In Vitro Model of Neurotrauma Using the Chick Embryo to Test Regenerative Bioimplantation

Effective repair of spinal cord injury sites remains a major clinical challenge. One promising strategy is the implantation of multifunctional bioscaffolds to enhance nerve fibre growth, guide regenerating tissue and modulate scarring/inflammation processes. Given their multifunctional nature, such implants require testing in models which replicate the complex neuropathological responses of spinal injury sites. This is often achieved using live, adult animal models of spinal injury. However, these have substantial drawbacks for developmental testing, including the requirement for large numbers of animals, costly infrastructure, high levels of expertise and complex ethical processes. As an alternative, we show that organotypic spinal cord slices can be derived from the E14 chick embryo and cultured with high viability for at least 24 days, with major neural cell types detected. A transecting injury could be reproducibly introduced into the slices and characteristic neuropathological responses similar to those in adult spinal cord injury observed at the lesion margin. This included aligned astrocyte morphologies and upregulation of glial fibrillary acidic protein in astrocytes, microglial infiltration into the injury cavity and limited nerve fibre outgrowth. Bioimplantation of a clinical grade scaffold biomaterial was able to modulate these responses, disrupting the astrocyte barrier, enhancing nerve fibre growth and supporting immune cell invasion. Chick embryos are inexpensive and simple, requiring facile methods to generate the neurotrauma model. Our data show the chick embryo spinal cord slice system could be a replacement spinal injury model for laboratories developing new tissue engineering solutions.


Introduction
There is a global drive to develop therapies which can safely and effectively restore function of the injured spinal cord.Given the complexity of spinal cord injury (SCI), the major challenge is to develop combinatorial therapies which can achieve multiple clinical goals (e.g., promoting and directing regeneration of nerve fibers, modulation of immune responses, degradation of inhibitory scar tissue).Bioscaffolds offer a chemical and physical environment to support neural cell growth.They can be made from natural polymers to mimic native tissue extracellular matrix or synthetic polymers with linked agonists to improve neural tissue growth.In addition, modified bioscaffolds can secrete molecules

Abstract
Effective repair of spinal cord injury sites remains a major clinical challenge.One promising strategy is the implantation of multifunctional bioscaffolds to enhance nerve fiber growth, guide regenerating tissue, and modulate scarring/inflammation processes.Given their multifunctional nature, such implants require testing in models which replicate the complex neuropathological responses of spinal injury sites.This is often achieved using live, adult animal models of spinal injury.However, these have substantial drawbacks for developmental testing, including the requirement for large numbers of animals, costly infrastructure, high levels of expertise, and complex ethical processes.As an alternative, we show that organotypic spinal cord slices can be derived from the E14 chick embryo and cultured with high viability for at least 24 days, with major neural cell types detected.A transecting injury could be reproducibly introduced into the slices and characteristic neuropathological responses similar to those in adult spinal cord injury observed at the lesion margin.This included aligned astrocyte morphologies and upregulation of glial fibrillary acidic protein in astrocytes, microglial infiltration into the injury cavity, and limited nerve fiber outgrowth.Bioimplantation of a clinical grade scaffold biomaterial was able to modulate these responses, disrupting the astrocyte barrier, enhancing nerve fiber growth, and supporting immune cell invasion.Chick embryos are inexpensive and simple, requiring facile methods to generate the neurotrauma model.Our data show the chick embryo spinal cord slice system could be a replacement spinal injury model for laboratories developing new tissue engineering solutions.

Plain language summary
Spinal cord injury can be highly debilitating for patients and carers.Repair of the spinal cord is therefore a major clinical goal but is challenging owing to the complex pathology of the injury site.Researchers need to mimic this complexity in the laboratory to test new therapies in realistic environments.Traditionally, researchers use live, adult animal models to achieve this, which are highly traumatic, variable, and costly.Here, we show an alternative system based on slices of spinal cord tissue from chick embryos that could partially replace live adult animal testing.We show slices can be grown in a dish and reproducibly injured, with complex pathology replicated.Further, we modified cell responses in the injury through interface with a novel implant, potentially improving repair.The model is cost-effective, simple, and associated with less animal suffering than live adult animal experiments.
Before E15, the chick embryo spinal cord is considered "regeneration permissive" and displays nerve outgrowth and neural progenitor proliferation in response to transection injury.Further, chicks with spinal cords transected before E15 can operate their lower limbs at hatching.Post E15, transections to the chick spinal cord have impaired regeneration, and the chicks show a lack of function of the lower limbs at hatching (Hasan et al., 1993;Shimizu et al., 1990;Whalley et al., 2006).In addition, microglia are present in the developing cord from E3 (Calderó et al., 2009).Microglia are nervous system immune cells and key mediators of inflammatory tissue responses to biomaterials.Use of the chick embryo may also contribute to the drive to reduce, refine, and replace animal use (the 3Rs) within the neural tissue engineering field.Prior to E15, the chick embryo is not protected under Directive EU/2010/63 and is not considered a protected species according to the Animals (Scientific procedures) Act of 1986 (UK), so could be used as a replacement strategy for animal experiments.Moreover, due to its development in ovo, the embryo can be accessed without harming the mother, and a batch of eggs usually comes from more than one hen, improving genetic diversity in comparison to use of rodents.Further, chick embryo culture is cost-effective as eggs are also significantly cheaper than rodents (ca.£0.80 versus £8 each) and require limited infrastructure and expertise in handling, so protocols could be adopted by regenerative neurology laboratories worldwide (Fig. 1).
Elegant studies have previously demonstrated successful organotypic culture of chick embryo spinal cord slices (Tubby et al., 2013;Yang et al., 2019).Whilst important, these were focused on aspects of spinal cord development and were not investigating responses to injury.Further, the models employed transverse slices not able to recapitulate nerve fiber tract regeneration, were cultured for short periods of time (up to 7 days), and did not inves-mal numbers required to generate statistically significant data, and are associated with substantial ethical implications in introducing long-term spinal injury.
Organotypic slices of rodent spinal cord tissue have been proposed as an "intermediate stage" between cell culture and adult animal experimentation to screen promising materials in neurorepresentative environments.In culture, these slices maintain in vivo cell types and cytoarchitecture.Further, a transecting injury can be introduced, and the tissue displays typical pathological responses to the injury such as glial scarring, immune cell infiltration, and limited nerve fiber outgrowth.Using this system, Weightman et al. (2014) demonstrated that aligned electrospun scaffolds implanted into injured rodent spinal cord slices were able to direct regeneration, highlighting the model's utility.More recently, Walsh et al. (2023) used transverse mouse spinal cord slices cultured on hydrogels to demonstrate a translationally relevant, 3D printed GelMA hydrogel did not cause adverse astrocyte and microglia reaction when interfaced with nervous tissue.However, overwhelmingly tissue is currently derived from rodents, which have high maintenance costs, requiring housing facilities, dedicated care personnel, and food.In addition, slices are generally derived from one litter, limiting genetic variability within biological repeats.Further, if embryonic rodent tissue is required, the mother also has to be sacrificed to access the tissue, adding ethical considerations.
As an alternative to rodent-derived tissue, the chick embryo is a well-defined model of the mammalian nervous system extensively employed in neurodevelopmental research (Himmels et al., 2017).In terms of the spinal cord, all major central nervous system cell types are established by embryonic days 9-10 (E9-10).Further, vascularization and neuroarchitecture, such as cortico-spinal tracts, are also observed at this time (Ferretti and Whalley, 2008).so we chose the latest time point the embryo is considered to be non-protected (E14).We subsequently received approval to sacrifice the embryos under Schedule 1 procedures from the local Animal Welfare and Ethical Review Body.Sacrifice was by decapitation, in accordance with approved methods of sacrifice for vertebrate embryos in the Animals (Scientific Procedures) Act 1986.

Experiment overview and time points
Spinal cords were dissected from E14 chick embryos, sliced longitudinally, and cultured.These slices were injured at two different time points: 4 days post dissection (I4) and 12 days post dissection (I12).Each of these groups were split and fixed at a further two different time points: 4 days post injury (denoted I4F4 and I12F4) or 12 days post injury (denoted I4F12 and I12F12).The viability of each of these time points was controlled with an uninjured spinal cord slice that was fixed at the same time.These time points were required to investigate the effects of an early versus late injury and whether the response to injury changed over time.For bioimplantation experiments, slices were injured at 4 days and tigate glial cell roles, all of which are important when considering modelling traumatic injury.Despite its potential, the utility of the chick embryo has never been assessed for generating slice models of SCI within which new tissue engineering strategies can be tested.
In this study we aimed to assess (i) the feasibility of establishing organotypic spinal cord slices from chick embryo tissue, (ii) the neuropathological responses to a transecting injury in organotypic spinal cord slices from chick embryo tissue, and (iii) the feasibility of implanting a bioscaffold and the subsequent regenerative response in the injury.

Ethics
Given the aim of the experiment is to test biomaterial modulation of neuropathological responses, we wanted to ensure all relevant cell types were present with established cytoarchitecture in place, the spinal cords were completely sectioned with a standard scalpel aided with a needle following the marks of the double-bladed scalpel.The slice debris between the two injury margins was removed with tweezers.For the implantation experiments, a 250 μm thick slice of uncoated and acellular Hemopatch™ (Hemo-patch™ was kindly donated by Baxter Healthcare Technologies) was implanted into the lesion.Hemopatch™ is a clinical grade bovine collagen scaffold used as a sealant in neurosurgical procedures.We have previously shown its utility for supporting neural transplant populations (Mogas Barcons et al., 2021), highlighting potential for repair mediating implantation in sites of neurological injury.To achieve implantation, under sterile conditions, a 0.5 cm 2 piece of Hemopatch™ was cut with a scalpel and then chopped with a McIlwain tissue chopper set to 250 μm.The piece was then placed adjacent to the slice and further cut to the size of the lesion with a scalpel.Then, it was implanted to bridge the gap between the two sides of the lesion (Fig. 2I-L).Finally, the slices were kept in the incubator and fed with culturing medium (50% change) every 2-3 days for 20 days until they were transferred into 24-well plates for fixation.

Imaging
Images of spinal cord organotypic slices were taken under a fluorescence microscope (Leica DMC 2500 LED) equipped with a CCD camera (DFC350 FX) unless stated otherwise.The software used for imaging was Leica Application Suite X v.1 (2017).All spinal cord slice images were taken at consistent light exposure parameters throughout the experiments.
immediately implanted with the scaffold.These slices were then cultured to the maximum time point from previous experiments, 20 days post-injury/implantation (24 days total culture).The overall experimental scheme is shown in Figure 2, which shows spinal cord extraction, spinal cord slicing, spinal cord culture and injury, and finally, biomaterial implantation into the injury site.Biomaterial implantation showed good integration with injured tissue, determined as cellular infiltration into the bioimplant at the lesion margins, as shown in Figure 2I-L.These steps are described in more detail in the rest of the methods.

Establishing organotypic slices from chick embryo spinal cords
Fertilized Bovans Brown eggs (Henry Stewart & Co.) were incubated at 37°C and 47% humidity in a specialized egg incubator (Brinsea) (Incubation parameters were chosen as indicated by Brinsea).After 14 days, the embryos were extracted, sacrificed, and a 1.5-2 cm section of the spinal column was dissected out.The spinal cord was ejected by a stream of slicing medium (HEPES (2.5%; BioSera L0180-100), EBSS (97.5%;Invitrogen 24010-043)) injected through the spinal canal.All spinal cords were kept in slicing medium on ice until use (up to a maximum of 1 h).
The spinal cord slice organotypic culture and lesioning protocols were adapted from Weightman et al. (2014).Here, each spinal cord was sliced longitudinally at 350 µm thickness with a McIlwain tissue chopper, collected in a small Petri dish containing ice-cold slicing medium, and kept on ice for 30 min.Around 3 intact slices could be generated from each spinal cord.Meanwhile, the culture chambers were prepared.A permeable insert (Millicell, PICM0RG50) was placed in a small Petri dish, and three 0.45 µm pore, 0.5 cm 2 size membranes (Omnipore JHWP04700) were placed on top of the insert.The insert provides a raised platform so slices can be cultured at an air:medium interface.The space between the dish and the insert was filled with 1.1 mL culturing medium (MEM (Invitrogen 42360-024), horse serum (25%; BioSera S0910-500), EBSS (19%), glucose (36 mM), pen-strep (1%; Fisher 11528876).After 30 min, the spinal cords were collected with a wide-bore transfer pipette and placed on the membranes (one spinal cord slice/membrane).Such membranes have previously been used for successful rodent spinal cord slice culture (Weightman et al., 2014) and facilitate manipulation of one spinal cord at a time, e.g., for transfer to a well for individual staining.Finally, the small Petri dishes were placed on a large square Petri dish for easier transportation and for better sterile protection and kept inside an incubator at 5% CO 2 and 37°C.The slices were fed every other day by 50% medium change until fixation at the required time point (all steps shown in Fig. 2A-H).For fixing, slices were washed in PBS once, incubated in 4% PFA for 1 h at room temperature, washed three times with PBS, and then kept at 4°C for further staining.

Introduction of a transecting lesion in the organotypic spinal cord slices and implantation of Hemopatch™
Transecting lesions were induced in sterile conditions under a dissection microscope.A double-blade scalpel was used to mark the lesion edges to ensure equal injury size in all slices.Then, overlaid onto each image.Three images were obtained for each slice, and the total number of lectin-positive cells per area was averaged.The results were statistically analyzed with a Kruskal-Wallis test with Dunn's correction.

Statistical analysis
All statistical analysis was performed and graphs were plotted using GraphPad Prism 9 (version 9.5.1).Data are presented as mean ± standard error of the mean (SEM) unless otherwise stated.Experimental "n" number is reported where each "n" is one separate slice.

Organotypic spinal cords could be derived from the chick with lesioning producing characteristic neuropathological responses
Initial experiments showed all major neural cell types, astrocytes, neurons, oligodendrocytes, and microglia, could be detected within the cultured chick embryo spinal cords (Fig. 3).Nerve fiber networks staining positive for Tuj-1 were observed in all slices.These fibers displayed both random orientation and occasional stereotypical aligned orientation (Fig. 3A).MBP staining was observed in all slices and in general appeared fibrous in nature (Fig. 3B).MBP staining was often coincident with nerve fiber staining, following the nerve fiber patterns (Fig. 3C).In around 20% of slices, tracts of myelin staining were observed running longitudinally along the spinal cord slice (Fig. 3D).A dense network of GFAP-positive astrocytes was detected in all slices (Fig. 3E).Finally, lectin staining was detected (Fig. 3F).However, in our hands, lectin staining was capricious and required multiple attempts at staining to achieve sufficient data for quantification purposes.When successful, lectin-positive microglia could be observed at regular intervals throughout the spinal cord tissue and displayed a ramified morphology, associated with an exploratory (non-inflammatory) phenotype (Fig. 3F).
Within lesioning experiments, all spinal cords showed calcein-AM staining, an indicator of cellular health, throughout the tissue, regardless of whether they had been lesioned or not (Fig. 4A,B).At high magnification it was possible to see single dead cells spread across the spinal cord (Fig. 4C), which were mostly accumulated outside the edges of the slice or at injury margins.Approximately between 60 and 80% of the spinal cord was stained with calcein-AM.There were no statistically significant differences in overall slice viability between injured and uninjured slices or between time points (Fig. 4D).At each time point, lesion widths appeared to be consistent in size, and after quantification there were no statistical differences in lesion width between groups, with an average width ranging from ca. 400 to 500 µm (Fig. 4E).
During microscopic observation it was noted that the lesion margins of spinal cords were stained more intensely with anti-GFAP than the center of the spinal cord (Fig. 5A-D).Further, at the lesion margin, GFAP-positive cell morphologies were characterized by elongated somas and processes, distributed in parallel to each other and perpendicular to the lesion (Fig. 5E), in contrast 2.8 Quantification and statistical analysis of organotypic spinal cord slices Slice health Images of stained spinal cords taken at 5x magnification were used for this analysis.First, the images for each channel (red and green for ethidium homodimer staining and calcein-AM staining respectively) were converted to 8-bit.The optical density (OD) background value was obtained for each channel as the mean OD of 5 different areas of the image outside the spinal cord.Then, the mean OD value of the full image for each channel was obtained and corrected by subtracting the background.The corrected values were added to obtain a total fluorescence value and, finally, a viability index was obtained as the percentage of corrected viable OD relative to the total OD value.The survival data for transection experiments was analyzed with a two-way ANOVA and Tukey correction.The survival data for the implantation experiments was analyzed with a Kruskal-Wallis test with Dunn's correction.

GFAP expression
5x images of GFAP-stained spinal cords were converted to 8-bit.Using ImageJ, a vertical line, perpendicular to the injury margin, was drawn from the lesion margin towards the center of the slice, and the OD profile of the line was generated.The average OD was then calculated at three regions: 0-10 µm, 50-60 µm, and 100-110 µm.The process was repeated three times for each slice to obtain an overall slice average at each region.Then, the background was calculated by averaging the mean OD value of three different regions outside the slice in each image, which was subtracted from the previously averaged slice OD values to obtain a corrected OD value for each region of each spinal cord.A repeated measures two-way ANOVA with Sidak's multiple comparison test was used to statistically analyze the data.

Axonal infiltration into the lesion
For injury alone experiments, the total number of axons infiltrating the lesion was analyzed.20x magnification consecutive images of the whole width of the lesion were fused together in a panoramic image using Microsoft PowerPoint 2016 and opened in Im-ageJ.All axons that were sprouting from the lesion margin were counted and divided by the full width of the slice (mm).However, this protocol was not usable after biomaterial implantation due to the intricate web of axonal infiltration within the material, meaning it was difficult to accurately count single axons.Therefore, the total length of all infiltrated axons in the lesion was calculated and compared between unimplanted and implanted slices.Here, three 40x images were taken at random areas within the lesion and uploaded to the NeuronJ plugin in ImageJ.The axons were traced, and the total length of traced axons in µm was obtained.The results from the three fields were averaged to generate an axonal length measurement for each slice.A Kruskal-Wallis test with Dunn's correction was performed for statistical analysis.

Lectin expression
Lectin quantification was achieved by counting the number of lectin-positive cells that were detached from the spinal cord slice per unit area (mm 2 ) at the lesion site using a standard size grid similar at all time points (Fig. 5I).Approximately 20 axonal fibers were counted per mm from one side of the lesion with a relatively high variation amongst slices (between 20 and 60 fibers/mm) within the same time point (Fig. 5J).The total length of sprouting axons per field was also highly variable (from 0 to 6000 µm), with no statistical differences between time points (Fig. 5K).

Bioimplantation modulates neuropathological responses in injured chick embryo spinal cord slices
In order to examine whether biomaterials could be implanted into the injury site and potentially modulate regenerative responses, slices were injured as above, and half were implanted with Hemo-patch™ with half remaining as controls (non-implanted).Similarly to the previous experiments, the spinal cords in the control with the random network distal from the lesion (Fig. 5E inset).Microglia, detected as lectin-positive cells, were seen infiltrating the lesion gap.The cells infiltrating the lesion presented an amoeboid morphology while cells visualized within the center of the slice appeared more ramified (Fig. 5F).Tuj-1 staining showed limited axonal sprouting into the lesion gap.The nerve fibers extended in a haphazard fashion with no particular direction favored and were never observed to reconnect both sides of the lesioned spinal cord (Fig. 5G).OD analyses revealed a significantly higher GFAP expression between 0-10 µm from the lesion margin than at 50-60 µm and 100-110 µm from the lesion margin for the spinal cords that were fixed 12 days post injury.This higher GFAP expression was not seen for those that were fixed at 4 days post injury (Fig. 5H).The number of infiltrating microglia cells per unit area was positive cells expressed more GFAP than those within the tissue.However, their morphology and cytoarchitecture differ from untreated slices as seen by the loss of polarity towards the lesion and the lack of a clear band of orientated astrocytes expressing high levels of GFAP (Fig. 6A,B).Biomaterial implantation induced high microglial infiltration into the implanted biomaterial compared to infiltration within the lesioned area alone.Visual analysis group had a higher expression of GFAP at the edge of the lesion, with the cells apparently aligned at 90° to the lesion (Fig. 6A,C).However, when Hemopatch™ was implanted, GFAP expression was not higher at the lesion edge than at the center of the spinal cord, and the orientation of GFAP-positive cells seemed random at all distances from the lesion (Fig. 6B,C).A visual assessment of the treated spinal cord slices indicates that infiltrating GFAP- We also report high axonal infiltration within the implanted biomaterial compared to the untreated lesion, similar to previously published in vivo (Galli et al., 2018;Wu et al., 2018;Yang et al., 2015) and in vitro research (Guijarro-Belmar et al., 2019;Weightman et al., 2014).We were unable to distinguish any possible physical reconnection between axons from each end of the lesion.Nevertheless, the high number of infiltrating axons is suggestive of a potential pro-regenerative role for axonal growth (or regeneration permissive role) of Hemopatch™ implantation.An exciting further development could be to fuse the model with specialist multielectrode array systems capable of recording electrophysiological activity at multiple points across the whole spinal cord slice.These could measure recovery of spinal cord network activity across the lesion after injury and any therapeutic manipulation.
As far as we are aware, this is the first report of immune cell infiltration in an implant in the chick.Our data indicate that microglial infiltration in the lesion area is higher upon implantation.Moreover, morphological examination seemed to indicate that microglia display a reactive phenotype within the biomaterial.Our results are in consonance with previously published research with mouse organotypic models of transecting SCI (Weightman et al., 2014) and in vivo models of transecting SCI in rats and mice respectively (Domínguez-Bajo et al., 2020;Kourgiantaki et al., 2020).A reduction in microglial infiltration has only been reported in one study (Guijarro-Belmar et al., 2019), due to the functionalization of the implanted biomaterial with an Epac2 agonist.However, there were some technical issues with lectin staining, including high background and inconsistent staining success.Lectins can be less specific than antibodies and may highlight different features from microglia such as blood vessels (Sharma et al., 2022).Future work should aim to employ chick-specific microglial antibodies such as anti-chick CD45 (Calderó et al., 2009), which we could not source at the time of this study.
Establishing the slice model still requires sacrifice of the embryo.However, this can be performed at a stage where the animal is not considered protected (before third trimester) and before the embryo would be able to survive outside of the egg on its own (before E16).Further, as opposed to embryonic rodent models, there is no requirement to sacrifice the mother.The use of embryonic tissue does mean direct inference to adult nervous system injury cannot be made.Further, vascularization and the blood-brain barrier (BBB; important components to injury and regeneration pathology) are absent in the chick embryo slice model.To address these issues, ageing the slices through extended culture may allow for "older" tissue neuropathological responses to be assessed.Organotypic slices of neural tissue have been cultured for several weeks, indicating this is a possible strategy.Introduction of a BBB would be challenging.However, in terms of tissue engineering, implants are designed to come in direct contact with nervous tissue to elicit the mechanism of action.We have shown that nervous tissue responses can be observed using the chick embryo slice model, indicating the model has utility in investigating these specific mechanisms of action.Whilst the model will not be able to completely replace live adult animal models of SCI, it may offer a replacement with reduced ethical implications for research at early stages of the lectin-positive cells' morphology suggested most of the infiltrated cells to present an amoeboid or reactive phenotype (Fig. 6D-F).In terms of nerve fibers, implantation of Hemopatch™ resulted in robust axonal infiltration within the material with high numbers of nerve fibers detected in the lesion area, which were absent in controls (Fig. 6G,H).The total axonal length in each microscopic field was 10,000-fold higher in implanted versus control injury sites (Fig. 6I).

Discussion
Here we show for the first time that longitudinal 3D slices of chick embryo spinal cord tissue can be derived and cultured for extended periods of time (at least 24 days) in the lab.Major neural cell types are present with classical cytoarchitecture observed (nerve fiber tracts, myelin tracts, and interwoven astrocytes).Further, we show that neuropathological responses to transecting injury commonly associated with adult injury can be observed that are modified with bioimplantation of a clinical grade scaffold material.Chick embryos are simple to maintain and cost-effective.Therefore, the data show this model could have substantial impact in enhancing accessibility to high-validity neurological injury models for a range of laboratories.This includes neuroscience laboratories with little or no infrastructure for animal experimentation or to facilitate cross-disciplinary research with material engineering laboratories.
When considering neuropathological response, our data on GFAP expression post-injury are in line with previous rodent studies.For example, we only saw an increase in GFAP expression at the lesion margin compared to the slice body at 12 but not at 4 days post injury.Although highly variable in the literature, visible changes in GFAP expression are generally reported to occur several days after the injury (Bradbury and Burnside, 2019).In comparable experiments, Weightman et al. (2014) report GFAP upregulation at the injury margin at seven days post injury in rodent organotypic slice models of spinal cord injury.We also report that GFAP OD at the edge of the lesion was lower after biomaterial implantation than in injured, non-implanted slices.Moreover, astrocytes, which were polarized in a 90° angle in respect to the untreated lesion, appeared more disperse and less elongated on the implanted slices, suggesting a possible disruption of glial reactivity.These results are supported by studies which report a disruption of the glial scar either through an architectural change in vivo (Hurtado et al., 2011) or a decrease in GFAP expression in vitro (Guijarro-Belmar et al., 2019;Weightman et al., 2014).Therefore, even though there is a slower change in astrocyte reaction to the injury, the effects of bioimplantation on astrocyte regenerative responses can be assessed within the chick embryo model.Some studies demonstrate no difference in glial scar formation after biomaterial implantation (Domínguez-Bajo et al., 2020;Galli et al., 2018).Varied data across the literature could be due to the use of different types of biomaterials (i.e., aligned fibers, injectable hydrogels, implantable gels), stressing the importance of biomaterial screening tests before committing to higher complexity, live animal models. of therapy development.It must be noted though that, as scientific opinion on embryo sentience is constantly changing, ongoing review is required with relation to the method of embryo sacrifice.
There may be species-specific differences in neuropathological responses to injury/bioimplantation between chicks and humans and chicks and rodents.Therefore, some caution is needed when interpreting the data.Other approaches offer a means to address some of these issues.Researchers are developing organoid models of the spinal cord using human induced pluripotent stem cells (hiPSCs; Lee et al., 2022;Xue et al., 2022), which overcome species-specific differences.However, it is still difficult to mimic the cytoarchitecture of the spinal cord using organoids, especially longitudinal tracts.Further, common protocols for organoid development are extended (e.g., at least 6 weeks), and longer-term culture poses issues of necrosis in the hypoxic organoid center.In terms of bioimplantation, organoids may also pose a technical challenge in biomaterial interfacing given their small size (ca.1-3 mm in diameter).In general, hiPSC-derived cultures also lack the neural immune cells, a major component of the tissue reaction to biomaterials.Alternatively, there is also a possibility of introducing human cells into tissue slices through xenografting.A recent study by Ogaki et al. (2022) demonstrated the replacement of microglia in rodent organotypic slices with human hiPSC-derived microglia, which could be tested within the chick embryo slice model.Some materials used in the culture and analysis of the chick embryo spinal cords are derived from animals (e.g., horse serum, antibodies).Establishing the model using non-animal derived materials is an important next step.Serum has been shown to be crucial for maintaining CNS architecture in tissue slices (Yang et al., 2019), so it may be challenging to replace.However, there are several commercial serum alternatives that could be tested.There is also a drive towards using chemically defined medium for in vitro cell and tissue culture, so a chemically defined medium for chick embryo slice culture could be attempted.In terms of analytical procedures, more antibodies are now being synthesized using recombinant cell lines, so should be examined as replacements for any animal-derived antibodies.Further, classical histological dyes could be used to detect nerve fibers (e.g., silver staining), myelin (Luxol fast blue), and overall tissue changes (H&E staining).These may offer an alternative, animal-free route to analyze tissue changes after injury and biomaterial implantation.
In conclusion, we have been able to show that using a simple, cost-effective tissue model based on the chick embryo, researchers can potentially establish a model of high utility for testing new regenerative therapies for spinal cord injury.

Fig
Fig. 1: Schematic summarizing advantages of using the chick embryo to derive organotypic slice models of spinal cord injury versus rodents

Fig. 2 :
Fig. 2: Images depicting experimental overviewThe spinal column was dissected (A) and the spinal cord ejected using a stream of slicing medium (B).Multiple cords were collected in one Petri dish (C) before slicing using the McIlwain tissue chopper (D).Slices were then added to Petri dishes (E) containing pre-cut membranes (F).A transecting injury was introduced into the slice (G), and in some slices biomaterial implantation was performed (H).(I-L) Representative fluorescent and phase images showing bioimplant integration with the injured spinal cord slice tissue.Dashed line indicates outline of implant.Cellular infiltration is noted within the implant (I-K), the fibers of which can be seen under phase microscopy (L).

Fig. 4 :
Fig. 4: Slices retain high viability, and the lesion width is consistent over time (A-C) Representative double-merged fluorescent images showing live (green), dead (red), and DAPI (inset) staining of control (A) and injured (B) organotypic slices.Arrowheads indicate dead cells.The injury in (B) is in the bottom left of the image.(C) shows a high-magnification field of an injured slice, where a limited number of dead cells can be observed (arrowheads).(D) Bar chart showing quantification of live staining compared to dead staining, represented as a viability index.No statistical differences were found across all experimental conditions (Twoway ANOVA, p > 0.05, n = 3).(E) Bar chart showing quantification of lesion width at each experimental time point.No statistical differences in lesion width were detected across conditions (one-way ANOVA, p > 0.05, n = 10)

Fig
Fig. 5: Classical neuropathological responses are detected after traumatic injury in chick embryo spinal cord slices (A-D) Representative fluorescent images showing GFAP staining of spinal cord slices when injured at four days in vitro and cultured for 4 (A) or 12 (B) days more, or when injured after 12 days in vitro and cultured for 4 days (C) or 12 (D) days more.(E) Representative high-magnification fluorescent image depicting GFAP staining at the injury margin and in the slice body (inset).Note the aligned morphology of the astrocytes at the lesion margin, perpendicular to the injury.(F) Representative fluorescent image showing microglia infiltration into the injury site.(G) Representative image depicting Tuj-1-stained nerve fibers projecting into the injury.(H-K) Bar charts showing quantification of (H) GFAP intensity at different distances (0-10, 50-60 and 100-110 µm) away from the injury margin, (I) numbers of microglia infiltrating the injury site, (J) numbers of nerve fibers per mm projecting into the injury, and (K) average length of projecting nerve fibers.Statistical differences in (H) are highlighted as *, p < 0.05 and **, p < 0.01 (two-way ANOVA, n = 5-12).No statistical differences were found in I-K (one-way ANOVA, n = 5-12).

Fig. 6 :
Fig. 6: Bioimplantation modulates neuropathological responses at the injury site in injured chick embryo spinal cord slices All images are from 24 days in culture where slices were injured after 4 days and then cultured for a further 20 days.(A, B) Representative fluorescent images showing GFAP staining in (A) control, unimplanted slices and (B) slices with Hemopatch™ implanted into the injury.(A, B insets) High-magnification images of GFAP staining at the injury margin in each condition.(C) Bar chart indicating quantification of GFAP expression at different distances from the lesion margin.(D, E) Representative fluorescent images showing lectin staining of microglia in (D) control, unimplanted slices and (E) slices with Hemopatch™ implanted into the injury.Note, A and D are counterpart images from the same experimental slice, and B and E are also counterpart fluorescent images from the same experimental slice.(D, E insets) High-magnification images of lectin-stained cells within the injury in each condition.(F) Bar chart indicating quantification of numbers of microglia in the injury site.(G, H) Representative fluorescent images showing Tuj-1 staining of neurons and nerve fibers in (G) control, unimplanted slices, and (H) slices with Hemopatch™ implanted into the injury.(G, H insets) High-magnification images of Tuj-1 staining within the injury in each condition.(I) Bar chart indicating quantification of total axonal length per field within the injury site.In (C), statistical differences are highlighted as ***, p < 0.001 (two-way ANOVA, n = 4-5).In (F) and (I) statistical differences are **, p < 0.01 (unpaired t-test, n = 3).HP, Hemopatch™