Protocol for detection of glial complement expression in relation to amyloid plaques in mouse brain with combined FISH and IHC

Summary

Complement proteins contribute to neurodegeneration in Alzheimer’s disease (AD) and are secreted by glia surrounding beta-amyloid (Aβ) plaques. We present an optimized protocol for Aβ plaque detection with tyramide-digoxigenin signal amplification. This is combined with a multiplex mRNA fluorescence in situ hybridization (FISH) panel to assay glial-specific complement expression proximal to Aβ plaques in TauPS2APP mice. We describe steps for tissue preparation and mRNA detection. We then detail steps for the detection of Aβ plaques, image acquisition, and analysis.

Graphical abstract

Highlights

• •Detection of astrocytes and microglia-specific complement expression using mRNA FISH
• •Signal-amplified detection of Aβ plaques in the TauPS2APP mouse model of AD
• •Measurement of glial complement expression in relation to Aβ plaques using mRNA FISH
• •Workflow for whole-slide imaging and spatial gene expression analysis within a week

Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

Complement proteins contribute to neurodegeneration in Alzheimer’s disease (AD) and are secreted by glia surrounding beta-amyloid (Aβ) plaques. We present an optimized protocol for Aβ plaque detection with tyramide-digoxigenin signal amplification. This is combined with a multiplex mRNA fluorescence in situ hybridization (FISH) panel to assay glial-specific complement expression proximal to Aβ plaques in TauPS2APP mice. We describe steps for tissue preparation and mRNA detection. We then detail steps for the detection of Aβ plaques, image acquisition, and analysis.

Before you begin

Here, we present a protocol for optimized detection of complement expression in glial cells associated with Aβ plaques (Figures 2 and 3). This dual mRNA FISH and immunohistochemistry (IHC) protocol detects amyloid plaques via IHC, and glial expression of complement component mRNA (C4b and C1q) in TauPS2APP mouse brain hemisections. Astrocytes and microglia are detected using oligonucleotide probes for Slc1a3 and Tyrobp mRNA, respectively.

Figure 2.

Plaque associated astrocyte and microglia-specific C4b expression

(A) Hemisection from TauPS2APP mouse brain in 5-color fluorescence with astrocytes (Slc1a3, green), microglia (Tyrobp, red) and C4b (orange), and Aβ detected with 6E10 antibody (magenta) and nuclei counterstained with DAPI. (Inset) Cortical layer V/VI area with diffuse plaques varying in size, regions of interest (ROIs) indicated by boxes (1–2). Magnified ROIs show examples of C4b expression in individual astrocytes (arrowheads) and microglia (asterisks) associated with diffuse plaque signal. No digital processing was performed aside from standard adjustments to brightness and contrast. Scale bars: 250 μm and 20 μm.

Figure 3.

Plaque associated astrocyte and microglia-specific C1qa expression

(A) Hemisection from TauPS2APP mouse brain in 5-color with astrocytes (Slc1a3, green), microglia (Tyrobp, red) and C1qa (orange), and Aβ detected with 6E10 antibody (magenta) and nuclei counterstained with DAPI. (Inset) Cortical layer V/VI area with diffuse plaques varying in size, regions of interest (ROIs) indicated by boxes (1–2). Magnified ROIs show examples of absence of C1qa expression in individual astrocytes (arrowheads) and presence of C1qa puncta in microglia (asterisks) with diffuse plaque signal. No digital processing was performed aside from standard adjustments to brightness and contrast. Scale bars: 250 μm and 20 μm.

Creating and optimizing a dual mRNA FISH and IHC workflow with an imaging panel for 5-color detection can be challenging. mRNA FISH signal is improved by protease pretreatment, but this can degrade epitopes needed for antibody-based protein labeling. Direct fluorophore-conjugated secondary detection of Aβ post-FISH produces dim labeling of plaques that does not accurately capture their size and number. We employed tyramide-based amplification to improve antibody detection of plaques post-FISH and found that this improved the detection of diffuse plaques (Figures 1). We also describe spectral bleed-through present in dyes recommended by commercial assays and present an alternative with more desirable spectral properties.

Figure 1.

Signal amplification produces high signal-to-noise and sensitive detection of diffuse Aβ plaques in sequential FISH-IHC

Hemisections from TauPS2APP mice post-FISH for Slc1a3 (astrocyte marker, orange), with (A and B) immunostaining for Aβ plaques (white) with 6E10 primary antibody post-FISH using detection with Alexa 750-conjugated secondary antibody. Autofluorescence subtracted images from ROIs (boxes) show loss of diffuse plaque signal with fluorescence-conjugated secondary antibody (inset 1).

(C and D) Immunostaining with 6E10 primary antibody post-FISH using signal amplified with an anti-HRP conjugated secondary antibody and TSA-DIG, and detected with an anti-DIG Alexa 780-conjugated secondary antibody. Autofluorescence subtracted images from ROIs (boxes) show signal retention with the optimized TSA-DIG approach (inset 2). Diffuse plaque boundaries and intact fibrils are detected above tissue autofluorescence. Intensities adjusted to 0–3000 on a 16-bit scale for visualization. Scale bars: 1 mm (left panels) and 50 μm (insets).

We use a combination of readily available laboratory reagents and components from commercially available assays. This combined mRNA FISH (referred to as FISH hereafter) and IHC imaging panel could be applied to characterize the efficacy of complement-targeting therapies in AD models or study gene expression in concert with other protein aggregates relevant to Alzheimer’s disease pathology (e.g., tau, synuclein). For applications with other tissues, pretreatments, probes, antibodies, or imaging workflows some optimization may be required.

Institutional permissions

Any experiments on animal tissues must be performed in accordance with relevant institutional and national guidelines and regulations.

Harvest and processing of mouse brain tissue

Timing: 2–4 h

• 1.
  Harvest tissue from WT and TauPS2APP mice (12–22 months old).

  • a.
    Gather cold 1X PBS, crushed dry ice, wet ice, ethanol, blotting paper, paddle forceps, cryoprotectant and labeled cryomolds.
  • b.
    Prepare a slurry by combining dry ice and ethanol. Prepare anesthetic by dissolving 2,2,2-Tribromoethanol in Tert-amyl alcohol to make a 2.5% working solution per institutional animal care guidelines.
  • c.
    Administer anesthetic via intraperitoneal injection of working solution at 20 mL/kg.
  • d.
    Observe cardiac rhythm until mouse is under surgical plane of anesthesia. Confirm absence of pain response with a toe pinch.

    Note: Guidelines for usage concentration and dose may vary based on institutional IACUC animal welfare guidelines. The 2.5% Avertin solution must be used within 1 month and must be stored in the dark at 4°C.

  • e.
    Perfuse anesthetized mice with cold 1X PBS at 5 mL/min via transcardial perfusion. Do this in pairs (WT and TauPS2APP) and harvest whole brain tissue on ice.
  • f.
    Remove the brain and place a midsagittal cut using a fresh blade.
  • g.
    Place the hemibrains briefly on wet blotting paper to orient the tissue using paddle forceps.
  • h.
    Lower tissue into cryomold with cryoprotectant (Tissue-Tek O.C.T. Compound) filled to 0.5 cm height (approximately 1/3^rd of the size of the tissue).
  • i.
    Place the mold on the dry ice slurry while holding tissue in place until the tissue is frozen in place. This is apparent when the bottom surface of the mold is frozen and the tissue orientation is unaltered when paddle forceps are withdrawn.
  • j.
    Fill remaining volume with cryoprotectant to allow tissue block to freeze on a bed of dry ice. Store at −80°C.

    Note: TauPS2APP mice display frank amyloid pathology at ∼10 months of age and phenotypes progress with age, but acquire aging-related illnesses at 20 months. Since mice are lost to these illnesses at around 20 months, plan cohort sizes large enough to accommodate some loss.

    Note: Do not wash the tissue with PBS or any solvent prior to embedding. This interferes with the integrity of tissue embedding and causes sections to separate from embedding medium during sectioning.

    Note: Tissue blocks should be stored airtight containers and can be stored at −80°C for 1 year. No detectable differences were observed in experimental results during this period.

• 2.
  Section brain hemisections.

  • a.
    Gather blades and brushes inside cryostat and set the cryostat to −20°C.
  • b.
    Set a slide holder on a bed of dry ice to gather slide-mounted tissue as collected.
  • c.
    Use a blade to cut the mold and carefully release the tissue block.
  • d.
    Use the cryoprotectant to adhere tissue block to holder.
  • e.
    Wait to bring block to the internal cutting temperature of cryostat (∼15 min) before proceeding to next step.
  • f.
    Set the cutting thickness to 5–7 μm and place blade in holder.
  • g.
    Trim until tissue is visible and until desired region of interest is apparent.
  • h.
    Cut sections and collect with slide face-down, use a cold brush to orient the tissue before lifting with slide.
  • i.
    Collect slides in an airtight slide box on dry ice, store at −80°C until use.

Note: If cryostat temperature rises and the cutting chamber begins to warm, sections will not cut cleanly and will be difficult to mount.

CRITICAL: Slides should be at room temperature for sections to properly adhere.

CRITICAL: This step should be performed no more than 3 months prior to hybridization, ideally within the same month for maximal RNA preservation.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Purified anti-β-amyloid, 1–16 antibody, mouse IgG1, κ	BioLegend	Cat#SIG-39300
Anti-mouse IgG, HRP-linked antibody	Cell Signaling Technology	Cat#7076
Goat anti-mouse IgG (H + L) cross-adsorbed secondary antibody, Alexa Fluor 750	Thermo Scientific	Cat#A-21039
Chemicals, peptides, and recombinant proteins
PBS, pH 7.4	Gibco	Cat#10-010-023
Paraformaldehyde 16% aqueous solution EM grade	Electron Microscopy Sciences	Cat#15710-S
Tris hydrochloride (Tris HCl)	MilliporeSigma/Roche	Cat#10812846001
Tris base	MilliporeSigma	Cat#TRIS-RO
5 M NaCl	MilliporeSigma	Cat#S6546
Tween 20	MilliporeSigma	Cat#P1379
Low protein blocking buffer	eBioscience	Cat#00-4953-54
2,2,2-Tribromoethanol	MilliporeSigma	Cat#T48402
Ethyl alcohol, pure, 200 Proof	MilliporeSigma	Cat# E7023-4L
DAPI (4′,6-diamidino-2-phenylindole, dihydrochloride)	Thermo Scientific	Cat#62247
Trisodium citrate dihydrate	MilliporeSigma	Cat#106432
Sodium chloride	MilliporeSigma	Cat#567462
ProLong Gold antifade mountant	MilliporeSigma	Cat#P36930
Critical commercial assays
RNAscope Multiplex Fluorescent Reagent Kit v2	Advanced Cell Diagnostics	Cat#323100
RNAscope Probe - Mm-Slc1a3 - Mus musculus solute carrier family 1 (glial high affinity glutamate transporter) member 3 (Slc1a3) mRNA	Advanced Cell Diagnostics	Cat#430781
RNAscope Probe - Mm-Tyrobp-C2 - Mus musculus TYRO protein tyrosine kinase binding protein (Tyrobp), mRNA	Advanced Cell Diagnostics	Cat#408191-C2
RNAscope Probe - Mm-Tyrobp-C3 - Mus musculus TYRO protein tyrosine kinase binding protein (Tyrobp), mRNA	Advanced Cell Diagnostics	Cat#408191-C3
RNAscope Probe- Mm-C4b	Advanced Cell Diagnostics	Cat#445161
RNAscope Probe - Mm-C1qa-C2 - Mus musculus complement component 1 q subcomponent alpha polypeptide (C1qa) mRNA	Advanced Cell Diagnostics	Cat#441221-C2
RNAscope 3-plex Negative Control Probe	Advanced Cell Diagnostics	Cat#320871
RNAscope Probe Diluent	Advanced Cell Diagnostics	Cat#300041
SSC (10X)	Roche Diagnostics	Cat#950-110
Opal 520 Reagent Pack	Akoya Biosciences	Cat#FP1487001KT
Opal 570 Reagent Pack	Akoya Biosciences	Cat#FP1488001KT
Opal 690 Reagent Pack	Akoya Biosciences	Cat#FP1497001KT
TSA Cyanine 5 Reagent Pack	Akoya Biosciences	Cat#NEL705A001KT
Opal 780 Reagent Pack	Akoya Biosciences	Cat#FP1501001KT
Tyramide Amplification Buffer Plus	Biotium	Cat#22029-T
Experimental models: Organisms/strains
TauPS2APP (APP K670_M671delinsNL (Swedish), MAPT P301L, PSEN2 N141I), C57BL/6	F. Hoffmann-La Roche Ltd	Male and female brains, 10–12 months old
Software and algorithms
VS200 ASW V3.4.1	Olympus Scientific Solutions/Evident	https://www.olympus-lifescience.com/en/downloads/detail-iframe/?0[downloads][id]=847253921
Qupath	QuPath	https://qupath.github.io/
Other
Peel-A-Way Disposable histology molds	Ted Pella, Inc.	Cat#27118
Tissue-Tek O.C.T. Compound	Sakura	Cat#4583
Tissue-Tek Accu-Edge low profile blades	Sakura	Cat#4980
Leica CM3050 S Cryostat	Leica	Cat#CM3050S
Camel hair brushes, #1, less than 1.59 mm wide	Ted Pella, Inc.	Cat#11859
Camel hair brushes, #2, 1.59 mm wide	Ted Pella, Inc.	Cat#11860
PELCO Histo brush	Ted Pella, Inc.	Cat#11872
Camel hair brush, 1.91 cm (3/4″) W x 3.81 cm (1-1/2″)	Ted Pella, Inc.	Cat#11873
Electron Microscopy Sciences solvent resistant pen	Electron Microscopy Sciences	Cat#62053R
Binocular bright-field microscope	Thomas Scientific	Cat#20A00G368
Precision General purpose water baths	Thermo Scientific	Cat#1184L86
Fisherbrand Superfrost Plus microscope slides	Fisher Scientific	Cat#12-550-15
Blotting paper	VWR	Cat#6067600
Tissue-Tek vertical slide rack	Fisher Scientific	Cat#NC9837976
Tissue-Tek slide staining dish	Fisher Scientific	Cat#NC0731403
Vector Laboratories ImmEDGE hydrophobic barrier	Fisher Scientific	Cat#NC9545623
ACD HybEZ II Hybridization System (220v)With ACD EZ-Batch Slide System	Advanced Cell Diagnostics	Cat#321720
VS200 slide scanner	Olympus	N/A
X-Cite XYLIS LED Illumination System	Excelitas	Cat#XT720S
LED-DAPI-A	N/A	N/A
ORCA-Fusion Digital CMOS camera	Hamamatsu	Cat#C14440-20UP
LED-FITC-A	N/A	N/A
LED-TRITC-A	N/A	N/A
LED-Cy5-A	N/A	N/A
LED-Cy7-A	N/A	N/A
PC Workstation (minimum 128 GB RAM)	N/A	N/A

Materials and equipment

• •
  4% Paraformaldehyde (PFA) solution.

  • ○
    Dilute 16% PFA at 1:3 in 1XPBS.
  • ○
    Chill 200 mL 4% PFA to 4°C.

CRITICAL: Paraformaldehyde is toxic and a potential carcinogen. Wear gloves while handling container and uncap the bottle only under fume hood. Dispose the liquid and containers in a chemical fume hood following institutional procedures for hazardous chemical waste disposal.

• •
  Ethanol dilution.

  • ○
    Make 90%, 50%, and 70% Ethanol dilutions in MilliQ water.
  • ○
    Cap tightly to avoid evaporation loss.
• •20X SSC Wash Buffer.

Reagent	Final concentration (M)	Amount
Trisodium citrate dihydrate	0.3	88.20 g
NaCl	3	175.30 g
Milli-Q Water	N/A	1.00 L
Total	N/A	1.00 L

Store at RT for up to 6 months.

• •
  1X SSC Wash buffer dilution.

  • ○
    Dilute 20X SSC to 1X in Milli-Q water.
  • ○
    Use dilution for up to 1 month.

Note: We found preparing this wash buffer to be cost-effective. Alternatively, commercially available 10X wash buffer can be used to prepare a 1X working solution and used instead (See key resources table).

• •
  RNAscope Probe dilutions.

  • ○
    Warm probes at in a water bath at 40°C.
  • ○
    Dilute probes in probe diluent or C1 probe per RNAScope Multiplex V2 Manual.

Note: C1 probes are sold at the working dilution. C2, C3 and C4 probes are sold at 50x. If the probe is in C2, C3 or C4 channel, dilute the probe in Probe Diluent for use in a 1-plex experiment. In a multiplex experiment, C2, C3, C4 probes are diluted in the C1 probe.

• •
  Opal dye dilution for RNAscope Probes.

  • ○
    The dyes are contained within the respective reagent packs listed in the key resources table. Dilute each dye in Tyramide Amplification Buffer Plus at concentrations listed below.

    • -
      Opal 520: 1:750.
    • -
      Opal 570: 1:750.
    • -
      Tyramide Cyanine 5 : 1:3000.

Note: These concentrations will need to be optimized for every probe-dye pair based on the expected and observed mRNA expression level for optimal image acquisition, with higher expressors typically requiring lower Opal dye concentration. We tested concentrations in the range of 1:300 to 1:5000.

• •1 M Tris, pH 8.0.

Reagent	Final concentration (M)	Amount
Tris HCl	0.56345	88.80 g
Tris Base	0.43751	53.00 g
Milli-Q Water	N/A	1.00 L
Total	N/A	1.00 L

After Tris HCl and Tris base are added, Milli-Q water is added to dissolve and solution is adjusted to pH 8.0 with HCl. Store at RT for up to 6 months.

• •1X TBST Wash Buffer.

Reagent	Final concentration	Amount
1 M Tris, pH 8.0	10 mM	10.00 mL
Milli-Q Water	N/A	959.00 mL
5 M NaCl	0.15 M	30.00 mL
Tween 20	0.1%	1.00 mL
Total	N/A	1.00 L

Store at RT for up to 6 months.

• •
  anti-β-Amyloid, 1–16 Antibody solution.

  • ○
    Dilute at 1:1500 in Low Protein Blocking Buffer, prepare ∼750 μl for 8–10 slides.
  • ○
    Use diluted antibody for up to 1 month.
• •
  Anti-mouse IgG, HRP-linked Antibody.

  • ○
    Dilute anti-mouse HRP-linked antibody at 1:1000 in Low Protein Blocking Buffer, prepare ∼1 mL for 8–10 slides. Use dilution within 1 week.
• •
  Opal TSA-DIG reagent solution.

  • ○
    Dilute in Opal TSA-DIG (from the Opal 780 Reagent Pack) in Tyramide Amplification Buffer Plus.
  • ○
    Make fresh 30 min prior to use and discard unused dilution.
• •
  Opal Polaris 780 reagent solution.

  • ○
    Dilute Opal Polaris 780 antibody at 1:300 in Low Protein Blocking Buffer.
  • ○
    The concentration can be increased to 1:100 for very low abundance proteins.
• •
  DAPI solution.

  • ○
    Dilute DAPI in 1XPBS at 1:1000. Use the same day and discard unused dilution.
• •
  Olympus VS200 (and alternatives).

  • ○
    This protocol uses the Olympus VS200 slide scanner but alternative whole-slide systems can be used instead (Zeiss Axio Scan Z.1/Zeiss Axioscan 7, Hamamatsu S60) and wide field systems (Leica Thunder) as alternatives). However, the VS200 was chosen for its user friendly UI, sensitive ORCA-Fusion Digital CMOS camera, and efficient automatic workflow (faster scan times by approximately 4x within this workflow) and scalability (up to 210 slides per imaging session).
  • ○
    The standalone slide scanner was equipped with the X-Cite XYLIS XT720S light source, and filter sets with the following specifications: LED-DAPI-A Ex 392/23, Em 447/60, LED-FITC-A Ex 474/27, Em 525/45, LED TRITC-A Ex 554/23, Em 609/54, LED-Cy5-A Ex 635/18, Em 680/42, LED-Cy7-A Ex 735/28 , Em 809/81.
• •
  Qupath (and alternatives).

  • ○
    This protocol describes the use of Qupath to segment nuclei, cells and amyloid plaques. Other options include cell segmentation tools like Cell Profiler or custom end-to-end scripting with Image J. However, Qupath offers the advantage of built-in scripting to annotate nuclei, cells and mRNA puncta. More importantly, it allows easy visualization of detections across whole hemisections to confirm the accuracy of segmentation attempts during optimization.

Step-by-step method details

Tissue pretreatment

Timing: 1.5 h

This step involves protein cross-linking, delineating tissue sections with a hydrophobic barrier followed by quenching of autofluorescence, and protease digestion to facilitate optimal probe hybridization.

• 1.
  Perform protease and hydrogen peroxide treatments per the RNAScope Multiplex V2 Manual. The steps are described in brief below:

  • a.
    Remove slides from −80°C, and place in a Tissue Tek Vertical Slide Rack.
  • b.
    Immediately immerse slides in the pre-chilled 4% PFA at 4°C.
  • c.
    Rinse the slides in 1X PBS twice to remove fixative.
  • d.
    Serially immerse in slides in 50%, 70% and 100% EtOH for 2 min.
  • e.
    Repeat the immersion in 100% EtOH.
  • f.
    Dry slides for 5 min, draw a barrier around tissue sections and dry for another 5 min.
  • g.
    Place a wet blotting paper in the slide tray to create a humidified environment.
  • h.
    Pretreat the tissue with H₂O₂ at RT for 10 min and then wash twice with MilliQ water.
  • i.
    Incubate with Protease IV for 20 min at RT and wash with 1X PBS water twice for 2 min.

Note: Turn on the hybridization oven at the beginning of this section and set to 40°C.

Note: Place hybridization, amplification and HRP detection reagent at room temperature.

Note: Thin fresh frozen tissue sections are susceptible to protease overdigestion. See Troubleshooting 1.

Note: Maintain 0.5 cm distance around the tissue section when drawing the hydrophobic barrier. Do not touch the tissue section.

Probe and dye detection optimization

Timing: 2 days

This section describes the necessary controls to test the compatibility of spectral emission of the fluorescent dyes with microscope detection filters before running the multiplexed dual FISH-IHC experiment. Unlike confocal microscopes, wide-field microscopes used to rapidly acquire whole-slide images do not allow adjustments to detection bandwidth making bleed-through in adjacent channels a possibility.

• 2.Test each probe on its own in a single-channel probe test. Based on expected relative expression of the probes, perform 1-plex FISH with each probe per the RNAscope Multiplex V2 Manual.

Note: We found it useful to test probes paired with Opal 570, since this detection bandwidth and fluorophore offers good signal-to-noise ratio. Based on the relative expression of each probe, we paired it with a fluorophore for the multiplexed FISH experiment.

Note: If you expect the gene to be a high expressor, pair it with Opal 520 to detect expression. Tissue autofluorescence is highest in the FITC emission spectra and accumulates even in samples stored at 4°C, making it less suitable for detection of low-moderate expressors. For genes with intermediate or low expression, use Opal 570 and TSA Cyanine 5 to detect probes. See Troubleshooting 3.

• 3.Evaluate the signal in the all emission bandwidths for each dye. Ensure that the signal over noise captures the dynamic range of mRNA expression i.e., individual puncta should be visible after background subtraction (Figures 2 and 3 and Image acquisition).

Note: We found that the Opal dye 690 recommended within the commercial assays had significant bleed through in bandwidths for Cy7 and adjacent TRITC and FITC (Figure 4). We tested several commercially available cyanine 5 dyes (Opal 690 Reagent Pack, cat# FP1497001KT, Akoya; CF660R, Biotium cat# 92195; CF680R, Biotium cat#92196, Styramide 670, AAT Bioquest cat# 45047) and found TSA Cyanine 5 is a suitable alternative. We found that bleed through in most dyes used increased with target mRNA expression and dye concentration. See Troubleshooting 4.

Note: Some probes do not work well when placed in FITC or Cy5 detection bandwidths. In FITC, tissue autofluorescence competes with signal from FISH, producing background that is challenging to distinguish from true signal. Cy5 detection filters typically require higher gain and/or exposure, and may detect probes with sparse signal or may increase overall imaging time significantly. In our experiments, Opal 570 had consistent performance with all probes.

• 4.Proceed with the 3-plex FISH and sequential IHC for a set of slides from biological replicates and genotypes in use.

Figure 4.

Reduced bleed-through emission in FITC with use of TSA Cyanine 5

Single-plex FISH (A) detected with TSA Cyanine 5 and (B) detected with Opal 690 with fluorescence captured with Cy5 640/690 filter sets. Opal 690 has significant bleed through in Cy7, TRITC and FITC detection wavelengths. Intensities adjusted to 0–2000 on a 16-bit scale for visualization. Scale bars: 50 μm.

Probe hybridization and detection

Timing: 7 h

The steps below include hybridization of probes to target mRNA, amplification to create branching and sequential HRP-dependent deposition of dyes to detect probes. The manufacturer’s protocol below is used and probe-dye combinations and dye concentrations have been carefully optimized (See Figure 1).

RNAscope Multiplex V2 Manual: https://acdbio.com/system/files_force/UM%20323100%20Multiplex%20Fluorescent%20v2_RevB.pdf.

RNAscope Multiplex V2 Workflow: https://acdbio.com/ebook/introduction/rnascope-workflow.

• 5.Perform the hybridization and amplification steps in the ACD HybEZ II Hybridization Oven per RNAscope Multiplex V2 Manual. Rinse slides with 1X SSC Wash Buffer for 2 min. Repeat the rinse.

Note: The incubation time for hybridization is 2 h, and the chamber tends to dry out if moisture is not added during the incubation. Add ∼5 mL every 30 min.

• 6.Develop HRP-C1 with Opal 520, followed by a reaction termination step with the HRP Blocker using the steps in ACD HybEZ II Hybridization Oven per RNAscope Multiplex V2 Manual.
• 7.Wash slides with 1X SSC for 2 min and repeat twice between each step.
• 8.Similarly develop HRP-C2 and HRP-C3 signal sequentially with Opal 570 and TSA Cyanine 5 thereafter.

Note: Use the dye dilutions listed in materials and equipment. Each signal development step takes approximately 1 h and entails addition of HRP for the specific probe channel, addition of the dye and blocking the HRP-dependent deposition of dye with HRP blocker.

• 9.Rinse slides with 1X SSC Wash Buffer for 2 min. Repeat the rinse.
• 10.If performing IHC post-FISH do not counterstain the slides and do not add DAPI. Instead, proceed with the next section for immunohistochemistry. If only performing in situ hybridization, proceed to Step 13 to counterstain and mount.

Immunohistochemistry

Timing: 4 h

This section describes the detection of plaques with TSA-Digoxigenin (TSA-DIG) mediated signal amplification. The TSA-DIG reagent uses horseradish peroxidase to catalyze the deposition of digoxigenin (DIG) directly adjacent to the immobilized enzyme. The deposited labels are detected using anti-digoxigenin antibody conjugates like the Opal Polaris 780 reagent. The primary antibody used here can be substituted with antibodies that are compatible with protease pretreatment.

• 11.
  Block tissue sections.

  • a.
    After the last probe signal development step, wash slides twice for 2 min in 1X TBST Wash Buffer.
  • b.
    Return slides to humidified chamber and add ∼50 μL Low Protein Blocking Buffer per section and incubate for 30 min at RT. Do not rinse.

Pause point: It is possible to leave sections in blocking buffer at 4°C for 12–16 h stored in the dark.

• 12.
  Detect target protein.

  • a.
    Add anti-Aβ primary Antibody solution (see materials and equipment) for 45 min at RT.
  • b.
    Rinse slides with 1X TBST Wash Buffer for 2 min. Repeat the rinse three times.
  • c.
    Add HRP-conjugated secondary antibody for 30 min at RT.
  • d.
    Rinse slides with 1X TBST Wash Buffer for 2 min. Repeat the rinse three times.
• 13.
  Amplify primary antibody signal.

  • a.
    Add TSA-DIG diluted in TSA Amplification Buffer for 30 min at RT.

    Note: Other primary antibodies of interest can be combined with a HRP-secondary, followed by TSA-DIG and anti-DIG dye for a two-step amplification as described here (see Figure 1).

  • b.
    Rinse slides with 1X TBST Wash Buffer for 2 min. Repeat the rinse three times.
  • c.
    Add 100 μL HRP Blocker per section and incubate for 15 min at 40°C.
  • d.
    Rinse slides with 1X TBST Wash Buffer for 2 min. Repeat the rinse three times.
  • e.
    Add diluted Opal Polaris 780 reagent for 30 min at RT.
  • f.
    Rinse with 1X TBST for 2 min. Repeat three times.
• 14.
  Counter stain nuclei and mount samples.

  • a.
    Remove excess liquid and add diluted DAPI (see materials and equipment) to each slide. Incubate for 5 min at RT.
  • b.
    Remove DAPI rinse quickly with 1X TBST to remove excess reagent.
  • c.
    Place slides on absorbent paper and allow excess liquid to evaporate (for <5 min). Monitor carefully to prevent sections from drying.

    • i.
      Immediately add 50 μL Prolong Gold to each section and mount with a glass coverslip.
    • ii.
      Use one side of the slide as an anchor and slowly lower the coverslip to allow any air pockets to move towards the edge before releasing the coverslip onto the slide.
• 15.Wait 8–12 h before imaging to allow the mounting media to set. Store slides at 4°C in the dark and image within 1 week.

Whole-slide image acquisition

Timing: 35 min (per slide)

This section describes the whole-slide imaging workflow for optimal detection of FISH signal.

• 16.
  Image acquisition setup.

  • a.
    Start VS200 acquisition software and click on ‘Exchange Trays’ to load slides into the instrument.
  • b.
    Load single-channel control slides into the slide tray and place into the instrument loader.
  • c.
    Click on ‘Single Scan’ mode and choose a default scan profile in the ‘Fluorescence’ tab to set up imaging profile.
  • d.
    In the ‘Overview Type’ within the Overview tab, set Observation Type to ‘Fluorescence’ and the Observation Method to ‘DAPI’ using the lowest magnification objective (2x).

    Note: Default exposure time is 100.00 ms. Adjust exposure based on results from Probe and dye detection optimization.

  • e.
    Click on ‘Start Scan’ to image an overview of the current slide using the selected objective.

    Note: Increase or decrease the exposure time in the Overview tab if the overview image of tissue is too bright or dim to avoid issues with automatic tissue detection.

  • f.
    In the Detail tab, select the desired objective (20X).
  • g.
    Choose ‘Virtual-z’ under ‘z-plane’ to acquire z-stacks.
  • h.
    Set z-range and set Manual z-spacing as desired.

    Note: In our experiments, z-range of 5.00 μm and Manual z-spacing of 0.50 μm worked best for 5–7 μm sections. This will need adjustment for thicker tissue sections.

  • i.
    In the Channels tab, select ‘Add FL Channel’ and ‘Remove Channel’ to create a list of channels to acquire in order (DAPI, FITC, TRITC, CY5, CY7).

    Note: The first channel listed is used to focus the image.

  • j.
    Manually select and tune exposure times for each channel by clicking the channel in the channel list and selecting the ‘Manual Exposure’ button.
  • k.
    Press the ‘Start Live’ button in the Exposure tab and click the overview image to move to an area with signal.
  • l.
    Check the Histogram to adjust the exposure time so that signal is well represented within the 16-bit range.

    Note: Set exposure time so that the peak of histogram is at approximately ⅓ of the range to avoid over-saturation and bleaching.

    Note: Best practice when optimizing exposure times is to image a region of the tissue with known positive signal and manually inspect the pixel intensities that are the brightest and dimmest. This should be repeated in areas of highest and lowest expression within the section. Do this in the Additional Layouts -> Image Processing tab. Zoom in and check pixel intensity per channel. Pixel intensity is shown in the bottom right corner with channel and pixel intensity in 16-bit range (0–65,535). Adjust exposure time to capture the dynamic range of the signal.

• 17.
  Semi-automatic focusing and image acquisition.

  • a.
    In the Focusing tab, select ‘Semi-automatic’ or ‘Manual’ button in Focus mode to adjust the focus within the displayed area during acquisition.
  • b.
    Choose the directory to save image files and format of image name in the Naming and Saving tab.
  • c.
    Then, click on the ‘Save Scan Project’ on the upper right panel and save the scan project.
  • d.
    Acquire all channels on the single channel control slides to check for bleed through.

    CRITICAL: This is a critical step to ensure fluorescence signal with a given dye and dye concentration is not present in other bandwidths. This allows for true detection of colocalized signal in a multiplexed imaging experiment.

  • e.
    Press the ‘Detail’ button on the top panel and adjust the ROI and focus boxes. Press ‘Start Scan’ to acquire the image.

    Note: In semi-automatic mode, software will calculate the most optimal focusing plane for each focus box.

  • f.
    Manually adjust the ‘z’ value so that each focus box is in focus and press ‘Start scan’ again.
  • g.
    After scan is complete, save the image file.
  • h.
    Repeat steps with the control and experiment slides using the same parameters/scan profile.
  • i.
    If focus and exposure are optimal, go to Batch Scan in the main menu, load the scan profile and press ‘Start Scan’ to start generating image overviews.
  • j.
    Adjust the ROIs and focus box in the Details tab and press ‘Start Scan’ again to acquire 20X whole-slide images.

    Note: Tiling artifacts may appear if there are variations in signal throughout the section, or when signal-to-noise is low. See Troubleshooting 5.

    Note: As a general note on applying this workflow to other instrumentation, most whole slide imagers run on similar workflows that begin with the acquisition of an overview image. Image parameters (exposure time and focal plane) are adjusted based on histogram or pixel intensity. Small ROIs are imaged to observe the results of chosen settings. Then the entire specimen is imaged. Minor differences between slide scanners include the use of darkfield to image the overview as well as a few imaging parameters and UI differences. However these differences do not produce significant difference in workflows.

Expected outcomes

Recent studies implicate complement genes in the pathology of Alzheimer’s disease (AD) and other neurological diseases.^1,2,3 Activated microglia and astrocytes increase complement gene expression in AD disease models. Since complement factors are secreted, the use of RNA FISH to detect complement expression allows spatial resolution of cells that express complement. In particular, an assay that detects cell-type specific complement expression associated with plaques can be used to observe pathology and read-out effects from treatment targeting complement or other therapeutic targets.

Our five-color protocol offers high-resolution detection of 3-plex RNA FISH in concert with signal-amplified detection of Aβ plaques and nuclear detection with DAPI. The high signal-to-noise detection offers a wide dynamic range of that allows observation of small, diffuse and dense Aβ plaques (Figure 1). We apply the workflow to TauPS2APP hemisections, and show C4b and C1qa expression in plaque-associated astrocytes and microglia (Figures 2 and 3). The protocol includes analysis of complement expression in relation to Aβ deposits. The assay can easily be adapted to human brain tissue, retinal tissue or to detect other pathological deposits relevant to neurodegeneration.

Quantification and statistical analysis

Timing: 10 min (per image)

This section describes annotation of images, detection of plaques, nuclei and cells, and detection of FISH signal. The free software QuPath⁴ is available here: https://qupath.github.io/.

• 1.Open QuPath and click on ‘Create Project’ in the left panel image. Select all .vsi images to be analyzed.
• 2.Drag and drop images into project, select ‘Bio-formats’ under file format, select ‘Fluorescence’ for file type. Check ‘Yes’ to auto generate pyramids.
• 3.Adjust Channels LUTs using View>Brightness/Contrast, and save the LUT profile to apply to the rest of the images within the project.
• 4.
  Raise the minimum and lower maximum gray values in each channel in the LUT until clear but unsaturated signal is visible.

  • a.
    Note the highest background gray value for each channel. This is to view the images only and does not alter the data.
  • b.
    Verify this by changing the thresholds in the LUT to this value and visually confirming that background signal is eliminated.

Note: For our experiments, the background gray values in each channel were as follows: DAPI: 500, FITC: 5000, TRITC: 3000, Cy5: 1000, Cy7: 500.

• 5.Annotate region of interest using a geometric tool (Polygon/Rectangle/Circle).

Note: Rename annotations to easily identify ROIs and subset data after data export.

• 6.
  Determine nucleus, cell and spot detection parameters empirically.

  Note: For each of the objects (plaques, nuclei, cells and mRNA spots), some trial-and-error may be required to accurately detect signal above background throughout the tissue section.

  • a.
    To detect plaques:

    • i.
      Go to ‘Annotations’ in the left panel and click on the three vertical dots next to ‘Auto Set’. Add/Remove> Add Class ‘Plaques’.
    • ii.
      Create a pixel classifier that identifies plaque objects and places them within parent annotation hierarchy.
    • iii.
      Go to Classify>Pixel Classification>Create Thresholder.
    • iv.
      Save as ‘Identify Plaques’.

      Classifier parameter	Value
      Resolution	Full
      Channel	Cy7
      Filter	Gaussian
      Smoothing Sigma	0
      Threshold	1500
      Above threshold	Plaques
      Below Threshold	Ignore
      Region	Any annotation ROI

  • b.
    This takes FISH signal in all channels and creates detections at the same hierarchical level as plaque objects.
  • c.
    Open Automated Analysis>Script Editor and paste the script below in the open untitled script file.

    Note: The code below works for the paradigm described in our protocol. Additional analysis will require alterations to this code.

  • d.
    Export Detection Measurements.

    Note: Since all detections have a centroid, this approach makes distance from plaque analysis possible. The distances to plaques appear under a column titled ‘distance to Plaques (2D)’ within the detection measurements data table.

    Note: QuPath accepts .vsi z-stacks and single-plane images generated by Olympus microscopes (see Supported Image Formats for more information: https://qupath.readthedocs.io/en/stable/docs/intro/formats.html).

    Note: All input parameters for nuclei, cell and spot detection require user input. The recommended parameters work well for the tissue and detection reagents used in this protocol but optimization may be necessary for optimal segmentation in other workflows.

    // This code sequentially creates cell objects, detects spots and creates annotations for plaques

    // Before running this code, create subfolders named 'annotation_result' and ‘detection_result’ within the project directory

    setImageType('FLUORESCENCE');

    selectAnnotations();

    // Change the channel display color

    getCurrentImageData().removeProperty('qupath.lib.display.ImageDisplay')

    setChannelColors(getColorRGB(0, 0, 255), getColorRGB(0, 255, 0), getColorRGB(255, 170, 100), getColorRGB(255, 0, 0), getColorRGB(255, 0, 255))

    //Set channel names

    setChannelNames("DAPI","FITC","TRITC","CY5","CY7")

    // Run cell and spot detection

    runPlugin('qupath.imagej.detect.cells.WatershedCellDetection', '{"detectionImage":"DAPI","requestedPixelSizeMicrons":0.5,"backgroundRadiusMicrons":50.0,"backgroundByReconstruction":true,"medianRadiusMicrons":0.0,"sigmaMicrons":1.5,"minAreaMicrons":6.25,"maxAreaMicrons":400.0,"threshold":500.0,"watershedPostProcess":true,"cellExpansionMicrons":5.0,"includeNuclei":true,"smoothBoundaries":true,"makeMeasurements":true}')

    runPlugin('qupath.imagej.detect.cells.SubcellularDetection', '{"detection[Channel 1]":-1.0,"detection[Channel 2]":5000.0,"detection[Channel 3]":3000.0,"detection[Channel 4]":1000.0,"detection[Channel 5]":-1.0,"doSmoothing":false,"splitByIntensity":true,"splitByShape":true,"spotSizeMicrons":3.0,"minSpotSizeMicrons":1.0,"maxSpotSizeMicrons":5.0,"includeClusters":true}')

    // Run plaque pixel classifier

    createAnnotationsFromPixelClassifier("Identify plaques", 1.0, 1.0, "SPLIT")

    detectionToAnnotationDistances(true)

    // Export results - Annotations

    def name = getProjectEntry().getImageName() + '.txt' // defines var.name to extract current image name and create a .txt files with it

    def pathAnnotation = buildFilePath(PROJECT_BASE_DIR, 'annotation_result') // defines var. pathAnnotation as filepath (project base directory appended with annotation.results)

    mkdirs(pathAnnotation) // ensures file path exists

    pathAnnotation = buildFilePath(pathAnnotation, name) // layers the variable pathAnnotation by appending the current image name.text

    saveAnnotationMeasurements(pathAnnotation) // saves annotation measurements to file path

    print 'Results exported to ' + pathAnnotation // prints results exported filepath

    // Export results - Detections

    def pathDetection = buildFilePath(PROJECT_BASE_DIR, 'detection_result') mkdirs(pathDetection) // ensures file path exists

    pathDetection = buildFilePath(pathDetection, name) // appends the image file name to the path detection file path

    saveDetectionMeasurements(pathDetection) // saves detection measurements

    print 'Results exported to ' + pathDetection // prints results export to filepath

Limitations

Detection of Aβ plaques post-FISH yields consistent results using this protocol, with minimal variability between sections. However, detection of low abundance protein or sparse signal may be more susceptible to variation due to the protease pretreatment. This may require some optimization, including reduction of protease treatment duration, or increase in concentration of antibody detection reagents. If the target epitope is sensitive to cleavage by proteases, trial-and-error with several candidate antibodies may be a useful approach. However, one advantage in this paradigm is that detection of antibody signal is placed in the near-IR detection wavelength, and tissue autofluorescence is low in this bandwidth.⁵ Moreover, this protocol detects 5 colors with serial enzymatic amplification and fluorescent labeling, but other situ hybridization techniques that allow serial probing and quenching allowing for the detection of several hundred RNA species per cell.^6,7

Troubleshooting

Problem 1: Low FISH signal and loss of nuclear integrity

Protease over digestion can interfere with tissue morphology and lower detection of FISH signal (See tissue pretreatment, step 1).

Potential solution

• •Uneven section thickness: Thinner areas of tissue undergo more protease digestion. This can be circumvented by ensuring the blade is tightly secured and sectioning all samples to be stained in one session to reduce experimental variability.
• •Duration of digestion: Lowering the protease digestion duration by 5–10 min increments and assessing the effect on a low expressor probe to verify probe detection is unaltered.
• •Tissue thickness: Section tissue at 5, 10, 20 μm to compare the effects of protease digestion on antibody detection.
• •Protease type: The RNAscope Multiplex Fluorescent Reagent Kit v2 also contains an alternative milder protease Protease III. Pilot experiments with Proteinase III may be useful for tissues where Protease IV is suboptimal. Alternatively, Pepsin^8,9 can be useful to test, which is milder than the more commonly used protease in FISH, Proteinase K.

Problem 2: Low or uneven antibody signal

Antibodies targeting epitopes that are sensitive to protease digestion may produce uneven signal throughout the tissue section (See tissue pretreatment, step 1, and immunohistochemistry, step 11)

Potential solution

• •Increase antibody concentration: This approach may be effective for detecting protein aggregates or for target proteins that are highly expressed.
• •Choice of protease: Protease IV is stronger than Protease III. Switch to Protease III and test for 15 min and 30 min incubation times. Dilute Protease IV 1:1 in 1X PBS for and test 15, 20 min and 30 min incubations.
• •Choice of antibody: Test several antibodies in parallel post-FISH for susceptibility to protease digestion. Trial-and-error may produce better results if alternative epitopes are well preserved despite protease digestion.

Problem 3: FISH probe signal not detected

Low expressor mRNA signal may be difficult to detect (see probe hybridization and detection, step 1).

Potential solution

• •Evaluate if each of the probes work in a 1-plex FISH experiment compared with a negative control probe with any of available dyes and detection bandwidths (FITC, Cy3 or Cy5). Probe detection can vary based on the dye used for fluorescent detection.
• •Increase concentration of the fluorescent dye. For some low expressor mRNA targets, dye concentration of 1:250:1:500 is necessary to detect any signal.

Problem 4: Bleed through from fluorescent dyes

Commercially available Cyanine 5 dyes are often modified for improved quantum yield and photostability, some far-red cyanine dyes can be detected in the lower detection wavelengths with FITC/green filter sets.¹⁰ The Opal 690 recommended for use with the RNAscope Multiplex Fluorescent Reagent Kit v2 (Cat#323100) has significant bleed through in Cy7 and some bleed through in the Cy3 and Cy5 channels (Figure 4). TSA Cyanine 5 is a good alternative, and other cyanine dyes that have lower quantum yield but almost no bleed through are other options (see probe hybridization and detection, steps 5–7).

Potential solution

• •Lower Cyanine 5 dye concentration: Preliminary testing of the TSA Cyanine 5 in a single-channel control with dilutions of dye ranging 1:1500–1:6000. These slides should be imaged at exposures ranging from 25–200 ms to test FITC emission and Cy5 emission. The Cy5 dye concentration and FITC exposure should be selected at which bleed through in the FITC channel is below detected background signal.
• •Increase FITC dye concentration and lower FITC exposure: Produce samples labeled with FITC at 1:300–1:750 to create a FITC single channel control. Exposure time for FITC should be selected based on signal captured above background and the outcome in the preceding step.
• •Alternative far-red dyes: For high-expressor probes, alternative cyanine dyes with lower quantum yield and spillover into other bandwidths may be desirable.

Problem 5: Tiling artifacts create non-uniform imaging

Image artifacts such as image tiling are caused by uneven illumination of the sample across large imaging areas. Uneven or low signal intensity, which is common in in situ hybridization can worsen tiling artifacts and reduce quality of downstream analyses (see whole-slide image acquisition, step 15).

Potential solution

• •
  Shading correction is used to correct for tiling and should be performed before the whole-slide images are acquired. The instructions below can be applied within the VS200 acquisition software.

  • ○
    Start at Single Scan mode and select the Fluorescence scan profile used. Load sample with uniform signal for each respective channel.
  • ○
    In the Detail tab, select the channels for which corrected images will be acquired.
  • ○
    Click the ‘Calibrate Shading Correction’ button under the list of channels and the Shading Correction dialogue box will appear.
  • ○
    Select the check box ‘Skip acquisition of the dark current correction image’ since this correction image is acquired during instrument start-up and only needs to be acquired once.
  • ○
    Select the objective needed for calibration (20X) and select ‘Special’ in the Calibration method and ‘Next.’
  • ○
    Move field of view (FOV) in live mode to an area with uniform signal that fills up all four corners of the FOV and acquire flat field correction.
  • ○
    It is recommended that shading corrections are repeated when a change is made to the light path or sample, or if tiling recurs in the acquired images.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Melissa Gonzalez Edick (melisago@gene.com).

Technical contact

Information and resource requests should be directed to the technical contact, Sadhna Rao (raos27@gene.com).

Materials availability

All material is freely available upon request.

Data and code availability

The protocol includes all analyzed data shown in Figures 1, 2, 3, and 4.

Acknowledgments

We would like to thank Tiffany Wu, Miriam Baca, and Digital Pathology Imaging, Genentech, for technical support. The graphical abstract was created with BioRender.com.

Author contributions

Conceptualization, S.R. and J.E.H.; investigation, S.R. and J.Y.; writing – original draft, S.R. and J.Y.; writing – review and editing, S.R., J.Y., and J.E.H.

Declaration of interests

The authors declare no competing interests.