Bex-Plex Enables Simultaneous Visualization of Protein Markers and Ribonucleic Acid Transcripts for Deep Spatial Profiling of Intact Tissue

Summary

Here, we present a protocol for combining iterative protein staining (IBEX) with RNAscope in situ hybridization technology for the deep spatial phenotyping of intact fixed frozen tissues for high-resolution spatial profiling applications.

Abstract

Here, we present a protocol to enable simultaneous spatial profiling of protein and ribonucleic acid (RNA) markers at single-cell resolution in intact tissues. One of the core issues surrounding solid tumor cancers is the inability to predict patient responsiveness to therapeutic intervention. Deciphering these complexities requires quantitative measurement of cellular behaviors, interactions, and phenotypes within intact tissues. Specifically, the proteome and transcriptome provide overlapping but nonredundant information that can be critical to designing rational immunotherapeutics. However, there is not currently a straightforward and cost-effective way to deeply profile both the transcriptional and functional state of all cells in the tumor microenvironment while retaining their critical spatial localization information. To this end, we have successfully combined the IBEX (iterative bleaching extends multiplexity) platform with HiPlex RNAscope to simultaneously visualize high-plex protein markers and high-plex ribonucleic acid (RNA) transcripts with single-cell resolution. This approach incorporates a modified RNA target retrieval step to preserve nuclear morphology and enable accurate image alignment across cycles. “Bex-Plex” enables us to perform an unmodified IBEX protocol on fixed frozen tissues followed by the HiPlex RNAscope protocol, in which we use a modified RNA target retrieval to ensure nuclear-level alignment with IBEX cycles. This protocol is optimized for fixed frozen tissues and enables integrated spatial analysis of transcriptional and functional cellular states. The power to profile at this single-cell level not only provides a deep well of information about the tissue itself but can also empower predictions about the nuanced cellular interactions that drive therapeutic responsiveness, providing a framework for studying spatial biology in tumor microenvironments and related systems.

Introduction

Since their introduction in 2011, cancer immunotherapies that decrease the suppression of cytotoxic T cells by blocking inhibitory immune checkpoint molecules such as PDL-1 and CTLA-4 have emerged as the gold standard for solid tumor treatment¹. However, despite large-scale use in patients, the clinical efficacies of these drugs vary widely because many solid tumor cancers either intrinsically present with or eventually generate resistance to these therapies². In many cases this is due to the extremely heterogeneous nature of solid tumors, even among the same cancer type³. A more in depth understanding of the cell relationships that occur in solid tumors is key to generating improved treatment modalities. Thus, identifying the exact mechanisms and interactions that mount and sustain this resistance is at the forefront of cancer research. Here, we aim to develop a method that enables integrated spatial analysis of protein and RNA expression within intact tissues to better characterize these complex cellular interactions.

The tumor microenvironment (TME) is composed of the non-cancerous cellular and structural components that shape tumor growth and immune response, including cancer associated fibroblasts, infiltrating immune cells, vascular components, and the extracellular matrix⁴. Recently, it has become evident that the composition and organization of the TME strongly influence whether tumors are permissive or resistant to immune infiltration, one of the main factors dictating responsiveness to immunotherapy⁵. In many cancers, an “immune desert” phenotype, characterized by low numbers of T cells and the presence of highly immunosuppressive myeloid populations, correlates with poor responsiveness to checkpoint blockade therapies^6,7. Chemokine signaling governs many of these recruitment dynamics. For example, CXCL1 and CCL2 attract immunosuppressive myeloid cells, while the CXCR3-CXCL9/CXCl10 axis recruits both cytotoxic and regulatory T cells into the tumor⁸. Consequently, the niche-specific signaling environment can potently influence cell phenotype and function. For instance, in colorectal cancer patients, the CCL5-CCR5 axis controls the suppressive function of tumor associated macrophages (TAMs) and blocking this signal skews TAM toward an anti-tumorigenic phenotype and correlates with improved clinical responses⁹. Furthermore, direct interactions between immune cells, cancer associated fibroblasts, and/or tumor cells via ICAM and PDL-1 can induce immune exhaustion and blunt antitumor immunity¹⁰. Together, these findings underscore the importance of spatially resolved analyses to understand how the TME influences therapeutic efficacy.

Immunofluorescent (IF) imaging remains a foundational approach for visualizing protein expression and cell-cell interactions in tissue. The field has advanced from conventional 3-4 color IF to high-content imaging workflows that rely on iterative staining and imaging. In these methods, large datasets are generated via repeated cycles of antibody labeling, imaging, and fluorophore inactivation or removal so that additional markers can be stained and imaged¹¹. Iterative Bleaching Extends Multiplexity (IBEX) was recently introduced as a user-friendly and cost-effective iterative imaging platform¹². Compatible with the use of commercial antibodies and conventional microscopes, IBEX utilizes lithium borohydride (LiBH₄) to quench the fluorescence of most commonly used dyes without damaging tissue epitopes, allowing for the acquisition of up to 75 parameters for a single tissue¹³. The images obtained from this cyclic staining are aligned using the SimpleITK open-source software to produce composite datasets suitable for high-dimensional spatial analyses.

Complementary to protein-level phenotyping, spatial transcriptomic methods can provide valuable information about the transcriptional state of cells within their native environment. Notably, in-situ hybridization (ISH) is a popular technique to identify the localization of DNA or RNA sequences in a specific tissue¹¹. RNAscope is a well-established ISH approach that uses probes with a unique Z chemistry to specifically amplify RNA transcripts at single-molecule resolution and is compatible with many different sample types¹⁴. As a widely adopted and commercially available platform through Advanced Cell Diagnostics (ACD), RNAscope offers validated probes against a wide variety of targets, simplifying experimental design for diverse applications.

A more complete view of cell identity and functions can be gained by combining protein and RNA detection within the same tissue. Short-lived secreted proteins, such as cytokines and chemokines, traditionally lack reliable antibody staining by IHC, but information about their expression can be gained through the quantification of their mRNA transcripts. Previous attempts to co-detect RNA and protein, such as conventional chromogenic IHC or TSA-OPAL, have been limited by nonlinear signal amplification and time- consuming antigen retrieval protocols¹⁵. Compared to these approaches, the Bex-Plex method enables simultaneous high-plex protein and RNA detection while preserving spatial context and nuclear morphology for accurate image registration.

Here we present “Bex-Plex”, a protocol that integrates the IBEX and RNAscope platforms for simultaneous, quantifiable characterization of the transcriptional and translational states of cells within their native tissues. In Bex-Plex, fixed OCT embedded tissues undergo the standard IBEX protocol followed by a slightly modified RNAscope HiPlex workflow, wherein target retrieval is performed at 60°C with a 1 h incubation time as compared to the original 99°C incubation for 5 min. This preserves the architecture of the nuclei, allowing for an accurate alignment of fiducial markers when registering the separate images from the IF and ISH cycles into one. This protocol is best suited for studies requiring simultaneous spatial analysis of protein and RNA expression in intact tissues, particularly when tissue availability is limited or when preserving spatial architecture is critical, such as in tumor microenvironment analysis or immune profiling studies.

Using widely available reagents and compatible with most imaging platforms, Bex-Plex can be carried out by any technically trained research personnel familiar with fluorescent imaging and with access to a conventional microscope. The protocol presented here is optimized for fixed frozen samples. At present, the protocol has been validated in fixed frozen tissues, and further optimization may be required for application to other sample types such as FFPE or human clinical specimens.

This workflow will be invaluable for any context where tissue is limited or precious, such as tissue biopsies or animal experiments that have a limited sample size, as well as for the phenotyping of cells that may be difficult to extract using dissociative techniques. Bex-Plex provides simultaneous, quantifiable information about the transcriptional and translational profiles of single cells while retaining their spatial context, data that is uniquely poised to answer questions about the relationships and crosstalk between cells in their original environment.

Protocol

The procedures outlined in this protocol were approved by the Institutional Animal Care and Use Committee of the University of California, San Diego (protocol number S24049). Male and female mice 7-12 weeks of age on a C57BL/6 genetic background and housed under specific pathogen-free conditions were used for these experiments.

1. Tissue preparation

1. Euthanize experimental mice by approved primary and secondary measures (e.g. CO₂ euthanasia and cervical dislocation) and harvest tissues of interest into supplemented media.  
2. Prepare a fresh stock of tissue fixative by diluting a 4% stock solution of paraformaldehyde (PFA) to either 0.5% or 1% PFA in 1X PBS. Use 0.5% PFA for lower-volume tissues such as lymph nodes and 1% PFA for higher-volume tissues such as tumors..   
   CAUTION: Paraformaldehyde is flammable and toxic if ingested or swallowed. Handle with proper PPE and avoid contact with skin or eyes.
3. Transfer tissues into a 24 well plate with 2 mL fixative per well and incubate overnight at 4°C.
   Note: Incubation times over 16 h can lead to over-fixation issues.
4. Fill the wells of a new 24-well plate with 1X PBS and transfer tissues over from the fixative plate.
   Pause point: If needed, tissues can be stored in 1X PBS at 4°C for up to 7 days.
5. Trim excess fat/debris from the tissues one at a time using a dissection scope and tools. Add a small volume of 1X PBS directly to the tissue to prevent tissue dehydration while trimming.
6. Prepare a 30% w/v sucrose solution by diluting powdered sucrose stock in 1X PBS.
7. Transfer the tissues into a 24 well plate with 2 mL 30% sucrose per well and incubate overnight at 4°C to dehydrate. The tissues are adequately dehydrated when they sink to the bottom of the wells.
8. Fill labeled cryomolds with OCT and place the tissues one at a time into the molds. Tissue orientation will be origin dependent, but our standard recommendation would be for the largest surface area to be parallel with the cryomold bottom. Ensure the tissues are completely submerged in the OCT.
9. Place on dry ice for 10 min to freeze and store at -80°C.

2. Tissue sectioning

Note: All reagents and lab materials used for this protocol must be RNase free. Take care to prevent contamination when handling slides.

1. Coat RNase free microscope slides with 15 µL of chrome alum adhesive and incubate for 1 h at 60°C in an oven. Chrome alum coated slides can be stored at RT (room temperature) for up to a month before use.
2. Take the OCT embedded tissues out of the -80°C freezer and place in a -20°C cryostat/cryotome to equilibrate for at least 20 min.
3. Section tissues onto chrome alum coated slides at a thickness of 10–15 µm.
4. Store slides with cut sections at 4°C or, if staining the same day, let slides equilibrate to RT. Sectioned slides can be stored at 4°C for up to 5 days before use.

3. Protein staining

1. In a dunk tank, soak the slide to be stained in RNase-free 1X PBS for 5 min to remove excess OCT from the tissues.
2. Use a hydrophobic barrier pen to draw boxes around the tissues on the slide and let dry for 2 min.
3. Prepare a blocking solution of 10 µg·mL⁻¹ anti-mouse FC receptor block in RNase-free 1X PBS. Add the blocking buffer to the samples on the slide, using enough to create a bubble that completely covers the tissue (100–150 µL·section^-1).
   Note: Sometimes the addition of FBS or BSA (typically 1–2%) to the blocking buffer is required for difficult tissues and/or markers.
4. Incubate the slides at RT in a humid staining box for 1 h, then wash by completely submerging them in a dunk tank filled with RNase-free 1X PBS for 5 min. 
5. Prepare the primary antibody staining solution in RNase-free 1X PBS, planning for 100-200 µL per tissue section. Standard antibody dilutions range from 1:50-1:200, but it is advised to perform single stain titrations to identify the best dilution factor for a particular antibody and/or tissue type. Vortex and pulse spin the antibodies before pipetting to prevent fluorescent aggregates.
6. Remove slides from the wash and gently wick the excess liquid from the corner of sample boxes using an absorbent task wipe to avoid diluting the antibody solution. 
7. Apply the antibody solution in the same manner as the blocking solution, making a bubble over the tissue.  
8. Place slides horizontally in a 37°C oven and incubate for 1 h. Alternatively, Incubate slides at RT for 3–4 h in a humid box. Standard IF protocols already optimized for the desired tissue of interest may be used. Maintain slides in a humid environment during incubation to prevent dehydration. (e.g. use a wet paper towel in the bottom of the staining tray).
9. Wash the slides by submerging in RNase-free 1X PBS for 2 min. Repeat twice for a total of three washes. If necessary for your staining panel, prepare a secondary antibody staining solution in RNase-free 1X PBS during this time. Standard antibody dilutions range from 1:500-1:1000, but it is advised to perform single stain titrations to identify the best dilution factor for a particular antibody and/or tissue type.
10. Wick any excess PBS off the slides with an absorbent task wipe and add 100–200 µL of secondary antibody staining solution to each tissue section. Place slides horizontally in a 37°C in an oven and incubate for 30 min. Alternatively, slides may be stained at RT for 1–2 h in a humid box.
11. Wash the slides by submerging in 1X RNase-free PBS for 2 min. Repeat twice for a total of three washes.
12. Dilute DAPI nuclear dye to a final concentration of 0.2 µg·mL⁻¹ in RNase-free 1X PBS.and add 100–200 µL to each tissue section.
13. Incubate slides for 10 min at RT in a humid box, then wash once by submerging in RNase-free 1X PBS and dunking up and down briefly.
14. Wick the excess liquid off the slides with an absorbent task wipe and add 20–30 µL of aqueous-base antifade mounting medium to each section. Carefully place a coverslip on the slide, making sure to minimize any bubbles, and let the mounting medium cure for 15–30 min.
15. Image the stained slides. For a more detailed explanation of panel design considerations and image acquisition for IBEX, see the original published IBEX protocols^12,13.
    Pause point: Store mounted slides in the dark at 4°C for up to 3 days, though we recommend proceeding immediately to ensure minimal RNA degradation.

4. Fluorophore bleaching and iterative protein staining

1. Submerge the slides in RNase-free 1X PBS until the cover slip easily slides off without any application of force. Depending on how long the coverslip has been mounted this can take up to 16 h.
2. While the slides are soaking, Prepare a 1 mg·mL^⁻1 solution of LiBH₄ in RNase-free DI H₂O by gently mixing until fully dissolved and wait 10 min for it to activate. Activation has occurred when bubbles are present in the solution and activity lasts for up to 4 h.
   Caution: LiBH₄ is water reactive, avoid excess exposure of powdered stock to moisture and properly dispose of waste accordingly.
3. Wash the slides once for 2 min in RNase-free 1X PBS to remove any mounting medium residue, then pipette 100–150 µL of the LiBH₄ solution onto each tissue. Incubate in a humid box for 15 min.
   Note: Depending on the fluorophores used for staining, this incubation time may need to be lengthened (discussed at length in the IBEX methods papers)^12,13.
4. Wash the slides once by submerging in RNase-free 1X PBS and dunking up and down briefly.
5. Proceed to staining the slides with additional protein markers by following steps 3.5-3.15. It is not necessary to re-stain the nuclear dye because it does not bleach with LiBH₄. Repeat protein staining if necessary.
   Pause point: Mounted slides may be stored at 4°C overnight before bleaching the last protein/IBEX cycle and proceeding to RNA.
6. Immediately proceed to RNAscope HiPlex after bleaching the fluorophores from the final cycle. Although up to six cycles of standard IBEX have been performed without tissue or epitope loss, we do not recommend exceeding three cycles of protein staining for this protocol due to the potential for RNA transcript quality to decrease over longer periods of time.

5. Slide preparation for RNAscope HiPlex

1. Submerge the slides completely in RNase-free 1X PBS until the coverslip easily slides off without any application of force. Depending on how long the coverslip has been mounted this can take up to 16 h. Wash once for 2 min in RNase-free 1X PBS to remove any mounting medium residue.
2. Prepare 200 mL of 50% ethanol, 200 mL of 70% ethanol, and 400 mL of 100% ethanol.
3. Turn on a hybridization oven and set it to 60°C. Dilute the 10X Target Retrieval Reagent to a 1X working stock in RNase-free DI H₂O and prewarm by placing it in the oven until step 5.8.
   Note: Failure to pre-warm target retrieval reagent to 60°C will affect downstream registration of images from the separate cycles. The target retrieval reagent must remain at 60°C for the duration of the retrieval step.
4. Immerse the slides in 50% ethanol. Incubate for 5 min at RT.
5. emove the slides from the 50% ethanol and immerse in 70% ethanol. Incubate for 5 min at RT.
6. Remove the slides from the 70% ethanol and immerse in 100% ethanol. Incubate for 5 min at RT and repeat with fresh 100% ethanol.
7. Remove the slides from the 100% ethanol and let dry at RT for 5 min.
8. Add 100–150 µL of the 60°C 1X Target Retrieval Reagent to each tissue section and place the slides in the hybridization oven. Incubate at 60°C for 1 h.
9. Remove the slides and rinse by submerging in RNase-free DI H2O for 15 seconds. At this time, change the set temperature on the oven to 40°C.
10. Transfer the slides to 100% ethanol and incubate for 5 min, then dry at RT for a minimum of 5 min.
11. Apply Protease III to the slides, 1–2 drops per tissue section. Place the slides in the hybridization oven and incubate at 40°C for 30 min.
12. Submerge the slides in RNase-free DI H₂O and wash by dunking up and down briefly. Repeat with fresh distilled water for a total of 2 washes.

6. RNAscope reagent preparation and probe hybridization

1. Prepare the RNAscope HiPlex probe stocks and HiPlex probe diluent by warming them to 40°C for 10 min in an oven or water bath.
2. Briefly pulse spin down all 50X probe stocks. Prepare a working probe mixture by diluting each unique target probe to 1X with the RNAscope Probe Diluent. Plan to use 50–100 µL of probe mixture per tissue and mix it well by vortexing before use. The RNAscope HiPlex protocol can accommodate up to 3 cycles of 4 probes (12 probes total that are denoted T1-12). For guidance on probe design, consult the original published RNAscope protocol¹⁴.
3. Place the RNAscope HiPlex Amp 1–3 and HiPlex Fluoro T1-T4 reagents at RT to equilibrate.
4. Take the Protease III treated slides and pipette 50–100 µL of the probe mixture onto each tissue section. Place horizontally into the hybridization oven and incubate for 2 h at 40°C.
5. Wash the slides by completely submerging in 1X Wash buffer for 2 min. Repeat with fresh buffer for a total of 2 washes.
   Pause point: Slides may be stored in 5X SSC buffer at RT overnight, as described in the RNAscope protocol from ACD. If this pause point is used, note that two 2 min washes in 1X wash buffer will be required the following morning.

7. Probe amplification

Note: Ensure that all reagents are equilibrated to RT before applying to slides.

1. Add 1–2 drops of the HiPlex Amp1 reagent to the slides and incubate in the hybridization oven at 40°C for 30 min. Wash the slides by submerging completely in 1X Wash buffer for 2 min. Repeat for a total of 2 washes.
2. Add 1–2 drops of the HiPlex Amp2 reagent to the slides and incubate in the hybridization oven at 40°C for 30 min. Wash the slides by submerging completely in 1X Wash buffer for 2 min. Repeat for a total of 2 washes.
3. Add 1–2 drops of the HiPlex Amp3 reagent to the slides and incubate in the hybridization oven at 40°C for 30 min. Wash the slides by submerging completely in 1X Wash buffer for 2 min. Repeat for a total of 2 washes.
4. Add 1–2 drops of the HiPlex Fluoro T1-T4 reagent to the slides and incubate in the hybridization oven at 40°C for 15 min. Wash the slides by submerging completely in 1X Wash buffer for 2 min. Repeat for a total of 2 washes.
5. Add 3–4 drops of HiPlex DAPI reagent to each tissue section and incubate for 1 min at RT.
6. Remove the DAPI by wicking the liquid off the slides with an absorbent task wipe. Immediately add 100 µL of aqueous-base antifade mounting medium (can be the same as used during the IBEX staining steps) to each section. Carefully place a coverslip on the slide, making sure to minimize bubbles, and let the mounting medium dry for 15–30 min.
7. Proceed to imaging.
   Pause point: Mounted slides may be stored in the dark at 4°C for up to 3 days, though we recommend proceeding quickly to ensure minimal RNA degradation.

8. Cleaving the fluorophores and iterative RNA hybridization

1. To remove the cover slips from the slides, soak in 4X SSC buffer for at least 30 min at RT. Wash once by completely submerging the slides in fresh 4X SSC buffer and dunking up and down briefly.
2. Open a fresh ampule of cleaving solution and make a 10% working stock by diluting with 4X SSC buffer.
3. Place the slides in a humid box and add 3–4 drops of the 10% cleaving solution. Incubate at RT for 15 min, then wash the slides by completely submerging in 0.5% PBST for 2 min. Repeat with fresh PBST for a total of 2 washes.
4. To proceed with imaging cycle 2 of the RNAscope panel, equilibrate the HiPlex Fluoro T5-T8 reagent to RT.
5. Turn on the hybridization oven and set the temperature to 40°C.
6. Add the HiPlex Fluoro T5-T8 reagent to the slides and incubate in the hybridization oven at 40°C for 15 min. Wash the slides by submerging completely in 1X Wash buffer for 2 min. Repeat for a total of 2 washes.
7. Carefully wick the excess liquid from the slides. Immediately add 100 µL of mounting medium to each section and carefully place a coverslip on the slide. Let the mounting medium dry for 15–30 min and proceed to imaging.
8. Repeat steps 8.1–8.7 with the HiPlex Fluoro T9-T12 reagent for a total of 3 RNAscope cycles if appropriate. Note that while 3 cycles are possible with the RNAscope HiPlex protocol, the additional time required for protein staining with this protocol may affect RNA quality. Therefore, 3 cycles may not always be feasible depending on the specific transcripts being targeted.

9. Image processing

1. Individually process the raw files from each round and Perform image processing using consistent parameters, including background subtraction settings, smoothing filters, and registration transformation type.. Channels should be named with a consistent separator character (e.g. Cycle 1: CD45 AF532, where “:” makes the distinction between cycle and marker/fluor).
2. Launch the SimpleITK software (https://github.com/niaid/sitk-ibex) and click Imaris extensions > SimpleITK > Affine registration of z-stack using common channel.
3. Upload one image file from each of the multiplex cycles and enter the separator character for all named channels (e.g. “:”).
4. Select the channel to be used for image registration and the fixed image that all other images will be aligned to. It is recommended to use DAPI as the fiducial marker although any common named channel between all cycles can be used.
5. Click “Register” and then “Resample and save combined image”.
6. Proceed to downstream analysis with your software of choice. For more information on image processing and registration, reference the original IBEX protocol^12,13.

Results

Figure 1 : Overview of Bex-Plex workflow.
Fixed frozen tissues sectioned onto chrome alum-coated slides undergo up to 3 cycles of standard IBEX followed by up to three cycles of RNAScope HiPlex wherein the initial target retrieval step is performed at 60°C for 1 h. Images from separate cycles are then registered and aligned using a fiducial marker (e.g., DAPI) to create a multiplex dataset for downstream analysis. Please click here to view a larger version of this figure.

We utilized the protocol described above to profile the tumor microenvironment (TME) of a murine pancreatic ductal adenocarcinoma (PDAC) model (Figure 1). A tumor line isolated from Kras^LSL-G12D/+;Trp53^L/+;Pdx1-Cre;Rosa26^YFP/YFP (KPCY) mice, which spontaneously develop pancreatic cancer and recapitulate many of the key features of human PDAC, was subcutaneously injected into the flank of C57BL/6 mice⁷. Tumor-bearing mice were then treated with combination immunotherapy consisting of anti-PD^-1 (i.p. 200 µg•mouse^-1  , RMP1-14) and anti-CTLA-4 (i.p., 200 µg•mouse^-1, 9H10) administered day 10 post implantation and continued every 3 days until experiment endpoint. Agonistic anti-CD40 (i.p., 100 µg•mouse^-1, FGK45) was given as a singular dose day 13 post implantation. Together, we refer to this as “FPC,” as previously published7,16. Three tumors per experimental group were collected at 16 days post-implantation and fixed-frozen sections were mounted onto chrome alum coated slides for Bex-Plex.

Figure 2: Representative IBEX + RNAscope (Bex-Plex) results.
(A) Combined IBEX and RNAscope staining in pancreatic ductal adenocarcinoma (PDAC) tumor sections demonstrates co-detection of protein and RNA targets and alignment of fluorescence channels. Scale bar, 10 µm. (B) Quantification of T cell (CD3+ CD11b-), macrophage (CD11b+ CD68+), and myeloid (CD11b+ CD68- Ly6G +/-) immune compartments based on protein marker channel intensities. Graphs show the mean cell density for each cell type by treatment group, n=3 tumors. (C) Quantification of Pdl1 and Il10 transcript-expressing cells in the immune compartments previously identified by protein. Graphs show the mean cell density by treatment group, n=3 tumors. Please click here to view a larger version of this figure.

One cycle of protein staining was performed on three FPC- treated tumors and three isotype-treated tumors to assess effects on myeloid lineage cells between treatment groups. Seven parameters were acquired, though we focus here on three markers of interest: CD11b, Ly6G, and CD68 (Figure 2A). After image acquisition, the fluorophores from the protein staining were bleached and one round of RNAscope was performed using probes against Il10 and Pdl1 (Figure 2A). The separate images were then aligned using DAPI as the fiducial (Figure 2A). Quantitative mean fluorescence intensity of CD11b, CD68, Ly6G and CD3 revealed a trend toward fewer macrophages and myeloid cells in FPC-treated tumors relative to controls (Figure 2B). Quantitative comparisons were performed across all samples, and data are presented as mean values. Furthermore, the number of cells expressing Pdl1, Il10, or a combination of both transcripts in both the macrophage and myeloid compartments showed a downward trend with FPC treatment compared to controls (Figure 2C). These results indicate a potential reduction in immunosuppressive cell populations following treatment, consistent with expected biological responses.

Figure 3: Optimization of target retrieval conditions for nuclear preservation.
(A) Representative image of a tissue after RT target retrieval showing nuclear swelling and indistinct morphology. Examples of nuclear sizes are indicated by yellow outlines. Scale bar, 40 µm. (B) Representative image of a tissue after modified target retrieval showing improved nuclear morphology with adjusted temperature conditions. Examples of nuclear size indicated by yellow outlines. Scale bar, 40 µm. (C) Mean nuclear volume determined by DAPI-based segmentation on tissues that underwent RT target retrieval (not pre-warmed), 60°C prewarmed target retrieval, or under standard IF conditions with no target retrieval. Graph bars indicate mean and error bars show standard deviation. ****P < 0.0001 by ordinary one-way ANOVA (asterisks denote statistical significance threshold); n= 1639 cells per condition. (D) Heatmaps representing the correlation scores between three images before and after image alignment. Correlation scores are shown for a successful registration and a registration that failed to align due to differences at a nuclear level.

Importantly, when Bex-Plex was performed without the modified target retrieval described, the nuclei appeared enlarged with indistinct edges compared to the nuclei from a tissue where the modified target retrieval was correctly performed (Figure 3A and B). Use of room temperature target retrieval regent led to a 33% increase in mean nuclear volume based on DAPI staining compared to target retrieval performed using reagent that was prewarmed to 60°C (Figure 3C). Nuclear volume measurements were obtained from representative fields across samples using DAPI-based segmentation. Moreover, when this step was performed at 60°C using prewarmed retrieval reagent, the mean nuclear volume did not significantly differ from the mean nuclear volume after standard IF staining (Figure 3C). As DAPI is the most effective parameter to use as a reference for image alignment, this comparison (Figure 3) highlights the critical importance of the modified retrieval parameters to preserve nuclear morphology. Correct image registration is indicated by an alignment correlation score that is close to 1, indicating similarity between images of different cycles (Figure 3D, before vs successful).

Successful implementation of the protocol is indicated by preserved nuclear morphology, strong and specific RNA signal, and accurate alignment of protein and RNA channels. In contrast, suboptimal outcomes may include nuclear distortion, weak or diffuse RNA signal, and misalignment between imaging cycles (Figure 3D, before vs failed). As shown in Figure 3C, the nuclear morphology should be retained throughout the protocol, and will be both qualitatively and quantitatively apparent. Prior to Bex-Plex implementation, users should ascertain the ground truth for their tissues of interest via separate protein and RNA staining. Not only does this ensure optimized protein and RNA staining for the target tissues, but provides relevant spatial patterns that should be retained when combined for Bex-Plex. For example, CD11b colocalizes with both CD68 and Ly6G and shows cell membrane-localized staining (Figure 2A). This is expected, as CD11b is a pan myeloid cell marker and CD68 and Ly6G are markers for the macrophage and neutrophil myeloid subsets. Successful RNA detection is characterized by distinct punctate signals localized to individual cells, whereas suboptimal results may present as weak or diffuse staining. The RNA probe signal should show localization in and around the nucleus, depending on the transcript function. Structural protein markers such as vimentin could be included to improve cellular resolution of transcript localization by revealing cell membranes in complex tissues. Negative or control samples are expected to show minimal background staining and no nonspecific probe signal, confirming the specificity of both protein and RNA detection.

Discussion

The Bex-Plex workflow integrates the iterative protein staining platform, IBEX, with RNAscope in situ hybridization to enable simultaneous detection of proteins and RNA transcripts at single-cell resolution. The primary goal of this protocol is to provide a reproducible method for integrated spatial analysis of protein and RNA expression within intact tissues. The ability to measure both protein and RNA expression in the same tissue has previously presented technical challenges: high costs, complex antigen retrieval steps, limited multiplex capacity, resolution limitations (non-single cell), nonlinear signal amplification, and extensive bioinformatics pipelines for downstream processing ^17,18,19. Compared to existing approaches such as multiplex immunofluorescence, OPAL/TSA staining, and spatial transcriptomics platforms, Bex-Plex enables reproducible, quantitative, and high-plex simultaneous detection while preserving spatial context and nuclear morphology for accurate image alignment. Additionally, Bex-Plex can be performed in a standard lab environment and generates data at a single cell resolution that is straightforward to process. Leveraging this protocol to gain a comprehensive understanding of cell-cell interactions within a variety of tissue contexts, such as in the tumor microenvironment, will be essential for developing more effective immunotherapies and improving patient outcomes across disease states.

Our method is performed on fixed tissues embedded in OCT and follows the standard IBEX workflow without modification, followed by a slightly altered RNAscope HiPlex protocol. A key difference is the target-retrieval step at the beginning of the RNA hybridization workflow: instead of the manufacturer’s 99°C for 5 min, retrieval is performed at 60°C for 1 h. This adjustment preserves nuclear architecture, which is imperative for accurate image alignment between protein and RNA cycles. Critical steps of this protocol include maintaining consistent target retrieval temperature, efficient fluorophore bleaching using LiBH₄, and minimizing RNA degradation through rapid processing and use of RNase-free reagents. It is most important to ensure the use of pre-warmed target retrieval reagents at 60°C; failure to do so can cause nuclear swelling even with the altered incubation step.

While it has been shown that 15 min of incubation with LiBH₄ is adequate to inactivate many commonly used fluorophores, it is recommended to individually verify this process for each protein cycle before running the whole Bex-Plex protocol. In the initial IBEX publications it is well characterized that not all IBEX-compatible flurophores are equally bleachable¹². Additionally, our lab has observed that some high abundance structural markers such as vimentin may require a longer bleaching time. Confirming that the fluorophore and marker combinations being used will adequately bleach should prevent residual signal spill into subsequent protein and RNA cycles. The IBEX Community page has excellent resources for IBEX-compatible panel design and is specifically intended to reduce the burden on new users by indicating antibody efficacy and fluorophore bleaching times²⁰. Furthermore, iterative staining and bleaching should not affect RNA transcript quality, however it is important to complete protein cycles as fast as possible and use RNAse-free reagents to minimize any RNA degradation that may happen over time. Failure to do so will affect probe hybridization and lead to decreased signal for any RNAscope cycles performed, particularly for low abundance transcripts. Common issues encountered during the protocol include weak RNA signal, incomplete fluorophore bleaching, and misalignment between imaging cycles, which can be mitigated through optimization of staining conditions and imaging parameters.

It is imperative to limit the time spent between performing protein staining and RNA probe hybridization to less than 5 days. When performing Bex-Plex on multiple slides, it is highly advised to process one slide at a time to reduce variability in RNA degradation. Similarly, limit the time between probe hybridization and imaging. While we have successfully imaged up to 3 days post amplification, longer storage is not recommended. If there is reduced signal after multiple cycles of staining or hybridization, it may be necessary to increase speed throughout the protocol. Ideally, we recommend working as fast as possible, as delays reduce RNA integrity and can diminish amplification. Other approaches include reducing the number of protein cycles performed to more easily fit into the timeline constraints of the protocol or changing RNA probe panel design to include higher abundance targets. Before combining IBEX and RNAscope, validate the antibody and probe panels for each cycle separately to determine the expected signal quality for each cycle. In the event of nuclear swelling after the target retrieval, it is possible that the target retrieval reagent was not prewarmed properly. Ensure that the retrieval reagent is not allowed to cool down once it is diluted and prewarmed to 60^°C and that the temperature stays consistent during the 1 h on-slide incubation. Lastly, if image registration and alignment fails in the absence of nuclear swelling, it means that there was a mismatch in image acquisition parameters such as bit depth, voxel size, or Z position across cycles. In the event of such errors, it may be necessary to repeat the protocol with adjustments to these settings so that they are consistent. These troubleshooting strategies are essential for ensuring reproducibility and consistency across experiments.

The data generated by Bex-Plex enables an in-depth view of the spatial landscape of tissues on a previously unattainable level. However, a major limitation of this protocol is the fact that the total number of iterative protein cycles possible is lower than with IBEX alone due to gradual RNA degradation over time. While up to 6 cycles of standard IBEX have been performed, we do not recommend exceeding 3 cycles of protein staining for Bex-Plex. Even so, the protocol does accommodate the maximum number of RNAscope cycles and this level of multiplex depth is sufficient for most applications.

Additional limitations include variability across tissue types and potential autofluorescence or staining artifacts that may affect signal interpretation. More fragile tissues can sustain fewer cycles and different tissue types can have unique characteristics that affect staining. These limitations can be assessed via single cycle stains of protein or RNA and addressed prior to running Bex-Plex. Sectioning onto chrome alum-coated slides will substantially improve tissue adherence and structural integrity, but autofluorescence proves a consideration for all immunofluorescence-based techniques. Additionally, at present the protocol has not been validated in FFPE or human clinical samples and will require further optimization for such applications. However, given that IBEX and RNAscope are predominantly performed on formalin-fixed paraffin embedded (FFPE) tissues, it is feasible that Bex-Plex can expand to these sample types. Successful implementation will require further optimization to antigen retrieval and, potentially, target retrieval steps to preserve RNA quality.

Overall, Bex-Plex offers a straightforward, cost-effective, and broadly accessible method for high-plex spatial analyses. This method is particularly well suited for applications such as tumor microenvironment analysis, immune cell profiling, and spatial mapping of signaling interactions. This integrated view enhances the ability to interrogate mechanisms of immune regulation, tumor heterogeneity, and therapeutic responsiveness across a wide range of experimental models. Bex-Plex will be invaluable for contexts where it is necessary to study the crosstalk that occurs between cells in intact tissue, such as in the tumor microenvironment or in other tissue during inflammation and homeostasis. In particular, the ability to resolve spatial dynamics of cytokine signaling by probing for the transcripts of short-lived molecules in conjunction with phenotypic protein markers will provide a deeper understanding of cellular interactions across disease states. Future improvements may include increasing multiplex capacity, integration with automated image analysis pipelines, and adaptation for clinical biopsy samples by expanding the workflow to include FFPE tissues.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
1.5 ml microcentrifuge tubes	Biopioneer	MCNT-1.5F
200 proof EtOH	Fisher Scientific	BP28184
20X SSC buffer	Invitrogen	AM9765
24 well plates
Alexa Fluor 488 anti-mouse CD11b Antibody	Biolegend	101217
Alexa Fluor 523 anti-mouse CD3e Antibody	Invitrogen	58-0032-82
Alexa Fluor 647 anti-mouse Ly6G Antibody	Biolegend	127610
Alexa Fluor 700 anti-mouse CD68 Antibody	Biolegend	137025
BD Cytofix/Cytoperm fixation and permeablization solution	BD	554722
Chrome alum gelatin	Fisher	NC1692262
Confocal microscope (e.g Leica Stellaris)
Cryomolds	VWR	25608-922
Cryostat (e.g CryoStar NX50 Cryostat)	Thermofisher Scientific	957100
Cryostat blades (e.g Leica Low Profile Disposable Blades DB80LX)	Leica Biosystems	14035843496
DAPI	Thermofisher Scientific	62248
Dissection microscope (e.g Zeiss SteREO Discovery.V8)
Dissection tools
Dry Ice
Dunk tank for slide staining (e.g Simport Scientific EasyDip Slide Staining Jars)	Fisher Scientific	22-038-489
Fluoromount-G	Fisher	OB10001
Heated water bath
Histology brushes (e.g. Ted Pella Camel Hair Brushes)	Fisher Scientific	NC2023341
Hybridization oven (e.g. HybEZ II Hybridization System)	ACD	321710
Image analysis software of choice
ImmEdge Hydrophobic Barrier PAP Pen	Vector Laboratories	H-4000
Kimwipes	VWR	21905-026
Lithium borohydride	STREM Chemicals	93-0397
Micropipettes (P-10, P-20, P-200, P-1000)
Mini Microcentrifuge (e.g. Corning LSE Mini Microcentrifuge)	Corning	6770
Mouse BD Fc Block	BD	553142
No 1.5 Coverglass	VWR	48393-241
Pipette tips (P-10, P-20, P-200, P-1000)
RNAscope HiPlex Probe, mm-CD274-T7	ACD	420501-T7
RNAscope HiPlex Probe, mm-IL-10-T6	ACD	317261-T6
RNAscope HiPlex12 Reagents Kit v2 (includes detection kit, cleaving stock solution, protease reagents, target retreival reagents, and wash buffer)	ACD	324409
RNAscope Protease III	ACD	322337
RNase ZAP	Millipore Sigma	R2020-250
Simple ITK software: https://github.com/niaid/sitk-ibex
Slide staining tray/humidity chamber
Sucrose Bioxtra	Sigma	S7903-1KG
Superfrost Plus Microscope slides	VWR	48311-703
Tissue-Tek OCT	VWR	25608-930
Tween-20, Molecular Biology grade	Promega	H5152
UltraPure DNase/RNase free water	Fisher Scientific	10977023