|Home | About | Journals | Submit | Contact Us | Français|
Under physiologic conditions, microvascular oxygen delivery appears to be well matched to oxygen consumption in respiring tissues. We present a technique to measure interstitial oxygen tension (PISFO2) and oxygen consumption (VO2) under steady-state conditions, as well as during the transitions from rest to activity and back. Phosphorescence Quenching Microscopy (PQM) was employed with pneumatic compression cycling to achieve 1 to 10 Hz sampling rates of interstitial PO2 and simultaneous recurrent sampling of VO2 (3/min) in the exteriorized rat spinotrapezius muscle. The compression pressure was optimized to 120–130 mmHg without adverse effect on the tissue preparation. A cycle of 5 s compression followed by 15 s recovery yielded a resting VO2 of 0.98 ± 0.03 ml O2/100cm3 min while preserving microvascular oxygen delivery. The measurement system was then used to assess VO2 dependence on PISFO2 at rest and further tested under conditions of isometric muscle contraction to demonstrate a robust ability to monitor the on-kinetics of tissue respiration and the compensatory changes in PISFO2 during contraction and recovery. The temporal and spatial resolution of this approach is well suited to studies seeking to characterize microvascular oxygen supply and demand in thin tissues.
The oxygen (O2) supply and demand of skeletal muscle exists in a dynamic and responsive equilibrium within the interstitial fluid providing a key focal point of observation (Ferreira, Poole et al. 2005). At rest, O2 diffuses according to its partial pressure gradient from hemoglobin binding sites within red blood cells (RBCs) through the interstitium to match overall rates of mitochondrial oxygen consumption (VO2) within cells (Sarelius and Pohl 2010). The partial pressure of O2 in the interstitial space (PISFO2), which is dissociated from red blood cell (RBC) hemoglobin and free to diffuse to intracellular mitochondrial sinks without appreciable interference from myoglobin desaturation, is sensitive to rapid changes in the balance between supply and demand. Different approaches ranging from whole animals (Hoyt, Wickler et al. 2006, Rodrigues, Figueroa et al. 2007), isolated muscles (Goodwin, Hernandez et al. 2012); (McDonough, Behnke et al. 2005), to single fibers (Hogan 2001); (Kindig, Howlett et al. 2003); (Wust, van der Laarse et al. 2013) have sought to characterize the relationship of the partial pressure of O2 (PO2) and VO2. The description of PISFO2 supply/demand coupling for intact muscle preparations in vivo, however, remains incomplete.
When assessing the VO2 of a particular microvascular bed under physiological conditions, the difference in PO2 between the input arteries/arterioles and output veins/venules can be related to blood flow through Fick’s principle to provide a good description of O2 utilization. With regards to isolated, contracting skeletal muscle operating below the lactate threshold - an oft-studied model of PO2 and VO2 dynamics - it is acceptable to say that changes in VO2 are proportionate to energy demand (Gutierrez, Pohil et al. 1989, Poole and Richardson 1997). VO2 calculated by Fick’s principle is regarded as a metric of tissue/organ respiration and has yielded values in mammalian skeletal muscles at rest ranging from 0.054 to 3.1 ml O2/100 g*min (Edmunds and Marshall 2001); (McDonough, Behnke et al. 2005, Hoy, Peoples et al. 2009); (MacInnes and Timmons 2005). While fully practical for determining the amount of O2 consumed by a particular organ or tissue, Fick’s principle cannot directly deconstruct fiber/tissue-specific VO2 rates from these heterogeneous preparations nor indicate what spatial relations exist between VO2 and the gradients of O2 supply. Thus, for a variety of different muscle compositions a wide range of VO2 values have been reported. Additionally, there is limited temporal resolution since the hemoglobin O2 capacitance can act as a buffer between cellular respiration and changes in vascular O2 depletion.
Another approach to determining the rate of O2 extraction from a particular region of tissue is to arrest blood flow and chart the local oxygen disappearance curve (ODC). One approach uses PO2 microelectrodes (Buerk, Nair et al. 1986), but has limited spatial sensitivity and may cause tissue damage. A less invasive technique makes use of Near-Infrared Spectroscopy (NIRS) (for review see: (Ferrari, Muthalib et al. 2011). Briefly, NIRS works by illuminating the target tissue with wavelength pairs of 700–805 nm and 830–805 nm to assess hemoglobin and myoglobin oxygenation, blood volume dynamics (relative measurements of absorption at ~805 nm over time) and mitochondrial cytochrome aa3 oxygenation to assess both vascular O2 supply and tissue metabolic rate. For measurements of VO2, blood flow is arrested and the rate of oxyhemoglobin conversion to deoxyhemoglobin is related to a change in PO2 over time, thus providing indirect information on VO2 kinetics. This non-invasive approach has yielded resting VO2 values in human calf muscles around 0.20 ml O2/100 g*min (Cheatle, Potter et al. 1991); (De Blasi, Luciani et al. 2009), which fall into the lower range found using Fick’s principle, but have greater spatial and temporal resolution, as well as an assessment of mitochondrial respiration by which to compare the measured ODC. A lingering problem with these measurements is the muting of the ODC resulting from the buffering actions of hemoglobin and the sigmoidal shape of the ODC.
Phosphorescence Quenching Microscopy (PQM) is a well-established technique (Rumsey, Vanderkooi et al. 1988); (Smith, Golub et al. 2002) which associates the phosphorescence emission of a porphyrin dye to the PO2 of biological preparations with a high degree of spatial (~10 μm) and temporal (<0.1 s) resolution (Wilson, Vanderkooi et al. 1987). PQM, by virtue of its unintrusive nature, has been applied to measure the distribution of O2 in a number of microvascular preparations (Torres Filho and Intaglietta 1993); (Golub, Popel et al. 1997);(Behnke, Kindig et al. 2001, McDonough, Behnke et al. 2005) (Copp, Hirai et al. 2011). Similarly to NIRS, but without a dependence on hemoglobin, PQM it has been found to be a useful tool in making discrete, single measurements of VO2 by means of mechanically isolating small regions of tissue from the blood supply and observing the ODC (Golub, Tevald et al. 2011). Here we demonstrate a method of VO2 sampling utilizing PQM that, in addition to providing continuous readings of PISFO2, also allows for simultaneous, recurrent measurements of VO2 in a variety of metabolic states without interfering with microvascular function. This new technique allows for extended measurements that can capture many systemic and local events that might influence tissue oxygenation with a simultaneous assessment of energy demand.
The following protocols and experimental procedures were approved by the Virginia Commonwealth University Institutional Animal Care and Use Committee and are consistent with the National Institutes of Health guidelines for the humane treatment of laboratory animals, as well as the American Physiological Society’s Guiding Principles in the Care and Use of Animals. Fifty male Sprague Dawley rats (250–300 g; Harlan, Indianapolis, IN) were given an initial dose of ketamine/acepromazine (75 mg/kg and 2.5 mg/kg, respectively; i.p.). A femoral vein cannula was inserted to maintain a continuous plane of anesthesia (assessed to have heel response with minimal toe pinch response) with Alfaxalone acetate (Alfaxan, Schering-Plough Animal Health, Welwyn Garden City, UK) at 0.1 mg/kg/min. A carotid artery catheter with attached pressure transducer was inserted to provide continuous measurement of mean arterial pressure (MAP) (MP-150, BIOPAC Systems, Goleta, CA). A tracheal cannula was inserted to maintain airway patency. After experimentation, animals were euthanized with an overdose of Euthasol (150 mg/kg, pentobarbital component, intravenously; Delmarva, Midlothian, Virginia).
The rat spinotrapezius muscle was exteriorized as described by Gray (Gray 1973) with some modification to accommodate measurement of PISFO2 and VO2, on a thermostatic animal platform adapted for intravital microscopy (Golub and Pittman 2003). A longitudinal, midline incision beginning between the shoulder blades exposed the muscle, which was then separated from underlying fascia and connective tissue by blunt dissection. Careful attention was given to maintaining tissue hydration with phosphate-buffered saline (PBS) and to avoid stretching the muscle as it was mounted on the transparent, thermostatic pedestal of the animal platform. The preparation was then covered with a transparent, gas impermeable film (Krehalon, CB-100; Krehalon Limited, Japan) to isolate it from the atmosphere.
Measurements of PISFO2 and VO2 in the exteriorized spinotrapezius preparation were performed using an intravital microscope (Axioimager2m, Carl Zeiss, Germany) configured for PQM via epi-illumination through a 20×/0.8 objective (Plan-APOCHROMATE, Zeiss, Germany). The phosphor (Oxyphore R2; Oxygen Enterprises, Philadelphia, PA) was topically applied to the tissue and allowed to diffuse into the interstitium. The phosphor probe was excited by a xenon flash lamp (Model FX249, EG&G Electro-optics Co., Salem, MA), which delivered 0.5 J/flash for 4 μs (full width at one third maximum pulse height) to an octagonal region 300 μm in diameter at a frequency of 1 or 10 Hz. The excitation light pulse was passed through a filter cube consisting of a narrow-band filter (INTOR 525/70/75%, Intor Inc. Socorro, NM), which was specifically designed to excite the probe at 525 nm (Q-band), a dichroic mirror (Chroma 565 DCLP, Chroma Technology Corp. Bellows Falls, VT), and a wide-band filter (Oriel Cut-on >650nm, Newport Corp, Stratford, CT) for selective collection of phosphorescence emission. The phosphorescence signal was collected by a photomultiplier tube and sent to a modified amplifier (OP37EP, Analog Devices, Norwood, MA) which functioned as a current-to-voltage converter outfitted with a precision analog switch (ADG419BN, Norwood, MA) allowing for a gating time of 12 μs. The 12 μs delay between pulse initiation and data collection covered the 5 μs propagation delay from the plasma arc to the flash, the 4 μs light pulse, and 4–5 μs of any residual thermal tail from the flash lamp and lingering fluorescence signal from tissue or probe. Thus, the data collected were optimized to contain the highest contribution from the short phosphorescence lifetimes (i.e., highest PO2) without interference from fluorescent artifact, for reliable measurements of PISFO2. These signals were visually monitored in real-time with an oscilloscope (72-3060, Tenma, Springboro, OH) before they were passed to a 12-bit analog-to-digital converter (PC-MIO-16E-4, National Instruments, Austin, TX), and saved digitally on a Dell Optiplex PC (Round Rock, TX). The tissue’s microvasculature was visualized under transillumination with a light emitting diode (Luxeon V Star white, Quadica Developments Inc., Brantford, Ontario) that was fed through the microscope’s condenser. The image, under 20× magnification, was captured in real-time by a color CCD camera (KP-D20BU, Hitachi, Tokyo, Japan) and displayed on a flat-screen color video monitor (Model LN19A450C1D, Samsung, Japan). Transillumination was used for measurement site acquisition and to verify tissue flow and arrest during air bag compressions.
Pd-meso-tetra-(4-carboxyphenyl)porphyrin (Oxyphor R2) was obtained from Oxygen Enterprises (Philadelphia, PA) and conjugated to human serum albumin in PBS. The probe was then topically applied to the spinotrapezius muscle at a concentration of 10 mg/ml for 30 minutes. This allowed for deep, uniform diffusion of the probe into the muscle’s interstitium from where the phosphorescence signal exclusively emanated (Golub, Tevald et al. 2011). The preparation was then rinsed with PBS to remove unloaded probe. Probe preparation, loading, and measurements were all performed under dark conditions to minimize photobleaching.
A nonlinear fitting procedure (Levenberg-Marquardt) for the individual phosphorescence decay curves was based on the rectangular PISFO2 distribution model (Golub, Popel et al. 1997) and yielded the following fitting equation for the phosphorescence time course:
where t (μs) is the time from the beginning of the phosphorescence decay, I(t) (volts) is the phosphorescence decay curve, I0 (volts) is the magnitude of the phosphorescence signal at t = 0, M (mmHg) is the mean PO2 in the volume of detection, δ (mmHg) is the half-width of the PO2 distribution, and B (volts) is the baseline offset of the amplifier. The constants for Oxyphor R2 are K0 = 1.53×10−4 μs−1 and Kq = 4.3×10−4 μs−1 mmHg−1.
Based on a previously described technique (Golub, Tevald et al. 2011), an objective-mounted air bag was inflated to produce a focal compression of the underlying tissue that resulted in immediate blood flow arrest and blood extrusion from the measurement region. Prior to compression, the air bag was pressurized to 5 mmHg to preserve gentle contact between the Krehalon barrier and the tissue to extrude any excess fluid on the surface of the tissue that might interfere with phosphorescent lifetime measurements. Next, following baseline measurements, the bag was pressurized to 120–130 mmHg. The rapid arrest of blood flow (<1 s), RBC extrusion, and any lateral shifting was observed under transillumination for quality assurance. Following the required compression duration (see below), the pressure in the air bag was decreased to 5 mmHg and flow was restored. Compression cycling was controlled by a custom-built automatic dual-stage pressure regulator, which allowed for appropriately timed alternation between two pressure levels, high (120–130 mmHg) and low (~5 mmHg), accommodating the repetitive 5 seconds of compression and 15 seconds of recovery cycle (denoted “5 s × 15 s cycle” below). The “5 s × 15 s cycle” was based on pilot experiments where 5 s of compression were needed to obtain five descending PO2 values for accurate plotting of the ODC and calculation of VO2 from the linear component. 15 s of decompression then permitted PISFO2 recovery to P0 prior to compression. A digital display reported the current pressure inside the air bag and a toggle switch allowed for the selection of manual pressure delivery for longer durations of both high and low pressure.
From the moment of flow arrest, the O2 supply was stopped and the PISFO2 began to decrease in a linear fashion. This ODC was recorded and used to compute the VO2 of the underlying tissue as previously described (Golub, Tevald et al. 2011). Briefly, the rate at which PISFO2 changes under conditions of flow arrest depends on three factors which appear on the right hand side of Eq. 2 below: 1) the rate at which O2 is consumed through respiration in the excitation region; 2) the rate at which O2 is consumed by the method itself (i.e., photoconsumption); and 3) the rate at which O2 diffuses into the region of excitation from the surrounding tissue (i.e., “refill” or replenishment of consumed O2). Each flash of the excitation source gives rise to a phosphorescence decay curve and each curve can be analyzed by Eq. 1 to yield a best-fit value for PISFO2 (i.e., M in Eq. 1) in the excitation region. On a per flash basis, the change in PISFO2 can be described by Eq. 2:
where dPn/dn is the rate per flash of O2 disappearance from the excited region, Pn is the average PISFO2 within the 300 μm diameter measurement area after flash n, Vn is the contribution of cellular respiration per flash, K is the coefficient of O2 consumption by the PQM method, and Z is the coefficient of O2 inflow (refill) by passive diffusion at the periphery of the measurement area. The interaction of an excited phosphor molecule with O2 (collisional quenching) causes the conversion to singlet oxygen and thereby reduces the local PO2. O2 consumption by the method depends on concentration of phosphor, intensity of excitation light and PO2. Since measurements are typically made with constant phosphor concentration and excitation intensity (incorporated in the factor K), the contribution of O2 consumption by the method is expressed as the term -KPn, proportional to local PO2. Refill decreases as the radius of the illuminated measurement region increases. With a large excitation area of 221,841 μm2 and a small KPn, 0.5% per flash, the coefficient of refill (Z) was found to be insignificant for the initial, linear portion of the ODC and and Eq. 2 can be simplified to:
V0 (mmHg/flash) can then be converted to VO2 by accounting for both flash rate (F) and the solubility of O2 in the interstitial fluid (α = 39 nl O2/(cm3·mmHg))(Mahler, Louy et al. 1985) to yield:
Rearranging Eq. 3, the expression for transforming dPn/dn into a measurement of tissue oxygen consumption is as follows:
Phosphorescence photoconsumption is directly proportional to the amount of light injected into the system per unit time. Therefore, to assess the impact of these measurements on the observed ODC, two separate excitation frequencies were used to measure dPn/dn at random sites in intact spinotrapezius preparations. Using the 5 s × 15 s high/low pressure cycling method, a site was selected in the R2-loaded muscle and measured for 100 s at either 1 or 10 Hz excitation. Immediately following this a second measurement was made using the other frequency for 100 s. The differential ODCs were then compared as follows:
Where F is the flash rate for each respective measurement (1 or 10 Hz here) and dPn/dn is the rate of oxygen disappearance per flash.
The spinotrapezius muscle was prepared and mounted as described above. A site was selected that contained well-perfused vessels with diameters < 15 μm. Based on preliminary and concurrent measurements, a compression recovery profile of 5 s high pressure and 15 s low pressure was selected to assess the influence of increasing pressure on both dPn/dn and the PO2 recovery time course back to baseline. After a brief baseline measurement using PQM at low pressure (~5 mmHg) the sequence of high and low pressures was initiated. The first high-pressure compression was set to 40 mmHg and subsequently increased by 10 mmHg per measurement to a maximum of 130 mmHg.
PQM was used to monitor the changes in PISFO2 and VO2 between rest and contraction in the rat spinotrapezius muscle. Two silver wire electrodes were placed in parallel along the length of the muscle at its edges. This positioning was found to maximize the passage of current into the tissue without creating air pockets or other morphological distortions that might interfere with both PQM and/or VO2 sampling. Next, appropriate sites for measurement, as defined by a proximity of ~200 μm to a sufficiently large (40–100 μm diameter) arteriole and possessing sufficient capillary perfusion, were selected using transillumination at 10× magnification. The 20× objective with attached air bag was then used with transillumination to verify flow arrest under a brief (~2 s) compression at high pressure. Next, a short succession of PQM measurements at 1 Hz was performed to confirm that PISFO2 corresponded to the observed perfusion rates. The protocol for contraction measurements consisted of the following sequence: 60 s rest, 120 s of contraction (2 Hz stimulation train at 10 volts with 20 ms duration) and 120 s of recovery. PQM (1 Hz) and VO2 sampling (5 s × 15 s cycle) were run throughout the 5-minute time course. The concurrent profiles of PISFO2 and VO2 were then plotted for analysis.
Data within these experiments are shown as mean ± standard error unless otherwise noted in figures and tables. One-way ANOVA was conducted using JMP software version 8.0 (SAS Institute Inc., Cary NC) followed by Tukey’s multiple comparison test to determine significant differences between points in the time course of measurements. Statistical significance was taken to be P < 0.05.
Four spinotrapezius muscles at rest (11 sites each) were assessed for the differential rates of oxygen disappearance during compression at 1 and 10 Hz phosphorescence excitation rates. The change in -dPn/dn (Fig. 1) was then used to calculate the coefficient of photoconsumption (K) per the PO2. K was found to be 7.7±0.6 × 10−3 (n=44), which indicated that approximately 0.5% of the observed O2 consumption per flash during resting conditions was due to O2 consumption by PQM itself.
Nine resting state spinotrapezius muscles were subjected to stepwise 10 mmHg increases in air bag pressure from 40 to 130 mmHg and assessed for changes in dPn/dn (Fig. 2A) and the restoration of PISFO2 (see Fig. 2B) during the 5 s × 15 s compression/recovery protocol. A baseline measurement of PISFO2 at the standard low pressure of 5 mmHg preceded initiation of the 5 s × 15 s VO2 sampling protocol. PISFO2 recovery did not vary significantly over the range of increasing compression pressures. The linear drop in PISFO2 (dPn/dn) that occurred with each compression was reported as an absolute value in terms of mmHg/sec as shown in Figure 2B. There was a progressive rise in –dPn/dn from 3.2±0.2 mmHg/sec at 40 mmHg pressure to 4.4±0.2 mmHg/sec at 80 mmHg, above which –dPn/dn leveled off. There was no statistical difference in –dPn/dn between groups from 80 mmHg to 130 mmHg. There was a slight, but insignificant, elevation in –dPn/dn at 120 mmHg. To accommodate variable tissue thickness and ensure that both blood flow-arrest and RBC extrusion took place rapidly and effectively, a range of 120–130 mmHg was chosen as optimal for the 5 s × 15 s VO2 sampling protocol. Values of -dPn/dn were not converted to VO2 because some lower pressure measurements had incomplete blood flow-arrest and RBC extrusion.
The VO2 method necessitated a temporary reduction in local PISFO2 which introduced the potential for these brief compressions to create a cumulative disruption in microvascular O2 supply/demand dynamics. To address this, preliminary studies (data not shown) were first carried out under resting conditions to produce an optimal VO2 sampling rate that did not produce noticeable changes in microvascular oxygenation or tissue VO2. As a result, a 5 s compression duration was determined to be a sufficient amount of time for an accurate assessment of the rate of O2 disappearance while still remaining in a physiological time course of brief ischemia. Figure 3A shows the averaged, PISFO2 curves from eight animals (~14 sites per animal) after 15 s of recovery following 5 s of compression of the spinotrapezius muscle at rest. The 15 s profile of recovery showed no significant differences between baseline and subsequent measurements following compression in terms of PISFO2 and VO2 (see Fig. 3).
Early data analysis of randomized sites showed that for various levels of perfusion, as indicated by the PISFO2 prior to the rapid onset of compression (P0), there existed a range of PISFO2 where VO2 was dependent on PO2. Fourteen resting state spinotrapezius muscles (~10 sites per animal) using a 10 hz PQM excitation frequency were used to study the range of VO2 dependence on PISFO2 under resting conditions (Fig. 4). VO2 rose from 0.24±0.09 ml O2/100 cm3 min at a PISFO2 of 12.2±3 mmHg to 1.07±0.16 O2/100 cm3 min at a PISFO2 of 27.1±0.5 mmHg, which then remained within a range of 1.07 to 1.33 ml O2/100 cm3 min to a PISFO2 of 62 mmHg. At higher PISFO2 levels, VO2 appeared to drop off slightly. Statistical tests were run against the VO2 measurement at a PISFO2 of 12.2 mmHg and trend towards a critical point for VO2 dependence on PISFO2 between 25 and 30 mmHg. To limit any sort of edge effect, a 10 mmHg buffer was placed between the high end of this critical point so that only measurements of VO2 where P0>40 mmHg were included in the analysis of resting VO2. Occasions of apparent atmospheric contamination (P0>70 mmHg) were also excluded from analysis to guard against artifactually high PO2. For the range of P0 values above 40 mmHg (Fig 3A; P0= 59.8 ± 1.1 mmHg; n= 125) where the VO2 data could be reliably considered insensitive to PISFO2, the mean VO2 was measured to be 0.98 ± 0.03 ml O2/100 cm3 min (Fig. 3B; n= 102).
Electrical stimulation was applied to the spinotrapezius muscles of seven animals to assess VO2 and PISFO2 under dynamic conditions. Measurements consisted of the following series: 60 s baseline (allowing 3 measurements of VO2 and PISFO2) to provide an assessment of the resting state, 120 s of electrical stimulation, and a final 120 s of recovery without stimulation (Fig. 5). Contractile activity was isometric and confined to the period of electrical stimulation. Measurements of PISFO2 made during the 60 s baseline under resting conditions were found to be consistent with other measurements of resting skeletal muscle conditions reported here. The onset of electrical stimulation produced an immediate decline in PISFO2 (as assessed during the low pressure recovery period immediately prior to each high pressure compression), which significantly decreased from 67.7±2.0 mmHg (n=25) to 37.2±2.7 mmHg (n=25; p<0.0001) within 20 s. This reduction continued for the next 20 s and then began to rise slowly; however, the rise in PISFO2 above the minimum value at 40 s of stimulation was not statistically significant. PISFO2 remained near 30 mmHg for the duration of electrical stimulation. Following muscle stimulation, PISFO2 rose from the average value during stimulation of 35.1±3.3 mmHg (n=25) back toward the baseline value (Fig. 5A). The onset of electrical stimulation caused a significant increase in VO2 from 1.1±0.1 ml O2/100cm3 min (n=28) to 3.1±0.2 ml O2/100cm3 min (n=28) within 20 s. This increased consumption persisted for the duration of stimulation. VO2 did not change significantly between the final data point taken during stimulation (3.0±0.2 ml O2/100cm3 min; n=25) and the first time point of recovery (i.e., 20 s into the recovery period; 3.2±0.2 ml O2/100 cm3 min; n=28). Following this, however, the recovery phase saw a progressive decrease in VO2 back towards baseline with a final measured value of 1.6±0.1 ml O2/100 cm3 min over 120 s (Fig. 5B).
A new technique for concurrent measurements of microvascular O2 supply (related to PISFO2) and demand (related to VO2) dynamics under various metabolic conditions was developed and employed in the rat spinotrapezius muscle. This VO2 method was adapted from a previously described technique that utilized the pressurization of an airbag to rapidly arrest blood flow while simultaneously observing the resultant drop in PISFO2 with PQM (Golub, Tevald et al. 2011). The system, which was originally designed to record a single, protracted ODC, was extended to include an automatic dual-pressure regulator protocol that allowed for brief, recurrent ODC collection. The early part of the ODC was found to be linear reflective of tissue VO2. When corrected for a 0.5 or 5% photoconsumption per second of O2 at 1 and 10 Hz flash rates respectively, resting VO2 values corresponded to the range of 0.5 to 3.1 ml O2/100 cm3 min reported by other investigators of mammalian skeletal muscle respiration (MacInnes and Timmons 2005, McDonough, Behnke et al. 2005, Hoy, Peoples et al. 2009). The use of PQM to continuously sample PISFO2 in conjunction with the automatic compression/decompression cycling method to make serial assessments of VO2 afforded excellent temporal and spatial resolution that was able to track PISFO2 dynamics in a muscle during electrically stimulated isometric contractions.
Two confounding variables in localized VO2 measurements of intact organs are the O2 buffering capacity of hemoglobin-bound O2 in RBCs and dissolved O2 in the interstitium. By extruding RBCs from the measurement site and measuring PO2 directly in the interstitial space, the ODC can be used to determine local VO2 without interference from luminal O2 capacitance. In oxidative muscles the contribution of intracellular myoglobin to the O2 buffering capacity of cells cannot be ignored at intracellular PO2 values near and below the P50 of myoglobin (~ 5 mmHg); however, the PO2 range over which normoxic VO2 was determined in the studies presented here was well above this level. To mitigate the influences of RBCs and interstitial buffering, a transparent air bag mounted on a 20× microscope objective was chosen to provide focal pressure to the region of interest, while retaining the objective’s ability to perform both intravital and phosphorescence microscopic measurements. The necessary pressure to arrest flow and extrude RBCs for early studies was initially based on the rat’s mean arterial pressure – approximately 100 mmHg for 250 g male Sprague Dawley using this protocol (Song, Nugent et al. 2014)) - which dictated the air bag’s high-pressure setting to be 120 mmHg. Once applied to the tissue, compression was observed, both through transillumination and PQM, to cause rapid flow-arrest (< 1 s) and RBC extrusion. PQM was used to record the resultant ODC, whose initial component was linear. During extremely hypoxic conditions, a mono-exponential decay could be observed for the last 1–2 seconds of compression. In the case of lingering RBCs, the early part of the ODC would have registered as an exponential decay instead of a linear drop in the normoxic region. At the lowest excitation frequency of 1 Hz, 5 s of compression covered the linear portion of the ODC (P0= ~60 mmHg) and provided up to 5 data points, which were sufficient to produce an accurate linear fit of –dPn/dn for eventual conversion to VO2.
A cycling protocol of 5 s of high pressure followed by 15 s of recovery was found to: 1) restore PISFO2 back to baseline; 2) prevent cumulative effects of serial VO2 measurements from altering the skeletal muscle’s oxygen supply and demand interplay; and 3) provide a reasonable number (3) of VO2 measurements per minute. Given the rapid flow arrest and extrusion of RBCs brought on by pneumatic compression, a sufficiently linear portion of the ODC can be captured during shorter compressions at higher PISFO2 sampling rates. The 5 s duration of compression was chosen to provide the capability of measuring VO2 under conditions where the changes were not too rapid. Shorter compression durations down to ~2 s are likely possible in future refinements of the technique as 2–10 Hz PISFO2 sampling would be sufficient to obtain a linear trend of the early ODC without reducing the signal-to-noise ratio. Figure 4 shows data captured using a 10 Hz PISFO2 sampling rate and lies in good agreement with the 1 Hz data. 10 Hz was found useful during hypoxia when consumption rate shortened the linear ODC component.
To fully describe the effects of pneumatic compression on the microvasculature, a range of pressures from 40 to 130 mmHg was tested using the 5 s ×15 s sampling protocol for the recovery of PISFO2 (see Fig. 2B) and the rate of O2 disappearance during compression (-dPn/dn; see Fig. 2A). No significant difference was found for the recovery of PISFO2 over this range of compression pressures. It was expected that at some point, recovery would either show a pronounced hyperemic response or begin to show protracted recovery times if compression pressure brought about vascular damage. Neither situation was observed, though there did appear to be a statistically insignificant trend in the elevation of PISFO2 during recovery around 70 mmHg of pressure. Likely, the detrimental effects of compression would have occurred at higher pressures, but these were not tested, as pressures over 130 mmHg greatly increased the probability of air bag rupture.
Stepwise increases in compression pressure of 10 mmHg starting with 40 and ending with 130 mmHg (5 mmHg was considered low pressure for the purpose of extrusion of any surface fluid layer that might confound PQM measurements) showed the corresponding rise in –dPn/dn that one might expect as flow-arrest and RBC extrusion becomes increasingly complete until reaching a plateau. Compressions from 80 mmHg to 130 mmHg did not yield a significant change in –dPn/dn as shown in Figure 2A, which indicated that maximal flow arrest and RBC extrusion had been achieved. Measurements at 90 and 100 mmHg appeared to be slightly elevated over measurements made from 110 to 130 mmHg, but this difference was not statistically significant. As higher-end pressures did not appear to cause any adverse effects in either VO2 or the recovery of PISFO2, and the initial VO2 sampling interval was determined using 130 mmHg, the decision to use a range of compression pressures from 120 to 130 was deemed appropriate. This allowed for a justifiable adjustment in compression pressure to accommodate a range of tissue sizes and microvascular architectures. The effectiveness of flow arrest was always visualized prior to measurement and 120 to 130 mmHg had a consistently high success rate over the course of experimentation.
Another point of interest was the relevance of the –dPn/dn versus pressure profile on the microvasculature below complete flow arrest. As pressure rose, perfusion dropped with an apparent linear progression. If these pressures had been continued beyond 5 s, an eventual point where the reduction in perfusion could no longer keep up with O2 demand would be reached - inevitably resulting in cellular hypoxia. This sort of behavior has been found to occur in Compartment Syndrome (de Laet and Malbrain 2007) and a study looking into the pressure needed to surpass this “tipping point” in the development of hypoxic tissues using this method would be useful.
Photoconsumption of O2 by PQM is a documented phenomenon (Golub and Pittman 2012) and most readily observed at high excitation frequencies or in cases where microvascular O2 delivery was purposely stopped to observe the ODC for VO2 measurements. Use of PQM to determine local PO2 relies on the ability of O2 to quench the phosphorescence of an excited probe (Oxyphor R2 in this case). In doing so, O2 is converted to singlet oxygen, thereby consuming the original O2 molecule. The consumption rate for the linear portion of the ODC is constant, however, and easily included in the equation for VO2. When O2 flux, probe concentration, and measurement region are held constant, photoconsumption becomes dependent on light intensity. Using this fact, we obtained an optimal measurement for photoconsumption in the spinotrapezius preparation by varying the flash rate during flow arrest and accounting for the contribution of photoconsumption to the measured –dPn/dn. Other probe-loaded media (i.e., gels, dead muscle tissue, microcapillary tubes) proved too difficult to translate to the physiologically intact situation, despite being excellent models of the photoconsumption effect. In the case of the R2 probe at 10 mg/ml which diffused into the spinotrapeizius muscle, our reported value of K to account for photoconsumption of oxygen by PQM was 7.7±0.6 × 10−3 (see Eqs. 3 and 6). New values for K need to be determined for different experimental conditions. In cases where PQM is used to measure a well-perfused region at low excitation frequencies (≤1 Hz), the value of K will be small and unlikely to affect the measured PISFO2.
VO2 determined by the procedure described above predominantly represents mitochondrial respiration where O2 functions as the terminal electron acceptor in the mitochondrial electron transport chain. In addition to energy production (ATP and heat), mitochondria produce small quantities of reactive oxygen species, reported as anywhere from 0.15 to 2% of mitochondrial VO2 under physiological conditions (St-Pierre, Buckingham et al. 2002); (Misra, Sarwat et al. 2009); (Tahara, Navarete et al. 2009). This percentage may rise in times of mitochondrial dysfunction, as seen during ischemia/reperfusion injury.
Measurements included in this study were performed under normoxic conditions (P0≥40 mmHg) where the ODC was observed to remain independent from PISFO2 during compression. Our definition of normoxia was a conservative limit based on measurements of VO2 vs PISFO2 (Fig. 4), which suggest a critical point somewhere between 25 and 30 mmHg for the interstitial space. We added an additional 10 mmHg buffer to avoid any edge effect from reducing our assessment of VO2. We then used electrical stimulation of the spinotrapezius muscle as a model to demonstrate the utility of PQM in simultaneous recording of VO2 - PISFO2 kinetics. Stimulation of the muscle showed an immediate decrease in PISFO2 which leveled off by the 20 s time point, in good agreement with the rapid adjustment of blood flow to meet sudden increases in metabolic demand (Naik, Valic et al. 1999); (Clifford and Hellsten 2004); (Mihok and Murrant 2004). VO2 increased 3-fold over baseline 20 s following the onset of contraction, a time course consistent with modeled and published data (Kindig, Richardson et al. 2002, Jones, Grassi et al. 2011). Following the termination of stimulation, VO2 slowly returned to baseline over 120 s reflecting the accumulated oxygen debt from the 120 s of stimulation.
Several limitations exist when utilizing this technique to study VO2 on kinetics in whole muscle: 1) As VO2 significantly increases, PISFO2 may fall below the 40 mmHg threshold, which could decouple measured VO2 from energy production as glycolysis and lactate production increase, and 2) Measurements of VO2 lose accuracy as PISFO2 drops below 10 mmHg due to reduced collection of the linear ODC component. VO2 measurements during isometric contraction, as presented here for demonstration of the technique’s capacity to track VO2 and PISFO2 dynamics, need to be interpreted with their corresponding PISFO2 data, which may impose a physiologically valid restraint on the increase in VO2. An increase of the sampling rate to 2–10 Hz would greatly increase hypoxic accuracy after correction for photobleaching and is suggested for adapting this technique to various tissues and experimental designs. Additionally, for measurements where VO2 is limited by PISFO2, the 5 s × 15 s compression cycling might produce a cumulative effect of reducing O2 supply versus a situation where PISFO2 were recorded without compression. This may manifest itself as increased lactate production and/or decreases in PISFO2 values over time. Since the total amount of PISFO2 delivery interruption can be extracted from the tracing and recovery intervals, the impact of compression can be calculated and measurements corrected accordingly.
In conclusion, serial measurements of VO2 represent a robust method for assessing oxygen supply and delivery dynamics in thin tissues with minimal intrusiveness. It satisfies the need for better spatial and temporal resolution in animal models and is adaptable to applications such as: exercise on-kinetics, ischemia/reperfusion injury, compartment syndrome, and the impact of pharmacologic agents on microvascular performance and tissue respiration.
This work was supported in part by grant R01 HL18292 from the National Heart, Lung and Blood Institute.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.