Mitochondria are organelles in almost all eukaryotic cells, performing essential functions such as ATP production, calcium signaling, lipid biogenesis, oxidative stress regulation, and apoptosis. Mitochondrial diseases are a complex and heterogeneous group of multisystemic disorders caused by mitochondria or nuclear DNA mutations, which lead to mitochondrial dysfunction or impaired energy production. Tissues with high-energy requirements are affected in mitochondrial pathologies, particularly the nervous system, whose metabolic needs are even higher during embryonic development. Patients develop early-onset neurological manifestations and neurodevelopmental disorders that are currently incurable (Falk, 2010).

The pathophysiological mechanisms underlying most mitochondrial diseases causing neurodevelopmental disorders are poorly understood. The main obstacle for researchers is the availability of suitable neurological models that recapitulate the phenotype of these complex diseases and allow for the successful translation of results to patients.

In this review, we summarize the role of mitochondria in neural function and development and various protocols for brain organoid generation. Finally, we focus on published cerebral organoid models of mitochondrial diseases and neurodevelopmental disorders, exploring their utility for drug screening techniques. We approach this from the perspective of the advantages and disadvantages of these cerebral models for studying this range of diseases.

Neurodevelopment refers to the processes involved in the generation and maturation of the nervous system (Gage and Temple, 2013). Neurogenesis or the generation of new neurons from neural stem cells (NSCs) to neural progenitor cells (NPCS), which differentiate into several neural lineages (neurons, astrocytes, and oligodendrocytes), is crucial during the early stages of neurodevelopment, but also in adult tissues (Urbán and Guillemot, 2014; Mira and Morante, 2020; Figure 1). Since the discovery of human embryonic stem cells (hESC), numerous advances have been made, generally focused on cell fate regulation. Several studies have demonstrated a central role of mitochondria in NSCs function and differentiation during embryonic and adult neurogenesis. The role of the mitochondria as energy producers in the form of ATP is one of the factors explaining the association of this organelle with neurodevelopment (Bélanger et al., 2011; Magistretty and Allaman, 2015; Arrázola et al., 2019; Rangaraju et al., 2019). Metabolic programming, defined as significant organ adaptations during the early stages of development, enables the organism to survive nutritional challenges. While embryonic NPCs primarily rely on glycolysis to meet their energy requirements, a metabolic shift toward oxidative phosphorylation (OXPHOS) from stem cells and neural progenitors to mature neurons are critical for proper human neurogenesis and involve specific metabolic regulators, including hexokinase II, lactate dehydrogenase and pyruvate kinase (Wang et al., 2016; Zheng et al., 2016; Figure 1). Evidence shows that mitochondrial function and structure in adult NPCs regulate specific metabolic changes along this neurogenic lineage (Lorenz et al., 2017; Fawal and Davy, 2018; Coelho et al., 2022). In addition, other mitochondrial functions, such as redox signaling, metabolic signaling, and the regulation of gene expression and epigenetics, may impact neurogenesis.

During neurogenesis, mitochondrial dynamics (fission, fusion, and transport) control metabolic adaptations and cell fate (Khacho et al., 2016; Figure 1). The transition to more compromised state cells or NPC is accompanied by mitochondrial fragmentation through NRF2-mediated retrograde signaling. This pathway triggers the suppression of self-renewal and cell differentiation. The transport and proper localization of mitochondria are determining factors in mature neurons. Mitochondria are very dynamic organelles that must travel long distances through the axon, involving microtubules (molecular motors of kinesin and dynein) (Misgeld and Schwarz, 2017; Sheng, 2017; Gumy and Hoogenraad, 2018). The movement of mitochondria, both anterograde and retrograde, is tightly regulated to ensure proper distribution and turnover of these organelles. In mature neurons, mitochondria must be correctly anchored in dendrites and synapses to effectively control the energy supply required for synaptic function and excitability (by local ATP synthesis and Ca buffering) (Rangaraju et al., 2019). In fact, some studies suggest that local Ca²⁺⁺ levels regulate mitochondrial distribution along axons (Greer and Greenberg, 2008; MacAskill et al., 2009; Datta and Jaiswal, 2021). Communication between mitochondria and other organelles, mainly endoplasmic reticulum (ER) and lysosomes, is crucial to regulating various cellular processes through Mitochondria-associated Membranes (MAMs). ER-Mitochondria contact sites play a specific role in mitochondrial fission regulation, calcium metabolism, and lipid transport (Herrera-Cruz and Simmen, 2017; Vallese et al., 2020). Although the importance of mitochondria in the development and function of the nervous system is widely accepted, further studies are required to understand how the mitochondrial networks contribute to dendritogenesis, cell fate, connectivity, and plasticity.

Understanding neurogenesis’s physiopathological mechanisms is essential in neurodevelopmental disorders and mitochondrial dysfunction (Khacho et al., 2019; Oyarzábal et al., 2021). As genetic knowledge of these diseases’ advances, associations between genes encoding mitochondrial proteins and clinical manifestations such as encephalopathy or intellectual delay, among others, are becoming clearer.

Certain mitochondrial disorders differentially affect various brain regions. Like other organs, the nervous system is composed of specific cell types with distinct metabolic needs and characteristics, which can lead to mitochondrial heterogeneity (Rangaraju et al., 2019; Menacho and Prigione, 2020). Evidence suggests that inter- and intra-neuronal mitochondrial diversity may explain the susceptibility of certain neuronal populations to specific mitochondrial disorders. Indeed, fragmented mitochondria have been observed in glycolytic neural stem cells, whereas tubular and interconnected mitochondrial networks are observed in neurons relying on OXPHOS metabolism (O’Brien et al., 2015; Khacho et al., 2016). In comparison to the individual and smaller mitochondria observed in the axons of cortical neurons, elongated and dense mitochondrial networks have been observed in their dendrites (Popov et al., 2005; Chang and Reyolds, 2006; Dickey and Strack, 2011). Various authors have demonstrated differences in the mitochondrial transcriptome and proteome across different populations of neuronal cells (Wynne et al., 2021). For example, there are differences in Mitochondrial Calcium Uniporter (MCU)-mediated calcium buffering capacity between granules and Purkinje cells in the cerebellum, with the former exhibiting greater capacity. This coincided with increased expression of MCU (Fecher et al., 2019). While it is acknowledged that mitochondrial function is associated with the structure and morphology of these organelles, it is important to comprehend how biochemical and structural dysfunctions of mitochondria affect specific cellular populations.

The international neuroscience community has managed, through in vivo and ex vivo data, to develop an infrastructure for mapping the brain and elucidating the mechanisms involved in neural pathways and their associated disorders. Projects such as Multiomic Human Brain Cell Atlas, Human Brain Project (HBP) (S. LaFee, N. Mlynaryk), EBRAINS Cellular Level Simulation Platform the Blue Brain project (Markram et al., 2015), allow to create computational simulations and in silico modeling using AI, neuroinformatics and biocomputing. These computational modeling tools of the brain would allow scientists to validate their results in silico.

Traditionally used disease models to study neurodevelopmental pathologies with mitochondrial involvement and their limitations are described below:

Mitochondrial functions are tightly controlled and coordinated by nDNA (encodes approximately 1500 mitochondrial proteins) and mtDNA (encodes 13 proteins of the electron transport chain complexes (ETCs), 22 transferents RNA (tRNAs) and 2 ribosomal RNA (rRNAs). The OXPHOS system is composed of five multi-subunit complexes (Complexes I-V; CI-CV), located in the inner mitochondrial membrane. Mitochondria rely on the import of approximately 250 additional nuclear encoded proteins required for OXPHOS function. Each cell contains hundreds to thousands of mtDNA copies, entirely inherited maternally. Unlike the nDNA, the copy number of mtDNA and the proportion of mutated copies (heteroplasmy) varies between cells. Heteroplasmy can be both cell type and mutation specific and the increase in mutated mtDNA copies can lead to diseases. Several factors can alter the degree of heteroplasmy, which can affect embryogenesis and genetic drift across generations. To date, a multitude of genes associated with mitochondrial diseases have been identified in both genomes. The strategies for modeling nDNA or mtDNA mutations are different. Specific and effective methodologies for targeted mutagenesis or mtDNA modification have not yet been established (Hu et al., 2019).

Traditionally, fibroblasts or peripheral blood lymphocytes from patients, and cytoplasmic hybrid (cybrids) cells have been the primary 2D-cell models for studying mitochondrial diseases (Wilkins et al., 2014; Sazonova et al., 2018). Research can be conducted using these primary cells isolated directly from patients and healthy donors or established cell cultures deposited in cell banks. Fibroblasts are the predominant group of cells in the human body, located in almost every tissue and organ. They play a role in numerous functions in connective tissue, being the main source of extracelular matrix (ECM). Fibroblasts are derived from embryonic mesoderm. Primary fibroblasts are easily obtained during muscle or skin biopsy and fibroblasts culture is highly proliferative, providing a renewable source of cells in vitro (Hu et al., 2019).

Cytoplasmic hybrid (cybrid) are cell lines developed to eliminate the interference of nuclear background and characterize the physiopathology of mDNA mutations. For that, the nucleus was removed from patient-derived cells such as immortalized lymphocytes or fibroblasts, and cytoplasts were fused with cells lacking mtDNA. Other types of cancer or immortalized cell lines have been used as models of mitochondrial diseases (Wilkins et al., 2014).

However, these cell lines have certain limitations, which are described below:

The specific factors and pathways regulating cellular differentiation, migration, spatial and morphogenic patterns and neurological development, involved in the physiopathology of neurodevelopmental disorders and mitochondriopathies, are not recapitulated in 2D cell cultures. On the other hand, studies and modeling of mitochondrial diseases are complex, mainly due to the complexity of mitochondrial genetics and the interaction between mutations in the mitochondrial genome and nuclear genome and it forces us to explore models that are more physiological and closer to the patient, such as hESC and organoids. The development of 3D cultures and brain organoids derived from patient cells is driving the modeling of such complex and rare diseases, presenting an opportunity to overcome the limitations of 2D cultures and animal models, which are far removed from the patient. However, the coexistence of both 2D and 3D cultures, depending on the research objective, is essential for advancing the study and development of treatments for human diseases. All these limitations and advantages of the 2D models are summarized in Table 1.

Human pluripotent cells, such as Embryonic Stem Cells (ESCs) and induced Pluripotent Stem Cells (iPSCs), have great potential for human rare disease research. Several authors to differentiate these stem cells have applied previous knowledge from the field of developmental biology (Clevers, 2016). However, thanks to Eiraku and Sasai’s research, in 2012 the first in vitro neural organogenesis model systems emerged. These systems were able to replicate not only the cellular differentiation process observed in adherent (monolayer) cultures of stem cells, but also the spatial and morphogenic patterns involved in the three-dimensional (3D) development of certain neural structures (Eiraku and Sasai, 2012). Late, Lancaster and Knoblich (2014) took this approach to a higher level, generating the so-called “cerebral organoids”, which were described as an in vitro 3D-culture system capable of generating different brain regions, with a brain-like organization in human development (Lancaster et al., 2013).

Currently, the term “organoid” refers to cellular aggregates derived from primary adult tissues or stem cells capable of self-organizing and forming organotypic structures in vitro (Shankaran et al., 2021). Some of the specific brain functions that are critical to be represented in brain organoids to model mitochondrial diseases are mainly neurogenesis, synapsis and function of glial cells (CNS homeostasis maintenance). For instance, it is important that the brain organoids are composed of more than one cell type (1) since some studies affirm that astrocytic mitochondria metabolize long-chain fatty acids more efficiently than neuronal mitochondria (Fecher et al., 2019). This is of interest because fatty acids seem to contribute to the cellular energy metabolism of the brain, although in a lower proportion than glucose (the major oxidative fuel for the brain) (Ebert et al., 2003). On the other hand, it is also essential that brain organoids show specific functions (2), such as synaptic function. For example, in neurons, mitochondria play key roles in energy metabolism and Ca²⁺ regulation, necessary for axonal and dendritic development, axonal regeneration and synaptic function (Rangaraju et al., 2019). Likewise, it is key that brain organoids present a cellular organization similar to that of the brain (3) for proper cellular functionality. For instance, in the brain, it has been described that neurons operate in a strong physical and metabolic association with astrocytes and glial cells, and the dysregulation of this cooperation contributes to the development of brain disease (Bonvento and Bolaños, 2021). In addition, through a mechanism called transmitophagy, the neurons can secrete damaged mitochondria that can be taken up and degraded by adjacent astrocytes (Davis et al., 2014), highlighting the need for adequate cellular and tissue organization in the brain. Since the pathways involved in mitochondrial metabolism, calcium buffering, mitochondrial dynamic (mitochondrial trafficking through the axons, mitophagy, fusion and fission), regulation of apoptosis and oxidative stress are essential to an adequate differentiation, development and maturation of CNS, the recapitulation of all these processes in brain organoids is a major challenge.

Unlike other types, human brain organoids are usually generated only from human pluripotent stem cells (both hESCs and hiPSCs), mainly due to the low regenerative capacity of the adult human brain (Tang et al., 2022). Pluripotent stem cells can differentiate into more specialized cells derived from the three germ/embryonic layers (endoderm, ectoderm, and mesoderm) (Liu et al., 2020). On the one hand, hESCs are stem cells derived from the human blastocyst’s inner cell mass (ICM). They can be cultured in vitro indefinitely and retain the expression of several key epiblast markers, such as the transcription factors Oct4 and Sox2 (Nair and Gongora, 2020; Weatherbee et al., 2021). On the other hand, hiPSCs are stem cells derived from the in vitro reprogramming of human fibroblasts. They emerged in 2007 thanks to the research of Yamanaka and Takahashi, who discovered the four transcription factors (Oct4, Sox2, Klf4, and c-Myc, abbreviated as “OSKM”) necessary for human fibroblasts reprogramming (Takahashi et al., 2007). Both hESCs and hiPSCs are two-dimensional (2D) cultures, and they are used in biomedical research as an excellent experimental platform for studying human development, disease modeling, drug screening, and cell therapy (Yang et al., 2021; Figure 2).

Brain organoids recapitulate some features of the fetal human brain, not only at the cellular level but also in structure, tissue organization, and developmental trajectory. Their applications have continued to expand, and protocols for generating human cerebral organoids from ESCs and iPSCs are growing exponentially. Many authors consider human brain organoids to be an excellent model for studying several aspects related to the biology of human brain development (Di Lullo and Kriegstein, 2017; Qian et al., 2019). Furthermore, the possibility of generating human brain organoids from hiPSCs derived from patients (e.g., with a neurodevelopmental disease) allows in vitro modeling of several brain disorders (while maintaining the genomic context) (Sidhaye and Knoblich, 2021), which would contribute to the study and understanding of the mechanisms involved in mitochondrial diseases and neurodevelopmental disorders (Soldner et al., 2011).

Studies on how mitochondrial dysfunction affects specific pathways involved in developing and maintaining the nervous system need to be addressed. Neurological 3D models for overcoming the limitations of 2D cultures and animal models are expanding enormously. The ability to replicate, in the laboratory, the early stages of human brain development to understand how mitochondrial dysfunction affects its normal development represents a great hope for researchers and patients of this group of rare diseases. Currently, the challenge is to standardize the protocols for the generation of brain organoids, to automate and scale the processes (from generation to the analysis of large amounts of data), and to efficiently translate results to patients. Cooperation between disciplines, such as neurological development, mitochondrial diseases, and engineering in consortia with private companies, is essential to obtain meaningful and reliable results.

Diseases caused by alterations in mitochondrial DNA (mtDNA) and nuclear DNA (nDNA) have proven challenging to model due to their inherent complexity, particularly the difficulties in manipulating mtDNA (Schon et al., 2020). This hampers the generation of diverse in vitro and in vivo models. However, recent advancements in iPSCs modeling, particularly with 3D organoids, have provided opportunities to develop innovative models for these diseases. Brain organoids enable the assessment of specific mtDNA and nDNA mutations during neurodevelopment in a human disease context. They can uncover molecular mechanisms involved in pathology and evaluate the efficacy of targeted treatments or gene therapy efficacy in vitro. The brain organoids and NPCs generated to date as models of mitochondrial diseases affecting the nervous system are discussed below and summarized in Table 2. Most of these models focus on diseases caused by mutations in mitochondrial DNA.

One of the more relevant questions for the researchers is whether organoids might replace or reduce the use of animal models in drug screening and cancer immortalized cell lines in high-throughput screening (HTS) (Entzeroth et al., 2009). The generation of organoids to model diseases and their use in screening drugs or cell therapy represent a great expectation for researchers and patients (Costamagna et al., 2021; Lampart et al., 2023). However, although in the last 10 years many new drugs have been developed to treat different human diseases, in the field of mitochondrial pathologies with neurodevelopmental disorders, the discovery of new compounds or treatments has not been very effective. Despite efforts made by the scientific community, it is estimated that 95% of drugs fail clinical trials (Wong et al., 2019). Disease models that recapitulate pathophysiological mechanisms and phenotypes of mitochondrial and neurodevelopmental disorders remain a fundamental bottleneck in order to identify new compounds and successfully translate them to clinical practice.

In this section, we review the main limitations and advantages of using brain organoids in high-throughput screening (HTS) or drug repurposing (DR) to discover new treatments for mitochondrial diseases that cause neurodevelopmental disorders.

Studies about HTS or DR in brain organoid models for neurodevelopmental diseases and mitochondrial disorders have not yet been reported, which to some extent reflects the enormous difficulty and limitations of brain organoids in modeling these groups of diseases and their applications in drug discovery.

High-throughput screening (HTS) is an automated/robotic system to test up to hundreds of thousands or even more compounds in vitro, to identify new potential drugs (Cota-Coronado et al., 2020; Durens et al., 2020). This system is one of the most efficient and quick processes and provides starting points for drug design. In general terms, the main requirements for HTS assays are summarized in Figure 4.

Traditionally, drug discovery and HTS assays rely on cancer immortalized cell lines, whose energetic requirements are less dependent on oxidative mitochondrial phosphorylation (OXPHOS) (Horvath et al., 2016). Furthermore, this cellular model lacks functional and phenotypical properties of neuron cells and the brain. Genetic and epigenetic abnormalities resulting from long-term in vitro culture in such an unphysiological environment also contribute to substantial changes in the physiological characteristics of immortalized lines. This has an impact on the heterogeneity of cultures and results between passages and laboratories, in the same cell line. Therefore, they are not an effective model for drug discovery model for mitochondrial pathologies. On this basis and considering that brain organoids are 3D cytoarchitectures resembling the embryonic human brain, many researchers and companies are developing protocols and pipelines for the generation, maintenance and analysis of brain organoids to model neurodevelopmental diseases (Lee et al., 2017; Yang et al., 2023; Zhou et al., 2023). However, it is necessary to address new technical approaches that allow us to evaluate the mitochondrial function and structure in these 3D models.

There is evidence that the response to therapy in patients correlates with effects in corresponding patient-derived organoids, such as pancreatic, ovarian, uterine, colon, gastrointestinal, brain, and cancer (Vlachogiannis et al., 2018). In addition, the beneficial effects of drugs, such as PTC-124 and Ataluren, in animal models of cystic fibrosis were not demonstrated in human intestinal organoids, which is consistent with the results in two-phase clinical studies. Although there are still challenges to overcome, patient-derived organoids represent a potential preclinical model (Zainal Abidin et al., 2017). The power of HTS in organoid-based drug discovery have been highlighted in cancer and to lesser extent in neurodegenerative diseases, such Alzheimer’s and Parkinson’s (Krall et al., 2020; Park et al., 2021; Renner et al., 2021; Jeong et al., 2023). Nevertheless, based on reports, neurodevelopmental disorders with mitochondrial involvement are not being covered sufficiently.

HTS assays can be target- or phenotypic-based screening. For most rare diseases, pathological mechanisms and molecular targets are unknown, so target-based screening may be impossible. Phenotypic screening results are more attractive for this kind of disease, since it only requires the monitoring of some specific biological functions after compounds administration (Khurana et al., 2015). In an organoid-based HTS assay, it is crucial to establish the phenotypic criteria to classify the control and pathological conditions. These phenotypic features must be measurable, easily detected, in addition to matching the clinical symptoms. Ultimately, viability and metabolic activity are two of the most important measurable parameters in HTS assays in different organoids, but the information obtained from such measurements is very limited. Among the parameters or readouts to establish the phenotypic deficits for mitochondrial and neurological disorders are organoid size, cell migration, gene expression, cortical lamination, neurogenesis, axonal growth, electrophysiological parameters, oxidative stress, oxygen consumption or ATP production. Some authors have shown that parameters of mitochondrial functions used in 2D cultures, specifically ATP production, can be measured in a 96-well plate in neuronal organoids. Oxidative stress and mitochondrial potential measurements have also been assayed in 96-well plates, in this case in retinal organoids (Nzou et al., 2018; Le et al., 2021; Ludikhuize et al., 2021).

Currently, protocols for the generation of 3D models have advanced remarkably; however, their use in high-throughput drug screening is still limited, especially for neurodevelopmental and mitochondrial diseases, as mentioned above. Complexity of mitochondrial diseases causing neurodevelopmental disorders, lack of profitability of industry derived from the low frequency of these diseases and the difficulty that still exists in developing brain organoids reproducible, reliable and scalable are some of the limitations. Although innovations such as vascularization, genome editing, omics-techniques, high-content imaging (HCI), machine-learning-based and automated systems are already a reality in organoids, protocols need to continue evolving to improve tissue communication, cell patterning signals, body axis simulation and spatial-temporal control of differentiation (Lokai et al., 2023). Engineering and instrumentalization advances to reconstitute brain-mimetic microenvironments will improve the development of efficient and reliable platforms to model and drug discovery of human brain diseases (Patel et al., 2024). Organoid on-chips are microfluidic chamber devices that provide a periodic flow and mimic the cerebrospinal and interstitial spaces. Brain organoids can be inserted into these chips, facilitating the oxygen supply and nutrient/waste exchange by controlling the environmental conditions of brain organoids (Cho et al., 2021; Zhou et al., 2021; Saorin et al., 2023). Microfluidic chips lead to a significant reduction of cell death and consequently an improvement of maturity, although their use is not enough extended.

Aspects such as production efficiency and organoid homogeneity, but also culturing conditions and handling are important challenges in drug screening. Optimization and standardization of reliable protocols are crucial to obtain meaningful results. Renner et al. (2020) have design an automated HTS workflow in 96-well formats (“automated midbrain organoids,” AMOs) to generate and analyze human midbrain organoids. Automation of all processes, from generation, maintenance, immunostaining, tissue clearing and HC imaging, allows scaling and homogenization of brain organoids, as well as drug testing (Renner et al., 2020). High-throughput screening (HTS) assays require the generation, maintenance, analysis, and interpretation of a large number of samples. Consequently, the analysis of data and results is complex and time-consuming. Fortunately, the progression of artificial intelligence (AI) tools already allows us to process and classify large datasets, as well as correlate results obtained even from different areas (Renner et al., 2021; Metzger et al., 2022).

Drug repurposing or drug repositioning is defined as the discovery of new applications of approved or investigational compounds. This process saves time, and it is more cost-effective than the development of new orphan treatments, representing a significant advantage for rare diseases and patients (Shah et al., 2021; Kakoti et al., 2022). However, in many areas of research, repurposed drugs may not be effective in clinical trials or may not maintain the benefit-injury equilibrium (Fetro and Scherman, 2020). Drug repurposing has been used for years, but new approaches have been developed to make them more systematic and efficient. As we mentioned before, drug screening in organoid models has been developed on a larger scale for diseases such as cancer, and other common pathologies (Mao et al., 2023; Mertens et al., 2023). Recently, Spelier et al. (2023) have developed a drug repurposing assay, in a 384-well screening format, using cystic fibrosis patient-derived intestinal organoid.

Another aspect to consider is the legal and regulatory requirements, which should be carefully analyzed before any study of drug repurposing. Computational drug repurposing, because of the progression in HTC technologies, and omics-based data related to disease indications, are considered as cutting-edge techniques for drug discovery (Emon et al., 2017; Paranjpe et al., 2019; Hendrickx et al., 2020). We wanted to focus on the use of these methodologies for drug discovery in rare diseases and particularly neurological diseases with mitochondrial alterations (Hodos et al., 2016). While these disciplines continue to advance, specific experimental and economic barriers arising from these complex diseases need to be overcome.

Advancements in the generation of brain organoids represent a revolution in the research of mitochondrial diseases affecting neurodevelopment. However, despite the increasing number of brain organoids for modeling rare diseases, the techniques for metabolic and structural analysis of mitochondria have not yet been sufficiently developed. Indeed, most readouts in these 3D models are based on structural characteristics.

There is a promising journey ahead. Brain organoids fill a significant gap in research of human neurological disorders. Nevertheless, there are several theoretical and practical limitations to understanding the role of mitochondria in the development of the neurological system and the progression of diseases. Currently, there is a lack of precise information about patterns of neural maturation in humans, and the pathophysiological mechanism underlying most mitochondrial diseases affecting the nervous system remains unknown. Regarding practical obstacles, heterogeneity of size and morphology, survival, handling and scalability stand out. Some of these limitations may be overcome by the unification of protocols and the adoption of engineering solutions, automated liquid handling and high content analysis. These factors are essential for drug screening. Current studies have focused on structural readouts, such as survival and neural differentiation, gene expression, size, cortical lamination, etc. However, aspects such as communication between different neural cell types and vascularization are being addressed from various scientific and technological fields.

The cooperation between private companies and researchers of different disciplines, as well as discussions on these approaches, are essential for the use of brain organoids as a proper model of mitochondrial diseases and neurodevelopmental disorders.

RC: Writing – original draft, Writing – review and editing. EG-M: Writing – review and editing. ES: Writing – review and editing. MB: Writing – original draft, Writing – review and editing. BM-D: Writing – original draft, Writing – review and editing. CS-O: Writing – review and editing, Funding acquisition, Supervision. IL: Writing – original draft, Writing – review and editing. MC-A: Writing – original draft, Writing – review and editing.