1. Human Biology and Medicine
  2. Neuroscience
Download icon

Mapping oxygen concentration in the awake mouse brain

  1. Declan G Lyons
  2. Alexandre Parpaleix
  3. Morgane Roche
  4. Serge Charpak  Is a corresponding author
  1. Institut National de la Santé et de la Recherche Médicale, U1128, France
  2. Université Paris Descartes, France
Research Article
  • Cited 26
  • Views 2,543
  • Annotations
Cite this article as: eLife 2016;5:e12024 doi: 10.7554/eLife.12024

Abstract

Although critical for brain function, the physiological values of cerebral oxygen concentration have remained elusive because high-resolution measurements have only been performed during anesthesia, which affects two major parameters modulating tissue oxygenation: neuronal activity and blood flow. Using measurements of capillary erythrocyte-associated transients, fluctuations of oxygen partial pressure (Po2) associated with individual erythrocytes, to infer Po2 in the nearby neuropil, we report the first non-invasive micron-scale mapping of cerebral Po2 in awake, resting mice. Interstitial Po2 has similar values in the olfactory bulb glomerular layer and the somatosensory cortex, whereas there are large capillary hematocrit and erythrocyte flux differences. Awake tissue Po2 is about half that under isoflurane anesthesia, and within the cortex, vascular and interstitial Po2 values display layer-specific differences which dramatically contrast with those recorded under anesthesia. Our findings emphasize the importance of measuring energy parameters non-invasively in physiological conditions to precisely quantify and model brain metabolism.

https://doi.org/10.7554/eLife.12024.001

eLife digest

Brain cells need a constant supply of oxygen to fuel their activities. This oxygen is delivered by the flow of blood through the vessels in the brain. If the blood flow to brain tissue is cut off as happens in stroke, or if an individual stops breathing, the brain becomes deprived of oxygen and brain cells will be damaged and die. To better understand how the brain works in health and disease, scientists need to learn how much oxygen the blood must deliver to the brain tissue to adequately support the activities of brain cells.

Many studies have measured oxygen levels in the brain. However, these studies have looked only roughly and taken measurements from large areas of the brain, or they have involved animals receiving anesthesia, which can alter blood flow and oxygen use in the brain. Recently, scientists discovered that they could measure oxygen concentration at high detail in the brain of anesthetized rodents with a specialized microscope, by using molecules that emit light at a rate that depends on the local oxygen concentration.

Now, Lyons et al. have shown that this same technique can be used in mice that are awake. First, a piece of the skull was replaced with glass to create a small transparent window. Then, the animals were allowed to recover for a few weeks, and were trained to get them used to how they would be handled during the experiments. After this period, the oxygen concentrations and blood flow in different parts of the mouse brains were measured in fine detail using the microscope while the animals were awake and relaxed.

The experiments showed that oxygen levels in awake resting mice are actually lower than in anesthetized mice, and that oxygen levels differ between different parts of the mouse brain. This first detailed look at oxygen levels in the brain of awake animals will likely lead to more studies. For example, future studies may look at how quickly the brain uses oxygen under normal conditions and what happens in the brain during disease.

https://doi.org/10.7554/eLife.12024.002

Introduction

To understand the relationship between brain oxygenation and diseases associated with hypoxia or ischemia, it is important to first determine what fixes the resting value of tissue Po2, that is, the concentration of oxygen in the interstitium that bridges oxygen delivery from erythrocytes to oxygen consumption by mitochondria. Numerous methods have been used to monitor brain oxygenation, and the most spatially resolved approaches have long relied on fine Clark-type electrodes (for review see Ndubuizu and LaManna, 2007), which have been used to report resting-state Po2 and local oxygen consumption in various brain regions (Lecoq et al., 2009; Masamoto et al., 2003; Offenhauser et al., 2005; Thompson et al., 2003) during neuronal activation. However, these electrodes are invasive, do not faithfully report Po2 in vessels and cannot easily be used to determine Po2 in physiological conditions, that is, in awake, unstressed animals, avoiding the use of anesthetics. As anesthetics affect resting and evoked neuronal and astrocyte activity, arterial blood pressure and cerebral blood flow, the physiological values of cerebral interstitial Poand their relationship to blood flow parameters in capillaries remain unknown.

Recently, a two-photon phosphorescent probe PtP-C343 has been generated (Finikova et al., 2007, 2008) and two-photon phosphorescence lifetime microscopy (2PLM) has been used to obtain depth-resolved, micron-scale measurements of Po2 in the anesthetized rodent brain (Devor et al., 2011; Lecoq et al., 2011; Parpaleix et al., 2013; Sakadzić et al., 2010; Sakadžić et al., 2014). In addition, by detecting single red blood cells (RBCs) during Po2 measurement, we demonstrated the possibility of simultaneously monitoring blood flow and Po2 in capillaries (Lecoq et al., 2011) and of detecting erythrocyte-associated transients (EATs), Po2 fluctuations associated with each individual erythrocyte flowing in capillaries, which were first reported in mesentery capillaries (Golub and Pittman, 2005). We showed that in olfactory bulb glomeruli of anesthetized mice, one parameter of EATs, the Po2 in between two red blood cells (Po2InterRBC), is at equilibrium with, and thus reports, the Po2 in the nearby neuropil (Parpaleix et al., 2013). This result implied that measurements of Po2InterRBC could provide a powerful tool to non-invasively map local interstitial oxygen concentration in the brain of awake animals.

Here, we report that in both the olfactory bulb glomerular layer and the somatosensory cortex of unstressed, awake, resting mice, the interstitial Po2 (equivalent to Po2InterRBC) has the same mean value of ~23 mm Hg, spanning over a range of about 40 mm Hg. This contrasts with the large differences of capillary hematocrit and RBC flow values observed between the two brain regions. In addition, we find that in the cortex capillary and interstitial Po2 values display layer-specific differences, being lower in layer I than in layer II/III or layer IV. We also find that hemoglobin in brain capillaries is highly saturated with oxygen. Finally, we show that in both brain regions, the interstitial Po2 is much lower during wakefulness than under isoflurane anesthesia.

Results

Habituation of mice to head-restraint

To ensure that we measured Po2 in real physiological conditions, that is, in awake, unstressed animals, each animal was habituated to all the conditions present during 2PLM Po2 measurements for several weeks prior to the experiments (see methods for detailed training procedures). In brief, over the course of 2–3 days, each mouse was habituated to handling, and trained to run on a treadmill placed in its cage. A titanium bar was then surgically attached to the cranium and then a cranial window implanted over the region of interest, either the olfactory bulb or the somatosensory cortex. Then, over 2–4 weeks, the mouse was progressively habituated to being head-fixed, via the attached bar, in the dark, below the objective of the two-photon microscope, and with the limbs and body free to move on the treadmill. Habituation was achieved when the animal remained calm for periods >1 hr in the set-up with short bouts of running (~30 s). On the day of recording, the animal was briefly anesthetized (2% isoflurane, <5 min) and the oxygen sensor PtP-C343 was injected intravenously. The animal was returned to its home cage, and after a delay of 90–120 min, Po2 recordings sessions of 1–3 hr commenced. Each animal underwent 1–3 recording sessions per day over the course of 2–7 days, with breaks of at least several hours between each session. Note that similar Po2 values were obtained from one day to the next and between sessions occurring the same day (without reinjection of PtP-C343).

Characterization of EAT properties (Po2 values) in the olfactory bulb of awake resting mice

Using our previous approach (Lecoq et al., 2011; Parpaleix et al., 2013), we characterized EATs in 38 capillaries (n = 5 animals) from the glomerular layer of awake resting mice (Figure 1). Po2 measured at the RBC border (Po2RBC) was significantly larger than at mid-distance between two RBCs (Po2InterRBC). Po2Mean, which was intermediate between these two values, is the average Po2 measured in a capillary without taking into account the existence of EATs, and is the only capillary Po2 value that has commonly been reported in the brain (Sakadzić et al., 2010; Vovenko, 1999). Several measurements were made in each capillary but the average values of Po2RBC, Po2InterRBC and Po2Mean were similar whether calculated on all measurements or on all capillaries (262 measurements: Po2RBC = 60.5 ± 0.9 mm Hg; Po2InterRBC = 23.4 ± 0.5 mm Hg; Po2Mean = 32.8 ± 0.7 mm Hg. 38 capillaries: Po2RBC = 60.6 ± 2.3 mm Hg; Po2InterRBC = 23 ± 1.5 mm Hg; Po2Mean = 32.7 ± 1.9 mm Hg) (Figure 1B). The standard deviation of measurements (SD) made in a given capillary, during single or multiple recording sessions were modest for Po2InterRBC (mean SD = 2.6 ± 0.3 mm Hg) and slightly larger for Po2RBC (mean SD = 5.4 ± 0.5 mm Hg). Figure 1C illustrates that average Po2 values masked the large span of all values measured. This was particularly true for Po2RBC, for which values frequently exceeded 70 mm Hg. Overall, these data show that in the glomerular layer of the awake resting mouse, interstitial Po2 ranges from 15 to 35 mm Hg in about 82% of our measurements. We then investigated whether the range and fluctuations of Po2values depend on two vascular parameters, RBC blood flow and hematocrit, in the same capillaries.

Erythrocyte-associated transients (EATs) in the olfactory bulb glomerular layer of the awake mouse.

(A) Left panel, schematic diagram of the organisation of the olfactory bulb. OSN: olfactory sensory neuron, PG: periglomerular neuron, M: mitral cell, ONL: olfactory nerve layer, GL: glomerular layer, EPL: external plexiform layer, MCL: mitral cell layer. Right panel, top, schematic illustrating the 2PLM Po2 measurement procedure in capillaries. Bottom, diagram showing Po2 values extracted from EATs. The continuous trace represents the Po2 profile relative to the RBC border in one selected capillary: Po2 at the RBC border (Po2RBC, in this case 47.2 mm Hg), Po2 at distance from a RBC (Po2InterRBC, in this case 8.6 mm Hg) which gives an estimate of Po2 in the interstitium of the glomerular layer, and average Po2 in the capillary (Po2Mean, in this case 19 mm Hg). (B) Multiple (~4–8) measurements were made in each capillary. The Po2InterRBC is significantly lower than both Po2Mean and Po2RBC, whether calculated on all measurements (the mean value is the average of all measurements pooled from all capillaries assessed, n = 262, left panel), or on the mean values from each capillary (the mean value is the average of the single mean values for each of the capillaries, n=38, right panel). SD (the average of the SD values for each capillary, presented as mean standard deviation ± SEM) illustrates the fluctuations of Po2 values in each capillary across the multiple measurements. Data presented as mean ± SEM. *p<0.05, ***p<0.001, Kruskal-Wallis test with 2-tailed Dunn's multiple comparison post-hoc test. (C) Frequency distributions of all measurements of Po2RBC (left panel), Po2InterRBC (middle panel), and Po2Mean (right panel). 5 mm Hg bin width. For all plots n = 5 mice.

https://doi.org/10.7554/eLife.12024.003

Capillary blood flow and hematocrit in the olfactory bulb of the awake resting mouse

Mean capillary RBC flow and hematocrit values were 30.6 ± 2.6 cells/s and 34.6 ± 1.8%, respectively (Figure 2A). Both RBC flow and hematocrit displayed a wide range of values (Figure 2B) with a positively skewed frequency distribution of RBC flow. Simultaneous measurements of Po2 and blood flow parameters revealed that although interstitial Po2 (Po2InterRBC) is correlated with both RBC flow and hematocrit (r = 0.3091, p<0.0001 and r = 0.4365, p<0.0001, Spearman r correlation, respectively), these relationships are non-linear. In particular, Po2InterRBC is mostly independent of both RBC flow below 60 cells/s and hematocrit from 20 to 50% (Figure 2C and E), increasing only at high values of both parameters. This stable region covers 82.4% of our measurements. In contrast, Po2RBC increased with RBC flow and hematocrit at low values, becoming stable above 20 cells/s and 30%, respectively (Figure 2C and E). Note that Po2Mean increased with both RBC flow and hematocrit values (RBC flow: r = 0.5116, p<0.0001; hematocrit: 0.6752, p<0.0001, Spearman r correlation, Figure 2D and F).

Relationships of capillary blood flow and hematocrit to Po2 values, in the olfactory bulb glomerular layer of the awake mouse.

(A) RBC flow and hematocrit calculated from the mean values of each capillary (n = 38). SD (the average of the SD values for each capillary, presented as mean standard deviation ± SEM) illustrates the fluctuations of RBC flow and hematocrit in each capillary. (B) Frequency distributions of RBC flow and hematocrit (5% bin). (C) Distribution of all Po2RBC and Po2InterRBC measurements as a function of RBC flow. Note that the Po2InterRBC is independent of RBC flow below 60 cells/s while Po2RBC increases with RBC flow below 40 cells/s. (D) Po2Mean as a function of RBC flow. (E) Distribution of Po2RBC and Po2InterRBC as a function of hematocrit. Po2InterRBC is independent of hematocrit from 20 to 50%. Po2RBC increases with hematocrit at low values. (F) Po2Mean as a function of hematocrit. Bar graph data presented as mean ± SEM. *p<0.05, **p<0.01, ***p<0.001. Kruskal-Wallis test with 2-tailed Dunn's multiple comparison post-hoc test. For all plots n = 5 mice.

https://doi.org/10.7554/eLife.12024.004

Brain oxygenation is greatly enhanced by isoflurane anesthesia

Isoflurane is a volatile anesthetic that is commonly used in the study of brain activation and metabolism. It differently affects regional cerebral blood flow in humans (Ramani and Wardhan, 2008), and modulates neurovascular coupling in a concentration-dependent fashion (Masamoto et al., 2009) as well as the relationship between spontaneous or evoked neuronal activity with BOLD signal (Aksenov et al., 2015). However, its effects on brain oxygenation have only been investigated using approaches with low spatial resolution and which did not allow simultaneous measurement of blood flow (Liu et al., 1995; Ortiz-Prado et al., 2010). We performed paired measurements of EATs and flow parameters in a set of capillaries, both when the animals were awake and when they were anesthetized with isoflurane (~0.75% as measured at the animal’s snout, delivered in air with no supplementary O2). As can be seen from Figure 3A, isoflurane significantly increases all capillary Po2 values (Po2Mean, Po2RBC and Po2InterRBC). This effect was present in all but one of the capillaries tested. Although reduced neuronal activity (and hence O2 consumption) in the isoflurane anesthetized state (Aksenov et al., 2015) is likely to play a role, it appears that this elevation of Po2 largely resulted from an increase in RBC flow, as the increase in Po2 values observed is in accordance with that which would be predicted from the observed increase in RBC flow based on the relationship presented in Figure 2C and D. This increase in capillary RBC flow rate is related to isoflurane’s vasodilatory effects on large vessels (Figure 3B) (Koenig et al., 1994). Consequently, cellular processes, including neuronal activity in response to odor, occur at a much higher oxygen concentration during isoflurane anesthesia than in the awake, resting state, and this difference is likely to be exacerbated when isoflurane is delivered in gas mixtures where [O2] is greater than 21% (as in many other studies). Since the olfactory bulb glomerular layer has a specific neuronal and vascular organisation that could be a main determinant of Po2 values, we extended our investigation of tissue oxygenation to the cerebral cortex, due to its importance to higher cognitive functions and its use as an ischemic model.

Figure 3 with 1 supplement see all
Isoflurane changes the brain oxygenation state.

Po2 and RBC flow were compared in the same sets of olfactory bulb glomerular layer capillaries when the animal was awake, and when the animal was anesthetized with isoflurane (0.75%, delivered in air, no oxygen added). (A) Isoflurane anesthesia increased all RBC flow and Po2 values as compared to the awake state (n = 142 measurements in each condition, 16 capillaries from 3 mice). Bar graph data presented as mean ± SEM., ***p<0.001, paired 2-tailed Wilcoxon signed rank test. (B) Isoflurane anesthesia induced a large dilation of pial vessels (a: artery, v: vein). See also Figure 3—figure supplement 1.

https://doi.org/10.7554/eLife.12024.005

Capillary flow and Po2 values in the somatosensory cortex of the awake resting mouse

We made measurements of all Po2 and blood flow parameters in capillaries distributed from the cortical surface to a depth of 410 µm in the fore- and hind-limb regions of the somatosensory cortex. Due to the greater inter-capillary distance in the cortex than in the glomeruli (Lecoq et al., 2009; Sakadžić et al., 2014), and as Po2 gradients have previously been observed around blood vessels in anesthetized animals (Devor et al., 2011; Sakadzić et al., 2010; Sharan et al., 2008), it is likely that there are tissue regions, in areas far from the nearest capillary, in which the Po2 is lower than what is reported by Po2InterRBC. Nonetheless, based on our previous results (Parpaleix et al., 2013) in olfactory bulb glomeruli and on theoretical predictions (Lücker et al., 2014), we expect that Po2InterRBC reports steady-state tissue Po2 up to a radius of approximately 10 µm from a capillary. Furthermore, recent work, that models hematocrit distribution in large microcirculatory networks and accurately replicates physiological RBC distribution, predicts that the distribution of oxygen in tissue volumes supplied by these networks is largely homogeneous (Gould and Linninger, 2015). We thus consider that our measured values of Po2InterRBC represent a significant proportion of the range of Po2 values present in the cortical interstitium. Our Po2 measurements revealed that the averages and ranges of Po2 values in cortical capillaries (Figure 4A and B) were similar to those measured in the olfactory bulb glomerular layer (Po2RBC = 66.3 ± 1.6 mm Hg, Po2InterRBC = 23.3 ± 1.1 mm Hg, and Po2Mean = 36.3 ± 1.3 mm Hg), indicating that at rest in the awake state, these two brain areas have similar levels of both RBC, capillary and pericapillary oxygenation. In contrast, the average RBC flow and hematocrit values (41.9 ± 1.8 cells/s and 47.3 ± 0.9%, respectively) were significantly higher than those found in the olfactory bulb glomerular layer (Figure 4B) with the hematocrit values being higher than previously reported levels in the cerebral cortex of anesthetized animals (Hudetz, 1997). The lower hematocrit levels in glomerular layer capillaries could be related to differences in the bulb cerebrovascular supply which, in contrast to the cortex (Blinder et al., 2013), is poorly established (Coyle, 1975).

The relationship of Po2 to RBC flow and hematocrit in the somatosensory cortex of the awake mouse.

(A) Average values and SD of Po2 parameters in somatosensory cortex capillaries, calculated from the mean values from each capillary (81 capillaries, 528 measurements, SD = the average of the SD values for each capillary, presented as mean standard deviation ± SEM). (B) Distribution of all capillary Po2 values averaged in (A). Frequency distribution histogram of local tissue Po2 (Po2InterRBC) in inset. 5 mm Hg bin. (C) Average values and SD of RBC flow and hematocrit calculated from the mean values from each capillary. (D) Distribution of all Po2RBC and Po2InterRBC measurements as a function of RBC flow. Note that for most values (from 20 to 60 cells/s), both Po2InterRBC and Po2RBC increase with RBC flow. (E) Po2Mean as a function of RBC flow. (F) Distribution of all Po2RBC and Po2InterRBC measurements as a function of hematocrit. For most values (from 20 to 60% ), both Po2InterRBC and Po2RBC increase with hematocrit. (G) Po2Mean as a function of hematocrit. Bar graph data presented as Mean ± SEM. *p<0.05, **p<0.01, ***p<0.001. Kruskal-Wallis test with 2-tailed Dunn's multiple comparison post-hoc test.

https://doi.org/10.7554/eLife.12024.007

As in the olfactory bulb, both Po2RBC and Po2InterRBC were correlated with RBC flow rate and hematocrit (Po2RBC with RBC Flow and hematocrit: r = 0.3824, p<0.0001 and r = 0.2315, p<0.0001, Spearman r correlation, respectively; Po2InterRBC with RBC Flow and hematocrit: r = 0.4283, p<0.0001 and r = 0.5018, p<0.0001, Spearman r correlation, respectively). Po2InterRBC (Figure 4D and F) increased with RBC flow, from 20 to 60 cells/s, that is, over the majority of the measurements, becoming stable thereafter. It also increased with hematocrit. Similarly, Po2RBC increased with both RBC flow and hematocrit. Note that Po2Mean, was correlated with and increased through the entire range with both RBC flow and hematocrit (RBC flow: r = 0.4670, p<0.0001; hematocrit: 0.5945, p<0.0001, Spearman r correlation, Figure 4E and G).

Laminar organisation of capillary blood flow and Po2 values in the superficial layers of the somatosensory cortex

Several studies, using polarographic electrodes or 2PLM in anesthetized rodents, have reported that vascular and interstitial Po2 varies with cortical depth (Devor et al., 2011; Masamoto et al., 2003; Sakadzić et al., 2010). As anesthetics could differently modulate synaptic activity and blood flow in layers I to IV, it is difficult to predict the extent to which Po2 depth variations occur in the awake animal. We thus first measured Po2 in descending and ascending large vessels and then compared Po2and blood flow parameters of capillaries from layers I to IV (from the surface to 410 µm in depth). Capillaries less than 60 μm below the surface were considered to be in layer I, those from 90 to 260 μm below the surface were classified as layer II/III capillaries and those deeper than 340 μm were considered layer IV capillaries (capillaries from 260 to 340 µm were not considered due the ambiguity of their location). Penetrating vessels (arterioles and venules) were traced from their point of descent below the surface down along their main trunk until they ramified into smaller vessels (or descended below 410 µm, which was the maximum depth at which we made measurements). Note that in these vessels EATs were not detectable due to the close apposition of RBCs. The mean Po2 was 69.2 ± 1.4 mm Hg for arterioles and 39.9 ± 1.3 mm Hg for venules and, in contrast to what was reported in anesthetized animals (Devor et al., 2011; Sakadzić et al., 2010), no gradient was observed with depth (Figure 5A and B).

Depth profiles of vascular and local tissue oxygenation in the somatosensory cortex of the awake mouse.

(A) Left panel, maximum intensity projection of superficial portion of the vasculature of the somatosensory cortex, with boxed area highlighting a penetrating arteriole shown in the XZ projection (right panel). (B) Left panel, Po2 values in penetrating arterioles (red) and venules (blue) as a function of depth from the cortical surface. Each line represents a single vessel. Right panel, mean of all Po2 values recorded from vessels as a function of depth. Note the absence of Po2 gradients with depth (50 µm bin size, 136 measurements in 11 arterioles, 148 measurements in 14 venules, from 6 mice.) Data presented as mean ± SD (C) Comparison of Po2 values (left panel), RBC flow and hematocrit (right panel) in layers I, II/III and IV. Note that capillary Po2Mean and Po2InterRBC are higher in layers II/III and IV than in layer I, although there are no significant differences in either blood flow parameter. Data presented as mean ± SEM. * p<0.05, **p<0.01, Kruskal-Wallis test with 2-tailed Dunn's multiple comparison post-hoc test. n = Layer I: 113 measurements in 17 capillaries, Layer II/III: 230 measurements in 41 capillaries, Layer IV: 151 measurements in 15 capillaries (D) Po2RBC and Po2InterRBC as a function of hematocrit in layers I, II/III and IV.

https://doi.org/10.7554/eLife.12024.008

Laminar analysis of capillary blood flow and Po2 values revealed some specific differences: although the hematocrit and blood flow values were similar in all three layers (Figure 5C, right panel), all Po2 values were lower in layer I than in layer II/III (Figure 5C, left panel). The low Po2InterRBC values 14.7 ± 1.7 mm Hg suggests that interstitial Po2 is correlated with the capillary density which is lower in layer I than II/III (Blinder et al., 2013; Sakadžić et al., 2014). The correlated analysis of Po2 with hematocrit revealed that low Po2InterRBC values were present at a wide distribution of hematocrit levels (Figure 5D, left panel). In addition, even though the average Po2RBC was higher in layer II/III than in layer IV, it was independent of hematocrit in both layers (Figure 5D). Finally, the effects of isoflurane in the cortex (Layer I, Figure 3—figure supplement 1) were similar to those observed in the olfactory bulb. Thus, in addition to its direct effects on neurons, isoflurane increases oxygen delivery to the entire brain.

Regions of low tissue Po2 are present at rest

Po2InterRBC values show a wide distribution in both the olfactory bulb glomerular layer and the somatosensory cortex (Figure 1C and 4B). Notable in both structures is the presence of measurements of Po2InterRBC, and hence local tissue Po2, that were <15 mm Hg. In the olfactory bulb glomerular layer these low Po2InterRBC capillaries (n=5) were found to have a large range of RBC flow rates but low hematocrit levels (<25%, Figure 6A and 6B). In contrast, in somatosensory cortex capillaries (n = 13) in which the Po2InterRBC value was <15 mm Hg, neither RBC flow rates nor hematocrit levels were notably low (Figure 6C and 6D). Instead, these low Po2InterRBC values were mostly found in capillaries in layer I (10 of 13 capillaries), suggesting that the presence of low tissue Po2 values results from several factors.

In the awake mouse, low interstitial Po2 in the olfactory bulb glomerular layer is associated with low hematocrit capillaries.

Po2 values from all glomerular layer capillaries with average Po2InterRBC values of <15 mm Hg (n = 24 measurements in capillaries) are plotted as a function of RBC flow (A) and hematocrit (B). The RBC flow rates in these capillaries were distributed across a wide range, whereas in all cases capillary average hematocrit was <25%, suggesting that, in the glomerular layer, areas of low interstitial Po2 are supplied by capillaries with relatively low hematocrit values (hematocrit of these capillaries = 18.6 ± 1.4%, n = 5; hematocrit of all other capillaries = 37 ± 1.7%, n = 33; p = 0.002, unpaired t-test. mean ± SEM). Conversely, cortical capillaries with average Po2InterRBC <15 mm Hg (n = 13 capillaries, 101 measurements) had wide ranges of both RBC flow (C) and hematocrit (D). However, the majority (10 of 13) of these capillaries were located in layer I. In all plots, single measurement values and mean ± SD of all measurements in each capillary shown.

https://doi.org/10.7554/eLife.12024.009

RBC hemoglobin is largely saturated in cerebral capillaries

Given that Po2RBC should be representative of the Po2 level inside RBCs, we used known values of the Hill coefficient and P50 for mouse hemoglobin (Uchida et al., 1998) to estimate hemoglobin saturation (So2) in capillaries. In both the olfactory bulb glomerular layer and the somatosensory cortex (Figure 7A and B), measured So2 is >50% in the vast majority of cases (94.7% and 98.5% of measurements in the glomerular layer and the cortex, respectively). This shows that in the resting brain, the majority of hemoglobin exists as oxyhemoglobin. Furthermore the presence of numerous measurements of So2 >85% in both structures suggests that hemoglobin saturation is nearly maximal in a significant proportion of capillaries (~10% of capillaries in the glomerular layer, ~20% in the somatosensory cortex). Note that, if So2 were to be inappropriately estimated from Po2Mean values, a very different distribution of hemoglobin saturation would be derived (Figure 7C). This demonstrates that in capillaries, an accurate measurement of Po2RBC is a prerequisite for estimating So2.

In the awake mouse, the majority of hemoglobin in cerebral capillaries is oxygenated.

(A) Left panel, frequency distribution of Po2RBC values measured in the glomerular layer of awake mice, which were used to compute So2 values (right panel). (B) Equivalent Po2RBC (left panel) and So2 distributions (right panel) from the cerebral cortex. In both structures > 90% of the measurements have So2 values of >50% (94.7% and 98.5% of measurements in the glomerular layer and the cortex respectively) with ~60% of So2 values in the cortex being >75%. (C) Left panel, frequency distribution of Po2Mean values from cerebral cortex capillaries, with the So2 distribution that would be computed were these lower Po2 values considered to represent those present at the hemoglobin molecules in RBCs (right panel). Glomerular layer: n = 262 measurements in 38 capillaries. Somatosensory cortex: n = 528 measurements from 81 capillaries. Bin size 5 mm Hg and 5% for Po2 and So2 respectively.

https://doi.org/10.7554/eLife.12024.010

Discussion

In the olfactory bulb and somatosensory cortex of awake resting mice, we report similar average values of interstitial Po2 (~23 mm Hg). In the awake rodent, brain oxygenation has previously only been investigated with lower resolution approaches. Liu et al. (Liu et al., 1995) used electron paramagnetic resonance (EPR) oximetry with a lithium phthalocyanine crystal being implanted in the cerebral cortex of rats 24 hr prior to the measurements. They reported that restrained and untrained rats have a cerebral Po2 of about 34 mm Hg, which, surprisingly, was reduced by isoflurane anesthesia (1% in 21% O2) to about 24 mm Hg. This resting value (34 mm Hg) is significantly higher than our average interstitial Po2 value, most probably due to the effects of the stress resulting from restraint in the absence of habituation. The effect of isoflurane, which decreased Po2 is intriguing, and indicates that the main effect of isoflurane in this case was to abolish the Po2 augmenting effects of stress. Interestingly, increasing the delay from crystal implantation to 7 days and introducing some habituation to restraint reduced cerebral Po2 to 27 mm Hg (Dunn et al., 2000). The best demonstration of physiological measurements of Po2 has been performed in unrestrained rats, in which brain Po2was measured with an implanted fiber optic probe measuring quenching of an oxygen sensor (Ortiz-Prado et al., 2010). Although the probe was very invasive, it reported an average bulk tissue Po2 value of 25 mm Hg, a value very similar to ours. We are thus confident that Po2InterRBC is an excellent reporter of the interstitial Po2, that our extensive habituation procedures are efficient, and that this study gives the first non-invasive and physiological values of Po2, hematocrit and capillary blood flow in the rodent brain.

Our results show a number of differences from those previously reported with high-resolution measurement techniques in anesthetized animals. Tissue Po2 values have previously been reported to be in the range of ~5–100 mm Hg, with the higher values occurring in regions close to pial arterioles (Devor et al., 2011; Sakadzić et al., 2010). Our values of tissue Po2 (Po2InterRBC) lie in the lower end of this wide range, do not include measurements from the capillary-free peri-arteriolar regions, and so are likely to be reflective of the Po2 levels that exist in the majority of the tissue, away from large arterioles. Considering intravascular Po2, our measurements of capillary Po2Mean (average ≈36 mm Hg in the cortex) seem to be broadly in agreement with those reported by Sakadzić et al., 2010 (~25–35 mm Hg), but our recorded values of capillary RBC Po2 (Po2RBC, average ≈66 mm Hg in the cortex), are dramatically higher than those published by Sakadzić et al., 2014, where the most frequent values were ~25–35 mm Hg. A prominent discrepancy between the results of previous studies and our findings in awake, resting animals relates to changes in cortical oxygenation with depth. Many previous studies report drops in interstitial Po2 with increasing depth in the cortex (Cross and Silver, 1962; Devor et al., 2011; Masamoto et al., 2003; Nair et al., 1975; Ndubuizu and LaManna, 2007; Whalen et al., 1970), with Po2 in layer I being higher than in underlying layers. In contrast, our measures of Po2InterRBC suggest that interstitial Po2 is lower in layer I than in either layer II/III or IV. These conflicting patterns of laminar variations of Po2 also exist in comparisons of the mean vascular Po2 values (Po2Mean), where previous 2PLM studies have reported decreases in mean vascular Po2 in concert with decreases in penetrating arteriole and venule Po2 (Devor et al., 2011; Sakadzić et al., 2010) with increasing depth in the cortex. In the present study, we see an increase in capillary Po2Mean from layer I to deeper layers, and no gradient in penetrating vessel Po2 with depth. It is possible that the deviation from normal physiological conditions inherent in anesthesia and acute surgical preparation, which we avoid with our approach, leads to the emergence of these disagreements.

Although average Po2InterRBC values in both the cortex and glomerular layer were ~23 mm Hg, in both structures a number of capillaries were found to have Po2InterRBC values <15 mm Hg, indicating the existence of regions in both structures where the interstitial Po2 is close to reported values of Po2 below which cellular respiratory rate is strongly dependent on Po2 (~10 mm Hg) (Kasischke et al., 2011). Furthermore, examples were found in both brain regions where the Po2InterRBC was below the value of ~3.4 mm Hg that has previously been reported as the critical Po2 in brain tissue (Kasischke et al., 2011). Similarly low tissue Po2 measurements have previously been observed (Ndubuizu and LaManna, 2007; Whalen et al., 1973), but their existence has been interpreted as being related to disruption of normal tissue physiology in the experimental preparations (Wilson et al., 2012). However, the presence of such low values in our awake preparations indicates that, surprisingly, at least some small regions of the brain can subsist at very low Po2 values.

Our finding of high Po2RBC and corresponding capillary So2 (estimated using the parameters used in Sakadžić et al., 2014) values in both the olfactory bulb and the somatosensory cortex (Figure 7) differs from recently published findings in the anesthetized, ventilated mouse (Sakadžić et al., 2014). We find that capillary So2 in both structures is generally high, indicating that the majority of hemoglobin in these vessels exists as oxyhemoglobin, and thus that capillaries are capable of supplying very significant quantities of oxygen to support neural function.

In conclusion, the present study establishes, for the first time, accurate and precise values of physiological Po2 in the vasculature and interstitium of mouse cerebral grey matter. As it is known that O2 concentration is a critical parameter in determining the properties of neuronal function (Huchzermeyer et al., 2008, 2013; Ivanov and Zilberter, 2011), and neurovascular interactions (Gordon et al., 2008), these values provide a standard on which future research can rely to provide relevant, physiologically accurate conditions of oxygenation in which to investigate such processes.

Materials and methods

Experimental procedures

Animal preparation and surgery

All animal care and experimentation was performed in accordance with the INSERM Animal Care and Use Committee guidelines. Adult C57BL/6 mice (3–6 months old, 25-–35 g, both males and female, housed in 12-hr light-dark cycle) were used in this study (n = 5 mice for olfactory bulb capillaries, n = 3 mice for cortex capillaries). Chronic glass cranial windows were implanted over the area of interest, either the olfactory bulb or the somatosensory cortex, with great care taken to not disturb the dura mater. The animals received anti-inflammatory and analgesic treatment (Carprofen, one daily 0.15 mg subcutaneous injection, administered pre-surgically and for the three days post-surgery), antibiotics (Ceftrioxone, one daily 0.25 mg subcutaneous injection, administered pre-surgically and for the three days post-surgery) and dexamethasone to reduce cerebral edema (one daily 60 µg subcutaneous injection on the day before surgery, directly before the surgery and the first post-surgical day). Surgical anesthesia was induced with ketamine-xylazine (100 mg and 8 mg per kg body mass, respectively). During surgery, the mice breathed a mixture of air and supplementary oxygen (the final inhaled proportion of oxygen was ~30%) and the body temperature monitored by a rectal probe and maintained at ~36.5°C by a feedback-controlled heating pad. A craniotomy was performed with great care taken not to apply great pressure to the bone and the area was regularly flushed with cool aqueous buffer solution to avoid damage or heating of the underlying tissue. Cover glass (~150 µm thick) was used for the window and sealed in place with photopolymerizable dental cement (Tetric EvoFlow, Ivoclar Vivadent, Schaan, Principality of Liechtenstein), which was also used to form a head-cap in which a titanium head-bar was also embedded.

Habituation of mice to head restraint

In the weeks preceding surgery, the animals were supplied with a treadmill (Fast-Trac, Bio-Serv, Flemington, New Jersey) in their cages that was similar to that which forms part of the restraint apparatus used during 2PLM recording (see below). In the days that preceded the surgery, the mice were gently habituated to handling, and provided with treats (sugar pellets, Bio-Serv). 2–3 days after the surgery restraint habituation began. The frame used for head restraint during the habituation and imaging consists of a metal frame in which the mouse’s head-bar is secured, and a treadmill wheel (similar to that in the animal’s home cage). This configuration allows the animal limbs and body to move freely while the head is stably fixed. All restraint-habituation sessions were carried out in the set-up that was used for imaging, with the animal kept in the near-dark at all times. Habituation sessions were performed multiple times per day over the course of 2–4 weeks, with the duration increasing from 5 min to >1 hr. Animals were considered ready for use in experiments when they could be easily fixed in the recording apparatus while awake, and their behavior during the sessions consisted of short bouts of locomotion (~30 s) separated by longer periods of stillness (5–15 min) during which measurements were performed.

On the days on which recordings were taken, the animals were briefly anesthetized with isoflurane (2%, in air. Total duration of anesthesia <5 min), and PtP-C343 solution (Mw≈65 kDa, 2.5 mM in 0.9% saline) was introduced intravenously via retro-orbital injection. Fluorescein isothiocyanate dextran (Mw = 150 kDa) was often co-administered with PtP-C343 to enhance imaging contrast. The final PtP-C343 concentration in the plasma is estimated to be ~100 µM. After the injection, the mice were returned to their home cages for ~90 min, before being brought back to the experimental room, for ~30 min before the start of the recording session. We implemented this delay of ~2 hr between the injection and the following recording session to avoid potential confounding effects of the brief exposure to isoflurane or the effects of the injection itself. We found that the concentration of PtP-C343 in the blood typically remained sufficient for Po2 measurements to be made over 1–3 days after an injection, allowing for multiple recording sessions to be made before reinjection was necessary. Each recording session lasted from 1–3 hr, depending on the type of measurements performed and the behavior of the animal, and each animal underwent 1–3 recording sessions per day over the course of 2–7 days, with breaks of at least 1–3 hr between each session. The data derived from each of the recording sessions across the 1–2 days post-injection was similar, indicating that no lingering effects of the injection procedure or brief isoflurane anesthesia had affected the measurements in the first recording session post-injection.

2PLM setup and procedure

The 2PLM setup is as described in Lecoq et al. (2011), and Parpaleix et al. (2013). Briefly, the output from a Ti:sapphire laser (Mira, Coherent, Santa Clara, California; λ = 850 nm, 120 fs pulse width, 76 mHz) is gated by an acousto-optic modulator (AOM, AA Optoelectronic, Orsay, France; MT110-B50-A1.5-IR-Hk). This allows for repetitive cycles of alternating 'on' and 'off' periods to be generated, that correspond to periods of excitation and recording from the PtP-C343 probe in the sample. The excitation period was 24 µs and the recording period was 225 µs, for a total cycle period of 250 µs and a repetition rate of 4 kHz. The scanning of the excitation light by the galvanometric mirrors (Cambridge Technology, Bedford, Massachusetts) is synchronized with the gating of the laser output by the AOM, using custom-built electronics and LabVIEW software (National Instruments, Austin, Texas). The excitation light was focused with a water-immersion objective lens (either x63 Leica [Wetzlar, Germany] or x40 Olympus lens [Olympus, Tokyo, Japan]).

Emitted photons were divided by a dichroic mirror (cut off wavelength = 560 nm). The green channel was bandpass filtered (HQ 520/40 m, Chroma Technology Corp, Bellows Falls, Vermont). The red channel had two shortpass filters (FF01-750/sp-25, Semrock, Rochester, New York), and one bandpass filter (HQ 680/60 m-2P, Chroma Technologies Corp), and a red-sensitive photomultiplier tube (PMT, R10699, Hamamatsu, Hamamatsu City, Japan). PMT signals were amplified with custom-build electronics and sampled at 1.25 MHz by an acquisition card. Photons detected in the green channel during the on phase of the recording cycle were used to detect the borders of RBCs and extract RBC flow rates and hematocrit values (see below). Phosphorescence decays detected during the 'off' period were averaged over a number of cycles (~5000–50,000 decays) and the lifetime of the phosphorescence determined by fitting a single exponential curve to the data. The first 6–7 bins (~5 µs) after the end of the on-phase were discarded. This lifetime measurement is then converted to a value of Po2 using a calibration curve. We used the nonlinear least-squares method and the Marquardt-Levenberg algorithm to obtain the decay lifetimes and associated standard errors. We used a Stern-Volmer–like calibration curve to convert the phosphorescence lifetimes into the corresponding Po2 values. This analysis was performed using custom build software (see below) developed using LabView (National Instruments).

Capillary blood flow analysis and EAT properties extraction

Initially, fast 3D image acquisition was performed to identify and target vessels of interest. Po2 measurements were then made by focusing the excitation point in blood vessels of interest, and recording the fluorescence and phosphorescence emitted from 10,000–40,000 cycles of excitation and detection of phosphorescence decays (2.5–10 s of measurement) at each point. Typically 4–8 measurements were performed in each capillary, with the total number of decays recorded for each capillary being around 60,000–100,000.

Details of the methods of determining EAT properties, RBC flow rates and hematocrit can be found in Parpaleix et al. (2013) and Lecoq et al. (2011), see especially Figures 1 and 3 therein). In brief, the fluorescence recorded during the excitation period of the cycle was analyzed and the passage of RBCs through the excitation point detected based on the associated dips in fluorescence intensity. These transient reductions in fluorescence intensity were detected and quantified using a binary threshold method, with the borders of the RBC corresponding to the full width at half maximum of the fluorescence dip. This allowed for the measurement of both capillary RBC flow rate (the number of RBCs detected per unit time) and hematocrit (the number of decay cycles originating from within RBCs as a percentage of the total number of decay cycles). The borders of the detected RBCs were used as time-stamps, and the phosphorescence decays recorded were binned according to their distance in time from the borders of the nearest RBC. The Po2RBC value was calculated from the average lifetime of decays recorded in the 1–4 ms around the border of the RBC, whereas the Po2InterRBC was determined from average value of decays at mid-distance between RBCs (averaged over a window of at least 5 ms). The EAT-extraction custom-built software is now available online at the following address: https://github.com/charpak-lab/EAT-detection. Hemoglobin saturation (So2) was estimated from Po2RBC using the Hill equation, with a Hill coefficient (2.59) and P50 (40.2 mm Hg), which are accurate for C57BL/6 mice (Uchida et al., 1998) and have previously been used to make estimations of mouse cerebral So2 from Po2 data (Sakadžić et al., 2014).

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
    Critical state of energy metabolism in brain slices: the principal role of oxygen delivery and energy substrates in shaping neuronal activity
    1. A Ivanov
    2. Y Zilberter
    (2011)
    Frontiers in Neuroenergetics, 3, 10.3389/fnene.2011.00009.
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
    Large arteriolar component of oxygen delivery implies a safe margin of oxygen supply to cerebral tissue
    1. S Sakadžić
    2. ET Mandeville
    3. L Gagnon
    4. JJ Musacchia
    5. MA Yaseen
    6. MA Yucel
    7. J Lefebvre
    8. F Lesage
    9. AM Dale
    10. K Eikermann-Haerter
    11. C Ayata
    12. VJ Srinivasan
    13. EH Lo
    14. A Devor
    15. DA Boas
    (2014)
    Nature Communications, 5, 10.1038/ncomms6734.
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38

Decision letter

  1. David Kleinfeld
    Reviewing Editor; University of California, San Diego, United States

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your work entitled "Mapping oxygen concentration in the awake mouse brain" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by David Kleinfeld as the Reviewing Editor and Timothy Behrens as the Senior Editor.

The reviewers have discussed the reviews with one another and the Reviewing editor has drafted this decision to help you prepare a revised submission. All reviewers are enthusiastic about this work. Each raised a number of questions. I ask you address each question by clarifying the manuscript. No new data is required. I have summarized the key questions:

1) The hematocrit is reported to be significantly higher in cortex than in olfactory bulb (subsection “Capillary flow and Po2 values in the somatosensory cortex of the awake resting mouse”). This needs to be discussed – how is this difference produced? Is it somehow produced by plasma skimming at a series of junctions of the vessels supplying the olfactory bulb? As this is an acute preparation, one needs to be certain that this is not due to some small amount of tissue damage when making the imaging window.

2) Please show a raw trace of the Po2 profile between RBCs, or state explicitly that the line in the inset to panel A IS raw data (and not a schematic); why is that line continuous to the left and dotted on the right hand side?

3) The authors repeatedly stress that previous measurements using two-photon phosphorescence only report Mean Po2. Could the authors please include a comparison of their Mean Po2 values against those made by other groups also using two-photon phosphorescence probe?

4) Please explain the bars in Figure 1B far better. State completely clearly that Po2 is the mean (over vessels) of a single time averaged value obtained from each vessel, while SD is mean (over vessels) of the standard deviation of the Po2 in each vessel over time (if that is what it is!).

5) The Po2 seems (subsection “Laminar organisation of capillary blood flow and Po2 values in the superficial layers of the somatosensory cortex”) to be determined by the local capillary density. Does this imply that local O2 consumption is independent of laminar position (one might have expected vessel density to be proportional to local O2 usage)? Is it possible to derive, from the laminar distribution of [O2] and the blood flow reaching each lamina, how much O2 is consumed by each cell layer?

6) A more thorough clarification of the argument of how isoflurane increases interstitial oxygen concentration would strengthen the manuscript. The current explanation links increased bloodstream oxygenation to the vasodilatory effect of isoflurane (subsection “Brain oxygenation is greatly enhanced by isoflurane anesthesia”). However, the increase in RBC flux (Figure 3A) resulting from vasodilation would lead to a decrease in oxygen unloading from RBCs, as evidenced by the increase in RBC Po2 (Figure 3A). If less oxygen is unloaded from the red blood cells, then less oxygen diffuses through the vessel wall, leading to a decrease in tissue oxygenation. This is contradicted by the statement that cellular processes occur at a higher oxygen concentration during isoflurane anaesthesia. This matter is further confused by the delivery of isoflurane in a 21% oxygen gas mixture.

7) Theoretical models are cited as evidence that InterRBC Po2 is in equilibrium with the surrounding tissue (subsection “Capillary flow and Po2 values in the somatosensory cortex of the awake resting mouse”). The theoretical models rely on accurate hematocrit prediction to determine the distribution of oxygen to the surrounding tissue. The relationship between radial oxygen gradients around capillaries, hematocrit and RBC velocities have been discussed in (I. Gould, A. Linninger, "Hematocrit Distribution and Tissue Oxygenation in Large Microcirculatory Networks", Microcirculation, 22, p1-18, 2015). The authors may consider inclusion of this reference to support the argument.

8) The authors mention that they anaesthetise all mice before initiating even awake imaging experiments. Is it possible to head fix the mice without using anaesthesia? Is the anaesthesia necessary to do the retro-orbital injection of the fluorescein dextran and oxygen probe? It seems awake mice could be tail/vein injected. This might also affect the results since there could residual effects of anaesthesia. At the very least this confound should be acknowledged.

The reviewers have also identified some technical and typographic errors that you should attend to. See the full reviews attached below. We look forward to a revised submission together with a tally of brief answers to each of the seven points above. This is beautiful work and we hope to see it published soon.

Reviewer #1:

This manuscript provides important data (obtained using a relatively non-invasive technique) characterising the O2 level at different parts of the vasculature in awake non-anesthetized animals. Although mainly descriptive, the data will be of great interest to all those working on the control of cerebral blood flow, energy supply to the brain, brain metabolism and BOLD fMRI signals, because all of these subjects depend critically on the local O2 concentration, which is poorly defined in vivo. Previous work has only provided similar data on anesthetized animals (or with more invasive techniques in freely moving animals), and it turns out, importantly, that the values are very different in the unanesthetized animal. This paper is therefore a significant step forward, and I support its publication.

There are some areas where the discussion should be improved, as follows.

1) The hematocrit is reported to be significantly higher in cortex than in olfactory bulb (subsection “Capillary flow and Po2 values in the somatosensory cortex of the awake resting mouse”). This needs to be discussed – how is this difference produced? Is it somehow produced by plasma skimming at a series of junctions of the vessels supplying the olfactory bulb? Why might that occur more for the supply to the olfactory bulb than for that to the cortex?

2) The Po2 seems (subsection “Laminar organisation of capillary blood flow and Po2 values in the superficial layers of the somatosensory cortex”) to be determined by the local capillary density. Does this imply that local O2 consumption is independent of laminar position (one might have expected vessel density to be proportional to local O2 usage)? Is it possible to derive, from the laminar distribution of [O2] and the blood flow reaching each lamina, how much O2 is consumed by each cell layer?

3) It is surprising that this study finds a lower Po2 in the surface layers of the cortex, whereas Devor and Sakadzic found a decrease in [O2] with depth. Is there any methodological difference that could account for this, e.g. O2 penetration into the brain from the cortical window in the earlier studies)? Was the laminar profile of O2 affected by isoflurane?

4) Figure 1A. Please show a raw trace of the Po2 profile between RBCs, or state explicitly that the line in the inset to panel A IS raw data (and not a schematic); why is that line continuous to the left and dotted on the right hand side?

5) Figure 1B. Please explain the bars far better. State completely clearly that Po2 is the mean (over vessels) of a single time averaged value obtained from each vessel, while SD is mean (over vessels) of the standard deviation of the Po2 in each vessel over time (if that is what it is!). Please try out your text on a naive reader so that it is comprehensible.

6) Figure 3 legend. Stress that the measurements in A are from the glomerular layer, unlike the vessel image on the right which is pial.

7) Figure 6. Add labels for Olfactory bulb and Cortex to the panels so the reader can grasp what is plotted at a glance.

8) In the first paragraph of the subsection “Laminar organisation of capillary blood flow and Po2 values in the superficial layers of the somatosensory cortex” state explicitly how deep capillaries "deeper than 410 um" could be.

Reviewer #2:

The paper reports micron scale mapping of oxygen concentration, hematocrit, and RBC flux in both awake and anaesthetized mice using the two-photon phosphorescence probe PtP-C343. Additionally, these measurements are replicated in both the primary somatosensory cortex (n=3) as well as the olfactory bulb (n=5). Measurements of erythrocyte-associated transients were measured in individual capillaries, and oxygen tension was reported in RBCs (RBC Po2), in between two red blood cells (InterRBC Po2) and the mean between these two values (Mean Po2). These measurements are report for each layer of the cortex.

This manuscript provides an impressively thorough analysis of the correlation between hematocrit, RBC flux and the three oxygen concentration measurements in the bloodstream. Additionally, the comparison between awake and anaesthetized mice clearly demonstrates the increased oxygen concentration in the bloodstream. Additionally, inclusion of a comparative study with previous measurements of bloodstream oxygen concentration in both the awake and anaesthetized mice would strengthen the argument that isoflurane increases interstitial Po2. The explanation of how isoflurane increases interstitial Po2 is left somewhat unclear.

The reviewers recommend that this manuscript be published with minor revisions.

Technical Comments:

1) A more thorough clarification of the argument of how isoflurane increases interstitial oxygen concentration would strengthen the manuscript. The current explanation links increased bloodstream oxygenation to the vasodilatory effect of isoflurane (subsection “Brain oxygenation is greatly enhanced by isoflurane anesthesia”). However, the increase in RBC flux (Figure 3A) resulting from vasodilation would lead to a decrease in oxygen unloading from RBCs, as evidenced by the increase in RBC Po2 (Figure 3A). If less oxygen is unloaded from the red blood cells, then less oxygen diffuses through the vessel wall, leading to a decrease in tissue oxygenation. This is contradicted by the statement that cellular processes occur at a higher oxygen concentration during isoflurane anaesthesia. This matter is further confused by the delivery of isoflurane in a 21% oxygen gas mixture.

2) The authors repeatedly stress the importance of differentiating between RBC Po2, InterRBC Po2, and Mean Po2 – stating that previous measurements using two-photon phosphorescence only report Mean Po2. Could the authors please include a comparison of their Mean Po2 values against those made by other groups also using two-photon phosphorescence probe?

3) Similarly, the authors cite previous low-resolution studies on the effects of isoflurane on brain tissue oxygenation, but the manuscript is lacking a side by side comparison with previous studies. An inclusion of the previous measurements would strengthen the conclusion made in this paper.

4) Theoretical models are cited as evidence that InterRBC Po2 is in equilibrium with the surrounding tissue (subsection “Capillary flow and Po2 values in the somatosensory cortex of the awake resting mouse”). The theoretical models rely on accurate hematocrit prediction to determine the distribution of oxygen to the surrounding tissue. The relationship between radial oxygen gradients around capillaries, hematocrit and RBC velocities have been discussed in (I. Gould, A. Linninger, "Hematocrit Distribution and Tissue Oxygenation in Large Microcirculatory Networks", Microcirculation, 22, p1-18, 2015). The authors may consider inclusion of this reference to support the argument.

Reviewer #3:

The authors describe methods and results using 2-photon phosphorescence lifetime mediated measurements of tissue oxygenation in vivo. The authors extend their exciting work demonstrating that local O2 tension changes over the time course of each RBC passing (up to 100 pass per second, with each making a local O2 transient). This work extends a proven strategy of using compounds with a phosphorescence lifetime that is oxygen sensitive. The approach enables probe-specific calibration of oxygen levels using the phosphorescence lifetime. The methods (O2 probe) although not developed by Charpak`s lab, is applied quite effectively here. The authors extend previous work by David Boas and make measurements of oxygenation from awake animals. The work examines different cortical layers and contrasts with olfactory bulb. The analysis and presentation of the data is strong. Novel data includes capillary specific measurements and relationships between local oxygenation and red blood cell flux. The work also is significant for examining potentially confounding effects of well-used anaesthesia such as isoflurane. In this way the current work strikes an important cautionary tone with respect to anesthesia.

1) The authors’ analysis is impressive and I don`t have any real concerns with it, since I am familiar with this literature. However, for those which have seen these phosphorescence measurements for the first time, I think it would be important to show supplementary data illustrating how the measurements are obtained. This could be the calibration curve or some raw data showing how Po2 and velocity/flux are calculated in a single experiment.

2) I would also like to see the authors give some measurements for arteriole and vein diameters with isoflurane anaesthesia vs. the control awake state, the example images point towards a robust finding.

3) The authors mention that they anaesthetise all mice before initiating even awake imaging experiments. Is it possible to head fix the mice without using anaesthesia? Is the anaesthesia necessary to do the retro-orbital injection of the fluorescein dextran and oxygen probe? It seems awake mice could be tail/vein injected. This might also affect the results since there could residual effects of anaesthesia. At the very least this confound should be acknowledged.

4) The authors should comment on how their results relate to recent work from David Boas lab which propose that significant exchange of oxygen with tissues can occur through veins.

5) The authors should also comment on whether there are changes on efficacy of O2 signal overtime. From the protocol it sounds as though this probe must have an extremely long half-life. Why this probe would not be cleared by the liver is uncertain, its pharmacokinetics should be further investigated or at least if there is information this should be discussed so others can have a better feel for how long after injection can quantification of O2 be done.

6) I thought the idea that some regions of cortex could have very low flow and oxygenation to be interesting, however this is also an acute preparation and one needs to be certain that this is not due to some small amount of tissue damage when making the imaging window.

https://doi.org/10.7554/eLife.12024.012

Author response

The reviewers have discussed the reviews with one another and the Reviewing editor has drafted this decision to help you prepare a revised submission. All reviewers are enthusiastic about this work. Each raised a number of questions. I ask you address each question by clarifying the manuscript. No new data is required. I have summarized the key questions:

1) The hematocrit is reported to be significantly higher in cortex than in olfactory bulb (subsection “Capillary flow and Po2 values in the somatosensory cortex of the awake resting mouse”). This needs to be discussed

how is this difference produced? Is it somehow produced by plasma skimming at a series of junctions of the vessels supplying the olfactory bulb? As this is an acute preparation, one needs to be certain that this is not due to some small amount of tissue damage when making the imaging window.

The fact that cortical capillary hematocrit is higher than that measured in the olfactory bulb glomerular layer is indeed an interesting finding. However, the mechanisms that underlie this difference are not easy to determine. We agree that differential hemoconcentration or hemodilution in vessels supplying the recorded capillaries could account for some of this difference. A modelling approach, such as that developed by Gould and Linninger (2015), could provide insight into this issue, however a definitive answer to this question remains elusive, as the details of the blood supply to the olfactory bulb are unknown. This is a weakness of the model, and we are just beginning a collaboration to investigate the anatomical organisation of the bulb vasculature. We hope to be able to answer this question in the future. We have added a short note on this point in the first paragraph of the subsection “Capillary flow and Po2 values in the somatosensory cortex of the awake resting mouse“.

Concerning the second point:

Conscious of the fact that a chronic cranial window preparation might affect our Po2 measurements, when starting this project, we initially employed a chronic thinned-skull window preparation (Drew et al. 2010), which avoided opening the cranium. However, after a number of months using this technique, we determined that the phosphorescence decays detected and the data recorded in this preparation were not compatible with high quality recordings, so we switched to using implanted glass cranial windows. Nonetheless, the facts that: 1. Our approach of placing the cover glass into the craniotomy and directly on to the dura mater surface seems to avoid the commonly described degradation (due to inflammation) and subsequent recovery of cranial window quality over the course of 2-3 weeks (Holtmaat et al. 2009), and 2. Our measurements were made many days after the chronic cranial window surgery suggest that we largely avoided damage to the tissue that would cause our results to deviate significantly from normal physiology.

2) Please show a raw trace of the Po2

profile between RBCs, or state explicitly that the line in the inset to panel A IS raw data (and not a schematic); why is that line continuous to the left and dotted on the right hand side?

The solid black trace in the lower right panel of Figure 1A was indeed data, drawn from recordings in anaesthetised mice as reported in Parpaleix et al. (2013). It has now however been replaced by a trace representing data from one capillary recorded in the awake mouse for this study. The continuous line represents the data trace, showing the profile of Po2 relative to the border of an RBC. The dotted portion of the trace is a mirror image of the data trace, and is used in the schematic solely for the purposes of illustration. The figure legend has been amended accordingly.

3) The authors repeatedly stress that previous measurements using two-photon phosphorescence only report Mean Po2. Could the authors please include a comparison of their Mean Po2 values against those made by other groups also using two-photon phosphorescence probe?

A paragraph has been added to the Discussion (second paragraph) to compare our recorded values of Po2 to those previously made with 2PLM by other groups.

4) Please explain the bars in Figure 1B far better. State completely clearly that Po2 is the mean (over vessels) of a single time averaged value obtained from each vessel, while SD is mean (over vessels) of the standard deviation of the Po2

in each vessel over time (if that is what it is!).

The necessary clarifications have been made in the legend for Figure 1.

5) The Po2 seems (subsection “Laminar organisation of capillary blood flow and Po2

values in the superficial layers of the somatosensory cortex”) to be determined by the local capillary density. Does this imply that local O2 consumption is independent of laminar position (one might have expected vessel density to be proportional to local O2 usage)? Is it possible to derive, from the laminar distribution of [O2] and the blood flow reaching each lamina, how much O2 is consumed by each cell layer?

Although our findings show that, at rest, cortical laminar Po2 is correlated to local capillary density, these data do not demonstrate any causal relationship, nor do they imply that the local rate of oxygen consumption is independent of the vessel density. Our lab is beginning to investigate the respective importance of cellular activity and delivery of oxygen in regulating tissue Po2 per layer. We plan to examine the effects of blocking pre/post-synaptic activity and firing on local tissue Po2. This research will help to clarify the factors that determine laminar and regional Po2.

We have amended the text of the manuscript to remove the implication of a causal relationship between laminar Po2 and capillary density (subsection “Laminar organisation of capillary blood flow and Po2 values in the superficial layers of the somatosensory cortex”, second paragraph).

6) A more thorough clarification of the argument of how isoflurane increases interstitial oxygen concentration would strengthen the manuscript. The current explanation links increased bloodstream oxygenation to the vasodilatory effect of isoflurane (subsection “Brain oxygenation is greatly enhanced by isoflurane asesthesia”). However, the increase in RBC flux (Figure 3A) resulting from vasodilation would lead to a decrease in oxygen unloading from RBCs, as evidenced by the increase in RBC Po2

(Figure 3A). If less oxygen is unloaded from the red blood cells, then less oxygen diffuses through the vessel wall, leading to a decrease in tissue oxygenation. This is contradicted by the statement that cellular processes occur at a higher oxygen concentration during isoflurane anaesthesia. This matter is further confused by the delivery of isoflurane in a 21% oxygen gas mixture.

It is true that increased RBC velocity (mm s-1) will lead to a decrease in oxygen unloading from each RBC. However, the overall increase in the RBC flux (cells s-1) that we measured will mean that although less oxygen is unloaded from each individual RBC, the total quantity of oxygen diffusing through the vessel wall could increase, leading to an increase in tissue oxygenation. This can be thought of as being somewhat similar to the increase in both RBC hemoglobin saturation and tissue oxygenation that is observed in a glomerulus during functional hyperaemia (Figure 4D, Parpaleix et al. 2013).

Although isoflurane is commonly delivered by others in either 100% oxygen or an air-oxygen mixture (where [O2]>21%), in our study we delivered it in air without additional oxygen ([O2]≈21%, (Line 142-143 of the track-changes version of manuscript). Our statement that “this difference is likely to be exacerbated when isoflurane is delivered in gas mixtures where [O2] is greater than 21%” was not made in reference to our study, but rather to the delivery of isoflurane in an air-oxygen mix by other researchers. This is now better clarified in the text.

7) Theoretical models are cited as evidence that InterRBC Po2 is in equilibrium with the surrounding tissue (subsection “Capillary flow and Po2

values in the somatosensory cortex of the awake resting mouse”). The theoretical models rely on accurate hematocrit prediction to determine the distribution of oxygen to the surrounding tissue. The relationship between radial oxygen gradients around capillaries, hematocrit and RBC velocities have been discussed in (I. Gould, A. Linninger, "Hematocrit Distribution and Tissue Oxygenation in Large Microcirculatory Networks", Microcirculation, 22, p1-18, 2015). The authors may consider inclusion of this reference to support the argument.

We thank the reviewer for this advice, and have cited this paper accordingly (subsection “Capillary flow and Po2 values in the somatosensory cortex of the awake resting mouse”, first paragraph).

8) The authors mention that they anaesthetise all mice before initiating even awake imaging experiments. Is it possible to head fix the mice without using anaesthesia? Is the anaesthesia necessary to do the retro-orbital injection of the fluorescein dextran and oxygen probe? It seems awake mice could be tail/vein injected. This might also affect the results since there could residual effects of anaesthesia. At the very least this confound should be acknowledged.

For awake recording sessions in our trained animals, we could head-fix the mice without using anaesthesia and without generating any stress. Technically, we preferred using a brief anaesthesia (< 5minutes) to perform a rapid retro-orbital injection rather than the tail/vein injection. However, to avoid any potential confound related to this brief exposure to isoflurane (2%) we:

A) Ensured a delay of ~2 hours between the period of isoflurane exposure and the initiation of the first set of recordings;

B) Performed multiple sessions of recordings over 1-2 days without reanaesthetising and injecting another bolus of PtP-C343 and fluorescein dextran. The data gathered in each of these sessions was comparable, suggesting that even the first set of recordings after the injection procedure was not affected by the exposure to isoflurane.

Author response image 1 shows Po2 values recorded from a single cortical capillary during 3 separate recording sessions across 2 days following the retro-orbital injection, demonstrating that Po2 values from this capillary were similar at ~2 hours (Session 1), ~7 hours (Session 2) and ~24 hours (Session 3) after the initial 5 minute exposure to isoflurane.

We have amended the text of the paper to better reflect our procedures for injection of PtP-C343, head-fixation and awake recordings (subsection “Habituation of mice to head restraint“).

Reviewer #1:

[…]

3) It is surprising that this study finds a lower Po2

in the surface layers of the cortex, whereas Devor and Sakadzic found a decrease in [O2] with depth. Is there any methodological difference that could account for this, e.g. O2 penetration into the brain from the cortical window in the earlier studies)? Was the laminar profile of O2 affected by isoflurane?

Sakadžić et al. (2010, 2014), and Devor (2011) sealed their cranial windows before their Po2 recordings, thus limiting the likelihood of oxygen diffusion into the brain directly from the atmosphere during the experiment. It is thus most likely that the decrease in Po2 with depth they reported is related to either the effects of anaesthesia, acute surgical procedures, or indeed a combination of these factors.

Our goal was mostly to stress the global effects of isoflurane. We thus unfortunately did not assess the effect of isoflurane on the laminar Po2 profile in the cortex, limiting paired measurements in capillaries from layer I.

[…]

6) Figure 3 legend. Stress that the measurements in A are from the glomerular layer, unlike the vessel image on the right which is pial.

The necessary alteration has been made to the figure legend.

[…]

8) In the first paragraph of the subsection “Laminar organisation of capillary blood flow and Po2 values in the superficial layers of the somatosensory cortex” state explicitly how deep capillaries "deeper than 410 um" could be.

410 µm was the maximum depth at which we made measurements, and this fact has now been clarified in the manuscript (subsection “Laminar organisation of capillary blood flow and Po2 values in the superficial layers of the somatosensory cortex”, first paragraph).

Reviewer #2:

Technical Comments:

[…]

2) Similarly, the authors cite previous low-resolution studies on the effects of isoflurane on brain tissue oxygenation, but the manuscript is lacking a side by side comparison with previous studies. An inclusion of the previous measurements would strengthen the conclusion made in this paper.

Rather than present a side by side comparison of our data with that of previous studies (as in a table), we present a significant discussion of the data from these lower resolution studies (citing their reported values of Po2. Discussion, first paragraph).

Reviewer #3: 1) The authors’ analysis is impressive and I don`t have any real concerns with it, since I am familiar with this literature. However, for those which have seen these phosphorescence measurements for the first time, I think it would be important to show supplementary data illustrating how the measurements are obtained. This could be the calibration curve or some raw data showing how Po2

and velocity/flux are calculated in a single experiment.

Rather than showing supplementary data, we now explicitly refer the reader to figures from both Parpaleix et al. (2013) and Lecoq et al. (2011), which present the measurement techniques we employ in great detail (subsection “Capillary blood flow analysis and EAT properties extraction”, second paragraph).

2) I would also like to see the authors give some measurements for arteriole and vein diameters with isoflurane anaesthesia vs. the control awake state, the example images point towards a robust finding.

Although we are able to show example images, and noticed the obvious dilation of pial vessels during isoflurane experiments, regretfully we did not systematically collect paired data on vessel diameter and cannot give serious statistics.

5) The authors should also comment on whether there are changes on efficacy of O2 signal overtime. From the protocol it sounds as though this probe must have an extremely long half-life. Why this probe would not be cleared by the liver is uncertain, its pharmacokinetics should be further investigated or at least if there is information this should be discussed so others can have a better feel for how long after injection can quantification of O2 be done.

See our response to the Key question 8. Further information on the interval following PtP-C343 injection during which accurate Po2 measurements can be made has been added to the manuscript (subsection “Habituation of mice to head restraint”, second paragraph).

https://doi.org/10.7554/eLife.12024.013

Article and author information

Author details

  1. Declan G Lyons

    1. Institut National de la Santé et de la Recherche Médicale, U1128, Paris, France
    2. Laboratory of Neurophysiology and New Microscopies, Université Paris Descartes, Paris, France
    Contribution
    DGL, Conducted the experiments, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-1775-4653
  2. Alexandre Parpaleix

    1. Institut National de la Santé et de la Recherche Médicale, U1128, Paris, France
    2. Laboratory of Neurophysiology and New Microscopies, Université Paris Descartes, Paris, France
    Contribution
    AP, Conducted the experiments, Conception and design, Acquisition of data
    Competing interests
    The authors declare that no competing interests exist.
  3. Morgane Roche

    1. Institut National de la Santé et de la Recherche Médicale, U1128, Paris, France
    2. Laboratory of Neurophysiology and New Microscopies, Université Paris Descartes, Paris, France
    Contribution
    MR, Conducted the experiments, Acquisition of data
    Competing interests
    The authors declare that no competing interests exist.
  4. Serge Charpak

    1. Institut National de la Santé et de la Recherche Médicale, U1128, Paris, France
    2. Laboratory of Neurophysiology and New Microscopies, Université Paris Descartes, Paris, France
    Contribution
    SC, Conducted the experiments, Conception and design, Analysis and interpretation of data, Drafting or revising the article
    For correspondence
    serge.charpak@parisdescartes.fr
    Competing interests
    The authors declare that no competing interests exist.

Funding

Institut National de la Santé et de la Recherche Médicale

  • Serge Charpak

Fondation Leducq

  • Serge Charpak

European Research Council (ERC-2013-AD6 (339513))

  • Serge Charpak

Ecole de Neuroscience, Paris Ile-de-France

  • Serge Charpak

Fondation pour la Recherche Médicale (Equipe FRM)

  • Serge Charpak

Ecole de Neuroscience, Paris Ile-de-France (Doctoral Fellowship)

  • Declan G Lyons

Fondation pour la Recherche Médicale (FDT20130928252)

  • Declan G Lyons

Agence Nationale de la Recherche (ANR 12 BSU4-0017-01)

  • Serge Charpak

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We would like to particularly thank Sergey Vinogradov for providing PtP-C343. We also thank Emmanuelle Chaigneau, Mathieu Ducros, Yannick Goulam-Houssen and Sergei Sasnouski for software support, along with David Attwell, Ravi Rungta and Etienne Audinat for their critical comments.

Ethics

Animal experimentation: All animal care and experimentation was performed in accordance with the INSERM Animal Care and Use Committee guidelines (protocole number CEEA34.SC.122.12 and CEEA34.SC.123.12)

Reviewing Editor

  1. David Kleinfeld, University of California, San Diego, United States

Publication history

  1. Received: October 1, 2015
  2. Accepted: January 25, 2016
  3. Accepted Manuscript published: February 2, 2016 (version 1)
  4. Version of Record published: February 23, 2016 (version 2)

Copyright

© 2016, Lyons et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 2,543
    Page views
  • 632
    Downloads
  • 26
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, Scopus, PubMed Central.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Human Biology and Medicine
    2. Neuroscience
    Xunde Xian et al.
    Research Article Updated
    1. Human Biology and Medicine
    2. Neuroscience
    Lukas Hofmann et al.
    Research Article Updated