Immunohistochemistry for the detection of neural and inflammatory cells in equine brain tissue

Tissue samples

Immunohistochemistry protocols were developed on diseased and normal horse tissues. Diseased horses consisted of clinically affected West Nile virus (WNV), both naturally and experimentally infected, and Sarcocystis neurona infections (Beckstead, 1994; Gutierrez et al., 1999; Porter et al., 2003; Seino et al., 2007). Neurologically, normal horses were obtained by owner surrender for humane euthanasia due to loss of use. Brain, spinal cord, lymph node, spleen, thymus, and liver were collected and archived from these animals under University of Florida (Gainesville, FL) Institutional Animal Care and Use Committee protocols #F077, #F093, #D163, and #4109. Tissues were fixed in 10% neutral buffered formalin and processed into paraffin-embedded blocks approximately one-week after fixation. Initial evaluation of antibody binding was tested on non-infected equine lymph node and spleen for lymphocytic targets, liver and thymus for tissue macrophage targets, and brain for microglia, astrocytes, and neurons.

Tissue processing

The invariable IHC procedures for all protocols included sectioning FFPE tissues at 5 μm and placing them on positively charged glass slides. The slides were soaked in xylene (Fischer Scientific, Pittsburg, PA, USA) three times for 5 min to remove paraffin. These sections were then rehydrated through a gradient of ethanol (Fischer Scientific) for 5 min in each concentration, 100%, 100%, 95%, and 70% ethanol, followed by de-ionized water. In order to reduce the volume of the reagents tested and liquid loss, tissues were encircled with a hydrophobic barrier pen (ImmEdge™ Pen, Ted Pella Inc., Redding, CA, USA).

Antigen unmasking

Three methods of heating slides for investigating heat induced epitope retrieval (HIER) effectiveness included using a pressure cooker, microwave, and a double boiler. Pressure cooking was performed at 125 °C for 30 s followed by 90 °C for 10 s (Matyjaszek et al., 2009; Grosche et al., 2012), or microwaving was performed for 10 min (Kumar & Rudbeck, 2009) using a commercial countertop GE 1000W oven. For double boiling, two tissue-slides were floated back-to-back in 25 ml of retrieval solution in a 50 ml plastic conical tube. Conical tubes were then placed in pre-warmed water of a 250 ml glass beaker on a hotplate. Water temperature was maintained at 90 °C. After approximately 5 min of warming the retrieval solution, HIER was timed for 10 min. Conical tubes were then removed from the double boiler and allowed to cool for 15 min at 27 °C. Tissues were rinsed in deionized water three times for 2 min.

Heat induced epitope retrieval buffers were tested primarily using the double boiler system. Regents included two commercial citrate buffers, Epitope Retrieval Solution pH 6 (Novacastra, Leica, Newcastle Upon Tyne, UK) and Target Retrieval Solution pH 6 (Dako, Glostrup, Denmark), and an ethylenediaminetetraacetic acid (EDTA) solution buffered at pH 9 (10 mM Tris Base, 1 mM EDTA solution, 0.05% Tween 20, and NaOH to titrate to pH 9). A 1 × concentration of each solution was freshly made by diluting stock solutions with deionized water. For proteolytic epitope retrieval, tissues were treated with 200 ug/ml proteinase K solution (Tris HCL 100 mM pH 8.2, Tween 20, and Proteinase K (Ambion, Foster City, CA, USA)) for 10 min at 37 °C.

Endogenous peroxidase blocking

Peroxidase neutralizing solutions that were tested included two prepared solutions of hydrogen peroxide (H₂O₂), 3% and 0.3% H₂O₂, and a ready-to-use commercial reagent, Peroxidase Block (Novolink™ Polymer Detection System; Leica, Wetzlar, Germany). Solutions containing 3% and 0.3% H₂O₂ were made fresh for each staining attempt by diluting 30% H₂O₂ (Fischer Scientific) in 1 × phosphate buffered saline (PBS) (10 × PBS, Fischer Scientific). Tissues were immersed in peroxidase blocking solution for 5 min followed by two, 5 min rinses in PBS.

Non-specific protein blocking

Non-specific blocking techniques included four commercial reagents and one lab prepared solution. Commercial reagents included 10% Normal Goat Serum (Invitrogen, Frederick, MD, USA), Protein Block (Novolink™ Polymer Detection Systems; Leica, Wetzlar, Germany), Novocastra™ IHC/ISH Super Blocking Solution (Leica), and Novocastra™ Liquid Serum, Normal Goat Serum Blocking Reagent (Leica, Wetzlar, Germany). Additionally, a 5% goat serum solution was prepared by diluting Immunopure® Goat Serum (ThermoFischer Scientific, Waltham, MA, USA) in 1 × PBS. Protein blocking reagents were applied for 20 min with the exception of Novolink™ Protein block for only 5 min. Blocking reagents were removed without rinsing before adding primary antibody.

Primary antibodies

Primary antibodies tested targeted cell populations of astrocytes, microglia, neurons, T lymphocytes, B lymphocytes, and macrophages (Table 1). All primary antibodies were diluted in one of two commercial diluents (IHC Diluent (Novacastra, Leica); Dako Antibody Diluent (Dako, Glostrup, Denmark)). Two-fold serial dilutions of each antibody were tested to determine an optimal staining range. If the signal was weak to absent or background staining was present, additional dilution tests were performed until optimal staining was achieved. Antibodies were applied in a 37 °C, humidified incubator for 60–120 min or overnight (∼16 hrs) at 4 °C in a humidified, covered dish. After primary antibody incubation, slides were washed three times for 5 min each in 1 × PBS before secondary antibody application.

Negative controls for each primary antibody consisted of either an isotype-matched negative primary control (MCA928, AbD Serotec, Kidlington, UK) for monoclonal antibodies or rabbit serum for polyclonal antibodies.

Detection systems

Three commercial, horseradish peroxidase (HRP)-linked conjugate detection kits were utilized by following manufacturer’s instructions. These kits included the Vectastain® ABC Kit–Mouse IgG (Vector Laboratories, Burlingame, CA, USA), the Vectastain® ABC Kit–Rabbit IgG, and the Novolink™ Polymer Detection System (Leica, Wetzlar, Germany).

Lastly, a substrate-chromogen (Vector NovaRED Peroxidase Substrate, Vector Laboratories) was applied for 10 min, rinsed with de-ionized water for 5 min, counterstained with Lab Vision™ Mayer’s Hematoxylin (ThermoFisher Scientific) for 1 min, and rinsed in running tap water for 2 min. All sections were then dehydrated through an increasing gradient of ethanol, 50%, 70%, 95%, and 100%, for 2 min each. Slides were cleared in xylene three times for 3 min each before coverslipping with Permount™ mounting medium (Fischer Scientific, Hampton, NH, USA).

Antigen unmasking

Of the reagents tested, low pH citrate based solutions resulted in superior staining (Table 2). Target Retrieval Solution pH 6 by Dako was sufficient for most antibodies; however, use of Epitope Retrieval Solution pH 6 by Leica resulted in positive staining when the Dako product failed to produce staining with the CD79αy⁺ antibody. Proteinase K solution was tested with macrophage antibodies and resulted in increased staining intensity of MAC387⁺ antibody.

Various issues were encountered with the method of applying heat. Utilizing, one antibody, CD79αy⁺, only the double boiling method resulted in consistent positive staining of lymph nodes. Pressure cooking and microwaving resulted in limited to no staining or uneven staining, respectively. In addition, tissue disruption was minimal with the double boiling method. One antibody, Iba-1⁺, did not require heat retrieval and any HIER resulted in non-specific background staining.

Endogenous peroxidase blocking

In FFPE tissues, 3% H₂O₂ solution was reliable for reducing background staining without disruption of the tissue. There was no difference between commercial and lab diluted reagents.

Non-specific protein blocking

Multiple combinations of IHC antibody diluents and non-specific protein blocking reagents were tested. No difference in staining intensity was noted between of either diluent used. Among protein blocking solutions, 10% Normal Goat Serum, Protein Block from the Novolink™ Polymer Detection System, and 5% Immunopure® Goat Serum in PBS solution were all equal in blocking non-specific staining.

Specificity and sensitivity of monoclonal and polyclonal antibodies

Several antibodies were tested for lymphocyte specific CD antigens including CD3⁺, CD4⁺, CD5⁺, CD8⁺, CD20⁺, CD21⁺, and CD79αy⁺ (Table 1). Of the two CD3⁺, pan-T cell markers examined, rabbit anti-human polyclonal antibody (A0452, Dako) appropriately stained lymph node cortex (Fig. 1A), periarterial lymphatic sheaths of spleen, and perivascular cuffs in WNV infected brain (Fig. 1B). Four CD4⁺ T-helper cell antibodies were tested. Only mouse, monoclonal anti-equine antibody (HB61A, VMRD, Pullman, WA, USA) at 1:25 dilution positively stained lymph node cortex; however, background staining was high when applied to brain tissues. Two CD8⁺ cytotoxic T cell markers were investigated, but neither marker had reactivity in FFPE tissue.

Antibodies against CD5⁺, CD20⁺, CD21⁺, CD79αy⁺, and IgG (H+L) (Table 1) were tested for identification of B cell populations. A putative lymphocyte marker, CD5⁺ (B29A, VMRD), reportedly selects for B cells in equine tissues. Although this antibody intensely stained the germinal centers of FFPE lymph nodes, cortical staining was also noted. Because of this, CD5⁺ (B29A) was unreliable for distinction between B cell and T cell populations. Additionally in brain tissue, this antibody resulted in non-specific background staining, which could not be resolved. No staining was achieved with CD20⁺, CD21⁺, or any IgG (H+L) antibodies. Anti-human CD79αcy⁺ (HM57; Dako, Glostrup, Denmark) monoclonal antibody at 1:100 successfully stained lymph node germinal centers with no background staining in equine brain (Figs. 1C and 1D).

Multiple macrophage-targeting antibodies were investigated. RAM 11 (Dako, Glostrup, Denmark), AM-3K (TransGenic Inc., Kobe, Japan), and CD68⁺ (KP1; Leica, Wetzlar, Germany) antibodies had no reactivity with control tissues; however, MAC387⁺ (Leica, Wetzlar, Germany) was reactive. This marker positively stained control hepatic and thymic macrophages (Fig. 2A) with limited staining noted in lymph node sections. Based on cell morphology, polymorphonuclear and mononuclear cells also stained positively due to lack macrophage-specificity of the antibody. No reactivity was noted in normal equine brain. The MAC387⁺ cell population was distinct from the distribution of CD3⁺ lymphocytes in brain sections with inflammation due to WNV infection (Fig. 2B). Additionally, this macrophage antibody had little to no cross reactivity with brain microglia since few MAC387⁺ cells were visualized within glial nodules of WNV⁺ brain.

To characterize reactive gliosis, antibodies against both microglial and astrocytic markers were tested. In non-infected brain tissue, Iba-1⁺ antibody stained scattered resident microglia with low intensity, yet microglial processes were visualized (Fig. 2C). Staining intensity was higher in WNV-infected brains (Fig. 2D). Astrocyte populations were identified with antibody against an intermediate filament, glial fibrillary acid protein (GFAP), specific to astrocytes in CNS tissues. This antibody reacted strongly with astrocytes in both normal and infected brains with distinct astrocytic processes notable. GFAP⁺ astrocytes were distributed within the brain parenchyma, at the glia limitans, along brain vasculature, and in areas of gliosis of WNV⁺ brains (Fig. 3).

Three antibodies that target neurofilament heavy-chain proteins (NF-H) of neuronal axons successfully stained equine brain tissue. Swine derived, NAP4 (EnCor Biotechnology Inc., Gainesville, FL, USA) antibody, was superior and stained neurons in both normal (Fig. 4) and infected brains. Mouse monoclonal Neu-N antibody (A60, Millipore, Billerica, MA, USA), which targets neuronal perikaryon, was also tested but without successful reactivity in equine tissue.

Successful manual IHC protocols for these antibodies in equine neural tissue were identified (Table 2). It should be noted that overnight, refrigeration with CD3⁺, CD79⁺, MAC387⁺, GFAP⁺, and Iba-1⁺ primary antibodies will successfully stain tissues and may provide flexibility in laboratory scheduling if single-day protocol completion is not possible.