Neurospheres grown as 3D cultures constitute a powerful tool to study glioma biology. Here we present a protocol to perform immunohistochemistry while maintaining the 3D structure of glioma neurospheres through paraffin embedding. This method enables the characterization of glioma neurosphere properties such as stemness and neural differentiation.
Analysis of protein expression in glioma is relevant for several aspects in the study of its pathology. Numerous proteins have been described as biomarkers with applications in diagnosis, prognosis, classification, state of tumor progression, and cell differentiation state. These analyses of biomarkers are also useful to characterize tumor neurospheres (NS) generated from glioma patients and glioma models. Tumor NS provide a valuable in vitro model to assess different features of the tumor from which they are derived and can more accurately mirror glioma biology. Here we describe a detailed method to analyze biomarkers in tumor NS using immunohistochemistry (IHC) on paraffin-embedded tumor NS.
Gliomas are primary solid tumors of the central nervous system classified by their phenotypic and genotypic characteristics according to the World Health Organization1. This classification incorporates the presence or absence of driver mutations, which represent an attractive source of biomarkers2. Biomarkers are biological characteristics that can be measured and evaluated to indicate normal and pathological processes, as well as pharmacological responses to a therapeutic intervention3. Biomarkers can be detected in tumor tissue and cells derived from glioma, enabling its biological characterization in different aspects. Some examples of glioma biomarkers include mutated isocitrate dehydrogenase 1 (IDH1), a mutant enzyme characteristic of lower-grade gliomas associated with better prognosis4. Mutated IDH1 is frequently expressed in combination with TP53 and alpha-thalassemia/mental retardation syndrome X-linked (ATRX)-inactivating mutations, defining a specific glioma subtype4,5. ATRX-inactivating mutations also occur in 44% of pediatric high-grade gliomas2,6 and have been associated with aggressive tumors and genomic instability6,7. In addition, pediatric diffuse intrinsic pontine gliomas (DIPG) are differentiated in subgroups according to the presence or absence of a K27M mutation in histone H3 gene H3F3A, or in the HIST1H3B/C gene1,6. Therefore, the introduction of molecular characterization permits the subdivision, study, and treatment of gliomas as separate entities, which will more allow the development of more targeted therapies for each subtype. In addition, the analysis of biomarkers can be used to evaluate different biological processes such as apoptosis8, autophagy9, cell cycle progression10, cell proliferation11, and cell differentiation12.
Genetically engineered animal models that harbor the genetic lesions present in human cancers are critical for the study of signaling pathways that mediate disease progression. Our laboratory has implemented the use of the Sleeping Beauty (SB) transposase system to develop genetically engineered mouse models of glioma harboring specific mutations that recapitulate human glioma subtypes13,14. These genetically engineered mouse models are used for generating tumor-derived NS, which enable in vitro studies in a 3D system, mirroring salient features present in the tumors in situ15. Also, the orthotopical transplantation of NS into mice can generate secondary tumors that retain features of the primary tumor and mimic the histopathological, genomic, and phenotypic properties of the corresponding primary tumor15.
In serum-free culture, brain tumor cells with stem cell properties can be enriched, and in the presence of epidermal growth factors (EGF) and fibroblast growth factors (FGF), they can be grown as single cell-derived colonies such as NS cultures. In this selective culture system, most differentiating or differentiated cells rapidly die, whereas stem cells divide and form the cellular clusters. This allows the generation of a NS culture that maintains glioma tumor features16,17,18. Glioma NS can be used to evaluate several aspects of tumor biology, including the analysis of biomarkers that have applications in diagnosis, prognosis, classification, state of tumor progression, and cell differentiation state. Here, we detail a protocol to generate glioma NS and embed the 3D cultures in paraffin to be used for IHC staining. One advantage of fixing and embedding glioma NS is that the morphology of NS is maintained better compared to the conventional cytospin method19, in which glioma NS are subjected to stressful manipulation for cell dissociation and become flattened. In addition, embedding quenches any endogenous fluorophore expression from genetically engineered tumor NS, enabling staining across the fluorescent spectra. The overall goal of this method is to preserve the 3D structure of glioma cells through a paraffin embedding process and enable characterization of glioma neurospheres using immunohistochemistry.
All methods described here have been approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Michigan.
1. Generation of Brain Tumor Neurospheres Derived from a Mouse Model
2. Maintaining and Passaging the NS
NOTE: Prevent overgrowth and maintain cells at relatively high density. Confluence is determined by spheres size (black centers of large spheres mean overgrowth). The growth rate of the NS will depend on the oncogenes used to generate tumors.
3. Freezing NS for Long-term Storage
4. Fixation of Tumor NS
5. Paraffin Embedding of Tumor NS
6. Sectioning of Paraffin-embedded NS
7. Immunohistochemistry (IHC) of Paraffin-embedded NS
To monitor tumor development and evaluate if the tumor has reached the size necessary to generate NS, we analyzed the luminescence emitted by tumors using bioluminescence. This allows the study of transduction efficiency in the pups and tumor progression by the luminescence signal of luciferase (Figure 1A and 1B). When the tumor size reaches a signal of 1 x 107 photons/s/cm2/sr (Figure 1B), the brain can be processed for NS generation. Another advantage of this system is that the coding sequence of different fluorescent proteins can be added to the plasmids used for gene delivery and tumor generation. We can then confirm the expression of the inserted genetic alterations, using a dissecting microscope equipped with a fluorescent lamp (Figure 1C). The NS generated from wt-IDH1 (Figure 1D) and mIDH1 (Figure 1E) tumors are identified by the characteristic spherical morphology, and the expression of fluorophores associated with the specific genetic lesion (i.e., GFP associated with shATRX and shp53, and Katushka associated with mIDH1; Figure 1D and 1E). To proceed with the IHC on NS, cells are fixed, then embedded in agarose and paraffin (Figure 2). Embedding the NS in paraffin blocks has several advantages, including the preservation of 3D structures of the NS and allowing for long-term storage. The embedded NS can be stained for specific glioma biomarkers. Using this technique, we analyzed the expression of ATRX, Ki67 and OLIG2 (oligodendrocyte differentiation marker) (Figure 3A-3C). We confirmed the low expression levels of ATRX protein in our NS cultures (Figure 3A), which is a feature of the LGG glioma subtype currently being studied4. Also, Ki-67, a well-described biomarker of cell proliferation, was positive in the mIDH1 NS (Figure 3B), demonstrating active proliferation, which is a key feature of tumor cells. Interestingly, cells localized in the center of the larger glioma NS clusters were negative for Ki-67 (Figure 3B), indicating that they were not proliferating, unlike the Ki-67 positive cells localized to the periphery. This phenomenon may be due to the fact that peripheral cells do not have access to nutrients from the growth media. When they reach this larger size, the glioma NS should be dissociated and passaged. Finally, wt-IDH1 NS were positive for Olig2, a transcription factor participating in oligodendrocyte differentiation (Figure 3C).
Figure 1: Glioma neurospheres generated from GFP- and Katushka-positive sleeping beauty (SB) brain tumor. (A) Luminescent signal of a neonate mouse injected with SB-transposase plasmid in combination with plasmids harboring genetic lesions to induce glioma (NRAS/shP53/shATRX/IDH1-R132H). (B) Luminescence indicating tumor development at 132 days post-injection (DPI). (C) A bright-field and fluorescent image of a brain harvested from a mouse that reached the end-point stage. The tumor area is delineated by the dotted line and positive for the fluorescent reporters, GFP and Katushka. (D and E) Tumor neurospheres generated from SB brain tumors expressing GFP associated to shP53 and shATRX; and Katushka associated to IDH1-R132H. Scale bar = 100 µm. Please click here to view a larger version of this figure.
Figure 2: Representation of the workflow for paraffin-embedded glioma NS. NS are collected from a T-25 flask into a 15 mL conical tube and fixed in 10% formalin. Then, NS are washed with HBSS and embedded in 2% agarose. Polymerized agarose containing glioma NS are placed in a cassette for tissue processing and paraffin embedding. Please click here to view a larger version of this figure.
Figure 3: Representative results of immunohistochemistry with HRP-DAB detection of biomarkers. (A-C) IHC performed on paraffin-embedded mIDH1 glioma NS stained for ATRX (A) and Ki67 (B), and wt-IDH1 glioma NS stained for OLIG2 (C). Red arrows indicate positive cells for the biomarker staining. Scale bar = 50 µm. Please click here to view a larger version of this figure.
In this article, we detail a versatile and reproducible method to perform immunohistochemistry on paraffin-embedded glioma NS, which maintain the characteristic 3D structure of tumor cells growing in the brain in situ. This technique possesses the following advantages over using cells plated onto glass or plastic surfaces: i) preservation of the 3D structure of NS; ii) avoiding of stress of cell dissociation before analysis of protein expression; iii) quenching of any endogenous fluorophore expression from genetically engineered tumor NS, enabling staining across the fluorescent spectra; and iv) long-term storage.
A critical step of this technique is to ensure that the cells are re-suspended well in the agarose to ensure that they are homogenously distributed and not clumped together. One limitation of this technique is the time-consuming steps related to cell culturing and NS processing before the IHC staining procedure. A second limitation is the high number of cells needed to execute this method. If the number of cells is less than 1 x 106 cells, there will be too few cells per field of view NS clusters in each section. Therefore, it could be difficult to reach the ideal number of cells if the glioma cell culture is slow-growing. We used a confluent T-25 flask containing at least 3.0 x 106 cells. The number of cells used can be modified as desired; however, it is necessary to obtain a NS pellet of appropriate size for the embedding procedure. After this, the embedded NS can be manipulated as a regular paraffin-embedded block, and it can then proceed to sectioning and IHC staining to analyze protein expression by immunofluorescence or IHC using DAB staining.
After sectioning of the blocks and following the protocol for IHC, sections should be counterstained with hematoxylin (blue). If tissue expresses the target protein, a brown signal should be visible (Figure 3). If after IHC the sections are too brown (high background), this could be due to 1) non-specific biding of the secondary antibody, 2) an incomplete quenching of the endogenous peroxidase, or 3) insufficient blocking. These issues can be solved by 1) using a negative control slide (incubated only with a secondary antibody) to check non-specific binding of the secondary antibody, 2) increasing the peroxidase quenching time, or 3) increasing the blocking incubation period. On the other hand, if no signal is detected, it is possible that the epitopes are still masked, which may be related to the type of buffer used for antigen retrieval. In our experience, we have had satisfactory staining using citrate buffer, but this can be tissue- and antibody-dependent, and different formulations for antigen retrieval buffers can be used such as 0.1 M Tris-HCl (pH 8.0) buffer21. These are some of the most common issues related to IHC staining, but if other difficulty arises, troubleshooting should be carried out as with any other IHC (i.e., trying different dilutions of the primary and secondary antibodies, different blocking reagents, different blocking times, etc).
An alternative method to evaluate protein expression using paraffin-embedded cells was previously described by Hasselbach et al.20. This method does not include the agarose steps, which in this protocol enable adequate special distribution of the 3D cell clusters. We also describe an accurate method for NS culturing, passaging, and freezing for long-term storage. This method also illustrates how to collect, fix, and embed NS without altering 3D structure (Figure 2).
Since the NS can be generated from glioma tissue obtained from patients and glioma animal models, this technique constitutes a powerful tool that can be used to analyze biomarker expression in different glioma suptypes and validate the original models. Here we describe the technique with NS originated from genetically engineered mice using the SB transposase system (Figure 1). However, this technique can also be used to analyze protein in paraffin-embedded human glioma cells that grow as 3D cultures without modifications to the protocol, other than the cell culture conditions. So, this method may also be applied to 3D cultures obtained from transplantable animal models, human PDX models, and other genetically engineered models. Lastly, this method may be used with other types of cancer, both mouse models and human cancers, in addition to gliomas.
The authors have nothing to disclose.
This work was supported by National Institutes of Health/National Institute of Neurological Disorders & Stroke (NIH/NINDS) Grants R37-NS094804, R01-NS074387, R21-NS091555 to M.G.C.; NIH/NINDS Grants R01-NS076991, R01-NS082311, and R01-NS096756 to P.R.L.; NIH/NINDS R01-EB022563; 1UO1CA224160; the Department of Neurosurgery; Leah's Happy Hearts to M.G.C. and P.R.L.; and RNA Biomedicine Grant F046166 to M.G.C. F.M.M. is supported by an F31 NIH/NINDS-F31NS103500. J.P. was supported by a Fulbright-Argentinian Ministry of Sports and Education Fellowship.
Coated microtome blade HP35 | Thermo Scientific | 3150734 | |
Microtome RM 2135 | Leica | MR2135 | |
Formalin solution, neutral buffered, 10% | Sigma-aldrich | HT501128-4L | |
Rabbit polyclonal anti-ATRX, | Santa Cruz Biotechnology | sc-15408 | IHC, 1:250 dilution |
Rabbit polyclonal anti-Ki-67 | Abcam | Ab15580 | IHC, 1:1000 dilution |
Rabbit polyclonal anti-OLIG2 | Millipore | AB9610 | IHC, 1:500 dilution |
Goat polyclonal anti-rabbit biotin-conjugated | Dako | E0432 | IHC, 1:1000 dilution |
Vectastain Elite ABC HRP kit | Vector Laboratories Inc | PK-6100 | |
BETAZOID DAB CHROMOGEN KIT | Biocare medical | BDB2004 L/price till 12/18 | |
N-2 Supplement | ThermoFisher | 17502048 | |
N-27 Supplement | ThermoFisher | A3582801 | |
Accutase® Cell Detachment Solution | Biolegend | 423201 | |
AGAROSE LE | GoldBio | A-201-1000 | |
Genesee Sc. Corporation | Olympus 15 ml | 21-103 | |
Genesee Sc. Corporation | TC-75 treated Flask | 25-209 | |
Genesee Sc. Corporation | TC-25 treated Flask | 25-207 | |
DMEM/F12 | Gibco | 11330-057 | NS media |
HBSS | GibcoTM | 14175-103 | balanced salt solution |
C57BL/6 | Taconic | B6-f | C57BL/6 mouse |