Laser microdissection (LMD) is a sensitive and highly reproducible technique that can be used to uncover pathways that mediate glioma heterogeneity and invasion. Here, we describe an optimized protocol to isolate discrete areas from glioma tissue using laser LMD followed by transcriptomic analysis.
Gliomas are primary brain tumors characterized by their invasiveness and heterogeneity. Specific histological patterns such as pseudopalisades, microvascular proliferation, mesenchymal transformation and necrosis characterize the histological heterogeneity of high-grade gliomas. Our laboratory has demonstrated that the presence of high densities of mesenchymal cells, named oncostreams, correlate with tumor malignancy. We have developed a unique approach to understand the mechanisms that underlie glioma's growth and invasion. Here, we describe a comprehensive protocol that utilizes laser capture microdissection (LMD) and RNA sequencing to analyze differential mRNA expression of intra-tumoral heterogeneous multicellular structures (i.e., mesenchymal areas or areas of tumor invasion). This method maintains good tissue histology and RNA integrity. Perfusion, freezing, embedding, sectioning, and staining were optimized to preserve morphology and obtain high-quality laser microdissection samples. The results indicate that perfusion of glioma bearing mice using 30% sucrose provides good morphology and RNA quality. In addition, staining tumor sections with 4% Cresyl violet and 0.5% eosin results in good nuclear and cellular staining, while preserving RNA integrity. The method described is sensitive and highly reproducible and it can be utilized to study tumor morphology in various tumor models. In summary, we describe a complete method to perform LMD that preserves morphology and RNA quality for sequencing to study the molecular features of heterogeneous multicellular structures within solid tumors.
Gliomas are the most aggressive primary tumors of the central nervous system. They are highly invasive and heterogeneous1. Analysis of cellular and molecular components of the tumor will reveal novel therapeutic targets.
Among different methods currently available, laser capture microdissection (LMD) of frozen brain tumor tissue is a cost-effective, reliable technique that allows the isolation of discrete anatomical areas or specific cell populations from tumor tissues to study their molecular profile2,3. LMD allows the analysis of mRNA gene expression profiles of selected single cells or multicellular structures4,5. LMD can be utilized to gain in-depth mechanistic knowledge about the molecular events that take place during tumor progression. Improvement in processing of tumor tissues is necessary to obtain optimal optical resolution of tissue morphology and RNA-quality6. Although paraformaldehyde fixation is the best option for morphological analysis, RNA quality is affected and degraded under these conditions, resulting in poor RNA quality for RNA-seq analysis. The use of frozen tissue sections avoids ice crystal formation, which could break cell membranes and produce holes within cells, and remains the best option for RNA-Seq analysis7.
Here, we describe an optimized section fixation and staining method to process frozen mouse brain tumor tissues for LMD. To prevent ice crystals from forming in the tissue, we perfused mice with a solution of 30% sucrose. This solution disrupts interactions between polar water molecules and prevents the formation of ice crystals, preserving the tissue morphology. Tissue staining is necessary to differentiate and obtain specific population of cells or anatomically distinct areas within the tumor. It is essential to fix and stain the tissue with innocuous dyes to maintain RNA integrity. It has been previously shown that staining tissue with hematoxylin/eosin (H&E) deteriorates RNA integrity8. We fixed and stained the tissue of interest with ethanol, Cresyl violet 4% and eosin Y 0.5% solutions. Cresyl violet is an acidophilic dye that stains the cell nucleus with a dark blue color. Eosin Y is a basophilic dye that stains basic components of the cells, providing a distinction between cytoplasm and other cellular structures8. Both dyes are soluble in ethanol and do not deteriorate RNA quality. To avoid tissue damage and maintain high optical resolution of the cellular structures, we mounted the tissue sections prior to LMD9.
All methods described here that use experimental animals have been approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Michigan.
NOTE: Glioma neurospheres generated from a GEMM or stable lines can be used for intracranial tumor engraftment in mice10 and processed for LMD and RNA sequencing. These cells constitutively express firefly luciferase and GFP proteins, which will be further utilized for tumor growth analysis and localization.
1. Generation of intracranial mouse glioma model from neurospheres derived from genetically engineered glioma models
2. Animal perfusion and brain preservation
3. Cryopreservation of brains harboring glioma tumors
4. Sectioning frozen brain tumor tissues
5. Fixation and staining of cryopreserved brain tissue sections
6. Laser capture microdissection
NOTE: A laser capture microdissection microscope needs to be utilized to laser microdissect specific areas of interest within the tumor tissue. To minimize the time for tissue laser microdissection, have the LMD microscope prepared before fixation and staining.
7. RNA isolation of micro-dissected glioma tissue
8. RNA quality control, library preparation and RNA-Seq analysis
Our laboratory has generated a genetically engineered mouse models (GEMMs) using the sleeping beauty transposase system (Figure 1A). This system incorporates specific genetic alterations into the genome of neural progenitor cells in neonatal mice. These altered progenitor cells form endogenous glioma tumors. Plasmid sequences used to generate the tumors were: (1) pT2C-LucPGK-SB100X for Sleeping Beauty transposon & luciferase expression, (2) pT2-NRASSV12 for NRAS expression, (3) pT2-shp53-GFP4 for p53 knock-down and GFP protein expression, and (4) pT2-shATRx-GFP4 for ATRX knock-down. Plasmids were injected into the lateral ventricle of 1-day old neonatal pups as described previously11. Plasmid uptake and tumor formation was monitored via in vivo bioluminescence imaging system. Once tumor-bearing mice displayed signs of tumor burden, they were sacrificed. Tumors were either used to generate neurosphere cultures or directly cryo-preserved for LMD processing (Figure 1A).
Cell cultures started from the GEMM were used to generate a translatable glioma model (Figure 1B). Glioma neurospheres derived from GEMM tumors were cultured and implanted intracranially into the striatum of immune-competent mice. Single cell suspensions obtained from neurospheres culture were used to generate glioma tumors by intracranial implantation as described before by our laboratory10,11,12. This methodology allows the careful quantification of the number of cells to be implanted per mouse (30,000 cells/1 µL/mouse). This protocol permits the reproducibility of the results between different experimental implantations. However, implantation of neurospheres could be an alternative option to generate mouse glioma tumors and subsequent LMD and RNA-Seq analysis. Nevertheless, this method is considered to be more accurate. Tumor progression was monitored by in vivo bioluminescence spectrum imaging system. Mice displaying signs of tumor burden were transcardially perfused and the brain was cryo-preserved for LMD processing (Figure 1C).
Tissue sections were laser microdissected to characterize the transcriptome of multicellular structures within gliomas. The perfusion, freezing and embedding procedures described were optimized to preserve tissue morphology and obtain good quality RNA after laser microdissection. Different perfusion approaches were evaluated in order to acquire tissues with superior morphology and RNA integrity (Table 10). To dissect the areas of interest, it was necessary to stain the tissue with innocuous dyes for RNA (Figure 2). We observed that perfusing tumor bearing mouse with Tyrode's solution for 5 min, and then 30% sucrose for 15 min, followed by overnight storage of the dissected brain in 30% sucrose preserves morphology and RNA integrity of the tumor tissue (Figure 3D). Perfusion with 30% sucrose solution prevented ice-crystal formation within the tissue. Although, paraformaldehyde tissue fixation resulted in high quality tissue morphology, RNA integrity was negatively affected (Figure 3B). Other approaches such as Tyrode's solution for 5 min or Tyrode's solution for 5 min + 30% sucrose solution for 15 min (Figure 3A and C) did not affect the RNA quality. Under these conditions, brains were not preserved in 30% sucrose overnight; we observed reduced resolution in tissue morphology.
We performed various staining techniques followed by RNA integrity quality control analysis. We observed that 4% Cresyl violet and 0.5% eosin Y staining was sufficient to identify glioma multicellular structures and maintain RNA integrity. Cresyl violet is an acidophilic dye that stains the nucleus of cells with a dark blue color. Eosin Y is a basophilic dye that stains basic components of the cells.
We observed that if the tissue was not mounted with mounting medium, the sections became dehydrated and the morphology deteriorated (Figure 4A). To maintain high quality tissue morphology, we mounted the tissue with mounting medium. We observed that 15% mounting medium dissolved in water (30 µL in 200 µL of water) maintained high quality tissue morphology (Figure 4B).
In Figure 5A, areas of glioma heterogeneity are shown. Images were acquired using a laser capture microdissection microscope preceding dissection. Red lines depict areas with abundant mesenchymal cells (elongated cells). Blue lines depict areas with no mesenchymal cells (rounded cells) (Figure 5A, middle image). Figure 5B shows images of laser micro-dissected tumor areas (red lines) and normal brain tissue areas (blue lines). ROI selection is made to dissect areas of interest (Figure 5A and B, lower images). We can observe in these images the success in the dissection of the selected areas.
RNA extraction was performed using a commercial kit. It was determined that a total area of dissected tissue between 2.5 x 106 – 7 x 106 µm2 was required for mRNA extraction and cDNA library preparation for subsequent RNA sequencing. RNA quality control determined a RIN between 6 to 7 in glioma tissue after laser microdissection. A RIN of 6 was determined to be appropriate for cDNA library preparation. Following RNA extraction, a kit for RNA isolation at picomolar concentrations was utilized to generate a cDNA library suitable for next generation sequencing.
0.25 ng – 10 ng (1-8 µL) of RNA |
4 µL of the 5x First-Strand Buffer |
1 µL of the Oligos Mix V2 |
Nuclease-free water to final volume of 13 µL |
Table 1: RNA quality control and library preparation: Master Mix preparation for step 8.3.
4.5 µL Mix V2 |
0.5 µL RNase inhibitor |
2 µL Reverse Transcriptase |
Final volume of 7 µL |
Add the 7 µL to the RNA sample tube |
Table 2: RNA quality control and library preparation: Master Mix preparation for step 8.5.
42 °C for 90 min |
70 °C for 10 min |
4 °C hold |
Table 3: RNA quality control and library preparation: Thermocycler conditions for step 8.5.
2 µL nuclease-free water |
25 µL 2x PCR Buffer |
1 µL DNA Polymerase for each reaction |
Mix gently and spin down briefly |
Add 28 µL of the master mix to each of the sample tubes |
Add 1 µL of each 5' and 3' primer to each tube |
Mix the reactions gently and spin down briefly |
Table 4: RNA quality control and library preparation: Master Mix preparation for step 8.6.
1 cycle |
94 °C for 1 min |
5 cycles |
98 °C for 15 s |
55 °C for 15 s |
68 °C for 30 s |
1 cycle |
68 °C for 2 min, 4 °C hold |
Table 5: RNA quality control and library preparation: Thermocycler conditions for step 8.6.
16.8 µL nuclease-free water |
2.2 µL 10x R Buffer |
1.5 µL R v2 |
1.5 µL R-Probes v2 |
Table 6: RNA quality control and library preparation: Master Mix preparation for step 8.18.
37 °C 60 min |
72 °C 10 min |
4 °C hold |
Table 7: RNA quality control and library preparation: Thermocycler conditions for step 8.20.
26 µL nuclease-free water |
50 µL CB PCR Buffer |
2 µL PCR2 Primers v2 |
2 µL DNA Polymerase |
Table 8: RNA quality control and library preparation: Master Mix preparation for step 8.21.
1 cycle |
94 °C for 1 min |
X cycles |
98 °C for 15 s |
55 °C for 15 s |
68 °C for 30 s |
1 cycle |
68 °C for 2 min, 4 °C hold |
Table 9: RNA quality control and library preparation: Thermocycler conditions for step 8.21.
Step 1 | Step 2 | Step 3 | |
Method 1 | Tyrode's 15' | — | — |
Method 2 | Tyrode's 5' | Sucrose 30% 15' | — |
Method 3 | Tyrode's 5' | 4% PFA 10' | — |
Method 4 | Tyrode's 5' | 30% Sucrose 15' | Brain in 30% Sucrose Overnight |
Table 10: Methods for different perfusion approaches.
Figure 1: Mouse glioma models used for laser microdissection. (A) Sleeping beauty GEMM used to develop de novo glioma. Images display tumor progression: (i) plasmid injection into the lateral ventricle of neonates, (ii) tumor burden. (B) Generation of transplantable glioma model using glioma cell cultures. Neurosphere cells cultured from GEMM are implanted intracranially into the striatum. (C) Depiction of cryo-sectioning of the brain embedded in coronal orientation following perfusion. Please click here to view a larger version of this figure.
Figure 2: Schematic representation of the method used to stain cryo-preserved tissue sections with Cresyl violet and Eosin. Slides used for staining were handled with tools cleaned with RNase-free water. All solutions were prepared with RNase/DNase free water on day of staining. Cresyl violet stains nuclei violet and eosin Y stains cytoplasm pink. Dyes were dissolved in 70% ethanol. Please click here to view a larger version of this figure.
Figure 3: Comparison of various perfusion approaches to preserve tissue morphology and RNA integrity. H&E images represent comparative morphology of brain tissue that was subjected to perfusion under different methods: (A) Tyrode's solution for 15 min, (B) Tyrode's solution for 5 min, 4% PFA for 10 min, (C) Tyrode's solution for 5 min, 30% sucrose for 15 min, and (D) Tyrode's solution for 5 min, 30% sucrose for 15 min with overnight immersion in 30% sucrose. RNA quality assessment is displayed for each perfusion method. Plots depict 18s and 28s rRNA peaks and the initial marker peak. The ratio of 18s and 28s rRNA is used to determine RNA quality. A gel image of the RNA fragments is displayed to the right of the plots. RNA quality was assessed using the RIN values. RNA quality was high in methods A, C, and D. Tissue morphology was superior in methods C and D. Please click here to view a larger version of this figure.
Figure 4: Representative images of mounted and non-mounted glioma tissue. (A) Representative images of tissue stained with H&E. This tissue was left unmounted and became dehydrated displaying poor morphology with breaks in the tissue. (B) Representative images of tissue stained with H&E and mounted with mounting medium. Please click here to view a larger version of this figure.
Figure 5: Representative images of glioma tissue used for laser microdissection. (A-B) Representative images of tissue stained with H&E and areas of multicellular structures selected for LMD. (A) On the left, areas of elongated cells (red) and control areas (blue) were selected for LMD. (B) On the right, areas of collective invasion (red) and control areas (blue) were selected for LMD. (C) Representative RNA quality control for laser microdissected areas. A total area of 6.5 x 106 µm2 was microdissected. The analysis shows RNA concentration of 2,346 pg/µL and RNA Integrity Number (RIN): 6.8. Please click here to view a larger version of this figure.
Understanding the molecular mechanisms underlying glioma heterogeneity and invasion are of critical importance to uncover novel therapeutic targets13. In this manuscript, we describe a detailed and optimized method to analyze the molecular landscape of glioma heterogeneity and invasion using laser capture microdissection (LMD) followed by transcriptomic analysis.
Laser capture microdissection (LMD) can be used to identify different areas or single cells within the tumor, providing a specific sample to further analyze the molecular pattern maintaining the spatial context of the tumor14. This technique is reliable and low-priced compared with other methods used to analyze the spatial transcriptome in solid tumors15. We analyzed glioma heterogeneity and invasion in genetically engineered mouse tumor models or intracranial implantable models using LMD. These models recapitulate the salient characteristics of human gliomas, allowing the study of glioma heterogeneity and invasion10,12.
Maintaining high-quality tumor tissue morphology and RNA integrity is one of the limitations for LMD. To improve tissue morphology, we perfused mice with Tyrode's solution for 5 minutes, followed by 30% sucrose dissolved in the same solution. Once the brain was dissected, we preserved it in 30% sucrose dissolved in RNase/DNase-free water overnight or until the brain reached the bottom of the storage container. These steps significantly improve the morphology of the tissue and reduced the formation of ice-crystals in the tissue during cryo-preservation. Although human glioma tissue was not used for this protocol, incubation of human samples in 30% sucrose solution overnight could be a feasible methodology to improve cryosections morphology. Another possible limitation for LMD is the preservation of RNA integrity post-staining8. Although other research teams have performed laser microdissection on glioma tissue followed by RNA-seq analysis they do not illustrate any RNA and/or morphology nor do they comment on morphological quality, or particular controls for glioma frozen sections16. In this protocol accurate morphological identification was essential in order to laser microdissect precise macrocellular structures within gliomas. We observed that fixing glioma tissue with ethanol solution and staining it with 4% Cresyl violet and 0.5% eosin Y dissolved in ethanol-maintained RNA quality and enabled the microscopic identification of single cells and multicellular structures. We demonstrated that mounting glioma sections with mounting solution is a critical step which prevents cracking and fissure formation in the tissue. Please note that the mounting medium needs to be prepared in water, as using the mounting medium dissolved in ethanol will result in poor tissue morphology. Laser microdissection has to be performed as fast as possible in an RNase-free environment. We also recommend to section up to 2.5 x 106 µm2 total tumor/tissue area in order to obtain appropriate amounts of RNA, both for RNA quality control and transcriptomic analysis.
LMD enables the analysis of the molecular signaling pathways that regulate glioma heterogeneity and invasion. This analysis could reveal novel potential targets for diagnosis, prognosis and future translational development in preclinical glioma models.
The authors have nothing to disclose.
Work was supported by National Institutes of Health, (NIH/NINDS) Grants: R37-NS094804, R01-NS105556, R21-NS107894 to M.G.C.; (NIH/NINDS) Grants R01-NS076991, R01-NS096756, R01-NS082311 to P.R.L.; (NIH/NIBI): R01-EB022563; (NIH/NCI) U01CA224160; the Department of Neurosurgery, Rogel Cancer Center at The University of Michigan, ChadTough Foundation, and Leah's Happy Hearts Foundation to M.G.C. and P.R.L. RNA Biomedicine Grant F046166 to M.G.C. National Institutes of Health, UL1 TR002240 for Michigan Institute for Clinical and Health Research (MICHR), Postdoctoral Translational Scholars Program (PTSP) grant, Project F049768 to A.C. University of Michigan Forbes Cancer Research Institute, a Physician-Scientist Award from Research to Prevent Blindness, Inc. (RPB), grant R01 EY022633 from the NEI of the NIH (AK), and an unrestricted grant from RPB to the Department of Ophthalmology and Visual Sciences. This research utilized the Vision Research Core (P30 EY007003), and the Cancer Center Research Core (P30 CA046592). AK is supported by the Mrs. William Davidson Emerging Scholar Award from the A. Alfred Taubman Medical Research Institute.
Accutase Cell Detachment Solution | Biolegend | 423201 | |
Animal-Free Recombinant Human EGF | Peprotech | AF-100-15 | |
Antibiotic-Antimycotic (100X) | Gibco | 15240062 | |
B-27 Supplement (50X), serum free | Gibco | 17504044 | |
Buffer RLT | Qiagen | 79216 | |
Corning PCR Tubes | Sigma Aldrich | CLS6530 | |
Cresyl Violet Acetate | Sigma Aldrich | C5042 | |
DMEM/F12 – Dulbecco's Modified Eagle Medium: Nutrient Mixture F-12 | Gibco | 11330057 | |
Eosin Y | Sigma Aldrich | E4009 | |
HiSeq 4000 | Illumina | N/A | |
Laser Microdissection (LMD) System | Leica | LMD7000 | |
N-2 Supplement (100X) | Gibco | 17502048 | |
Normocin – Antimicrobial Reagent | Invivogen | ant-nr-1 | |
Peel Away Disposable Embedding Molds | Electron Microscopy Sciences | 70182 | |
PEN Membrane Glass Slide (2 µm) | Lieca | 1150518 | |
Pinpoint Solution | Zymo Research | D3001-1 | |
Recombinant Human FGF-basic | Peprotech | 100-18B-1MG | |
Research Cryostat | Leica | CM3050s | |
RNaseZap RNase Decontamination Solution | Fisher Scientific | AM9780 | |
RNeasy Plus Micro Kit | Qiagen | 74034 | |
SMARTer Stranded Total RNA-Seq Kit v2 – Pico Input Mammalian | Takara Bio | 634411 | |
Tissue-Plus O.C.T. Compound | Fisher Scientific | 23-730-571 |