Protocol for preparing formalin-fixed paraffin-embedded musculoskeletal tissue samples from mice for spatial transcriptomics

Summary Here, we present a protocol for using spatial transcriptomics in bone and multi-tissue musculoskeletal formalin-fixed paraffin-embedded (FFPE) samples from mice. We describe steps for tissue harvesting, sample preparation, paraffin embedding, and FFPE sample selection. We detail procedures for sectioning and placement on spatial slides prior to imaging, decrosslinking, library preparation, and final analyses of the sequencing data. The complete protocol takes ca. 18 days for mouse femora with adjacent muscle; of this time, >50% is required for mineralized tissue decalcification. For complete details on the use and execution of this protocol, please refer to Wehrle et al.1 and Mathavan et al.2

Bone adaptation and healing are spatially and mechanically controlled processes 3 involving crosstalk of multiple tissues. 4Preservation of the spatial context represents a transformative advancement in cellular profiling technologies (e.g., RNA sequencing) with the potential to supplant established histology-based methods, such as in situ hybridization, immunohistochemistry and dissociated single-cell techniques.Whereas spatial transcriptomics analyses have successfully been applied in many research areas, 5 its use in bone has been limited due to the calcified nature of the samples often associated with tissue detachment from slides.Recently, spatial transcriptomics approaches have been described for cryo-embedded skeletal samples from P0 calvaria and during digit regeneration in mice. 4,6,7To allow for broader application of spatial transcriptomics in musculoskeletal research, we have now established a protocol for FFPE calcified tissue samples from mice.
Here we describe a detailed protocol that permits the use of spatial transcriptomics (Visium, 103 Genomics) in bone and multi-tissue musculoskeletal FFPE samples from mice.The protocol below describes the specific steps for samples from an externally-stabilized femur defect model in mice.However, we have also used this protocol for bone samples (vertebra, intact femur), muscle samples (m.quadriceps femoris) and multi-tissue samples (intact femur with adjacent muscle).The protocol is structured into seven steps ranging from tissue harvesting and preparation to the final analyses of the sequencing data.The complete protocol takes ca.18 days for mouse femora with adjacent muscle; of this time >50% is required for mineralized tissue decalcification.Potential pause points are indicated in the detailed description of the procedure.

Institutional permissions
The FFPE embedded musculoskeletal tissue samples used for protocol establishment were obtained from fracture healing experiments in mice.For these experiments the mice were either purchased from the distributor Janvier (Saint Berthevin Cedex, France) or bred in house at the ETH Phenomics Center (EPIC).All animal procedures were approved by the Commission on Animal Experimentation (license numbers: ZH181/2015 and ZH229/2019; Kantonales Veterina ¨ramt Zu ¨rich, Zurich, Switzerland).We confirm that all methods were carried out in accordance with relevant guidelines and regulations (Swiss Animal Welfare Act and Ordinance (TSchG, TSchV)) and reported considering ARRIVE guidelines.10] Prior to starting this protocol, all relevant permissions for performing animal experiments and for obtaining musculoskeletal samples from mice need to be acquired.Note: Remove as much muscle tissue surrounding the femur as wanted.

Preparation and tissue harvesting
Note: Femur and m.Quadriceps can be kept together throughout decalcification.
Note: if using bones from fracture healing studies with external fixation, the fixator can either be removed now or kept in place until Paraffin embedding.

CRITICAL:
Steps h-o should be done as fast as possible (max.10 min) to avoid RNA degradation.

STEP-BY-STEP METHOD DETAILS
Major Step 1: Sample treatment and paraffin embedding (for workflow see Figure 1) Major Step 1 describes the fixation, decalcification and embedding of bone and multi-tissue musculoskeletal samples from mice.Special considerations when processing samples from externally-stabilized femur defect models in mice are indicated in the notes related to sub-step 6.
1. Fix the tissue in 10% neutrally buffered formalin for up to 24 h (ca.16  Note: for muscle tissue: cross-sections are easier to cut compared to longitudinal sections.
Note: for bone tissue: aligning the femur axis to the embedding mold allows for easier sectioning compared to oblique embedding.
Note: if external fixation is used to study bone healing: for embedding of bones with the external fixator still being attached, hold the external fixator with tweezers while embedding.Fill Paraffin until the bone is covered without yet reaching the fixator body, do not let the Paraffin completely dry.Carefully remove the pins of the fixator, completely fill the mold with paraffin and place the histology cassette on top of the paraffin.the Visium FFPE tissue preparation guide available at: https://www.10xgenomics.com/support/spatial-gene-expression-ffpe/documentation/steps/tissue-prep/visium-spatial-gene-expressionfor-ffpe-tissue-preparation-guide.9. Place FFPE block on 4 C plate or on metal block in freezer (<5 min).10.Set-up microtome and insert blade.11.Clamp FFPE block in microtome.12. Align complete surface of FFPE block to the blade (90 ). 13. Trim FFPE block until all tissues on block are exposed/representative of tissues later to be placed on the Visium slide.14.Cut a section of 10 mm (without water flow turned on) and immediately place the rolled-up section with autoclaved tweezers into a pre-cooled autoclaved Eppendorf tube.15.Repeat step 14 3x, to pool 4 sections of 10 mm per FFPE tissue block in one Eppendorf tube.16.Immediately proceed to RNA isolation or intermediately store the Eppendorf tube with the pooled sections at À80 C -Pause point (days -weeks).17.Isolate RNA using the RNeasy FFPE kit (QIAGEN) -Pause point.18. Immediately proceed to DV200 measurements or intermediately store the RNA at À80 C -Pause point (days -weeks).19.Determine the DV200 using a TapeStation and select samples meeting the requirements of the specific spatial transcriptomics platform (e.g., DV 200 R 50% for Visium FFPE spatial transcriptomics) -Pause point (days -weeks).
Major Step 3: Sample sectioning and placement on visium FFPE slide (for workflow see Figure 2B) Timing: 1 day

Major
Step 3 provides details on the sample sectioning and placement on the Visium FFPE slide.We have included detailed steps relevant for processing bone and multi-tissue musculoskeletal samples from mice.CRITICAL: For the post hybridization and ligation washes take care to slowly add the wash buffer to the side of the gasket to avoid detachment of the tissue.

Major
Step 6: 10x Visium spatial transcriptomic data analyses and visualization (for workflow see Figure 3) Timing: 1 day 41.Install software and dependency packages, download the reference genome, probe set reference file, and the slide layout GPR file for the mouse model from the weblink provided under software and dependency requirements.42.Use Space Ranger's mkfastq module to perform the process of separating Visium-prepared raw base call (BCL) files produced by Illumina sequencers into FASTQ files.43.Align the unprocessed Visium sequence data to the mouse reference genome, then generate spatial features, count matrices, and filter barcodes using the count pipeline within Space Ranger.44.Perform spot-level expression data analysis using Seurat (R toolkit for single-cell genomics) 11 or Scanpy (Python-based platform). 1245.Perform quality control and basic filtering of spots based on total counts and expressed genes.
Further exclude spots with mitochondrial and ribosomal reads from the analysis.46.Normalize and scale the data and identify significant variable genes using SC Transform. 137.Apply the dimensionality reduction technique to the dataset and elbow plots to identify the appropriate number of principal components for clustering.48.To identify clusters of spots, employ a clustering algorithm that optimizes shared nearest neighbor (SNN) modularity.49.Visualize spots clustering using uniform manifold approximation and projection for dimension reduction (umap). 1450.Overlay umap gene expression clusters with histology image.51.Perform further downstream analysis such as clustering and visualization, cell-specific marker genes, cell-type identification, segmentation and visualization in the spatial domain, spatial trajectory inference, cell-cell interaction, identification of spatially variable genes, differential gene expression analysis, enrichment analysis of spatial expression data etc.52.Perform tissue annotation using qupath (https://qupath.github.io/https://qupath.github.io/)and extract the spot coordinates and barcodes using Loupe Browser (https://www.10xgenomics.com/support/software/loupe-browser/latest).

EXPECTED OUTCOMES
The developed protocol enables the application of FFPE musculoskeletal tissue samples from mice to spatial transcriptomics.Our motivation was to assess spatially-resolved gene expression within fractured bone and the surrounding muscular tissue in FFPE samples from mice.To achieve this, we developed a spatial transcriptomics protocol for musculoskeletal tissue overcoming current restrictions in mineralized samples.The protocol allows to meet quality measures (e.g., DV 200 can be kept >50 for bone+/-muscle during decalcification), we highlight critical steps (timing of tissue harvest; tissue embedding and sample orientation important for later sectioning) and indicate instruments and tools helpful for the processing and the application of musculoskeletal tissue to spatial transcriptomics.[17]

LIMITATIONS
The protocol has been established for musculoskeletal samples from mice.All pretreatment steps, including the decalcification period, were optimized for samples from an externally-stabilized femur defect model in mice.While we have also used this protocol for bone samples (vertebra, intact femur), muscle samples (m.quadriceps femoris) and multi-tissue samples (intact femur with adjacent muscle) from mice, application of the protocol to other musculoskeletal sites and species may require further optimization (e.g., addition of RNase inhibitor during sample pretreatment).

TROUBLESHOOTING Problem 1
Inhomogeneity in Paraffin of the embedded samples (related to Step 1).This can occur when the Paraffin was added several times with the first layers already being cooled down when further layers were added.

Potential solution
Make sure to have no phase separation of the Paraffin during embedding.Do not let layers of Paraffin cool down before adding further Paraffin

Problem 2
Low DV 200 ratio (related to Step 2).This can be caused by the pre-treatment, the duration of sample storage and the exposure to RNases during sectioning.

Potential solution
Shorten the pre-treatment by determining the minimum required decalcification time to allow sectioning of the samples.Avoid long-term storage.Use sterile instruments and RNase away.Switch to spatial transcriptomics workflows targeted towards transcriptome recovery from degraded samples.

5 .
Place histology cassettes into tissue processor with the following program: 1. Flushing with 70% Ethanol at 37 C for 1 h, 2. Flushing with 70% Ethanol at 37 C for 1 h, 3. Rinsing with 100% Ethanol at 37 C for 30 min, 4. Rinsing with 100% Ethanol at 37 C for 30 min, 5. Dehydration in 100% Ethanol at 37 C for 1 h, 6. Clearing with Isopropanol at 40 C for 1 h, 7. Clearing with Isopropanol at 45 C for 1 h, 8. Infiltration with Paraffin at 62 C for 4.5 h. 6. Remove samples from tissue processor and proceed to Paraffin embedding.

7 .
Place mold on 4 C plate for several hours (ca.3-7 h). 8. Remove FFPE sample from mold and store at 4 C in fridge -Pause point (days-years).Major Step 2: FFPE sample selection and preparation for sectioning (for workflow see Figure 2A) Timing: 1 day Major Step 2 describes the sample selection based on RNA quality as well as the preparation for sectioning of bone and multi-tissue musculoskeletal samples from mice.The steps are based on

Figure 1 .
Figure 1.Overview of sample treatment and paraffin embedding workflow consisting of fixation, decalcification, paraffin infiltration and paraffin embedding Created with BioRender.com.

Figure 2 .
Figure 2. Workflow for spatial transcriptomics of musculoskeletal FFPE samples from mice (A) Pre-treatment: Sectioning, pooling and RNA quality check (B) Spatial transcriptomics: section scoring, placement on capture area, H&E staining and imaging prior to library prep and RNA-seq.

Figure 3 .
Figure 3. Comprehensive analysis pipeline for spatially-resolved transcriptomics data (A) The tissue sections are placed on the Visium slide.Barcoded capture probes store spatial information, which is added to the captured transcript before sequencing.Stained tissue sections are examined by microscopy to acquire bright field imaging.(B) Sequence whole-transcriptome gene expression libraries on sequencing instrument.(C) The sequencing data is used as input for demultiplexing and transcript quantification using space ranger pipeline.(D) The count data, together with the image data, are used as inputs to perform various downstream analyses.
Make circumferential skin incision with scissors around right femur.i. Incise muscles of upper hindlimb toward hip joint.j.Ex-articulate hip joint via manually rotating.k.Remove right hind limb and place on lid of Petri dish.l.Remove outer pair of gloves.m.Hold the tibia with tweezers on the Petri dish (cranial part of tibia facing down).n.Use a scalpel (scalpel blade 11 or 15) to disconnect the tibia and the femur in the knee joint.o.Place femur +/-surrounding muscle in pre-cooled formalin solution.
b. Fill Falcon tubes with ca. 12 mL of 10% neutral buffered Formalin solution and place them on ice add Formalin concentration.