Tyramide Signal Amplification Permits Immunohistochemical Analyses of Androgen Receptors in the Rat Prefrontal Cortex

Abstract

Research on neural androgen receptors (ARs) has traditionally focused on brain regions that regulate reproductive and aggressive behaviors, such as the hypothalamus and amygdala. Although many cells in the prefrontal cortex (PFC) also express ARs, the number of ARs per cell appears to be much lower, and thus, AR immunostaining is often hard to detect and quantify in the PFC. Here, we demonstrate that biotin tyramide signal amplification (TSA) dramatically increases AR immunoreactivity in the rat brain, including critical regions of the PFC such as the medial PFC (mPFC) and orbitofrontal cortex (OFC). We show that TSA is useful for AR detection with both chromogenic and immunofluorescent immunohistochemistry. Double-labeling studies reveal that AR+ cells in the PFC and hippocampus are NeuN+ but not GFAP+ and thus primarily neuronal. Finally, in gonadally intact rats, more AR+ cells are present in the mPFC and OFC of males than of females. Future studies can use TSA to further examine AR immunoreactivity across ages, sexes, strains, and different procedures (e.g., fixation methods). In light of emerging evidence for the androgen regulation of executive function and working memory, these results may help understand the distribution and roles of ARs in the PFC.

Introduction

Immunohistochemical studies examining androgen receptors (ARs) in the rodent brain have traditionally focused on regions associated with control of reproductive and aggressive behaviors, such as the medial preoptic area (MPOA), ventromedial hypothalamus (VMH), medial amygdala (MeA), lateral septum (LS), and bed nucleus of the stria terminalis (BNST). In these regions, the number of ARs per cell is high, and AR+ cell nuclei are readily distinguished.^1–5 Recent studies have examined ARs in regions associated with working memory and executive function, where ARs may have lower densities per cell. In such regions, like the medial prefrontal cortex (mPFC), orbitofrontal cortex (OFC), and isocortex,⁶ the number of ARs per cell is lower, and AR+ cell nuclei are harder to detect and quantify.

Many factors can influence the quality of steroid receptor immunostaining. First, tissue fixation can contribute to differences in signal strength. Butler et al.⁷ found that estrogen receptor (ERα) immunoreactivity in the rat cerebral cortex could be increased by fixing tissue with acrolein or a paraformaldehyde (PFA) and glutaraldehyde mixture, rather than PFA alone. Second, injecting animals with testosterone before tissue collection can improve AR immunostaining when using the anti-AR antibody, PG-21.^8,9 Third, for weak signal that is due to masked antigens, antigen retrieval methods are often successful for increasing the signal.^10,11 Fourth, for weak signal that is due to low abundance of antigen, tyramide signal amplification (TSA) methods are often highly effective.^12,13

First described by Bobrow et al.,¹⁴ TSA relies on the rapid deposition of a reporter molecule, biotinylated tyramine, at the site of the antigen and adjacent proteins to achieve amplification of signal.^15,16 TSA can be applied to both chromogenic and immunofluorescent immunohistochemistry and typically results in a 10- to 100-fold increase in signal.¹⁷

TSA has been used in immunostaining for ERα and ERβ,^18–20 glucocorticoid receptor,²¹ and progesterone receptor²² in both neural and non-neural tissues. TSA has also been used with AR immunostaining in muscle,^23,24 testis,^23,25 prostate,²⁶ blood vessels,²⁷ kidney, spleen, heart, and intestine,²³ and neural tissue including the spinal cord²³ and brain.^18,23,28

However, to our knowledge, TSA has never been used to examine ARs in brain regions, such as the prefrontal cortex (PFC), that have low amounts of ARs per cell. Although few studies have examined ARs in the PFC, there is strong evidence that androgens exert important effects on higher order cognitive functions mediated by the PFC.^6,29,30 Therefore, we evaluated the effect of TSA on AR immunostaining in the rat brain, particularly in prefrontal cortical regions with lower numbers of ARs per cell. The first study assessed the effect of TSA in a chromogenic immunohistochemistry protocol. The second study assessed the effect of TSA in a fluorescent immunohistochemistry protocol and included double-labeling to characterize AR+ cells as neuronal or glial.

Materials and Methods

All animal procedures complied with the Canadian Council on Animal Care guidelines for animal care and were approved by the University of British Columbia Animal Care Committee.

Study 1: Tyramide Signal Amplification Applied to Chromogenic Immunohistochemistry of ARs

Animals

Subjects were adult male Fischer 344 × Brown Norway F1 hybrid rats (F344/BN; National Institute on Aging, Taconic Farms) that were young (5 months, n=12) or aged (22 months, n=12). Rats were group-housed with rats of similar age for the first week after arrival. All cages contained Aspen chip bedding (NEPCO, Warrensburg, NY), paper towel, and a PVC pipe for environmental enrichment. Rats were given ad libitum access to standard lab chow (Rat Diet 5012; LabDiet, St. Louis, MO) and water. Rats were kept on a 12 hr:12 hr light:dark cycle at an average temperature of 21C and relative humidity of 40–50%.

Tissue Fixation

Subjects were euthanized by isoflurane inhalation and rapid decapitation. The brains were removed and cut in half along the sagittal plane. One half of each brain was immersion fixed in 4% PFA in phosphate-buffered saline (PBS) for 4 hr at room temperature, and then cryoprotected in a 30% sucrose solution at 4C for 48 hr or until they had sunk. Brains were then placed in aluminum foil packets, frozen on powdered dry ice, and stored at −80C until sectioning. Brains were sectioned on a cryostat into five series of 40-µm coronal sections. Free-floating sections were stored in an antifreeze solution (5-g polyvinylpyrrolidone, 150-g sucrose, 150-ml ethylene glycol, and 250-ml 0.1-M PBS, adjusted to 500 ml with dH₂O³¹) at −20C until immunohistochemical processing.

Immunohistochemistry

Of the five tissue series, two series were used here. Each series underwent immunohistochemical processing separately. All animals within a given series were processed together in one run (n=12 aged, n=12 young). The first series, hereafter referred to as TSA−, was stained for ARs using a protocol that did not include TSA. This protocol was extensively optimized under a number of different conditions to maximize sensitivity. The second series, hereafter referred to as TSA+, was stained for ARs (8 months after the TSA− series), using a protocol that included TSA. Because a TSA step was introduced, minor adjustments to the protocol were required to prevent increased background. For both series, the primary antibody was a monoclonal anti-AR antibody [EPR1535(2)] raised in the rabbit (ab133273; Abcam, Inc., San Francisco, CA). This antibody has been previously validated for specificity by Western blot,³² staining a single band at 110 kDa (manufacturer’s technical information³²). It has also been validated in testicular feminization mutant (Tfm) rat tissue lacking fully functional ARs, where Tfm rats showed an absence of AR staining³² (D. K. Hamson, personal communication, October 14, 2016). This antibody likely recognizes both bound and unbound ARs. It recognizes an epitope located within the first 30 amino acids of the AR protein (as per manufacturer), which is very similar to the epitope recognized by PG-21, a widely used anti-AR antibody³³ previously shown to recognize both bound and unbound ARs.³⁴ The secondary antibody was a biotin-SP-conjugated affinity-purified donkey anti-rabbit IgG (711-065-152; Jackson ImmunoResearch Laboratories, West Grove, PA).

TSA− (Control) Series

Sections were washed for 4 × 5 min in 0.1-M Tris-buffered saline (TBS), and then incubated in 0.5% hydrogen peroxide (H₂O₂) for 30 min to quench endogenous peroxidase activity. Sections were then washed for 5 × 5 min in 0.1-M TBS, and then for 5 min in 0.1-M PBS with 0.3% Triton X-100 and 0.1% gelatin (PBS-GT). Next, sections were blocked in 10% normal donkey serum (017-000-121; Jackson ImmunoResearch Laboratories) in PBS for 2 hr. Sections were then immediately incubated in primary antibody solution (1:200 in PBS-GT) for 1 hr at room temperature and then for 3 nights at 4C. Sections were then washed for 9 × 5 min in PBS-GT, and then incubated in a secondary antibody solution (1:500 in PBS-GT) for 1 hr at room temperature and then overnight at 4C. Following the secondary antibody incubation, sections were washed for 9 × 5 min in PBS-GT, then incubated in AB solution in TBS-T (PK-6100 Standard, VECTASTAIN ABC Kit; Vector Laboratories, Burlingame, CA) for 1 hr at room temperature, and then washed in PBS-GT for 4 × 5 min and TBS for 10 min. Sections were incubated in DAB and nickel substrate solution (Vector Laboratories Peroxidase Substrate Kit SK-4100) for 10 min at room temperature. Finally, sections were mounted onto gelatin-subbed slides, dehydrated, and cover-slipped.

TSA+ Series

This series followed a similar protocol, with the following differences. First, sections were incubated in a 2% H₂O₂ solution. Second, sections were blocked in 8% tryptone (J859; AMRESCO, Inc., Solon, OH) in PBS. Third, the primary antibody incubation was shortened to 24 hr at room temperature. Fourth, the secondary antibody was used at a concentration of 1:2000 in 0.8% tryptone in PBS. Fifth, sections also underwent a modified AB incubation schedule to accommodate the addition of TSA. Sections were incubated in AB solution (in 0.8% tryptone in PBS) for 30 min at room temperature, washed in PBS-GT for 3 × 5 min, and then incubated in a biotin tyramide solution (“BT”; made as described by Adams¹⁵; 15-ml PBS with 45-µL BT and 5-µL 30% H₂O₂) for 10 min, then washed in PBS-GT for 3 × 5 min, incubated in AB solution for 30 min, and washed in PBS-GT for 3 × 5 min. Finally, the DAB incubation time was extended to 20 min.

Additional Control Tissue

Additional AR immunostaining of intact and castrated male rat brain tissue was performed to confirm the specificity of the primary antibody. Previous studies have demonstrated that castration reduces neural AR immunoreactivity (AR-ir) to low or non-detectable levels.^3,5,35 Subjects were male Long–Evans rats (Charles River, Saint-Constant, Quebec, Canada) that were 2 months old when received. Rats were housed in the same conditions as described above. One week after arrival, rats underwent gonadectomy (n=1) or sham surgery (n=1). Six weeks following surgery, rats were euthanized by isoflurane inhalation and transcardially perfused with 0.9% NaCl in a 0.1-M phosphate buffer (pH 7.4), followed by 4% PFA. Following perfusion, brains were extracted and post-fixed in 4% PFA for 4 hr at room temperature, and then cryoprotected in a 30% sucrose solution at 4C for 72 hr or until they had sunk. Brains were frozen and sectioned, and AR immunohistochemistry with TSA was performed as described above.

Measurement of Immunoreactivity

Photomicrographs were taken using a Nikon Digital Sight DS-U1 camera and Nikon Eclipse 90i microscope (20× objective) using NIS-Elements Basic Research software (Nikon Canada, Inc., Richmond, British Columbia, Canada). All images were taken by one experimenter (K.L.L.), using identical resolution, gain, and exposure settings. Images were taken from serial brain sections—three MeA sections, four mPFC sections, and five lateral OFC sections—with one image taken per section. For each image, a background mean intensity (BMI) value was calculated by averaging six intensity measurements taken from random locations within each brain area where specific staining was not present. The BMI for each image was used to generate a threshold value that was 1.25 or 1.5 times as dark as the BMI. Only staining that was at or above threshold was quantified. Thresholds were applied to a defined region of interest (ROI) superimposed onto each photomicrograph within the limits of each brain region, and the program calculated the area percentage of the ROI that was above threshold (%AR-ir). The ROI had an area of 135,000 µm², and its placement within each photomicrograph was guided by the position of major neuroanatomical landmarks, such as the anterior forceps of the corpus callosum. AR-ir values were measured in two areas with low AR abundance per cell (mPFC and OFC) and in one area with high AR abundance per cell (MeA).

Statistical Analysis

Data were analyzed using SPSS Statistics software (Version 23.0 for Mac OS X; Chicago, IL). Pearson correlations between TSA+ and TSA− AR-ir values were computed for the OFC and MeA. The paired samples t-test was used to compare means of TSA+ and TSA− AR-ir in the mPFC, OFC, and MeA.

Study 2: Tyramide Signal Amplification Applied to Fluorescent Immunohistochemistry of ARs

Animals

Subjects were male and female adult (9–10 weeks old) Long–Evans rats (Charles River; n=3 per sex; as in the study by Ferris et al³⁶). Rats were separated by sex for housing in two colony rooms and group-housed in clear cages with Aspen chip bedding. Rats were given ad libitum access to standard lab chow (Rat Diet 5012; LabDiet) and water, and kept on a 12 hr:12 hr light:dark cycle. Rats were handled daily for 1 week before the start of the experiment.

Perfusion and Tissue Fixation

Subjects were euthanized via overdose of chloral hydrate (140 mg/kg intraperitoneal) and then transcardially perfused with 0.9% saline (60 ml) and 4% PFA (120 ml). Following perfusion, brains were extracted and post-fixed in 4% PFA for 4 hr at room temperature, and then cryoprotected in a 30% sucrose solution at 4C for 72 hr or until they had sunk. Brains were then flash frozen on powdered dry ice and stored at −80C. Brains were sectioned on a cryostat into 40-µm coronal sections, and free-floating sections were stored in an antifreeze solution at −20C until immunohistochemical processing.

Double-Label Immunofluorescence

Brain tissue from six animals (n=3 female, n=3 male) was processed in one run to fluorescently co-label ARs and NeuN, with and without TSA, as well as ARs and GFAP, with and without TSA.

In all conditions, sections were washed in PBS for 3 × 5 min, and then incubated in 0.5% H₂O₂ in PBS for 30 min. Unless otherwise stated, all subsequent washes were in PBS. Sections were washed for 3 × 5 min, then blocked in 5% normal donkey serum (Jackson ImmunoResearch Laboratories) in PBS-T (PBS with 0.2% Triton X-100) for 2 hr. Sections were then incubated in the primary antibodies, anti-AR (1:200) and anti-NeuN (1:500) in PBS-T, or anti-AR (1:200) and anti-GFAP (1:500) in PBS-T. The anti-AR antibody was the same as in study 1. The anti-NeuN antibody (MAB377; Millipore, Temecula, CA) and anti-GFAP antibody (MAB360; Millipore) were both monoclonal. Following primary antibody incubation for 24 hr at room temperature, sections were washed for 3 × 5 min, and then incubated in biotinylated donkey anti-rabbit IgG secondary antibody (same as in study 1) in PBS-T for 1 hr at room temperature and then overnight at 4C. After incubation, sections were then washed in PBS for 3 × 5 min. For TSA+ treatments only, sections were incubated in AB solution for 30 min, washed in PBS for 3 × 5 min, incubated in BT for 10 min, and washed in PBS for 3 × 5 min. For fluorophore labeling in all treatments, sections were incubated in streptavidin-conjugated Alexa Fluor 488 (1:300, S32354; Invitrogen, Carlsbad, CA) and Alexa Fluor 647 donkey anti-mouse IgG (1:500, A31571; Invitrogen) in PBS-T for 3 hr. Finally, sections were washed in PBS for 3 × 5 min, and mounted onto gelatin-coated slides and cover-slipped with ProLong Gold Mountant with 4′,6-diamidino-2-phenylindole (P36931; Invitrogen).

Confocal Scanning Laser Microscopy

A confocal laser scanning microscope (TCS SP8; Leica Microsystems, Wetzlar, Germany) was used to acquire Z-section images of tissue sections. High-magnification images (1024 × 1024 pixels) were acquired using a multiline argon ion laser and water-immersion 25× objective. For AR signal, emitted fluorescence was detected at 488 nm in the green channel. For NeuN or GFAP signal, emitted fluorescence was detected at 647 nm in the far red channel. In addition, to avoid any possible cross-talk between the two color channels, a sequential scanning mode was used during image collection. Final images were prepared using ImageJ and Adobe Photoshop software. Images were adjusted for brightness and contrast to improve image quality and to more effectively show co-localization. For a given region, similar changes were applied consistently to TSA+ and TSA− images, with the exception of CA1 of the hippocampus, for which it was not possible to apply equivalent adjustments to both TSA+ and TSA− images because of differing levels of background staining. Approximately 200 cells in each region (mPFC, OFC, and CA1 of the hippocampus) were examined by visual observation in each animal for co-localization of ARs with NeuN or GFAP in a single cell. The percentage of cells with co-localization was calculated as a proportion of the total number of cells counted.

Results

Study 1: Tyramide Signal Amplification Applied to Chromogenic Immunohistochemistry of ARs

Effect of TSA in Hypothalamus, Amygdala, and Lateral Septum

We first examined the hypothalamus, amygdala, and LS, which are areas that are well known to contain many AR+ cells and high amounts of ARs per cell. In this study, with or without TSA, the overall distribution of AR staining in these areas was qualitatively consistent with previous reports of AR protein and mRNA.^2,37 In addition, with or without TSA, omitting the primary antibody abolished AR staining.

As expected, AR-ir was clearly detectable in the amygdala, LS, and hypothalamus in control (TSA−) sections (Fig. 1A, C, and E). Compared with TSA− sections, TSA+ sections had more AR immunostaining. TSA treatment increased %AR-ir in the MeA, LS, and VMH (Fig. 1B, D, and F). For example, in the MeA, %AR-ir above threshold significantly increased from 6.56 ± 0.61% in TSA− sections to 17.62 ± 1.65% in TSA+ sections, paired t-test, t(22) = −8.19, p<0.0001 (Fig. 4). Furthermore, in the MeA, a within-subject comparison revealed that AR-ir values from the TSA+ and TSA− series were positively correlated (r = 0.64, p=0.001).

Figure 1.

Effect of TSA on AR immunoreactivity in the amygdala, lateral septum, and hypothalamus. Chromogenic immunohistochemistry with DAB was used (Study 1). All photos were taken from the same subject, without TSA (TSA−) or with TSA (TSA+). TSA+ sections (B, D, F) had higher AR immunoreactivity than TSA− sections (A, C, E), although AR staining was easily detectable without TSA. Regions shown are the MeA (A, B), LS (C, D), and VMH (E, F). Subject is an adult male Fischer 344 × Brown Norway F1 hybrid rat. Scale bar represents 100 µm. Abbreviations: TSA, tyramide signal amplification; AR, androgen receptor; MeA, medial amygdala; LS, lateral septum; VMH, ventromedial hypothalamus.

Figure 4.

Effect of TSA on AR immunoreactivity in the MeA, OFC, and mPFC. TSA increased AR immunoreactivity in all three regions. Data are mean ± SEM. n=21–24 total subjects. Abbreviations: TSA, tyramide signal amplification; AR, androgen receptor; MeA, medial amygdala; OFC, orbitofrontal cortex; mPFC, medial prefrontal cortex. ***p<0.0001.

Effect of TSA in Hippocampus

AR-ir has also been reported in the CA1 region of the hippocampus.^3,5,38 As expected, AR-ir was clearly detectable in the CA1 region in TSA− sections (Fig. 2A). AR-ir was even greater in the CA1 in TSA+ sections (Fig. 2B).

Figure 2.

Effect of TSA on AR immunoreactivity in the hippocampus. In the CA1 region of the hippocampus (A, B), a known AR-positive region, there was strong AR immunoreactivity in both TSA− (A) and TSA+ (B) sections, with stronger staining in the TSA+ section. In contrast, in the DG of the hippocampus (C, D), a known AR-negative region, there was no AR immunoreactivity in either TSA− (C) or TSA+ (D) sections. Note that the photomicrographs of CA1 and DG were taken from the same tissue section. Subject is an adult male Fischer 344 × Brown Norway F1 hybrid rat. Scale bar represents 100 µm. Abbreviations: TSA, tyramide signal amplification; AR, androgen receptor; DG, dentate gyrus.

Previous studies have not found AR-ir in the dentate gyrus of the hippocampus.^2,37 Here, we did not detect AR-ir in the dentate gyrus, with or without TSA treatment (Fig. 2C and D). The dentate gyrus serves as a useful negative control, especially because the dentate gyrus and CA1 were present on the same sections.

Effect of TSA in the PFC

In the PFC, which includes the mPFC and OFC, in TSA− sections, AR-ir was very faint and was not sufficiently above the background to quantify (Fig. 3A and C). In TSA+ sections, there was a robust increase in AR staining intensity in both the mPFC and OFC and numerous cells were above the background (Fig. 3B and D).

Figure 3.

Effect of TSA on AR immunoreactivity in the PFC. In TSA− sections (A, C), AR immunoreactivity was faint and difficult to detect above background. In contrast, in TSA+ sections (B, D), AR immunoreactivity was clearly detectable and quantifiable. Regions shown are the prelimbic portion of the mPFC (A, B) and the lateral OFC (C, D). Subject is an adult male Fischer 344 × Brown Norway F1 hybrid rat. Scale bar represents 100 µm. Abbreviations: TSA, tyramide signal amplification; AR, androgen receptor; PFC, prefrontal cortex; mPFC, medial prefrontal cortex; OFC, orbitofrontal cortex.

In the mPFC, AR-ir above threshold significantly increased from 0% in TSA− sections to 2.13 ± 0.38% in TSA+ sections, paired t-test, t(20) = −5.66, p<0.0001 (Fig. 4). In the OFC, AR-ir above threshold significantly increased from 0.76 ± 0.11% to 3.54 ± 0.49%, t(23) = −6.06, p<0.0001 (Fig. 4), and a within-subject comparison revealed AR-ir from the TSA− and TSA+ series showed a trend to be positively correlated (r = 0.39, p=0.06).

Omitting the primary antibody abolished AR staining in the mPFC and OFC in the TSA− and TSA+ sections.

Additional Control Tissue

A comparison of TSA+ AR staining in intact and castrated male rats showed a large reduction in AR staining in the castrated rat across several brain regions, including the MeA, CA1 of the hippocampus, and the mPFC (Fig. 5). In particular, in the mPFC, AR staining in the castrated rat was extremely faint and almost undetectable (Fig. 5F). The castrated subject serves as a negative control to validate the specificity of the primary antibody.

Figure 5.

AR immunoreactivity in an intact male rat and a castrated male rat. Regions shown are the MeA (A, B), the CA1 region of the hippocampus (C, D), and the mPFC (E, F). In these regions, the intact male (A, C, E) had stronger AR immunoreactivity than the castrated male (B, D, F). Subjects are young adult male Long–Evans rats. Tissue was collected at 6 weeks following sham surgery or gonadectomy. Scale bar represents 100 µm. Abbreviations: AR, androgen receptor; MeA, medial amygdala; mPFC, medial prefrontal cortex.

Study 2: Tyramide Signal Amplification Applied to Fluorescent Immunohistochemistry of ARs

Effect of TSA

In Study 2, in TSA− sections, AR-ir was not visible in any brain regions (Fig. 6A, C, and E). The staining difference between Studies 1 and 2 in TSA− sections was likely due to the more sensitive enzyme-catalyzed chromogenic detection system included in Study 1. Note that in TSA− sections from Study 2, some background staining was present, especially at tissue edges (Fig. 6A). In addition, a number of other differences between Studies 1 and 2 such as strain and fixation methods could have contributed to the staining difference.

Figure 6.

Effect of TSA on AR immunoreactivity with fluorescent immunohistochemistry (Study 2). Each pair of photos was taken from the same subject, without TSA (TSA−) or with TSA (TSA+). In TSA− sections (A, C, E), specific AR immunoreactivity was not detectable, although there was some background staining at the tissue edges. In contrast, in TSA+ sections (B, D, F), AR immunoreactivity was clearly detectable. Regions shown are the prelimbic portion of the mPFC, with the medial edge of the tissue section visible at left (A, B), OFC (C, D), and CA1 region of the hippocampus (CA1; E, F). In CA1, arrow denotes pyramidal cell layer. Images have been adjusted similarly for brightness and contrast. These photomicrographs were taken from the same subject, an adult male Long–Evans rat (n=3 subjects total). Scale bar represents 100 µm. Abbreviations: TSA, tyramide signal amplification; AR, androgen receptor; mPFC, medial prefrontal cortex; OFC, orbitofrontal cortex.

Nonetheless, in TSA+ sections, AR-ir was indeed detectable, and the overall pattern of AR staining closely matched the pattern seen in TSA+ sections from Study 1 (Fig. 6B, D, and F).

As expected, the NeuN and GFAP signals were not affected by TSA treatment.

AR+ Cell Phenotype

In the mPFC, OFC, and CA1 region of the hippocampus, 100% of AR+ cells were NeuN+, consistent with a neuronal phenotype (Fig. 7). However, not all NeuN+ cells were AR+. Furthermore, 0% of AR+ cells in these regions were GFAP+ (n=3 males; >600 cells total; Fig. 8).

Figure 7.

Cellular co-localization of ARs and NeuN in the prefrontal cortex and hippocampus. Regions shown are the mPFC (A–C), OFC (D–F), and CA1 region of the hippocampus (CA1; G–I). AR immunoreactivity is shown in green (A, D, G), and NeuN immunoreactivity is shown in red (B, E, H). In merged images (C, F, I), co-localization is shown in yellow. In all three regions, all AR+ cells were NeuN+. All images have been adjusted similarly for brightness and contrast. Subject shown is an adult male Long–Evans rat (n=3 subjects total). Scale bar represents 100 µm. Abbreviations: AR, androgen receptor; mPFC, medial prefrontal cortex; OFC, orbitofrontal cortex.

Figure 8.

Lack of cellular co-localization of ARs and GFAP in the prefrontal cortex and hippocampus. Regions shown are the mPFC (A–C), OFC (D–F), and CA1 region of the hippocampus (CA1; G–I). AR immunoreactivity is shown in green (A, D, G), and GFAP immunoreactivity is shown in red (B, E, H). In merged images (C, F, I), co-localization (if present) would be shown in yellow. In all three regions, no AR+ cells were GFAP+. All images have been adjusted similarly for brightness and contrast. Subject shown is an adult male Long–Evans rat (n=3 subjects total). Scale bar represents 100 µm. Abbreviations: AR, androgen receptor; mPFC, medial prefrontal cortex; OFC, orbitofrontal cortex.

Sex Differences in AR-ir in the PFC

AR staining was present in the mPFC and OFC of intact males (n=3; Fig. 9A and C). In contrast, AR staining was not detectable in the mPFC or OFC of intact females, even in TSA+ sections (n=3; Fig. 9B and D). However, AR-ir was present in other regions of the female brain. For example, in females, there was moderate to high intensity of AR staining in the MeA and CA1 of the hippocampus.

Figure 9.

Sex difference in AR immunoreactivity in the prefrontal cortex. Regions shown are the prelimbic portion of the mPFC (A, B) and OFC (C, D). In these regions, males (A, C) had detectable AR immunoreactivity, but females (B, D) did not. Females did have detectable AR immunoreactivity in other regions, such as CA1 of the hippocampus (not shown). All images have been adjusted similarly for brightness and contrast. Subjects are Long–Evans rats, three males and three females total. Scale bar represents 100 µm. Abbreviations: AR, androgen receptor; mPFC, medial prefrontal cortex; OFC, orbitofrontal cortex.

Discussion

Here, we evaluated the effect of TSA on AR immunostaining in the rat brain, especially in the PFC, using (1) chromogenic immunohistochemistry with DAB and (2) double-label fluorescent immunohistochemistry. TSA greatly increased AR staining intensity using chromogenic or fluorescent immunohistochemistry. The effects of TSA were clear in rats of different ages (in Study 1) or different strains (in Studies 1 and 2), suggesting wide applicability of TSA. In particular, the use of TSA allowed clear detection and quantification of ARs in the PFC (e.g., mPFC and OFC) and thus facilitates the study of androgen action in the rat cerebral cortex.

TSA Is Important for Visualizing ARs in the PFC

Evidence for the presence of ARs in the adult rat brain comes from studies using multiple techniques, including autoradiography and binding assays,^39–41 immunohistochemistry and Western blots with polyclonal antibodies,^37,42 in situ hybridization,^2,43,44 and RT-PCR.⁴⁵ Such studies are in clear accordance on AR presence in regions such as the VMH, MPOA, BNST, and MeA. However, for prefrontal cortical regions such as the mPFC and OFC, there is far less information available, although there are a few studies that find ARs in the cerebral cortex.^2,6,37,46

Reduced information about the PFC could be due, in part, to greater difficulty in visualizing ARs, as they are less abundant per cell, especially in comparison with the strong staining intensity seen in the hypothalamus and limbic system. Importantly, even though ARs per cell are less abundant in the cerebral cortex, there is a large population of AR+ cells in the cerebral cortex,⁶ suggesting androgen modulation of cortical functions. A few studies have examined the effects of androgens on the PFC and on behaviors mediated by the PFC, such as working memory and attentional set-shifting.^6,30,47–49 Consequently, the ability to detect and quantify ARs in the PFC is critical. Here, we found TSA makes a large difference and can be useful for studies that wish to detect or quantify ARs in the rat cerebral cortex.

Using within-subject comparisons, AR-ir values in chromogenic TSA− and TSA+ sections (Study 1) were positively correlated, suggesting that TSA increases staining intensity consistently across animals. This is an important consideration for group comparisons. The strength of this positive correlation appeared lower in the OFC than in the MeA. This is likely due to the difficulty of quantifying AR-ir in the OFC from TSA− sections, as faint staining led to more measurement variability. This was not a problem with the MeA from TSA− sections.

Phenotype of AR+ Cells in the PFC

In the mPFC and OFC, 100% of AR+ cells were NeuN+ and 0% were GFAP+, consistent with a neuronal phenotype. Identical results were obtained in the CA1 region of the hippocampus. These data are in agreement with several studies showing neuronal ARs in the rat brain, including the cerebral cortex, forebrain, BNST, MeA, and CA1 of the hippocampus.^37,46,50,51 Studies in primates also report similar findings⁵² suggesting that in the adult mammalian brain, AR expression is primarily in neurons.⁵³

However, ARs can also be expressed by glial cells in specific contexts.⁵³ In rats, this expression is age and region specific, such that AR+ glia have been observed in the cerebral cortex of postnatal day 10 pups, and in the arcuate nucleus of the hypothalamus and posterodorsal MeA of adults.^53,54 AR expression in microglia is observed after traumatic brain injury,¹⁸ at extranuclear sites in astrocytes in the hippocampus,⁵⁵ and in primary cultures of rat astrocytes.^56,57

Sex Differences in AR Expression

Studies of sex differences in neural AR expression have produced mixed results. Some studies have shown that males have higher levels of ARs than females in regions associated with reproductive behavior, such as the BNST, VMH, LS, and preoptic area.^34,58 However, previous evaluations of the cerebral cortices have not detected any sex differences in AR protein or mRNA.^2,37,46 Our detection of sex differences in AR staining in the PFC in intact subjects diverges from these results, although strain and age differences may be responsible. Our results are consistent with the idea that circulating T levels upregulate AR levels in the PFC, as is the case in some other brain regions.³⁴

Potential Pitfalls and Limitations of TSA

TSA can increase background staining.⁵⁹ Therefore, it is essential to optimize the primary antibody concentration and blocking conditions. Here, 8% tryptone in PBS was effective in blocking nonspecific background staining and outperformed more common blocking agents, such as normal serum and skim milk. Moreover, 0.8% tryptone was added to other protein-containing solutions, including the secondary antibody and AB solutions. Tryptone consists of partially digested peptides from casein and provides strong blocking against background resulting from TSA-based immunohistochemistry.¹¹ Here, the combination of tryptone blocking and a highly specific monoclonal antibody against ARs permitted the use of TSA with minimal background staining.

Another potential concern with TSA is that the amplified signal is not specific. In this case, for several reasons, it is unlikely that TSA generated nonspecific staining. Known AR-negative regions such as the dentate gyrus of the hippocampus^2,37 did not contain AR-ir, with or without TSA treatment (Fig. 2). In addition, the laminar staining pattern of ARs in the PFC, observable only with TSA, is consistent with previous work.⁴⁶ Other studies using TSA with a wide range of antibodies also show that staining specificity is unaffected by amplification, under the appropriate conditions.¹⁷ Finally, a comparison of TSA+ AR staining in intact and castrated male rats revealed a large reduction in AR staining following castration (Fig. 5). Castration has been shown to reduce AR-ir to levels ranging from low to absent.^3,5,35 Consequently, the castrated subject serves as a “negative control” and provides further evidence that the observed AR staining is specific. Castration reduced AR-ir across regions of high AR densities per cell like the MeA, as well as in regions of low AR densities per cell like the mPFC, where AR staining was extremely faint.

Finally, it is important to note that the goal of this study was simply to compare AR signal with and without TSA, particularly in the rat PFC. We do not directly compare chromogenic and fluorescent immunohistochemistry, tissue fixation methods, or different rat strains. Using TSA, future studies can examine AR-ir in the PFC and further examine differences between the sexes, across ages, and across different fixation methods. We conclude that the addition of TSA to an AR immunostaining protocol dramatically increases staining intensity in the rat brain. This has particularly important implications for studying brain regions, such as the PFC, with low numbers of ARs per cell. TSA allows detection and quantification of signal that was previously not quantifiable, thereby permitting a better understanding of AR distribution and regulation in the cerebral cortex. These findings should prove useful for elucidating the effects of androgens on cognition and executive function.