Minimizing photodecomposition of flavin adenine dinucleotide ...

Journal of Microscopy, Vol. 264, Issue 2 2016, pp. 215–223

doi: 10.1111/jmi.12436

Received 1 March 2016; accepted 30 May 2016

Minimizing photodecomposition of flavin adenine dinucleotide fluorescence by the use of pulsed LEDs ´ CS‡ J . R O¨ S N E R ∗ , ∗∗ , A . L I O T T A †, ∗∗ , E . A . A N G A M O ∗ , C . S P I E S †, U . H E I N E M A N N ∗ & R . K O V A ∗ Neuroscience Research Center, Charité Universit¨ atsmedizin, Berlin

†Department of Anesthesiology and Intensive Care Medicine, Charité Universitätsmedizin, Berlin ‡Institute for Neurophysiology, Charité Universitätsmedizin, Berlin

Key words. Brain slice, energy metabolism, evoked field potentials, flavin adenine dinucleotide, hippocampus, photobleaching. Summary Dynamic alterations in flavin adenine dinucleotide (FAD) fluorescence permit insight into energy metabolism-dependent changes of intramitochondrial redox potential. Monitoring FAD fluorescence in living tissue is impeded by photobleaching, restricting the length of microfluorimetric recordings. In addition, photodecomposition of these essential electron carriers negatively interferes with energy metabolism and viability of the biological specimen. Taking advantage of pulsed LED illumination, here we determined the optimal excitation settings giving the largest fluorescence yield with the lowest photobleaching and interference with metabolism in hippocampal brain slices. The effects of FAD bleaching on energy metabolism and viability were studied by monitoring tissue pO2 , field potentials and changes in extracellular potassium concentration ([K+ ]o ). Photobleaching with continuous illumination consisted of an initial exponential decrease followed by a nearly linear decay. The exponential decay was significantly decelerated with pulsed illumination. Pulse length of 5 ms was sufficient to reach a fluorescence output comparable to continuous illumination, whereas further increasing duration increased photobleaching. Similarly, photobleaching increased with shortening of the interpulse interval. Photobleaching was partially reversible indicating the existence of a transient nonfluorescent flavin derivative. Pulsed illumination decreased FAD photodecomposition, improved slice viability and reproducibility of stimulus-induced FAD, field potential, [K+ ]o and pO2 changes as compared to continuous illumination. Introduction The fluorescence emission of the electron carrier molecule flavin adenine dinucleotide (FAD) depends on its redox state. ∗∗

J. Rösner and A. Liotta contributed equally to this work.

Correspondence to: Richard Kovacs, Institute for Neurophysiology, Charité Universitätsmedizin Berlin, Chariteplatz 1, 10117 Berlin, Germany. Tel: +49 30 450 528357; fax: +49 30 450 528 962; e-mail: [email protected] C 2016 The Authors C 2016 Royal Microscopical Society Journal of Microscopy

The oxidized form of FAD contains an isoalloxazine chromophore that emits fluorescence (emission peak at 515 nm) when excited with blue light, whereas the reduced form, FADH2 is not fluorescent. Therefore, precious insights into the redox state of this electron carrier can be obtained in living tissue simply by monitoring autofluorescence at the appropriate wavelengths (Chance & Williams, 1955). Although several flavin-containing enzymes undergo redox changes during metabolism, the vast majority of dynamic flavin-dependent fluorescence signal originates from mitochondrial metabolism (Kunz & Kunz, 1985). Monitoring FAD fluorescence – along with the redox-dependent fluorescence of another electron carrier; nicotinamide adenine dinucleotide (NADH) – has been widely used in primary cell cultures, living tissue slices and also in vivo to gather information on oxidative energy metabolism (Hassinen, 1986; Duchen & Biscoe, 1992; Combs & Balaban, 2001; Schuchmann et al., 2001, Shibuki et al., 2003; Shuttleworth et al., 2003; Reinert et al., 2004; Foster et al., 2005; Gerich et al., 2006). The advantage of recording FAD fluorescence over NADH lies in the difference of their excitation spectra. NADH fluorescence is excited with potentially phototoxic UV light (excitation maximum; 365 nm), which on the other hand, has a restricted tissue penetration (Chance et al., 1962). Phototoxicity is an important issue in living preparations, such as primary cell cultures, brain slices or slice cultures. Another disadvantage of NADH is its involvement in a multitude of mitochondrial and cytosolic reactions that complicates the interpretation of the changes in NADH fluorescence (Brennan et al., 2006). Although FAD fluorescence would be an ideal candidate for monitoring energy metabolism-related changes in intracellular redox potential, its use has been hampered by the fact that FAD fluorescence is less intense as compared to NADH, necessitating higher excitation light intensity (Aubin, 1979). Another difficulty is the strong photobleaching, observed in fluorescent microscopy applications using lasers or arc lamps for excitation. One exception is represented by multiphoton microscopy, which, however, requires expensive equipment

216

J . R O¨ S N E R E T A L .

and is labour-intensive (Huang et al., 2002). Metabolismdependent shifts of redox potential are usually slow, lasting from seconds to several minutes, necessitating exposure to excitation light for several minutes (Huchzermeyer et al, 2008; Rösner et al., 2013). FAD fluorescence decline due to illumination is ascribed to reversible photobleaching and photodecomposition, similar to most exogenous fluorescent probes used for live-cell imaging. However, as FAD is intricately embedded in the metabolic network of cells, its photodecomposition also affects the metabolic flux of electrons as well as its synthesis and catabolism. As a consequence, tissue viability could be impaired by long-time FAD fluorescence measurements. With respect to data analysis, several strategies exist for correction of the bleaching of FAD fluorescence. Quite often, a simple linear correction is used, following the steady state of the first exponential decay component. Although this approach is a reasonable compromise, it still accepts considerable fluorescence losses occurring at the beginning of an experiment. An alternative method is to take a baseline trace without any stimulus followed by a trace of the same length with an appropriate stimulation. Subsequently, the baseline trace is subtracted to achieve the real kinetics of stimulusinduced redox changes. Although this method yields good results, there is no evidence that the second trace would have exactly the same baseline as it had in the first run – especially, for long recordings where the amount of bleaching during the control trace is not negligible anymore. Some studies did not correct for bleaching at all and are restricted to the fast components of the FAD signal. However, slow recovery of FAD to the baseline following minute-long changes could not be determined without knowing the exact kinetics of photobleaching. None of the aforementioned correction strategies addresses the consequences of photodecomposition. Thus, instead of correcting for, the optimal solution would be to minimize bleaching while achieving the same fluorescence yield. This would improve viability of the preparation and allow for determination of the real kinetics of slow redox changes. The opportunity of defining illumination protocols with arbitrary pulse length and frequency came along with the advent of narrow band LEDs in fluorescence microscopy (Rösner et al., 2013). Here, we sought to determine the optimal illumination protocol by using a narrow band LED as a light source driven by a commercially available trigger box. Kinetics of FAD fluorescence decline in hippocampal brain slices were compared with continuous and pulsed illumination protocols. In a subsequent set of experiments, we determined the dependence of fluorescence decay on the pulse length and duration. Potential changes in viability and excitability of the slices were tested by monitoring neuronal activity-dependent changes in FAD fluorescence and tissue oxygen consumption as well as evoked potentials and changes in extracellular potassium concentration ([K+ ]o ) at different time points of long recordings.

Experimental methods Slice preparation and maintenance For this study, 13 adult (4-6 weeks old) male Wistar rats were used. Animals were housed and sacrificed in accordance with the Helsinki declaration and institutional guidelines as approved by local authorities (LAGeSo, T0096/02). The animals were decapitated under deep anaesthesia with isoflurane (2% v/v) and laughing gas (N2 O, 70%). The brain was rapidly removed and transferred to cold and gassed (95% O2 and 5% CO2 ) artificial cerebrospinal fluid (aCSF) containing (in mM): 129 NaCl, 21 NaHCO3 , 10 glucose, 3 KCl, 1.25 NaH2 PO4 , 1.6 CaCl2 and 1.8 MgCl2 . Osmolarity and pH were 295–305 mOsm L−1 and 7.35–7.45, respectively. Horizontal hippocampal slices of 400 μm thickness were prepared with a Leica VT1200 S vibratome (Wetzlar, Germany) and stored in an interface chamber with continuous aCSF perfusion (flow rate of 2 mL min−1 , temperature 34–35°C). After 2 h of recovery, slices were transferred to a submerged chamber for fluorescence, oxygen and electrophysiology recordings (flow rate 10 mL min−1 , temperature 34–35°C). The high flow rate permitted partial oxygen pressure (pO2 ) of 100 mmHg at the recording depth in spite of the submerged conditions.

Electrophysiology, oxygen and fluorescence recordings Simultaneous [K+ ]o and field potential measurements were performed using double-barrelled ion-sensitive microelectrodes built according to the protocol described in Liotta et al. (2012). The Potassium Ionophore I 60031 (Fluka, Buchs, Switzerland) was used accordingly. All experiments were performed in the hippocampal area CA1. In order to induce activity-dependent changes in FAD fluorescence, we applied 2 s, 20 Hz stimulus trains (pulse duration 100 μs) with a bipolar platinum wire electrode to the stratum radiatum thereby activating Schaffer collaterals (intensity evoking 50–75% of maximal population spike amplitude). The extracellular potassium ion concentration (in mM) was calculated by using the modified Nernst equation as described previously. Tissue pO2 and stimulus-induced pO2 responses were measured by using Clark-style oxygen sensor microelectrodes (tip: 10 μm; Unisense, Aarhus, Denmark) placed near to the ionsensitive microelectrode at 50 μm depth from the surface. The oxygen electrodes were polarized overnight and two point calibrated in aCSF gassed with 50% and 95% O2 prior to each recording session (Liotta et al., 2012). FAD fluorescence was recorded using a custom-built equipment with a light emitting diode (LED, 460 nm, Lumen, Prior Scientific, Rockland, MA, USA) and a photomultiplier tube (PMT, Seefelder Messtechnik, Germany) by using a filter set (exciation BP 475/50 nm; dichroic mirror HC BS 506, ¨ emission BP 540/50, AHF Tubingen, Germany). We used a C 2016 The Authors C 2016 Royal Microscopical Society, 264, 215–223 Journal of Microscopy

MINIMIZING PHOTODECOMPOSITION OF FLAVIN ADENINE

20× water immersion objective (N.A.: 0.5) focusing at the stratum pyramidale of CA1 (PMT field of view 1.2 mm). The LED intensity was set for all recordings at 18%, resulting in 2.39 mW radiant power (53 mW cm−2 ) below the objective and was triggered externally with a Master 8 (A.M.P.I., Jerusalem, Israel). The gain of the PMT was set to 400 V for all experiments. Fluorescence was either recorded with continuous or pulsed mode of the LED. The impact of pulsed excitation on photobleaching of FAD was studied either with increasing pulse duration (1, 5, 10 and 50 ms) at constant frequency (5 Hz) or with variable frequency (1, 5, 10 and 50 Hz) with fixed pulse duration (5 ms). For comparison of FAD fluorescence bleaching between pulsed and continuous illumination, baseline recordings with continuous light exposition and pulsed light at 5 Hz were performed. In order to determine the effects of photobleaching of FAD on slice excitability and oxidative metabolism, tissue pO2 was recorded simultaneously in these experiments. Short stimulus trains (20 Hz/2 s) were applied to the Schaffer collaterals 5 and 20 min after starting fluorescence recordings. In some of these experiments, we performed simultaneous electrophysiological (field potential and [K+ ]o ) recordings in addition to pO2 and FAD fluorescence monitoring.

Data acquisition and analysis Data acquisition was performed using Spike2, analog signals of the electrodes were low pass filtered (3 kHz for field potential and 1 kHz for K+ ) and digitalized with a CED micro1401 (Cambridge Electronic Design Ltd., Cambridge, UK). Data analysis and statistics were performed using Spike2, Excel 2010 (Microsoft) Origin (Origin Lab Corporation) and SPSS (IBM). Data are presented as mean and SEM. Fluorescence is shown either as absolute intensity (Photomultiplier output in Volts) or relative fluorescence intensities in percentage of the baseline intensity (f/f0 ). F0 was either taken as an average intensity of 20 s period prior to the stimulus or at the beginning of the respective trace (comparison of frequency/duration dependence of photobleaching). Continuous intensity plots from recordings with pulsed light were generated by using the ´XYMeasurements´ function of Spike2 (Method: peak find, Minimum step 0.18 s). Decay time constants (τ ) were calculated for each experiment with Origin using the fit function for firstorder exponential decay (y = y0 ± Ae−x/t ). The fit was accepted when the squared correlation coefficient (R2 ) was larger than 0.9. Only in experiments concerning the application of 1 ms light pulses, when the signal did not reach saturation during the duration of the pulse, were fits with lower R2 values (0.5) accepted. Between and within group differences were tested for significance either by Student´s t-test for normally distributed data or by Friedman test followed by Wilcoxon test with Bonferroni correction. Changes were stipulated to be significant when p was 10 Hz temporal resolution, experimental protocols abstain of dark intervals or keep them minimal, which corresponds to potentially injurious continuous illumination conditions. Alternatively, CCD acquisition rates are decreased to 0.5–0.2 Hz in order to counterbalance the bleaching during long exposure time of individual images in a sequence, thereby losing high frequency information. The present protocol could be easily adopted for FAD imaging with fast EMCCD/sCMOS cameras by using synchronized acquisition and LED triggering, thereby minimizing the interference of bleaching with the experimental data. C 2016 The Authors C 2016 Royal Microscopical Society, 264, 215–223 Journal of Microscopy


219

Fig. 2. Dependence of the recovery kinetics after photobleaching on previous bleaching periods (A) Example traces of photobleaching with increasing frequency at a fixed pulse duration of 5 ms. ABC marks the 30 s breaks between individual traces. Note that the largest part of the bleaching did not recover during this period. (B) Excerpts of the 30 s breaks marked with A, B and C. Arrowheads mark the recovery during the 30 s interval between individual traces. (C) Plot of the recovery amplitude (PMT readout in mV) between pulsed illumination periods with increasing frequency. Recovery amplitude positively correlates with the frequency of the prior bleaching period * = p < 0.05. (D) Example traces showing the photobleaching at 1 Hz pulse frequency with or without previous bleaching, arrowheads marking the 30 s breaks between individual traces. Note that fluorescence kinetics changes its sign if it was preceded by an intense bleaching due to 100 Hz pulsed illumination protocol.

We concluded that 1 Hz pulse frequency was the most appropriate to decrease photobleaching and minimize photodecomposition. However, the application of 5 Hz/5 ms pulses seems to be a reasonable compromise for sufficient temporal resolution and reduced bleaching during prolonged FAD fluorescence recordings in brain slices. Dependence of fluorescence recovery after photobleaching on pulsed illumination frequency Although the FAD fluorescence recovered during the 30 s interval between individual traces, it never reached the initial fluorescence level, indicating lasting loss of the chromophore (Fig. 2A). Remarkably, the recovery during the break depended on the frequency of the pulses during the previous illumination period (Figs. 2B, C). There was only a slight recovery of FAD fluorescence between the end of the 1 Hz and the beginning of the 5 Hz traces. With increasing pulse frequency up to 100 Hz, the amplitude of recovery becomes larger, albeit still incomplete due to the increased rate of photobleaching at high frequency illumination (Figs. 2C, D). Switching back to 1 Hz after the 100 Hz illumination period, the fluorescence stayed below the level of the initial 1 Hz trace. However, instead of photobleaching, we observed an ongoing run up in FAD fluorescence, indicating that intense bleaching activates additional recovery mechanisms that were absent during the first recording period at 1 Hz pulse frequency

C 2016 The Authors C 2016 Royal Microscopical Society, 264, 215–223 Journal of Microscopy

(Fig. 2D). In spite of this slow increase, the fluorescence following 100 Hz illumination did not recover within the 5 min recording to the initially observed level. Thus, a significant part of the photobleaching was irreversible, likely due to photodecomposition of FAD molecules. Photochemical degradation of flavins is the result of intramolecular photoreduction in a oneelectron transfer process from the ribityl-side chain resulting in other intermediates that are susceptible to further photolysis (Heelis, 1991; van den Berg, 2001). The reversible component of photobleaching could be explained by assuming a light-induced transition of the chromophore to a nonfluorescent derivative in addition of genuine photodecomposition of FAD. Upon continuous illumination, a certain fraction of the fluorochrome molecules is always converted into a dark state where they are not available for further fluorescence cycles (van den Berg, 2001). Following cessation of the illumination, these fluorochromes may return gradually to the excitable ground state, whereas photodecomposition would be mostly irreversible as suggested by experiments in which 1 Hz illumination followed the 100 Hz bleaching period. The time course of recovery of the FAD fluorescence was in seconds to minute range that is several orders of magnitude longer than that would be expected if the excited triplet life time of the flavin isoalloxazine would determine the dark state of FAD. Excited triplets of flavin can be effectively quenched by various electron donors (ascorbic acid, amines or amino acids) in the presence of oxygen (Görner, 2007). The

220


resulting semiquinone radical anion, yielded by the electron transfer from an appropriate donor following photon absorption, might be one component of the reversible nonfluorescent FAD population. Finally, the light-induced conversion of FAD results in the reduced, nonfluorescent flavin hydroquinone via the semiquinone intermediates. Reoxidation of the molecule and regaining the fluorescent state might be achieved by the activity of the mitochondrial electron transport chain or by direct enzymatic equilibration with the NAD± /NADH redox couple (Combs & Balaban, 2001). The kinetics of these processes would fit to the observed slow recovery of FAD in our system. Indeed, fluorescence recovery after photobleaching of another electron carrier, NADH, depended on the availability of metabolic substrates in a cell culture system (Rösner et al., 2013). Remarkably, FAD bleaching kinetics with 1 Hz pulsed illumination depended on the presence of prior high frequency illumination. Following intense photobleaching at 100 Hz pulse frequency, the FAD fluorescence gradually recovered when monitored at 1 Hz. The positive deflection of the fluorescence curve might indicate adaptive processes – such as enhanced de novo synthesis – counterbalancing the photodecomposition of FAD. In the next series of experiments, we investigated whether the reduced photobleaching due to pulsed illumination would also improve viability and responsiveness of brain slices.

Comparison of the effects of continuous and pulsed illumination on respiration To compare the effects of continuous light exposure with pulsed illumination (5 Hz frequency and 5 ms pulse duration), we performed baseline recordings for 25 min in two groups of slices with the same LED power and PMT gain (Fig. 3). As expected, pulsed illumination resulted in less photobleaching than continuous illumination (88.2 ± 2.1% of the initial fluorescence for pulsed LED vs. 57.3 ± 3.0% for continuous LED illumination after 25 min exposition, n = 6 and 7, respectively, p < 0.001, independent t-test, Figs. 3A, B). Similar to the short-term (5 min) recordings, an initial exponential decay represented the largest part of photobleaching with continuous illumination, followed by a slower, almost linear decay. The initial exponential component was significantly prolonged in case of pulsed illumination (Fig. 3A). The decay time constant was significantly shorter during continuous illumination (338.3 ± 17.8 vs. 467.8 ± 22.3 s for continuous vs. pulsed illumination, respectively, n = 6 and 7, p = 0.002, independent t-test, Fig. 3B). As mitochondrial FAD is an essential electron carrier for oxidative energy metabolism, we asked whether the decreased photobleaching would have a positive effect on oxygen consumption of the brain slice. This can be investigated by monitoring tissue pO2 , because in the presence of continuous carbogen (95% O2 /5% CO2 ) supply to the aCSF, any change in

Fig. 3. Effects of continuous and pulsed illumination on FAD bleaching and respiration (A) Plot of the normalized fluorescence change during long-term recording (25 min) with continuous (black) and pulsed (5 Hz/5 ms) illumination (red). The continuous illumination was accompanied by increased photobleaching. Note the difference in the initial exponential component of the fluorescence decline. (B) Plot of the relative decay and the decay time constant under different illumination protocols. The decay time during pulsed illumination (red) was significantly longer than during continuous illumination (black). * = p < 0.05. (C) Continuous illumination and photobleaching were associated with an increase in tissue partial oxygen pressure (pO2 ), whereas pO2 tended to decrease during recordings with pulsed illumination (black). (D) Comparison of the initial and terminal pO2 values after 25 min of continuous (black) or pulsed (red) illumination protocols. * = p < 0.05.

pO2 indicates a change in respiration (Huchzermeyer et al., 2008; Liotta et al., 2012). In slices exposed to continuous illumination, we observed an increase of baseline tissue pO2 (from 119.7 ± 10.2 to 147.7 ± 16.8 mmHg, n = 7, p = 0.02, paired t-test, first and the last 20 s, Figs. 3C, D) indicating decreased oxygen consumption. In contrast, pulsed illumination was accompanied by a decrease of tissue pO2 (from 118.8 ± 4.9 to 91.6 ± 6.0 mmHg, n = 6, p = 0.001, paired t-test, first and the last 20 s, Figs. 3C, D). Decreased respiration with continuous illumination could be a sign of disturbances in the function of the electron transport chain due to restricted availability of FADmediated electron supply. In addition, photodecomposition of other flavin-dependent enzymes might negatively interfere with cell viability. Unfortunately, measuring solely the C 2016 The Authors C 2016 Royal Microscopical Society, 264, 215–223 Journal of Microscopy


221

Fig. 4. Comparison of the effects of continuous and pulsed illumination on FAD fluorescence transients (A) Overlay of two recordings of FAD fluorescence with pulsed (black) and continuous illumination (grey). Note the prominent photobleaching and a decay of the absolute amplitude of the FAD transients when continuous light was applied. (B) Comparison of evoked biphasic FAD fluorescence changes as recorded in the 5th and 20th minute with pulsed illumination protocol. (C) Comparison of the peak and undershoot components of the FAD transient recorded at the 5th and 20th minute. The peak and the undershoot of the FAD signals remained stable (n.s.: not significant). (D) Comparison of evoked biphasic FAD fluorescence changes as recorded in the 5th and 20th minute with continuous illumination protocol. (E) Comparison of the peak and undershoot components of the FAD transient recorded at the 5th and 20th minute. The peak and the undershoot of the FAD signals showed a clear rundown with continuous illumination. * = p < 0.05.

changes in pO2 did not allow for distinguishing between reversible disturbances in respiration and protein/cell damage. As neuronal function relies almost exclusively on oxidative energy metabolism, we assessed how photobleaching in brain slices influences neuronal responsiveness. Comparison of the effects of continuous and pulsed illumination on FAD fluorescence transients and neuronal excitability If photodecomposition of FAD affects neuronal excitability, long-term fluorescence recordings might falsify simultaneously obtained electrophysiological data. In the next set of experiments, we compared the impact of different illumination protocols on stimulus-induced changes in FAD, pO2 and C 2016 The Authors C 2016 Royal Microscopical Society, 264, 215–223 Journal of Microscopy

[K+ ]o . Repetitive stimulation- (20 Hz/2 s) induced responses were compared between the beginning and the end of the fluorescence recording, either with continuous or with pulsed illumination (Fig. 4A). Stimulus trains induced a rise in [K+ ]o with subsequent enhancement of oxidative energy metabolism in order to restore transmembrane ion gradients (Liotta et al., 2012). Accordingly, tissue pO2 decreased transiently, whereas FAD fluorescence underwent a biphasic change due to the initial oxidation (peak) and late reduction (undershoot) of this electron carrier (Shuttleworth et al., 2003; Rösner et al., 2013). As shown in Figure 4, repeated stimulus trains at 20 min intervals induced biphasic FAD fluorescence transients that remained stable in case of pulsed illumination (peak: 2.95 ± 0.5 vs. 3.08 ± 0.4%, p = 0.44, paired t-test

222


Fig. 5. Comparison of the effects of continuous and pulsed illumination on evoked population spike amplitude, pO2 and [K+ ]o transients (A1–3) Absolute population spike amplitude and pO2 -decay and [K+ ]o rise during FAD transient remained stable over 20 min during recordings with pulsed illumination (n.s.: not significant). (B1–3). When continuous light was applied, the amplitude of the population spike, the stimulus-induced [K+ ]o and pO2 -decay decreased significantly. * = p < 0.05. C. Plot of normalized changes of the data presented in A and B. The application of continuous light generated a rundown of the measured values indicating changes in slice vitality. The FAD signals shown in Figure 3 were recorded simultaneously with the presented data.

and undershoot: 6.99 ± 2.4 vs. 6.48 ± 1.8%, p = 0.88, n = 10, Wilcoxon test, Figs. 4B, C). On the contrary, with the continuous illumination regimen, we observed a rundown of the biphasic FAD fluorescence transients (peak: 2.98 ± 0.5 vs. 2.12 ± 0.4%, p < 0.001, n = 11, paired t-test and undershoot: 8.96 ± 1.5 vs. 5.35 ± 1.2%, p = 0.002, n = 11, Wilcoxon test, Figs. 4D, E). Field potential, [K+ ]o and pO2 changes showed only a minimal decay between the first and second stimulus train when exposed to pulsed illumination (population spike amplitude: 2.30 ± 0.4 vs. 2.27 ± 0.5 mV, p = 0.76; [K+ ]o : 0.64 ± 0.1 vs. 0.60 ± 0.14 mM, p = 0.5, n = 8; pO2 : 32.26 ± 5.2 vs. 32.77 ± 5.3 mmHg, p = 0.5, n = 8, paired t-test, respectively, Figs. 5A1-3 and C). The population spike amplitude (3.08 ± 0.4 vs. 2.35 ± 0.3 mV, p = 0.004, n = 10, paired t-test), the rise in [K+ ]o (0.87 ± 0.1 vs. 0.75 ± 0.1 mM, p = 0.002, n = 11, paired t-test) and the change in pO2 (25.54 ± 6.8 vs. 20.32 ± 6.2 mmHg, n = 9, p = 0.005, Wilcoxon test) decreased significantly from the first to the second stimulus train (Figs. 5B1-3 and C). Thus, pulsed illumination not only reduced photobleaching but it also improved the reproducibility of stimulus-induced changes in FAD, pO2 and [K+ ]o indicating a positive impact on tissue viability in general. In addition to the photodecomposition of FAD, ongoing metabolic dysfunction of the brain slices could be due to the formation of semiquinone and peroxide radicals following the reaction of the flavin semiquinone radical with oxygen in carbogen-saturated aCSF (Görner 2007). The electron transfer from a donor to the excited triplet state of FAD yields a

semiquinone radical anion and the donor radical cation that could convert oxygen either into superoxide, followed by dismutation to peroxide, or to singlet molecular oxygen (Zhang & Görner, 2009). Consequently, flavins can be used as sensitizers of photodamage of different proteins. Formation of radical oxygen species is an important issue in the brain slice preparation as complex neuronal network activity can be only maintained in the presence of hyperoxic conditions (Huchzermeyer et al., 2008). Indeed, preincubation of slice cultures with a free radical scavenger before monitoring the fluorescence of another electron carrier, NADH, increased the amplitude of the signals and delayed their decay in the presence of recurrent epileptiform activity (Kovács et al., 2002). Remarkably, stimulus-induced extracellular potassium changes declined significantly in the presence of continuous illumination indicating decreased tissue excitability likely due to ongoing cell loss. Although it was not tested for, it is has been described for several preparations, such as primary cell cultures, slice cultures or acute slices that continuous illumination negatively interferes with viability (Rodrigues et al., 2011; Penjweini et al., 2012). Conclusions Here, we describe an optimized excitation protocol for monitoring redox-dependent changes in FAD fluorescence in living tissue by using pulsed LEDs. The enhanced efficacy of our illumination protocol minimized photobleaching and the interference of FAD photodecomposition with energy metabolism



thereby improving slice viability during long imaging sessions. Thus, the use of pulsed LED at 5–10 Hz instead of continuous illumination yields the same fluorescence with an appropriate temporal resolution and with largely improved viability of the tissue. As LED excitation light sources become increasingly cheap, pulsed illumination protocols could be adopted to a multitude of in vitro and in vivo preparations (Mayevsky et al., 2011). Acknowledgement This work was supported by EU grant FP7 Desire (Grant Agreement n.: 602531-1) and by DFG grant He1128/18-1 as well as by EXC Neurocure (EXC 257) to UH and by the DFG grant Ko3814/1-1 to RK. We thank the Clinical Science Program of the Charité for support to AL. The authors are deeply grateful to Janine Tyra and Tanja Specowius for their technical and administrative assistance. References Aubin, J.E.J. (1979) Autofluorescence of viable cultured mammalian cells. Histochem. Cytochem. 27(1), 36–43. Brennan, A.M., Connor, J.A. & Shuttleworth, C.W. (2006) NAD(P)H fluorescence transients after synaptic activity in brain slices: predominant role of mitochondrial function. J. Cereb. Blood Flow Metab. 26(11), 1389–1406. Chance, B. & Williams, G.R. (1955) Respiratory enzymes in oxidative phosphorylation. II. Difference spectra. J. Biol. Chem. 217, 395–407. Chance, B., Cohen, P., Jöbsis, F. & Schoener, B. (1962) Intracellular oxidation–reduction states in vivo. Science 137, 499–508. Combs, C.A. & Balaban, R.S. (2001) Direct imaging of dehydrogenase activity within living cells using enzyme-dependent fluorescence recovery after photobleaching (ED-FRAP). Biophys. J. 80(4), 2018–2028. Duchen, M.R. & Biscoe, T.J. (1992) Mitochondrial function in type I cells isolated from rabbit arterial chemoreceptors. J. Physiol. 450, 13–31. Foster, K.A., Beaver, C.J. & Turner, D.A. (2005) Interaction between tissue oxygen tension and NADH imaging during synaptic stimulation and hypoxia in rat hippocampal slices. Neuroscience 132, 645–665. ¨ Gerich, F.J., Hepp, S., Probst, I. & Muller, M. (2006) Mitochondrial inhibition prior to oxygen-withdrawal facilitates the occurrence of hypoxiainduced spreading depression in rat hippocampal slices. J. Neurophysiol. 96, 492–504. Görner, H. (2007) Oxygen uptake after electron transfer from amines, amino acids and ascorbic acid to triplet flavins in air-saturated aqueous solution. J. Photochem. Photobiol. B. 87(2), 73–80. Hassinen, I.E. (1986) Reflectance spectrophotometric and surface fluorometric methods for measuring the redox state of nicotinamide nucleotides and flavins in intact tissues. Methods Enzymol. 123, 311–320. Heelis, P.F. (1991) Chemistry and biochemistry of flavoenzymes, I (ed. by F. ¨ Muller), p. 171. CRC Press, Boca Raton, FL.


223

Huang, S., Heikal, A.A. & Webb, W.W. (2002) Two-photon fluorescence spectroscopy and microscopy of NAD(P)H and flavoprotein. Biophys. J. 82, 2811–2825. Huchzermeyer, C., Albus, K., Gabriel, H.J., Otáhal, J., Taubenberger, N., Heinemann, U., Kovács, R. & Kann, O. (2008) Gamma oscillations and spontaneous network activity in the hippocampus are highly sensitive to decreases in pO2 and concomitant changes in mitochondrial redox state. J. Neurosci. 28(5), 1153–1162. Kovács, R., Schuchmann, S., Gabriel, S., Kann, O., Kardos, J. & Heinemann, U. (2002) Free radical-mediated cell damage after experimental status epilepticus in hippocampal slice cultures. J. Neurophysiol. 88(6), 2909–2918. Kunz, W.S. & Kunz, W. (1985) Contribution of different enzymes to flavoprotein fluorescence of isolated rat liver mitochondria Biochim. Biophys. Acta 841, 237–246. Liotta, A., Rösner, J., Huchzermeyer, C., Wojtowicz, A., Kann, O., Schmitz, D., Heinemann, U. & Kovács, R. (2012) Energy demand of synaptic transmission at the hippocampal Schaffer-collateral synapse. J. Cereb. Blood Flow Metab. 32(11), 2076–2083. Mayevsky, A., Walden, R., Pewzner, E. et al. (2011) Mitochondrial function and tissue vitality: bench-to-bedside real-time optical monitoring system. J. Biomed. Opt. 16(6), 067004. Penjweini, R., Loew, H.G., Hamblin, M.R. & Kratky, K.W. (2012) Longterm monitoring of live cell proliferation in presence of PVP-Hypericin: a new strategy using ms pulses of LED and the fluorescent dye CFSE. J. Microsc. 245(1), 100–108. Reinert, C., Dunbar, R.L., Gao, W., Chen, G. & Ebner, T.J. (2004) Flavoprotein autofluorescence imaging of neuronal activation in the cerebellar cortex in vivo. J. Neurophysiology 92, 199–211. Rodrigues, R.M., Macko, P., Palosaari, T. & Whelan, M.P. (2011) Autofluorescence microscopy: a non-destructive tool to monitor mitochondrial toxicity. Toxicol. Lett. 206(3), 281–288. Rösner, J., Liotta, A., Schmitz, D., Heinemann, U. & Kovács, R. (2013) A LED-based method for monitoring NAD(P)H and FAD fluorescence in cell cultures and brain slices. J. Neurosci. Methods 212(2), 222– 227. Schuchmann, S., Kovacs, R., Kann, O., Heinemann, U. & Buchheim, K. (2001) Monitoring NAD(P)H autofluorescence to assess mitochondrial metabolic functions in rat hippocampal-entorhinal cortex slices. Brain Res. Protoc. 7(3), 267–276. Shibuki, K., Hishida, R., Murakami, H. et al. (2003) Dynamic imaging of somatosensory cortical activity in the rat visualized by flavoprotein autofluorescence. J. Physiol. 549(3), 919–927. Shuttleworth, C.W., Brennan, A.M. & Connor, J.A. (2003) NAD(P)H fluorescence imaging of postsynaptic neuronal activation in murine hippocampal slices. J. Neuroscience 23, 3196–3208. van den Berg, P.A., Widengren, J., Hink, M.A., Rigler, R. & Visser, A.J. (2001) Fluorescence correlation spectroscopy of flavins and flavoenzymes: photochemical and photophysical aspects. Spectrochim. Acta A Mol. Biomol. Spectrosc. 57(11), 2135–2144. Zhang, Y. & Görner, H. (2009) Flavin-sensitized photo-oxidation of lysozyme and serum albumin. Photochem. Photobiol. 85(4), 943– 948.