Detection of G Protein-Coupled Receptor Complexes in Postmortem Human Brain by Proximity Ligation Assay

Abstract

Combining immunological and molecular biological methods, the antibody-based proximity ligation assay (PLA) has been used for more than a decade to detect and quantify protein-protein interactions, protein modification and protein expression in situ, including in brain tissue. However, the transfer of this technology to human brain samples requires a number of precautions due to the nature of the specimens and their specific processing. Here we used the PLA brightfield detection technique to assess the expression of dopamine D2 receptor (D2R) and adenosine A2A receptor (A2AR), and their proximity in human postmortem brains, and we developed a systematic random sampling method to help quantify the PLA signals.

Introduction

Numerous biochemical, biophysical and immunological approaches are routinely used to study protein-protein interactions. However, it is still challenging to work with endogenous proteins in brain preparations to assess protein-protein interaction in situ with high resolution. Proximity ligation assay (PLA) is an immunoassay coupled with DNA rolling circular amplification (RCA), and provides an easy-to-use method that has been increasingly adopted for the detection of antigen proximity in situ. PLA allows the detection and quantification of interactions between molecules with high sensitivity. PLA provides greater resolution regarding the possible interaction of target molecules than traditional immunohistochemistry (IHC) assays. Furthermore, PLA does not require specialized equipment other that used in routine immunohistology studies. Since Fredriksson and his colleagues reported this technique in 2002 (Fredriksson et al., 2002), PLA has been widely used to study protein-protein interactions, protein-nucleic acid interactions, protein modification, and protein expression (Gomez, Shankman, Nguyen, & Owens, 2013; Gu et al., 2013; Gullberg et al., 2004; Lonskaya, Desforges, Hebron, & Moussa, 2013; Roussis, Guille, Myers, & Scarlett, 2016; Soderberg et al., 2006; Trifilieff et al., 2011). This technology was successfully commercialized, and the Duolink PLA kits, that were initially produced by Olink (Sweden), are currently provided by Sigma-Aldrich. Duolink PLA Fluorescent detection kits are available for fluorescent labeling for Green (490/520nm), Orange (542/562nm), Red (593/622nm), and FarRed 646/664nm), whereas the DuoLink PLA Brightfield Detection kit uses horseradish peroxidase (HRP) and its substrate, NovaRed, for brightfield microscopy.

In this assay, a pair of PLA probes (probe plus and probe minus), which are antigen specific antibodies conjugated with a short strand of nucleotides, target the molecules of interest either directly (primary antibody conjugated PLA probes) or indirectly (secondary antibody-conjugated PLA probes). Subsequently, short linker DNA, which are partially complementary to PLA probes “probe plus” and “probe minus”, and DNA ligation mixture is added to tissue sections to produce circular DNA template for RCA, which occurs only if the probe plus and probe minus are in close proximity. RCA produces hundreds of replicates of the DNA template, which can be hybridized with fluorophore- or Horseradish peroxidase (HRP)-conjugated detection probes, and consequently produce PLA puncta in situ (Sigma-Aldrich, 2015). Previous studies suggested that a pair of primary antibody conjugated PLA probes only generates signal when these two antibodies are within roughly 30 nm, including the size of two antibodies and the oligonucleotides connecting them (Soderberg et al., 2006). More recent studies using secondary antibody-conjugated indicate that the maximum distance between the PLA probes could be about 40 nm, based on the length of fully stretched DNA probes (Sigma-Aldrich).

In this protocol, we use secondary antibody-conjugated PLA probes and Brightfield PLA (PLA_BF) to study the interaction between the dopamine D2 receptor (D2R) and the adenosine A2A receptor (A2AR) in postmortem human brain. The dopamine system is involved in many functions, such as locomotion, motivation and reward, and learning. The D2 receptor (D2R) plays a critical role in dopamine transmission and is the target of multiple therapeutics for Parkinson disease and schizophrenia (Beaulieu & Gainetdinov, 2011; Urs, Peterson, & Caron, 2017). Numerous studies have shown that D2R-dependent signaling (Ferre et al., 2016) and related functions (Collins et al., 2012; Pardo et al., 2012) can be modulated by the activity of the A2AR in medium spiny neurons, possibly through heteromers formed by D2R and A2AR. We show through PLA that a fraction of these two receptors exist in close proximity in native brain tissue from both rodents and humans (Trifilieff et al., 2011; Zhu et al., 2019), which is of course a necessary prerequisite for the formation of functional oligomers.

In order to assess the proximity and potential interaction between D2R and A2AR, we performed “dual PLA”, in which two primary antibodies specific for each receptor and raised in different species were used. Rabbit anti-D2R and mouse anti-A2AR primary antibodies target D2R and A2AR, and two species-specific PLA probes, anti-rabbit probe minus and anti-mouse probe plus subsequently bind the primary antibodies (Fig. 1B). We also used “single PLA”, in which only one primary antibody was used to assess the expression of D2R or A2AR individually. For example, rabbit anti-D2R primary antibody binds to D2R, and then polyclonal anti-rabbit PLA probe plus and probe minus are added to recognize the primary antibody (Fig. 1A). In this case, since the secondary antibodies are polyclonal, complementation of the probe plus and probe minus can result from their binding to a single and/or two neighboring anti-D2R antibodies. Thus, puncta from dual recognition PLA correlate close proximity between two different antigens while puncta from single recognition PLA correlates expression of a single target protein (Trifilieff et al., 2011).

Figure 1.

Diagram of PLA_BF with secondary antibody-conjugated PLA probes. Single PLA was performed with one primary antibody (i.e. rabbit anti-D2R or mouse anti-A2AR) and a pair of PLA probes to the primary antibody (A). Dual PLA was performed with two primary antibodies from two different species and the species specific PLA probes (B). For a discussion of the impact of using polyclonal antibodies on the possible binding configuration, see the main text (Introduction and Understanding the PLA results).

The basic PLA_BF protocol described herein is applicable to thin paraffin sections and frozen sections from fixed or snap-frozen fixed tissue mounted on glass slides. The PLA results can be visualized and assessed with a regular brightfield microscope directly or imaged by whole slide scanning. The protocols describe the methods of Luxol fast blue/cresyl violet staining, and systematic random sampling of whole slide scanned images. We recently implemented a machine learning-based quantification program, Bayesian optimized PLA signal sorting (BOPSS), which can be used for automatic quantification of PLA puncta (Zhu et al., 2019).

Strategic Planning

1. Controls: nonspecific staining that looks similar to PLA signal is quite common with PLA, and can be caused by many issues, including tissue quality, nonspecific binding from primary antibody or PLA probes, expired PLA kits, etc. Well-designed controls are mandatory to validate the specificity of the PLA signal. See the Critical Parameters section concerning controls.

2. Antibodies: We recommend validating the specificity of the primary antibodies and optimizing the working conditions through traditional IHC. Ideally, if the epitopes of interest are conserved in human and mouse, the specificity and working conditions of primary antibodies in PLA should be tested through IHC (and ideally with single PLA as well) with mouse brain tissue first. Validating the antibody specificity with gene knockout mice is highly recommended. Antigen retrieval may be required, depending on specimen preparation and antibodies.

3. Sample format. To obtain better staining results, we recommend using thin sections (< 10 μm). Free-floating staining is more adapted for thick sections (around 30 to 50 μm).

4. Detection method and sample preparation. The detection method needs to be carefully chosen after considering the aim of the study, the quality and preparation method of the specimen, data analysis method, and nonspecific background. Although, theoretically, both brightfield and fluorescence methods are applicable in many cases, we have found that PLA brightfield is highly preferable for the processing of adult human brain samples (Zhu et al., 2019).

5. Imaging. 40x objectives (at least 20x) and z-step scanning imaging is strongly recommended. If whole slide scanning is used, the imaging parameters should anticipate the counting procedure that will be employed.

6. Analysis of the results. Systematic random sampling (SRS) is recommended for oversized and heterogeneous samples, such as human brain slices.

Basic Protocol 1: Sample Preparation and Sectioning for PLA_BF

PLA has been widely used in cell culture and animal tissue sections, but seldom in human tissues, especially the brain. For IHC, human brains are typically fixed by immersing in fixative, embedded in paraffin, and prepared for further assays as thin sections on glass slides. We describe herein a protocol for preparing mounted paraffin sections.

All procedures involving postmortem human tissue or animals must be performed in accordance with National Regulations and Guidelines and approved in advance (e.g. Institutional Review Board and Institutional Animal Care and Use Committee). The human subject study in this protocol was approved by the Institutional Review Board of the New York State Psychiatric Institute. These protocols for embedding and sectioning are adapted from the Abcam Immunohistochemistry Application Guide/Fixation, embedding and sectioning/Paraffin-embedded tissue (Abcam).

Basic Protocol 1 Materials

Protocols for preparation the solutions are listed after the basic protocol)

* Chemicals and solutions that must be handled in a laminar hood.

• Postmortem human brain tissue (Institute for Forensic Medicine, Ss. Cyril & Methodius University, Skopje, Macedonia)

• Phosphate buffered saline (PBS) pH 7.4.

• 4% phosphate buffered formalin (see recipe)

• 4% paraformaldehyde (PFA) in PBS, pH 7.4 (see recipe)

• PBS with 0.02% sodium azide

• Ethanol (anhydrous denatured, histological grade)

• Deionized water (DI H₂O)

• Xylene*

• Paraffin wax

• Scalpel handle (#4) and matched blades (interchangeable)

• Micro dissecting forceps (straight, serrated, 4”)

• Pencil

• Paraffin embedding cassettes

• Histology Specimen Processor (Tissue-Tek VIP, Histotronix)

• Paraffin Wax Tissue Embedding Work Station (Tissue Embedding Center, Reichert-Jung)

• Microtome (RM2155, Leica)

• Positively charged glass slides

Basic Protocol 1 Steps

Paraffin embedded specimen preparation

• 1Fixation: Slice cerebral hemispheres coronally at 2 cm intervals and immerse in phosphate-buffered formalin at 4 °C for 5 days. Rinse the fixed sample with PBS and store in PB buffer at 4°C with 0.02% sodium azide until embedding.

Transfer postmortem unfixed specimen on ice to preserve tissue. The volume of fixative should be greater than 10 times the tissue volume.

• (Optional) When collecting mouse brain, euthanize mice by cervical dislocation, or other approved method. Cut off the head and dissect the brain out on ice. Rinse briefly with ice-cold PB, and immerse the specimen in 4% paraformaldehyde (PFA) at 4 °C in PB for 1 to 3 days. Rinse and store fixed brains in PB with 0.02% sodium azide until embedding. The volume of fixative should be greater than 10 times the tissue volume.

  Notice: Transcardial perfusion is optional, but might be required for labile antigens, such as phosphorylated proteins. Note that this may lead to non-specific nuclear background (Zhu et al., 2019).

  • 2
    Trim fixed specimen and place in pencil labeled tissue-embedding cassettes.
  • 3
    Following the procedure of your Tissue Processor and Embedding Workstation, dehydrate and embed the specimen blocks.
  • 4
    Store paraffin tissue blocks at RT, avoiding direct light and humidity.

Sectioning of Paraffin-embedded tissue block

Attention: Expertise is required to perform paraffin sectioning, especially for oversized specimen such as human brains. Well-prepared sections are critical for successful PLA, because the sectioning issues, e.g. bubbles, wrinkles or folded sections, will likely cause uneven staining and nonspecific positive signals.

• 5Set a water bath with ultrapure water at 40 to 45 °C.
• 6Place paraffin tissue blocks on ice before sectioning.
• 7Place the blade in the holder. Follow the manufacturer’s instructions to set the blade.
• 8Insert and orientate the paraffin tissue block. The blade should cut straight across the block.
• 9Cut a few thin sections and adjust the positions of the blade and the tissue block if necessary.
• 10Trim the block with thick sections (10 to 30 μm) until close to the region of interest.
• 11Cut serial sections of suitable thickness (<10 μm), pick the sections out of the water bath and mount the sections on positively charged glass slides.
• 12Let the slides dry at room temperature overnight.
• 13Store the slides in sealed slide boxes at room temperature, avoiding humidity and direct light exposure.

Basic Protocol 2: PLA_BF Staining of Brain Tissue

We have found that the commonly used fluorescence-based PLA (PLA_FL) is not suitable for human brain samples due to autofluorescence issues, which mostly originate from lipofuscin aggregation as a result of aging (Zhu et al., 2019). In addition, in PLA_FL assays, human and some adult rodent tissues often display dense autofluorescent dots inside the soma, which can overlap with actual PLA signal and therefore reduce signal/background ratio and affect quantification. Unlike PLA _FL, the PLA_BF signal is visualized by HRP substrate colorization, viewed and imaged with a brightfield microscope, thereby avoiding autofluorescence. In addition, stained slides of PLA_BF can be stored at room temperature (RT) for longer periods of time than PLA_FL and are not subject to photobleaching.

The basic protocol we describe herein uses PLA_BF and secondary antibody-conjugated PLA probes and Brightfield PLA (PLA_BF) to assess the proximity between dopamine D2 receptor (D2R) and adenosine A2A receptor (A2AR) in human striatum (Zhu et al., 2019).

Basic Protocol 2 Materials

Protocols for preparation of the solutions are listed after the basic protocol)

* Chemicals and solutions that must be handled in a laminar hood.

• Sections from formalin-fixed, paraffin-embedded human brain tissue mounted on glass slides

• Phosphate buffer (PB) pH 7.4

• Xylene *

• 100%, 95%, 80%, 75% and 50% ethanol (anhydrous denatured, histological grade)

• Deionized water (DI H₂O)

• Tris buffer with sodium (TBS), pH 7.4

• Tris buffer with sodium and tween-20 (TBS-T), pH 7.4

• 10% Tween-20

• Antigen retrieval buffer: 10 mM Sodium Citrate, pH 6, with 0.02% Tween-20

• Quenching buffer: 1 % hydrogen peroxide (H₂O₂) in TBS (Store hydrogen peroxide in the dark at 4 °C)

• Antibodies: Dopamine receptor D2R was probed using a rabbit polyclonal anti-D2R (Millipore Cat# ABN462, RRID: AB_2810225) and adenosine receptor A2AR was detected using a mouse monoclonal anti-A2AR antibody (Millipore 05717, RRID: AB_11213750). These antibodies were previously validated for specificity using knockout mice lacking the targeted receptors (Biezonski, Trifilieff, Meszaros, Javitch, & Kellendonk, 2015; Trifilieff et al., 2011).

• PLA probes (Duolink): rabbit plus, rabbit minus, mouse plus, mouse minus

• PLA detection kit for brightfield (Sigma Duolink)

• PLA wash buffer A (See recipe or Sigma)

• Permount mounting solution (toluene solution) *

• Slide staining set (at least 10 jars) with removable slide holder (Tissue Tek)

• Liquid blocker Pap Pen for immunostaining (Sigma)

• Freezer block for enzymes

• Forceps for picking slides

• Glass slide incubation boxes (home-made: In a Thermo Scientific™ Nunc™ Square BioAssay dish, lay up to three 1000 ul pipette tip racks on two pieces of Kimwipes, and wet the Kimwipes with DI H₂O before incubation)

• Coverglass (The thickness of No.1 coverglass is 0.13 to 0.17 mm. The optimal choice of coverglass depends on the model of microscope lens.)

• Microwave oven or hot plate

• Humidity controlled Incubator at 37 °C

• Brightfield microscope with 4x, 20x and 40x lens

Basic Protocol 2 Steps

General rules for PLA:

• Do not let the tissue or sections dry at any time during the procedure until mounting.

• The minimal volume of incubation for each step should be 40 μl for a tissue section of about 1 cm².

• Use enough fixative solution and washing buffer. Generally, 10 ml solution per 1 slide in deparaffinization, rehydration and all washing steps.

• Use a level to confirm that the surface on which the slides are incubated is horizontal.

• Use autoclaved DI H2O.

Attention: make sure all sections are fully immersed in solutions

• 1
  Prepare 10 slide staining jars with removable slide holder (Tissue Tek).

  • Two for xylene (in fume hood)
  • two for 100% ethanol
  • two for 95% ethanol
  • one for 70% ethanol
  • one for 50% ethanol
  • two for TBS
• 2Label the slides with a pencil and load them in a removable rack.
• 3Incubate the slides in xylene at RT for 2 × 5 minutes. No washes between step 3 to 8.
• 4Transfer and incubate the slides in 100% ethanol for 2 × 5 minutes.
• 5Transfer and incubate the slides in 95% ethanol for 2 × 5 minutes.
• 6Transfer and incubate the slides in 70% ethanol for 5 minutes.
• 7Transfer and incubate the slides in 50% ethanol for 5 minutes.
• 8Transfer and rinse the slides twice with TBS for 2 × 5 minutes.

(Optional) Heat-induced antigen retrieval (antigen unmasking)

Attention: Antigen retrieval is required for some antibodies, such as anti-D2R and anti-A2AR for paraffin embedded tissue. Antigen retrieval solutions and methods must to be optimized for specific antibodies.

• 9Transfer the slides to the antigen retrieval buffer (10 mM sodium citrate buffer at pH 6.0, add 0.02% Tween-20 freshly) in a microwave-safe container of appropriate size, e.g. a beaker or Tissue Tek staining box. Do not preheat the antigen retrieval buffer.
• 10Boil the slides in the antigen retrieval buffer in a microwave at high power and keep the buffer boiling for 3 minutes. Optional: you can stop the microwave and reduce power to 70% to avoid buffer evaporation. Alternatively, a hot plate can be used to boil the antigen retrieval buffer.
• 11Turn off the microwave and let cool for 5 minutes.

If too much evaporation occurs during step 9), add antigen retrieval buffer to ensure sections are immersed in the buffer.

• 12Boil the buffer again for 3 minutes.
• 13Keep the slides in the hot antigen retrieval buffer until they cool to room temperature.

Optional: Put the staining box in a water bath at room temperature until the buffer cools to RT. In either case, the buffer should be at room temperature before the slices are transferred.

• 14Rinse the slides in TBS at room temperature for 2 × 5 minutes.

Quenching of endogenous peroxidase activity

• 15Incubate the slides in quenching buffer (TBS with 1% hydrogen peroxide) at room temperature for 30 minutes.
• 16Rinse the slides in TBS-T twice at room temperature for 5 minutes each.

Blocking

• 17Remove the slides from the TBS-T rinse and remove the liquid from slides with Kimwipes.
• 18
  Draw a circle around the tissue with a liquid blocker PAP pen and add the blocking buffer (included in the PLA Probe kit) to cover the sample.

  The PAP pen circle provides a water repellent barrier to prevent the solution from spreading over the whole slide. Attention: do not let the tissue sections dry. Avoid touching the tissue sections with the PAP pen, and leave some space between the PAP pen line and the sample; otherwise, the hydrophobic barrier from the PAP pen can prevent subsequent reactions. If this happens, wash the slide with 0.1M TBST buffer three times.

CAUTION: In the following incubation steps, especially those at 37 °C, the edges of the samples will dry if there is too little solution. Remember, the buffer volume should be sufficient to cover the entire section for all the following incubation procedures.

• 19Place the slides in a sealed glass slide incubation boxes with wet Kimwipes (An ICH slide staining tray or home-made chamber with slide racks). Make sure all the slides are placed horizontally and all the sections are evenly covered by the blocking buffer. The wet Kimwipes help to maintain humidity in the staining box and reduce evaporation.
• 20Incubate the slides at room temperature for 1 hour.

Primary antibody incubation

Prepare the primary antibody working solution during the blocking step.

• 21Dilute the primary antibody in Antibody Diluent (included in the PLA Probe kit).

For D2R-A2AR dual recognition PLA, add both anti-A2AR (1.67 μg/ml) and anti-D2R (5 μg/ml) antibodies. For single recognition PLA, use Anti-A2AR (1 μg/ml) or Anti-D2R (1.67 μg/ml), respectively. For the dual PLA negative control for PLA probes, add only one of the two primary antibodies. Primary antibody concentration needs to be optimized for each specific assay.

• 22Tap off the blocking buffer and replace with the diluted primary antibody.
• 23Place the slides in the sealed glass slide incubation boxes at 4 °C, making sure that all the slides are placed horizontally and all sections are evenly covered by the antibody incubation buffer. Incubate the slides with the primary antibodies at 4 °C for 16 hours or longer (overnight).

PLA-Probe incubation

The following steps are adapted from the Duolink PLA protocol for brightfield detection (Sigma-Aldrich)

• 24Remove the primary antibody solutions and rinse the slides three times in TBS-T for 10 mins each (10 ml buffer per slide).

Wash dual PLA sections, single PLAs and negative controls in separate staining jars, if possible. PLA is very sensitive and the contamination of primary antibodies can cause false positive signals.

• 25During the TBS-T washing step, prepare the PLA probe mixture: mix the probe stock solution and dilute two PLA probes (one probe plus and one probe minus) 1:5 in the antibody diluent (from the PLA probe kit) for each PLA, and keep the mixture on ice.

For example, mix anti-mouse plus and minus for single PLA against A2AR. Mix anti-rabbit plus and minus for single PLA against D2R. Mix anti-mouse plus and anti-rabbit minus for dual PLA against A2AR and D2R. Always transfer the PLA probes on ice, and do not let the mixture of PLA probes sit at RT.

• 26Drain the slides and remove the liquid around the sections with Kimwipes or with slow vacuum flow. Make sure to keep the sections wet. Draw a PAP pen circle again if necessary.
• 27Add the mixture of PLA probes and place the slides in the sealed glass slide incubation boxes.
• 28Place the incubation boxes in a preheated humidity chamber (e.g. tissue culture incubator) at 37 °C for one hour.

Ligation

• 29Thaw the 5X ligation buffer on ice and vortex it.
• 30Dilute the ligation buffer 1:5 in autoclaved DI H₂O and vortex.

Take the volume of ligase (1/40) into account when calculating the amount of water in the final 1X amplification solution. Wait to add the ligase until immediately before addition to the samples. The ligation buffer contains ATP and is sensitive to repeated freeze-thaw cycles. Aliquot the ligation buffer if necessary.

• 31Tap off the PLA solutions from the slides and rinse the slides in 1X PLA wash buffer A for 3 × 10 mins.
• 32After the last rinse, add the ligase to the ligation solution from step 30) at a 1:40 dilution and mix well by pipetting (transfer ligase in a frozen block at −20 °C).
• 33Add the ligation–ligase solution to the sections.
• 34Place the slides in the glass slide incubation boxes and place them in a preheated humidity chamber (e.g. tissue culture incubator) at 37 °C for 30 min.

Amplification

• 35Dilute the Duolink In Situ Amplification stock buffer at 1:5 in autoclaved DI H₂O and mix.

Take the volume of Polymerase (1/80) into account when calculating the amount of water in the final 1X amplification solution. Wait to add the polymerase until immediately before the addition to the samples and mix well by pipetting.

• 36Tap off the ligation-ligase solution from the slides.
• 37Rinse the slides in 1X PLA Wash buffer A for 3 × 5 mins.
• 38During the final rinse, add the Duolink in situ Polymerase to the amplification solution prepared in step 35) at a 1:80 dilution and mix by pipetting.
• 39Add the amplification–polymerase solution prepared in step 38) to the samples, and place the slides in the glass slide incubation boxes.
• 40Incubate in a preheated humidity chamber for 120 min at 37 °C.

HRP labeling

• 41Dilute the 5X brightfield detection stock at 1:5 in autoclaved DI H₂O and mix by brief vortexing.
• 42Tap off the amplification–polymerase solution from the slides.
• 43Rinse the slides in 1X PLA wash buffer A for 3 × 5min.
• 44Add the 1X detection solution prepared in step 41).
• 45Save 5 μl of the diluted detection solution for HRP–NovaRed test.
• 46Place the slides in the glass slide incubation and incubate for 60 minutes at room temperature.

Detection with HRP substrate NovaRed

Optional: If a large volume of staining solution is required, NovaRed staining kit (NovaRed Peroxidase HRP Substrate) is available from Vector Laboratories.

• 47Dilute the reagents A (1:70), B (1:100), C (1:100) and D (1:50) in autoclaved DI H₂O. This is the HRP–NovaRed reaction solution.
• 48Perform an HRP–NovaRed quality test. Mix 5 μl of the diluted NovaRed solution (the ABCD solution prepared above) with 5 μl detection solution saved in step 41) and monitor the color. The colorless solution should darken quickly. Otherwise, check and prepare the solutions in step 41) and 47) again.
• 49Tap off the detection solution from slides and rinse the slide with PLA wash buffer A for 3 × 5 min.
• 50Add the HRP–NovaRed reaction solution to each sample.
• 51Incubate the slides for 5 to 10 minutes at room temperature (Note: the reaction time needs to be optimized).
• 52Stop the reaction by placing the slide in DI H₂O and rinse 2 × 2 mins.
• 53(Optional) Add the Duolink in situ Nuclear Stain to each sample and incubate the slides at room temperature for 2 minutes. The nuclear counterstaining is helpful to define the outline of specimen during imaging.
• 54Gently rinse the slides under running tap water for 10 mins.

Dehydration and mounting

• 55Rinse the slides in fresh 95% ethanol solution for 2 × 2 minutes.
• 56Rinse the slides in 100% ethanol 3 × 2 minutes.

Dehydration is critical to get clear PLA results and sharp images. Make sure to use fresh 100% ethanol for this step. If the bottle of 100% ethanol is not sealed well, do not use it for dehydration.

• 57Incubate the slides in a fresh xylene solution for 3 × 5 minutes (under a laminar hood). Use fresh xylene for this step.
• 58Mount the slides with a minimum volume of Permount (or other toluene-based non-aqueous mounting medium). Avoid bubbles and extra Permount.
• 59Dry the slides under a fume hood for at least 2 days before imaging or checking the results with a brightfield microscope (20x lens at least).

Basic Protocol 3: Image Acquisition and Result Analysis

D2R and A2AR are heterogeneously expressed throughout the striatum, and the large size of human brain sections make quantification challenging. In order to analyze the results in an unbiased and efficient manner, we adapted a systematic random sampling method from stereology (Gundersen, Jensen, Kieu, & Nielsen, 1999) for whole slide scanned images of the brain sections, and perform automatic PLA puncta quantification with BOPSS (Zhu et al., 2019).

Attention: If whole slide scanning is used, the magnification and numerical aperture (NA) of the objective lens, the NA of the condenser, the immersion medium, and the number of imaging planes should be compatible with the counting procedure that will be employed.

Basic Protocol 3 Materials

• Whole slide scanner system (Leica SCN 400 system)

• Whole slide scanned images of LFB/CV staining and PLA_BF (on serial sections)

• Transparent sheets for printer

• Aperio ImageScope (Leica Biosystems Pathology Imaging)

• Adobe Photoshop (Adobe)

• Adobe Illustrator (Adobe)

• BOPSS (MATLAB)(Mészáros, 2019)/ ImageJ (Schneider, Rasband, & Eliceiri, 2012) / ICY (de Chaumont et al., 2012)

• Optional: Stereo Investigator (MBF Bioscience)

Basic Protocol 3 Steps: Systematic random sampling for whole slide scanning images

Procedures (adapted from (Zhu et al., 2019))

1. Make an outline printout for each sample.

   1. Open a virtual whole slice image (a .scn file) of LFB/CV staining (Fig. 2A) in Aperio ImageScope (Leica).
   2. Zoom to have a full view of the brain section that fits the screen, and export this full view image as a tiff file with the Snapshot function.
   3. Open this full view image of LFB/CV staining in Photoshop (Adobe), create a new transparent layer, and draw the outline of this brain section and its sub-territories, i. e. region of interest (ROI) and landmarks to orient the section) on this new layer.
   4. Close the layer for LFB/CV image, and print only the layer for outline on a transparent sheet without scaling (Fig. 2B). Make an outline printout for each sample.

Figure 2.

Design of the systematic random sampling for PLA_BF (Zhu et al., 2019). Luxol fast blue/cresyl violet (LFB/CV) staining shows the white matter and the grey matter in a full view image of the brain section (A). An outline of the ventral striatum sub-region (B) is drawn based on the LFB/CV stained image (A) and printed on transparency file. The overlap of the sampling grid (C) and the outline (B) dividing the region of interest (ROI) into a number of evenly distributed sampling areas (D). One counting locus was selected in each sampling area (indicated by *) and the 40x image of this counting locus was exported for quantification. The counting locus was only used for quantification when it is inside the ROI. Scale bar, 5 mm.

Note: Because all images exported in this protocol are produced by screenshot, it is critical to keep the monitor size and the mode of view in Aperio ImageScope (e.g. full view or regular view) consistent in the sampling process. Otherwise the size of exported images will vary, which will disrupt the consistency of sampling and subsequent quantification.

• 2.
  Make a sampling grid (Fig. 2C) that divides the section or ROI into a series of sampling areas of the same size. Only a small zone, which we refer to as a counting locus, of a sampling area will be analyzed to obtain representative data for the entire sampling area.

  1. In Adobe Illustrator, create a grid image (16.5 mm x 16.5 mm/cell in this protocol) of the same size as the outline image.
  2. Print the grid on transparency film without scaling. This printout is a sampling grid (Fig. 2C), and each cell of the grid corresponds to one sampling area (Fig. 2D). The size of the grid and its cells should be suitable to cover an entire ROI and create a proper number of sampling areas for the ROI.
• 3.Select and mark the counting locus. Select a specific location within a sampling area (cell) as the counting locus, and mark it on the sampling grid. Repeat this for all sampling areas. The location of the counting locus can be anywhere within the sampling area, as long as it is consistent across all sampling areas.
• 4.
  Sampling.

  1. Open a whole slice image (scn file) of PLA_BF in Aperio ImageScope in the same mode of view as in step 1).
  2. Zoom the PLA_BF image to have a full view that fits your screen (0.4x in this study).
  3. Overlap and mount the outline printout of this section on screen. Mount the sampling grid on top of the outline to cover the entire ROI. Now the ROI in the PLA_BF image is defined by the outline and divided into a series of sampling areas by the sampling grid.
  4. Turn on the Pan Navigation in Aperio ImageScope, the “hand” in the tool bar, and now the mouse indicator changes from an arrow to a hand.
  5. Place the index finger of the mouse indicator at a counting locus that is labeled in step 3) and zoom the image to full magnification (40X in this protocol) by double left-click. Do not move the mouse until you get the image of full magnification. Then you should have a fully magnified image of the counting locus (Optional for z stack images: choose the layer with most PLA puncta in focus).
  6. Export this 40X image using the Snapshot function and save it as a tiff file. Thus, a section of a sampling area, which is represented by the 40x image of the counting locus, is selected for further analysis.
• 5.Zoom the PLA_BF image back to the full view (0.4x) again and repeat step 4) for all counting loci within the ROI. For sampling areas that are partially covered by the ROI, we only analyze those when their counting loci are inside the ROI.
• 6.Organize the sampled images (step 4) by specimen, ROI and type of PLA assay. Examine the quality of the images. Unfocused or oversaturated images will not be analyzed correctly. Always make a copy for all saved images.
• 7.After collecting the sampled images, perform automatic PLA puncta quantification with an automated PLA signal quantification. We recommend the customized MATLAB package, BOPSS (Mészáros, 2019; Zhu et al., 2019). BOPSS is a machine learning based and user-defined-parameter free method, detecting PLA signal by color- and size-based segmentation and Naïve Bayesian classifier. , ICY_Spot Detector (de Chaumont et al., 2012) or ImageJ_Particle (Schneider et al., 2012) are open source software, but need image transformation and parameter optimization. We have previously discussed the advantages and limitations of these quantification methods in details (Zhu et al., 2019).

Support Protocol: Luxol fast blue/ cresyl violet staining (LFB/CV)

This protocol is adapted from: http://www.ihcworld.com/_protocols/special_stains/fast_blue.htm

Materials

• Mounted sections on glass slides

• Histology staining rack

• Staining jars

• Incubator

• Absolute ethanol

• 95% and 70% ethanol

• Xylene

• DI water or distilled water

• 0.1% Luxol Fast Blue

• 0.5% Cresyl Fast Violet

• Lithium Carbonate (Sigma Aldrich: 62470)

• 0.05% Lithium Carbonate

• Permount mounting medium

Procedures

To define the sub-territories of interest in the striatum, use one of the serial sections for LFB/CV staining to stain the white matter and the gray matter. The myelin and phospholipids will be stained blue to green, and the Nissl substance (neurons) will be stained violet. The entire procedure is at room temperature unless specified.

• 1Deparaffin and rehydrate paraffin sections following the instructions in Basic Protocol 2.

Luxol Fast Blue staining.

• 2Place the slides in Luxol fast blue solution and incubate at 58 °C overnight. Cap or seal the staining jar to prevent evaporation.
• 3Rinse the slides once with 95% ethanol.
• 4Rinse the slides once in DI water or distilled water.
• 5Place the slides in the 0.05% lithium carbonate solution for 10 seconds.
• 6Rinse in 70% ethanol twice (new ethanol solution for each).
• 7Repeat step 5) to 6) until there is a sharp contrast between the blue white matter and the colorless gray matter. Check slides under microscope and do not over differentiate. The number of cycles may vary.

Cresyl Violet staining.

• 8Place the slides in 0.1% cresyl violet solution for 5 minutes.
• 9Rinse the slides in DI water or distilled water for 3 × 1 minute.
• 10Place the slides in 70% ethanol (differentiation) for 5 minutes or until background in the gray matter is colorless. The time of differentiation may vary.
• 11Dehydration. Rinse the slides in fresh 95% ethanol solution for 2 × 2 minutes and in 100% ethanol 3 × 2 minutes.
• 12Incubate the slides in a fresh xylene solution for 3 × 5 minutes (under a laminar hood).
• 13Mount the slides with Permount.

Reagents and Solutions Preparation

Use high purity water, i.e. deionized water or distilled water, to prepare the following solutions.

• For 1 L PBS: Dissolve 9 g NaCl, 1.92 g NaH₂PO₄, and 4.83 g Na₂HPO₄ in 800 ml DI H₂O. Adjust pH to 7.4 and adjust final volume to 1 L. Store at 4°C for up to 1 year. Concentrations: 0.05 M phosphate buffer; 0.15 M NaCl.

• Paraformaldehyde (PFA), 4%, in PBS For 100 ml: Heat 80 ml PBS (see recipe) in a beaker to 60°C on a hotplate with a magnetic stirrer. Weigh 4 g PFA powder, add it to the heated PBS in the hood, and mix solution with a stirrer. PFA will not dissolve at once. Increase pH with1 M NaOH until the solution becomes clear. Let the dissolved PFA/PBS cool down to room temperature and adjust pH to 7.4 with HCl. Adjust volume to 100 ml and filter solution to remove undissolved PFA. Store at 4°C for up to 1 week or aliquot and store frozen for up to 1 month; avoid repeated thaw/freeze cycles. Wear proper personal protective equipment and perform this procedure in a fume hood. Use freshly prepared 4% PFA in PBS for good fixation. Optional: Commercially prepared PFA solutions are also available at various concentrations.

• 1 L 4% Phosphate buffered formalin: Dissolve 4 g anhydrous NaH₂PO₄ (FW 120) and 6.5 g anhydrous Na₂HPO₄ (FW 142) in 800 ml DI H₂O, add 100 ml concentrated formalin (40%) and adjust final volume to 1 L. Adjust amount if using hydrated Na₂HPO₄ and NaH₂PO₄.

• 1 L 0.1M PB buffer with 0.02% sodium azide. Add 0.05 mL 5% sodium azide (BICCA) to 1 L 0.1M PB buffer.

• 1 L 10 mM Sodium Citrate: Dissolve 2.58 g sodium citrate (FW: 258) in 800 mL DI H₂O, adjust pH with HCL to 6, and add DI H₂O to 1L. Store up to 6 months at 4 °C.

• 200 ml antigen retrieval buffer (10 mM Sodium Citrate containing 0.02% Tween-20): Add 400 ul 10% Tween-20 before use.

• 1 L 1 M NaCl: dissolve 58.4 g sodium chloride (FW 58.4) in 1 L DI H₂O, sterilize by autoclave and store at RT, or at 4 °C for long term storage. Discard if precipitate accumulates.

• 1 L 1 M Tris-HCL solution (pH 7.4): Dissolve 121.1 g of tris base in 800 ml of H₂O and add 70 ml concentrated HCl (*) (Handle the concentrated HCl with extreme care in a fume hood). Cool the solution to room temperature (RT) and adjust the pH to 7.4 with 1 M HCL. Adjust the volume of the solution to 1 liter with H₂O. Sterilize by autoclaving and store at RT for short term and at 4 °C up to one year. Because the pH of the Tris buffer is temperature sensitive, always adjust the pH and use the Tris buffer at room temperature.

• TBS (0.1M Tris, 150 mM NaCl): For 1 L TBS, add 100 mL 1M Tris (pH7.4) and 150 mL 1M NaCl, adjust the volume to 1 L. Store at RT for up to one month.

• TBS-T: For 1 L TBS-T, add 2 mL Triton-X100 to 998 mL TBS. Store at RT for up to one week.

• 10 % Triton-X100: Add 10 Add 10 ml Triton-X100 to 90 ml DI H₂O, and mix well. Store up to 12 months at 4 °C.

• 10% tween-20: Add 10 ml tween-20 to 90 ml DI H₂O, and mix well. Store up to 12 months at 4 °C.

• PLA wash buffer A: To prepare 1L buffer A, add 10 ml 1M tris-HCL (pH7.4), 150ml 1 M NaCl, and 5 ml 10% Tween-20 to 800 ml DI H₂O, adjust the volume to 1L. Store up to 1 week at RT. PLA wash buffer A is also available from Sigma.

• 0.1% Luxol fast blue solution: Dissolve 0.1g Luxol fast blue (molecular biology standard) in 100 ml 95% ethyl alcohol with 0.5 ml glacial acetic acid

• 0.1% Cresyl violet solution: Dissolve 0.1 g cresyl violet (cresyl fast violet) in 100 ml DI water. Add 10 drops of glacial acetic acid just before use and filter.

• 0.05% Lithium carbonate solution: Dissolve 0.05 g Lithium carbonate in 100 ml DI water.

Commentary

a. Background Information

A PLA signal is formed and amplified at the locus of ligation between a pair of plus and minus probes, under controlled reaction conditions. PLA signals are individual puncta, representing the loci of proteins or protein-protein interaction. Thus, quantifying the level of expression or interaction becomes easy through quantifying the number (rather than the size) of PLA puncta, with additional spatial information. In addition, PLA signals are amplified up to 1000× by DNA rolling circle amplification; therefore this method is sensitive enough to assess endogenous protein-protein interactions in both cell cultures and tissue sections.

In addition to the paraffin-embedded tissue described in this protocol, PLA_BF can be applied in a variety of formats including frozen sections from fixed, snap frozen samples or free-floating sections of fixed samples.

b. Critical parameters

Antibodies

The specificity and optimal concentration of primary antibodies, tissue quality and general PLA reaction conditions can be validated through regular immunostaining. Generally, antibody reaction conditions for immunostaining are applicable to PLA. However, because of signal amplification with PLA, single recognition PLA may require reducing antibody concentration, as signal saturation can result in overcrowded signal that is difficult to quantify. While in dual PLA, if the signal is low, increased concentration of primary antibody may be required for optimal signal, although it is critical to use primary antibodies in a range where they remain specific to the antigen being detected.

Tissue fixation

Formaldehyde (e.g. formalin and PFA) fixes tissue by cross-linking the proteins, which can mask the epitopes and prevent antibody binding. Therefore, if samples were fixed with formaldehyde, antigen retrieval may be required for good PLA signals. Besides the heat-induced epitope retrieval (HIER) in sodium citrate buffer used in this protocol, there are many other antigen retrieval methods, such as HIER in different kinds of buffer (e.g. 1% sodium dodecyl sulfate in phosphate buffered saline), proteolysis induced epitope retrieval, etc. A specific antigen retrieval method should be tested for a specific antibody, considering the epitopes and the fixation method of specimen.

Non-specific signal

As with all immunological techniques, non-specific staining and background should be reduced to obtain good signal/noise ratio, which is especially important for automatic software-based PLA puncta counting. As previously reported (Zhu et al., 2019), we found that some specimen preparation methods, such as fixation after transcardial perfusion with 4% PFA or snap freezing, cause nonspecific background in nuclei of mouse brain tissue. Therefore, we recommend fixing the tissue through direct immersion to avoid this potential issue, if the target antigen of interest is stable during slow fixation.

The controls

Well-planned controls are critical for PLA assay. Samples devoid or depleted of the protein(s) of interest, such as knockout or knockdown animal models or tissues that do not express the protein(s) of interest, are ideal negative controls. Negative controls omitting the primary antibodies or the PLA-probes are necessary for quality control of the PLA probes and PLA detections kits, as various factors, e.g. nonspecific binding of PLA probes and expired PLA products, can cause nonspecific binding and signal. However, those technical negative controls cannot validate the specificity of PLA signals, which could be caused by nonspecific binding of primary antibodies. Step-by-step positive controls are helpful to ensure the PLA procedure is performed correctly, such as regular immunostaining with the same primary antibodies and specimen to verify the antibodies and tissue quality, and the HRP-substrate test to confirm the HRP reaction.

c. Troubleshooting

Problem	Common Causes	Suggestions
No signal	Primary antibody reaction failed.	Try antigen retrieval and increase primary antibody concentration.
No signal	HRP substrate reaction failed. The detection solution or the HRP substrate/ NovaRed solution (from PLA-Brightfield box B) is not prepared correctly, or some of the component (A, B, C, D) expired, e.g. H2O2.	Prepare fresh HRP substrate/ NovaRed solution. Perform the HRP–NovaRed test). Detection solution should be prepared freshly.
No signal	Ligation failed.	Do not use expired kits. The ATP in ligation buffer is sensitive to repeated thaw and freeze cycle. Aliquot the ligation buffer upon receiving the kits. All enzymes should be handled and transferred in a −20 °C frozen block.
No signal or very few and small puncta	Failed or inefficient amplification	Do not use expired kits. Store and use the amplification buffer and enzyme properly. All enzymes should be handled and transferred in a −20 °C frozen block. Increase the time of amplification.
Strong PLA signal in negative controls without primary antibody	Expired PLA probes or detection kits can cause nonspecific binding.	Do not use expired kits.
Saturated signal in experimental samples	Over-concentrated primary antibody or PLA-probes	Reduce the concentration of primary antibodies (or PLA-probes), or the time of amplification.
Abnormal increased signal in some area	Wrinkled/ uneven sections, or unleveled surface during incubation	Check the quality of the sections and the surface for incubation.
PLA puncta in blank area, e.g. inside the cross section of vessels or lateral ventricular area	Insufficient washing	Increase washing time and the volume of washing buffer

d. Understanding PLA results

Fig. 3 shows representative results of single PLA signal for D2R and A2AR, dual PLA signal for D2R-A2AR and negative controls for PLA in human brains. PLA is designed for qualitative analysis and relative quantification. PLA assays produce signal/puncta between plus and minus PLA probes in close proximity. The puncta of single recognition PLA indicate expression of a target protein, and the puncta of dual recognition PLA indicate close proximity between two different antigens, and therefore the likely interaction of two target proteins directly or in a complex. Posttranslational modification, such as phosphorylation, can also be detected with dual PLA between a phospho-specific antibody and a second antibody to a different epitope in the protein of interest. With the PLA_BF kit from Sigma, the puncta are red-brown, from HRP–NovaRed reaction. PLA probes commercially available include polyclonal secondary antibodies against mouse, rabbit, or goat. We observe puncta of varied size, especially when the primary antibodies are polyclonal. The increased size of puncta could reflect increased clustering of the protein/complex. However, except when using customized monoclonal antibody conjugated probes that each can bind a single epitope, it is difficult to determine why some puncta are larger than the others. Therefore, we quantify the number of PLA puncta, independently of the size.

Figure 3.

Representative images of single PLA, dual PLA, and negative controls from human brain sections. Expression of D2R, A2AR (by single PLA) and A2AR-D2R (by dual PLA) was detected in the striatum (A-C). None or very few PLA puncta were detected in the negative controls that omitted the primary antibody in the single PLA (D-E) or anti-A2AR in the dual PLA (F). Scale bar, 25 μm. The arrows indicated representative PLA puncta. The images were obtained with an Olympus BX51 and a QImaging 1600X1200 pixel 8-bit color CCD camera controlled by Stereo Investigator (MBF Bioscience).

To analyze a large amount of PLA data from samples of high heterogeneity, such as human brains, it is important to sample the data with an unbiased and efficient approach. We recommend the systematic random sampling for whole slide scanning images (Basic Protocol 3), or manual counting of PLA puncta (with at least 20x lens) by with an Optical Fractionator in Stereo Investigator (MBF Bioscience, stereo Investigator). Systematic random sampling is a sampling method used in stereological procedure, in which samples are selected from a large population according to a random starting point but with a fixed, periodic interval. In this protocol, the cell size of the sampling grid determined the distance between two adjacent samples and the number of samples selected in a ROI. The selection of counting loci in the sampling area is random, but equally likely for all foci within an ROI. This provides a systematic pattern of sampling, as the distance between the counting loci is constant while the sampling areas of fixed size are evenly distributed throughout the ROI (West, 2012). A suitable adjustment of parameters, such as smaller but denser sampling areas for more heterogeneous specimen, is needed to obtain reliable and representable results.

The PLA results can be analyzed by puncta density measurement for relative quantification. For example, we used a machine learning-based algorithm (BOPSS) to quantify puncta density of each PLA assay (D2R, A2AR, and D2R-A2AR) automatically, and analyzed the density of PLA puncta and the ratio of dual PLA over single PLA (D2R-A2A/ A2AR, D2R-A2A/ D2R) (Zhu et al., 2019) for comparison of D2R-A2AR complexes. In addition to the absolute density of D2R-A2AR dual PLA puncta, the relative level of D2R-A2AR over A2AR (or D2R) provides information regarding the proportion of D2R-A2AR complexes, which might vary among individual subjects.

e. Time considerations

The time of specimen fixation varies from one day to one week, depending on the size of the tissue block. Specimen pretreatment and paraffin embedding take two days. It may take a couple of hours or longer to section one paraffin embedded tissue block, depending on the quality of specimen, the number of sections, and expertise. We recommend performing primary antibody incubation at 4 ^oC overnight. Therefore it takes 2 days and 5– 8 hours per day to perform PLA_BF assay on paraffin sections, depending on the number of slides.