In preclinical drug development, the gold standard is still to utilize animal models, usually mice or rats, and this is also true in brain trauma research. The development of in vitro injury models for neurological diseases, more specifically for central nervous system (CNS) trauma, usually mimics human conditions to better understand the specific elements of the injury and to test the efficacy of potentially promising medicines [4,7]^[1][2]. Current in vitro models for studying traumatic brain injury are not effective enough, because they do not accurately simulate all the multifaceted and heterogeneous aspects of trauma. However, in vitro models have certain advantages over animal testing, such as decreased costs, higher throughput capabilities, and in the end bigger control over experimental conditions.

Mechanical types of injuries used to mimic in vitro trauma include mechanical stretch, transection or scratch, blunt impact, and compression. The most important events included in traumatic brain injury (TBI) pathophysiology examined by in vitro models are membrane disruptions that lead to ionic dysregulation, then inflammation, damage of microtubules and axons, and in the end cell death. In in vitro studies, it is also important to apply controlled and repeatable injuries to cells, to adequately mimic the microenvironment of brain cells and disruptions of the mentioned microenvironment [4]^[1].

Stretch-induced injury models are the most widely used in vitro TBI models [8]^[3]. One of the simplest and accepted methods is the use of stretch in cultured murine brain cells via a cell-injury controller. Using stretch to culture brain cells helps study cellular and molecular events involved in TBI, including the blood−brain barrier (BBB) disruption. The principle of the assay is to induce injury to cultured cells through the delivery of a controlled pulse of compressed nitrogen gas to the cultured cells in the medium. Cells are cultured on specific plates with elastic membrane bottoms, and the controller enables the regulation of pressure strength, which determines the extent of injury to adhered cultured cells.

Transection and scratch injury models are used to analyze trauma-induced axotomy and to test the therapeutics efficacy of therapies aimed at promoting axonal regeneration. Primary axotomy, relatively rare in TBI in comparison to spinal cord injury, is investigated by the induction of transection injury, which not only mimics primary physical injury, but it also leads to the activation of secondary injury responses similar to in vivo conditions. The microenvironment of the injured brain is affected through the promotion of glutamate-induced excitotoxicity, release of pro-inflammatory cytokines, expression of growth factors, axonal growth-inhibiting molecules, changes in cell metabolism, and production of reactive oxygen species [9]^[4]. The scratch assay is a simple transection in vitro model where primary neurons, astrocyte cultures, or immortalized cells are scraped using a pipette tip. It is commonly used to induce astrocytic reactivity and assess the response of these cells, as they are extremely important in wound closure and healing processes [10]^[5].

One of the mechanisms involved in brain trauma pathophysiology is the process of cavitation. It is a process of vaporization, bubble generation, and bubble implosion that results from decreases and increases in pressure. The so-called “flyer-plate model” is a model that represents such a type of trauma. It is an in vitro high-energy model of trauma, and the principle of the method is to hit the bottom of a cell culture causing cavitation and consequently creating shock waves inside the well and the medium. This model is useful for analyzing the cellular responses to micro cavitation, particularly in neuronal cells [11]^[6].

A model that is also used to investigate neurotrauma is “brain-on-a-chip”. It consists of 3D cell cultures, and it is used in an effort to model the physiological responses of brain tissue in a microfluidic environment. 3D culturing methods represent a good model to determine the characteristics of the glial scar, the main feature of secondary injury. The main use of the model is the high throughput screening of compounds in pharmacological and toxicology studies. It is also a good model to study diseases process, by adding free radicals, causing inflammation or using modified cell lines to stimulate diseases. This type of model requires mechanical injury at the microscopic level of axons and neurons, so the credibility of such an injury is questionable and quite difficult to carry out [3]^[7].

Overall, in vitro CNS trauma models are very valuable in clarifying brain pathologies after the primary and secondary injury to explore new clinical treatments. Nevertheless, there are some important limitations, such as the differences between cultured cells and tissues and their matching in vivo counterparts due to variances in the microenvironment. The problem which occurs is how to prepare ex vivo brain or spinal cord slices, without damaging cells and tissues and affecting cellular and molecular responses due to experimental injury procedures and therapy treatments [12]^[8].

2. In Vitro Models of Cerebral Hypoxia/Ischemia

A cascade of cellular events that begin with a loss of oxygen, followed by energy depletion, excitotoxicity, and subsequent complex changes in tissue metabolic activity, are main characteristics of the cerebral hypoxia/ischemia pathophysiology. Intense post-ischemic inflammation is mediated by the activation of various pro-inflammatory cytokines and chemokines, and perturbed mitochondrial function is responsible for the apoptotic pathways activation [13]^[9]. Several in vivo models have been established to mimic clinical conditions of global (e.g., cardiac arrest) or focal (e.g., stroke) cerebral hypoxia/ischemia. Although in vivo models most realistically mimic the clinical conditions, including reperfusion, in vitro models of cerebral hypoxia/ischemia are important for understanding and elucidating the complexity of the pathophysiological cascade of biochemical and molecular mechanisms involved in these processes [14]^[10]. There are several possibilities to induce hypoxia/ischemia in vitro, but the most commonly used models are inhibition of cellular metabolism by chemical or enzymatic blockade and oxygen-glucose deprivation (OGD) [15]^[11]. The introduction of new technologies allows better modeling of ischemia-reperfusion injury in vitro by using OGD media flow perfusion methods or different cell culture platforms [7,16,17]^[2][12][13]. Inhibition of cellular metabolism can be triggered by various chemicals that interfere with the electron transport chain and lead to the energy deficiency that occurs in the initial phase of cerebral hypoxia/ischemia. The most used chemical inhibitors include antimycin, rotenone, 2-deoxyglucose, or sodium azide. In addition, it is possible to induce cell injury by using NMDA or glutamate receptor agonists to mimic in vivo excitotoxic conditions that result in a substantial extracellular increase in glutamate [18,19]^[14][15]. The enzymatic methods used to induce in vitro hypoxia/ischemia conditions are the glucose oxidase/catalase systems, consisting of catalase and 2-deoxyglucose. The advantages of both methods are their relatively simple and accessible methodology and the ability to rapidly gain insight into the specific mechanisms involved in the pathophysiological cascade during the hypoxia/ischemia process. However, in vitro assays cannot provide insight into the complexity of the processes that occur under in vivo conditions. A particular problem is the lack of appropriate possibilities to test the processes occurring during reperfusion in vivo. Oxygen-glucose deprivation (OGD) is the most commonly used and relevant method to create in vitro hypoxia/ischemia-like conditions that mimic stroke. This is usually conducted by exposing cells to glucose-free media and displacing oxygen with a nitrogen/carbon dioxide mixture in a hypoxia chamber. This model allows mimicking reperfusion conditions by reintroducing glucose with a return to atmospheric oxygen. OGD was described to induce neuronal depolarization within 10 min of onset. Within 30 min, there was depolarization of astrocytes and acute cell swelling followed by apoptotic and excitotoxic necrotic cell death, consistent with observations of ischemia-reperfusion injury in vivo. Experiments performed in vitro with hypoxia alone better represent cerebral hypoxic conditions such as carbon monoxide poisoning than ischemic stroke, because they mimic conditions in which blood flow is maintained [22]^[16]. Many in vitro studies have shown that hypoxia alone causes dramatic changes in endothelial cell actin cytoskeleton (EC) and tight junction protein localization in BBB models. The majority of experiments were performed with the immortalized BV-2 microglia cell line, a proven replacement for primary microglia [23]^[17]. These cells spontaneously show a dual phenotype with predominant growth in an amoeboid cell form cultured under standard conditions of 10% DMEM (DMEM supplemented with 10% v/v FCS, fetal calf serum). 

3. Cell Culture Platforms Used in Traumatic Brain Injury and Brain Ischemia In Vitro Models

Cellular in vitro platforms used up until now to model TBI and stroke are brain slices, organotypic cell cultures, primary neuronal cells, immortalized cell lines, and different types of stem cells of human and rodent origin [2,24,25]^[18][19][20]. In the in vitro studies of TBI, most used cells are the ones of rodent origin, and of those rats account for about 70 % [2]^[18]. When it comes to the use of cells of human origin, it has been reported that they were used in only about 15% of studies. In the studies that used human cells, the researchers have most frequently used immortalized cell lines, followed by primary cells, and induced pluripotent stem cells (iPSCs).

3.1. Primary Cell Lines

Even though the use of the so-called monocultures (cell cultures that consist of a single cell type) does not provide with the information on the complex tissue and organ reactions to noxious events or the effects of pharmacological interventions, they still give an important insight into cell-specific responses as well as the reaction of particular cell types to neuroprotective agents. Most commonly, rodent (rats more often than mice)-derived neuronal cells are used. However, an important part of the tissue response to ischemia or mechanical damage belongs to glial cells, i.e., astrocytes, microglia, and oligodendrocytes. In studies related to the BBB reaction, endothelial cells have also been used. From the technical standpoint, primary cell isolation and preparation can be viewed as time-consuming, and the purity of cell cultures might be challenging to achieve, and this needs to be considered regarding the reproducibility of the results. Additionally, primary cells are dissociated from either embryonic tissue or from the animals sacrificed in early postnatal days, so cells need to be maturated during a period of time.

3.2. Immortalized Cell Lines

The use of established cell lines has plenty of advantages in in vitro studies. These cells are highly proliferative, and they offer high reproducibility with the possibility of easy genetic manipulation. Additionally, many of used immortalized cell lines are of human origin, i.e., with human genetic backgrounds. However, these types of cells might require differentiation protocols for them to reach necessary morphological and/or physiological characteristics. In addition, immortalized cell lines have an oncogenic origin, and their main characteristic is the high proliferation rate, something that is clearly not a feature typical for cells of the CNS origin. As is the case with dissociated primary cultures, immortalized cell line cultures lack in that they are unable to imitate higher-dimension interactions between cells as well as to consider the influence of the extracellular matrix (ECM) environment the cell reactions to injurious events.

3.3. Co-Cultures, 3D Culture Models, and Brain Organoids

By combining different cell types, studies are able to imitate to higher-degree complex interactions that occur in the in vivo conditions. As the human brain is built of different cell types—neurons, astrocytes, microglia, oligodendrocytes, pericytes, and the epithelial cells, combining them in a culture provides a more useful system for studying complex cell-to-cell interactions that occur in the CNS, in both physiological and pathological conditions.

Co-culturing of cells can be achieved in a 2D cell culture environment, but a more representative approach involves the use of 3D cell culture models that allow establishment of multiple interactions between different cell types and the ECM. Adding multidimensionality to in vitro systems also enables cells to develop distinct phenotypes that are more physiologically relevant. Major advantage of using 3D cultures in in vitro research is the ability to reconstruct 3D organization of cells, and it represents an important step in an effort to imitate normal cell-to-cell and cell-to-ECM interactions. Additional benefit of using 3D in vitro models is the increased viability of difficult-to-culture cells, improved cell-type-specific function and gene expression, and the accumulation of secreted factors in the ECM that could have pathological effects and cannot be studied in 2D cell culture. Cells grown in 3D cultures can self-organize and differentiate, and they allow highly scalable and high-throughput analyses of cell responses and can be used to obtain electrophysiological network activity outputs. Different types of 3D systems for culturing cells include the scaffold-free, scaffold based, and hybrid culture strategies. Scaffolds are structures that are made of biopolymers organized in a way to imitate the physiological ECM. They are matrices that can be made of hydrogels or can be of solid, porous, and fibrous build. Other than providing the structural support, scaffolds can be enriched with different molecules, e.g., growth factors, thus adding to the similarity of the cultured environment to in vivo conditions. In both scaffold-based and scaffold-free systems, cells can be cultured in 3D structures. Multicellular aggregates called spheroids are 3D structures that can mimic various as well as tumors [38,39,40]^[21][22][23]. Neurospheres are neuronal aggregates created from neural progenitor cells that can be manipulated to generate brain-region-specific cell types (e.g., neurons and astrocytes). They have proven to be useful in the neurodevelopment and neurodegenerative diseases research, but one major limitation is the creation of the necrotic core in the central part of the spheres that occurs due to the insufficient perfusion and the lack of vascularization.

3.4. Organotypic Slice Cultures

Organotypic slices are used also in the TBI and stroke research as a tool to study the effects of injury on cells that preserve neuronal connections [49]^[24]. They not only allow analysis of electric activity in circuits and measurement of calcium changes during injury, but are also amenable to interventions, e.g., by using the optogenetic approach. In the brain ischemia research, hippocampal organotypic slices are frequently used, and they proved to be useful in studying the pathophysiology of stroke [50^[25][26],51], neuron-glia interactions [52^[27][28],53], as well as a useful platform for the screening of the therapeutic interventions, both pharmacological and non-pharmacological [54,55]^[29][30]. However, the technique of obtaining the tissue slices requires that it is cut out, and this trauma itself could be used as an injury model [25]^[20]. As an example, it has been found that organotypic slices develop epileptiform activity after a week in culture that resembles changes related to the post-traumatic epilepsy [56]^[31].

3.5. Human Induced Pluripotent Stem Cells

The use of iPSCs technology, specifically human iPSCs, has many advantages in the different-disease/disorder research. This technology was first described in 2007, when the adult human fibroblasts were reprogrammed back to a pluripotent state using specific factors [57]^[32]. These cells maintain the genetic features of their parent cells, but with the added property of being able to proliferate easily with the additional possibilities of genetic manipulation. Even though human iPSCs can be used as a powerful tool to study diseases in many different organs, they are especially useful in neurological disorders research. The reason for this is the fact that it is particularly challenging to obtain human neuronal tissues and cells and also because of the distinctive properties of human CNS. Some of the advantages in using human iPSCs are the ability to derive specific cell types (neurons, astrocytes, and microglia) from controls and individuals suffering from a certain disorder/disease. Thus, it creates an ideal environment to screen on-target drug effects. They mimic brain development and pathologies better than both human immortalized cancer cell lines and primary rodent cell cultures.