Novel Application of Tyramide Signal Amplification (TSA): Ultrastructural Visualization of Double-labeled Immunofluorescent Axonal Profiles

Volume 48(1): 153–161, 2000 The Journal of Histochemistry & Cytochemistry

http://www.jhc.org

TECHNICAL NOTE

Novel Application of Tyramide Signal Amplification (TSA): Ultrastructural Visualization of Double-labeled Immunofluorescent Axonal Profiles András Büki, Susan A. Walker, James R. Stone, and John T. Povlishock Department of Anatomy, Medical College of Virginia, Virginia Commonwealth University, Richmond, Virginia (AB,SAW,JRS,JTP) and Department of Neurosurgery, University Medical School of Pécs, Pécs, Hungary (AB)

Fluorescent immunocytochemistry (FICC) allows multiple labeling approaches when enzyme-based techniques are difficult to combine, such as in double-labeling experiments targeting small-caliber axonal segments. Nevertheless, the conversion of FICC to a product visible at the electron microscopic (EM) level requires labor-intensive procedures, thus justifying the development of more user-friendly conversion methods. This study was initiated to simplify the conversion of FICC to EM by employing the unique properties of tyramide signal amplification (TSA), which allowed the simultaneous targeting of a fluorescent tag and biotin label to the same antigenic site. Briefly, one of two antigenic sites typically co-localized in damaged axonal segments was visualized by the application of a fluorescent secondary antibody, with the other tagged via a biotinylated antibody. Next, an ABC kit was used, followed by the simultaneous application of fluorophore–tyramide and biotin–tyramide. After temporary mounting for fluorescent digital photomicroscopy, sections were incubated in ABC and reacted with diaminobenzidine before EM analysis. Double-labeling fluorescent immunocytochemistry with TSA clearly delineated damaged axonal segments. In addition, these same axonal segments yielded high-quality EM images with discrete electron-dense reaction products, thereby providing a simple and reproducible means for following fluorescent analysis with EM. (J Histochem Cytochem 48:153–161, 2000)

SUMMARY

The combination of fluorescent tags with distinct emission spectra and the use of enzyme-based staining procedures constitute the most widely used tools in double/multiple-labeling immunocytochemistry (ICC). Although enzyme-based multiple-labeling techniques are excellent for the analysis of axodendritic and axosomatic reactions or for immunocytochemical co-localization of nuclear and cytoplasmic antigens, the spatial resolution of these techniques is generally insufficient to detect antigens co-localized in the same axonal segment. Fluorescent markers are more appropriate for such studies. However, their use necessitates further conversion methods when subsequent electron microscopic (EM) investigations are needed to confirm their Correspondence to: John T. Povlishock, PhD, Dept. of Anatomy, Medical College of Virginia, Virginia Commonwealth U., Richmond, VA 23298-0709. Received for publication June 22, 1999; accepted August 31, 1999 (9T5014). © The Histochemical Society, Inc.

0022-1554/00/$3.30

KEY WORDS tyramide signal amplification axonal injury immunofluorescence double labeling trauma calpain

precise sites of localization (Sandell and Masland 1988; Lubke 1993; Todd 1997; Humbel et al. 1998). Recently, several conversion methods have been described, including photoconversion of fluorescent tracers with DAB (Sandell and Masland 1988; Lubke 1993), the use of anti-fluorescein antibodies combined with streptavidin–gold (Shiosaka et al. 1986) or immunogold (Luby–Phelps et al. 1984), and the use of fluorescent tags directly (Powell et al. 1997) or indirectly (Powell et al. 1998) coupled with gold particles. Although the use of laser scanning confocal microscopy (LSCM) has offered new possibilities for the conversion of fluorescent multiple labeling ICC to EM (Sun et al. 1995; Todd 1997), the technique is rather complex and the more restricted availability of the instrument limits its employment. Therefore, the vast majority of present techniques that allow combined fluorescent/ultrastructural visualization require timeconsuming and labor-intensive procedures and/or the 153

154 use of special technical tools that constitute obstacles to their routine laboratory use. Recently, in our laboratory we have successfully utilized a combination of the peroxidase-based chromogens benzidine dihydrochloride (BDHC) and diaminobenzidine (DAB) or Vector VIP for both LM and EM co-localization of reaction products linked to calciuminduced, calpain-mediated spectrin proteolysis (CMSP) and cytoskeletal change in traumatically injured axonal segments (Büki et al. 1999a). However, despite our success in these approaches, we wished to utilize double-label fluorescent approaches that would provide better chromogen penetration and higher spatial resolution and reliability while offering a more simplified method of processing and evaluation than the use of enzyme-based chromogenic immunohistochemical techniques. Having demonstrated that traumatic axonal injury leads to activation of the neutral protease calpain, while also establishing the spatiotemporal correlation between local intra-axonal CMSP and neurofilament compaction (NFC) as markers of traumatic axonal injury (Povlishock et al. 1997; Büki et al. 1999a,b; Okonkwo et al. 1999), we sought to utilize these specific changes as targets for new LM double labeling, a fluorescent approach that is also compatible with EM analysis. Our goal was to improve the sensitivity and utility of our experimental methods, thereby saving time and, more importantly, reducing the number of animals required for unraveling the pathobiology and treatment of traumatic axonal injury. Recognizing that the TSA technique allows the simultaneous application of a fluorescent tag [fluorophore (rhodamine)–tyramide] and biotin (biotin–tyramide) to the same antigenic site, we describe here a substantial modification of the conventional TSA technique that yields a reliable and simple method that can be employed for fluorescent double/multiple-labeling experiments combined with subsequent pre-embedding immunoelectron microscopic analysis of the labeled structures. We also comment on the utility of this method in the study of axonal profiles, in which it combines the advantages of sensitive signal amplification with a user-friendly conversion method for immunoelectron microscopy.

Materials and Methods Induction of Traumatic Brain Injury For induction of focal axonal injury, a well-characterized rodent model of impact acceleration head injury was employed as described by Marmarou et al. (1994). Eight Sprague– Dawley rats (Charles River Laboratories; Raleigh, NC) weighing 368–396 g were used for the experiments. For induction of anesthesia, the animals were exposed to 4% isoflurane (Iso Flo; Abbott Laboratories, North Chicago, IL) in a bell jar for 5 min and then intubated and ventilated with a

Büki, Walker, Stone, Povlishock mixture of 1–2% isoflurane in 30% O2 and 70% N2O. Next, the skull between the coronal and lambdoid sutures was exposed with a midline incision. A metallic disk-shaped helmet 10 mm in diameter was firmly attached to this point of the skull using dental acrylic. The animal was placed in a prone position on a foam bed with the metallic helmet centered under the edge of a plexiglas tube and was prevented from falling by two belts secured to the foam pad. Brass weights weighing 450 g were allowed to fall from a height of 200 cm through the plexiglas tube directly to the metallic disk fixed to the animal’s skull, a setting that produces neither cerebral contusion nor subdural hemorrhage. After the injury the animal was immediately ventilated with 100% O2. The helmet was removed and the skull was investigated for any sign of fracture which, if found, would have disqualified the animal from further evaluation. The scalp wound was sutured, with the animal remaining on artificial ventilation until it regained spontaneous breathing, followed by sacrifice at the predetermined survival periods of 30 (three rats) and 180 min (three rats) after injury. Two sham-injured animals (one for each survival period) were treated in the same way as the other three animals per group but were not injured.

Physiological Assessments The respiratory status was monitored through pulse oximetry via the footpad and/or the tongue. In addition, brain temperature was monitored by a temporalis muscle probe and core temperature was determined by a rectal probe.

Immunohistochemistry At the designated survival time, the rats were reanesthetized with an overdose of sodium pentobarbital and transcardially perfused with 4% paraformaldehyde and 0.1% glutaraldehyde in Millonig’s buffer. The brains were immersed in the same fixative overnight (16–18 hr). On the basis of previously published observations concerning the topography of diffusely injured axons in rat brain (Povlishock et al. 1997), a midline 5-mm-wide block of the brain containing the medullary pyramids was removed using a sagittal brain blocking device to include the region extending from the interpeduncular fossa to the first cervical segment. The tissue was sectioned with a Vibratome Series 1000 (Polysciences; Warrington, PA) at a thickness of 30 m and was collected in 0.1 M phosphate buffer. Then the sections were rinsed three times for 10 min in PBS, placed in an Na-citrate buffer (pH 6.0), and transferred to a programmable magnetron-powered 900-W PELCO 3460 microwave oven (Ted Pella; Redding, CA). This laboratory microwave oven is equipped with a load cooler and a computerized temperature monitoring system, and was operated with 70% energy cycling over two 5-min periods during which tissue temperature was never allowed to rise above 40C. This approach was based on our recent finding (Stone et al. 1999) that the use of microwave energy without the generation of considerable heat significantly enhances immunoreactivity while also suppressing background staining. This controlled temperature approach also allows the preservation of excellent ultrastructural immunohistochemical detail, unlike traditional approaches that may cause tissue damage via the generation of significant

155

Conversion of Immunofluorescence to EM by TSA heat (Stone et al. 1999). After microwave processing, sections were incubated for 35 min with 0.2% Triton X-100 (Sigma Chemical; St Louis, MO) in a 10% mixture of normal goat serum (NGS) and normal donkey serum (NDS) (Sigma) in PBS. After two quick rinses in PBS containing 1% NGS/NDS, sections were incubated with a mixture of the primary antibodies, Ab38 and RMO-14, overnight (12–16 hr). The rabbit polyclonal antibody Ab38 (applied here at a dilution of 1:18,000) targets the NH2-terminal fragment of the -subunit of brain spectrin on its cleavage by the calcium-activated protease calpain, allowing the identification of the precise locus of any calcium-mediated events linked to the proteolysis of spectrin at both the LM and EM level (Siman et al. 1989; Roberts–Lewis et al. 1994). The mouse monoclonal antibody RMO-14 (1:500) has been described to exclusively target rod domains of compressed neurofilament M-subunits (NF-M) visualized only on alteration of their side arms, with subsequent exposure of their previously masked epitopes (Lee et al. 1987; Pettus et al. 1994; Povlishock et al. 1997; Okonkwo et al. 1998). Because such NFC and side-arm modification routinely occurs in foci of traumatically induced axonal damage (Maxwell et al. 1997; Povlishock et al. 1997; Okonkwo et al. 1998), this antibody delineates the precise sites of axonal injury in which the activated CMSP is also found (Büki et al. 1999a). After rinses in 1% NDS/NGS, sections were incubated in a solution of coumarin-labeled donkey anti-mouse immunoglobulin (1:200; Jackson Immunoresearch Laboratories, West Grove, PA) and biotinylated goat anti-rabbit immunoglobulin (1:400; Vector, Burlingame, CA) for 60 min, followed by three 10-min rinses in PBS. In all subsequent steps of tissue processing, direct illumination of the sections was avoided.

Tyramide Signal Amplification Based on attempts to simplify the TSA method and make it more cost-effective, while also providing better preservation of ultrastructural detail several modifications of the original protocol (for details see the commercial protocol, Renaissance Kit, NEN Life Sciences Products, Boston, MA) were introduced in our laboratory. Specifically, after incubation in an avidin–biotin–peroxidase complex (ABC standard Elite kit; Vector; dilution 1:200) with rinsing in TNB blocking buffer (Renaissance Kit, NEN Life Science) for 20 min and in PBS twice for 10 min, a 1:1 mixture of rhodamine–tyramide and biotin–tyramide (Renaissance Kit, NEN Life Science), both diluted 1:300, was applied. A 1:6 mixture of amplification diluent (Renaissance Kit, NEN Life Science) and PBS constituted the solvent for the tyramide stock solution, together forming the working solution. The working solution was applied in mid-sized tissue culture wells containing six tissue sections each (12-well/tray; Falcon Multiwell Tissue Culture Plate, Becton Dickinson Labware; Becton Dickinson, Lincoln Park, NJ), 400 l working solution/well, for 12 min, followed by three 10-min rinses in PBS. The dilution of the amplification diluent in PBS to 1:6, as well as the substitution of the detergent-containing rinsing buffer (Tween is suggested in the original protocol) with PBS was aimed to prevent unnecessary membrane damage and reduce the background reactivity. Next, the sections were mounted on premium microscope slides (Fisher Scientific; Pittsburgh, PA) and coverslipped using 50% glycerol dissolved in double-

distilled water. The slides were immediately transferred to a Nikon Eclipse E 800 biological research microscope (Image Systems; Columbia, MD) equipped with a Sony Catseye digital camera (Image Systems). Images of immunofluorescent axonal profiles were digitally captured and archived. All the sections analyzed were mounted and coversliped separately, with each assigned a serial ID number. After completion of digital acquisition, coverslips were removed and the sections were re-rinsed and incubated in an avidin–biotin–peroxidase complex at a dilution of 1:200. The sections were then processed for visualization of the immunohistochemical complex using 0.05% diaminobenzidine (DAB) (Sigma) and 0.01% H2O2 in 0.1 M phosphate buffer. Although in theory one could argue that the previously applied peroxidase molecule found in the ABC complex, used to activate tyramide, could also serve as the catalyst for the chromogenic reaction with DAB, this approach consistently yielded such a weak signal as to preclude its usefulness for routine LM/EM studies.

Ultrastructural Analysis After the above prepared sections were dehydrated and flatembedded between plastic slides in Medcast resin (Ted Pella), areas for EM investigation were selected by repeated, detailed analysis involving comparisons with the previously captured digital images. To draw a precise correlation between these samples assessed either with fluorescence or routine light microscopy, we employed anatomic landmarks such as the vicinity of blood vessels, tissue edges, and the relative distance between labeled axonal segments. Once identified, homologous immunopositive foci prepared for ultrastructural analysis were trimmed, mounted on plastic studs, and sectioned using an LKB Ultratome at a thickness of 70– 100 nm. A few semithin sections were mounted and coverslipped for comparison to the digital images while also serving as topographic guides for the ultrastructural analysis. Thin sections were picked up on Formvar-coated slotted grids, then stained in 5% uranyl acetate in 50% methanol for 2 min and 0.5% lead citrate for 1 min. Alternatively, some sections were not stained to provide additional evidence for antibody specificity. Ultrastructural analysis was carried out using a JEOL-1200 electron microscope.

Immunohistochemical Controls The antibodies employed have been extensively characterized in the existing literature and widely used (Lee et al. 1987; Roberts–Lewis et al. 1994; Saatman et al. 1996; Posmantur et al. 1997; Povlishock et al. 1997; Büki et al., 1999a). Nevertheless, additional control studies were performed to achieve further rigor. To this end, all immunocytochemical reactions included sections incubated without the addition of the primary or secondary antibody or with the secondary alone with or without TSA in concert with multiple dilution strategies to obtain optimal staining.

Results Fluorescent and Routine LM Observations

In accord with the results of our previously published studies, traumatically injured axonal profiles displayed

156 focal axonal immunofluorescence representing neurofilament compaction (RMO-14) and calpain-mediated spectrin proteolysis (Ab38), with the highest proportion of these profiles found in the corticospinal tract and in the medial longitudinal fasciculi (Povlishock et al. 1997; Okonkwo et al. 1998,1999; Büki et al. 1999a,b). Thirty min after injury the immunoreactive axonal segments appeared sharply delineated as swollen and occasionally vacuolated profiles. At 180 min after injury the majority of the immunolabeled axonal foci appeared as large, swollen axonal bulbs indicating axonal disconnection. The majority of the visualized axonal profiles displayed both co-localized coumarin immunofluorescence, representing NF alteration, and rhodamine fluorescence, delineating calpain-mediated spectrin proteolysis at all time points examined. In addition, the localization of the individual fluorescent markers was easier to recognize and more precise than that achieved by the previously employed peroxidase enzyme-based intra-axonal chromogens. Specifically, by applying a 2.5- to fourfold lower concentration of the primary antibody (Ab38) than employed in traditional immunohistochemical investigations and a sixfold lower concentration of the tyramide solution than originally recommended for this technique, we recognized rhodamine–TSA to be precisely localized to the focus of the axonal damage, without background immunoreactivity in any other tissue elements and/or emission spectra. Furthermore, this protocol did not compromise the sharp detection of isolated axonal profiles single labeled with the RMO-14 antibody (coumarin fluorescence) but not labeled with rhodamine, a finding also consistent with our previous observations (Büki et al. 1999a) (Figures 1A–1C). The digital comparison of homologous fluorescent and brightfield images revealed the consistent localization of the reaction products and also confirmed the accuracy of the conversion method itself. Specifically, the DAB reaction product (Figure 1D) was precisely localized to the same sites containing rhodamine labeling. Furthermore, none of the single-labeled axonal profiles displaying NFC alone were DAB-positive, emphasizing the reliability of the method chosen. Ultrastructural Observations

In this approach, the overall ultrastructural appearance of the tissue described above was well-preserved despite these additional tissue-processing procedures. At the EM level, damaged immunoreactive axons displayed the well-established repertoire of traumatic axonal injury, including neurofilament compaction, mitochondrial swelling, loosening of the myelin sheath, and formation of periaxolemmal spaces (Christman et al. 1994; Maxwell et al. 1997; Povlishock et al. 1997) (Figures 2A, 2B, 3A, and 3B). CMSP immunoreactiv-

Büki, Walker, Stone, Povlishock ity was observed as an electron-dense DAB reaction product in axonal foci where neurofilaments were clearly compacted. At 30 min after injury the electrondense DAB granules were predominantly localized in the subaxolemmal compartment (Figures 2A and 2B), whereas at 180 min after injury the altered cytoskeletal elements in the entire damaged axonal segment were covered with reaction product (Figures 3A and 3B). The TSA method provided excellent sensitivity and specificity at the ultrastructural level, with the immunoreactivity exclusively localized in the injured axonal segments and with no spreading or diffusion into noninjured axons (Figures 2 and 3). Moveover, the ultrastructural appearance of the DAB reaction product was identical to or perhaps even more intense than that detected in our previous studies (Povlishock et al. 1997; Büki et al. 1999a), indicating that neither signal amplification or the slightly electron-dense tyramide itself complicated the detailed ultrastructural analysis of immunoreactive damaged axons. Immunohistochemical Controls

The exclusion of either the primary or the secondary antibody from the immunohistochemical reaction resulted in lack of immunoreactivity. Gradual overdilution of the immunosera caused disappearance of the immunostaining. Simultaneous or sequential application of immunosera in the case of double labeling F-IHC resulted in equally powerful detection of both antigens. The TSA kit alone did not contribute to any specific immunohistochemical reaction that might have influenced the analysis. Elimination of the rhodamine– tyramide step resulted in the lack of fluorescent labeling of CMSP-IR axonal profiles, while the absence of biotin–tyramide resulted in the lack of specific DAB staining (i.e., lack of conversion) at the LM level without affecting the fluorescence signal (rhodamine label) indicating CMSP-IR.

Discussion In a series of labor-intensive studies employing the combination of enzyme-based immunohistochemical reaction products (DAB/VVIP and BDHC/VVIP), we have recently demonstrated the co-localization of calpain-mediated spectrin proteolysis (CMSP) with markers of traumatically induced cytoskeletal alteration, providing evidence for the contribution of CMSP to the pathogenesis of traumatically induced axonal injury (TAI) (Büki et al. 1999a). In this study we utilized these observations as a template for establishing a new double-labeling light microscopic technique that was also compatible with EM investigation of the doublelabeled axonal profiles. In this search for a doublelabeling method to provide excellent spatial resolution at both light and electron microscopic levels, we con-

Conversion of Immunofluorescence to EM by TSA Figure 1 Fluorescent light micrographs demonstrating coumarin-labeled (A) axonal profiles in the medial longitudinal fasciculus which delineate neurofilament compaction (RMO-14 immunoreactivity). Also in this same field, rhodamine–tyramide-labeled axonal segments displaying calpain-mediated spectrin proteolysis (Ab38 immunoreactivity) (B) can be seen 30 min after injury. Digital overlays (C) of these fields (A,B) facilitates the selection of double-labeled axonal profiles for further analysis at the ultrastructural level. Note that only the rhodamine-labeled profiles display the DAB chromogen after signal conversion (D). The axonal segment marked with an arrow is shown in Figure 2 at the ultrastructural level. Bars  15 m.

157

158

Büki, Walker, Stone, Povlishock Figure 2 (A) Ultrastructural analysis of the axonal profile double labeled in Figure 1 reveals neurofilament compaction and DAB distribution primarily along the axolemma, delineating calpain-mediated spectrin proteolysis (CMSP) localized to the subaxolemmal network (arrows) 30 min after injury. White asterisk marks the concomitant neurofilament compaction; M indicates loosening of the myelin sheet. Note the lack of immunoreactivity in the neighboring noninjured axonal segments (black asterisk). High-power electron micrograph from the same axonal field clearly demonstrates the subaxolemmal distribution of CMSP immunoreactivity (arrows), the substantial alterations in the axonal cytoskeleton (white asterisk), the myelin sheath (M), and the mitochondrial swelling (double arrow) (B). Bars  750 m.

Conversion of Immunofluorescence to EM by TSA

159

Figure 3 Tyramide signal amplification of the immunohistochemical reaction reliably demonstrates calpainmediated spectrin proteolysis in axonal foci undergoing diffuse axonal injury 3 hr after injury (A,B). Note that at this time the immunoreactive material covers the cytoskeletal elements and/or fills the axoplasm (white asterisks) while remaining absent in the adjacent noninjured axonal profiles (black asterisk). This fact and the fine structure of the electron-dense reaction product are well illustrated in A, which is an unstained (nonosmicated) section. Neurofilament compaction (black asterisk) and the accumulation of organelles in the axoplasm (arrows) due to the halt of axoplasmic transport can also be detected (B). Bars  750 m.

sidered the application of immunofluorescent markers. However, we realized that despite their excellent spatial resolution, the conversion of immunofluorescent signals to EM was inherently difficult. This limi-

tation was overcome by our recognition of the advantages of the recently developed TSA technique, also described as catalyzed reporter deposition (Bobrow et al. 1989,1992; de Haas et al. 1996; Hunyady et al.

160 1996; van Gijlswijk et al. 1996,1997; Van Heusden et al. 1997). Essentially, in the TSA technique the biotinor fluorophore-conjugated tyramide is “attracted” to the antigenic site activated by the peroxidase component of the immunolabeling complex. The activated tyramide instantly and covalently binds to nearby tyrosine molecules, and this single enzymatic reaction greatly (1–400-fold) amplifies the signal intensity of the immunocytochemistry (Bobrow et al. 1989,1992; Hunyady et al. 1996; Erber et al. 1997; van Gijlswijk et al. 1997). Soon after the technique was first described, its widespread utilization (Van Heusden et al. 1997; Kaufmann et al. 1998; Kressel 1998; van de Corput et al. 1998; Speel et al. 1998,1999) and further modifications in its application highlighted its usefulness and importance, particularly in paradigms involving fluorescent double labeling with antibodies derived from the same species (Hunyady et al. 1996; Shindler and Roth 1996; Teramoto et al. 1998). Recently, this technique has also been used for pre- and postembedding immunoelectron microscopy, an extremely labor-intensive approach (Schofer et al. 1997; Stanarius et al. 1997; Punnonen et al. 1999), with some suggesting that the electron-dense tyramide can also be used for EM without concomitant DAB chromogen application (Mayer and Bendayan 1997). In the present study, the combination of the conventional immunofluorescence (coumarin-labeled marker for neurofilament compaction) and fluorophore-labeled tyramide (rhodamine marker for calpain-mediated spectrin proteolysis) yielded high-quality fluorescent labeling that consistently denoted sites of injury previously described in this model of TAI. Further, the conversion of the tyramide label to an electron-dense DAB reaction product allowed this same fluorescently tagged structure to be easily carried forward to the EM level. We emphasize that the reaction presented here does not involve a direct conversion of the rhodamine tag but rather proceeds in relation to the co-localized biotin–tyramide. This strategy provided extremely high resolution and fidelity because virtually all the axonal profiles detected with the rhodamine tag were also labeled with DAB. It should be noted that none of the axonal profiles displaying single labeling with the coumarin marker alone demonstrated DAB reactivity. This finding further underlines the specificity and fidelity of the technique. As noted, at the ultrastructural level the tissue was well preserved, with the suggestion that the intensity of the staining was even more pronounced than that achieved through the tradititional HRP–DAB methods. Although we are uncertain as to why this was the case, we postulate that the concomitant presence of the electron-dense tyramide molecule itself enhanced the density of the overall reaction product (Mayer and Bendayan 1997). In relation to the observed electron-dense reaction, we also note that the reaction product appeared to be

Büki, Walker, Stone, Povlishock of consistent size and density, showing consistent localization. To be candid, we were surprised by this finding because the histochemistry underlying tyramide amplification involves a tyramide–tyrosine covalent binding that is not directly determined by immunological principles (Bobrow et al. 1989,1992; van Gijlswijk et al. 1996,1997). In this scenario, one would predict a more widespread electron-dense reaction product, reflecting more random amplification. Why our method is not confounded by this liability is unknown. However, we speculate that our modification of the original commercially available TSA protocol at several key steps may have provided optimal signal-to-noise ratio at both the LM and EM level while achieving an optimal cost–benefit ratio. Also of note in this process is the fact that the simultaneous visualization of the biotin molecule by the DAB chromogen enables the investigator to overcome the problem of fluorescent signal fading over time. We believe that our study convincingly demonstrates the versatility of the TSA method. Therefore, in addition to being an excellent technique for augmenting immunocytochemical reactions, it is also considered an outstanding tool for immunofluorescent double/multiple labeling, with the possibility of a simple signal conversion method for EM. Furthermore, because the simultaneous application of biotin– and rhodamine–tyramide facilitates ultrastructural analysis of double-labeled immunofluorescent profiles, this opens up new potential for the future use of this sensitive, versatile signal amplification technique. Acknowledgments Supported by grants NS 20193 and by the Martin Rodbell Fellowship. We thank Dr Robert Siman for kindly donating the Ab38 and Dr John Q. Trojanowski for the RMO-14 antibody. We also thank Lynn Davis, Thomas Coburn, and Judy Williamson for excellent technical support.

Literature Cited Bobrow MN, Harris TD, Shaughnessy KJ, Litt GJ (1989) Catalyzed reporter deposition, a novel method of signal amplification. Application to immunoassays. J Immunol Methods 125:279–285 Bobrow MN, Litt GJ, Shaughnessy KJ, Mayer PC, Conlon J (1992) The use of catalyzed reporter deposition as a means of signal amplification in a variety of formats. J Immunol Methods 150:145–149 Büki A, Okonkwo DO, Povlishock JT (1999b) Postinjury cyclosporin A administration limits axonal damage and disconnection in traumatic brain injury. J Neurotrauma 16:511–521 Büki A, Siman R, Trojanowski JQ, Povlishock JT (1999a) The role of calpain-mediated spectrin proteolysis in traumatically induced axonal injury. J Neuropathol Exp Neurol 58:365–375 Christman CW, Grady MS, Walker SA, Holloway KL, Povlishock JT (1994) Ultrastructural studies of diffuse axonal injury in humans. J Neurotrauma 11:173–186 de Haas RR, Verwoerd NP, van der Corput MP, van Gijlswijk RP, Siitari H, Tanke HJ (1996) The use of peroxidase-mediated deposition of biotin–tyramide in combination with time-resolved fluorescence imaging of europium chelate label in immunohis-

Conversion of Immunofluorescence to EM by TSA tochemistry and in situ hybridization. J Histochem Cytochem 44:1091–1099 Erber WN, Willis JI, Hoffman GJ (1997) An enhanced immunocytochemical method for staining bone marrow trephine sections. J Clin Pathol 50:389–393 Humbel BM, de Jong MD, Muller WH, Verkleij AJ (1998) Preembedding immunolabeling for electron microscopy: an evaluation of permeabilization methods and markers. Microsc Res Tech 42:43–58 Hunyady B, Krempels K, Harta G, Mezey E (1996) Immunohistochemical signal amplification by catalyzed reporter deposition and its application in double immunostaining. J Histochem Cytochem 44:1353–1362 Kaufmann O, Baume H, Dietel M (1998) Detection of oestrogen receptors in non-invasive and invasive transitional cell carcinomas of the urinary bladder using both conventional immunohistochemistry and the tyramide staining amplification (TSA) technique. J Pathol 186:165–168 Kressel M (1998) Tyramide amplification allows anterograde tracing by horseradish peroxidase-conjugated lectins in conjunction with simultaneous immunohistochemistry. J Histochem Cytochem 46:527–533 Lee VM, Carden MJ, Schlaepfer WW, Trojanowski JQ (1987) Monoclonal antibodies distinguish several differentially phosphorylated states of the two largest rat neurofilament subunits (NF-H and NF-M) and demonstrate their existence in the normal nervous system of adult rats. J Neurosci 7:3474–3488 Lubke J (1993) Photoconversion of diaminobenzidine with different fluorescent neuronal markers into a light and electron microscopic dense reaction product. Microsc Res Tech 24:2–14 Luby–Phelps K, Amato PA, Taylor DL (1984) Selective immunocytochemical detection of fluorescent analogs with antibodies specific for the fluorophore. Cell Motil 4:137–149 Marmarou A, Foda MA, van den Brink W, Campbell J, Kita H, Demetriadou K (1994) A new model of diffuse brain injury in rats. Part I: Pathophysiology and biomechanics. J Neurosurg 80:291–300 Maxwell WL, Povlishock JT, Graham DL (1997) A mechanistic analysis of nondisruptive axonal injury: a review [published erratum appears in J Neurotrauma 1997;14:755]. J Neurotrauma 14:419–440 Mayer G, Bendayan M (1997) Biotinyl-tyramide: a novel approach for electron microscopic immunocytochemistry. J Histochem Cytochem 45:1449–1454 Okonkwo DO, Büki A, Siman R, Povlishock JT (1999) Cyclosporin A limits calcium-induced axonal damage following traumatic brain injury. Neuroreport 10:353–358 Okonkwo DO, Pettus EH, Moroi J, Povlishock JT (1998) Alteration of the neurofilament sidearm and its relation to neurofilament compaction occurring with traumatic axonal injury. Brain Res 784:1–6 Pettus EH, Christman CW, Giebel ML, Povlishock JT (1994) Traumatically induced altered membrane permeability: its relationship to traumatically induced reactive axonal change. J Neurotrauma 11:507–522 Posmantur R, Kampfl A, Siman R, Liu J, Zhao X, Clifton GL, Hayes RL (1997) A calpain inhibitor attenuates cortical cytoskeletal protein loss after experimental traumatic brain injury in the rat. Neuroscience 77:875–888 Povlishock JT, Marmarou A, McIntosh T, Trojanowski JQ, Moroi J (1997) Impact acceleration injury in the rat: evidence for focal axolemmal change and related neurofilament sidearm alteration. J Neuropathol Exp Neurol 56:347–359 Powell RD, Halsey CM, Hainfeld JF (1998) Combined fluorescent and gold immunoprobes: reagents and methods for correlative light and electron microscopy. Microsc Res Tech 42:2–12 Powell RD, Halsey CM, Spector DL, Kaurin SL, McCann J, Hainfeld JF (1997) A covalent fluorescent-gold immunoprobe: simultaneous detection of a pre-mRNA splicing factor by light and electron microscopy. J Histochem Cytochem 45:947–956 Punnonen EL, Fages C, Wartiovaara J, Rauvala H (1999) Ultrastructural localization of beta-actin and amphoterin mRNA in

161 cultured cells: application of tyramide signal amplification and comparison of detection methods. J Histochem Cytochem 47:99–112 Roberts–Lewis JM, Savage MJ, Marcy VR, Pinsker LR, Siman R (1994) Immunolocalization of calpain I-mediated spectrin degradation to vulnerable neurons in the ischemic gerbil brain. J Neurosci 14:3934–3944 Saatman KE, Bozyczko–Coyne D, Marcy V, Siman R, McIntosh TK (1996) Prolonged calpain-mediated spectrin breakdown occurs regionally following experimental brain injury in the rat. J Neuropathol Exp Neurol 55:850–860 Sandell JH, Masland RH (1988) Photoconversion of some fluorescent markers to a diaminobenzidine product. J Histochem Cytochem 36:555–559 Schofer C, Weipoltshammer K, Almeder M, Wachtler F (1997) Signal amplification at the ultrastructural level using biotinylated tyramides and immunogold detection. Histochem Cell Biol 108:313–319 Shindler KS, Roth KA (1996) Double immunofluorescent staining using two unconjugated primary antisera raised in the same species. J Histochem Cytochem 44:1331–1335 Shiosaka S, Shimada S, Tohyama M (1986) Sensitive double-labeling technique of retrograde biotinized tracer (biotin-WGA) and immunocytochemistry: light and electron microscopic analysis. J Neurosci Methods 16:9–18 Siman R, Noszek JC, Kegerise C (1989) Calpain I activation is specifically related to excitatory amino acid induction of hippocampal damage. J Neurosci 9:1579–1590 Speel EJ, Hopman AH, Komminoth P (1999) Amplification methods to increase the sensitivity of in situ hybridization. Play card(s). J Histochem Cytochem 47:281–288 Speel EJ, Saremaslani P, Roth J, Hopman AH, Komminoth P (1998) Improved mRNA in situ hybridization on formaldehyde-fixed and paraffin-embedded tissue using signal amplification with different haptenized tyramides. Histochem Cell Biol 110:571–577 Stanarius A, Topel I, Schulz S, Noack H, Wolf G (1997) Immunocytochemistry of endothelial nitric oxide synthase in the rat brain: a light and electron microscopical study using the tyramide signal amplification technique. Acta Histochem 99:411–429 Stone JR, Walker SA, Povlishock JT (1999) The visualization of a new class of traumatically injured axons through the use of a modified method of microwave antigen retrieval. Acta Neuropathol (Berl) 97:335–345 Sun XJ, Tolbert LP, Hildebrand JG (1995) Using laser scanning confocal microscopy as a guide for electron microscopic study: a simple method for correlation of light and electron microscopy. J Histochem Cytochem 43:329–335 Teramoto N, Szekely L, Pokrovskaja K, Hu LF, Yoshino T, Akagi T, Klein G (1998) Simultaneous detection of two independent antigens by double staining with two mouse monoclonal antibodies. J Virol Methods 73:89–97 Todd AJ (1997) A method for combining confocal and electron microscopic examination of sections processed for double- or triplelabelling immunocytochemistry. J Neurosci Methods 73:149– 157 van de Corput MP, Dirks RW, van Gijlswijk RP, van de Rijke FM, Raap AK (1998) Fluorescence in situ hybridization using horseradish peroxidase-labeled oligodeoxynucleotides and tyramide signal amplification for sensitive DNA and mRNA detection. Histochem Cell Biol 110:431–437 van Gijlswijk RP, Wiegant J, Raap AK, Tanke HJ (1996) Improved localization of fluorescent tyramides for fluorescence in situ hybridization using dextran sulfate and polyvinyl alcohol. J Histochem Cytochem 44:389–392 van Gijlswijk RP, Zijlmans HJ, Wiegant J, Bobrow MN, Erickson TJ, Adler KE, Tanke HJ, Raap AK (1997) Fluorochrome-labeled tyramides: use in immunocytochemistry and fluorescence in situ hybridization. J Histochem Cytochem 45:375–382 Van Heusden J, de Jong P, Ramaekers F, Bruwiere H, Borgers M, Smets G (1997) Fluorescein-labeled tyramide strongly enhances the detection of low bromodeoxyuridine incorporation levels. J Histochem Cytochem 45:315–319