Combining Multiplex Fluorescence In Situ Hybridization with Fluorescent Immunohistochemistry on Fresh Frozen or Fixed Mouse Brain Sections

Summary

This protocol describes a method for combining fluorescence in situ hybridization (FISH) and fluorescence immunohistochemistry (IHC) in both fresh frozen and fixed mouse brain sections, with the goal of achieving multilabel FISH and fluorescence IHC signal. IHC targeted cytoplasmic and membrane attached proteins.

Abstract

Fluorescent in situ hybridization (FISH) is a molecular technique that identifies the presence and spatial distribution of specific RNA transcripts within cells. Neurochemical phenotyping of functionally identified neurons usually requires concurrent labelling with multiple antibodies (targeting protein) using immunohistochemistry (IHC) and optimization of in situ hybridization (targeting RNA), in tandem. A "neurochemical signature" to characterize particular neurons may be achieved however complicating factors include the need to verify FISH and IHC targets before combining the methods, and the limited number of RNAs and proteins that may be targeted simultaneously within the same tissue section.

Here we describe a protocol, using both fresh frozen and fixed mouse brain preparations, which detects multiple mRNAs and proteins in the same brain section using RNAscope FISH followed by fluorescence immunostaining, respectively. We use the combined method to describe the expression pattern of low abundance mRNAs (e.g., galanin receptor 1) and high abundance mRNAs (e.g., glycine transporter 2), in immunohistochemically identified brainstem nuclei.

Key considerations for protein labelling downstream of the FISH assay extend beyond tissue preparation and optimization of FISH probe labelling. For example, we found that antibody binding and labelling specificity can be detrimentally affected by the protease step within the FISH probe assay. Proteases catalyze hydrolytic cleavage of peptide bonds, facilitating FISH probe entry into cells, however they may also digest the protein targeted by the subsequent IHC assay, producing off target binding. The subcellular location of the targeted protein is another factor contributing to IHC success following FISH probe assay. We observed IHC specificity to be retained when the targeted protein is membrane bound, whereas IHC targeting cytoplasmic protein required extensive troubleshooting. Finally, we found handling of slide-mounted fixed frozen tissue more challenging than fresh frozen tissue, however IHC quality was overall better with fixed frozen tissue, when combined with RNAscope.

Introduction

Proteins and mRNAs that neurochemically define subpopulations of neurons are commonly identified with a combination of immunohistochemistry (IHC) and/or in situ hybridization (ISH), respectively. Combining ISH with IHC techniques facilitates the characterization of colocalization patterns unique to functional neurons (neurochemical coding) by maximizing multiplex labelling capacity.

Fluorescent ISH (FISH) methods, including RNAscope, have higher sensitivity and specificity compared to earlier RNA detection methods such as radioactive ISH and non-radioactive chromogenic ISH. FISH enables visualization of single mRNA transcripts as punctate stained spots¹. Furthermore, the RNAscope assay allows an increased number of RNA targets to be labeled at a time, using different fluorophore tags. Despite these advantages, technical limitations may affect the number of fluorophores/chromogens that can be used in a single experiment. These include availability of microscope filter sets; such considerations are compounded when neurochemical identification uses combined FISH and IHC, compared to using each technique in isolation, since inherent steps optimal for one method may be detrimental to the other.

Previous application of FISH combined with IHC has demonstrated the expression of specific cellular targets in human B-cell lymphomas², chick embryos³, zebrafish embryos⁴, mouse retina⁵ and mouse inner ear cells⁶. In these studies, tissue preparation was either formalin-fixed paraffin embedded (FFPE)^2,3,5 or fresh whole mount^4,6. Other studies applied chromogenic RNAscope on fixed mouse and rat brain preparations^7,8,9. In particular, Baleriola et al.⁸ described two different tissue preparations for combined ISH-IHC; fixed mouse brain sections and FFPE human brain sections. In a recent publication, we combined FISH and fluorescent IHC on fresh frozen sections, to simultaneously visualize low abundance mRNA (galanin receptor 1, GalR1), high abundance mRNA (glycine transporter 2, GlyT2) and vesicular acetylcholine transporter (vAChT) protein¹⁰ in the brainstem reticular formation.

The nucleus of the solitary tract (NTS) is a major brain region involved in autonomic function. Located in the hindbrain, this heterogeneous population of neurons receives and integrates a vast number of autonomic signals, including those that regulate breathing. The NTS harbors several neuronal populations, which may be phenotypically characterized by the expression pattern of mRNA targets including GalR1 and GlyT2 and protein markers for the enzyme tyrosine hydroxylase (TH) and the transcription factor Paired-like homeobox 2b (Phox2b).

The RNAscope proprietor recommends fresh frozen tissue preparations, but tissue prepared by whole animal transcardial perfusion fixation, along with long term cryoprotection (storage at -20 °C) of fixed frozen tissue sections, is common in many laboratories. Hence, we sought to establish protocols for FISH in combination with IHC using fresh frozen and fixed frozen tissue preparations. Here, we provide for fresh frozen and fixed frozen brain sections: (1) a protocol for combined FISH and fluorescent IHC (2) a description of the quality of mRNA and protein labelling produced, when utilizing each preparation (3) a description of the expression of GalR1 and GlyT2 in the NTS.

Our study revealed that, when combined with RNAscope methodology, IHC success varied in fresh frozen and fixed frozen preparations and, was dependent upon localization of the target proteins within the cell. In our hands, membrane bound protein labelling was always successful. In contrast, IHC for cytoplasmic protein required troubleshooting even in cases where the cytoplasmic protein was overexpressed in a transgenic animal (Phox2b-GFP)¹¹. Finally, while GalR1 is expressed in non-catecholaminergic neurons in the NTS, GlyT2 expression is absent in the NTS.

Protocol

A summary of tissue pre-processing steps may be found in Figure 1. All procedures were carried out in compliance with the Animal Care and Ethics Committee of the University of New South Wales in accordance with the guidelines for the use and care of animals for scientific purposes (Australian National Health and Medical Research Council).

1. Sample preparation of fresh frozen brain tissue

1. Transcardial Perfusion
   1. Prepare heparinized (2500 U/L) 0.1 M phosphate buffer (PB), pH 7.5. Make dry ice ethanol slurry by mixing dry ice with ethanol. This will have a temperature of approximately −72 °C and will be used for immediate freezing of the harvested tissue.
   2. Euthanize adult C57BL/6 and Phox2b-GFP¹¹ (Mouse Genome Informatics database ID MGI:5776545) mice by anesthetizing with sodium pentobarbital (70 mg/kg, i.p.), using a 27.5 inch needle gauge.
      CAUTION: Pentobarbital is a barbiturate. It is acutely toxic in high doses and may cause death by respiratory arrest. Consult local healthcare, legal and material safety guidelines before use.
   3. Expose the heart and cannulate the left ventricle with a drawing-up needle (23 inch gauge). Perform transcardial perfusion with heparinized 0.1 M PB until the blood clears (2-3 minutes) at a flow rate of 11-13 mL/min. Determine blood clearing by monitoring the coloration of the liver and the effusate from the right atrium¹².
   4. Isolate the brain from the skull cavity, immediately embed it in Optimal Cutting Temperature Compound (OCT) in a cryomold or aluminum foil and place it on the dry ice ethanol bath. Store the frozen embedded tissue in an airtight container at - 80 °C for up to 3 months.
2. Sectioning of fresh frozen tissue
   1. Set the cryostat temperature to -20 °C. Leave the OCT-embedded tissue and a cryostat chuck in the cryostat for ~30 minutes to allow for equilibration to the new temperature.
      NOTE: Keep the tissue frozen at all times; transport the tissue from the -80 °C freezer to the cryostat on dry ice.
   2. Secure the tissue to the pre-chilled cryostat chuck using OCT compound. In this protocol, tissue blocks were mounted onto the chuck in the coronal plane.
      NOTE: Trim excess OCT from the tissue, using a razor blade, to minimize the amount of OCT being cut by the cryostat and subsequently transferred onto the glass slide.
   3. Cut 14 µm thick coronal sections and mount them onto charged glass microscopy slides.
      1. Warm the slides to room temperature before mounting the sections. Once the section has been mounted, keep the slides in a slide box in the cryostat.
      2. If more than one section needs to be mounted on one slide, warm the area for the second section by placing a finger on the opposite side of the slide for 5-10 seconds to aid adherence of the section to the slide. A cold tissue section will not attach to a cold slide. The sections should adhere to the slides flat; folding will cause them to fall off the slides during wash steps.
      3. If cracks are noticed in the sections, increase the cryostat temperature by 1-5 °C to avoid this. It is particularly important to place tissue sections in close proximity to one another on the same slide. This will prevent wastage of FISH probes and reagents during the assay.
   4. Store tissue sections mounted onto glass slides in an air-tight container at -80 °C for up to 6 months.
      NOTE: Keep the sections frozen at all times and avoid freeze thaw cycles, to prevent RNA degradation. Transport the slide box from inside the cryostat to the -80 °C freezer on dry ice.
3. Fixation of fresh frozen tissue
   1. On the day the FISH probe assay is to be performed, prepare 4% paraformaldehyde (PFA) in 0.1 M PB, pH 7.5 (4% PFA solution). Filter by passing through filter paper (Grade 1: 11 µm, Table of Materials) in a Buchner funnel or crucible filter.
      CAUTION : PFA is harmful and toxic by skin contact or inhalation. All procedures with PFA solution should be performed in a fume hood cabinet. PFA solution waste should be disposed of carefully following institutional safety protocols.
   2. Cool the 4% PFA solution to 4 °C. Transport the slide-mounted tissue from the -80 °C freezer in dry ice and immediately immerse it in the pre-chilled fixative for 15 minutes.
      NOTE: It is important that this fixation step does not exceed 15 minutes as over-fixation will result in non-specific background labelling.
4. Dehydration of fresh frozen tissue
   1. Dehydrate tissue sections by submerging the slides in graded concentrations of ethanol. In a Coplin jar, first submerge in 50%, then 70% and finally absolute ethanol, for 5 minutes each at room temperature. Repeat the final absolute ethanol incubation a second time.
   2. Air dry slides and, outline the group of sections using a hydrophobic barrier pen, ensuring that the internal area is kept to a minimum.
      NOTE: Make sure that the glass slide is completely dry before drawing the hydrophobic barrier. The hydrophobic barrier should surround the tissue sections completely without gaps and must be dry before further processing.

2. Sample preparation of fixed frozen brain tissue

1. Transcardial perfusion fixation
   1. Euthanize mice by anesthetizing with sodium pentobarbital (70 mg/kg, i.p) followed by transcardial perfusion, first with 0.1 M PB then 4% PFA solution. Fix with 10 minutes of perfusion at 11-13 mL/min.
   2. Isolate the brain from the skull cavity following perfusion-fixation and submerge overnight in 4% PFA solution, at 4 °C.
2. Tissue sectioning of fixed tissue
   1. Rinse the brain in sterile 0.1 M phosphate buffered saline (PBS) before removing the meningeal layers, with the aid of a dissecting microscope, using fine forceps.
   2. Cut the brain precisely into blocks (separate the brainstem from the forebrain prior to vibratome sectioning) using a brain matrix (Table of Materials). Specifically, cut the brainstem caudally at the pyramidal decussation and dissect away the cerebellum. Similarly, cut the forebrain immediately rostral to the optic chiasm.
   3. Secure the tissue onto a vibrating microtome chuck using cyanoacrylate and embed in 2% agar solution.
   4. Cut 30 µm thick tissue sections using a vibrating microtome and store cut sections in cryoprotectant solution (30% RNase free sucrose, 30% ethylene glycol, 1% polyvinylpyrrolidone (PVP-40), in 0.1 M PB, pH 7.4). Tissue sections may be stored in cryoprotectant at -20 °C for up to 6 months.
3. Preparation of fixed sections prior to FISH
   1. On the day of FISH, wash free floating sections three times, for 10 minutes per wash, to remove the cryoprotectant solution. To wash, place sections in 0.1 M PBS in a 12 well cell culture plate and agitate on a rotating platform shaker (90 - 100 rpm).
   2. After washes, use a paintbrush to mount sections on glass microscopy slides and air dry for at least 2 hours.
      NOTE: The sections should adhere flat onto the slides as any pronounced folds will cause them to detach during washes.
   3. Using a hydrophobic barrier pen, draw a barrier around the sections to restrict the FISH reagents to the sections. Once more, it is important to minimize the internal area of the outline drawn with the barrier pen.
      ​POSSIBLE BREAK POINT: The sections could be stored at room temperature, overnight, to continue the assay the next day.

3. FISH assay

NOTE: The rest of the protocol applies to both fresh frozen and fixed frozen tissue.

1. Prepare the reagents and instruments for hybridization and amplification steps.
   1. Set a benchtop incubator and water bath to 40 °C.
   2. Prepare a humidified, light-protected chamber for incubating slides. Humidification prevents drying of tissues - slides are securely located above a moist reservoir. Ideally, the chamber is made of heavy-duty polystyrene, it is light-proof and airtight to maintain a saturated water vapor atmosphere. Closure of the chamber relies on minimal friction to avoid movement. We used a slide box lined with wet laboratory wipes (Table of Materials) at the bottom. Place the slide box inside the incubator to prewarm it to 40 °C.
   3. Warm the 50x Wash Buffer (Table of Materials) and probes to 40 °C for 10 minutes, using the water bath, then cool to room temperature.
   4. Prepare 1 L of 1x Wash Buffer from the 50x stock concentration.
   5. Prepare probe mixture (Table of Materials): the C1 probe is ready to use at stock concentration whereas C2 and C3 probes are shipped as 50x concentration and require dilution with the diluent supplied in the kit.
      NOTE: Probe mixtures can be stored at 4 °C for up to 6 months.
2. Protease treatment
   1. Incubate sections with Protease III (Table of Materials) at room temperature for 30 minutes.
      NOTE: Ensure that Protease III and incubation reagents in downstream processes (probe mixture, amplification solutions, blocking buffer and antibody sera) cover the sections entirely. A pipet tip may be used to spread the reagent onto the section to cover the entire area inside the hydrophobic barrier.
   2. Wash slides twice with 0.1 M PBS, for 2 min each time, in a large plastic square Petri dish. Here a 245 mm x 245 mm square bioassay dish was used (Table of Materials). Hold from one side of the dish and tilt gently 3-5 times. After washes, flick excess 0.1 M PBS from slide and immediately add the next reagent. Do not let tissue sections dry.
      NOTE: During each wash, the slides are immersed in solution at room temperature. This is the workflow for all subsequent wash steps. The fixed 30 µm thick sections dislodge from slides more easily than 14 µm thick sections, be gentle during the washes.
3. Hybridization and amplification
   1. After washing off the protease solution, place the slides in the humidified, prewarmed chamber. Incubate sections with probe mixture (Table of Materials) for 2 hours at 40 °C inside a benchtop incubator.
      ​NOTE: Ensure there are at least 2 sections set aside for positive and negative control probes to assess sample RNA quality and optimal permeabilization. Positive control probes target house-keeping genes; here, these were a cocktail of RNAs targeting ubiquitin C (UBC; high-abundance), peptidylpropyl isomerase B (PPIB; moderate-abundance) and RNA polymerase 2a (POLR2A; low-abundance). Negative control probes target the bacterial 4-hydroxy-tetrahydrodipicolinate reductase (DapB) gene, which is normally absent in mouse brain samples. Positive DapB signal indicates non-specific signal and/or bacterial contamination of the sample.
   2. Following hybridization with the probe mixture, the signal amplification steps consist of incubation with Amp 1-FL (30 minutes), then with Amp 2-FL (15 minutes), followed by Amp 3-FL (30 minutes) and finally Amp 4-FL (15 minutes) - each at 40 °C. Using the dropper bottles provided, cover tissue sections with amplification solution. Proceed to IHC assay following the last amplification step.
   3. Rinse slides with Wash Buffer twice for 2 minutes between probe hybridization and each amplification step.

4. IHC Assay

1. IHC blocking step
   1. To prevent non-specific binding of antibodies, incubate the sections for 1 h at room temperature with blocking solution containing 10% normal horse serum, 0.3% Tween20 in 1x TBSm (50 mM Tris-Cl, pH 7.5, 150 mM NaCl, 0.05% merthiolate) following the FISH assay. Prepare primary antibodies in a dilution buffer containing 1x TBSm, 5% normal horse serum and 0.1% Tween20. Primary antibody suppliers are listed in the Table of Materials.
2. Immunohistochemistry
   1. Remove excess blocking buffer by flicking the slide and incubate sections with primary antibodies overnight at 4 °C.
   2. Wash slides 3 times (5 minutes each) with 1x TBSm and incubate with secondary antibody in diluent containing 1x TBSm, 1% normal horse serum and 0.1% Tween20 for 2 hours at room temperature. Secondary antibodies used in this protocol are listed in the Table of Materials.
   3. Wash slides 3 times with 1x TBSm (5 minutes each) before coverslipping with mounting medium with or without DAPI (Table of Materials).

5. Imaging

1. Examine immunostaining under an epifluorescence microscope equipped with a camera (see Table of Material for details). Acquire representative images at 20x magnification and save as TIFF files.
2. Export representative images into an image processing software (Table of Materials) for brightness/contrast adjustment to increase the clarity and to reflect true rendering.

6. OPTIONAL: Quantitative analysis of target transcripts

NOTE: This is a methods article and quantitative results are not provided. The method of quantification presented here is sourced from Dereli et al.¹⁰.

1. Acquire images from the regions of interest as explained in 5.1 and apply the same microscope and camera settings (such as exposure time and light intensity) to all images of the same fluorophore.
2. Plot the neuronal profiles using an image analysis software (Table of Materials).
3. Align the sections with reference to Bregma level according to a stereotaxic brain atlas¹³.
4. Apply the same brightness and contrast to all the images of the same fluorophore. Only consider the neurons with DAPI-stained nuclei.
5. Manually count the number of mRNA, protein expressing, mRNA/mRNA, protein/protein and mRNA/protein coexpressing cells within the region of interest.
6. To decrease bias in the experimental results, have the person quantifying experimental outcomes blinded to the experimental groups.
7. Apply Abercrombie correction¹⁴ to total cell counts by using the following Abercombie equation:
   Corrected cell count = manual cell count x section thickness / (section thickness + nuclear size)
   For example, for 14 µm thick sections, the average nuclear width is calculated to be 7.7 ± 0.3 µm and average section thickness is 14 ± 1 µm based on 30 cells and 10 sections respectively in 5 animals¹⁰. According to the Abercrombie equation, corrected cell count would be manual cell count x 14/(14+7.7).

Figure 1: Parallel workflow of tissue pre-processing steps for both fresh-frozen and paraformaldehyde fixed tissue. Processing steps for fresh-frozen tissue are displayed in the red outlined boxes, whereas those for paraformaldehyde (PFA) fixed tissue are displayed in the blue outlined boxes. Please click here to view a larger version of this figure.

Figure 2: Summary of combined FISH probe and immunohistochemistry procedure. Following tissue pre-processing, the slide mounted tissue is encircled using a hydrophobic barrier pen, as seen in the first frame, and incubated in a protease solution at room temperature. Following washes, tissue is transferred to a benchtop incubator for hybridization for 2 hours before sequential amplification steps. The in situ hybridization system utilizes a proprietary 'Z probe' design, preamplifiers and amplifiers as seen in frames 3-6⁶. Once tissue has undergone FISH probe processing, it is washed before blocking with normal horse serum. The primary antibody incubation is carried out overnight at 4 °C to maximize antibody-antigen binding. The secondary antibody incubation (2 hours) was carried out at room temperature. Please click here to view a larger version of this figure.

Results

Here, we outline a method for combining multiplex FISH with fluorescent IHC to localize mRNA expression for GalR1 and GlyT2 using fresh-frozen and paraformaldehyde fixed tissues respectively in the mouse NTS. A pipeline of the tissue processing, FISH and IHC procedures described in the methods is displayed in Figure 1 and Figure 2. Table 1 provides a summary of the FISH probe and antibody combinations used in each figure.

Control probes are routinely assayed concurrently with target probe, to ensure integrity of the workflow and confirm sample quality. The absence of DapB labelling confirms sound tissue quality and integrity, and the absence of bacterial contamination (Figure 3A). Labelling from positive control probes targeting ubiquitin C (UBC, high abundance), peptidylpropyl isomerase B (PPIB, moderate abundance) and RNA polymerase 2a (POLR2A, low abundance) mRNA confirms RNA integrity and the signal observed between assays may be used to calibrate inter assay variability (Figure 3B). To validate FISH probe expression, we used control tissues that have been previously described to express the mRNA transcript. For example, GalR1 mRNA expression, was confirmed to be positive in the thalamus as previously described^10,15. Phox2b mRNA distribution was additionally verified by colabelling with Phox2b antibody; we confirmed that FISH labelling was present only in neurons that were also positively stained using the Phox2b antibody (Figure 5).

To distinguish GalR1+ neurons in the NTS from neighbouring nuclei, we used additional neurochemical markers. TH, Phox2b or Phox2b-GFP immunoreactivity (Figure 4-6), and Phox2b FISH (Figure 5 and Figure 6) differentiated the NTS from other nuclei in the dorsal brainstem as NTS neurons have previously been reported to express Phox2b and TH^16,17. Since the NTS is nestled by cholinergic nuclei - it lies dorsal to the hypoglossal nucleus and dorsal motor nucleus of the vagus (DMNX), and ventral to the vestibular nucleus - we co-labelled with the cholinergic marker vAChT¹⁸ (Figure 4). Therefore, the expression of GalR1 within the NTS was assessed in relation to TH and Phox2b, whilst vAChT labelling aided spatial orientation with respect to rostrocaudal, dorsoventral and mediolateral coordinates. We found all TH immunoreactive and GalR1 mRNA positive neurons in the NTS were Phox2b-GFP immunoreactive, but not all Phox2b-GFP immunoreactive neurons in the NTS were TH immunoreactive or GalR1 mRNA positive (Figure 4). Also, we demonstrated that mRNA for the low abundance receptor GalR1 was absent in TH and vAChT immunoreactive neurons.

In fresh frozen preparations, when combined with the FISH probe assay, IHC success was dependent on subcellular location of the target protein. For example, vAChT (a synaptic vesicle membrane-bound protein) was clearly immunolabelled, whereas TH and GFP (cytoplasmic proteins) were indefinitely immunolabelled and only faintly observed (Figure 4). We describe this indefinite labelling as 'flocculent' because cells lacked a clear outline and proved difficult to distinguish from the background. On the same fresh frozen tissue section, GalR1 FISH probe labelling of cytoplasmic GalR1 mRNA was punctate and clearly observed (Figure 4).

Furthermore, since the TH and vAChT antibodies are raised in the same host, both proteins were labelled using the same secondary antibody and therefore the same color fluorophore (excitation light: 594). They are easily distinguished for two reasons: they never co-label in the same neurons, and the subcellular localisation is different for these proteins; vAChT in vesicles exhibiting a punctate appearance, and TH in the cytoplasm and neuronal processes.

To support our hypothesis that IHC quality (in fresh frozen preparations) is dependent on protein subcellular localisation, we compared labelling for Phox2b mRNA (located in the cytoplasm), GFP (over-expressed in cytoplasm) and Phox2b protein (primarily found in the nucleus) in neurons. As expected, our results show overlap of Phox2b mRNA, GFP and Phox2b antibody labelling in individual neurons of the NTS (Figure 5). Cells with cytoplasmic mRNA labelling corresponded with cells exhibiting nuclear labelling of the Phox2b protein providing validation of the combined FISH-IHC method. Although cytoplasmic Phox2b-GFP had a flocculent appearance, nuclear Phox2b protein signal was clear and specific. In conclusion, when combined with FISH on fresh frozen preparations, membrane-bound proteins including vAChT and Phox2B exhibit higher quality immunolabelling than cytoplasmic proteins.

In contrast, IHC was reliable irrespective of subcellular localization, when performed on fixed frozen sections in combination with FISH. Multiplex FISH for GlyT2 mRNA and Phox2b mRNA was successful, as shown in Figure 6. GlyT2 mRNA positive neurons were located ventral to the NTS and not within the NTS. GlyT2+ and Phox2b+ neurons did not colocalize. A subpopulation of Phox2b+ NTS neurons was TH immunoreactive and none contained GlyT2 mRNA. TH immunoreactive neurons are apparent on the same tissue section, exhibiting positively labelled soma and neuronal processes (Figure 6). This contrasts with the 'flocculent' appearance of TH immunoreactive neurons in fresh frozen tissue sections. Thus, the fixed frozen preparation described here is an alternative method of tissue preparation which enables reliable targeting of cytoplasmic proteins immunohistochemically, in combination with RNAscope.

Figure 3: Representative microscopic images from coronal mouse forebrain sections at the level of the lateral septum (Bregma 1.1 to -0.1) showing labelling of positive and negative control probes. (A) A lack of signal following ISH with bacterial 4-hydroxy-tetrahydrodipicolinate reductase (DapB) confirms the absence of background signals. (B) Labelling with positive control probes targeting ubiquitin C (UBC), peptidylpropyl isomerase B (PPIB) and RNA polymerase 2a (POLR2A) illustrates the signal to be expected from high, moderate and low abundance targets respectively. Scale bars are 50 µm. All images were acquired with 20x objective. Please click here to view a larger version of this figure.

Figure 4: Representative microscopic images of a fresh frozen coronal brainstem section from a Phox2b-GFP mouse showing combined labelling of GalR1 mRNA (FISH) and 3 proteins (IHC) in the nucleus of the solitary tract (NTS) region. Insets in A are enlarged in B. GalR1 mRNA is indicated by punctate FISH probe labelling (arrowheads). Antibodies targeting the cytoplasmic proteins GFP and tyrosine hydroxylase (TH) exhibited "flocculent" labelling (arrows). Vesicular acetylcholine transporter (vAChT) immunoreactivity is demonstrated (red punctate labelling) in the hypoglossal nucleus (XII). Scale bars are 100 µm in A and 25 µm in B. All images were acquired with 20x objective. Other abbreviations: area postrema (AP), medial vestibular nucleus (MVe). Please click here to view a larger version of this figure.

Figure 5: Representative microscopic images of a fresh frozen coronal brainstem section from a Phox2b-GFP mouse, illustrating targeting of Phox2b in the nucleus of the solitary tract (NTS) with three different approaches: Phox2b mRNA (FISH), GFP (IHC) and Phox2b protein (IHC). Phox2b protein is localized to the nucleus. Insets in A are enlarged in B. Arrows indicate neurons that are triple labelled with Phox2b probe (orange-550), GFP antibody (green-488) and Phox2b antibody (red-647). Scale bars are 100 µm in A and 25 µm in B. All images are acquired with a 20x objective. Other abbreviations: area postrema (AP). Please click here to view a larger version of this figure.

Figure 6: Representative images from fixed frozen coronal brainstem sections demonstrating successful FISH combined with reliable immunolabelling of cytoplasmic proteins (tyrosine hydroxylase [TH]). Double FISH showing glycine transporter 2 (GlyT2-red-647, filled arrowheads) and Phox2b (yellow-550, arrows) mRNA labelling in nucleus of the solitary tract (NTS) region. FISH was combined with IHC for TH protein (blue-346, empty arrowheads). Insets in A are enlarged in B. Scale bars are 25 µm. All images were acquired with a 20x objective. Please click here to view a larger version of this figure.

		Primary antibody or RNAscope probe	Secondary antibody or Amp 4-FL-Alt Display module	Excitation (nm)	Tissue preparation
Figure 3	probe	POLR2A (C1)	Amp 4-FL-Alt B Display module	647	fresh frozen
	probe	PPIB (C2)	Amp 4-FL-Alt B Display module	488
	probe	UBC (C3)	Amp 4-FL-Alt B Display module	550
	probe	DapB (C1, C2, C3)	Amp 4-FL-Alt B Display module	647, 488, 550
	DAPI			346
Figure 4	antibody	rabbit-anti-GFP	donkey-anti-rabbit	488	fresh frozen
	antibody	sheep-anti-TH	donkey-anti-sheep	647
	antibody	goat-anti-vAChT	donkey-anti-goat	647
	probe	GalR1 (C1)	Amp 4-FL-Alt B Display module	550
	DAPI			346
Figure 5	antibody	rabbit-anti-GFP	donkey-anti-rabbit	488	fresh frozen
	antibody	mouse-anti-Phox2b	donkey-anti-mouse	647
	probe	Phox2b (C2)	Amp 4-FL-Alt A Display module	550
	DAPI			346
Figure 6	antibody	mouse-anti-TH	donkey-anti-mouse	346	fixed
	probe	GlyT2	Amp 4-FL-Alt A Display module	647
	probe	Phox2b	Amp 4-FL-Alt A Display module	550

Table 1: FISH probe, antibody and corresponding flurophore combinations used in Figures 3-6.

Discussion

In the neurosciences, FISH and IHC are routinely used to investigate the spatial organization and functional significance of mRNA or proteins within neuronal subpopulations. The protocol described in this study enhances the capacity for simultaneous detection of mRNAs and proteins in brain sections. Our combined multiplex FISH-IHC assay enabled phenotypic identification of distinct neuronal subpopulations in the NTS in both fresh frozen and fixed brain preparations. FISH-IHC in fixed frozen tissue preparations produced reliable IHC outcomes. For example, multiplex FISH for low and high abundance mRNAs (GalR1 and GlyT2 respectively) and IHC (targeting tyrosine hydroxylase) revealed that GalR1 and GlyT2 are expressed in non-catecholaminergic NTS neurons. IHC for TH was not successful in fresh frozen tissue, highlighting the limited capacity for FISH-IHC in fresh frozen preparations.

ISH may be more appropriate than IHC in a range of scenarios. First, IHC may not perform well when detecting low abundance proteins, such as receptors. Using ISH to target relatively higher abundance mRNAs for these proteins improves detectability¹. Second, proteins such as neuropeptides are often trafficked to the axonal terminals following translation in the cell soma¹⁹. When neuropeptides are targeted with IHC, the axonal processes and terminals of cells label with the antibody, but not the soma, reducing the capacity to identify the cell of origin or perform quantitative analysis of the number of cells. However, since mRNAs which code for all proteins are found localized to the soma, the ISH technique is advantageous. Finally, antibodies are not readily available for some protein species, or the available antibodies are appropriate for other proteomics techniques (e.g., western blot) but not IHC. In these circumstances, mRNA labelling methods prove useful. A caveat is that mRNAs may not always be translated into protein, and so they only provide a proxy for protein identification. Since commercial FISH kits can be costly and the ISH probes are less likely to be commercially available compared to antibodies, combining FISH with IHC presents a cost and time effective strategy for increasing the number of targets that can be labelled simultaneously.

Fresh frozen versus fixed tissue preparation was a factor conferring successful IHC following FISH probe assay. We tested IHC using antibodies targeting nuclear, vesicular and cytoplasmic proteins and found reliable labelling of membrane-bound proteins (vAChT and Phox2b) on fresh frozen samples but not cytoplasmic proteins (TH and GFP). Coexpression of Phox2b protein and mRNA with 'flocculent' Phox2b-GFP labelling validated that neurons expressing the transcript also expressed the related protein, confirming the neurochemical identity of the neurons (Figure 5). In contrast, fixed frozen tissue preparations yielded reliable IHC labelling regardless of subcellular localization of the antigen. Previous studies have demonstrated that protease (e.g., pronase^8,20) pretreatment can have a detrimental effect on IHC. The contents of the protease solution utilized in the RNAscope protocol is proprietary, and permeabilization by protease is recommended for RNAscope probe access into cells. Labelling of cytoplasmic proteins using the antibodies described here has been previously verified on free floating 30 µm fixed frozen mouse brain sections^10,21,22. We slide-mounted 30 µm thick fixed samples and performed the FISH-IHC protocol, as opposed to the 14 µm thick fresh frozen sections recommended by the manufacturer. In the absence of assay modifications, or alteration of other variables (antigen retrieval, higher antibody concentration, change of protease), reliable IHC was achieved on thick, fixed samples with demonstrated labelling of the cytoplasm and axonal processes together with FISH probe labelling (Figure 6). While similar approaches were employed by other research groups^7,8,9, the current study achieved combined ISH-IHC on neurons and in a fluorescent set-up.

There were a series of critical steps in the methods to take note of. For the fresh frozen preparation, fixation time should not exceed 15 minutes; longer fixation times elicited higher background labelling. The protease step was optimized since tissues of different thickness and from various organs require different types of protease to achieve permeabilization. Fixed frozen sections adhere less to glass slides and dislodge more easily during wash steps. Hence, extra care must be taken in manual handling of fixed frozen sections, to avoid tissue loss or damage.

Although we found combined FISH and IHC to be an effective strategy, the disadvantages include cost and technically demanding assay when combining the two methods. One limitation of the study is that a side-by-side comparison of the two tissue preparation protocols was not performed. Also, our evaluation of results was limited by the number of channels the epifluorescent microscope could accommodate; the set-up allowed a maximum of 4 channels at a given time: 346, 488, 550 and 647 nm (excitation light). We were able to achieve multiplex labelling of 5 targets by labelling two proteins with different subcellular localizations using the same flurophore (Figure 4, Table 1). By using a confocal microscope, discrete excitation of many additional fluorophores may be used for individual protein labelling via IHC, or for imaging of fluorescent molecules expressed by transgenes.

Combined FISH and fluorescent IHC can reduce the reliability of each technique in isolation. In the future, we aim to improve cytoplasmic protein labelling on fresh frozen tissue with an antigen retrieval treatment²³. Previous studies show that heat induced antigen retrieval increases accessibility of the protein epitope^24,25,26. Heat treatment cleaves the crosslinks and methylol groups of the protein and unfolds the antigens in tissues, exposing epitopes which would otherwise be hidden in the tertiary protein structure under biological conditions. This accessibility may improve the success of protein labelling^26,27. Additionally, we will target different epitopes of the same cytoplasmic protein to determine if the success of protein-antibody labelling depends on the specific antibody clones used.

In conclusion, combined FISH and IHC is useful for neurochemical identification of heterogenous populations of cells in the brain, such as those in the NTS. This study presents two protocols assaying different mouse brainstem tissue preparations - fresh frozen, or fixed - for simultaneous multiplex fluorescent labelling of mRNA and proteins in situ. Both protocols may be widely applied to detect the expression pattern of low abundance mRNAs, such as GalR1. Thick (30 µm) fixed frozen preparations permeabilized with protease conferred more reliable cytoplasmic protein detection and more tissue handling challenges, when compared to thin (14 µm) fresh frozen preparations.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
ANIMALS
C57BL/6 mouse	Australian BioResources, Moss Vale	MGI: 2159769
Phox2b-eGFP mouse	Australian BioResources, Moss Vale	MGI: 5776545
REAGENTS
Cyanoacrylate	Loctite
Ethylene Glycol	Sigma-Aldrich	324558
Heparin-Sodium	Clifford Hallam Healthcare	1070760	Consult local veterinary supplier or pharmacy.
Lethabarb (Sodium Pentabarbitol) Euthanasia Injection	Virbac (Australia) Pty Ltd	N/A	Consult a veterinarian for local pharmaceutical regulations regarding Sodium Pentabarbitol
Molecular grade agarose powder	Sigma Aldrich	5077
OCT Compound, 118mL	Scigen Ltd	4586
Paraformaldehyde, prilled, 95%	Sigma-Aldrich	441244-1KG
Polyvinylpyrrolidone, average mol wt 40,000  (PVP-40)	Sigma-Aldrich	PVP40
ProLong Gold Antifade Mountant	Invitrogen	P36930	With or without DAPI
RNAscope Multiplex Fluorescent Reagent Kit (up to 3-plex capability)	Advanced Cell Diagnostics, Inc. (ACD Bio)	ADV320850	Includes 50x Wash buffer and Protease III
RNase Away	Thermo-Fisher Scientific	7003
Tris(hydroxymethyl)aminomethane	Sigma-Aldrich	252859
Tween-20, for molecular biology	Sigma-Aldrich	P9416
EQUIPMENT
Benchtop incubator	Thermoline scientific micro incubator	Model: TEI-13G
Brain Matrix, Mouse, 30g Adult, Coronal, 1mm	Ted Pella	15050
Cryostat	Leica	CM1950
Drawing-up needle (23 inch gauge)	BD	0288U07
Hydrophobic Barrier Pen	Vector labs	H-4000
Kimtech Science Kimwipes Delicate Task Wipes	Kimberley Clark Professional	34120
Olympus BX51	Olympus	BX-51
Peristaltic pump	Coleparmer Masterflex	L/S Series
Retiga 2000R Digital Camera	QImaging	RET-2000R-F-CLR	colour camera
SuperFrost Plus Glass Slides (White)	Thermo-Fisher Scientific	4951PLUS4
Vibrating Microtome (Vibratome)	Leica	VT1200S
Whatman qualitative filter paper, Grade 1, 110 mm diameter	Merck	WHA1001110
SOFTWARES
CorelDRAW	Corel Corporation	Version 7
FIJI (ImageJ Distribution)	Open Source/GNU General Public Licence (GPL)	N/A	ImageJ 2.x: Rueden, C. T.; Schindelin, J. & Hiner, M. C. et al. (2017), "ImageJ2: ImageJ for the next generation of scientific image data", BMC Bioinformatics 18:529, PMID 29187165, doi:10.1186/s12859-017-1934-z   and Fiji: Schindelin, J.; Arganda-Carreras, I. & Frise, E. et al. (2012), "Fiji: an open-source platform for biological-image analysis", Nature methods 9(7): 676-682, PMID 22743772, doi:10.1038/nmeth.2019
PRIMARY ANTIBODIES
Anti-Tyrosine Hydroxylase Antibody	Millipore Sigma	AB1542	Sheep polyclonal (1:1000 dilution), RRID: AB_90755
Anti-Tyrosine Hydroxylase Antibody, clone LNC1	Millipore Sigma	MAB318	Mouse monoclonal (1:1000 dilution), RRID: AB_2201528
Anti-Vesicular Acetylcholine Transporter (VAchT) Antibody	Sigma-Aldrich	ABN100	Goat polyclonal (1:1000 dilution), RRID: AB_2630394
GFP Antibody	Novus Biologicals	NB600-308	Rabbit polyclonal (1:1000 dilution), RRID: AB_10003058
Phox2b Antibody (B-11)	Santa Cruz Biotechnology	sc-376997	Mouse monoclonal (1:1000 dilution), RRID: AB_2813765
SECONDARY ANTIBODIES
Alexa Fluor 488 AffiniPure Donkey Anti-Rabbit IgG (H+L) (min X Bov, Ck, Gt, GP, Sy Hms, Hrs, Hu, Ms, Rat, Shp Sr Prot)	Jackson ImmunoResearch	711-545-152	Donkey anti-Rabbit (1:400 dilution), RRID: AB_2313584
AMCA AffiniPure Donkey Anti-Sheep IgG (H+L) (min X Ck, GP, Sy Hms, Hrs, Hu, Ms, Rb, Rat Sr Prot)	Jackson ImmunoResearch	713-155-147	Donkey anti-Sheep (1:400 dilution), RRID: AB_AB_2340725
Cy5 AffiniPure Donkey Anti-Goat IgG (H+L) (min X Ck, GP, Sy Hms, Hrs, Hu, Ms, Rb, Rat Sr Prot)	Jackson ImmunoResearch	705-175-147	Donkey anti-Goat (1:400 dilution), RRID: AB_2340415
Cy5 AffiniPure Donkey Anti-Mouse IgG (H+L) (min X Bov, Ck, Gt, GP, Sy Hms, Hrs, Hu, Rb, Rat, Shp Sr Prot)	Jackson ImmunoResearch	715-175-151	Donkey anti-Mouse (1:400 dilution), RRID: AB_2619678
Cy5 AffiniPure Donkey Anti-Sheep IgG (H+L) (min X Ck, GP, Sy Hms, Hrs, Hu, Ms, Rb, Rat Sr Prot)	Jackson ImmunoResearch	713-175-147	Donkey anti-Sheep (1:400 dilution), RRID: AB_2340730
RNASCOPE PROBES
Galanin Receptor 1 oligonucleotide probe	ACDBio	448821-C1	targets bp 482 - 1669 (Genebank ref: NM_008082.2)
Glycine transporter 2 oligonucleotide probe	ACDBio	409741-C3	targets bp 925 - 2153 (Genebank ref: NM_148931.3)
Phox2b oligonucleotide probe	ACDBio	407861-C2	targets bp 1617 - 2790 (Genebank ref: NM_008888.3)