Reply to Letter to the Editor: Perfusion controls muscle glucose uptake by altering the rate of glucose dispersion in vivo

Letter to the EditorReply to Letter to the Editor: Perfusion controls muscle glucose uptake by altering the rate of glucose dispersion in vivoP. Mason McClatchey, Ian M. Williams, Zhengeng Xu, Nicholas A. Mignemi, Curtis C. Hughey, Owen P. McGuinness, Joshua A. Beckman, David H. Wasserman, David C. Poole, Thorbjorn Akerstrom, Daniel Goldman, Graham M. Fraser, and Christopher G. EllisP. Mason McClatcheyDepartment of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nashville, Tennessee, Ian M. WilliamsDepartment of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nashville, Tennessee, Zhengeng XuDepartment of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nashville, Tennessee, Nicholas A. MignemiDepartment of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nashville, Tennessee, Curtis C. HugheyDepartment of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nashville, Tennessee, Owen P. McGuinnessDepartment of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nashville, Tennessee, Joshua A. BeckmanDepartment of Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee, David H. WassermanDepartment of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nashville, Tennessee, David C. PooleDepartments of Anatomy & Physiology and Kinesiology, Kansas State University, Manhattan, Kansas, Thorbjorn AkerstromDepartment of Nutrition, Exercise and Sports, Section of Integrative Physiology, University of Copenhagen, Copenhagen, Denmark, Daniel GoldmanDepartment of Medical Biophysics, Schulich School of Medicine & Dentistry, University of Western Ontario, London, Canada, Graham M. FraserDivision of BioMedical Sciences, Faculty of Medicine, Memorial University of Newfoundland St. John's, Newfoundland, Canada, and Christopher G. EllisDepartment of Medical Biophysics, Schulich School of Medicine & Dentistry, University of Western Ontario, London, CanadaPublished Online:18 Feb 2020https://doi.org/10.1152/ajpendo.00508.2019MoreSectionsPDF (1 MB)Download PDF ToolsExport citationAdd to favoritesGet permissionsTrack citations ShareShare onFacebookTwitterLinkedInEmail to the editor: We thank Keske et al. (13a) for their interest in the paper of McClatchey et al. (23) and we are grateful for the opportunity to respond to their letter. Keske et al. revisit a discrepancy in the interpretation of experiments claiming measurement of skeletal muscle capillary recruitment. Their letter is enlightening because it reveals the bases of misinterpretation that have led to the misconception that contrast enhanced ultrasound (CEU) can distinguish skeletal muscle capillary recruitment from other forms of flow redistribution.Keske et al. contend that a high percentage of capillaries are not perfused in the basal state and that hyperinsulinemia recruits these capillaries. McClatchey et al. (21, 23) have now twice shown that ~95% of capillaries are perfused in the basal state. A corollary to this is that there is no meaningful margin for capillary recruitment. Keske et al. assert that the studies of McClatchey et al. are limited because there is a "wealth of data" obtained using noninvasive methods as well as intravital microscopy (IVM) "demonstrating only 30–50% of muscle capillaries" are perfused at rest. In our response we will explain that the noninvasive methods make indirect measurements that have not been validated for the purpose of measuring capillary recruitment. IVM methods are direct visual observations and application of these techniques over the last three decades overwhelmingly support the findings of McClatchey et al. (21, 23). A more accurate summary of the differences in the proportion of perfused capillaries is that indirect measurements have been interpreted as claiming that less than 50% of capillaries are perfused at rest, whereas direct visualization shows that nearly all capillaries are perfused. In our response we explain the reason that CEU is not currently useful for measuring capillary recruitment and emphasize the need to validate indirect measurements.The content in the letter by Keske et al. is similar to that published in previous forums (3, 7, 8). McClatchey et al. anticipated that the findings would provoke this response and designed experiments to address the concerns that had been expressed previously. Despite this, we still saw that ~95% of capillaries are perfused in the basal state. This result agrees with studies showing that when 1) the microcirculation is directly visualized and 2) muscle oxygenation is appropriately maintained nearly all capillaries are perfused at baseline (1, 4, 14, 15, 21, 25, 26, 31). McClatchey et al. undertook experiment 3 in response to a reviewer who wished to know why results obtained using IVM appeared to contradict the interpretation of CEU data. The results of this experiment clearly demonstrated that the concentration of microbubbles used for CEU is too low to populate all capillaries [Fig. 9 of McClatchey et al. (23)]. We will elaborate upon the mathematics that make this experimental observation totally predictable later in our response.Keske et al. primarily defends the capillary recruitment-based interpretation of CEU data rather than make an in-depth analysis of the experiments of McClatchey et al. (23). The concerns expressed by Keske et al. were already addressed in the published paper. We will cite the relevant sections in McClatchey et al. (23) and briefly reiterate. First, muscle was not exteriorized (as inaccurately stated in the letter) but, rather, simply exposed as described (Section 2.1b). The muscle remained in its anatomical position, but the skin and fascia were dissected away from the field of vision. Temperature, Po2, and normal innervation are maintained using this preparation (21, 23, 24, 36–38). Anesthesia is a legitimate consideration, and we previously reported subtle perfusion differences with different anesthetics (21). However, as stated in the discussion (4th paragraph) of McClatchey et al. (23), "using IVM techniques ensuring normal muscle oxygenation, high baseline capillary recruitment is consistently observed across multiple species (including mice, hamsters, and rats), multiple anesthetic regimens (including isoflurane, ketamine/xylazine, pentobarbital), and multiple surgical preparations (4, 14, 15, 21, 25, 26, 31)". Moreover, the same result has been observed in experiments that require neither anesthesia nor dissection (34). Most recently, Akerstrom et al. (1) published a thorough paper that reached the same conclusion as McClatchey et al. (23). The surgical preparation used in this paper neither exteriorizes the muscle nor disrupts the surrounding fascia. Thus, although the potential experimental confounders suggested by Keske et al. may influence aspects of the microcirculation, they do not appear to influence the proportion of capillaries perfused at baseline.We would like to specifically address an argument of Keske et al. that, we feel, distracts from this discourse. Keske et al. reference papers from the literature as visual evidence of capillary recruitment (9, 12, 16, 19, 32). The most recent of these references was published nearly 30 yr ago. We share their appreciation of this older work. However, there are two key issues that cannot be dismissed. First, scientists did not cease visualizing the microcirculation after the earlier investigations cited by Keske et al. They continued with the benefits of improved technology. The overwhelming majority of studies in the three decades that followed did not support the capillary recruitment hypothesis as the proportion of perfused capillaries is high at baseline (1, 4, 14, 15, 21, 23, 25, 26, 31). Second, even those papers selected by Keske et al. contradict or do not address their argument. The papers cited by Keske et al. were published in 1991 or before. One of these, the study of Lindbom (11), could not be accessed and is not addressed. Experimental support for capillary recruitment is scant in the remaining four papers. In the study by Segal (32), only the capillaries that were not perfused at rest were considered and the proportion of perfused to nonperfused capillaries was not quantified. Segal states that "substantially higher rbc flux and content were observed but could not be quantified due to the inability to resolve higher values using frame-by-frame analysis" (32). Thus, the author fairly acknowledges that only capillaries with especially low flow velocities were characterized (this is reflected by the low flow velocities reported in this paper). The resulting data are consistent with McClatchey et al. (Fig. 5D), which shows that low flow rates correspond to a decreased proportion of perfused capillaries (23). The other three papers cited by Keske et al. were from Duling and colleagues (9, 12, 16). The earliest of these does indeed report a large number of nonperfused capillaries in resting muscle (12). However, the second paper discovered that this effect only occurred at oxygen levels that are now known to be supra-physiological (16, 26). By the time Duling and colleagues published the third paper cited by Keske et al., oxygen tension was better controlled, and the first sentence of the Results states that "under control conditions, all capillaries originating from a terminal arteriole were perfused" (9). This conclusion is the opposite of what Keske et al. assert and the same as McClatchey et al. and the vast majority of the literature report (1, 4, 14, 15, 21, 23, 25, 26, 31).Keske et al. continue by asserting that "increasing velocity in a fixed number of capillaries will not change the number of microspheres in that capillary bed." This assertion is demonstrably false. Keske et al. err by applying standard tracer theory to particulates, which experience different microvascular dynamics from those of soluble tracers. Phase separation effects result in differential distribution of particulates (e.g., microspheres, microbubbles, RBCs, etc.) relative to plasma at microvascular bifurcations based on the properties of both the particulates themselves as well as the endothelial surface layer (2, 6, 11, 22, 27, 35, 39). Particulates are preferentially distributed to high velocity central capillaries. This is a reason for the well known hematocrit heterogeneity in capillaries (10, 28–30), which changes in proportion to flow velocity. These physical principles apply to microbubbles/microspheres as well. It is very possible that the increased CEU signal Keske et al. attributes to capillary recruitment is instead caused by more microbubbles flowing through a finite number of high-velocity capillaries. Physical principles aside, McClatchey et al. prove experimentally that increased velocity within a fixed number of capillaries can increase the number of microspheres in the capillary bed (Fig. 9E of McClatchey et al.), and this effect is independent of plasma-perfused volume or surface area. Meanwhile, the compelling studies of Akerstrom et al. show that insulin homogenizes blood flow rather than recruiting capillaries (1). Akerstrom et al. perform simulations predicting that microsphere-perfused volume and microsphere concentration both vary with perfusion homogeneity/heterogeneity. This supports the findings of McClatchey et al. (23) and explains the results typically obtained with CEU.Importantly, CEU dependency on the velocity and homogeneity/heterogeneity of blood flow is in complete accord with previous theoretical publications (20, 22). These theories successfully predicted group differences in muscle metabolism and substrate dispersion by extrapolating from blood flow and its distribution as demonstrated in Fig. 7 of McClatchey et al. (23). We are not aware of a similar validation of the capillary recruitment-based paradigm for interpretation of CEU data. It seems that evidence for the capillary recruitment paradigm has come from a circular process of naming the signals resulting from indirect techniques (CEU, laser-Doppler flowmetry, positron emission tomography, and 1-methylxanthine metabolism) "capillary blood volume" and concluding that capillary recruitment must have occurred because the so-named "capillary blood volume" increased. The legitimacy of the underlying data is not in dispute. However, the assumption that these data measure capillary blood volume is difficult to accept as none of these techniques have been validated for this purpose and direct observation do not generally show capillary recruitment (1, 4, 14, 15, 21, 23, 25, 26, 31).We appreciate that there are also limitations in our own experimental models and theoretical constructs, and that these limitations may yet prove to be important. However, we can assert with certainty that measurement of the proportion of perfused capillaries with current CEU techniques is statistically implausible (5th paragraph of the Discussion of McClatchey et al.) and that a flow velocity- and perfusion heterogeneity-based model better reconciles CEU and IVM findings. As indicated in McClatchey et al. (23), this updated paradigm does not detract from CEU as a valuable noninvasive technique nor does it challenge the concordance of CEU with measures of muscle metabolism. Whatever the CEU signal represents is clearly relevant. However, CEU is fundamentally incapable of distinguishing between capillary recruitment and other forms of hyperemia or flow redistribution.To distinguish capillary recruitment, CEU microbubbles would need to distribute volumetrically in the capillary bed in proportion to either RBCs or plasma. Such a proportionality has not been demonstrated under any conditions, let alone metabolic transients. The dose of microbubbles commonly used for CEU creates a ratio of a single microbubble for every ~6,000 RBCs (33). It is highly improbable at this ratio that microspheres can capture the dynamics of whole blood. In experiment 3 of McClatchey et al. (23), microspheres were injected to create a ratio of 1 microsphere for every ~2,400 RBCs. RBC flow rate was measured as 17 RBCs per second per capillary for 15 s. It can be calculated that 255 RBCs per capillary were then measured (17 RBC/s × 15 s). From this, and the RBCs to microsphere ratio, we can calculate that there is an ~11% chance of observing a microsphere in any one capillary during data acquisition. In a CEU experiment where a RBCs to microbubble ratio of 6,000 is used, the probability of a microbubble appearing in any one capillary during 15 s of data acquisition would be 4%. Basic calculations, which are also outlined in the Discussion (5th paragraph) of McClatchey et al. (23), show that it is a near certainty that many capillaries perfused with plasma and RBCs will not contain a microbubble (23). It is irrelevant to these mathematical relationships whether one is comparing RBC number to CEU microbubbles or latex microspheres of similar size and concentration. These calculations are also consistent with our experimental observations of microsphere distribution in vivo (Fig. 9, A–C, in McClatchey et al. and Fig. 1, A and B). Moreover, as addressed in the Discussion (3rd paragraph) of McClatchey et al. (23), capillary recruitment-based interpretation of CEU data yields patently nonphysiological values for capillary volume density in skeletal muscle (5). A flow velocity- and perfusion homogeneity/heterogeneity-based interpretation of CEU data is capable of accounting for the concordance of CEU data with muscle metabolism without requiring similar intellectual contortions. The capillary recruitment hypothesis is not required to explain the results from CEU (or any other technique cited by Keske et al.) and it is uniquely ill-founded when comparing to direct visual observations of the microcirculation.Fig. 1.Intravital microscopy (IVM; A–D) and contrast-enhanced ultrasound (CEU; E) images of the muscle circulation. A: map of plasma-perfused capillaries in mouse skeletal muscle. B: map of 3 µm (microbubble-sized) microsphere-perfused capillaries in the same field of view (FOV) as A. The quantity of microspheres given was comparable to the amount given in CEU experiments. Only ~50% of plasma-perfused capillaries contain microspheres. C: map of basal RBC distribution in rat skeletal muscle. D: map of RBC distribution during hyperinsulinemia in the same FOV as C. E: CEU image of a skeletal muscle before, during 10% isometric exercise, and immediately after isometric exercise in the same FOV. A and B are borrowed with permission from McClatchey et al. (23), and E is borrowed with permission from Krix et al. (17).Download figureDownload PowerPointWhy is it important to know whether capillary recruitment occurs or not? Among many compelling reasons to describe accurately microvascular dynamics is to recognize that the proportion of capillaries perfused under basal conditions is a hallmark of disease (13, 25, 26, 31). The pathophysiological changes detected by CEU are likely to hold functional importance. There may be disease states characterized by a decrease in the proportion of perfused capillaries. However, despite over 100 yr of investigation (18), capillary recruitment has yet to be successfully targeted for either diagnostic or therapeutic purposes.In closing, contemporary techniques capable of resolving individual capillaries do not detect capillary recruitment in normoxic muscle (1, 4, 14, 15, 21, 23, 25, 26, 31). However, they do show concordance with CEU data that are interpreted using a paradigm based on flow velocity and perfusion heterogeneity (1, 20). The capillary recruitment-based interpretation of the CEU signal is incongruent with both basic mathematics and direct observations of the microcirculation (1, 23). In the interest of full disclosure, the authors of McClatchey et al. began their studies intending to build upon the capillary recruitment paradigm. During the experimental process, it was realized that the premise underlying capillary recruitment was fundamentally flawed. This led to an overhaul of planned studies and a renewed focus. As the Royal Society's motto entreats us, "Nullius in Verba": Take no one's word for it, see it for yourself. In this spirit, we have included a composite figure to show raw data on which this debate is based. Figure 1, A and B, compares IVM-visualized plasma-capillaries to microsphere-perfused capillaries, respectively, in the same field of vision (23). A dose of microspheres comparable to the dose given in CEU experiments populates ~50% of capillaries. Figure 1, C and D, shows IVM-visualized RBC distribution before and during a hyperinsulinemic euglycemic clamp, respectively, in the same field of vision. There is no capillary recruitment during the hyperinsulinemic euglycemic clamp. Figure 1E shows example raw CEU images of a skeletal muscle before exercise, during isometric exercise, and after isometric exercise (17). The question to ask is whether raw data, such as these, can distinguish capillary recruitment from hyperemia or blood flow distribution when given the physical and mathematical limitations that have been outlined?DISCLOSURESNo conflicts of interest, financial or otherwise, are declared by the authors.AUTHOR CONTRIBUTIONSD.G., G.M.F., and C.G.E. prepared figures; P.M.M. and D.H.W. drafted manuscript; P.M.M., I.M.W., Z.X., N.A.M., C.C.H., O.P.M., J.A.B., D.H.W., D.C.P., T.A., D.G., G.M.F., and C.G.E. edited and revised manuscript; P.M.M., I.M.W., Z.X., N.A.M., C.C.H., O.P.M., J.A.B., D.H.W., D.C.P., T.A., and G.M.F. approved final version of manuscript.REFERENCES1. Akerstrom T, Goldman D, Nilsson F, Milkovich SL, Fraser GM, Brand CL, Hellsten Y, Ellis CG. Hyperinsulinemia does not cause de novo capillary recruitment in rat skeletal muscle. Microcirculation. In press. doi:10.1111/micc.12593. Crossref | PubMed | Google Scholar2. Barber JO, Restrepo JM, Secomb TW. Simulated red blood cell motion in microvessel bifurcations: effects of cell-cell interactions on cell partitioning. Cardiovasc Eng Technol 2: 349–360, 2011. doi:10.1007/s13239-011-0064-4. Crossref | PubMed | Google Scholar3. Barrett EJ, Keske MA, Rattigan S, Eringa EC. Rebuttal from Eugene J. Barrett, Michelle A. Keske, Stephen Rattigan and Etto C. Eringa. J Physiol 592: 5137–5138, 2014. doi:10.1113/jphysiol.2014.284604. Crossref | PubMed | ISI | Google Scholar4. Cabrales P, Vázquez BY, Tsai AG, Intaglietta M. Microvascular and capillary perfusion following glycocalyx degradation. J Appl Physiol (1985) 102: 2251–2259, 2007. doi:10.1152/japplphysiol.01155.2006. Link | ISI | Google Scholar5. Chadderdon SM, Belcik JT, Smith E, Pranger L, Kievit P, Grove KL, Lindner JR. Activity restriction, impaired capillary function, and the development of insulin resistance in lean primates. Am J Physiol Endocrinol Metab 303: E607–E613, 2012. doi:10.1152/ajpendo.00231.2012. Link | ISI | Google Scholar6. Chien S, Tvetenstrand CD, Epstein MA, Schmid-Schönbein GW. Model studies on distributions of blood at microvascular Am J Physiol Link | ISI | Google Rattigan S, Barrett EJ, There is capillary recruitment in skeletal muscle during J Appl Physiol (1985) Link | ISI | Google Rattigan S, Barrett E, M. word on There not capillary recruitment in skeletal muscle during J Appl Physiol (1985) Link | ISI | Google Duling A study of the functional capillary perfusion in muscle. Crossref | PubMed | ISI | Google Duling of the measurement of flow heterogeneity in muscle. Crossref | PubMed | ISI | Google red cell distribution in to 100 Crossref | PubMed | ISI | Google B, Duling between muscle and Am J Physiol Link | ISI | Google Exercise in skeletal muscle and Am J Physiol Physiol Link | ISI | Google Keske MA, Barrett EJ, Lindner CD, D, SM, Rattigan S, Eringa EC. Perfusion controls muscle glucose uptake by altering the rate of glucose dispersion in Am J Physiol Endocrinol | ISI | Google A of the microcirculation in the rat and Crossref | PubMed | ISI | Google muscle capillary from rest to for oxygen J Appl Physiol (1985) Link | ISI | Google B, Duling oxygen during muscle in the of capillary recruitment, functional and Crossref | PubMed | ISI | Google Krix MA, S, H. in the of skeletal muscle due to isometric exercise by contrast-enhanced J Crossref | PubMed | ISI | Google A. a to the of the capillaries. Physiology or Medicine Lindbom Microvascular blood flow distribution in skeletal muscle. intravital study in the Physiol PubMed | Google McClatchey A for predicting and the of microvascular Crossref | PubMed | ISI | Google McClatchey of microvascular Crossref | PubMed | ISI | Google McClatchey The endothelial glycocalyx blood flow distribution within the Am J Physiol Physiol Link | ISI | Google McClatchey Perfusion controls muscle glucose uptake by altering the rate of glucose dispersion in Am J Physiol Endocrinol Metab Link | ISI | Google McClatchey CL, changes in the microvascular of skeletal muscle insulin during Am J Physiol Endocrinol Metab Link | ISI | Google P, Y, of on capillary in skeletal muscle. Am J Physiol Physiol Link | ISI | Google muscle capillary contemporary observations and Physiol Crossref | PubMed | ISI | Google P. cell distribution at microvascular Crossref | PubMed | ISI | Google P. of the for microvessel Am J Physiol Physiol Link | Google Secomb TW. of the PubMed | ISI | Google Secomb TW. Microvascular blood in vivo and the endothelial surface Am J Physiol Physiol Link | ISI | Google of on skeletal muscle capillary at rest and during J Appl Physiol (1985) Link | ISI | Google Segal Microvascular recruitment in for Am J Physiol Physiol Link | ISI | Google L, B, Rattigan S, Exercise in Microvascular Perfusion and Molecular Crossref | PubMed | ISI | Google JR. perfusion in skeletal muscle. Am J Physiol Physiol Link | ISI | Google Intaglietta dispersion in Am J Physiol Physiol 2014. Link | ISI | Google McClatchey In Crossref | PubMed | ISI | Google D, skeletal muscle capillaries by J Crossref | PubMed | ISI | Google McClatchey insulin is impaired in skeletal muscle capillaries of T, J. red blood cell through microvascular bifurcations: cell cell and hematocrit Crossref | PubMed | ISI | Google for and other H. Wasserman, of Molecular Physiology and Biophysics, Vanderbilt School of Medicine, Nashville, to controls muscle glucose uptake by altering the rate of glucose dispersion in vivo Feb of and from this & the in 1

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.