INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

INTRACELLULAR FREE CA²⁺ concentration ([Ca²⁺]_i) not only plays an important role in excitation-contraction coupling in muscles but also in signal transduction (second messenger) modulating a wide range of cellular functions. Binding of free Ca²⁺ to a variety of control and regulatory proteins in the cytosol activates these signaling mechanisms. Usually the number of activated protein molecules is a rather steep function of the concentration of intracellular free Ca²⁺ (14, 36). Because intracellular Ca²⁺ is one of the main regulators of enzymes involved in energy metabolism, it must play an indirect but pivotal role in local blood flow regulation. The relationship between [Ca²⁺]_i and local flow, however, has not been established yet.

Early noninvasive optical attempts to determine [Ca²⁺]_i were based on Ca²⁺-dependent fluorescence of such rare natural dyes as aequorin and obelin and resulted in semiquantitative estimations of peak [Ca²⁺]_i in the range of 0.5 µM to several micromolar concentrations during activation of the tissue (10, 33). This technique, although subject to certain limitations (4, 11), was found to be very useful in a large number of cellular studies. Because of methodological problems, whole organ studies were extremely difficult to carry out with these dyes. Hence, ion-sensitive, intracellular microelectrode techniques were developed parallel to the aequorin fluorescence method to assess [Ca²⁺]_i in single cells (2), in whole organs like the brain (38), and in cardiac (18) and skeletal (8) muscles. The selectivity and sensitivity of the optical approach were substantially improved by the introduction of the 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid family of fluorescent dyes like quin 2, fura 2, indo 1, fluo 3 (15, 42), and especially their membrane-permeable acetoxymethyl esters (AM), which are cleaved and trapped inside the cells.

In the past decade, many fluorescent tracer studies were performed and have revealed a plethora of new information on [Ca²⁺]_i function. An overwhelming portion of these studies was carried out on isolated or cultured cells. The fura 2 and indo 1 ratiometric techniques were also adapted for isolated organ models like the intact perfused heart (5, 23), isolated skeletal muscle (3), and the isolated perfused liver (35).

Quantitative data obtained from single cell or cell culture experiments, however, can be translated to intact tissues or blood-perfused organs only with substantial uncertainties. Although perfused, isolated organs provide a more realistic model to study intracellular processes, lack of blood circulation and the absence of neurohumoral signaling and regulatory mechanisms still might induce substantial shifts in the intracellular ion milieu. In conclusion, although isolated or cultured cell and perfused organ studies provide indispensable information on the possible role and importance of [Ca²⁺]_i in signal transduction, its actual role and importance should be ultimately tested in intact organs in vivo.

The primary causes for paucity in the literature of [Ca²⁺]_i studies performed in vivo are methodological of origin, the most important one being loading of the cells with the fluorescent dye in intact blood-perfused tissues. Although the AM forms of fluorescent dyes readily cross cell membranes and are trapped inside the cell, most of the fluorescent tracers will bypass the targeted cells when loaded via the feeding artery.

Single wavelength (intensity) fluorometric techniques are very sensitive to the stability of the total indicator concentration in the sampled tissue volume, the constancy of nonspecific optical properties of the tissue, and last, but not least, the stability of the size and shape of the sampled volume. Neither of these conditions are strictly met in vivo (12, 24). Blood vessels in both the illumination and fluorescence light path will inevitably distort the measured signals due to changes in local blood volume and in the degree of hemoglobin oxygenation (6, 16). For decades investigators have developed different correction techniques to compensate for the most important errors caused by these artifacts (7, 12, 16, 20, 37, 44). Correction methods, however, were less successful when the contribution of optical tissue artifacts to the fluorescence signal was substantial. A more promising correction technique, based on multiwavelength or spectral information, was proposed by LaManna et al. (22). A significantly improved correction method was described by Koretsky et al. (21), who measured changes in NADH fluorescence of the isolated perfused heart and used the fluorescence signal emitted by an internal standard at an isosbestic wavelength to compensate for artifacts caused by motion and tissue density. Substantial differences in optical and chemical properties between internal standard and fluorescent indicator (solubility, compartmentalization, leakage, pH sensitivity, etc.), however, limit the use of such internal standards.

More recently dual wavelength excitation or emission ratio techniques have been developed to determine [Ca²⁺]_i (43). This approach is based on the use of two closely related fluorescence signals derived from the same tracer molecule. In this approach, when signal levels are high enough to give acceptable signal-to-noise ratio, absolute fluorescence amplitudes are of no critical importance, because the real information is derived from the ratio of amplitudes that is much less sensitive to these artifacts. Because of major differences in spectral properties of blood and blood-free tissue, however, this method is still sensitive to hemodynamic artifacts and may not be satisfactory for in vivo measurements.

In a previous approach we developed an intravital microscopic technique to quantitate changes in intracellular NADH levels in the exteriorized, blood-perfused rat spinotrapezius muscle (40, 41). Fluorescence was sampled in very small, principally avascular tissue volumes to minimize changes in apparent fluorescence due to hemodynamic artifacts. The findings in these studies indicate that at practically avascular tissue sites, it is indeed possible to avoid major hemodynamic artifacts and to assess real changes in NADH concentration.

In the present study, direct superfusion of the exteriorized, blood-perfused skeletal muscle of the rat was used as a simple effective way to load in vivo skeletal muscle cells with indo 1-AM. Once the dye is cleaved to indo 1 and trapped inside the cells, fluorescent images could be collected with a good signal-to-noise ratio using a sensitive video camera. With this imaging ratiometric approach, we were able to assess in vivo [Ca²⁺]_i levels, predominantly independent of common artifacts.

It was evaluated whether with this technique real time changes in [Ca²⁺]_i, as induced by simple interventions, could be followed in vivo. The ratiometric data with a temporal resolution on the order of 0.5 min were converted into [Ca²⁺]_i assuming a [Ca²⁺]_i of 100 nM for resting muscle and linearity of the ratio versus [Ca²⁺]_i relationship in the range from 80 to 800 nM as verified by an in vitro calibration curve. [Ca²⁺]_i levels were increased by inducing ischemia and applying a calcium ionophore (ionomycin) and decreased by means of manganate to completely quench the Ca²⁺-dependent fluorescence.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Fluorescence ratio-imaging methods require image acquisition at various wavelengths. By taking the ratio of two images, one obtains a measure of activity independent of the actual individual fluorescence intensities. The use of a rotatable filter wheel equipped with the appropriate interference filters provides a simple way for serial image collection at multiple wavelengths.

Measuring Technique

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Schematics of measuring system. Fluorescence microscope with illumination unit consisting of a Hg-arc lamp (A), field diaphragm (B), filter box (C), and condenser (D); and imaging unit consisting of objective lens (F), beam switch (G), filter box with a rotating filter wheel (H), video-rate CCD cameras (I), and special cooled slow scan DCCD camera (J). The latter is controlled by an IBM compatible personal computer. Exteriorized spinotrapezius muscle is mounted in a temperature-controlled transparent chamber (E) held by a pedestal on the microscope stage.

The ratio method requires a camera with very low noise characteristics. This requirement is fulfilled by the computer-controlled, highly sensitive, cooled, slow-scan DCCD camera, which has excellent stability, linearity over five orders of magnitude, close to uniform spatial and temporal sensitivity, and extremely low noise (0.6-2 electrons/pixel · s, at 50°C). The sensitivity of the camera makes the use of an image-intensifier unit unnecessary. Images (512 × 512 pixels) are digitized by a 16-bit analog-to-digital (AD) converter running at a sampling rate of 150 kHz. The 16-bit resolution of the AD converter enables a dynamic resolution of better than 1:10,000 following background correction. The signal-to-noise ratio can be significantly improved by collecting the signal for an extended time period (1-60 s). The wide dynamic range of the DCCD camera was utilized by using a sampling period of 0.01 s for topological (brightfield) images and 10 s for fluorescence images to obtain an appropriate signal-to-noise ratio. Because of this relatively slow-image acquisition, however, this configuration can only be used if changes in tissue fluorescence are slow compared with the acquisition times.

Preparation and Procedure

The left spinotrapezius muscle was carefully exteriorized on one side and mounted in an optically transparent perfusion chamber, leaving the vascular and nervous supply of the muscle intact. This muscle contains mixed fibers, both glycolytic and oxidative types, which are present approximately in the same percentage. Another important feature of this muscle is its arcading arteriolar system, which compensates for ligation of one major feeding artery during the exteriorizing process. The muscle was stretched to its approximate resting length before it was secured in the chamber. In the chamber this muscle preparation still has a close-to-normal circulation and metabolic state (40).

A modified gelatin-containing Krebs-Henseleit (GKH) solution was used to perfuse the chamber. The composition of the solution was (in mM) 131.9 NaCl, 4.7 KCl, 22 NaHCO₃, 1.17 MgSO₄, and 2.0 CaCl₂, with 1% gelatin. The superfusion solution was equilibrated with a mixture of 95% N₂-5% CO₂, and its temperature was kept at 35°C. All drugs were dissolved in GKH.

The microcirculation of the exteriorized muscle was microscopically inspected with the use of the video-rate CCD camera. Only muscle preparations with good quality microcirculation, as judged by the presence of intermittency of capillary flow, the absence of areas without perfusion, the absence of microhemorrhages, and the visibility of cross striations in the skeletal muscle cells were used. An equilibration period of ~30 min was allowed, during which an area of the midportion of the muscle was selected without major vessels in its cross section to avoid interference with hemoglobin (hemodynamic artifacts) caused by significant blood volume changes in the region of interest.

The fluorescent probe indo 1-AM (Molecular Probes) was dissolved first in dry dimethyl sulfoxide in a 1 mg/ml concentration with the addition of Pluronic F-127 (10%) and then sonicated into GKH solution to obtain a final concentration of 35 µM. For in vitro calibration, the water-soluble pentapotassium salt of indo 1 was used. Ionomycin (Sigma) was prepared similarly to indo 1-AM in a final concentration of ~15 µM. The quenching solution contained 5 mM MnCl₂ dissolved directly in saline, containing NaCl (0.9%) to avoid precipitation of calcium in GKH.

Serial fluorescent images of a small preselected region (~320 µm × 320 µm) of the muscle were collected by the cooled digital CCD camera at wavelengths of 400, 506, and 465 nm. From this region, well-defined representative regions of interest (ROI) containing no major vessels and only a few small ones (capillaries, used as landmarks) were selected from the topological image for quantitative evaluation of the 400 nm/506 nm fluorescence ratio. The edge of the ROI corresponded to the edge of muscle fibers. All statistical analyses of the fluorescence and transmittance intensities were performed on this ROI. The size of the ROI was generally <100 µm × 100 µm. From each individual fluorescence and ratio image, the mean and SD values were calculated. Coefficients of variation were typically <10%. Before the 400 nm/506 nm ratio images were computed, single wavelength fluorescence images were corrected for tissue autofluorescence by pixel-to-pixel subtraction of the dye-free fluorescence images. Ratio images were then computed by pixel-by-pixel division of the corrected images. Numeric values of the fluorescence ratio were calculated by dividing the mean values as calculated from the 400- and 506-nm image pairs. To compensate for NADH-increase-induced autofluorescence shifts, the approximate NADH changes, as obtained from the dye-free experiments at these two wavelengths, were used for corrections when appropriate. Because only negligible NADH-induced shifts in tissue autofluorescence are predicted from the emission spectrum of NADH at 400 nm, corrections were only performed at 506 nm.

All fluorescence signals collected in dye-loaded tissues represent a sum of Ca²⁺-dependent indo 1 and Ca²⁺-independent background autofluorescence. Images captured at 400 and 506 nm were used for ratiometric computations, whereas images collected at 465 nm, which is an isosbestic wavelength for Ca²⁺-dependent changes, were used to estimate the effect of "dye-leakage." The influence of NADH level changes on fluorescence intensities was also determined in several "nonloaded" experiments (see below for explanation).

Before the tissue was loaded with dye, a brightfield image of the basic topology of the selected area was captured at a 400-nm illumination wavelength (topology). Then the illumination was switched to 365 nm, and a set of serial fluorescence images of the same site was collected at wavelengths of 400, 506, and 465 nm (dye free). The filters were moved manually with image-capturing times of 10 s each, resulting in a total capturing cycle of ~1.5 min.

During the experiment, a similar set of serial images was captured at the end of each intervention to follow [Ca²⁺]_i kinetics. After the collection of the first (dye free) set of serial images, the dye was infused into the tissue chamber at a rate of 0.5 ml/min for 10 min. Then the infusion was stopped and the muscle bathed in the loading solution for another 20 min. Loading (i.e., superfusion of the tissue with indo 1-AM) hence lasted for 30 min. During the first 15 min of loading, the fluorescence signals reached almost their maximal level, and very little loading took place during the second 15 min. Then the nonloaded indo 1-AM was removed by superfusion with dye-free GKH solution. Washout was terminated after ~15 min when the fluorescence signals had stabilized. At that instant the various protocols were started.

Experimental Groups

Group 1 (dye + interventions): After stabilization of the fluorescence signal, a second set of serial images was collected (Loaded). After image collection, ischemia was induced by completely arresting flow in the main feeding artery in the proximal portion of the muscle, using a simple occluder mounted on a micromanipulator. Completion of stopflow was verified with the video-rate camera. At the end of the ischemic cycle (30 min), a third set of serial images was collected (Ischemia). The occluder was then removed and reperfusion started. Immediate recovery of the muscle microcirculation was verified. At the end of the reperfusion period (20 min), a fourth set of serial images was collected (Reperfusion). Next the muscle was superfused with ionomycin (15 µM) containing GKH. After about 15 min, a fifth set of serial images was captured (Ionomycin). Finally, MnCl₂ (5 mM) in saline was superfused (for ~5 min) to quench indo 1 fluorescence to recover dye-free autofluorescence, and the last set of serial images was collected (Manganate).

Group 2 (no dye + interventions): In this group, a superfusion solution containing solvent but no dye was used ("nonspecific autofluorescence changes" tests). All perturbations (ischemia, reperfusion, ionomycin, and manganate treatment) were performed and images were collected as in group 1.

Group 3 (dye + no interventions): In this group, no interventions were performed following dye loading and washout ("dye-leakage" tests). Fluorescence and transmittance images were collected at 0, 5, 15, 30, 60, 90, and 120 min after the termination of the loading process. These instants approximate the protocols in the other two groups.

Calibration

Image Processing

Statistics

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

A topological (brightfield) image of the preselected area of a muscle is shown in Fig. 2. Microvasculature and single muscle fibers with their cross striations can be clearly distinguished. Distinct vascular structures were used as landmarks to identify the ROI.

View larger version (187K): [in this window] [in a new window]   Fig. 2.   A typical topological (brightfield) image. Size of the area of muscle surface shown (i.e., field of view, FOV) is ~320 µm × 320 µm; image was captured at 0.01-s shutter speed. Both illumination and detection wavelength were 400 nm. Rectangular dotted box in midportion of image is region of interest (ROI) selected for final image processing.

A set of single wavelength images collected in the same experiment and related ratio images are shown in Fig. 3. These pictures were individually photographed from the monitor screen. Consequently, their brightness cannot be considered quantitative. The native (dye free) autofluorescence images of the preselected area are shown in Fig. 3, row A. All images were corrected against "closed-shutter background." Neither of the images in Fig. 3, however, were corrected for changes in tissue autofluorescence (i.e., NADH redox changes). In Fig. 3, row B, single wavelength fluorescence and ratio images are shown in the dye-loaded state (at the end of the washout period). Fluorescence intensity in the images was significantly increased compared with the dye-free state. In the resting skeletal muscle, the fluorescence ratio image and consequently [Ca²⁺]_i is relatively uniform from fiber to fiber.

View larger version (144K): [in this window] [in a new window]   Fig. 3.   Single wavelength fluorescence images from same experiment as shown in Fig. 2. Pictures were photographed individually from the computer screen. Muscle was illuminated at 365 nm. Sets of serial images were collected at 400, 506, and 465 nm (shown from left to right). Shutter speed was set to 10 s. Single wavelength images were corrected against "closed shutter background." Neither of the images in Fig. 3 was corrected for "dye leakage" or "nonspecific" changes in tissue autofluorescence (i.e., NADH redox changes) or for nonspecific changes in tissue absorbance (i.e., myoglobin redox changes, etc.). Ratio images (on right side of each row) were calculated by pixel-to-pixel division of (background and autofluorescence corrected, but NADH uncorrected) images collected at 400 and 506 nm, respectively. A single-step smoothing routine was performed on all images, using a 3 × 3 mask with the following weights: 0.25 (central pixel), 0.125 (four pixels on the side), 0.0625 (four pixels in the corners). Identical intensity (gray) scales were used for images, captured at same individual wavelengths, and for ratio image set (i.e., images in columns). Image rows represent images collected/calculated in dye-free (A; tissue autofluorescence) and loaded (B; control) state, at end of ischemia (C) and reperfusion periods (D) and after ionomycin treatment (E).

Pictures taken at the end of the ischemic and the reperfusion period are shown in rows C and D of Fig. 3, respectively. During 30 min of ischemia, fluorescence increased slightly at 400 nm but decreased significantly at 506 nm, resulting in a change in the ratio image. These changes indicate a substantial shift in [Ca²⁺]_i. The relatively even fluorescence ratio increase reflects a rather homogeneous rise in [Ca²⁺]_i. During the reperfusion period, fluorescence intensities at both 400 and 506 nm decreased with about the same extent, suggesting that changes in [Ca²⁺]_i are relatively moderate. This incomplete recovery of [Ca²⁺]_i during the 20-min reperfusion period is even better revealed by the ratio image.

Images captured after ionomycin treatment are shown in Fig. 3, row E. After ionomycin treatment, the substantial increase in [Ca²⁺]_i as indicated by the ratio image is caused by a moderate increase at 400 nm combined with a decrease at 506 nm. Manganate treatment eliminated most of the fluorescence (not shown). Intensities at both 400 and 506 nm returned to dye-free autofluorescence levels. This complete quenching effect of the manganate on the fluorescence signal above the background is direct proof that indo 1 is the ultimate source of tissue fluorescence detected above dye-free levels.

Figure 4 shows the apparent (uncorrected) tissue fluorescence levels at the three wavelengths for the experiments in group 1, as related to the background autofluorescence as measured in the nonloaded state. In dye-loaded muscles, fluorescence at 400 and 506 nm rose by ~80 and 440%, respectively. The increase at 465 nm was even higher (~500%). Tissue fluorescence at all wavelengths showed a trend of decreasing in time. This is probably due to a time-dependent decrease in intracellular dye concentration (leakage). During interventions, fluorescence changes at the Ca²⁺-dependent wavelengths (400 and 506 nm) were similar to those described in Fig. 3, indicating significant changes in [Ca²⁺]_i during perturbations.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Summary of apparent single wavelength fluorescence levels obtained in group 1 experiments. Illumination of tissue was at 365 nm. Fluorescence intensities are shown related to tissue autofluorescence (dye-free levels). Actual levels at 400, 465, and 506 nm are not only dependent on tissue intracellular Ca2+ concentration ([Ca2+]i) but also on intramitochondrial [NADH] and actual tissue dye concentration, steadily decreasing in time as a consequence of both passive leakage and active ionic pump. * Significantly different from loaded state, P < 0.01, Student's t-test.

Tissue fluorescence determined at the isosbestic wavelength (465 nm) for Ca²⁺-dependent changes also steadily decreased in time. During ischemia a relative increase and after ionomycin treatment a relative decrease was observed at this wavelength. Time-independent shifts in fluorescence at this wavelength are most probably due to changes in tissue NADH concentration, a large elevation during ischemia and a significant decrease following ionomycin treatment. The relative importance of these shifts in the experiments in group 1 can be readily estimated from the experiments in group 2 (see Fig. 5). The ischemia-induced rise at 465 nm is ~55% of the autofluorescence, whereas the ionomycin-induced fall in autofluorescence at the same wavelength is close to 20%. Compared with total intensity changes, however, it is evident that this magnitude of autofluorescence shifts can cause only moderate, but not necessarily negligible, shifts in the total fluorescence of dye-loaded cells.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Summary of apparent single wavelength fluorescence levels obtained in group 2 experiments. Conditions and perturbations were identical as in group 1, except that only solvent was used. Loading caused an ~20% nonspecific decrease at all wavelengths. Ischemia induced a small increase at 400 nm, but a significant (probably NADH-dependent) rise was found at 506 and 465 nm (40 and 55%, respectively). Reperfusion completely reversed this effect. Ionomycin treatment induced both a nonspecific signal decrease (~15%) and NADH-oxidation (~15%). These values can be used to correct for NADH-dependent shifts in dye-loaded muscles. * Significantly different from loaded state, P < 0.01, Student's t-test.

Figure 6 summarizes the ratiometric data obtained in the experiments in group 1 on indo 1-loaded uncorrected fluorescence, background autofluorescence (estimated from the experiments in group 2), and the indo 1-loaded NADH-corrected fluorescence during the various interventions. Because the dye-free ratio image does not provide Ca²⁺-related information, we used the ratio values calculated for the loaded state as 100% reference levels for all other ratio calculations.

View larger version (23K): [in this window] [in a new window]   Fig. 6.   Ratio values obtained in indo 1-loaded and nonloaded muscles (group 1 and group 2 experiments). For loaded muscles, both apparent and corrected values are shown. Ratio values are normalized to those obtained in loaded state. Ischemia induced a substantial (~100%) increase in fluorescence ratio and consequently in [Ca2+]i. In contrast to NADH shifts, this increase was not completely reversed even after 20 min of reperfusion. In that instant, [Ca2+]i was still significantly elevated in muscles. Ionomycin caused a large (at least 10-fold) increase in the ratio values (and consequently in [Ca2+]i), which cannot be simply quantitated because these values are on the saturating part of the in vitro calibration curve. Effect of NADH level shifts can be evaluated by comparing uncorrected and corrected values. The larger the intracellular dye concentration is, the smaller (but in vivo not negligible) the effect of NADH redox state on the ratio value. * Significantly different from loaded state, P < 0.01, Student's t-test.

The average indo 1 fluorescence ratio in the control state (Loaded) was 0.20 ± 0.05. Ischemia caused a significant rise of ~80% in the apparent ratio to 0.35 ± 0.07. During ischemia, however, a Ca²⁺-independent increase in tissue autofluorescence at both 400 and 506 nm due to a mitochondrial NADH shift has to be considered. A satisfactory correction for this NADH-related change can be achieved by using an estimate of the dye-free data for autofluorescence increase (see Fig. 5). The NADH-corrected ratio values in Fig. 6 were calculated based on the estimation of a 40% increase in autofluorescence at 506 nm. The effect of NADH changes on the 400-nm fluorescent signal was found to be only ~2-5%, and no corrections were performed at this wavelength. After the correction for the NADH-induced autofluorescence shifts, the calculated ischemiainduced rise in [Ca²⁺]_i was found to be slightly but not significantly higher compared with the noncorrected value (0.41 ± 0.05). During the reperfusion period, the fluorescence ratio moderately decreased to 0.29 ± 0.06 compared with the ischemic situation, but the ratio values obtained were still significantly (~50%) above the control values. Because NADH fluorescence completely returned to the control level during the first minutes of reperfusion, no correction for NADH changes was needed at this point. Ionomycin treatment induced a very large increase in the fluorescence ratio, close to 1,000% of control (1.98 ± 0.31). However, because ionomycin also caused a significant (~15%) fall in tissue autofluorescence at 506 nm (see Fig. 6), most probably due to decreased tissue NADH levels, these ratio values tend to overestimate [Ca²⁺]_i and, like the ischemic ratio values, have to be corrected. The NADH-corrected ratio values (1.54 ± 0.23) showed a smaller but still pronounced increase in [Ca²⁺]_i, of at least 750%, following ionomycin treatment. However, because ionomycin also causes a significant (10-30%) nonspecific decrease in tissue optical density, the corrected ratio values are still only semiquantitative estimations of the magnitude of [Ca²⁺]_i changes.

The in vitro calibration between [Ca²⁺]_i and fluorescence signal ratios indicated quasilinearity of the system in the range between 80 and 800 nM [Ca²⁺]_i. This quasilinearity can be well utilized for quantitative evaluations of absolute [Ca²⁺]_i levels during perturbations. Assuming an average [Ca²⁺]_i of 100 nM in the resting skeletal muscle, calculation of the average [Ca²⁺]_i leads to an NADH shift-corrected estimate of 201 ± 13 nM after 30 min of ischemia, 143 ± 31 nM at the end of 20 min reperfusion, and at least 753 ± 111 nM after ionomycin treatment.

To evaluate differences between individual muscles in dye-leakage rate to calculate time factors for the kinetics of the process and to get a clue on possible dye-removing mechanisms, all individual "leakage" curves measured at the isosbestic 465 nm were plotted on a linear scale against time (not shown). Leakage rate was higher immediately after the washout process, probably reflecting removal of nonloaded interstitial dye from the extracellular space. After ~30 min, the decrease in dye fluorescence was steady. This second slower decay in fluorescence intensities is more likely due to the slow removal of the intracellular indo 1 pool. Both monoexponential and proportional approximations seem to fit well to the curves except for their first 30-min portion. The linear approximations for mean values at all wavelengths provided a half-life time of ~1.6 h for dye leakage (~30% decrease/hour). With the use of a monoexponential fit, the estimated dye-leakage rate was slightly lower (~25% decrease/h).

Figure 7 shows the stability of the ratio value for these experiments over a 2-h period without interventions. As can be seen in Fig. 7, in contrast to the significant decrease in intracellular dye concentrations and as indicated by the steadily falling fluorescence signals, only minor shifts in the calculated ratio values, and, hence, in [Ca²⁺]_i can be observed.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   In contrast to the major time-dependent dye concentration changes, stability of calculated ratio values is excellent. This figure also proves that the signal-to-noise ratio of fluorescence intensity measurements by the cooled slow scan detector (DCCD) is good enough to enable reliable ratio value determination even at relatively small tissue indo 1 concentrations, when Ca2+-dependent fluorescence signals are only fractions (30-50%) of tissue autofluorescence intensities. These findings can be especially important in long-term protocols when neither use of ionic pump blockers nor reloading muscle is acceptable.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The present study demonstrates that in vivo skeletal muscle cells can be loaded with a Ca²⁺-sensitive, cell-permeable dye, indo 1-AM, by superfusing the exteriorized blood-perfused spinotrapezius muscle. The combination of intravital microscopy and the indo 1-AM ratiometric fluorescence method proved to be a valuable tool to study [Ca²⁺]_i at the single fiber level. Relatively slow physiological or pathological changes in [Ca²⁺]_i can be continuously monitored in this preparation.

This method only requires a regular fluorescence microscope and a highly sensitive, cooled DCCD camera. The only important limitation is its relatively slow acquisition speed; at low light levels it takes several seconds to capture and transfer good-quality fluorescence images.

In the indo 1 fluorescence ratio measurements, the excitation wavelength of 365 nm is relatively far from the peaks of the absorption spectra of the dye (349 nm for free and 331 nm for Ca²⁺-bound dye). Therefore, the use of 365 nm for excitation results in an apparent higher sensitivity of the fluorescence signal measured at 506 nm (free dye) than at 400 nm (Ca²⁺-bound dye). This has the advantage that no special (and expensive) ultraviolet optics are required in the microscope system. Although excitation of the tissue at 365-nm wavelength decreases absorption quantum efficiency, it will also decrease harmful ultraviolet load of the tissue.

Perfusing the chamber with dye containing GKH for ~30 min was found to be sufficient to load the muscle cells. After removal of unbound dye, tissue fluorescence had stabilized at a level approximately one- to sevenfold of autofluorescence. During the actual experiment, only a slow but permanent decrease was observed at all wavelengths, mainly caused by leakage and to a lesser extent by photobleaching. Despite this leakage, resulting in a decrease in intensity of ~70% after 2 h, the data presented in Fig. 7 clearly indicate that even this relatively low intracellular level is high enough to enable reliable ratio determinations. Dye leakage may be stopped by blocking the anion carrier activity responsible for leakage with probenecid or sulfathiazole. The use of such drugs, however, might alter the metabolism of muscle cells, making the interpretation of the fluorescence signals more difficult.

In vivo fluorescence studies are intrinsically sensitive to multiple optical artifacts such as tissue movements, autofluorescence, inner filter effect, bleaching, dye concentration changes (leakage), and blood volume changes in the sampled tissue volume (hemodynamic artifact) (12, 16). To yield reliable quantitative data, these artifacts should be eliminated, or at least their overall effect should be minimized to an acceptable level.

Movements of the resting spinotrapezius muscle are only caused by breathing of the animal. Its effect is decreased significantly by proper mounting of the preparation. Residual movements can be partially compensated for by careful topology-based (best fit) selection of the region of interest. Because collection of a single fluorescence image took ~10 s, some movement artifact might still be present in our images.

The relatively low fluorescence intensities of the dye-free images at both 400 and 506 nm (Fig. 3, row A) suggest that at these wavelengths autofluorescence in the muscle is moderate but cannot be neglected. Autofluorescence at 465 nm, the peak of the NADH emission spectrum, is somewhat higher. Therefore, the most important cause of the changes in tissue autofluorescence at 400 and 506 nm wavelengths is the varying NADH/NAD(P)H content of the cells. The contribution of other sources of tissue fluorescence (like collagen) should be minor considering the changes in the signal (40). Because levels of NADPH in the muscle are low, we considered NADH as the "sole" source of autofluorescence. Because the quantum efficiency of the membrane-bound (mitochondrial) NADH is about two orders of magnitude higher than that of free (cytoplasmic) NADH, the autofluorescence mainly represents mitochondrial, membrane-bound NADH (29, 31). Consequently, beside the time-dependent leakage of the dye, the fluorescence-intensity changes at these wavelengths are caused mainly by [Ca²⁺]_i changes. Because these changes may also be subject to changes in tissue NADH levels, the ratio images may be influenced by these changes. Corrections at 400 nm were not performed, because the emission spectrum of NADH predicts negligible changes at this wavelength.

The effect and significance of the NADH fluorescence changes in the single wavelength images were investigated in the images captured at 465 nm, because this wavelength is isosbestic for the [Ca²⁺]_i-dependent (indo 1) fluorescence. In the present study, the changes in indo 1 fluorescence were generally substantially greater than the calculated changes in NADH fluorescence. Loading the tissue with large quantities of indo 1 would obviously negate the influence of NADH fluorescence. This, however, is not desirable because of the [Ca²⁺]_i-buffering effect of the chelator dyes.

The myoglobin content of the muscle fibers is most likely responsible for the "nonspecific" changes in absorption (optical density) in the avascular tissue (i.e., in the magnitude of the inner filter effect). Shifts in the redox state of myoglobin will inevitably alter the apparent absorbance of the tissue (13). From our previous study (40), this effect can be estimated to be about 10% for NADH fluorescence, which is the autofluorescence in the present study. In indo 1-loaded cells, this artifact is likely to be lower because of its larger signal.

The inner filter effect is steeply (exponentially) increasing with increasing optical pathlength; thus the overall magnitude of this effect is likely to be limited in this model because the muscle is very thin (only 200-300 µm). As a consequence of loading by superfusion, the dye concentration may be larger in the superficial than in the medial cell layer. Therefore, the fluorescence reaching the camera is probably mostly emitted from the uppermost cell layers.

The leakage of dye from the cells has a considerable effect on the fluorescence signal at both 400 and 506 nm (Fig. 4) but had very little effect on the ratio value, as can be concluded from the data presented in Fig. 7. This excellent stability of the ratio, despite significant changes in single wavelength fluorescence intensities, suggests that the calculated numerical values for the ratio in our experiments are not sensitive to the substantial decrease in absolute dye concentration in muscle cells, at least during the time period of observation of ~2 h. The time constant for a decrease in dye concentration (~1.6 h), determined from these experiments, is in good agreement with the time constant derived from the experiments in group 1 but is substantially longer than the values obtained in isolated liver studies (35).

A substantial portion of the leakage is a consequence of basic physiological processes like passive diffusion and active anionic pumping. In many in vitro studies, probenecid has been used to block this anionic pump (28, 30, 45). Because the effect of indo 1 concentration changes can be nearly completely eliminated by the ratiometric method, we did not block the pump at this point.

In principle, the decay of the fluorescence signal could be explained by photobleaching. Ruttner et al. (35) have studied this phenomenon in a perfused liver model and found that by blocking the anion carriers by probenecid, the half-life time of the dye in the cells increased about threefold (from 18 to 60 min), demonstrating that the most important mechanism responsible for the decay of the fluorescent signals is indeed an active pump. Another indirect proof for the effect of bleaching being moderate is that in our experiments relatively low and intermittent excitatory light levels were used and that changing the exposure protocol had little effect on the decrease in signal, i.e., changing the length of illumination from <10 to 30% of the total time had little effect on the decay of the fluorescence signals (at the isosbestic 465 nm 33 and 30%/h, respectively).

Because no direct calibration technique for intact organs is as yet available, in the present study calibration was performed in vitro, using ultraviolet cuvettes. Because there are significant differences between transparent Ca²⁺ solutions and opaque tissue layers in most optical variables (e.g., density, inner filter effect), numerical values for fluorescence ratios, obtained from this calibration process, cannot be directly used for calculations of in vivo [Ca²⁺]_i changes. Because the concentration range in which the ratio of the fluorescence intensities as determined is linearly related to the Ca²⁺ concentrations in the samples, this range can be assessed. The results of the calibration confirm the existence of a close to linear relationship in the [Ca²⁺]_i range of 80-800 nM.

It is difficult to use an established reference value for resting skeletal muscle [Ca²⁺]_i (1, 17, 26, 27, 46). The [Ca²⁺]_i during the various interventions were calculated assuming a [Ca²⁺]_i of 100 nM [a frequent value found in the literature (e.g., 25)] for resting skeletal muscle. The calculated shift in [Ca²⁺]_i to ~200 nM during ischemia reflects the possible role of calcium in mediating ischemic cell injury. A number of investigators agree that ischemia and/or reperfusion elicits a substantial increase in [Ca²⁺]_i in skeletal and cardiac muscles (9, 19, 32, 34, 39). The detailed mechanism of the kinetics of this Ca²⁺-homeostasis disturbance, however, is as yet unclear. The ability to perform [Ca²⁺]_i determinations in vivo is likely to be an asset in evaluating the [Ca²⁺]_i alterations during ischemia and/or reperfusion in this situation.

During postischemic reperfusion, [Ca²⁺]_i decreased to an estimated level of 145 nM, indicating that the long-lasting increase in [Ca²⁺]_i during reperfusion is not unique to cardiac muscle and that similarly elevated postischemic [Ca²⁺]_i levels occur in skeletal muscle. During ionomycin treatment, [Ca²⁺]_i increased to at least 750 nM, which is well above the maximal physiological levels in resting muscle. Because our in vitro calibration exhibits saturation at such high [Ca²⁺]_i levels, the ionomycin-induced [Ca²⁺]_i increases may be well underestimated by the simple linear approximation. Because ionomycin causes large heterogeneous shifts in nonspecific optical properties of the cells as well (e.g., apparent density changes), which can hardly be estimated, quantitative [Ca²⁺]_i calculations are further compromised under these circumstances. After manganate treatment, fluorescence returned to the approximate autofluorescence level, which is compatible with the quenching properties of 5 mM manganate for indo 1-dependent fluorescence.

A simple way to estimate indo 1 fluorescence at any wavelength is to subtract the dye-free autofluorescence intensity from the actually measured intensity. As can be seen in Fig. 5, however, ischemia induced an NADH-dependent rise in tissue autofluorescence at 506 nm. This NADH rise should be corrected for in indo 1-loaded muscles before calculation of ratio values for the postischemic state. The correction step for an estimated 40% increase in tissue autofluorescence at 506 nm resulted in only a minor change in ratio values.

The present study was confined to measurements at rest. The method described, however, can be extended to the assessment of [Ca²⁺]_i in contracting muscle, using gated illumination. In this way the effects of changes in optical density can be avoided.

In conclusion, this study demonstrates that the indo 1 ratiometric technique, in combination with a quantitative intravital microscopic approach, enables the assessment of [Ca²⁺]_i in vivo and the evaluation of changes in [Ca²⁺]_i during interventions.