INTRODUCTION

Top Abstract Introduction Results Discussion References

RECENTLY, WE DEMONSTRATED that microspectrophotometry combined with digital image processing enables quantitation of the fractional oxygen binding of myoglobin (S_Mb) with a subcellular spatial resolution in a single individual cardiomyocyte of the rat (24). Using this technique, we demonstrated significant radial PO₂ gradients in the intracellular space when the mitochondrial respiration was increased approximately eightfold by the use of an uncoupler of oxidative phosphorylation (24). Because of the considerable intracellular PO₂ (PO_2,i) gradients, the core of the cell appeared severely hypoxic, whereas myoglobin (Mb) near the sarcolemma is almost fully saturated with oxygen. The important question that remains to be addressed is whether these radial PO_2,i gradients limit the mitochondrial oxidative phosphorylation.

In the present study, we sought to visualize the possible intracellular heterogeneity of mitochondrial oxygenation in a single rat cardiomyocyte in exact same condition as we demonstrated the radial PO_2,i heterogeneities. Blue autofluorescence emitted from the heart on epi-illumination of ultraviolet (UV) (~360 nm) light can be a good index of mitochondrial reduced nicotinamide adenine dinucleotide (NADH) and reduced nicotinamide adenine dinucleotide phosphate (NADPH) (5, 7, 14, 19). In addition to the whole organ, mitochondrial NAD(P)H fluorescence has been demonstrated in an isolated individual cardiac myocyte (8, 9, 11, 20, 26). NADH is the primary source of reducing equivalent for the respiratory chain that is finally oxidized by molecular oxygen at the terminal enzyme of the respiratory chain, cytochrome aa₃. Consequently, occurrence of NADH in mitochondria tells us whether the diffusional oxygen supply to mitochondria is adequate to match the cellular oxygen requirement (3, 4, 22). Hence, NADH fluorescence measured with a subcellular spatial resolution would reveal the intracellular heterogeneity of oxygen supply to mitochondria.

In the present study, to extend our knowledge about regulation of diffusional oxygen supply to cardiac mitochondria, we quantitated NAD(P)H fluorescence with a subcellular spatial resolution in a single individual cardiomyocyte of the rat. We found significant heterogeneities of the autofluorescence in the intracellular space that mirror image the radial PO_2,i gradients observed in the previous experiment.

Methods

Fifteen microliters of the cell suspension (~10⁵ cells/ml) were placed on the glass slide of an air-tight measuring cuvette. The depth of the suspension medium was ~190 µm. The measuring cuvette provides inlet and outlet ports for superfusion gas. A computer-controlled gas blender consisting of mass-flow controllers (STEC-320, STEC, Kyoto, Japan) supplied mixed gas, via a humidifier, to the inlet port of the cuvette at 2 ml/min. Oxygen concentration of the mixed gas was continuously monitored at the outlet port using a polarographic oxygen electrode (model 17026, Instrumentation Laboratory, Lexington, MA), which had been calibrated to humidified standard gases.

Eng et al. (8) demonstrated that intact single isolated cardiomyocytes of the rat show a broad fluorescence spectrum centered at ~450 nm on UV excitation. The spectrum was quite similar to that of the isolated rat heart mitochondria. We carried out microfluorometry using a BH2-RFCA microscope (Olympus, Tokyo, Japan) equipped with a 100-W mercury lamp and a dichroic filter (360-390 nm/>430 nm excitation/emission wavelength). Autofluorescence emitted from an individual cardiomyocyte was captured using a 10-bit cooled digital charge-coupled device (CCD) camera (C4742, Hamamatsu Photonics, Hamamatsu, Japan). Because the fluorescence signal of the individual cell was weak, the gain of the video amplifier was set at maximum and the exposure time was also set at maximum (1 s). We consecutively captured five fluorescence images from one cell and integrated them to improve the signal-to-noise ratio. The digitized image was stored in a computer for future image processing. Frame size was 640 × 480 pixels, and one pixel corresponded to 0.11 µm with the use of a ×40 objective lens (numerical aperture = 0.55, LWDCDPlan ×40, Olympus, Tokyo, Japan).

We first evaluated photo bleaching of the cellular fluorescence. The cell was incubated with 2 mM NaCN under nonlimiting oxygen (i.e., superfusion by air) to significantly augment the NADH fluorescence (8, 11). We measured the autofluorescence of a cell three times, at 3-min intervals. The UV excitation light was blocked between each measurement by a shutter. Total UV-exposure duration for a cell was 15 s (1 s exposure × 5 frames integration × 3 measurements). We also measured fluorescence intensity of the adjacent region without any cellular material and designated it as the blank image. Cellular autofluorescence intensity was determined as the difference from the blank image. Extracellular Ca²⁺ concentration was 1.0 mM.

Second, we examined changes in the fluorescence to pharmacological interventions of the respiratory chain under nonlimiting oxygen (superfusion by air). Sodium cyanide blocks oxidation of NADH by inhibiting electron transport of the respiratory chain, thus maximizing the NADH fluorescence (8, 11). In contrast, the uncoupler of oxidative phosphorylation maximally increases NADH oxidation, thus significantly decreasing NADH fluorescence (8, 20, 26). We compared the NAD(P)H fluorescence of untreated (control), 2 mM NaCN-treated, and 1 µM CCCP-treated cells in either the absence or presence of extracellular Ca²⁺ (1.0 mM).

Finally, we examined heterogeneities of NAD(P)H fluorescence within a single cardiomyocyte with a subcellular spatial resolution. In the previous study, we demonstrated significant radial PO_2,i gradients when superfusing 1 µM CCCP-treated cells with 2.09 or 3.14% O₂ gas at 27°C (24). The volume-averaged S_Mb was similar to that in vivo at normal and elevated oxygen consumption (S_Mb = 0.5 ~ 0.7) (6, 10, 29). We conducted the present NAD(P)H fluorescence measurements exactly in the same condition. Namely, the cell was incubated with 1 µM CCCP and superfused with a gas mixture containing either 2.09 or 3.14% O₂ in N₂ at 27 ± 1°C. Additional measurements were conducted using superfusion gas containing 1.00, 1.50, 2.62, 4.18, or 5.23% O₂ in N₂. The cell was first superfused with a gas mixture of the desired O₂ concentration, and autofluorescent measurement was carried out (1 s exposure and 5 frames integration). The shutter was then closed and the superfusion gas switched to 99.999% N₂ gas. After 3 min, the second measurement of the fluorescence was conducted. Finally, the blank image was obtained, and the cellular NAD(P)H fluorescence was determined as the difference from the blank image. To quantitate the microheterogeneities of the NAD(P)H fluorescence, the first NAD(P)H fluorescence image was divided by the second (anoxic) NAD(P)H fluorescence image. This procedure normalizes regional variations of mitochondria/pyridine nucleotides density. With an assumption that the photo bleaching is uniform all over the cell, we did not compensate the fluorescence ratio image for photo bleaching. To examine whether the NAD(P)H fluorescence heterogeneities depend on oxygen flux in the cell, the same experiment was repeated in the control (without treatment with CCCP) cell. Fractional oxygen concentration of the superfusion gas for this experiment was 0.25, 0.50, 0.96, 1.05, 2.09, 3.14, or 4.18%.

	RESULTS

Top Abstract Introduction Results Discussion References

Photo bleaching of the cellular autofluorescence was evaluated by triple 5-s exposures to the UV light in 2 mM NaCN-treated cells. The autofluorescence decreased by 17 ± 6 and 31 ± 6% at the second and the third exposures, respectively (means ± SD; n = 9). From these results, we believe that calculations of the fluorescence intensity from more than three consecutive measurements in one cell are not appropriate because of the low signal-to-noise ratio resulting from photo bleaching.

Changes in the cellular autofluorescence to pharmacological alterations of mitochondrial redox state were examined at nonlimiting PO₂. Fluorescence intensities (in arbitrary unit) of the untreated cell were 80 ± 17 (means ± SD; n = 10) and 99 ± 15 (n = 10) for nominally free and 1.0 mM extracellular Ca²⁺, respectively. Inhibition of electron transport by 2 mM NaCN increased the autofluorescence to 116 ± 31 (n = 9) and 150 ± 28 (n = 11) for nominally free and 1.0 mM extracellular Ca²⁺, respectively. In contrast, 1 µM CCCP significantly suppressed the fluorescence to 40 ± 8 (n = 10) and 38 ± 6 (n = 11) for nominally free and 1.0 mM extracellular Ca²⁺, respectively.

Extracellular Ca²⁺ significantly affected the fluorescence intensity in the control and NaCN-treated cells (P < 0.05, Student's t-test). The level of NADH reduction in the control cell was calculated assuming that NaCN caused maximal reduction, whereas CCCP caused maximal oxidation (8, 11). The quiescent single cardiomyocyte in the present study was 53% reduced regardless of the extracellular Ca²⁺.

We exposed 1 µM CCCP-treated cells to the various PO₂ levels of superfused gas. Five measurements were conducted for each of the seven oxygen levels. We first calculated the cellular fluorescence intensity averaged over the cell, followed by compensation for photo bleaching. As indicated in Fig. 1, average fluorescence intensity increased as oxygen concentration of the superfusion gas was decreased to <3.14%. Similarly, average fluorescence intensity of the control cell significantly increased but at a much lower extracellular oxygen level (superfusion gas O₂ <0.50%).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Changes in cellular autofluorescence plotted against fractional oxygen concentration of superfusion gas. Autofluorescence intensity was averaged over the cell and normalized to the autofluorescence at anoxia (superfusion with <0.001% O2 gas). Photo bleaching was compensated. Horizontal dotted lines represent average fluorescence level at nonlimiting PO2 determined by separate experiments (fluorescence intensity normalized by that in the presence of 2 mM NaCN). Data are means ± SD; n = 5.

Finally, we determined subcellular variations of NAD(P)H fluorescence using the digital image processing. Figure 2 demonstrates validity of the image-processing technique in a 1 µM CCCP-treated cell. The difference in the fluorescence intensities between the anoxic cell (Fig. 2B) and the cell exposed to 2.09% O₂ (Fig. 2A) was difficult to detect. In contrast, ratio imaging indicated in Fig. 2D clearly visualized intracellular heterogeneities of NAD(P)H fluorescence. CCCP-treated (1 µM) cells analyzed by this way showed similar NAD(P)H fluorescence patterns depending on the superfusion gas O₂. Figure 3 summarizes the representative intracellular patterns of NAD(P)H fluorescence for superfusion gas containing 5.23% (B), 4.18% (C), 3.14% (D), 2.09% (E), and 1.00% (F) O₂. During exposure to 5.23% O₂, the fluorescence was low but uniform over the cell. As the extracellular oxygen level was lowered, heterogeneous fluorescence patterns became apparent. The fluorescence intensity was significantly higher in the intracellular space away from the sarcolemma. The heterogeneity was detectable even in the relatively well-oxygenated cell (Fig. 3C, superfusion by 4.18% O₂ gas). For extracellular oxygen levels at which we previously demonstrated the presence of radial PO_2,i gradients (superfusion gas O₂ 3.14% and 2.09%), heterogeneous NAD(P)H fluorescence patterns were clearly demonstrated.

View larger version (131K): [in this window] [in a new window]   Fig. 2.   Representative data demonstrating quantitation of NAD(P)H fluorescence in a single cell. A and B: raw fluorescence images of a cell superfused with 2.09% and <0.001% O2 gases, respectively. Fluorescence intensities are represented in shades of gray (calibration bar at top). B was compensated for photo bleaching. C: transmitted light image of test cell (scale bar 10 µm). D: NAD(P)H fluorescence pattern calculated from ratio imaging technique (A divided by B) representing relative fluorescence intensity of 2.09% O2 superfused cell. Photo bleaching was not compensated. Data are represented in pseudo colors (calibration bar bottom). Cell was treated with 1 µM CCCP.

View larger version (92K): [in this window] [in a new window]   Fig. 3.   Representative data demonstrating heterogeneities of NAD(P)H fluorescence within a cell. A: NAD(P)H fluorescence of a control (untreated) cell superfused with 0.25% O2 gas. B-F: cells treated with 1 µM CCCP to increase oxygen flux. Superfusion gas %O2: 5.23% (B), 4.18% (C), 3.14% (D), 2.09% (E), 1.00% (F). Fluorescence intensity was normalized to that of anoxia (superfusion with <0.001% O2 gas) and represented in pseudo colors (calibration bar on right). Photo bleaching was not compensated. A 10-µm calibration bar appears in each panel.

In the control cell, the heterogeneity of NAD(P)H fluorescence was never detected (Fig. 3A) for all of the superfusion gas levels examined in the present study.

	DISCUSSION

Top Abstract Introduction Results Discussion References

In the previous study (24), we demonstrated that steep radial PO_2,i gradients may be produced within an actively respiring, metabolically uncoupled, single isolated cardiomyocyte. The profile of PO_2,i was consistent with the prediction by mathematical models (12). We conducted the present study to further determine whether such steep radial PO_2,i gradients directly limit mitochondrial respiration. To assess the mitochondrial oxygenation with a subcellular spatial resolution, we measured the autofluorescence of a single cardiomyocyte. Previous studies (8, 9) demonstrated that the broad spectrum centered at ~450 nm on UV (~350 nm) excitation of an isolated single cardiomyocyte of the rat can be an indicator of the relative level of mitochondrial NADH with a minor contribution of NADPH (7, 8). In the mitochondrial respiratory chain, oxidation of NADH is the major source of the electron supply to molecular O₂ via the respiratory chain. Thus the NADH/NAD⁺ redox couple indicates whether diffusional oxygen supply to the mitochondria is adequate (3, 4). It is also likely that the elimination of the respiratory control by CCCP would allow accurate estimation of oxygen availability at the mitochondria by NADH fluorescence if the substrate supply is nonlimiting.

We showed that the fluorescence intensity averaged over the cell reasonably responded to pharmacological alterations of the mitochondrial redox state. At nonlimiting PO₂, the present untreated quiescent cardiomyocyte was 53% reduced at 27°C. This value is comparable to that reported by Eng et al. (8) in the similar preparation (44 ± 6%), whereas White and Wittenberg (26) reported a somewhat lower value (27 ± 3%). Eng et al. (8) also examined the topological correlation between the pattern of the autofluorescence and the localization of mitochondria in a single cardiomyocyte. They demonstrated that the intracellular fluorescence structure (i.e., longitudinal stripes), which responded to metabolic transition caused by cyanide, correlated with the distribution of mitochondria determined either by a specific fluorescence probe 2-(4-dimethyl aminostyryl)-N-methylpyridinium iodide or electron micrograph. This was further verified by Richmond et al. (20) where they resolved the distribution of the fluorescence pattern with a higher spatial resolution by a mathematical deconvolution technique. In the present study, a similar fluorescence pattern was in fact observed (Fig. 2), and the fluorescence changed according to the alterations of the mitochondrial redox state as expected. Therefore, the present fluorescence imaging is the visualization of a mitochondrial active NADH pool with a subcellular spatial resolution.

We did not conduct measurements of NAD(P)H fluorescence and S_Mb in the same cell, because the measurement of S_Mb is quite difficult and time consuming. Instead, we carried out the NAD(P)H measurement separately in the exactly same condition as we visualized the radial PO_2,i gradients. In the previous study (24), we found that the average S_Mb in 1 µM CCCP-treated cells at 27°C was 0.4-0.7 when the cell suspension in the measuring cuvette was superfused with 2.09% or 3.14% O₂ gas. The average S_Mb is comparable to that of the in vivo beating heart at normal oxygen consumption (6, 10, 29). This S_Mb value has been demonstrated to be maintained when the cardiac work is significantly elevated (10, 15). In these conditions, significant gradients of S_Mb were demonstrated along the short axis of the cell. Calculation of PO_2,i from the measured S_Mb revealed the presence of very steep gradients of PO_2,i near the sarcolemma, whereas the gradients were very shallow near the center of the cell. We speculated that these shallow PO_2,i gradients might be accounted for in part by the suppression of mitochondrial respiration caused by insufficient diffusional oxygen delivery to these regions: the hypoxic core (24). Present measurements in the exact same condition demonstrated the presence of significant gradients of NAD(P)H fluorescence that mirror image the intracellular gradients of S_Mb/PO_2,i.

Radial gradients of mitochondrial NAD(P)H fluorescence were detected only in cardiomyocytes treated with CCCP and subsequently exposed to mild hypoxia. This is consistent with the idea that treatment with CCCP stimulates mitochondrial oxygen consumption to the near-maximal level secondary to abolition of the mitochondrial respiratory control and produces large diffusional oxygen concentration gradients from the sarcolemma to the cell core. On the other hand, abolition of the mitochondrial respiratory control eliminates regulation of mitochondrial respiration by the cellular energy state. Hence, whether the uncoupled cardiomyocyte that lacks such regulation by the cellular energy state is an adequate model for the normal coupled cardiomyocyte with the same oxygen consumption should be discussed here.

In the coupled cardiomyocyte, mitochondria located near the central core and exposed to a lower PO₂ than the periphery may respire more slowly [e.g., half-maximal oxygen consumption at PO₂ (P₅₀) of 0.3 Torr, see Ref. 21]. Partial suppression of mitochondrial respiration near the cell core would thus tend to prevent the formation of steep radial oxygen concentration gradients. In contrast, in the uncoupled cardiomyocyte, mitochondria are insensitive to the cellular energy state and respire at near-maximal rates until oxygen is almost completely depleted (P₅₀ < 0.04 Torr, see Ref. 21). Consequently, if compared at the same rate of respiration, larger intracellular oxygen gradients may be produced in the uncoupled cell.

However, as oxygen flux increases, the effect on mitochondrial respiration of the difference in the P₅₀ between the coupled and uncoupled mitochondria is surpassed by increased diffusional oxygen concentration gradients, particularly in large cells such as the cardiomyocyte. Rumsey et al. (21) demonstrated that uncoupling isolated cardiomyocytes increased oxygen consumption 11-fold with a 4-fold increase in the P₅₀ for cellular respiration. Because uncoupling of oxidative phosphorylation should decrease the mitochondrial P₅₀ (due to lowered cellular energy level), they concluded that at the resting metabolic level the regulation of cellular respiration depends on the cellular energy state, whereas at significantly increased oxygen consumption rate, mitochondrial respiration is dominated by the oxygen concentration gradients from extracellular space to mitochondrial membrane. Therefore, CCCP-treated cardiomyocytes used in the present study that have high oxygen consumption but lack the respiratory control by the cellular energy state may be a model of intracellular oxygen transport of the normal cell working at a significantly elevated workload.

Measurement of NADH using autofluorescence may be affected by light absorption of intracellular pigments such as cytochromes and Mb (inner filter effect) (1, 17, 18). Increase in the autofluorescence in the cell core thus might be attributable to light absorption of deoxymyoglobin locating in these regions. This hypothesis, however, is not compatible with the present finding because deoxymyoglobin absorbs more fluorescence than oxymyoglobin for wavelengths >430 nm. The inner filter effect was also examined in cardiomyocytes in which Mb was inactivated by 2 mM NaNO₂. We found significant intracellular heterogeneities of the autofluorescence that were similar to those demonstrated in the absence of NaNO₂. Thus we conclude that heterogeneities of mitochondrial oxygenation may be produced in actively respiring cardiomyocytes, presumably resulting from insufficient diffusional O₂ delivery due to a relatively large oxygen diffusion resistance in the intracellular space.

Regional inhibition of oxidative metabolism arising from insufficient oxygen supply to mitochondria was also demonstrated at relatively high oxygen levels. For example, NAD(P)H fluorescence was significantly increased near the center of 1 µM CCCP-treated cells superfused with 3.14% O₂ gas (Fig. 3D). Figure 4 (adapted from Fig. 1 of Ref. 24) is an example of S_Mb measurement in a 1 µM CCCP-treated cell superfused with 3.14% O₂ gas. The nadir S_Mb of this particular cell was 0.55. The value corresponds to a PO_2,i of 4.0 Torr, which is far above the apparent Michaelis constant (K_m) of cytochrome aa₃ (13). Furthermore, the increase in NAD(P)H fluorescence was detectable in the intracellular space even a few micrometers away from the sarcolemma (Fig. 3D) where Mb may be only partially deoxygenated (Fig. 4). Chance (2) proposed a concept called gradient coherence as concurrent deoxygenation of Mb and reduction of cytochrome (aa₃). The gradient coherence was reported by Kennedy and Jones (16) in isolated single rat cardiomyocytes and by Tamura et al. (25) in isolated perfused rat heart. The present results appear to support the coherency between Mb oxygenation and mitochondrial oxidation. The gradient coherence may arise from changes in the apparent K_m of mitochondria that depend on cellular energy level (27, 28) and/or O₂ concentration gradients between cytosolic space and mitochondrial inner membrane (16), where the latter becomes dominant when mitochondrial respiration is stimulated (21). Gradient coherence at increased oxygen consumption thus implies that diffusion resistance between cytosol and mitochondrial inner membrane may be a factor that significantly affects mitochondrial oxygenation. We speculate that the effect of the PO₂ gradients between these two compartments would be more evident at regions near the cell core, where PO_2,i is lower (due to radial PO_2,i gradients) compared with the region near the sarcolemma.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   An example of intracellular profile of fractional myoglobin saturation (SMb, ) of a single cardiomyocyte that was incubated with 1 µM CCCP and superfused with 3.14% O2 gas at 27°C. Radial gradients of SMb were fitted to a hyperbolic curve (solid curve) and subsequently converted to intracellular PO2 (PO2,i, dashed curve) using calibration factors depicted in Fig. 6 of Ref. 24. Width of this particular cell was 23.9 µm. Figure was adapted from Fig. 1 of Ref. 24.

With a nonlimiting oxygen supply, mitochondrial oxygen consumption is principally set by the cellular energy state. In addition, our results suggest that oxygen diffusion resistance in the intracellular space is one of the important factors that limit mitochondrial oxygen consumption in the uncoupled cell. Regional PO_2,i decreases as the oxygen molecule travels from the sarcolemma toward the cell core (Fick's law of diffusion). Consequently, at increased cellular oxygen demand, the center of a cell is more subject to hypoxia, and mitochondrial respiration at this region may be suppressed (hypoxic core). Reduced regional oxygen consumption would conversely elevate the PO_2,i, resulting in partial restoration of mitochondrial respiration at the hypoxic core. Thus the regional oxygen metabolism and regional PO_2,i are determined from the interaction between diffusional oxygen transport (Fick's law) and mitochondrial respiration that depends on mitochondrial PO₂ (Michaelis-Menten relationship). The cell core is most susceptible to this diffusion limitation of oxygen metabolism. Therefore, intracellular diffusion of oxygen is one of the regulatory parameters for oxidative phosphorylation in uncoupled actively respiring cells.