HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: 92/1/199    most recent 01275.2003v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (31) Citing Articles via Google Scholar Google Scholar Articles by Reinert, K. C. Articles by Ebner, T. J. Search for Related Content PubMed PubMed Citation Articles by Reinert, K. C. Articles by Ebner, T. J.

Flavoprotein Autofluorescence Imaging of Neuronal Activation in the Cerebellar Cortex In Vivo

Department of Neuroscience, University of Minnesota, Minneapolis, Minnesota 55455

Submitted 31 December 2003; accepted in final form 20 February 2004

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Autofluorescence has been used as an indirect measure of neuronal activity in isolated cell cultures and brain slices, but only to a limited extent in vivo. Intrinsic fluorescence signals reflect the coupling between neuronal activity and mitochondrial metabolism, and are caused by the oxidation/reduction of flavoproteins or nicotinamide adenine dinucleotide (NADH). The present study evaluated the existence and properties of these autofluorescence signals in the cerebellar cortex of the ketamine/xylazine anesthetized mouse in vivo. Surface stimulation of the unstained cerebellar cortex evoked a narrow, transverse beam of optical activity consisting of a large amplitude, short latency increase in fluorescence followed by a longer duration decrease. The optimal wavelengths for this autofluorescence signal were 420–490 nm for excitation and 515–570 nm for emission, consistent with a flavoprotein origin. The amplitude of the optical signal was linearly related to stimulation amplitude and frequency, and its duration was linearly related to the duration of stimulation. Blocking synaptic transmission demonstrated that a majority of the autofluorescence signal is attributed to activating the postsynaptic targets of the parallel fibers. Hypothesized to be the result of oxidation and subsequent reduction of flavoproteins, blocking mitochondrial respiration with sodium cyanide or inactivation of flavoproteins with diphenyleneiodonium substantially reduced the optical signal. This reduction in the autofluorescence signal was accomplished without altering the presynaptic and postsynaptic components of the electrophysiological response. Results from reflectance imaging and blocking nitric oxide synthase demonstrated that the epifluorescence signal is not the result of changes in hemoglobin oxygenation or blood flow. This flavoprotein autofluorescence signal thus provides a powerful tool to monitor neuronal activity in vivo and its relationship to mitochondrial metabolism.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
The autofluorescence of nicotinamide dinucleotide (NADH) and flavoproteins has been used to monitor aerobic energy metabolism and indirectly monitor neuronal activity by exploiting the close coupling between neuronal activity and mitochondrial metabolism (Duchen 1992; Mironov and Richter 2001; Schuchmann et al. 2001). Neuronal activation is generally accompanied by an influx of Ca²⁺ ions into the cytoplasm. This increase in cytoplasmic Ca²⁺ is partially absorbed by mitochondria, and leads to a short-term increase in the oxidation of reducing equivalents, including NADH and flavoproteins, followed by a longer-duration reduction. This oxidation and reduction can be monitored through changes in the fluorescence of flavoproteins (excitation 430–500 nm, emission 520–590 nm; Chance et al. 1968), which fluoresce when oxidized and do not when reduced, yielding a biphasic response consisting of a brief increase in fluorescence followed by a decrease in fluorescence (Duchen 1992). More commonly, the fluorescence of NADH (peak excitation 340–360 nm, peak emission 450–480 nm), which fluoresces when reduced and does not when oxidized, has been used to monitor the initial oxidation as a transient decrease in fluorescence (Chance et al. 1962; Duchen 1992; Lewis and Schuette 1976).

The potential advantages of using these autofluorescence signals to monitor neuronal activity and/or mitochondrial metabolism are clear. Because these fluorophores are endogenous in all tissues, there is no need to introduce voltage, Ca²⁺, or pH-sensitive dyes, many of which are toxic or difficult to get into intact preparations (Cohen et al. 1978; Ebner and Chen 1995; Grinvald 1985; Lieke et al. 1989). Autofluorescence signals also do not have the spatial limitation of the hemodynamic response–based intrinsic optical signal, which is restricted to the resolution of the capillary bed supplying the tissue under examination (Erinjeri and Woolsey 2002; Frostig et al. 1990; Malonek and Grinvald 1996). Because these autofluorescence signals are mitochondrial in origin, there is the potential to resolve the activity of single cells (Duchen 1992).

Earlier investigations using autofluorescence signals primarily relied on monitoring NADH and were conducted in cell cultures or brain slices (Duchen 1992; Mironov and Richter 2001; Schuchmann et al. 2001). A recent study demonstrated that NADH autofluorescence imaging can be used to monitor the spatial and temporal properties of neuronal activity in the hippocampal slice (Shuttleworth et al. 2003). The question is whether these signals could be useful for monitoring neuronal activation in vivo. In the cerebral cortex, the NADH signal evoked by direct stimulation is weak, with an amplitude of 0.5–0.8% F/F and averaging multiple trials is needed to increase the signal-to-noise ratio (Lothman et al. 1975; Rosenthal and Jobsis 1971). Consequently, NADH autofluorescence has only found limited use in vivo, such as monitoring pathophysiological processes such as spreading depression or epileptiform activity (Hashimoto et al. 2000; Jobsis et al. 1971; Mayevsky and Chance 1974; O'Connor et al. 1973).

The flavoprotein-mediated signal in vitro had previously been reported to be 50–100 times smaller than the NADH signal (Aubin 1979), seemingly limiting its utility. However, recent reports have shown large-amplitude autofluorescence signals at flavoprotein wavelengths in both hippocampal and somatosensory cortical slices (Shibuki et al. 2003; Shuttleworth et al. 2003). Flavoprotein-mediated autofluorescence has also been used to monitor neuronal activity in the somatosensory cortex in vivo (Shibuki et al. 2003). The present study further characterized this intrinsic fluorescence signal in the cerebellar cortex in vivo, using the parallel fiber–Purkinje cell circuit. The wavelength selectivity and biphasic time course of the autofluorescence signal evoked by surface stimulation, as well as its dependency on stimulus parameters, were systematically examined. Previous studies in vivo did not dissociate the effects of mitochondrial blockers on oxidative metabolism from their effects on synaptic transmission or neuronal excitability (Shibuki et al. 2003). The present study aimed to demonstrate the flavoprotein origin of the signal by blocking mitochondrial respiration without altering cerebellar cortical neuronal excitability. The contributions of changes in hemoglobin oxygenation and blood flow on the autofluorescence signal were also assessed. An abstract of some of these results has been presented (Reinert et al. 2002).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Animal preparation

All animal experimentation was approved by the Institutional Animal Care and Use Committee of the University of Minnesota and conducted in conformity with the American Physiological Society's Guiding Principles in the Care and Use of Animals. Experimental details on the animal preparation and optical imaging techniques have been provided in previous publications (Chen et al. 1998; Gao et al. 2003a) and therefore are only described briefly. Adult male FVB mice (Charles River Laboratories, Wilmington, MA), ages 3–10 mo, were anesthetized by intramuscular injection of 1.0 ml/kg of a ketamine (60 mg/ml) –xylazine (3 mg/ml) cocktail, supplemented as needed by 0.5 ml/kg doses. Animals were mechanically ventilated and paralyzed with an intramuscular injection of gallamine triethiodide (0.05 ml, 20 mg/ml). The animal was placed in a stereotaxic frame and body temperature was feedback-regulated through a rectal temperature probe connected to a heating pad. The electrocardiogram was monitored to assess the depth of anesthesia. A craniotomy exposed Crus I and II. A watertight chamber of acrylic was built up around the exposed cortex and filled with an artificial Ringer solution gassed with 95% O₂-5% CO₂.

In some experiments CNQX (6-cyano-7-nitroquinoxaline-2,3-dione disodium salt), NaCN (sodium cyanide), DPI (diphenyleneiodonium chloride), or L-nitroarginine methyl ester (L-NAME) were added to Ringer solution and perfused into the chamber. All drugs were obtained from Sigma Chemical (St. Louis, MO). For NaCN and DPI a range of concentrations were tested (100 µM to 1 mM for NaCN and 100 nM to 100 µM for DPI) to find a concentration that blocked the optical response without affecting the field potential responses to surface stimulation. The final concentrations of 250 µM NaCN and 50 µM DPI were found to successfully block the optical response while leaving the field potential responses intact for 30 min. Higher concentrations affected both the optical signals and field potentials, and lower concentrations were insufficient to completely block the optical response.

Electrical stimulation and electrophysiological monitoring techniques

Parallel fiber stimulation was delivered through a paralyene-coated tungsten microelectrode (1–3 M) placed just below the cerebellar surface. Typical parameters for surface stimulation were a train of 200-µA, 100-µs pulses at 10 Hz for 10 s. To evaluate the dependency of the autofluorescence signal on stimulation parameters, amplitude (50–300 µA), frequency (5–20 Hz), and train duration (1–20 s) were systematically varied in separate experiments. In some experiments, simultaneous extracellular recordings of the evoked field potentials were obtained from the molecular layer with glass microelectrodes (2 M NaCl, 2–5 M) using conventional electrophysiological techniques (Chen et al. 2001). These field potentials were digitized (50 kHz), averaged on-line, and stored for additional off-line analysis. The presynaptic component of the field potential, consisting of the initial triphasic peak–trough–peak (P₁, N₁, P₂) wave, was measured as the difference between the maximum of P₁ and the minimum of N₁. The postsynaptic component, the subsequent negativity (N₂), was measured as the absolute value of its amplitude relative to the baseline preceding the stimulus artifact (Chen et al. 2001; Eccles et al. 1967).

Optical imaging

Images of the cerebellar surface were acquired by fixing the stereotaxic frame to an X-Y stage mounted on a modified Nikon epifluorescence microscope fitted with a 4x objective and a Quantix cooled charge coupled device (CCD) camera with 12-bit digitization (Roper Scientific, Tucson, AZ). A 100-W mercury–xenon lamp (Hamamatsu Photonics) powered by an Opti Quip power supply (model 1600) was used as the light source. The images were binned 2 x 2 for image frames of 265 x 256 pixels, resulting in a final pixel resolution of about 10 x 10 µm. In one series of experiments, the excitation and emission filters, and the dichroic mirror were systematically changed to determine the wavelength sensitivity of the autofluorescence signal. After determining the optimal wavelengths, remaining experiments were performed with a band-pass excitation filter (455 ± 35 nm), an extended reflectance dichroic mirror (500 nm), and a >515 nm long-pass emission filter. A typical acquisition protocol included a series of 20 control frames followed by series of 150 experimental frames, with an exposure time of 1 s for each frame. Surface stimulation was initiated at the onset of the experimental frames.

In some experiments, the autofluorescence signal was compared with the optical signal obtained with neutral red (Chen et al. 1998, 2001). In these experiments, two 0.2 ml intraperitoneal injections of a 35 mM solution of neutral red (3-amino-m-dimethylamino-2-methylphenazine hydrochloride) were used to stain the brain. For neutral red imaging the excitation filter was 551 ± 5 nm, the long-pass emission filter was >590 nm, and the dichroic mirror was 565 nm. The autofluorescence signal was also compared with the hemodynamic intrinsic optical signal (Frostig et al. 1990; Grinvald et al. 1986; Malonek et al. 1997) at the wavelengths used for excitation and emission of the autofluorescence signal. To monitor reflectance changes in these experiments the dichroic mirror was replaced with a half-silvered mirror (50/50), and the excitation and emission filters were both replaced with band-pass filters for either 420–490 nm or 510–560 nm. Because the background intensity was high in reflectance mode, a neutral density filter (ND32) was used to reduce the input light and a glass coverslip was placed over the exposed cerebellar cortex to reduce glare. Use of the coverslip and neutral density filter had no effect on the autofluorescence signal. Because the amplitude of these intrinsic reflectance signals were generally small, 4 trials were averaged to improve the signal-to-noise ratio.

Data analysis

The first step in the analyses of the optical signals was to generate a series of difference images by subtracting the average of 18 control frames, referred to as the control average, from each control and experimental frame. These difference images were then divided by the control average, resulting in images in which the intensity value of each pixel reflects the F/F change in fluorescence intensity relative to the average of the control frames. To quantify the amplitude and time course of the autofluorescence response, first a primary region of interest (ROI) was defined that consisted of a 5-pixel-wide line placed on the center of the evoked optical beam. Second, 2 background ROIs, each 5 pixels wide, were defined on both sides of the optical beam offset by about 500 µm. The average F/F value within each of the ROIs was then determined and the F/F values for the background ROIs were averaged together and subtracted from the average F/F for the primary ROI to correct for fluctuations in background fluorescence or light intensity. The result was a F/F for the optical response along the beam that reflected only the time course of the localized optical response.

As described in the RESULTS, the autofluorescence signal consisted of an initial increase in fluorescence (light phase) followed by a decrease in fluorescence (dark phase). The next step was to quantify the peak amplitude of each phase by averaging the 3 s around the peak of the light phase and 20 s around the minimum of the dark phase trough. This average was then used as a measure of the amplitude of the light and dark phases of the signal. To analyze the effects of drugs on the light and dark phases of the signal these amplitudes were evaluated by an ANOVA with randomized block design followed by Duncan's post hoc test. To study the effects of stimulus duration, durations of the light and dark phases were also determined. The duration of the light phase was defined as the period during which the optical response was significantly increased above the background fluorescence during the control period. The duration of the dark phase was defined as the period from the time at which the amplitude of the optical response dropped below control levels to the time at which it returned to the baseline. Because changing the duration of stimulation dramatically affected the shape and duration of both phases of the optical response, averaging several seconds around the peak was not accurate, particularly at the shorter stimulus durations. The peak value of the light phase provided a more consistent measure because it was relatively unaffected by the changes in shape and duration. Similar effects of stimulation duration on the dark phase were accounted for by averaging the 6 and 10 s around the minimum of the trough for the 1- and 2-s stimulations, respectively, and 20-s averages for all others. The resulting values for amplitude and duration of the light and dark phases and their relationship to stimulus parameters were evaluated using simple linear regressions.

For displayed images, the frames with the maxima of the light phase (typically at 3 s after stimulation) and the minima of the dark phase (typically at 40 s after stimulation) were selected from the series of F/F images. These images were then scaled to ±3% F/F.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Without the presence of any contrast agents or optical dyes, stimulation of the cerebellar surface in the mouse evoked a beamlike optical response when imaged using excitation at 420–490 nm and emission >515 nm (Fig. 1A). This autofluorescence signal was biphasic, consisting of a rapid, initial increase in fluorescence (light phase) followed by a more prolonged decrease in fluorescence (dark phase). As shown in Fig. 1B for a stimulation train consisting of 200-µA, 100-µs pulses at 10 Hz for 10 s, the amplitude of the optical response peaked at 1.57% F/F for the light phase. This light phase was initiated within 50 ± 30 ms of stimulation onset when examined at faster exposure times (data not shown). The subsequent dark phase persisted for about 100 s, reaching an amplitude of –1.56% F/F at 58 s after the onset of the stimulation. Because of this large amplitude, the optical signal was easily detected without averaging. Note that in the transition from the light phase to the dark phase, the optical response did not necessarily change in a uniform manner along the length and width of the beam. The light phase of the optical beam extended the length of the surface of the folium in the field of view (2.5 mm) and had a width of 120 µm. The width and length of the evoked optical beam are consistent with activation of parallel fibers induced by surface stimulation and their postsynaptic targets (i.e., Purkinje cells and cerebellar interneurons) (Eccles et al. 1967; Ito 1984).

View larger version (55K): [in this window] [in a new window]   FIG. 1. Example of the autofluorescence signal evoked by surface stimulation. A: image on the left of the background fluorescence shows a raw image of the exposed cerebellar cortex with Crus I and II labeled. Subsequent frames are subtracted F/F images showing the optical response to surface stimulation (200-µA, 100-µs pulses at 10 Hz for 10 s) at 3 time points (3, 16, and 40 s) from the onset of stimulation. Images were scaled to ±3% F/F. Orientation of the image is shown in the rightmost image and scale bar is shown in the left most image. B: time course of the optical response. Graph shows the change in fluorescence, F/F, of the optical signal along the length of the optical beam during the same imaging trial as in A. Bar below the trace denotes the stimulation period (stim).

View larger version (73K): [in this window] [in a new window]   FIG. 2. Dependency of the autofluorescence signal on excitation and emission wavelengths. Light and dark phase responses (the top and bottom half of each pair of images, 3 and 40 s after stimulation onset, respectively) to surface stimulation are shown for a series of excitation (x-axis) and emission wavelengths (y-axis). An appropriate dichroic mirror was chosen for each excitation filter (400 nm for 330–80 nm excitation, 455 nm for 400–40 nm excitation, and 500 nm for 450–90 nm). For all experiments surface stimulation parameters were 200-µA, 100-µs pulses at 10 Hz for 10 s. Optimal wavelength ranges for the autofluorescence signal of interest and a lower wavelength signal are outlined in dashed and solid boxes, respectively.

View larger version (68K): [in this window] [in a new window]   FIG. 3. Comparison of the autofluorescence signal with the neutral red signal. A–B: light and dark phase of the autofluorescence signal evoked by surface stimulation (200-µA, 100-µs pulses at 10 Hz for 10 s). Imaging was performed using an excitation filter of 420–490 nm, long-pass emission filter >515 nm and a 500-nm dichroic mirror. Images in A show the light phase and dark phase at 3 and 40 s after stimulation onset, respectively. Time course of the optical signal is shown in B. C–D: similar images at 3 and 40 s after stimulation onset showing the optical signal and time course in response to surface stimulation when imaging at 330- to 380-nm excitation and >420-nm emission with a 400-nm dichroic mirror. Note the change in scale in D. E–F: optical response using neutral red, showing images at 3 and 40 s after stimulation onset and time course of the optical signal. Imaging was performed at wavelengths optimal for neutral red with excitation at 546 ± 5 nm, >590-nm emission and a 565-nm dichroic mirror.

View larger version (20K): [in this window] [in a new window]   FIG. 4. Effect of stimulus parameters on the autofluorescence optical signal. For each parameter of interest the average amplitude ± SD is shown during the light phase (gray bar) and dark phase (black bar) of the optical response. A: signal amplitude as stimulation amplitude was varied from 50 to 250 µA in 50-µA increments (100-µs pulses, 10 Hz, 10 s). B: signal amplitude as stimulation frequency was varied from 5 to 20 Hz as indicated using a 10-s stimulation train of 200-µA, 100-µs pulses. C and D: signal amplitude and duration as stimulation duration was varied from 1 to 20 s using a stimulation train of 200-µA, 100-µs pulses at 10 Hz.

The effect of stimulus duration on the autofluorescence signal was also examined by testing train durations of 1 to 20 s (10-Hz, 200-µA, 100-µs pulses). As shown in Fig. 4C, the amplitude of the light phase did not increase with stimulus duration, maintaining an average value of 1.57 ± 0.45% F/F, and the linear regression did not yield a significant fit (R² = 0.01, F-test, P = 0.55, slope = 0.007% F/F · s^–1, n = 6). At 10-Hz stimulation the response amplitude would not necessarily be expected to reflect the duration of the stimulus. However, the duration of the light phase closely followed and was linearly related to the duration of the stimulation (Fig. 4D, R² = 0.83, F-test, P < 0.01, slope = 0.78 s · s^–1). In contrast, the amplitude of the dark phase displayed a significant proportional increase with train duration, increasing from –0.29 ± 0.19% F/F at 1 s to –2.00 ± 0.71% F/F at 20 s (R² = 0.71, F-test, P < 0.01, slope = –0.094% F/F · s^–1, n = 6). The duration of the dark phase also increased with stimulation duration but persisted much longer (Fig. 4D, R² = 0.62, F-test, P < 0.01, slope = 3.72 s · s^–1). The differences in the responses of the light phase and dark phase with the duration of stimulation suggest that the 2 phases may have different cellular origins.

The responses to a small number of pulses in a short-duration train provide additional information on the size, time course, and input–output properties of the signal. As shown in Fig. 5, A and B when imaging at faster frame rates (100-ms exposure), a single surface stimulation pulse evoked a statistically significant increase in F/F relative to the control frames (P < 0.01, Student's t-test). The optical response is just detectable in the images. Two or 3 stimulation pulses at 100 Hz generated robust, biphasic optical responses, showing that both the light phase and the dark phase are very sensitive to high-frequency stimulation.

View larger version (67K): [in this window] [in a new window]   FIG. 5. Optical response evoked by high-frequency burst stimulation. A: series of light and dark phase images showing the responses to a single pulse and 2 and 3 pulses at 100 Hz (200-µA, 200-µs pulses). In these studies the frame exposure time was 100 ms. B: time courses of the optical responses for the images shown in A. Arrow below the x-axis denotes the time of stimulation.

View larger version (107K): [in this window] [in a new window]   FIG. 6. Effects of Ca2+ removal and blocking synaptic transmission on optical response to surface stimulation. A: series of images showing the light and dark phases (labeled L and D, 3 and 40 s after stimulation onset, respectively) of the optical response to surface stimulation in normal Ringer solution, in 0 Ca2+ Ringer solution and its washout, and then 50 µM CNQX in normal Ringer solution, followed by an additional washout. B: average ± SD (n = 4 animals) F/F response amplitudes for the light (open bar) and dark (black bar) phases of the optical response in control, 0 Ca2+ Ringer solution and washout, and then CNQX and its washout. Asterisk (*) denotes significant change from control (ANOVA followed by Duncan's post hoc, P < 0.05).

View larger version (76K): [in this window] [in a new window]   FIG. 7. Effect of blocking mitochondrial metabolism on the autofluorescence signal. A: examples of the light and dark phase (labeled L and D, 3 and 40 s after stimulation onset, respectively) optical responses to surface stimulation in normal Ringer solution, NaCN (250 µM), and washout. B: average ± SD F/F (n = 3 animals) for the light (open bar) and dark phases (black bar) for control, 250 µM, NaCN, and washout. * denotes significant change from control (ANOVA followed by Duncan's post hoc, P < 0.05). C: background intensity (arbitrary units) for control, NaCN, and washout expressed as a % of control. Note the significant decrease in the presence of NaCN relative to the control. D: top traces are examples of the field potentials evoked by surface stimulation (average of 100) for control, NaCN, and washout. P1, N1, P2, and N2 components are labeled. Bottom: average ± SD (n = 4 animals) of the amplitude of the presynaptic (P1–N1 amplitude) and postsynaptic component (N2 amplitude).

View larger version (69K): [in this window] [in a new window]   FIG. 8. Effect of blocking flavoproteins on the autofluorescence signal. A: examples of the light and dark phase (labeled L and D, 3 and 40 s after stimulation onset, respectively) optical responses to surface stimulation in normal Ringer solution, 50 µM diphenyleneiodonium chloride (DPI), and after washout. B: average ± SD F/F for the light phase (open bar) and dark phase (black bar) for control DPI and washout. * denotes significant change (P < 0.05) from control (ANOVA followed by Duncan's post hoc comparison, n = 5 animals). C: top traces are example field potentials evoked by surface stimulation for control, DPI, and washout. Bottom: average ± SD of presynaptic (P1–N1) and postsynaptic (N2) response amplitudes as described in Fig. 7 (n = 3 animals).

View larger version (33K): [in this window] [in a new window]   FIG. 9. Comparison between the autofluorescence signal and intrinsic reflectance signals. A: single trial (4 gray traces) and averaged (black traces) signal time course for the reflectance signal recorded at 420–90 nm. B: single trial (4 gray traces) and averaged (black traces) signal time course for the reflectance signal recorded at 510–60 nm. C: average ± SD (n = 4 animals) F/F light (open bar, 3 s) and dark (black bar, 40 s) phase optical responses for the autofluorescence signal and R/R light (open bar, 1 s) and dark (black bar, 30 s) phases optical responses for the reflectance signals recorded at 420–90 and 510–60 nm (200-µA, 100-µs pulses at 10 Hz for 10 s).

View larger version (22K): [in this window] [in a new window]   FIG. 10. Dependency of the autofluorescence signal on hemodynamics. A: average ± SD (n = 3 animals) F/F light (open bar) and dark (black bar) phase optical responses in normal Ringer solution, L-nitroarginine methyl ester (L-NAME), and after washout. For each condition the optical response was examined at 3 stimulation amplitudes (100, 200, and 300 µA, as indicated in the open bars) using 100-µs pulses at 10 Hz for 10 s. B: example of the reflectance signal at 510–560 nm before and after application of L-NAME (300-µA, 100-µs pulses at 10 Hz for 10 s).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

Electrical stimulation of the unstained cerebellar surface resulted in a transverse, "beamlike" optical response during epifluorescence imaging using excitation at 420–490 nm and emission >515 nm. This intrinsic autofluorescence signal consisted of an initial, rapid increase followed by a slower, prolonged decrease in fluorescence. The beam extended the entire length of the visible folium and was consistent with the length of the parallel fibers in the mouse (Soha et al. 1997). Therefore the origin of the signal is consistent with the activation of the parallel fibers and their postsynaptic targets, Purkinje cells, and inhibitory interneurons (Eccles et al. 1967; Ito 1984).

The amplitude and time course of the optical signal were dependent on stimulation parameters. Peak amplitudes of both the light and dark phases were linearly related to stimulation amplitude (50–250 µA) and stimulation frequency (5–20 Hz). At 10 Hz the amplitude of the initial increase in fluorescence was not affected by changes in stimulation duration but the duration of the light phase closely followed the duration of the stimulus train. At 100 Hz, very brief stimulation evoked strong responses, and increasing the number of stimulation pulses increased the amplitude of both the light and dark phases. Several factors are likely to contribute to the increased responses with a high-frequency burst including presynaptic facilitation of transmitter release from parallel fibers (Dittman et al. 2000; Isope and Barbour 2002) and activation of postsynaptic mGluR1 receptors (Batchelor and Garthwaite 1997; Tempia et al. 1998). The input–output characteristics and the robust responses to brief stimulation bursts suggest that the autofluorescence signal is closely coupled to neuronal activity.

The amplitude, timing, and sensitivity of the autofluorescence signal further suggest its utility for monitoring neuronal activity. Signal amplitude is quite large, with the peak of the light phase averaging about 1.25% F/F for a train of 200-µA, 100-µs pulses at 10 Hz for 10 s (Fig. 4) and can be detected without averaging in response to a single stimulation pulse (Fig. 5). The optical response is substantially larger than the intrinsic optical signal (Frostig et al. 1990; Grinvald et al. 1986; Malonek et al. 1997) and most voltage-sensitive dye signals obtained in vivo in mammalian preparations (Cohen et al. 1978; Ebner and Chen 1995; Grinvald 1985; Lieke et al. 1989). The peak light phase response is somewhat smaller, but much faster than the signal observed with neutral red (Fig. 3) (Chen et al. 1998, 2001). The properties of the autofluorescence signal, particularly the light phase, thus make it an excellent signal for monitoring neuronal activity. Autofluorescence signals have been used to monitor the responses to peripheral stimuli in the somatosensory cortex (Shibuki et al. 2003) and in the cerebellar cortex (Gao et al. 2003b).

Source of the autofluorescence

There are relatively few endogenous fluorophores present in neurons, mainly the aromatic amino acids and transmitters such as tryptophan and serotonin (Lillard and Yeung 1997; Maiti et al. 1997; Williams et al. 1999), and metabolic reducing equivalents like NADH (Aubin 1979; Chance et al. 1962) and flavoproteins (Chance et al. 1968; Duchen 1992) involved in mitochondrial metabolism. Of these, only NADH and flavoproteins have changes in fluorescence that are modulated by neuronal activity through the resultant change in mitochondrial metabolism, with NADH fluorescing when reduced and flavoproteins fluorescing when oxidized. The optimal wavelengths for the autofluorescence signal, excitation at 420–490 nm and emission at 500–570 nm, are consistent with the established ranges for flavoproteins (Benson et al. 1979; Duchen 1992).

When neurons depolarize, opening Ca²⁺ permeable channels leads to an increase in Ca²⁺ ions in the intracellular compartment (Fig. 11A). Shibuki and colleagues (2003) have already demonstrated in slices of auditory cortex that this Ca²⁺ increase precedes the onset of the autofluorescence signal. It has also been established that this increase in Ca²⁺ is partially taken up by mitochondria (Budd and Nicholls 1996) through uniport mechanisms driven by the –190 mV potential across the inner mitochondrial membrane (Gunter et al. 1998; Rizzuto et al. 2000). The uptake of Ca²⁺ depolarizes the mitochondria and induces calcium cycling (Fig. 11B), both of which reduce the proton gradient across the inner mitochondrial membrane, resulting in a short-term oxidation of the respiratory chain as it restores the gradient (Nicholls and Ferguson 2002). After the Ca²⁺ enters the mitochondrial matrix, it activates the pyruvate dehydrogenase complex (PDC, Fig. 11E) and other dehydrogenases associated with the tricarboxylic acid cycle (Hansford 1994; McCormack et al. 1990), resulting in a delayed increase in NADH along with reduction of the respiratory chain (Fig. 11F). This oxidation and reduction generate the biphasic time course of the autofluorescence signal, corresponding to the oxidation and subsequent reduction of flavoproteins (Duchen 1992). The putative NADH signal at 300–370 nm with the predicted inverse time course provided further support for the flavoprotein hypothesis, given that NADH fluoresces when reduced.

View larger version (34K): [in this window] [in a new window]   FIG. 11. Schematic of the sources of the autofluorescence signal. See text for detailed description. A: voltage-gated calcium channels and Na+-K+ pump activated by depolarization. B: Ca2+ entry into the mitochondria through a uniport mechanism and Ca2+ cycling. C: ATP synthase activated by ADP. D: glycolysis and lactate dehydrogenase (LD) provide NADH and pyruvate. E: pyruvate dehydrogenase complex (PDC) activated by Ca2+ converts pyruvate to fuel for the tricarboxylic acid cycle. F: respiratory chain complexes (I–IV) and cytochrome C (Cyt C) are oxidized and reduced. It is the oxidation and reduction of flavin adenine dinucleotide (FAD) and flavin mononucleotide (FMN), the flavin moieties in respiratory complexes I and II, that generates the autofluorescence signal. Other flavoproteins associated with lipoamide dehydrogenase complexes (not shown except for PDC) also participate in this reduction through the shared NADH/NAD+ pool.

Although the predominant theory of mitochondrial metabolism suggests that the signal is Ca²⁺ dependent (Duchen 1992; Shibuki 1989), the signal was not completely blocked by 0 Ca²⁺ Ringer solution. Similar observations have been made in the hippocampal slice concerning the NADH autofluorescence signal (Shuttleworth et al. 2003). One possibility is that the removal of Ca²⁺ from the bath and extracellular fluid is not complete or does not eliminate the release of Ca²⁺ from intracellular stores during depolarization. Another possibility is that the induction of mitochondrial metabolism also depends on the depletion of ATP and production of ADP by the Na⁺-K⁺ pump (Fig. 11A) that restores the ion gradients across the plasma membrane (Magistretti et al. 1999; Shuttleworth et al. 2003). This depletion of ATP and the production of ADP, which are translated across the inner mitochondrial membrane, increase the activity of ATP synthase and reduce the proton gradient (Fig. 11C), resulting in an immediate oxidation of the components of the respiratory chain (Nicholls and Ferguson 2002). The increased ADP also activates key enzymes in glycolysis (Fig. 11D) and the tricarboxylic acid cycle (Fig. 11E), resulting in increased NADH and reduction of the respiratory chain (Fig. 11F). Thus both increases in intracellular Ca²⁺ and ATP depletion can generate the biphasic time course of the autofluorescence signal. Ultimately, the induction of mitochondrial metabolism and the autofluorescence signal probably depend on a combination of intracellular Ca²⁺ increases and ATP depletion determined by the metabolic and excitatory requirements of individual neurons.

A more direct test of the mitochondrial origin of the signal is to block the electron transport chain. Cyanide blocks cytochrome C oxidase (Fig. 11, complex IV), the last enzyme in the chain, and produces a backup of electrons throughout the chain, maximally reducing all of the components in the chain, including the flavoproteins (Palmer 1993). Because the concentration of NaCN was adjusted to block the optical response, while leaving the pre- and postsynaptic components of the electrophysiological response intact, we can conclude that the reduction in the optical signal was neither the result of alterations in the excitability of the parallel fibers or Purkinje cells, nor caused by the cytotoxic effects of NaCN. There was also a significant decrease in the background fluorescence during the superfusion, consistent with a general reduction of flavoproteins in the electron transport chain across the exposed cerebellar cortex, which also recovered with washout. Both the effects of NaCN on the optical response and background fluorescence are completely consistent with a mitochondrial origin of this autofluorescence signal.

The direct role of flavoproteins as the source of the autofluorescence signal was also evaluated using DPI, which irreversibly inactivates flavoproteins (Fig. 11, FMN and FAD) by forming a covalent bond (Majander et al. 1994). In the somatosensory cortex in vivo, Shibuki and colleagues (2003) achieved a complete signal blockade, but the electrophysiological response was severely affected. By adjusting the concentration of DPI, we achieved complete blockade of both phases of the autofluorescence signal, while leaving the electrophysiological responses intact, thus leaving little question as to the source of the signal. Because this inactivation of flavoproteins is irreversible, the autofluorescence signal did not recover with repeated rinsing of the chamber. Because the excitability of the cerebellar cortical circuitry and mitochondrial metabolism were uncoupled by both NaCN and DPI, thus blocking the signal without affecting neuronal function, we conclude that mitochondrial flavoproteins are the source of the observed autofluorescence signal.

Several differences in the properties of the light and dark phases suggest different cellular origins. CNQX almost completely blocked the light phase of the autofluorescence signal, but reduced the dark phase by only 50%, demonstrating that the dark phase is not completely dependent on AMPA-mediated synaptic transmission. In contrast, 0 Ca²⁺ Ringer solution, which reduces the induction of mitochondrial metabolism along with blocking synaptic transmission, affected both phases equally. Both neurons and glia use mitochondrial metabolism that responds to calcium influx and ATP depletion during repolarization. Neurons are directly affected by stimulation, and have a very rapid metabolic response that closely follows stimulation. Glia are also depolarized in response to synaptic transmission and neuronal depolarization by sensing increases in extracellular glutamate and potassium, leading to metabolic induction on a slower time scale than that for neurons (Pellerin and Magistretti 1994; Takahashi et al. 1995). Although neurons are hypothesized to rely on the oxidative metabolism of lactate, glia primarily use glycolysis (Fig. 11D), leading to an accumulation of NADH and reduction of the electron transport chain (Magistretti 2000; Mangia et al. 2003). These 2 factors together suggest that the source of the light phase is primarily neuronal, whereas the dark phase has a large contribution from glia (Kasischke et al. 2003). The different cellular origin would also account for the differential effects of stimulus duration on the two phases of the signal. Because the light phase depends on the rapid response of neurons to the initial depolarization, successive stimulation at low frequency (i.e., 10 Hz) would have little effect in increasing its amplitude. In contrast, the buildup of metabolic demand in glia with successive stimulation would lead to the longer and larger dark phase observed with longer-duration stimulation.

A remaining concern is that the signal is contaminated with the intrinsic optical reflectance signal. This is especially significant because cytochromes and hemoglobin absorb light in both of the wavelengths used for excitation and emission of the autofluorescence signal (Frostig et al. 1990; Jobsis et al. 1977; La Manna et al. 1987; Malonek and Grinvald 1997). Monitoring reflectance changes at these wavelengths demonstrated that these signals were quite small and did not contribute significantly to the autofluorescence signal. The lack of an effect from blocking nitric oxide synthase with L-NAME suggests that changes in blood flow also did not contribute significantly to the autofluorescence signal.

This activity-dependent autofluorescence signal also raises a concern for recent efforts into the generation of transgenic mice expressing fluorescent proteins for monitoring various aspects of neuronal activation, especially the yellow variants of green fluorescent protein. Examples include green fluorescent protein–based voltage-sensitive fluorescent proteins (Knopfel et al. 2003), cyclic guanosine monophosphate indicators (Honda et al. 2001), and the Ca²⁺-sensitive chameleon constructs (Miyawaki et al. 1999). These proteins fluoresce in response to various changes evoked by neuronal activity at wavelengths that significantly overlap with those used to monitor the autofluorescence signal described in this study. This overlap in wavelengths presents a particularly difficult confound because, even if the transgenic protein is expressed and functional, distinguishing between the fluorescence generated by the transgenic protein and the autofluorescence signal may be difficult. This confound is magnified by the large amplitude and time course of the autofluorescence signal.

The time course, wavelength selectivity, and results of the pharmacological manipulations are all consistent with the hypothesis that the autofluorescence signal reflects the oxidation and reduction of mitochondrial flavoproteins. Because flavoproteins are a component of all tissues, this autofluorescence signal is likely to be a useful tool for monitoring neuronal activity without exogenous dyes throughout the nervous system. The large amplitude of the signal and linear relationship with stimulation amplitude and frequency demonstrate its utility for monitoring neuronal activation in vivo. Equally important, this autofluorescence signal can be used to monitor neuronal activity–induced changes in mitochondrial metabolism in vivo.

	GRANTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
This work was supported in part by National Institute of Neurological Disorders and Stroke Grant P01-NS-31318 and a grant from the Bob Allison Ataxia Research Center.

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
We thank Dr. Jan Dubinsky for helpful discussions, Y. Pan and L. Zhuo for animal preparation, M. McPhee for graphics, and B. Swanson for preparation of the manuscript.

Present address for R. Dunbar: Buena Vista University, School of Science, Storm Lake, IA 50588.

	FOOTNOTES

 
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Address for reprint requests and other correspondence: T. J. Ebner, Department of Neuroscience, University of Minnesota, Lions Research Building, Room 421, 2001 Sixth St. S.E., Minneapolis, MN 55455 (E-mail: ebner001{at}umn.edu).

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Aubin JE. Autofluorescence of viable cultured mammalian cells. J Histochem Cytochem 27: 36–43, 1979.[Abstract]

Batchelor AM and Garthwaite J. Frequency detection and temporally dispersed synaptic signal association through a metabotropic glutamate receptor pathway. Nature 385: 74–77, 1997.[CrossRef][Medline]

Benson RC, Meyer RA, Zaruba ME, and McKhann GM. Cellular autofluorescence—is it due to flavins? J Histochem Cytochem 27: 44–48, 1979.[Abstract]

Budd SL and Nicholls DG. A reevaluation of the role of mitochondria in neuronal Ca²⁺ homeostasis. J Neurochem 66: 403–411, 1996.[ISI][Medline]

Chance B, Cohen P, Jobsis FF, and Schoener B. Intracellular oxidation-reduction states in vivo. Science 137: 499–508, 1962.[Abstract/Free Full Text]

Chance B, Mela L, and Wong D. Flavins and flavoproteins. In: Proceedings of the 2nd Conference on Flavins and Flavoproteins, edited by Kunio Y. Tokyo: Univ. of Tokyo Press, 1968.

Chen G, Dunbar RL, Gao W, and Ebner TJ. Role of calcium, glutamate neurotransmission, and nitric oxide in spreading acidification and depression in the cerebellar cortex. J Neurosci 21: 9877–9887, 2001.[Abstract/Free Full Text]

Chen G, Hanson CL, and Ebner TJ. Optical responses evoked by cerebellar surface stimulation in vivo using neutral red. Neuroscience 84: 645–668, 1998.[CrossRef][ISI][Medline]

Cohen LB, Salzberg BM, and Grinvald A. Optical methods for monitoring neuron activity. Annu Rev Neurosci 1: 171–182, 1978.[CrossRef][ISI]

Dittman JS, Kreitzer AC, and Regehr WG. Interplay between facilitation, depression, and residual calcium at three presynaptic terminals. J Neurosci 20: 1374–1385, 2000.[Abstract/Free Full Text]

Duchen MR. Ca²⁺-dependent changes in the mitochondrial energetics in single dissociated mouse sensory neurons. Biochem J 283: 41–50, 1992.[Medline]

Ebner TJ and Chen G. Use of voltage-sensitive dyes and optical recordings in the central nervous system. Prog Neurobiol 46: 463–506, 1995.[ISI][Medline]

Eccles JC, Ito M, and