Immunology and Infection
Published: November 29th, 2018
We provide a method for the generation, cultivation and systematic analysis of organotypic slices derived from murine lung tumors. We also describe how to optimize for slice thickness, and how to select drug concentrations to treat tumor slices.
Organotypic primary tissue explant cultures, which include precision-cut slices, represent the three-dimensional (3-D) tissue architecture as well as the multicellular interactions of native tissue. Tissue slices immediately cut from freshly resected tumors preserve spatial aspects of intratumor heterogeneity (ITH), thus making them useful surrogates of in vivo biology. Careful optimization of tissue slice preparation and cultivation conditions is fundamental for the predictive diagnostic potential of tumor slice explants. In this regard, murine models are valuable, as these provide a consistent flow of tumor material to perform replicate and reproducible experiments. This protocol describes the culturing of murine lung tumor-derived tissue slices using a rotating incubation unit, a system that enables intermittent exposure of tissues to oxygen and nutrients. Our previous work showed that the use of rotating incubation units improves the viability of tissue compared to other culture methods, particularly floating slices and stagnant filter supports. Here, we further show that slice thickness influences the viability of cultured slices, suggesting that optimization of slice thickness should be done for different tissue types. Pronounced ITH in relevant oncogenic functions, such as signaling activities, stromal cell infiltration or expression of differentiation markers, necessitates evaluation of adjacent tissue slices for the expression of markers altered by drug treatment or cultivation itself. In summary, this protocol describes the generation of murine lung tumor slices and their culture on a rotating incubation unit and demonstrates how slices should be systematically analyzed for the expression of heterogeneous tissue markers, as a prerequisite prior to drug response studies.
Solid tumor tissues, including lung cancer, exhibit genetic and phenotypic heterogeneity, and harbor complex microenvironments1,2. The interplay between tumor cells and their surrounding microenvironment influence on drug sensitivity and resistance mechanisms3. This highlights the need for preclinical models that can accurately model biological complexities and functions acting in native tumors. Precision-cut slices immediately derived from fresh tumors provide a unique resource, as they have a principal capacity to represent in vivo biology, at least for a short window of time, including phenotypes spatially distributed in an individual tumor. Resected clinical tumors are one of the few personalized specimens that can be obtained from a cancer patient, and their diagnostic use deserves scrutiny.
The history of organotypic cultures dates back to the early 19th century, when human intracranial tumors were hand-cut into tissue pieces and cultured using the so-called hanging drop method. Tissue fragments were attached to a coverslip and allowed to dip into heparinized human plasma, after which the coverslips were inverted, sealed and cultured for several weeks4. Manually cut tumor pieces have since been cultured using a variety of other methods, such as on plasma clots5, in liquid media6, or on 0.45 µm pore size filters6. The term "organotypic" was first used in 1954, in a study on retinal differentiation of the chick embryo eye7. This was followed by studies that used lung and heart tissue explants derived from chick embryos8, and brain explants from adult rats9.
Various slicing methods have been described, namely manual choppers10,11, the Krumdieck tissue slicer12,13, and vibratomes11,14,15,16. The Krumdieck tissue slicer generates cylindrical tissue cores, which are then sliced into circular tissue slices using a microtome. A vibratome, on the other hand, uses a vibrating blade microtome. In a study on liver slices, it was shown that the Leica vibratome generates more reproducible and consistent slices compared with the Krumdieck slicer15. Slice thicknesses ranging from 250 to 500 µm have been used, and studies report maintenance of viability and morphological features even until 16 days14,10,17,18. However, tumors have variable metabolic profiles that can affect nutrient requirements, and parameters such as tissue stiffness and matrix composition can influence on aeration and nutrient flow. It is therefore likely that each tissue type requires optimization of slicing and culture conditions.
Different cultivation methods have been used to support slice cultures: i) stationary interphase culture, also called stagnant support culture, in which slices are placed on top of a semi-porous membrane insert immersed in the culture medium. In this, the top of the slices is exposed to the air, while the bottom is supplemented with nutrients via the porous insert14,19; ii) the Trowel method, originally developed to culture whole organs or embryonic tissue slices. In this, slices are placed on top of a cotton sheet or filter supported by a metal grid, and the filter is soaked in the culture medium. To keep the tissues moist, a thin layer of medium is added on top of the slices20,21,22. These first two are so called air-liquid interphase cultures; iii) roller tube cultures, in which slices are placed inside the flat side of a plastic tube containing medium, and slow tube rotation ensures that the tissue is covered with medium during the first part of a cycle, or aerated during the second23; iv) rotating incubation units, in which slices are intermittently exposed to medium with nutrients and aeration. Different from roller tubes, in this method slices are placed on top of porous titanium grids placed in 6-well plates with culture medium24.
Tissue slices derived from resected solid tumors logically present an attractive ex vivo model in which to test the treatment response of anti-cancer agents, as they permit the evaluation of viability, targeted pathway activity, and molecular profiles of a specific tumor in the presence of its native tumor microenvironment. However, to evaluate whether the drug responses measured in tumor slices are predictive of in situ responses, it is important to assess to what extent tissue slices preserve tumor-specific biological functions such as cell proliferation, histopathology-specific cell differentiation or oncogenic signaling activities. The impact of mechanical stress elicited during slice preparation, slice handling, or culture-induced adaptions on both the quality and biological functions of tissue slices are fundamental questions, tightly linked to the ability to implement tumor-derived slices for functional diagnostics.
Our IMI-funded consortium project PREDECT (http://www.predect.eu) set out to systematically address these fundamental questions, by studying slice explants from a variety of sources. Using slices derived from breast, prostate and lung cancer models, this joint effort utilized qualitative read-outs as well as quantitative hematoxylin and eosin (H&E) - based readouts to demonstrate a requirement for atmospheric oxygen and stagnant filter supports to sustain the viability of cultured slices until 72 h. Furthermore, immunohistochemistry (IHC) analyses on cultured slices revealed intra-slice viability gradients, evidenced as necrosis gradients in slices derived from murine non-small cell lung cancer (NSCLC), estrogen receptor (ER), HIF1α and γH2AX gradients in breast cancer slices, or androgen receptor (AR) expression gradients in prostate cancer slices25. Interestingly, intra-slice viability gradients in 24 h cultures of murine NSCLC were rescued by cultivation in a rotating incubation unit, and our recent study showed that viability was extended to 72 h26. Particularly the top side remained most viable26, endorsing that drug response analyses on slices are best carried out on this side of the tissue.
Even though it remains a question as to how far tissue slices can recapitulate in situ tumor functions, they have been extensively used to test responses to anti-cancer agents, including targeted drugs, monoclonal antibodies and chemotherapy agents10,11,13,14,18,27. We recently showed that murine NSCLC slices show dynamic changes in proliferation and oncogenic signaling activities following cultivation when compared to freshly cut 0 h slices26. This indicates that it is important to investigate whether the targeted in situ biological functions are appropriately preserved during cultivation, prior to perturbation studies. Despite these findings, we showed that tumor slices can model spatial response to targeted therapies, midst drug treatments remain brief (24 h) and are initiated at the onset of culturing26. The following protocol describes important validation aspects relevant to the establishment and analysis of tumor slice cultures, prior to their application in pharmacological drug testing.
All mouse experiments described in this study were performed by following the guidelines from the Finnish National Board of Animal Experimentation, and were approved by the Experimental Animal Committee of the University of Helsinki and the State Provincial Office of Southern Finland (License number ESAVI/9752/04.10.07/2015).
1. Preparations Prior to Slicing
2. Collection of Tumor-bearing Lungs
3. Generation of Precision-cut Lung Tumor Slices
4. Treatment of Tumor Slices with Small Molecule Inhibitors
5. Fixation and Processing of the Tissue Slices
6. Processing and Analysis of Formalin-fixed and Paraffin-embedded (FFPE) Tissues
7. Analysis of Tissue Viability and Biomarker Expression
Figure 1 represents the workflow for the generation, cultivation and analysis of precision-cut tissue slices derived from murine NSCLC tumors. For this demonstration, we utilized tumors from a genetically engineered mouse model (GEMM) harboring conditional activation of KrasG12D together with the loss of Lkb1 (also known as Serine/Threonine Kinase 11), hereafter called KL. Mouse breeding and lung tumor initiation was performed as described in26,28. Figure 2A demonstrates the effect of tissue slice thickness on the viability of slices cultured for 24 h using a rotating incubation unit. The results show that 160 µm thin slices contain large necrotic areas across the slice. In addition, 250 µm thin slices show a necrosis gradient across the slice compared to 200 µm thin slices. It is likely that the poor overall viability of the thinnest 160 µm is caused by technical handling during positioning of the slices on top of the grids, as these are fragile and tend to curl. On the other hand, when slices are too thick, they can become prone to deficiencies in oxygen or nutrient diffusion across the slices, which in murine NSCLC explants is evidenced as necrotic death gradient25. However, it should be noted that slices with variable thicknesses can be generated from one tumor, despite use of identical vibratome settings. It is therefore recommended to analyze multiple replicates from different tumor samples. Importantly, each tissue type requires slice thickness optimization to achieve maximum viability, as the tissue texture and hardness can affect oxygenation and nutrient flow. Figure 2B-2C illustrates quantitative IHC analyses of NKX2-1 expression, a marker of well-differentiated lung adenocarcinoma (AC) in samples cultivated up to 72 h and matched 0 h slices. Results show that NKX2-1 expression is not significantly altered in cultured slices as compared to 0 h uncultured slices, suggesting that the process of cultivation does not overtly affect the differentiation status of AC tumor tissue. Figure 2D demonstrates the utility of tumor tissue slices for assessing the effectiveness of targeted drugs. We recently showed that Kras mutant murine ACs exhibit high expression of phosphorylated ERK1/2 (marking increased MAPK pathway activity) when compared to adenosquamous (ASC) tumors, while expression of phosphorylated 4EBP1 (marking mTOR activity) is similarly expressed in both AC and ASC tumors26. To test if these pathways can be effectively targeted on tissue slices, KL AC tissue slices were treated with DMSO or titrated amounts of compounds, namely 0.1 – 1 µM dactolisib to target the mTOR pathway or 0.05–0.5 µM selumetinib to target the MAPK pathway. Results show that 1 µM dactolisib or 0.5 µM of selumetinib are effective in inhibiting the phosphorylation of 4EBP1 or ERK1/2, respectively. Furthermore, dose-dependent inhibition of the targeted phosphoproteins indicates that tissue slices can also be utilized to validate phosphorylation-specific antibodies.
Figure 1: Schematic representation of the workflow for establishment and analysis of murine NSCLC tumor-derived slice explants. (A) Schematic describing the collection and preparation of tumor-bearing lungs for slicing. Lung lobes are harvested from a mouse and tumor tissue is dissected away from normal tissue. The black arrowhead and asterisk indicate approximately 4 mm and 1 mm tumors, respectively. The white arrowhead indicates lung tissue glued to the surface of the specimen holder. The red arrow points at an additional piece of normal lung support tissue to retain the tumor in an upright position. (B) Vibratome slicing and collection of tissue slices. White arrow indicates the slicing direction. Collection of sequential slices into a 24-well plate containing cold HBSS + P/S. The slices can either be cultured for different time points (here, 24–72 h) to assess tumor-specific marker expression during cultivation (top row), or can be used to perform drug treatments. C: vehicle control, T: drug treatment (bottom row). (C) Placing the tissue slice for cultivation using rotating incubation units. Tilt the 6-well plate so that some medium covers the top of the grid, place the tissue slice in the middle of the grid on top of the medium, and spread the slice using forceps. Ensure that the 6-well plates are weight balanced for a smooth rotation cycle. X: indicates incorrect, and : indicates correct positioning of the slice. (D) Photograph of the FFPE block of a tumor slice. Black arrow points at paraffin-embedded tissue slice stained with hematoxylin. (E) Schematics showing the sectioning order of the slices in FFPE blocks; these sections can be processed to assess tissue viability and tumor-specific biomarker expression. Please click here to view a larger version of this figure.
Figure 2: Assessment of viability and histotype-specific marker expression, and targeted drug treatment on NSCLC tissue slices. (A) Representative H&E images of AC NSCLC slices of the indicated thicknesses cultured for 24 h. 200 µm thin slices maintain better viability compared to 160 µm or 250 µm thin slices. Dark blue represents H&E stained viable tissue, and pink indicates pseudocolored necrotic regions. Light blue indicates regions excluded from the analysis, either due to poor tissue quality or presence of fibrous stroma. T1 and T2 represent biological replicates derived from two different tumors. Scale bar = 500 µm. (B) Representative IHC images of NKX2-1 expression in AC slices cultured for the indicated time points. Arrow indicates the area shown in higher magnification. Results show that NKX2-1 expression is not altered in the cultured slices compared to 0 h slices. Scale bar = 500 µm and 50 µm for low and high magnifications, respectively. (C) Quantification of the data shown in (B). (D) Representative IHC images of phosphorylated 4EBP1 or ERK1/2 expression in 0 h slices, or slices treated with DMSO or titrated amounts of dactolisib (dact, top row) or phosphorylated ERK1/2 expression in 0 h slice, or slices treated with DMSO or selumetinib (sel, bottom row). Black square boxes indicate areas shown in higher magnification. Scale bar = 1 mm or 50 µm for low or high magnification, respectively. Please click here to view a larger version of this figure.
Various complex in vitro tumor models, including 3D cultures and organoids, have been developed to recapitulate the architecture and oncogenic functions of in vivo tumor tissue29,30. However, the establishment of 3D cultures or organoids involves tissue dissociation and selective growth of a single cell type or co-culture of a select few cell types in an artificial environment. As a consequence, such models incompletely capture the intricacies of tumor heterogeneity and tumor-stroma interactions. Organotypic tumor slices, on the other hand, maintain the tissue architecture and biological complexities of the in situ tumor, without extensive manipulation. This ability of tissue slices to model tumor cells in their native microenvironment renders them particularly attractive for preclinical studies. We previously reported an optimized workflow for the establishment and analysis of precision-cut tumor slices, and showed that, compared to filter supports, a rotating incubation unit improve the viability of short-term murine NSCLC slice cultures25,31. However, cultivation on rotating units is technically challenging and requires constant monitoring. We here present a protocol for tumor tissue slice generation and practical use of a rotating incubation unit to culture them, as well as accompanying methods to monitor the ability of slices to capture in situ tumor biology, a prerequisite prior to drug response testing.
Several critical steps in the protocol ensure tissue integrity and viability of the tumor slices. If normal lung tissue surrounds the tumor, the vibratome can generate slices with inconsistent thickness or damage the slices, due to differences in texture and stiffness between normal and tumor tissues. It thus is important to remove the surrounding normal lung tissue prior to tumor slicing. Another critical step is the slicing thickness, which should be carefully optimized for each tissue type. Furthermore, once sliced, it is critical that the slice is placed approximately in the middle of the grid, so to ensure accurate intermittent dipping in culture medium and oxygen exposure. Finally, it is important to follow the position of a slice during its rotation period, as a slice can drop down in to the medium; if this happens, further actions can be taken as explained in step 3.6 of the protocol.
In addition to tissue handling to assure integrity, there are also critical steps in the IHC analysis to interpret how the slice resembles the native tissue. Our previous study showed that murine NSCLC tumors exhibit pronounced intra-tumor spatial heterogeneity in oncogenic signaling activities26. This means that the use of spatially distinct tissue slices for controls or drug-treated samples can affect reliable experimental data interpretation, and hence closely adjacent slices should be used as controls and test samples. We further showed that while proliferation or oncogenic phosphoprotein expression in freshly cut uncultured 0 h slices were similar to in situ tumors, cultured slices showed altered oncogenic phosphoprotein expression, specifically altered p4EBP1 and pSRC, as well as altered proliferation analyzed by Ki67 IHC. Altered p4EBP1 expression was similarly detected in 24 h human NSCLC and prostate cancer slice cultures (Narhi et al., Supplementary Figure S3B-S3C, S5 and S7)26. These findings endorse that comparison of cultured slices with their nearest 0 h uncultured slice is critical to assess the preservation of in situ tumor functions in cultured slices.
Despite improving the viability of organotypic slices25, there are limitations with the rotator system in terms of technicalities. Placing the tissue slices onto titanium grids is more challenging compared to filter inserts, and a rotating incubation unit may not be available. As an alternative, stagnant filter supports can be used, but in that case only the air-exposed side of the slices should be analyzed, as air-to-filter gradients in viability and hypoxia measured by HIF1α expression are rapidly formed in filter-supported slice cultures (Davies et al., Figure 5 and Figure 7A-7B)25. We have further shown that tumor slice cultures can exhibit altered proliferation and oncogenic signaling activities compared to their native tumors26, possibly because of wound-healing responses or metabolic adaptation of the slices to ex vivo culture32. Although gross morphological features of the murine NSCLC tumors were maintained during 72 h cultivation, culture-induced proliferative changes may affect accurate grading of the cultivated slices. Thus, tissue slices should only be utilized for short-term functional studies.
Use of a rotating incubation unit at least partially rescues intra-slice viability or biomarker expression gradients, particularly during the first 24 h of culture. Once validated for integrity and function, this provides tissue material for functional studies, such as drug treatment studies. In addition to drug response profiling, altered target expression following drug treatment can also benefit antibody validation. This is particularly relevant for the detection of murine epitopes with mouse monoclonal antibodies, as these tend to give high staining background. In addition, well-validated antibodies are required to achieve reliable and reproducible data in diagnostic and clinical settings. Thus, modulation of the abundance or phosphorylation of relevant epitopes following drug treatment in tissues slices provides a handy practical application in antibody validation. A major advantage of tumor slice cultures is the ability to model spatially-distributed functions, including oncogenic signaling activities or drug response in tumor or stromal cells, which makes them an attractive ex vivo model. However, slices rotate during cultivation, and the process of cultivation can further damage the tissue particularly at the edges. It is therefore challenging to precisely overlay the biomarker-stained IHC images of 0 h slices with the necrotic regions detected in cultured slices, which compromises the ability to precisely link spatial biomarker activities to drug response. In addition, tumor-intrinsic, culture-induced and drug-induced necrotic responses are indistinguishable at least in murine NSCLC tissue slices, compromising accurate quantitation of spatial drug responses. Finally, the use of tumor slice cultures permits a researcher to the test multiple compounds on the same tumor, without a need to treat animals, thus refining, reducing, and replacing experiments on laboratory animals.
As a future application, the described protocol can be adopted to clinical solid tumor samples. Further tissue type-dependent modifications or optimizations are likely required, starting with adjustments to the vibratome settings including slice thickness and vibration speed to optimize these for tumor texture or stiffness. In addition, nutrient and growth factor requirement may vary for different tumor tissues. As an example, breast cancer slices have been cultured with insulin supplemented in the medium13,18. Given that limited tissue material is obtained during surgery or biopsy, optimization of patient-derived tumor slice cultures can be challenging due to difficulties in obtaining sufficient numbers of replicate samples. Furthermore, data reproducibility is also challenged by pronounced patient-to-patient sample heterogeneity, particularly in the percentage of tumor cells versus fibrotic regions or stromal infiltrates, as well as necrotic tissue components. Finally, application of tumor slices in diagnostic settings would require investigation of the extent to which drug responses in slice explants of pre-treatment biopsies matches to post-treatment in vivo responses.
Authors declare no conflict of interest.
This research work received financial support from Innovative Medicines Initiative Joint Undertaking grant agreement no 115188, the University of Helsinki Doctoral Programme in Biomedicine scholarships (A.S.N.), and the Sigrid Juselius and Orion-Farmos Foundations (E.W.V.). We sincerely thank our PREDECT tissue slice platform consortium members (www.predect.eu), namely John Hickman, Heiko van der Kuip, Meng Dong, Emma Davies, Simon Barry, Wytske van Weerden and Hanneke van Zoggel, for collaborative development of the slice technology. We thank Taija af Hällström and Siv Knaappila for support in setting up the rotator system, and Riku Turkki for support with the MATLAB analysis. Jouko Siro is thanked for capturing the Figure 1 photographs. We thank the FIMM WebMicroscope team for scanning histological slides, and the Laboratory Animal Centre for husbandry support.
|Hank’s Balanced salt solution (HBSS)
|Penicillin Streptomycin solution
|Thermo Fischer Scientific
|Ham's F-12 medium
|Thermo Fischer Scientific
|Thermo Fischer Scientific
|Thermo Fischer Scientific
|Cyanoacrylate adhesive (GLUture)
|Leica VT1200 S vibrating blade microtome
|Albamma Research and Development
|Slice incubation unit
|Albamma Research and Development
|10 cm tissue culture plate
|Neolus Hypodermic Needles
|50 mL falcon tube
|Trifold histo cassette paper
|KOS The microwave multifunctional tissue processor
|Thermo Fischer Scientific
|Superfrost Ultra plus slides
|Thermo Fischer Scientific
|Thermo Fischer Scientific
|Hematoxylin for H&E staining
|Hematoxylin for counter staining
|Lot Number: GR98031-12
|pERK 1/2 antibody
|Cell Signaling Technologies
|Lot Number: 17
|Cell Signaling Technologies
|Lot Number: 20
|BrightVision poly-HRP Goat anti-rabbit secondary antibody
|Mounting medicum pertex
|Dactolisib (NVP BEZ-235)
|Pannoramic 250 slide scaner
|3DHISTECH PANNORAMIC VIEWER
|Adobe Photoshop CS6
|CellProfiler cell image analysis software
Copyright © 2024 MyJoVE Corporation. All rights reserved