Laser Capture Microdissection of Mouse Embryonic Cartilage and Bone for Gene Expression Analysis

The musculoskeletal system is a multicomponent system composed of muscle, connective tissue, tendon, ligament, cartilage, and bone, innervated by nerves and vascularized by blood vessels¹. The skeletal tissues develop with increasing cellular heterogeneity and structural complexity. Cartilage and bone develop from the same osteochondroprogenitor lineage and are highly related. Embryonic cartilage and bone develop in association with muscles, nerves, blood vessels, and undifferentiated mesenchyme. Cartilage may also be surrounded by bone, such as Meckel's cartilage and condylar cartilage within the mandibular bone. These tissues are anatomically associated and interact with each other through extracellular signals during development. In the study of gene expression in the development of cartilage and bone, one obstacle is the heterogeneity of skeletal structures composed of multiple tissue types. Precise isolation of the specific tissue of interest is key for successful transcriptional analysis.

Laser capture microdissection (LCM) is a powerful tool to isolate cell types or regions of interest within heterogeneous tissues, and is reproducible and is sensitive to the single cell level². It can precisely target and capture cells of interest for a wide range of downstream assays in transcriptomics, genomics, and proteomics^3,4. The quality of the isolated RNA, DNA, or protein can be assessed with a bioanalyzer or equivalent platform. For example, RNA quality is indicated by the RNA integrity number (RIN)⁵.

Here, we provide a protocol for the rapid staining and isolation of cartilage and bone by LCM from fresh frozen tissues. We use the mouse embryo to demonstrate that this protocol yields high quality RNA for subsequent transcriptomic analysis, such as RNA sequencing (RNA-seq).

Tissues from mice were obtained in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and study protocols were approved by the Institutional Animal Care and Use Committee at the Icahn School of Medicine at Mount Sinai.

1. Preparation of Fresh Frozen Specimen

1. Dissect the embryo or tissue of interest. Embed the sample in a disposable embedding mold with optimal cutting temperature (OCT) compound. Adjust the orientation of the specimen with a tip or needle.
2. Rapidly freeze the samples in a dry ice/methyl-2-butane bath. Continue to the next step or store at -80 °C.
   NOTE: The protocol can be paused here.

2. Cryosectioning for Laser Capture Microdissection

1. Defrost the cryostat and clean internal surfaces with 70% ethanol. Treat the anti-roll plate and a pair of forceps with RNase decontamination agent. Set the cryostat to the desired cutting temperature (-18 to -22 °C). Prepare a lidded container with dry ice, placing a sheet of aluminum foil on the dry ice for temporary storage of the sectioned sample slides during cryosectioning.
   CAUTION: Dry ice is extremely cold. Always handle dry ice with care and wear insulated gloves whenever handling it.
2. Transfer the fresh frozen specimen into a bucket of dry ice. Place a layer of OCT compound onto a cryostat specimen holder and immediately place the specimen on top of the OCT with gentle pressure. When OCT is completely frozen, fasten the holder with the specimen into the cryostat cutting arm.
3. Leave the sample in the cryostat for 15 min to equilibrate to the cutting temperature.
4. Section the tissue and collect the slices on polyethylene naphthalate (PEN) membrane slides at a thickness of 12 μm. Align consecutive sections on the membrane. Once one slide is completed, allow sections to dry for a few minutes and transfer the slide onto the foil in the dry ice container until all the required slides are collected.
5. Store the slides at -80 °C in a slide box.
   NOTE:Section thickness is chosen to maximize the amount of tissue collected while allowing efficient cutting of the tissue, and may vary depending on tissue of interest. The protocol can be paused here. Keep slides free from RNase contamination. Avoid temperature changes. Although the RINs of RNA from slides stored for up to 6 months may still indicate high quality of the RNA, we recommend using slides as soon as possible.

3. Sample Preparation for Laser Capture Microdissection

1. Prepare solutions for staining and dehydration.
   1. Place 45 mL of 80% ethanol in each of three 50 mL centrifuge tubes on ice. Label them as #1, #2, and #3. Dilute ethanol with RNase/DNase free water.
   2. Place 45 mL of 95% ethanol in a 50 mL centrifuge tube on ice.
   3. Place 45 mL of 100% ethanol in a 50 mL centrifuge tube on ice.
   4. Place 45 mL of xylene in a 50 mL centrifuge tube at room temperature.
      CAUTION: Xylene should be used in a fume hood.
2. Thaw a PEN membrane slide briefly by placing it against a gloved hand, then wash twice for 30 s each in 45 mL of 80% ethanol (#1 and #2) with agitation in 50 mL centrifuge tubes to remove the OCT.
   NOTE: It is important to remove OCT before staining, to prevent OCT loosened during staining from obscuring the sections. Forceps can be used to facilitate removal of OCT that is not washed off by agitation.
3. Lay slide on a sheet of aluminum foil. Pipette 0.8 mL of 0.1% cresyl violet in 50% ethanol onto the slide and stain for 30 s. Wash the slide in 45 mL of 80% ethanol (#3) for 30 s, and then dehydrate by passage for 30 s each through 45 mL of 95% ethanol, 100% ethanol, and xylene in 50 mL centrifuge tubes.
4. Stand slides on a delicate task wiper for 5 min to drain xylene and dry sections.
   NOTE: Slides must be dried completely before LCM.

4. Laser Capture Microdissection

NOTE: Wear powder free nitrile gloves while performing LCM.

1. Turn on the LCM microscope, laser, and computer. Log on to computer and start the software. Turn the key to the "On" position to start up the laser. When the green light (Laser) is on, press the red button to activate the laser, and then the light (Laser) turns red indicating laser is ready to use.
   NOTE: The Laser needs approximately 10 min to warm up.
2. Set up the collection tubes.
   1. Insert the cap of a 0.5 mL polymerase chain reaction (PCR) tube into the collector.
      NOTE: Choose the matching collector for the desired PCR tubes (0.2 or 0.5 mL).
   2. Pipette 50 µL of extraction buffer (provided by RNA isolation kit listed in the Table of Materials) into the cap of the 0.5 mL collection tube.
   3. Insert the collection device into the microscope.
   4. In the Change Collector Device window, click on the Move to Reference Point button to move the collector to the reference point (RP). Adjust the focus to clearly view the RP. Click the arrows to move the RP to the center of the view.
3. Load specimen slides.
   1. Click the left Unload button in the toolbar to lower the slide holder.
   2. Place the slide into the holder, with the tissue section facing downwards.
      NOTE: This orientation ensures that the captured tissue sections fall into the cap of the collection tube by gravity.
   3. Insert the holder back into the stage.
4. Set up the laser parameters.
   1. Select the Control option of the Laser menu or click on the Laser icon.
   2. Adjust these parameters as desired: Power, the power of the laser; Aperture, the width of the laser; and Speed, the speed of cutting in Draw and Cut mode.
      NOTE: Test the laser in an area out of interest. Higher power or larger aperture makes it easier to cut but causes more damage to cells. Adjust the combination of power, aperture, and speed to obtain efficient cutting and minimal damage of the tissue. For example, Power = 40, Aperture = 5, and Speed = 7 for cartilage tissue on a 12 μm section.
5. Select and cut areas of interest.
   1. Start with 5x objective and find the areas of interest, and then switch to the objective (5x, 10x, or 40x) that best shows the area of interest.
      NOTE: After cresyl violet staining, cartilages are stained magenta and mineralized tissues appear brown or black, distinguishable from other tissues. Images can be saved using Save Image As from the File menu.
   2. Choose the tube for collection (A, B, C, D) by clicking the mark at the collector.
   3. Choose Move and Cut or Draw and Cut. In Move and Cut mode, the specimen is cut manually, using the mouse or the touch-screen pen to draw shapes freehand. In Draw and Cut mode, shapes may be drawn freehand with the mouse or touch-screen pen for subsequent cutting. Click the Start Cut button to initiate laser cutting.
   4. Once the cut is completed, the target tissue with PEN membrane falls into the cap of the collection tube by gravity. Repeat 4.5 to pool multiple areas of interest if needed.
      NOTE: Repeating cuts or increasing Power or Aperture may be necessary if the tissue does not drop into the cap of the collection tube due to incomplete cutting. Mineralized bone (brown or black) cannot be cut directly by laser.
6. Unload the collector and carefully close the PCR tube. Place the microdissected tissues in dry ice. Continue to the next step or store at -80 °C.
   NOTE: The protocol can be paused here. Repeat 4.2 to 4.6 for the next sample.

5. Lysis of Microdissected Tissues and RNA Isolation

1. Thaw the microdissected tissues at room temperature. Centrifuge briefly and incubate the samples for 30 min at 42 °C.
2. After incubation, spin the lysis buffer down into the 0.5 mL PCR tube. Perform DNase treatment and RNA extraction using an RNA isolation kit (see the Table of Materials) following the manufacturer's instructions.

Coronal sections of fresh frozen mouse tissues at E16.5 were used to demonstrate the isolation and collection of Meckel's cartilage (MC), condylar cartilage, and mandibular bone by LCM. Mouse embryos at E16.5 were dissected and embedded in cryogenic molds with OCT compound. Samples in molds were rapidly frozen in a dry ice and methyl-2-butane bath and stored at -80 °C.

To demonstrate cresyl violet staining of cartilage and bone, cryosectioning in the coronal plane was performed and samples were collected on microscope slides. Sections were washed, stained, and dehydrated following the protocol above (step 3.2–3.4). Slides were air dried and mounted with permanent mounting medium. All cartilages examined were stained magenta (MC, condylar cartilage, nasal septum cartilage, costal cartilage, cartilage primordium of presphenoid bone, cartilage primordia of radius and ulna) and all mineralized tissues were stained brown or black (Figure 1A–E). Both cartilage and bone were easily distinguished from other tissues at multiple anatomical sites.

For LCM, heads of embryos at E16.5 were sectioned in the coronal plane on PEN membrane slides at a thickness of 12 μm, and 6–8 consecutive sections were collected per slide and stored at -80 °C. OCT was removed and sections were stained with cresyl violet and dehydrated following the protocol described above (step 3.2–3.4). MC, condylar cartilage, and the mandibular bone regions were selected and isolated by LCM (Figure 2). The selected regions dropped into the lysis buffer in the cap of the collection tube. To obtain sufficient RNA for sequencing, we pooled 10 regions of MC, 10 regions of condylar cartilage or 4 regions of mandibular bone into each collection tube as one sample, respectively.

RNA was extracted using an RNA isolation kit following the manufacturer's instructions. Total RNA was analyzed using a bioanalyzer (Figure 3A–B). To test the effect of cresyl violet staining on the quality of RNA, we compared RNA from mandibular bone samples stained with cresyl violet and samples without staining (mineralized tissue is visible without staining, but cresyl violet staining enhances the visibility to distinguish bone from surrounding tissues). No significant difference in RNA integrity was observed between stained samples (n = 4) and samples without staining (n = 4), indicating the quick cresyl violet staining in this protocol has insignificant effect on RNA quality (P = 0.858, two-tailed Welch's t-test, Figure 3C). We used our protocol for LCM of various tissues at different developmental stages and RINs were measured, indicating high RNA quality (Figure 3D). Average yields of RNA from MC, condylar cartilage, and mandibular bone were 7.50 ± 1.45 ng, 12.55 ± 2.75 ng, and 33.02 ± 7.63 ng (Figure 3E) and the yield/area was 19.73 ± 3.82 ng/mm², 26.70 ± 5.84 ng/mm², and 17.23 ± 3.98 ng/mm², respectively (Figure 3F), without significant difference among tissues (MC versus condylar cartilage, P = 0.383; condylar cartilage versus mandibular bone, P = 0.260; MC versus mandibular bone, P = 0.674).

Libraries were prepared and sequenced as previously described^6,7. A representative cDNA size was approximately 500 bp (Figure 4A). RNA-seq data was analyzed with MultiQC⁸. We analyzed RNA-seq data from 18 LCM samples (MC1-6, Meckel's cartilage; C1-6, condylar cartilage; M1-6, mandibular bone). The mean quality values across each base position in the reads were generated by FastQC, indicating very good quality calls (Figure 4B). Read alignment was analyzed with Picard (Figure 4C). The reads showed high aligned percentages and the average percentage of aligned reads was 75%. Gene coverage was analyzed with Picard (Figure 4D), which indicated a good representation along the transcript from the 5' to the 3' end.

Differential gene expression analysis was performed^6,7. There were 4,006 genes significantly differentially expressed (P < 0.05) between the mandibular bone and MC (Figure 5A). Genes specific to osteoblasts or osteocytes (Col1a1⁹, Col1a2¹⁰, Dkk1¹¹, Dmp1¹², Dstn¹³, Runx2¹⁴, Sp7¹⁵, and Sparc¹⁶) were more highly expressed in the mandibular bone compared to MC, while chondrocyte-specific genes (Acan¹⁷, Col2a1¹⁸, Col9a1¹⁹, Col9a2²⁰, Col9a3²¹, Comp²², Lect1²³, and Sox5²⁴) were more highly expressed in MC compared to the mandibular bone (Figure 5B and Table 1). In addition, osteoclast markers such as Acp5²⁵, Csf1r²⁶, Ctsk²⁷, Itgb3²⁸, and Oscar²⁹ were also identified as more highly expressed in the mandibular bone compared to MC (Figure 5B and Table 1), indicating successful isolation of targeted tissues.

Figure 1: Representative cresyl violet staining of cartilage and bone in coronal sections of the mouse embryo at E16.5. (A) Meckel's cartilage and hemimandible. (B) Meckel's cartilage, condylar cartilage, and hemimandible. (C) Nasal septum and maxillae. (D) Cartilage primordium of presphenoid bone and palatine bone. (E) Cartilage primordium of radius, cartilage primordium of ulna, and costal cartilage. C = condylar cartilage; CC = costal cartilage; CP = cartilage primordium of presphenoid bone; M = hemimandible; MC = Meckel's cartilage; MX = maxilla; NS = nasal septum; PL = palatine bone; R = cartilage primordium of radius; U = cartilage primordium of ulna. Scale bar = 200 µm. Please click here to view a larger version of this figure.

Figure 2: Representative regions isolated by LCM and collected for RNA-seq. (A–C) Representative stained MC (A), condylar cartilage (B), and hemimandible with MC already isolated (C) in cryosection before LCM. (D–F) The regions in A, B, and C after targeted tissues were isolated by LCM. Scale bar = 200 µm. Please click here to view a larger version of this figure.

Figure 3: RNA quality and quantity of LCM samples. (A–B) Representative electropherogram (A) and the associated gel image (B) from a bioanalyzer for a mandibular bone sample. (C) RINs of total RNA from mandibular bone samples stained with cresyl violet (n = 4) or without staining (n = 4). (D) RINs of total RNA from different tissues isolated by LCM. Meckel's cartilage at E16.5 (n = 6); Condylar cartilage at E16.5 (n = 6); Mandibular bone at E16.5 (n = 6); Nasal septum cartilage at E14.5 (n = 6); Brain at E14.5 (n = 6); Brain at E16.5 (n = 7); Brain at E18.5 (n = 4). (E) Yield of total RNA from three tissues. Each MC or condylar cartilage sample was a pool of 10 microdissected regions of cartilage and each sample of mandibular bone was a pool of 4 microdissected regions of hemimandible. (F) The yield per unit area (ng/mm²) in each tissue. Data are mean ± s.e.m. Please click here to view a larger version of this figure.

Figure 4: The quality of libraries and RNA-seq data generated from LCM samples. (A) Representative cDNA sizes of a library from a mandibular bone sample determined by a bioanalyzer. (B–D) Quality control analysis of RNA-seq from 18 LCM samples (MC1-6, Meckel's cartilage; C1-6, condylar cartilage; M1-6, mandibular bone) by MultiQC. (B) The mean quality values across each base position in the reads were generated by FastQC. The background of the graph divides the y axis into very good quality calls (green), calls of reasonable quality (orange), and calls of poor quality (red). (C) Alignment of reads was analyzed by Picard. The summary is shown as the percentages of aligned reads. (D) Normalized gene coverage analyzed with Picard. The plot indicates the relative average coverage along the transcript from 5' end (left) to 3' end (right). Please click here to view a larger version of this figure.

Figure 5: Differential expression analysis of RNA-seq data from mandibular bone and MC isolated by LCM. (A) Hierarchical clustering of 4,006 genes significantly differentially expressed (P < 0.05) between the mandibular bone and MC. Three biological replicates were used for each tissue. M1-3, mandibular bone; MC1-3, MC. (B) Volcano plot showing fold changes and P-values of differentially expressed genes between mandibular bone and MC. Examples of highly differentially expressed cell-specific genes are shown: osteoblast and osteocyte markers in blue, osteoclast markers in green, and chondrocyte markers in red. Please click here to view a larger version of this figure.

Table 1: Differential expression of known genes in various cell types of bone and cartilage (mandibular bone versus MC).

LCM enables the isolation of enriched or homogenous cell populations from heterogeneous tissues. Its advantages include rapid and precise capture of cells in their in vivo context, while potential disadvantages include it being time consuming, expensive, and limited by the need for the user to recognize distinct subpopulations within a specified sample³⁰. This protocol provides details of LCM of mouse embryonic cartilage and bone, highlighting the use of cresyl violet staining in a rapid procedure to visualize cartilage and bone for precise tissue collection while maintaining high RNA integrity for subsequent analysis by RNA-seq. One limitation of this protocol is that the LCM system used here is not able to directly cut across mineralized tissues (e.g., the mandibular tissue in black in Figure 2C); as a result, the whole bone area needs to be dissected. Notably, the microdissected ossified regions isolated by LCM are not homogeneous, and include at least the subpopulations of osteoblasts, osteocytes, and osteoclasts (Table 1). Nevertheless, the enrichment by LCM is valuable as our previous study has shown that transcriptional changes in specific subpopulations such as osteoclasts in the microdissected bone tissue can still be detected by RNA-seq⁶.

For gene expression analysis, we have optimized the sample preparation and LCM procedure for high RNA quality and yield. Our protocol starts with fresh frozen tissues. Fresh frozen tissue sections allow for excellent quantity and quality of extracted RNA³, as mRNA is sensitive to standard methods of fixation³¹. RNA is quickly degraded by RNase contamination, and maintenance of an RNase-free environment throughout sample preparation, LCM, and RNA isolation is critical for successful application. Exposure to water during section processing is detrimental to RNA quality^31,32. Our protocol limits such exposure, with the highest level of aqueous exposure being 50% during cresyl violet staining. We have tested that a quick staining (30 s) with 0.1% cresyl violet in 50% ethanol gives a distinguishable color to cartilage and does not lower the RNA integrity for downstream analysis such as low input RNA-seq (Figure 3C and Figure 4). The yield/area (ng/mm²) is approximately 20 ng/mm² (Figure 3F), similar to or higher than previous optimized methods³³. According to the cell densities in MC and mandibular bone⁶, we estimate that with this protocol the average yield from one cell is approximately 5 pg RNA per cell, and 1–5 ng of total RNA can be extracted from 200–1,000 cells, which can be used for low-input RNA-seq^6,7,33. This yield efficiency is much higher than established LCM methods³⁴, and also superior or similar to recent optimized protocols^33,35.

The ability to distinguish different cell types in histological sections is essential for precise collection of specific tissues of interest. Cresyl violet is a hydrophilic, basic stain that binds to negatively charged nucleic acids³⁶. This property makes it useful for counterstaining with cell-type selectivity³⁷, and it is a standard histological stain for neurons³⁸. Cresyl violet has been used in LCM for a variety of tissues, providing good tissue morphology and high RNA quality^33,36,39,40. Compared with other staining methods, cresyl violet staining provides cytoplasmic and nuclear details, and a low RNA degradation rate³². We demonstrate here that cresyl violet staining is an easy and quick method to visualize cartilage and bone and that tissue stained with this method conserves high RNA integrity. Xylene treatment is commonly used for dehydration of the tissues before LCM^35,36,41,42. We use xylene to enhance the visualization of tissue morphology³⁵ which is essential to distinguish targeted tissues, especially on PEN membrane slides that are not as optically clear as plain glass slides. We found no significant occurrence of section loss from PEN slides during staining and washing steps, even with agitation to remove OCT.

Microdroplet or microfluidics-based single-cell RNA-seq (scRNA-seq) is a high-throughput method to analyze transcriptomes of tissues with heterogeneous cell types and has begun to be used in skeletal biology⁴³. In contrast to LCM, scRNA-seq involves enzymatic and/or mechanical disaggregation of tissue into single cells. Most protocols for single cell preparation require tissues to be incubated at room temperature or 37 °C for extended periods, which alters the transcriptome^44,45. In addition, isolation of single cells in cartilage and bone may be physically limited by skeletal element complexity of trabecularization and mineralization. It is still a technical challenge to isolate a sufficient number of viable cells from skeletal tissues that are accurately representative of the cellular diversity of tissues in vivo, and incomplete dissociation of the cells may cause bias in the detection of cell types⁴³. While scRNA-seq allows identification of distinct cell types, the sequencing depth is low, and typical sequencing approaches capture 3’ transcript ends and do not allow alternative splicing analysis. Bulk RNA-seq of LCM-derived tissue homogenizes cell differences, but allows comprehensive transcript detection and alternative splicing analysis. LCM is therefore especially useful for skeletal tissues that are not easily separable from surrounding tissues and internally complex, and fresh frozen preparation of the tissue can preserve both transcriptional profiles and RNA integrity for transcriptional analysis. Another weakness of scRNA-seq is that after single cell preparation, the spatial information of the cells is lost. A combination of LCM and scRNA-seq has been developed to permit the study of the transcriptome of a small sample from defined geographical locations (Geo-seq)⁴⁶, which is another approach of utilizing LCM to study regionalized gene expression.

In summary, this protocol provides details of optimized LCM of cartilage and bone, highlighting the use of cresyl violet staining in a rapid procedure to visualize cartilage and bone for precise tissue collection while maintaining high RNA integrity for subsequent analysis by RNA-seq. This protocol has been used successfully for LCM of cartilage and bone at different developmental stages for gene expression analysis^6,7, and also can be used for other tissues.