Organotypic slice culture based on in ovo electroporation for chicken embryonic central nervous system

Abstract Organotypic slice culture is a living cell research technique which blends features of both in vivo and in vitro techniques. While organotypic brain slice culture techniques have been well established in rodents, there are few reports on the study of organotypic slice culture, especially of the central nervous system (CNS), in chicken embryos. We established a combined in ovo electroporation and organotypic slice culture method to study exogenous genes functions in the CNS during chicken embryo development. We performed in ovo electroporation in the spinal cord or optic tectum prior to slice culture. When embryonic development reached a specific stage, green fluorescent protein (GFP)‐positive embryos were selected and fluorescent expression sites were cut under stereo fluorescence microscopy. Selected tissues were embedded in 4% agar. Tissues were sectioned on a vibratory microtome and 300 μm thick sections were mounted on a membrane of millicell cell culture insert. The insert was placed in a 30‐mm culture dish and 1 ml of slice culture media was added. We show that during serum‐free medium culture, the slice loses its original structure and propensity to be strictly regulated, which are the characteristics of the CNS. However, after adding serum, the histological structure of cultured‐tissue slices was able to be well maintained and neuronal axons were significantly longer than that those of serum‐free medium cultured‐tissue slices. As the structure of a complete single neuron can be observed from a slice culture, this is a suitable way of studying single neuronal dynamics. As such, we present an effective method to study axon formation and migration of single neurons in vitro.


| INTRODUCTION
Organotypic slice culture is an effective and well-established method of maintaining tissue ex vivo. Tissues are able to maintain normal intercellular contact and differentiation patterns, as well as near-normal physiological and morphological characteristics, thereby avoiding or reducing deficits between standard cell culture and animal modelling techniques. Organotypic slice culture techniques provide a good method for studying central nervous system (CNS) development, but the real breakthrough in this technology came from the roller-tube technique pioneered by Gähwiler. [1][2][3] Using this technique, slices derived from various brain regions have been kept in culture for up to several weeks. In 1991, Stoppini et al used a sterile, transparent and porous membrane to culture the rat hippocampus, achieving good results, and thus establishing a simple method for organotypic culture of nervous tissue. 4 Brain slice culture can successfully simulate the process of neural development similar to that observed in vivo, and provide experimental evidence for organ-like brain slices, as an alternative to the in vivo brain. 5 This method has been effectively applied to various fields of experimental neuroscience. [6][7][8][9] Today, organotypic brain slice culture is mainly applied to mouse and rat tissue, as well as for disease modelling, such as with transgenic mice. [10][11][12][13][14][15][16][17] However, there are few reports related to the study of organotypic slice culture using chicken embryos. 18,19 The chicken embryo is a good animal model, used for both developmental biology and neurobiology studies. Particularly for the study of embryonic CNS development, the chicken embryo has the advantages of being abundant, easy to manipulate and easy to access to collect material from. With the emergence of in ovo electroporation technology using the chicken embryo, the ectopic expression of exogenous genes can easily be achieved in the CNS. We thus developed an organotypic slice culture combined with in ovo electroporation technology to study the dynamic changes of neurons under exogenous gene expression during chicken CNS development. This method will provide a new strategy for the functional study of exogenous genes in the development of the chicken CNS.

| Embryo preparation
Fertilized eggs were obtained from a local farm and incubated (HWS-150 Incubator, JingHong, China) at 37.8°C and 65% humidity. The Hamburger and Hamilton system was used to stage the embryos. 21 Embryos were studied from stages 17 (E2.5) to 38 (E12), with at least five embryos collected from each stage.

| Spinal cord in ovo electroporation and tissue section
We used a pCAGGS-green fluorescent protein (GFP) plasmid to drive GFP expression as a cellular marker. All plasmids were purified using a plasmid extraction kit (Cwbio, China), as per the manufacturer's instructions, and diluted in water. The in ovo electroporation protocol used was modified from a previously published study and a stereomicroscope was used to assist with all steps of electroporation. 22 Briefly, fertilized eggs were incubated until stage 17 (E2.5) ( Figure 1A,B), after which 3-4 mL of albumin was removed without disrupting the yolk ( Figure 1C). The shell was cut carefully with a pair of curved scissors to construct a 1-2 cm diameter window without touching the embryo. A mixture containing 0.25 µg/µL of pCAGGS-GFP plasmid and Fast Green dye (0.01%) was injected and loaded into the spinal cord with a mouth pipette until the dye filled the entire space ( Figure 1D). Electrodes were then immediately placed in parallel on either side of the embryonic neural tube (Figure 1E). A total of six 18-volt pulses, lasting 60 ms and separated by a 100 ms pause, were emitted. After electroporation (CUY-21 Electroporator, Nepa Gene, Japan), the electrodes were carefully removed and the egg was sealed with tape ( Figure 1F). The eggs were then placed back into the incubator until they reach the

| Optic tectum in ovo electroporation and tissue section
The in ovo electroporation protocol was modified from a previous study. 23 Briefly, fertilized eggs were incubated until stage 17 (E2.5), after which 6 mL of albumin was removed from the egg without disrupting the yolk. 23 One day later (E3.5), the shell was cut carefully with a pair of curved scissors to construct a 1-2 cm diameter window without touching the embryo. A mixture, containing 0.25 µg/µL of pCAGGS-GFP plasmid and Fast Green dye (0.01%), was injected and loaded into one side of the optic tectum using a mouth pipette until the dye filled the entire space ( Figure 2A). Electrodes were then immediately placed in an overlapping position on both sides of the embryonic optic tectum ( Figure 2B). A total of six, 15-volt pulses were administered, lasting 60 ms each and separated by a 100 ms pause ( Figure 2B). After electroporation (CUY-21 Electroporator, Nepa Gene), the electrodes were carefully removed and the egg was sealed with tape. The eggs were then placed back into the incubator until they reach the desired stage 38 (E12) for sample collection ( Figure 2C), sectioning and slice culture.
At stage 38 (E12), GFP-positive embryos were selected using a stereo fluorescence microscope ( Figure 2D-F). The optic tectum section was cut according to two different methods. Initially, the 4% pre-cooled agar is cut into a square according to the size of the tissue ( Figure 2G), and pasted to the tray of the vibratory microtome using a strong adhesive ( Figure 2G). A groove was next cut into one side ( Figure 2H-I), with the optic tectum of the chicken brain towards the groove, and the sides of the olfactory bulb affixed to the tray ( Figure 2J). Another method was to take selected tissues and embed the selected tissues in 4% agar that had been previously pre-cooled to 40°C ( Figure 2J). The tissues were coronally sectioned on a vibratory microtome (VT1200S, Leica) and 300 μm thick sections were mounted on a membrane of millicell cell culture insert (PICM03050, Millipore) ( Figure 2K, L).
In the process of tissue sectioning, aCSF was added to the buffer plate of the vibratory microtome in order to protect the cut tissue ( Figure 2K-L).

| SLICE CULTURE
The Millicell cell culture insert well was placed in a 30 mm culture dish containing 1 ml of slice culture media. The dishes were transferred to an incubator at 37°C and 5% CO 2 . The slices were cultured for 24 h before media replacement, and then the media replenished once in every two days. Tissues were maintained for 7 days. The morphological structure of the cultured slice was analysed after 48 h by confocal microscopy (Olympus ix81, Japan). For BrdU labelling,

| Microscopy
The whole embryo or brain was imaged under a stereo fluores-

| Time-lapse imaging
For time-lapse imaging, we set up a climate chamber on an inverted confocal microscope platform. The temperature of the climate chamber was 37°C and contained 5% of carbon dioxide. In order to clearly observe the morphology of neurons, the confocal 40× long focus lens was used. A culture slice with complete morphology and GFP expression was selected from a culture dish under an inverted fluorescence microscope. The dish containing the membrane insert with the slice and containing 1 mL of slice culture medium was placed into the climate chamber of the confocal microscope. The resolution was set to 800 by 1000 pixels and the time-lapse series was initiated at a rate of one z-stack every 10 minutes. Time-lapse series were analysed using Photoshop software (Adobe).

| Combination of in ovo electroporation and slice culture for the study of gene function
We with the medium and exposed to the air so as to prevent tissue necrosis ( Figure 3J).

| Serum-free medium cultured slice neurons lose strict regulation
In

| Comparison of tissue morphology and neuronal structure between serum-free medium and 25% horse serum medium-added cultured slice
Over the course of the experiment, the tissue slices cultured in serumfree medium lost their original morphology and structure, and patterns of neuronal migration also lost their strict regulation. Therefore, we added 25% horse serum in serum-free medium for comparative purposes. We found that in serum-free medium, the adhesive between the tissue slice and the insert culture dish membrane was very weak and would fall off easily on rinsing ( Figure 6A-C). In addition, the edge of the tissue was not smooth enough ( Figure 6A). However, neurons transfected with GFP plasmids were clearly visible and distributed  Figure 6D). However, typical neurons can be clearly seen, suggesting that brain slices remain active in serum-free medium, but do not guarantee good morphological structure ( Figure 6E-F). Tissue slices were similarly cultured in the medium containing 25% horse serum. It was found that the cultured tissue slices adhered strongly to the insert culture dish membrane were not desquamated by rinsing or even blowing slightly, and did maintain good morphology and structure ( Figure 6G-I). Greater magnifications revealed that the DAPIstained nuclei had a layered structure ( Figure 6J), but that the layered structure was still different from that of the same time layer in vivo ( Figure 5A). Compared with serum-free medium, the length of GFP-labelled axons in 25% horse serum culture were significantly longer than those in serum-free medium ( Figure 6K-L), especially in the enlarged structure of a single neuron ( Figure 6M,N). To shed light on this, we analysed the length of neuronal axons. Results showed that the length of axons in the medium containing 25% horse serum was 358.78 ± 70.48, which was significantly (P < 0.001) larger than that in the medium without serum, while the length of axons in the serum-free medium was 170.06 ± 33.60 ( Figure 6O).

| Slice culture is suitable for the study of a single neuron dynamics
The analysis of single neurons in cultured-tissue reveals an integral neuronal structure, including a very obvious dendrite ( Figures 4G-L and 5F). A complete single neuron can be observed by a scanning confocal microscope ( Figure 5G-N). Usually, as neurons have three-dimensional spatial structure, this complete structure is difficult to observe in tissue sections ( Figure 5C). Indeed, during the process of sectioning, parts of the neurites are usually lost ( Figure 5C). Thus, in the observation process, incomplete structures are often acquired ( Figure 5C).
During slice culture, neural precursor cells can form neurons, and the protruding axons and the structure of the dendrites can be completely retained ( Figure 5F). Despite the thickness of the cultured tissue slice, it is not possible to observe complete neuronal structures at low

| Effects of in ovo electroporation on cell apoptosis and proliferation in slice cultures
In order to shed light on whether this technique will interfere with neuronal survival and apoptosis, we stained for caspase 3   X arrow shows NF overlapped with GFP).

| Comparison of IBa1 and GFAP expression in vivo and in cultured slice of mouse cerebral cortex
In the process of tissue slice culture, we were able to fully characterize various states of neuronal growth through the detection of key Compared with the established methods, we have combined in ovo electroporation with tissue slice culture to study the function of exogenous genes in the CNS during chicken embryo development.
At the same time, we succeeded in performing in ovo electroporation in the spinal cord and the optic tectum. In ovo electroporation was used in the spinal cord or the optic tectum prior to slice culture.
The embryo was then allowed to develop to a specific stage, and tis- Differences between the spinal cord and optic tectum mean that the methods of in ovo electroporation need to also be different, the main differences being electrode and electro parameters. There are also some differences in later stages of slice tissue culture highlighted in the methods section.
We have found that the morphological structure of the CNS is compromised during the process of serum-free medium culturing.
We further found that the neurons in the spinal cord lose their regular migration patterns and that the direction of commissural axons is changed, in addition the optic tectum losing its distinctive layered structure. This result was as predicted since, considering that cells are living and growing in culture medium, this process causes the organically generated regulatory function of various factors in the organism to be lost. In this experiment, we used serum-free medium.
Despite being in tissue slices, neurons that survive in the medium differ from those of the body as many functions are not completely reflective of the body's natural conditions. This is also a problem that needs to be taken into consideration in slice culture experiments.
There are differences between slice cultures in vitro and sections in vivo are not only in terms of gross morphology, but also in terms horse are significantly longer than those in serum-free medium.
The results of GFAP and Iba1 showed a significant difference between in vivo and in vitro-cultured brain slices. The number of GFAP-and Iba1-positive cells increased significantly in cultured-brain slices. GFAP is considered a marker for astrocytes and Iba1 a marker for microglia. Increased GFAP gene expression is a common feature of CNS injury and is usually used as a marker of nerve injury. 29 Microglia, the innate immune cells of the CNS, constantly survey CNS parenchyma for pathogens and cellular stress signals. 30 Many studies have shown that the expression of GFAP will change over the course of organotypic slice culture. 31,32 The changes in GFAP and Iba1 expression levels reflect differences between the brain slice culture process and in vivo conditions. However, they are also due to damage induced in the process of tissue sectioning. More research is needed to elucidate the differences between organotypic slice culture and in vivo conditions.

| CONCLUSION
We provide a method that combines in ovo electroporation and slice culture to study gene function in chicken CNS during embryonic development. The comparison of serum-free medium with in vivo and 25% horse serum medium shows that tissue culture and in vivo conditions differ. Adding horse serum is more conducive to the maintenance of tissue structure, and this method is suitable for the study of single neuronal dynamics.

CONFLI CT OF INTEREST
The authors declare that they have no known conflicts of interest associated with this publication.