2.1.

Ethics Approval

All procedures were performed in accordance with the National Institute of Health and the Canadian Council on Animal Care (CCAC) guidelines for the care and use of laboratory animals, after evaluation by the local Comité Institutionnel des Bonnes Pratiques sur les Animaux en Recherche, accredited by the CCAC.

2.2.

Cell Isolation

Cardiac myocytes were isolated from LV of Sprague-Dawley rats (13- to 14-week-old females, Charles River, Canada) sacrificed humanely by decapitation after retrograde perfusion of the heart with proteolytic enzymes, as previously reported.^20,21 Myocytes were maintained in storage solution at 4°C until used. Only cardiac myocytes showing clearly defined striations and edges were studied.

2.3.

NAD(P)H and Flavin Fluorescence

Endogenous fluorescence of NAD(P)H and flavins was recorded using spectrally resolved time-correlated single photon counting, as previously described.^14,22 In these experiments, a 375-nm, or a 435-nm, picosecond laser diode (BDL-375, BDL-435, Becker&Hickl, Boston Electronics, USA) was used as an excitation source with $\sim 1\mathrm{mW}$ output power. The laser beam was reflected to the sample through the epifluorescence path of the Axiovert 200M (Zeiss, Canada) inverted microscope. The emitted fluorescence was separated from laser excitation either by 395-nm dichroic and 397-nm long-pass filter (for excitation at 375 nm) or by 460-nm dichroic filter and 470-nm long-pass filter (for excitation at 438 nm) and detected by 16-channel photomultiplier array (PML-16, Becker&Hickl, Boston Electronics) attached to an imaging spectrograph (Solar 100, Proscan, Germany). Data were always evaluated from the first scan of each cell (measured for 30 s) to avoid artifacts induced by photobleaching.

2.4.

Solutions and Drugs

The storage solution contained (in mM) NaCl, 130.0; KCl, 5.4; ${\mathrm{MgCl}}_2·{6\mathrm{H}}_2\mathrm{O}$, 1.4; ${\mathrm{NaH}}_2{\mathrm{PO}}_4$, 0.4; creatine, 10.0; taurine, 20.0; glucose, 10.0; and 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 10.0; titrated to pH 7.30 with NaOH. The basic external solution contained (in mM): NaCl, 140.0; KCl, 5.4; ${\mathrm{CaCl}}_2$, 2.0; ${\mathrm{MgCl}}_2$, 1.0; glucose, 10.0; HEPES, 10.0, adjusted to pH 7.35 with NaOH. The intracellular media-mimicking solution contained (in mM) KCl, 140; NaCl, 10; glucose, 10; HEPES, 10, adjusted to pH 7.25 with NaOH. NADPH was added to the intracellular solution in concentrations ranging from 1 to 20 μM. NADPH was also produced from NADP-isocitrate dehydrogenase (ICDH) ($3.9\mathrm{U}/\mathrm{mL}$) by reaction of isocitrate (89 mM) and NADP (0.5 mM) with or without glutathione [glutathione disulfide (GSSG), 50 nM] and GR (0.5 or $1\mathrm{U}/\mathrm{mL}$). Cells were treated (in basic external solution for 30 min at 35°C) with 1,3-bis(2-chloroethyl)-1-nitroso-urea (BCNU, 100 μM) and/or HNE (10, 25, 50, or 100 μM), or ${\mathrm{H}}_2{\mathrm{O}}_2$ (1 μM), and recorded at room temperature. Rotenone (1 μM) with cyanide (4 mM) and/or 9,10-dinitrophenol (DNP, 50 μM) were applied for 5 to 10 min before recording; 3-β-hydroxybutyrate (BHB, 3 mM) was prepared freshly, while acetoacetate (AcAc, 150 μM to obtain a BHB/AcAc ratio of $20:1$, and 1.5 mM to obtain a BHB/AcAc ratio of $2:1$) was added from 250 mM stock solution. All chemicals were from Sigma-Aldrich (Canada), except for cyanide, which was from Fisher (Canada).

2.5.

Data Analysis

The data are reported as $\text{means}\pm \text{standard error}$ error of the means. Data were collected from at least three different animals per experimental group and compared by one-way analysis of variance; $p<0.05$ was considered statistically significant. Data were analyzed using SPCImage software (Becker&Hickl, Boston Electronics), Origin 7.0 (OriginLab, USA), laser scanning microscope Image Browser (Zeiss), and custom-written procedures for data correction and analysis written in C++. In our experiments, we typically collected a photon-counting histogram of spectrally resolved autofluorescence decay $P({\lambda }_j,t_k)$. The histogram was measured simultaneously on 16 spectral channels (18-nm wide) denoted as ${\lambda }_j$ and on 1024 temporal channels denoted as $t_k$, spaced equidistantly by 24.4 ps. Individual components of NAD(P)H fluorescence were determined by PCA, applying the linear unmixing approach to separate individual components in the recorded data, as previously reported.^13,14 Amplitudes of the resolved NAD(P)H fluorescence components were calculated by unmixing. The sum of amplitudes of all resolved components equals the total photon count in the measured experimental condition. Fluorescence lifetime was estimated using a monoexponential fitting procedure at the maximum intensity spectral channel (450 nm) for each decay component independently, with the assumption that the fluorescence decay is independent of emission wavelength. Intensity was calculated as $[(\text{component}\mathrm{amplitude})\times (\text{fluorescence}\text{lifetime})]$ for each component. Relative spectral contribution was calculated as an integral intensity, where the component amplitude was multiplied by the unitary reference spectrum reported in Ref. 13. This spectrum is composed from the one for NADH in organic solvents (represented by water) with maximum at 460 nm, for NADH in inorganic solvent (represented by glycerol) with maximum at 440 nm, and for flavin adenine dinucleotide (FAD) in water with maximum at 520 nm.

3.1.

NADPH Fluorescence Recording In Vitro

Fluorescence spectra and fluorescence lifetimes [Fig. 1(a)] of NADPH were studied in vitro in the intracellular media-mimicking solution (pH 7.25), as previously reported.¹² Steady-state emission spectra measured simultaneously at 16 acquisition channels were determined as the total photon counts on each spectral channel. Concentrations ranging from 1 to 20 μM of NADPH in intracellular media-mimicking solutions were used to test the dose-dependence of the NADPH fluorescence spectral intensity and fluorescence decays in vitro. As expected, the spectral intensity was linearly dependent on the NADPH concentration, as illustrated by the maximum emission wavelength of 450 nm [Fig. 1(a) right]; at the same time, we found no modification in the fluorescence decays [Fig. 1(a) middle]. Normalized spectra for NADPH concentrations from 1 to 20 μM overlaid perfectly (data not shown), confirming the same molecular origin. This result is in agreement with the effects induced by elevated NADH concentrations in aqueous solutions.^12,23

Fig. 1

NADPH fluorescence in vitro. (a) Concentration dependence of NADPH fluorescence in the intracellular media-mimicking solution (pH 7.25), excitation 375 nm. Background-corrected emission spectra (left), normalized fluorescence decays at 450 nm (middle), and the total photon counts at 450 nm (right) as a function of the NADPH concentration. (b) Viscosity dependence of NADPH (10 μM) fluorescence in intracellular solution. Background-corrected emission spectra (left), normalized fluorescence decays at 450 nm (middle), and the total photon counts at 450 nm (right) as a function of the glycerol concentration versus water. (c) ${\mathrm{NADP}}^{+}$-ICDH produced NADPH fluorescence in vitro in the absence or presence of GR (0.5 or $1.0\mathrm{U}/\mathrm{ml}$) in intracellular solution. Blank-corrected emission spectra (left), normalized fluorescence decays at 450 nm (right). (d) Schematic representation of the induced reaction.

To test the sensitivity to molecular environment, we have recorded NADPH fluorescence (10 μM) in solutions with different viscosity, achieved by the addition of glycerol (0, 25, 50, 75, and 100%) versus ${\mathrm{H}}_2\mathrm{O}$ [Fig. 1(b)]. The total photon counts, measured at the emission maximum (${\lambda }_{\mathrm{em}}=450\mathrm{nm}$), revealed exponential rise of the peak of NADPH fluorescence intensity with glycerol [Fig. 1(b) right]. We also noted a slight 10 nm blue spectral shift when NADPH fluorescence was recorded in a more viscous environment (data not shown). This result was due to the prolongation of the fluorescence lifetimes in a highly viscous environment [Fig. 1(b) middle] and is in agreement with previous findings,²⁴ showing that the sensitivity of NADPH fluorescence intensity to changes in its microenvironment can be related to changes in the molecular mobility and/or modification in the conformation of NADPH molecules in more viscous milieu.

NADP-ICDH has been shown to play an important role in increasing NADPH pool in heart mitochondria, especially during oxidative stress.^25,26 NADPH is a cofactor that is essential for glutathione system, one of the defense mechanisms against oxidative stress. It is used by GR to convert GSSG to reduced glutathione (GSH), which is needed to detoxify ${\mathrm{H}}_2{\mathrm{O}}_2$ and change it into water.²⁷ GR is a flavoprotein, with functions to regenerate antioxidant capacity, converting from GSSG to the sulfhydryl form GSH. As the most important nonprotein thiol source, GSH plays a key role in the maintenance of cellular redox status and antioxidant defense. We have tested the capability of the experimental setup to monitor the NADPH produced by ${\mathrm{NADP}}^{+}$-ICDH in vitro and the efficiency of GR to reverse this effect according to the schematic representation in Fig. 1(d). Our data revealed comparable spectral and lifetime characteristics of NADPH produced in vitro from NADP-ICDH and NADPH in intracellular media-mimicking solution. In the GR catalyzed reaction, the required molar ratio GSSG/NADPH is 1. We therefore investigated whether the binding of NADPH to GR has repercussions on its spectral and lifetime properties. Blank-corrected spectra [Fig. 1(c) left] showed that the addition of GR, in the presence of GSSG, reduced (at concentration $0.5\mathrm{U}/\mathrm{ml}$), or nearly completely abolished (at $1\mathrm{U}/\mathrm{ml}$), the NADPH fluorescence produced by NADP-ICDH, while no modifications of NADPH fluorescence decays by GR were noted [Fig. 1(c) right]. This observation of the loss in the fluorescence intensity is in agreement with the dehydrogenation of NADPH to ${\mathrm{NADP}}^{+}$ by GR [see schematic representation in Fig. 1(d)].

3.2.

HNE Decreases NAD(P)H Fluorescence in Living Cardiac Cells

LPO, which can be initiated by free radical attack, results in the production of toxic molecules such as HNE. Oxidative stress induced by HNE was evaluated by time-resolved spectroscopy of NAD(P)H fluorescence in living cardiac myocytes [Fig. 2(a)]. HNE (25 μM) induced clear decrease in the total photon counts of NAD(P)H fluorescence [Fig. 2(b)] and this effect was more important when higher HNE concentration (50 μM) was used. We have chosen a concentration of 10 to 25 μM of HNE, which produced no toxic cell shortening [see insets of Figs. 2(b) and 3(d)] and was also used by others to test bioenergetic function in response to oxidative stress²⁸ to evaluate HNE-induced NAD(P)H fluorescence modifications. In addition, a concentration of 50 μM, which was described to induce protein leak through uncoupling proteins (reviewed in Ref. 29) and which provoked shortening at the cell edges in 30% of tested cells [as illustrated in the insets of Figs. 2(b) and 3(d)], was also applied to evaluate the toxic effect of lipid peroxidation on NAD(P)H content; other oxidative stress agents, such as ${\mathrm{H}}_2{\mathrm{O}}_2$, provoked similar effects on the fluorescence, as well as on the membrane quality (data not shown). These results demonstrate that HNE is capable of inducing a concentration-dependent decrease in the cell NAD(P)H content.

Fig. 2

NAD(P)H fluorescence in living cardiac cells and its modulation by HNE. (a) Original recording of the time-resolved spectroscopy measurements of NAD(P)H fluorescence in living cardiac cells, excitation 375 nm (mean of 10 measurements; in the inset, representative transmission image of the single cardiomyocyte illumination, $\text{scale}=33\mu \text{m}$). (b) Background-corrected emission spectra recorded in living cardiac cells in rising concentrations of HNE; control, $n=78$; HNE, 25 μM, $n=71$; HNE, 50 μM, $n=19$; *$p<0.05$ versus control. (c) Individual NAD(P)H components resolved by linear unmixing of endogenous fluorescence excited at 375 nm: fluorescence decays of C1: component 1 bound NAD(P)H and C2: component 2 free NAD(P)H fluorescence is plotted in control conditions and in the presence of increasing HNE concentrations.

Fig. 3

Effect of HNE on flavin fluorescence in living cardiac myocytes. Effect of HNE (a) on fluorescence decays and (b) on the integral intensity spectra of the residual component, resolved by linear unmixing of endogenous fluorescence excited at 375 nm. (c) Original recording of the time-resolved spectroscopy of flavin fluorescence in living cardiac cells, excitation by 438 nm (mean of three measurements; in the inset, representative transmission image of the single cardiomyocyte illumination). (d) Effect of HNE (25 μM, $n=21$ and 50 μM, $n=5$) on flavin fluorescence (excited at 438 nm; in the inset, representative transmission images of the HNE-treated cardiac cells).

To better understand the mechanism underlying HNE effect on the NAD(P)H content in living cardiac cells, spectral linear unmixing of NAD(P)H fluorescence components was performed. We previously reported that the recorded fluorescence has mitochondrial origins and is composed of three individual components with separate spectral characteristics:^13,14 two NAD(P)H components [Fig. 2(c)] and one residual red-shifted flavin-like component [Fig. 3(a)]. The first blue-shifted component with longer fluorescence lifetime of $2.14\pm 0.08\mathrm{ns}$, $n=78$ [bound NAD(P)H], had characteristics of NAD(P)H component in a more viscous environment, while the second component with faster fluorescence lifetime of $0.55\pm 0.01\mathrm{ns}$, $n=78$ [free NAD(P)H] had characteristics of the NAD(P)H component in a less viscous environment.

Both resolved NAD(P)H fluorescence components decreased with rising HNE concentrations [Fig. 2(c): C1 and C2] due to lower component amplitudes [Fig. 4(a) left]. At high HNE concentrations (at 50 μM), deterioration of the cell membrane could be observed in about 20% of cells [see inset of Fig. 2(b)], and at these concentrations, we also noted a change in the fluorescence lifetime of the individual components [Fig. 4(a) middle], indicating modification in the cell environment. Despite the predominant effect of HNE on the amplitude of component 2, fluorescence intensity, calculated as $[(\text{component}\text{amplitude})\times (\text{fluorescence}\text{lifetime})]$, revealed that component 1 is responsible for the biggest part of the lowering of the NAD(P)H fluorescence with HNE [Fig. 4(a) right]. These modifications resulted in a clear decrease of the relative spectral contribution (calculated as the fluorescence amplitude multiplied by the unitary reference spectrum reported in Ref. 13) [Fig. 4(b)], suggesting an increase in the cell oxidation. The ratio of component 2 to component 1 (free/bound) amplitude, corresponding to the change in $\mathrm{NAD}(\mathrm{P})\mathrm{H}/\mathrm{NAD}{(\mathrm{P})}^{+}$ reduction/oxidation pair^18,19 and thus in the mitochondrial metabolic state, decreased with rising HNE concentrations [Fig. 4(c)]. These results were confirmed by mimicking oxidative stress conditions by the addition of ${\mathrm{H}}_2{\mathrm{O}}_2$ (1 μM) [Fig. 4(a)]. Resolved monoexponential decay for each NAD(P)H component corresponds well to the molecule in a specific state; this result is more precise than the one obtained in human myocytes,¹³ possibly due to higher signal intensities in rats when compared to human.

Fig. 4

Effect of HNE on individual components of NAD(P)H fluorescence. (a) The effect of HNE on the NAD(P)H fluorescence component amplitude (left), fluorescence lifetime (middle), and fluorescence intensity calculated as the component amplitude multiplied by the fluorescence lifetime of the resolved component 1 (gray squares) and component 2 (black stars) (right); the numbers represent cell number; *$p<0.05$ versus control. (b) Integral intensity, calculated for components C1 and C2 at respective spectral channels as the component amplitude multiplied by the unitary reference spectrum. (c) Ratio of component 2 and component 1 amplitude.

3.3.

HNE Increases Flavoprotein Fluorescence in Living Cardiac Cells

In addition to the two main NAD(P)H components, linear unmixing of the endogenous fluorescence also revealed presence of a small additional component. This component presented characteristics of the flavin fluorescence: emission maximum at 520 nm [Fig. 3(b)] and increase in the presence of uncouplers such as DNP (data not shown). HNE was capable of enhancing the contribution of this component in living cardiac cells [Fig. 3(b)]. To verify that this effect of HNE can indeed be attributed to flavoproteins, we have performed time-resolved spectroscopy experiments using 438-nm excitation [Fig. 3(c) and 3(d)] to favor the selective excitation of flavoproteins.³⁰ These data clearly demonstrated a rise in the flavoprotein fluorescence by HNE (25 and 50 μM) with maximum at 500 to 520 nm, which corresponds most likely to the mixture of electron transferring flavoproteins and free FAD.³¹

GR is one of the flavoproteins that could contribute to the observed fluorescence change in cardiac cells. The activity of GR can be selectively inhibited by the chemotherapeutic drug BCNU (carmustine). The in vitro inhibition of GR occurred after biochemical reduction of the enzyme with NAD(P)H.³² We have investigated the effect of HNE in the presence of the GR inhibitor BCNU (100 μM) (Fig. 5). In cells pretreated with BCNU (100 μM), we noted a decrease in the fluorescence intensity due to the reduced amplitude of component 2 [Fig. 5(a) and 5(b)]. In the presence of HNE (25 and 50 μM), BCNU induced an even further decrease of the cell autofluorescence by lowering the intensities of the two components [Fig. 5(a) and 5(b)]. We have also noted clear lowering of the component amplitude ratio and thus $\mathrm{NAD}(\mathrm{P})\mathrm{H}/\mathrm{NAD}{(\mathrm{P})}^{+}$ ratio following inhibition with BCNU [Fig. 5(c)].

Fig. 5

Responses of NAD(P)H fluorescence to glutathione reductase inhibition in the presence and in the absence of HNE. (a) The effect of GR inhibitor BCNU (100 μM) on background-corrected emission spectra in control (left) and in the presence of HNE, 25 μM (middle) and HNE, 50 μM (right). (b) The effect of BCNU (100 μM) on the component amplitude (left), fluorescence lifetimes (middle), and the fluorescence intensity (right); the numbers represent cell number; *$p<0.05$ versus control. (c) C2 to C1 amplitude ratio in the presence of BCNU.

3.4.

HNE Increases the Percentage of Oxidized Nucleotides

Percentage of oxidized nucleotides is an important parameter indicating cell oxidation state. To test whether HNE affects the percentage of oxidized nucleotides, cells were exposed to different concentrations of HNE and then to rotenone (1 μM) with cyanide (4 mM), or to DNP (50 μM), which were used to induce a fully reduced versus a fully oxidized state, respectively [Fig. 6(a)]. Rotenone in the presence of cyanide increased, while DNP decreased the fluorescence intensity of NAD(P)H [Fig. 6(a) and 6(c)]. Change in the percentage of oxidized nucleotides was then evaluated as fluorescence at (fully reduced—control)/(fully reduced—fully oxidized) state. Our data showed a clear rise in the percentage of oxidized nucleotides with rising concentrations of HNE [Fig. 6(b)], indicating an important cell oxidation in the presence of the LPO product. This result was verified by the addition of ${\mathrm{H}}_2{\mathrm{O}}_2$ (1 μM), a strong peroxidation agent, which was capable of inducing nearly 100% cell oxidation [Fig. 6(b)].

Fig. 6

Effect of HNE on the percentage of oxidized nucleotides in cardiac cells. (a) Background-corrected emission spectra of the NAD(P)H fluorescence in control conditions, $n=78$ and in fully reduced state after treatment with rotenone with cyanide (1 μM and 4 mM, respectively, $n=30$), or in fully oxidized state after treatment with DNP (50 μM, $n=15$) (left) and cells treated with HNE (25 μM, $n=74$) and after treatment with rotenone with cyanide (1 μM and 4 mM respectively, $n=25$), or with DNP (50 μM, $n=15$) (right). (b) Percentage of oxidized nucleotides, calculated as fluorescence at (fully reduced – control) / (fully reduced – fully oxidized) state in different HNE concentrations and/or in the presence of ${\mathrm{H}}_2{\mathrm{O}}_2$. (c) Component amplitude (left), fluorescence lifetimes (middle), and fluorescence intensity (right) of C1 and C2 in the presence of rotenone with cyanide (1 μM and 4 mM, respectively), or DNP (50 μM); the cell number is in brackets; *$p<0.05$ versus Control.

3.5.

HNE Lowers the Production of NADH

HNE has been reported to reduce the production of NADH in isolated cardiac mitochondria by acting through inhibition of α-ketoglutarate dehydrogenase.⁹ BHB oxidized into AcAc is known to produce NADH dependently on the BHB/AcAc ratio.³³ To investigate whether HNE is capable of affecting NADH production in living cardiac cells, we have therefore used different ratios of BHB/AcAc ($20:1$, condition favorable for NADH production, versus $2:1$, closer to physiological condition) (Fig. 7). As expected, we observed lower fluorescence when decreasing the BHB/AcAc ratio from $20:1$ to $2:1$, in accordance with lower NADH concentration in cardiomyocyte mitochondria [Fig. 7(a) and 7(b)]. This effect was also observable in the presence of HNE (25 μM). Maximal production of NADH was then calculated as fluorescence at [control-(BHB:AcAc $2:1$)]/[(BHB:AcAc $20:1$)—(BHB:AcAc $2:1$)]. In this way, we estimated maximal production of NADH to 97% in control cells before HNE application; in the presence of HNE (25 μM), it decreased to 28.6%. These experiments are in agreement with previously reported reduced NADH production by HNE.³⁴

Fig. 7

Effect of HNE on the NADH production in cardiac cells. (a) Background-corrected emission spectra of the NAD(P)H fluorescence in control conditions ($n=78$) and/or treated with BHB/AcAc $20:1$ ($n=35$), or BHB/AcAc $2:1$ ($n=18$) (left) and in cells treated with HNE (25 μM, $n=71$) alone and/or treated with BHB/AcAc $20:1$ ($n=16$), or BHB/AcAc $2:1$ ($n=16$) (right). (b) Component amplitude (left), fluorescence lifetimes (middle), and fluorescence intensity (right) of C1 and C2 in the presence of different BHB/AcAc ratio; the cell number is in brackets; *$p<0.05$ versus control.