Abstract
The fluorescent pigment lipofuscin accumulates with age in the cytoplasm of cells of the CNS. Because of its broad excitation and emission spectra, the presence of lipofuscin-like autofluorescence complicates the use of fluorescence microscopy (e.g., fluorescent retrograde tract tracing and fluorescence immunocytochemistry). In this study we examined several chemical treatments of tissue sections for their ability to reduce or eliminate lipofuscin-like autofluorescence without adversely affecting other fluorescent labels. We found that 1-10 mM CuSO4 in 50 mM ammonium acetate buffer (pH 5) or 1% Sudan Black B (SB) in 70% ethanol reduced or eliminated lipofuscin autofluorescence in sections of monkey, human, or rat neural tissue. These treatments also slightly reduced the intensity of immunofluorescent labeling and fluorescent retrograde tract tracers. However, the reduction of these fluorophores was far less dramatic than that for the lipofuscin-like compound. We conclude that treatment of tissue with CuSO4 or SB provides a reasonable compromise between reduction of lipofuscin-like fluorescence and maintenance of specific fluorescent labels.
Keywords
A
Materials and Methods
Experimental Animals, Retrograde Tract Tracing, and Tissue Preparation
Experimental protocols were approved by the Institutional Animal Care and Use Committee of the University of Minnesota. Tissue from two male Rhesus monkeys (Macaca mulatta, 4-5 kg body weight; Sierra Biomedical, Sparks, NV) were used in these studies. To study the effect of lipofuscin reduction on retrograde tract tracers, several tracers were injected into either monkeys or rats. The monkeys were maintained under isofluorane anesthesia and an incision was made at the level of the lumbar enlargement. A vertebra was partially removed to expose the spinal cord. A syringe (Hamilton Gastight; Reno, NV) equipped with a glass micropipette filled with retrograde tracer was introduced into the dorsal spinal cord. One monkey was injected with a mixture of 2% True Blue and 2% Fast Blue (Sigma Chemical; St Louis, MO); the other was injected with a 2% wheat germ agglutinin-horseradish peroxidase (WGA-HRP; Sigma) solution. Injection micropipettes were pulled from 1.5 mm outer diameter, 1.2 mm inner diameter glass capillary tubing using a micropipette puller (Narishige; Tokyo, Japan) and the tips trimmed to a diameter of ~50 μm. Pressure injections (100-500 nl at a rate of ~100 nl/min) were made into each monkey; after each injection, the micropipette was held in place for 5 min to reduce backwelling through the pipette tract. The monkeys were allowed to survive 5-10 days before they were sacrificed by vascular perfusion after being deeply anesthetized (10 mg/kg ketamine, 0.1 mg/kg butorphanol, and 0.5 mg/kg acepromazine). The monkeys were first perfused with calcium-free Tyrode's solution (116 mM NaCl, 5 mM KCl, 2 mM MgCl2, 400 μM MgSO4, 1.2 mM NaH2PO4, 26 mM NaHCO3, and 2.9 mM glucose), followed by fixative (4% w/v depolymerized paraformaldehyde in 160 mM phosphate buffer containing 14% v/v saturated aqueous picric acid, pH 6.9). After fixation, the animals were perfused with 5% w/v sucrose in 100 mM phosphate buffer, pH 7.2. The brains and spinal cords were removed and immersed in 5% sucrose overnight before sectioning. Tissue was sectioned either on a cryostat (Bright Instruments; Huntington, UK) at 10 μm and stored at −20C or on a freezing microtome (Leitz Instruments; Heidelberg, Germany) at 20-50 μm and stored in PBS at 4C for 2-3 days.
Two Sprague-Dawley rats (Harlan; Madison, WI) were used in this study. One rat (150 g body weight) was anesthetized with 75 mg/kg ketamine, 5 mg/kg xylazine, and 1 mg/kg acepromazine. Hypoglossal motoneurons in this rat were retrogradely labeled by injection into each side of the tongue with 500 nl 10% hydroxystilbamidine (Schmued and Fallon 1986; Wessendorf 1991) (Fluoro-Gold, FG; Fluorochrome, Englewood, CO). Injections were made at a rate of ~100 nl/min; the rat was allowed to survive 2-3 days before it was sacrificed. The second rat (about 7 months old, 300 g body weight) was used as a source of lipofuscin-containing tissue. Both rats were sacrificed by perfusion-fixation; cryostat tissue sectioning was accomplished as described for monkeys.
Lipofuscin is found throughout the neuraxis. Images obtained from cryostat sections of monkey neural tissue using a fluorescent microscope and UV filters. These images show the widespread distribution of autofluorescent material. There are few regions of the monkey CNS in which lipofuscin-like autofluorescence does not exist. Higher-magnification images (insets) were taken from the regions indicated by the boxes. The image from the medulla was taken from the reticular formation; that from the cerebellum was taken from the vermis; that from thalamus was taken from the region of the dorsomedial thalamic nucleus; that from the hypothalamus was taken from the lateral hypothalamic nucleus; and that from the cortex was from motor cortex. DRG, lumbar dorsal root ganglion. Bars = 50 μm; insets = 10 μm.
A waste portion of one human medulla was obtained under a protocol approved by the University of Minnesota Institutional Review Board. Sections were cut at a nominal thickness of 20 μm on a freezing microtome.
Immunocytochemistry
Sections were immunofluorescently labeled using either goat anti-calcitonin gene-related peptide diluted 1:200 (CGRP; a gift from Hunter Heath, Mayo Clinic, Rochester, MN), goat anti-serotonin diluted 1:100 (5-HT, 5-hydroxytryptamine, (Wessendorf and Elde 1985), or rabbit anti-μ opioid receptor diluted 1:1000 (MOR1; Arvidsson et al. 1995). Cryostat sections were incubated with primary antibodies overnight at 4C and washed in PBS (three times for 15 min) at room temperature (RT). The sections were then incubated with secondary antibodies for 2 hr at RT and washed in PBS (three times for 15 min). Fluorescent secondary antibodies (diluted 1:100) conjugated to Cy2 or cyanine 3.18 (Cy3) and were obtained from Jackson Immunoresearch Laboratories (West Grove, PA). Frozen microtome sections were incubated in primary antibodies at 4C for 48 hr and washed in PBS (three changes of PBS over 2 hr) at RT. The sections were then incubated with secondary antibodies at 4C for 48 hr and washed in PBS (three changes of PBS over 2 hr) at RT.
Attempted Reduction of Lipofuscin Autofluorescence
After review of the literature regarding the nature and composition of lipofuscin, several protocols were used in an attempt to reduce or eliminate the autofluorescence of lipofuscin. These included treating tissue sections with potassium permanganate alone (Barden 1978), potassium permanganate followed by oxalic acid (Barden 1978), sodium borohydride (Lillie and Pizzolato 1972), ferrous or ferric chloride (Kikugawa et al. 1995,1997), Nile Blue (Pearse 1972, 1985; Larsson 1993), hydrogen peroxide (Dowson 1983), potassium iodide (Lakowicz 1983), Sudan black (Meister et al. 1991; Romijn et al. 1999), cupric sulfate (Kikugawa et al. 1995,1997), sodium sulfate, cupric chloride or extraction by chloroform methanol (Csallany and Ayaz 1976). Of these chemical tissue treatments, only Sudan Black B, cupric sulfate, and cupric chloride treatments had appreciable effects on lipofuscin-like autofluorescence.
Histochemical Protocols to Reduce Lipofuscin Autofluorescence
Cupric Sulfate. After immunocytochemistry, the sections were removed from the PBS wash, dipped briefly in distilled H2O, and treated with CuSO4 (Fisher Scientific; Pittsburgh, PA) in ammonium acetate buffer (50 mM CH3COONH4, pH 5.0) for 10-90 min, dipped briefly in distilled H2O, and returned to PBS. The sections were either mounted with PBS/ glycerol containing 0.1% p-phenylenediamine (PPD; Johnson et al. 1982) or dehydrated through graded ethanols (50-100%), cleared in xylene, and mounted in DPX (Fluka Chemical; Ronkonkoma, NY). As noted by Kikugawa and co-workers (1997), cupric sulfate treatment was most effective under acidic conditions (pH ≤ 5.0). Although that group included 100 μM ethylendiamine tetra-acetic acid in the ammonium acetate buffer, our preliminary experiments determined that this compound was unnecessary for successful reduction of lipofuscin autofluorescence.
Lipofuscin has broad spectral properties. The image shows lipofuscin-like autofluorescence in cells of monkey inferior olivary nucleus. Autofluorescent material is visible (arrows) using UV filters, FITC filters, rhodamine filters, and Cy5 filters. Therefore, the presence of lipofuscin-like autofluorescence can mimic the appearance of virtually any fluorescent label. Bar = 50 μm.
The presence of lipofuscin can complicate the use of fluorescence immunohistochemical techniques. Images obtained from monkey spinal cord sections that were double labeled for 5-HT (Cy2; green) and the cloned μ-opioid receptor MOR1 (Cy3; red). False-positive double labeling for 5-HT and MOR1 in dorsal horn (
Sudan Black. After immunocytochemistry, the sections were treated with a solution of Sudan Black B (Allied Chemical; New York, NY) in 70% methanol for 5 min. If overstained, they were then differentiated by dipping in clean 70% ethanol until a desired level of staining was achieved and then were mounted with PBS/glycerol/PPD. Sudan black histochemical staining was not compatible with xylene-based mounting media (e.g., DPX) because SB is lipophilic and is removed from tissue on immersion in xylene.
Microscope and Filter Systems
The sections were examined with an Olympus BH-2 microscope equipped for reflected fluorescence illumination and digital imaging. Filter bandpasses were as follows: FG or True Blue-Fast Blue (wideband UV) 330-390-nm bandpass excitation filter and a 420-nm longpass emission filter; FITC/Cy2 (green fluorophores) 460-490-nm excitation and 510-550-nm emission; rhodamine/Cy3 (red fluorophores) 541-551-nm excitation and 572-607-nm emission; and Cy5 (deep red fluorophore) 615-635-nm excitation and 655-nm longpass emission. Black-and-white digital images were collected with a Cohu 4915 CCD camera (Cohu; San Diego, CA) and Power Macintosh 7100 computer (Apple Computer; Cupertino, CA) equipped with an image acquisition board (model LG-3; Scion Image, Frederick, MD) and using Scion Image version of the public domain NIH Image program (developed at the National Institutes of Health and available on the Internet at http://rsb.info.nih.gov/nih/image/). Color digital images were obtained using a cooled color CCD camera (Optronix; Schaumberg, IL) and acquired with MetaMorph software (Universal Imaging; West Chester, PA). Digital images were manipulated with Photoshop 5.0 software (Adobe Systems, San Jose, CA) using a Power Macintosh G3 computer and were printed with a Pictrography 3000 color printer (Fujix; Tokyo, Japan).
Results
General Observations
Lipofuscin-like autofluorescence was found throughout the neuraxis of macaques, including cortex, hippocampus, cerebellum, thalamus, hypothalamus, medulla, spinal cord, and dorsal root ganglia (Figure 1). There appeared to be lipofuscin-like pigments in cells of all sizes in most of these regions, although not all cells emitted autofluorescence. For example, in the cerebellum the majority of lipofuscin-like autofluorescence was observed surrounding Purkinje cells, whereas less was observed within the Purkinje cells themselves. Furthermore, little or no autofluorescence was observed in cells comprising the molecular and granular layers of the cerebellum.
The excitation and emission characteristics of the lipofuscin-like autofluorescence were sufficiently broad that they could complicate the use of fluorescence techniques (Figure 2). Lipofuscin-like autofluorescent pigments were visible under fluorescent filters for UV, fluorescein, rhodamine, and Cy5. Using the UV filter, the lipofuscin-like fluorophore was punctate and gold-yellow, which is similar in appearance to FG. Using other filters, it appeared either green (using the filters for fluorescein), red (using the filters for rhodamine), or deep red (using the filters for Cy5); the intensity of the lipofuscin-like compound could be as strong as that of the strongest immunofluorescence. Thus, lipofuscin-like autofluorescence could mimic the appearance of immunofluorescent labeling (Figure 3). In addition to the histochemical means to reduce lipofuscin autofluorescence mentioned above (see Materials and Methods), we attempted to reduce it using high-intensity, long-duration UV illumination (X60, 1.4 NA objective, 1 hr duration. Mounting medium was omitted from under the coverslip to allow free access of molecular oxygen). The lipofuscin-like fluorophore did not photobleach to an appreciable extent (Figure 4).
Lipofuscin cannot be photobleached. (Left) Lipofuscin-containing cell in monkey spinal cord. (Right) Same cell after 1 hr of UV illumination using a × 60, 1.4 NA oil-immersion objective. The autofluorescent material did not appreciably photo-bleach. Bar = 50 μm.
Concentration-dependent reduction of lipofuscin by cupric sulfate. Images obtained from monkey spinal cord sections that were incubated in either PBS, 50 mM ammonium acetate buffer alone (pH 5), or ammonium acetate buffer containing various concentrations of cupric sulfate. PBS and ammonium acetate buffer alone had no effect on lipofuscin-like autofluorescence. Low concentrations of cupric sulfate (≤ 100 μM) were insufficient to eliminate the autofluorescent pigments. However, higher concentrations (≥1 mM) of cupric sulfate gave significant reduction in lipofuscin-like autofluorescence. Bar = 100 μm.
The effect of Cu2+ or SO4 − on lipofuscin-like autofluorescence. Images obtained from sections of monkey medulla that were either untreated, incubated with 10 mM CuSO4, incubated with 10 mM CuCl2, or incubated with 10 mM Na2SO4. Sections incubated in CuCl2 showed reduction of autofluorescence similar to that observed using CuSO4. Na2SO4 had no noticeable effect on lipofuscin-like autofluorescence. Therefore, it appears that copper is the component responsible for reduction of lipofuscin-like autofluorescence. Bar = 50 μm.
Concentration-dependent reduction of lipofuscin by Sudan Black B. Images obtained from monkey spinal cord sections that were incubated in increasing amounts of Sudan Black B dissolved in 70% ethanol. Lower concentrations (≤ 0.01%) of SB provided insufficient reduction of autofluorescent pigments. However, higher concentrations (≥ 0.1%) of SB gave adequate reduction in lipofuscin-like autofluorescence. Bar = 100 μ
Effect of CuSO4 or Sudan Black on lipofuscin-like autofluorescence in human and aged rat CNS. Images obtained from 20-μm sections of inferior olivary nucleus from a human (top) and a 7-month-old rat (bottom). In the elderly human, CuSO4 reduced the intensity of the lipofuscin-like autofluorescence, whereas SB completely eliminated it. (It should be noted that lipofuscin-like autofluorescence in other regions in the human medulla was adequately reduced by CuSO4.) In the aged rat, both treatments were equally effective for eliminating lipofuscin-like autofluorescence. Bar = 50 μm.
Effect of CuSO4 or Sudan Black Treatment on Lipofuscin-like Autofluorescence
Depending on the concentration used, CuSO4 treatment greatly reduced or eliminated the lipofuscin-like autofluorescence in monkey spinal cord (Figure 5). At low concentrations (1-100 μM CuSO4), grains of yellow autofluorescent pigment could still be observed in neuronal cytoplasm. At higher concentrations (1-10 mM CuSO4), almost all of the autofluorescent material was eliminated from neuronal somata (Figure 5), although some yellow fluorescent material remained surrounding structures that resembled blood vessels (not shown). At the highest concentrations of CuSO4 (≥ 100 mM), all lipofuscin-like pigments were eliminated.
CuSO4 is composed of two ions, either or both of which might be responsible for its action. To determine which ion (Cu2+ or SO4 2-) was the active component, we treated tissue with either cupric chloride or sodium sulfate in ammonium acetate buffer as above. Incubation of monkey spinal cord sections in 10 mM CuCl2 eliminated the lipofuscin-like autofluorescence (Figure 6). However, incubation of tissue sections in 10 mM Na2SO4 had no effect on lipofuscin-like autofluorescence (Figure 6).
SB treatment of monkey CNS tissue also appeared to reduce or eliminate lipofuscin-like autofluorescence in a concentration-dependent manner (Figure 7). We found that concentrations of less than 1% SB were not sufficient to eliminate autofluorescent pigments. At the highest concentrations, 1-10% SB, all autofluorescent pigments were eliminated. To demonstrate the concentration-dependent effect of SB, the sections from the dilution series shown in Figure 7 were not destained (see Materials and Methods).
On the basis of the above experiments with monkeys, we tested the efficacy of the CuSO4 and SB treatments using CNS tissue collected from one elderly human and one rat (aged 7 months). In human medulla, CuSO4 treatment provided substantial reduction of lipofuscin-like autofluorescence. However, autofluorescence in many large cells of the inferior olivary nucleus remained. SB treatment was able to eliminate lipofuscin-like fluorescence in human tissue (Figure 8). In the rat, both treatments successfully eliminated the autofluorescent pigments (Figure 8).
Effect of CuSO4 or Sudan Black Treatment on Immunofluorescent Labeling
In preliminary experiments, we found that pretreatment of tissue by CuSO4 or SB before (rather than after) immunocytochemistry unacceptably reduced the intensity of immunocytochemical labeling of monkey tissue. Therefore, tissue treatment with SB or CuSO4 was performed after the final PBS wash (i.e., after the final rinsing of secondary antibodies). We found that CuSO4 reduced the intensity of immunofluorescence in a concentration-dependent manner. However, the effect of CuSO4 on lipofuscin was much more potent than its effect on immunofluorescence (Figure 9). At low concentrations (50 μ?-5 mM), the deleterious effect on immunofluorescence was much less pronounced; the intensity of immunocytochemical labeling was reduced but was readily visible. However, at the highest concentrations, (≥ 50 mM), it was difficult to visualize any of the immunocytochemical labeling.
Sudan Black treatment also reduced immunofluorescence in a concentration-dependent manner (Figure 10). Concentrations of ≤ 1% SB allowed visualization of all the fluorophores tested, whereas only Cy3 could be visualized at the highest concentration (10% SB).
Effect of CuSO4 or Sudan Black Treatment on Retrograde Tract Tracing
We also attempted to determine whether it was practical to reduce lipofuscin-like fluorophores using CuSO4 and SB in experiments in which fluorescent retrograde tract tracing dyes were used. In the rat, both CuSO4 and SB induced a slight reduction in the intensity of FG retrograde labeling in the hypoglossal nucleus (Figure 11). In monkeys, retrograde labeling of dorsal root ganglion cells by a mixture of True Blue and Fast Blue was slightly reduced using CuSO4 but completely eliminated using SB (Figure 11). Retrograde labeling of monkey bulbospinal neurons by WGA-HRP (observed using immunofluorescent labeling for WGA) was affected to the same degree as for immunocytochemical labeling (not shown, but see above).
Discussion
The interpretation of fluorescence microscopy of tissue from older animals in general, and from monkeys and humans in particular, has been made difficult by the presence of the autofluorescent pigment lipofuscin, which accumulates in the cytoplasm of cells as animals age. We and others (Correa et al. 1980; Partanen et al. 1980; Santer et al. 1980; Helen 1983; Kalyuzhny and Wessendorf 1998) have reported that the presence of lipofuscin can complicate the use of fluorescent immunocytochemical techniques because it fluoresces intensely using a variety of microscope filter systems. Therefore, it is difficult to distinguish specific labeling from that of lipofuscin.
Effect of CuSO4 concentration on immunofluorescent labeling. Images obtained from monkey spinal cord dorsal horn sections that were immunofluorescently labeled for CGRP (Cy2) and treated with various concentrations of CuSO4. Images were all obtained and processed identically. At lower concentrations of cupric sulfate (≤ 5 mM), a modest reduction in immunofluorescent labeling intensity was observed. This reduction could be adequately compensated for by increased exposure time (not shown). At higher concentrations (≥ 50 mM), immunofluorescent labeling intensity was markedly reduced. Bar = 50 μm.
Effect of Sudan Black B concentration on immunofluorescent labeling. Images obtained from monkey spinal cord dorsal horn sections that were immunofluorescently labeled for MOR1 (Cy3) and treated with various concentrations of SB. Images were all obtained and processed identically. At lower concentrations of SB (≤ 0.01%), a modest reduction in immunofluorescent labeling intensity was observed. As was the case with CuSO4 treatment, this reduction could be adequately compensated for by increased exposure time (not shown). At higher concentrations (≥ 1%), immunofluorescent labeling intensity was markedly reduced. Bar = 50 μm.
Because of the difficulties presented by lipofuscin-like autofluorescence, we were interested in finding a protocol that eliminated or reduced it and that was compatible with our fluorescent immunocytochemical and retrograde tract tracing procedures. Previously, it was suggested that lipofuscin-like autofluorescence could be reduced by the use of younger animals (Correa et al. 1980). We did not find this solution to be satisfactory because monkeys as young as 4 months of age possess considerable amounts of lipofuscin-like pigment (see, e.g., Brizzee et al. 1974). It has also been suggested that the use of custom-made barrier filters might decrease lipofuscin fluorescence (Partanen et al. 1980); this appears unlikely to be the case because of its broad excitation and emission characteristics. However, because lipofuscin is visible under all common fluorescence filters, it may be possible to determine whether fluorescence is not due to lipofuscin. For example, if a cell fluoresces using the fluorescein filters but not using the rhodamine filters, one can reasonably conclude that the fluorescence is not due to lipofuscin (Kalyuzhny et al. 1996). However, this approach is of no help when the cell of interest contains lipofuscin and is of only limited use for multicolor fluorescence microscopy.
Previous studies have reported that lipofuscin could be extracted with a mixture of chloroform and methanol (Nandy 1971; Taubold 1975; Csallany and Ayaz 1976; Barden 1980; Katz et al. 1996; Yin 1996). The results of these studies indicated that a blue fluorescent compound, possibly consisting of lipids, was the source of lipofuscin autofluorescence in tissues. However, it has also been shown that blue fluorescence can be induced artifactually through manipulations of the organic extract (e.g., light irradiation and chromatographic processes) (Eldred et al. 1982; Eldred and Katz 1989; Kikugawa et al. 1994). Recently, a series of reports has demonstrated that a lipofuscin-like compound is extractable by aqueous but not by organic solvents (Kikugawa et al. 1994, 1995, 1997). Furthermore, this aqueous phase retains several of the same properties of lipofuscin-like autofluorescence found in situ (e.g., broad excitation and emission spectra and yellow fluorescence under UV illumination). Therefore, it appears that these recent studies may have identified the bona fide component responsible for lipofuscin-like autofluorescence.
In the latter studies, it was found that treatment of the aqueous extract with certain metallic salts decreased or eliminated the yellow fluorescence, depending on the pH of the solution used (Kikugawa et al. 1995,1997). On the basis of these results, we similarly treated sections of monkey spinal cord to determine their ability to reduce lipofuscin-like autofluorescence. In these experiments, we found that treatment of tissue sections with cupric sulfate reduced or eliminated lipofuscin-like autofluorescence without excessively reducing immunocytochemical labeling. We are uncertain about the chemical mechanism by which cupric sulfate (or cupric chloride) acts to reduce lipofuscin fluorescence. Fluorescence requires a photon of light to be absorbed by a molecule, causing an electron to be raised from its ground state to its excited state. Fluorescence then occurs when that electron returns to its ground state, releasing energy in the form of another photon. Quenching of fluorescence (i.e., reduction of its intensity) occurs when something interferes with this process. It would be reasonable to hypothesize that either collisional quenching (in which case electrons are transferred from the lipofuscin to the Cu2+ ion), static quenching (in which a nonfluorescent complex is formed between Cu2+ and lipofuscin), or a combination of the two types of quenching occurs (Lakowicz 1983). It has been reported that Cu2 + is an excellent “electron scavenger,” suggesting that electrons could be transferred to it by collisional contact between it and a fluorescent molecule (Steiner and Kirby 1969). Transfer of electrons from the excited state of lipofuscin to Cu2+ would circumvent the emission of its fluorescence. Regardless of the mechanism, we are convinced that Cu2+ is important for reduction of lipofuscin-like fluorescence because (a) another copper compound (CuCl2) was shown to decrease its fluorescence and (b) SO4 −, in the form of Na2SO4, had no effect. Ferric or ferrous salts appeared to have little or no effect on lipofuscin-like fluorescence, in contrast to previous reports (Kikugawa et al. 1997). Other divalent cations (specifically Mn2+ and Mg2+) similarly had no effect (not shown).
Effect of CuSO4 or Sudan Black on fluorescent retrograde tract tracer labeling. (Top) Images of retrograde labeling of hypoglossal motoneurons obtained by injection of Fluoro-Gold into the tongue of a rat. At concentrations sufficient to eliminate lipofuscin-like pigments from monkey tissue sections, a slight reduction in FG intensity was observed (middle and right panels). (Bottom) Retrograde labeling of dorsal root ganglion neurons obtained by injection of True Blue/Fast Blue solution into the dorsal spinal cord of monkeys. Cupric sulfate reduced the intensity of True Blue/Fast Blue fluorescence but retrograde labeling was still readily visible (middle panel). However, SB completely eliminated True Blue-Fast Blue fluorescence (right panel). Bars = 50 μm.
In agreement with recent studies (Meister et al. 1991; Romijn et al. 1999), tinctorial staining of lipofuscin (for protocols see Pearse 1985) also appears to reduce lipofuscin-like autofluorescence. Of the dyes tested, Sudan Black was the most effective in reducing lipofuscin-like autofluorescence. Given the almost opaque character of its labeling, we think that SB may act by obscuring lipofuscin without interacting with it at the physical chemical level.
The little that is known regarding the chemical nature of lipofuscin has been reviewed by Pearse (1985). Our experiments may provide a few additional insights into the matter. First, several chemical “bleaching” protocols (e.g., potassium permanganate, sodium borohydride, hydrogen peroxide) to which conjugated double bonds would be susceptible but aromatic double bonds would not, had no effect on its fluorescence. It therefore appears that any carbon-carbon double bonds necessary for fluorescence in lipofuscin must be aromatic in character. Second, lipofuscin is unlikely to undergo free radical formation, because illumination of lipofuscin for extended periods did not cause appreciable photobleaching as would be expected if free radical reactions were altering the compound. Third, its very broad excitation and emission characteristics suggest that lipofuscin might be composed of a mixture of different, although probably related, fluorescent molecules. Finally, our data are consistent with the existence of different types of lipofuscin because, in aged human, CuSO4 only reduced the intensity of lipofuscin autofluorescence in the inferior olivary nucleus rather than eliminating it. Furthermore, lipofuscin-like labeling in vascular endothelium was less affected by CuSO4 than was neuronal lipofuscin. The existence of different types of lipofuscin-like substances has been discussed by Pearse (1985).
In summary, the experiments described here demonstrated that two methods that reduced the intensity of lipofuscin-like fluorophores in monkeys were compatible with common fluorescent immunocytochemical and retrograde tract tracing techniques. Substantial reductions of lipofuscin-like autofluorescence in monkey CNS tissue sections were obtained with CuSO4 or Sudan Black. Furthermore, these substances were shown to be compatible with a wide range of fluorophores commonly used for immunocytochemical labeling. Although CuSO4 or SB also reduced the intensity of immunocytochemical labeling, this reduction did not appear to decrease our ability to visualize specific labeling. In fact, the reduction of immunocytochemical labeling could be largely overcome by using longer exposures while imaging. The two techniques had differential effects on the fluorescence of retrograde tract tracers. In the rat, CuSO4 or Sudan Black slightly attenuated the intensity of FG labeling. However in monkeys, a mixture of True Blue and Fast Blue used as a retrograde tract tracer showed that CuSO4 only reduced the intensity of the specific labeling but SB completely eliminated it.
We conclude that treatment of monkey tissue with CuSO4 or SB strikes an acceptable balance between reduction of lipofuscin-like autofluorescence and retention of immunocytochemical labeling. Because SB is incompatible with some fluorescent retrograde tract tracers and with xylene-based permanent mounting media, we also conclude that the cupric sulfate protocol may be of more general utility for the reduction of lipofuscin-like autofluorescence. However, SB may be superior for eliminating lipofuscin-like autofluorescence in aged human CNS tissue.
Footnotes
Acknowledgments
Supported by PHS grant DA09642 from NIDA.