands. 89; rev. 107,125 It is also notable that capillaries may also target their O 2 delivery preferentially to 1 of 2 or 3 adjacent fibres by selectively embedding into the sarcolemma of that fibre. 263

O 2 Interdependence of Microvessels and Myocytes
Countercurrent capillary flow is present, but it not considered to contribute much to arteriovenous O 2 shunting. 86,264; rev. 253 However, capillaries with short RBC transit times, consequent either to minimal path lengths 265 or high RBC velocities, can potentially decrease fractional O 2 extraction. 253 That said, in human knee extensor muscles, it is possible to achieve ∼80% O 2 extraction at estimated mean capillary RBC transit times of only 0.1-0.2 s atQ's ∼4 l/kg/min. 126,127 This process may be aided by the close proximity of arterioles, tortuous capillaries and venules and their anastomoses facilitating diffusive interactions among adjacent capillaries, capillaries and arterioles as well as venules (Figure 4 #5): 136,137,266 A process where the interstitial space must constitute a conduit for O 2 around the periphery of the myocytes ( Figure 5, right panel). These "network" O 2 diffusional interactions are likely to be crucial to support the rapidVO 2 kinetics following exercise onset where, in young healthy fit individuals,QO 2 kinetics being at least as fast as mitochondrial VO 2 kinetics such that mitochondrial energetics, rather thaṅ QO 2 , constrainVO 2 kinetics. 1,2 Traditional Role of Mb Mb has been regarded as an O 2 reservoir and, particularly in red, highly oxidative muscles, a primary O 2 transporting molecule 267-269; rev. 270 with both functions determined by [Mb] and extent of desaturation during exercise. In keeping with an important role for Mb in intracellular O 2 transport Mb-O 2 conductance correlates with muscle oxidative capacity across species 271;rev. 272 and [Mb] with animal mass across several orders of magnitude body mass from pigeon to blue whale (i.e., psoas muscle data from 273; rev. 272 ). In addition, for horses and steers running atVO 2 max, muscles with the highest [Mb] were those calculated to have the greatest diffusion limitation (i.e., VO 2 /DO 2 > 1). 272 An extensive modeling literature also supports the importance of Mb-facilitated intramuscular O 2 transport. 274

Mb As a Facilitator of Intramuscular O 2 Diffusion
Although championed by the Wittenbergs 278-280; rev. 266; see also 253 Mb's function as an O 2 transporter is confounded by its large mass of 17,500 kDa and 3.5 nm diameter, which restricts mobility and compromises diffusion. 280; rev. 107,266 Considering its size and concentration the effective conductance of Mb-O 2 is so low that it is thought to approximate that of free O 2 . 272,276 Accordingly, whether Mb has a quantitatively important role in intramuscular O 2 transport has been questioned. 281 125 With Mb deoxygenating during exercise greater O 2 transport capacity would be recruited as the FCDR disappears thereby increasing muscle DO 2 . Although speculative, this Mb mechanism (reduced FCDR) may potentially explain a portion of the rest-exercise increase of DO 2 that cannot be attributed to events in the capillary (i.e., elevated hematocrit and increased longitudinal recruitment of endothelial surface area). 51,119,124,125 The fact that Mb-less mice can exercise 284 would seem to mitigate against an obligatory role for Mb. However, these animals do exhibit high in utero mortality, greater systemic hematocrit and coronary blood flow reserve and capillarity but impaired cardiac and exercise performance. 285,286 It is not presently known whether larger species of Mb knockouts with greater muscle fibre cross-sectional areas are viable.
Notwithstanding the above support for Mb's importance in facilitating intramyocyte O 2 transport there is evidence that diffusion distances, in-and-of themselves, might not limit O 2 delivery. If this were indeed the case one raison d'etre for Mb would not exist. Specifically, Hudlicka and colleagues 287 noted that muscles with contrasting intercapillary distances exhibited very similar muscle/mitochondria-specific metabolic rates. see also 288 289 Also, DO 2 atVO 2 max in dog gastrocnemiusplantaris is not related to capillary density and therefore mean intramuscular diffusion distances after immobilization or exercise training. 196 Unfortunately, those studies could not determine the RBC volumes within the capillary bed, and changes thereof, that constitute a primary determinant of DO 2 . 108,109 In addition to this circumstantial evidence questioning the central role for Mb-facilitated O 2 diffusion (and appreciable O 2 storage in terrestrial mammalian muscles) Gros et al. 276 have calculated that, given the root mean square displacement for Mb-O 2 between ∼4 to 12 μm, O 2 must combine and dissociate multiple times to traverse the necessary intramyocyte distances (e.g., up to 10 to 60 μm). Moreover, Mb-O 2 must cross various potential barriers such as sarcoplasmic reticulum, myofilaments and T-Tubules although M-Lines and Z-disks may not impede this process. 282,290; rev. 107 It is also pertinent that, whilst the low intramyocyte PO 2 's 135,140,143 might facilitate O 2 transport via Mb, in part by removal of the FCDR, absence of substantial intramyocyte PO 2 gradients, from subsarcolemmal to more capillary-distant cell regions, would presumably decrease the quantitative importance of passive relative to Mb-facilitated O 2 diffusion.

Contemporary Role for Mb-The Intramyocyte Power Grid
Both cryomicrospectrophotometric 135 ,254,291-293 and P-NMR 140,142 analyses, though spatially crude, support that intramyocyte PO 2 's remain >1 mmHg. This value is far greater than the "critical" PO 2 , below whichVO 2 is constrained (i.e., 0.1 to 0.5 mmHg). 294 In contrast, mitochondrialVO 2 max versus intramyocyte PO 2 curves generated by humans inspiring a range of hypoxic, normoxic, and hyperoxic gases (i.e., see Figure 8 in 272 ) support a far higher critical PO 2 ∼4 mmHg; as do measurements made by phosphorescence quenching in the rat spinotrapezius muscle. 295,296 Interestingly, spectroscopic determination of muscle [oxidized cytochrome C oxidase (a, a 3 )] atVO 2 max reached a nadir common with that for death or anoxia, 297 which were interpreted as evidence for an ∼2 mmHg cytosol-mitochondrial PO 2 gradient. 272 However, with an estimated mitochondrial surface area being >100 times that of the functional capillary (i.e., adjacent to flowing RBCs) the corresponding O 2 flux density would be tiny; requiring a disappearingly small transmitochondrial membrane(s) PO 2 gradient. 240 205 has insightfully termed the "NO shield" (see Figure 8). Although, in many circumstances, elevating [NO] and increasing its bioavailability is an important clinical goal, there is evidence that too much intracellular NO production, for instance by iNOS in response to sepsis, reduces oxidative function. 306 That increased iNOSgenerated NO in sepsis, and other diseases, might impair muscle energetics and contractile function by compromising the NO shield's ability to distribute oxidative function across the mitochondrial reticulum, is an intriguing mechanistic hypothesis.

Skeletal Muscle Mitochondria: Structure-Function and Relevance for Understanding Capillary-Mitochondrial O 2 Transport
In contrast to the classical depiction of mitochondria as independent "bean-shaped" organelles, in myocytes, of all mammalian fibre types, they constitute a contiguous interconnected reticulum (see Figure 7) that may extend from subsarcolemma to intermyofibrillar regions. [197][198][199][200][201][202][203] Moreover, the structure of the mitochondrial reticulum itself is not fixed but dynamically modulates, potentially becoming more interconnected during exercise. 204 This observation, if confirmed, may help explain how exercise promotes such large increases in intramyocyte O 2 transport and DO 2 . In addition, the composition of the mitochondrial reticulum is heterogeneous with respect to subsarcolemmal mitochondria exhibiting greater [cytochrome C oxidase complex IV] than their intermyofibrillar whereas [ATP synthase complex V] is homogeneous. 202; rev. 205 These observations underpin the capacity for H + -electrochemical (and likely lipid/free fatty acids and O 2 ) gradients to move spatially like a power grid. In this manner charge generated at one location can produce ATP remotely. The elegance of this system provides a robust protection against focal hypoxia limiting ATP generation and abrogates dependence on long O 2 diffusion pathways from capillary to mitochondria. Without the opportunities for the "lethal corners" of tissue anoxia hypothesized by Krogh, this schema helps explain the dissociation between capillary density (therefore diffusion distances) and DO 2 noted by Hepple and colleagues. 196; see also [287][288][289] Among species and muscle fibre types and across exercise training programs muscle DO 2 can correlate with capillary volume density. 232,307 As above, this observation is consistent with DO 2 being determined principally by the volume of RBCs adjacent to the contracting muscle fibers 108,109; see also 193 and mandates using 3D models with RBC volume fractions to calculate muscle oxygenation. 308 The DO 2 dependency on capillary RBC volume means that RBC distribution and hemodynamics can, via altered DO 2 , impact PimO 2 regulating muscle metabolic control and therefore contractile performance. 141,[247][248][249] In addition to upstream (i.e., extramyocyte) determinants of DO 2 it is important to consider all the potential pathways of intramyocyte O 2 transport ( Figure 11). We have, based upon capillary 3D geometry, made the case for the Hill model of O 2 diffusion being more appropriate than Krogh's "pinpoint" capillary O 2 supply. In addition, the demonstrated movement of O 2 among adjacent capillaries and other vessels supports that the interstitial space, which has a greater PO 2 than PimO 2 , 131-133 constitutes an alternative/additional pathway for trans-sarcolemmal O 2 flux. Such a mechanism could greatly increase the available surface area, which now becomes the entire sarcolemma, for O 2 diffusion (and hence DO 2 ) above that provided by the direct capillary-sarcolemmal "contact" considered by Ellis and colleagues when proposing the Hill model. 130 Thus, the geometry of intramyocyte O 2 transport is very different from that hypothesized in the Krogh model. Tissue and mitochondrial volume decrease with increasing distance from the sarcolemma (c.f., Figure 5, left and right) with O 2 moving around and into the myocyte at sites remote from capillaries. Intramyocyte PO 2 profiles are flat with PO 2 gradients not detectable. Within the myocyte there will be free O 2 and Mb-facilitated O 2 transport-though the absolute and relative importance of each is debated especially given the absence of substantive O 2 gradients. 276 In addition, Pias 309 has presented an intriguing, though mostly theoretical, case for O 2 transport tracking along high O 2 -conductance lipid membranes. Thus, there are at least 3 putative parallel paths for intramyocyte O 2 transport ( Figure 11). These pathways, conflated with the ability of the mitochondrial reticulum to track H + -electrochemical gradients deep into the fiber, [197][198][199][200][201][202][203] would, presumably, mitigate (or remove entirely) O 2 flux constraints resulting from capillary-mitochondrial O 2 diffusion distances.

Intravital Microscopy of Skeletal Muscle: Essential Quality Control
In contrast to the right side of Table 1 (opposing capillary recruitment) the left side exemplars studies in which it is concluded that most (or many) capillaries in resting skeletal muscle do not support RBC flow at rest and are thus available to be recruited during exercise. Whereas many papers interpret their indirect findings in terms of capillary recruitment without actually observing the capillary bed-a practice harshly criticized by the eminent microscopist Professor Eugene Renkin 310there are also intravital microscopy studies supporting that the majority of (or at least many) capillaries in resting muscle do not support RBC flow (see left side of Table 1). In this regard it is pertinent that capillaries are extremely delicate structures that can easily be damaged by the surgical intervention necessary to visualize the microcirculation and also the structure and function of which is impacted by muscle length (especially stretching), arteriolar control (neural and humoral), and their physicochemical microenvironment. The following are just a few of the experimental conditions, lack of attention to which, will decrease the proportion of capillaries supporting RBC flux and, as experienced by Krogh's stretched, hyperoxic frog tongue preparation, set the stage for contraction-induced capillary recruitment.
1. Anesthesia, especially in combination with systemic hypoxia, can cause hypotension decreasing muscleQ (and PmvO 2 ) via reduced driving pressure and also a sympathetically induced arteriolar vasoconstriction. 311 2. Local muscle hyperoxic and hypoxic conditions should be specifically avoided as they will produce nonphysiological vasoconstriction (hyperoxia) and muscle dilation (hypoxia).
The former will have direct consequences for decreasing the proportion of capillaries that support RBC flux. Phosphorescence quenching measurements have defined the microvascular and interstitial PO 2 's (i.e., PmvO 2 , PisO 2 ) in resting and exercising muscles across the fibre-type spectrum (see Figure 3 #4) [131][132][133] and it is now possible to set the appropriate physiological conditions, at least for PO 2 . 3. Sharp or blunt force trauma during surgery can prevent arterioles and capillaries flowing. Aggressive removal of overlying fascia to improve visual clarity may also damage the microcirculation and should be avoided/minimal. 4. Muscle stretch above 3 μm, often practiced to thin the muscle and improve microcirculation optical clarity, stretches the capillaries, reducing their luminal diameter and vascular conductance and stopping RBC flux in a substantial proportion of capillaries. 88 Stretching muscles also produces a reflex vasoconstriction, which compounds the hyporemia and capillary no-flow. 312 Attention to these and other physiological conditions (temperature, hormonal, fluid status) insofar as possible, is simply good science. Where such a standard is not met, the results must be interpreted accordingly. With respect to muscle oxygenation this is especially true, not just for microcirculatory, but also for metabolic control. Specifically, even disappearingly small changes in intramyocyte PO 2 (PimO 2 ), evoked consequent to altered PmvO 2 and PisO 2 (and thus PimO 2 ) can exert a commanding influence on myocyte energetics, substrate utilization and metabolic control. 247-249; rev. 313; see also 141,314 Moreover, inattention to establishing physiologic PO 2 profiles can destroy our ability to, for example, determine the effect of manipulating the nitrate-nitrite-NO pathway on contractile performance. 315

Conclusions
Effective progress in muscle microcirculatory research will be facilitated by: Advocating for rigorous quality control of intravital microscopy investigations and interpreting previous findings, theories, and models accordingly.
Recognizing that different organ microcirculations subserve wide-ranging functions and those elements critical for capillary hemodynamic control in muscle may work very differently for heart, mesentery, gut, and especially the brain.
Testing novel or extant hypotheses within the context of physiological control. For instance, as regards the role of pericytes in closing/opening capillaries, ask whether their response kinetics are sufficiently fast to support theQ and capillary hemodynamics demonstrated following the onset of contractions? In brain and heart the pericyte-associated constriction is slight and too slow to explain the hyperemia kinetics following contractions onset (c.f. in brain, 208,210,316 and heart, 211 with skeletal muscle 89 ). Is there really the possibility for unitary capillary control in muscle, or rather, is fine-tuning ofQO 2 -VO 2 matching achieved by ascending vasodilation that increases terminal arteriole diameter and thus RBC flux to all capillaries in a given capillary module. 102,106,[259][260][261][262] This latter process may be accentuated via venular signaling. 102,106 The elegant NIRS technology employed by Bowen and colleagues 317 suggests that temporalQO 2 -VO 2 mismatching inand-of-itself which leads to a deoxygenation overshoot does not appear to slowVO 2 kinetics; presumably limiting the functional consequences of such fine-tuning ofQO 2 -VO 2 matching. Whereas we have supported the case, in health, forQO 2 kinetics being as fast, or faster, than that ofVO 2 (see Myth #6 above) there is evidence during knee-extension exercise that estimated cap-illaryQO 2 kinetics may be slower (Harper et al. 318 ; but see also Schlup et al. 319 ) than that of the femoral artery and this phenomenon is expected to produce a deoxygenation overshoot. Future studies might judiciously account for capillary hemodynamics rather than relying on the tacit presumption that such kinetics are the same as those for arterialQ.
Recognizing that strong physiological evidence supports that capillary-mitochondrial diffusion distances do not determine DO 2 . Rather, a key determinant of DO 2 is the RBC volume of flowing capillaries, which may reflect the importance of newly discovered aquaporin and rhesus pores in the RBC membrane and also the high RBC-capillary endothelium O 2 flux density.
Understanding more about the role of the endothelial surface layer (glycocalyx) in determining capillary hematocrit and hemodynamics, especially during exercise, in health and perturbations of such in disease. In vivo NIRS measurements during heavy intensity exercise suggest that capillary hematocrit may not reach systemic levels even at very high musclė Qs. 120-122 243 Considering that O 2 can diffuse between arterioles, capillaries, and venules 136,137,266 and the presence of higher PisO 2 than (independently) measured for PimO 2 raises the intriguing possibility that O 2 movement through the interstitial space-around the myocyte periphery and among myocytes-may be important. If so, this provides additional support for the Hill (versus Kroghian "pinpoint") model of O 2 diffusion ( Figure 5) 107,130 and this capability would facilitateQO 2 -VO 2 matching irrespective of the marked heterogeneities of capillary RBC flux/hematocrit observed and potentially help explain the hyperbolic rise of O 2 extraction with increasingVO 2 . Of course, it remains to be determined to what extent a higherQO 2 -VO 2 (and thus PmvO 2 ) in one region might be beneficial if total muscleQO 2 is constant and such behavior therefore predicates a low PmvO 2 in another region.
Determining the quantitative importance of the multiple pathways for intramyocyte O 2 transport-aqueous and myoglobin-facilitated O 2 diffusion and, potentially, that O 2 tracks along lipid membranes during heavy and severe exercise ( Figure 11). This is also true for the role of the mitochondrial reticulum tracking charge potential across the myocyte and the putative role of deoxymyoglobin to enact local NO release thereby controlling cytochrome oxidase activity and protecting ATP production across the myocyte (Figure 8). Development of highly spatially resolved O 2 sensitive dyes with an active range ∼0.5 to 10 mmHg and rapid response in the subsecond range will be invaluable to this mission.
Incorporating the latest aspects of muscle capillary function models (e.g., Figure 4) into the design and interpretation of investigations. This is especially pertinent, not only in health, but in unveiling the mechanistic bases for dysfunction in diseases such as HF, diabetes, sepsis, and many others.