Chromogen-based Immunohistochemical Method for Elucidation of the Coexpression of Two Antigens Using Antibodies from the Same Species

Abstract

The authors established a chromogen-based, double immunolabeling method using antibodies from the same species without any unwanted cross-reactivity. In addition, time-consuming staining steps were shortened by using polymer-based secondary antibodies. Taking advantage of the nature of the chromogen 3-amino-9-ethylcarbazole (AEC), which is used as a horseradish peroxidase substrate for antibody detection, the AEC-derived signals in the first color development were easily eliminated by alcohol treatment. Therefore, the signals from the first staining did not interfere with those from the subsequent second staining, which used the chromogen 3,3′-diaminobenzidine. The co-localization of antigens within the same cell could be confirmed using this method, because cell images of the individual dye staining steps could be obtained and developed. The images from each step could be expressed in pseudo-colors in a dark field by using a computer. As a result, merged images could be constructed that resembled the images acquired by the fluorescent immunolabeling technique. The resolution of this method enabled analysis of the coexpression of two antigens in the same cell in the same section. The authors have named this staining technique the elucidation of the coexpression of two antigens in a cell using antibodies from the same species (ECSS).

Various methods have been developed to date to detect antigens in animal and human tissues. One of the most important of these methods for histological, pathological, and cell biological studies is the double immunohistochemical method, which allows analysis of the coexpression of two independent antigens in the same cell or in the same compartment of a cell. Today, double and even triple immunohistochemical methods are routinely used in biology labs.

In conventional immunohistochemistry, single immunohistochemical methods are used to detect one antibody-bound antigen using the chromogen 3,3′-diaminobenzidine (DAB), which is a horseradish peroxidase (HRP) substrate that is used for secondary HRP-conjugated antibody detection. When double immunohistochemical methods are required, fluorescent-labeled secondary antibodies are used. It is normally difficult to determine the coexpression of two antigens using dyes on the same tissue section because of the following limitations: (1) Two primary antibodies derived from different species have to be prepared, and (2) it is hard to distinguish the cellular area in which two antigens are coexpressed because the mixture of dyes interferes with observation of the cell structures. In the first case above, when using the same tissue section for multiple rounds of staining, there is often unwanted cross-reactivity between the primary antibody in the first round of staining (e.g. for Protein A) and the secondary antibody used in the second round of staining (e.g. for Protein B) and vice versa, particularly when the same species of primary antibody is used in the subsequent staining. This limitation is now being overcome by microwave treatment of the tissue between the two rounds of staining, also known as the antigen retrieval method, which inactivates the HRP activity of the secondary antibody and the antigenicity of both the primary and secondary antibodies used in the first round of staining (Lan et al. 1995). However, the second limitation is still a problem when using dye-based double immunohistochemical methods (Lan et al. 1995).

In the present study, we employed an HRP-conjugated polymer-based secondary antibody for double immunohistochemical staining and showed that the HRP activity and the antigenicity of HRP-conjugated polymer-based immunoglobulins could be inactivated by using conventional microwave treatment. We also showed that signals derived from the chromogen 3-amino-9-ethylcarbazole (AEC), which is another HRP substrate, could be easily extinguished by alcohol treatment and after the second color development. Elimination of the AEC-derived signals in this way made it possible to show only the signals derived from the second immunolabeling step. Finally, we showed that it is possible to express each signal in pseudo-color, so that coexpression and/or co-localization of antigens could be visualized in the stained images in a manner similar to when fluorescent-labeled double immunohistochemical methods are employed.

Materials and Methods

Tissue Sections

Eight-week-old male Sprague-Dawley rats weighing 250 to 300 g (SLC Co., Shizuoka, Japan) were anesthetized with intraperitoneally injected pentobarbital (5 mg/100 g body weight; Abbott Laboratories, North Chicago, IL). Animals were perfused transcardially with 30 ml of ice-cold saline, followed by ice-cold 4% paraformaldehyde–phosphate-buffered saline (PBS), pH 7.4. The spinal cords were then dissected, cut into pieces, and cryoprotected by dipping into PBS containing 30% sucrose for 16 hr at 4C. The tissue blocks were frozen on dry ice, and 30-µm sections were cut with a sliding microtome (SM2000R; Leica, Solms, Germany) and stored in tissue collection solution (25% glycerol, 30% ethylene glycol, 0.05 M phosphate buffer [PB], pH 7.4) at −80C until use. This study was approved by the Committee of Animal Ethics of the National Defense Medical College.

Double Immunohistochemical Method Using Two Primary Antibodies from the Same Species

A brief protocol for immunolabeling is outlined in Table 1.

Table 1.

Brief Protocol for ECSS

Condition		Experimental	No First Primary Antibody	No Second Primary Antibody	Neither Primary Antibodies
First step	MW treatment	+	+	+	+
	Primary antibody	Antibody A	−	Antibody A	−
	Secondary antibody	Secondary Ab	Secondary Ab	Secondary Ab	Secondary Ab
	Chromogen	AEC	AEC	AEC	AEC
Photo		+	+	+	+
Second step	MW treatment	+	+	+	+
	Primary antibody	Antibody B	Antibody B	−	−
	Secondary antibody	Secondary Ab	Secondary Ab	Secondary Ab	Secondary Ab
	Chromogen	DAB	DAB	DAB	DAB
Photo		+	+	+	+
Quenching AEC		+	+	+	+
Photo		+	+	+	+

AEC, 3-amino-9-ethylcarbazole; DAB, 3,3′-diaminobenzidine; ECSS, elucidation of the coexpression of two antigens in a cell using antibodies from the same species; MW, microwave; Ab, antibody. Quenching is performed with alcohol.

In the first staining step, sections were attached to MAS-coated slide glasses (S9442; Matsunami Glass, Osaka, Japan) in 0.25 M PB (pH 7.4) and dried for 24 h. Next, sections were subjected to microwave (MW) treatment. In brief, sections were treated by microwaving in 10 mM citrate buffer (pH 6.0) for 7 min at 100C, using a microwave oven (RE-T1; SHARP, Osaka, Japan) set at 700 W and 2.45 GHz and then cooled for 20 min at room temperature. After MW treatment, sections were washed several times in Tris-buffered saline (TBS) and blocked in TBS containing 5% normal goat serum (NGS) and 0.5% Triton for 30 min. The sections were then incubated with primary antibodies (Table 2) for 1 hr at room temperature. The primary antibodies were diluted with TBS containing 5% NGS and 0.5% Triton X-100 before use. As a control, the same concentration of normal mouse IgG was applied instead of the primary antibody. After the sections were washed three times with TBS containing 0.05% Tween (TBS-T), they were incubated for 1 hr in the HRP-labeled secondary antibody solution (Table 3) (DakoCytomation; Kyoto, Japan). After the sections were washed three times with TBS-T, signals were developed with AEC solution (K3464; DakoCytomation) following the manufacturer’s instruction. After washing with TBS, sections were covered with a few drops of TBS and temporarily mounted without excess pressure from the top to prevent the sections from attaching to the coverslips. The sections were then digitally photographed using a BX50 microscope (Olympus; Tokyo, Japan) equipped with a CCD camera (VB-7010; Keyence, Osaka, Japan). After photographs were taken, slide glasses were dipped in PBS in a coplin jar. The coplin jar was mildly agitated to remove coverslips, preventing any damage to the sections. Then, the sections were subjected to MW treatment as described above to inactivate immunoglobulin antigenicity and HRP activity.

Table 2.

Primary Antibodies Used in This Study

Antibodies	Product No.	Source	Note
Anti-GFAP	G-3893	Sigma	Mouse monoclonal
Anti-synaptophysin	573822	Calbiochem	Mouse monoclonal
Anti-SLC18a3		In-house	Rabbit polyclonal
Anti-neuN	MAB377	Chemicon	Mouse monoclonal

GFAP, glial fibrillary acidic protein; SLC18a3, vesicular acetylcholine transporter; neuN, neuronal nuclear antigen.

Table 3.

Secondary Antibodies Used in This Study

Antibodies	Product No.	Source	Note
EnVision+ System HRP-labeled polymer anti-mouse	K4000	DakoCytomation	Anti-mouse Igs
EnVision+ System HRP-labeled polymer anti-rabbit	K4002	DakoCytomation	Anti-mouse Igs

HRP, horseradish peroxidase.

In the subsequent round of staining, the sections were further stained to detect a different antigen by incubation with the second primary antibody for 1 hr, followed by its corresponding secondary antibody (DakoCytomation), as described above. As a control, the same concentration of normal mouse IgG was applied instead of the primary antibody. The color development was performed using 0.05% DAB solution in the presence of 0.003% H₂O₂ for 10 min at room temperature, instead of AEC. After washing with 0.1 M PB, the sections were temporarily mounted in TBS as described above, and the same area that was imaged after the first staining step was digitally photographed. After removing the coverslips, the sections were subsequently dehydrated through an ascending ethanol series and xylene and were covered with a coverslip, and photographed once more. AEC signals had disappeared in this step. The same area was then imaged again. The images acquired between first and second staining steps were used as references to check for the coexistence of multiple signals in the same cells.

Pseudo-color Imaging

The images obtained with AEC or DAB dyes were processed using Photoshop CS5 (Adobe Systems; Tokyo, Japan) as described below, so that they were expressed in pseudo-colors. The images were converted to gray-scale images and the scales were inverted to convert the bright-field images to dark-field ones. The images were then again converted to RBG images, and all the gains except for the gain of the desired color were omitted. For example, when green was selected for a pseudo-color, the gain of red and blue was omitted to get the green fluorescent–like images. Merged images were also produced to improve recognition of coexpressed and/or co-localized proteins in a manner similar to that used for fluorescent-labeled double immunohistochemical methods, using images that included both AEC and DAB as references.

Results

To determine whether AEC signals on immunohistochemically stained slides were removed by alcohol treatment, we first stained tissue sections with a mouse anti–glial fibrillary acidic protein (GFAP) antibody and developed signals using HRP-conjugated anti-mouse secondary antibody and the HRP substrate AEC (Fig. 1A). After MW treatment, the sections were then subjected to alcohol treatment. AEC signals could not be detected after the MW and alcohol treatments (Fig. 1B). To confirm the effectiveness of the MW and alcohol treatments, we next applied a second HRP substrate, DAB, to these slides. No HRP activity was detected with this substrate, indicating that all of the HRP activity was inactivated (Fig. 1C). We next tested whether the antigenicity of this primary antibody could be extinguished after MW and alcohol treatment. After the first antigen staining, using the same primary and secondary antibodies and AEC, (Fig. 1D), the sections were then subjected to MW and alcohol treatment followed by a second round of antigen staining, in which this primary antibody for a different antigen could be omitted, and only the anti-mouse secondary antibody and DAB were added to the tissue section. Any primary antibody that remained from the first round of staining that might have still been available for recognition by the second antibody should have been detected in this step. No unwanted residual signals derived from the primary antibody in the first step could be detected (Fig. 1E). However, if the same procedure was performed using an excess amount of the initial primary antibody for staining, an unwanted signal is detectable (Fig. 1F). These results suggest that AEC signals can be extinguished by MW and alcohol treatment.

Figure 1.

Effects of 100% ethanol treatment on 3-amino-9-ethylcarbazole (AEC)–derived signals and of microwave (MW) treatment on horseradish peroxidase (HRP) activity and antigenicity of primary antibodies. All sections were stained with mouse anti–glial fibrillary acidic protein (GFAP) antibodies (1:10,000), HRP-conjugated anti-mouse antibodies (DAKO, K4000), and AEC according to the first staining step described in Materials and Methods (A). Subsequently, the same sections were directly subjected to MW treatment, destained with 100% ethanol (B), and dipped in DAB staining solution (C). The other AEC-stained sections (D) were subjected to MW treatment and alcohol treatment, incubated again with HRP-conjugated secondary antibody (DAKO, K4000), and developed by DAB for 10 min (E). When sections were processed as other sections (E) had been done, except that excess primary antibody was applied (1:1000) in the first round of staining, unwanted cross-reactive signals were detected (F). Bar = 50 µm.

We next applied this staining method to double staining of slides using a mouse anti-synaptophysin antibody as the primary antibody in the first round of staining and a mouse anti-GFAP antibody as the primary antibody in the second round of staining (Fig. 2). In this double-staining method, the sections were microwaved between the first and second round of staining, but AEC signals were not extinguished by alcohol treatment until after the signals of the second staining step had been developed. Typical synaptophysin signals could be detected from sections to which anti-synaptophysin antibody was applied and signals were developed using HRP-conjugated secondary antibodies and AEC substrate (Fig. 2A, G). No signals were detected from sections that were stained using normal mouse IgG in place of the anti-synaptophysin antibody (Fig. 2D, J). Following microwave treatment, an anti-GFAP antibody was then used as the primary antibody in the second round of staining. This signal was developed with HRP-conjugated secondary antibodies and a DAB substrate (Fig. 2B, E). As control, some sections were stained using normal mouse IgG instead of anti-GFAP (Fig. 2H, K). Before the synaptophysin signals were extinguished with alcohol treatment, the signals from both dyes could be recognized (Fig. 2B, E, H). After alcohol treatment of double-stained sections, the AEC signals were clearly extinguished (compare Fig. 2C and F, I, and L, respectively). These results indicate that this method could be a useful tool for double immunolabeling using primary antibodies from the same species without any cross-reactivity.

Figure 2.

Development of the dye-based double immunolabeling method using mouse antibodies. Four sections, including controls, were prepared for this experiment. In the first round of staining, sections were stained with mouse anti-synaptophysin antibody (1:1000) and anti-mouse HRP-conjugated secondary antibody, and signals were developed by 3-amino-9-ethylcarbazole (AEC) (A). In the second round of staining, following microwave treatment, the same sections were stained with mouse anti–glial fibrillary acidic protein (GFAP) antibody (1:10,000), and anti-mouse HRP-conjugated secondary antibody, and signals were developed with 3,3′-diaminobenzidine (DAB) (B). The AEC signals from the first round of staining were quenched by alcohol treatment (C). As controls, some sections were stained using normal mouse IgG in place of the primary antibody in the first round of staining (D–F), in the second round of staining (G–I), or in the both rounds of staining (J–L). Each image was taken after the staining or quenching indicated above the images. Note that signals from AEC completely disappeared following alcohol treatment (compare G and I), without affecting signals from DAB (compare E and F). Bar = 50 µm.

Next, the versatility of this double immunohistochemical method was checked. For this purpose, the primary antibody in the first round of staining and the primary antibody in the second round of staining were switched. Following the first round of staining, typical GFAP signals, developed with HRP-conjugated secondary antibody and AEC, could be detected (Fig. 3A, E), and no signals were detected in the control sections (Fig. 3C, G). Following treatment of the GFAP-stained sections with MW, processing through all of the steps in the second round of staining, and subsequent alcohol treatment, only synaptophysin signals could be detected (Fig. 3B, D). In contrast, GFAP signals could not be detected in the control (Fig. 3F). These results suggest that the primary antibodies used in the two rounds of staining could be successfully switched, meaning that this staining technique does not depend on the sensitivities of the two primary antibodies.

Figure 3.

Versatility of the double-staining technique in terms of the primary antibodies used. Four sections, including controls, were prepared for this experiment. In the first round of staining, sections were stained with mouse anti–glial fibrillary acidic protein (GFAP) antibodies (1:10,000), and signals were developed with 3-amino-9-ethylcarbazole (AEC) (A). In the second round of staining, following microwave treatment, the same sections were stained with mouse anti-synaptophysin antibodies (1:1000), and the signals were developed with 3,3′-diaminobenzidine (DAB). The AEC signals were quenched after staining by alcohol treatment (B). As controls, some sections were stained using blocking solution in place of the primary antibody in the first round of staining (C and D), in the second round of staining (E and F), or in the both rounds (G and H). Each image was taken after the staining or quenching, as indicated above the images. Bar = 50 µm.

Finally, we wanted to determine if this method could be applied for the analysis of the coexpression of two antigens in the same cell. Sections were stained first with the rabbit anti-SLC18a3 antibody, and the signals were developed with the anti-rabbit HRP-conjugated antibody and AEC (Fig. 4A). Subsequently, the sections were subjected to MW treatment, the mouse anti-NeuN antibody was applied, with signals developed using the anti-mouse HRP-conjugated secondary antibody and DAB, and the slide was treated with alcohol (Fig. 4B). Pseudo-color images were produced from these images using a computer-based method (Fig. 4C, D). The merged image of these two images clearly showed the coexpression of these two antigens in the same cell (Fig. 4E). The order with which these two primary antibodies were used was also switched to determine the versatility of this method (Fig. 4F,G), and these images were also converted to pseudo-color (Fig. 4H,I, respectively). A merged image (Fig. 4J) was also created. SLC18a3 is one of the vesicular acetylcholine transporters and expressed in cholinergic neurons. So SLC18a3 should be expressed in the cell body of the motoneuron in the anterior horn (Fig. 4A, arrows) in addition to the presynaptic terminals surrounding the cell body. On the other hand, NeuN is not expressed in the presynaptic terminals but in the cell body of most neurons (Fig. 4F, arrows). We used these markers to test the coexpression of two antigens in a cell using antibodies from the same species (ECSS) with no unwanted cross-reactivity (in presynaptic terminals). Thus, the cell body of the motoneuron and presynaptic terminals offer good locations to test the coexpression of these markers and identify any potential cross-reactivity, respectively. As a result, a part of the universality of this technology was successfully shown. An overview of ECSS is outlined (Fig. 5).

Figure 4.

Determination of the coexpression of two antigens in a cell using pseudo-color imaging of the immunohistochemical staining to enhance cell identification. The immunohistochemically image obtained by elucidation of the coexpression of two antigens in a cell using antibodies from the same species (ECSS) in the first antigen round of staining using the anti-SLC18a3 antibody (1:200) and 3-amino-9-ethylcarbazole (AEC) (A), or the image obtained using the anti-neuN antibody (1:32,000) and 3,3’-diaminobenzidine (DAB) (B) were changed to a gray scale and expressed as red and green pseudo-colors, respectively (C and D, respectively). A merged image is shown in (E). Images where the primary antibodies were switched (F, G) were also subjected to the pseudo-color image processing (H–J). Note that both antigens are expressed at the position where motoneurons are considered to be located in the spinal cord (arrows). Bar = 50 µm.

Figure 5.

Overview of the elucidation of the coexpression of two antigens in a cell using antibodies from the same species (ECSS). Schematic representation of the novel double-staining technique for ECSS. The chromogen 3-amino-9-ethylcarbazole (AEC) was used as the horseradish peroxidase (HRP) substrate-staining dye at the end of the first staining step, and 3,3′-diaminobenzidine (DAB) was used at the end of the second staining step. The image at the end of the second staining step was used as a reference. Subsequently, AEC signals were erased by alcohol treatment. Both AEC and DAB images were converted to pseudo-color. Merged images were produced from these images using the double-stained image as a reference for exact relative locations.

Discussion

For the novel staining system described in this article, we used a polymer-based HRP-conjugated secondary antibody system to shorten time-consuming staining steps. In this system, both IgG antigenicity and HRP activity were found to be completely inactivated by MW treatment as long as an appropriate concentration of the first antibody was used. The fact that alcohol treatment after MW treatment completely extinguished the AEC signals suggested that the AEC dye may be a good candidate dye that can be easily removed when performing two different stainings of the same section.

A method to perform double immunolabeling using two primary antibodies from the same species without any resulting cross-reactivity between the antibodies was established in this study. Theoretically, multiple staining of the same section using more than two antigens would be possible, taking advantage of the feature of the AEC dye that facilitates such lack of cross-reactivity. This method was also found to be versatile with regard to the primary antibodies used in each step (Fig. 3). Regarding the optimal concentration of antibodies, anti-synaptophysin and anti-GFAP were used at 1:1000 and 1:10,000 as primary antibodies in this order, respectively (Fig. 2), and later, the order of the antibodies was switched (Fig. 3). Moreover, anti-SLC18a3 and anti-NeuN were used at 1:200 and 1:32,000 in this order, and then the order of primary antibodies was switched (Fig. 4). The fact that switching the order of two primary antibodies did not affect the results, even in the situation where the optimal concentrations of two primary antibodies differed exceedingly, suggests that ECSS is not dependent on the sensitivities of two primary antibodies.

It was also shown that the combination of mouse and rabbit antibodies could be applicable as primary antibodies in ECSS (Fig. 4) in addition to the combination of two mouse primary antibodies (Figs. 1–3). This method, in combination with pseudo-color imaging, made it possible to determine the co-localization of two antigens within a cell. Because AEC-stained and DAB-stained images are captured from the same section both separately and collectively, we can exactly merge images stained separately without any speculation, even when using frozen sections that could potentially shrink and/or expand during the staining steps.

There have already been some studies reported regarding immunolabeling methods that use two primary antibodies from the same species (Osamura et al. 1981; Tidman et al. 1981; Würden and Homberg 1993; Shindler and Roth 1996; Hontanilla et al. 1997; Tornehave et al. 2000). Each of these methods indicated some limitation in using these antibodies to determine the coexpression of two antigens in the same cell: (1) The adjacent sections where target cells are cut just in the center should be stained (Osamura et al. 1981), (2) the sensitivity of the second round of staining should be higher than the first one (Shindler and Roth 1996), (3) both primary antibodies have to be monoclonal antibodies and belong to different immunoglobulin subclasses (Tidman et a1. 1981), (4) both primary antibodies have to be labeled by haptens, such as biotin, before staining (Würden and Homberg 1993), and (5) moderate microwave treatment is needed to prevent antibodies from eluting (Tornehave et al. 2000). These limitations seemed to be overcome by ECSS, in which most of the primary antibodies would be available using the same section, regardless of the sensitivities and the subtypes of primary antibodies, without direct prelabeling and without any difficult techniques.

In our experiments, we focused on the coexpression of antigens in the anterior horn of the rat spinal cord. It was relatively easy to identify the regions where photos should have been taken repeatedly, because the area where one photo could have been taken covered the whole anterior horn using a ×20 objective lens. No special equipment or software for positioning slides was used. There were no limitations with regard to the size of area examined, so long as you can identify the landmark in your section between successive rounds of staining. No special instruments or expensive reagents were required, except for the HRP-conjugated secondary antibody, DAB, AEC, alcohol, and microscopes that are conventionally used in most biology labs.

The universality of ECSS remains to be examined in various aspects, although it has overcome some critical limitations. One possible limitation in the present method is the concentration of the primary antibody in the first round of staining. The concentration of the primary antibody used in each step should be diluted to a point that is sufficient to detect the signal. If the concentration of the primary antibody is too high, it will be hard to completely inactivate the antigenicity of the primary antibody by MW treatment, and this will result in high background signals in the second round of staining (Fig. 1F). However, this may not be a big problem because scientists routinely do this type of preliminary dilution experiments to determine the lowest concentration of antibody. Another possible limitation is that the “dense” reaction products that HRP first yielded might be a potential problem for ECSS. It is known that HRP reaction products sometimes structurally interfere with the penetration of a second set of antibodies (Sternberger et a1. 1979). The last possible limitation is that, after the second round of staining is finished, it seems harder to recognize the first AEC signal (Fig. 2H). This seems to be mainly because the color of the AEC signal changes from red to brownish, rather than because the signals become less intense. Investigators need to be accustomed to these color-changing signals that are used only as a reference to locate the single images exactly, as illustrated in Figure 5. We need to check the universality of ECSS one by one in future experiments.

Immunohistochemical methods are also important for pathological diagnosis in hospitals. In most cases, single immunohistochemical methods are employed to detect one antibody-bound antigen using the chromogen. ECSS would be a more attractive method if all the monoclonal antibodies accumulated so far as diagnostic markers could be available to determine the coexpression of two antigens. In this study, the only tissue examined was rat spinal cord. ECSS could contribute to the pathology and cell biology field in the future by examining a variety of tissues, including lymph nodes, skin, kidney, liver, and intestine. The ability to use false color images to reproduce some features of fluorescent co-localization would be advantageous, particularly if it is applicable to the situations where fluorescent-labeled technology is not available, such as for formalin-fixed paraffin-embedded tissues and/or virtual slide scanners, which have high throughput.

In this report, we successfully developed a dye-based double immunohistochemical labeling method using antibodies from the same species without any unwanted cross-reactivity (ECSS). The greatest advantage of this method is that images of the single-stained sections can be acquired with those of double-stained sections. These images from double-stained sections allow analysis of the exact relative locations of each signal derived from different dyes, so this method can allow identification of cells that express these antigens. Therefore, it should be possible to clearly judge whether two different antigens are coexpressed within a cell.