Neurodevelopmental disorders represent a group of heterogeneous conditions that affect the brain function and neurological development, causing impairments in cognition, communication, behavior, and motor functioning (1, 2). The symptoms and behaviors of neurodevelopmental disorders usually appear and diagnose during childhood, but typically persist across a lifespan (3). Individuals with these disorders often experience difficulties in socialization and a reduction in self-esteem and interpersonal skills (2). In the framework of the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), conditions fall under the umbrella of neurodevelopmental disorders include intellectual developmental disorders, communication disorders, autism spectrum disorder (ASD), attention deficit hyperactivity disorder (ADHD), specific learning disorder, and motor disorders (1). Among these conditions, ASD and ADHD are the two most common neurodevelopmental disorders observed in childhood. ASD is characterized by delayed or abnormal language development, deficits in social interaction, repetitive behaviors and restricted interest, whereas ADHD is characterized by severe deficits in attention with or without hyperactivity and impulsivity (1, 4). Despite the differences in the core symptoms and diagnostic criteria of ASD and ADHD, these two disorders frequently co-occur (5). In this context, a significant proportion (30–80%) of individuals with ASD present with ADHD, and ASD occurs in about 20–50% of ADHD individuals (6). The phenotypic overlap between ASD and ADHD has been explained, at least partly, by shared genetic factors (7–9). As with most complex diseases, other factors besides genetics such as environmental, infectious, and traumatic factors among others contribute to neurodevelopmental disorders (10). Studies have also suggested that common dysfunctional pathways involving mitochondrial activities could account for multiple clinical signs in neurodevelopmental disorders (11–13). Mitochondria produce most of the cellular energy through oxidative phosphorylation, and are also an important source reactive oxygen species (ROS). Beside energy production, mitochondria play a vital role in calcium homeostasis, innate immune, inflammatory responses, red-ox balance and apoptosis (14). Mitochondria have their own genome (mtDNA), which encodes 37 essential genes for proper mitochondrial and cellular functions (15). Each human cell contains multiple copies (about 1,000–10,000) of mtDNA (16). The mtDNA-cn reflects the abundance of mitochondria within a cell and can vary according to the cell’s energy requirement (16, 17). The mtDNA-cn may also differ according to age and general age-related decline in mtDNA-cn was observed across tissues and species (18). Moreover, a study by Chu et al. (19) showed that parental mtDNA-cn was significantly lower than that of their children. Several factors make the mtDNA particularly vulnerable to oxidative stress and other sources of genotoxic damage such as its close proximity to the site of ETC-mediated ROS production, lack of protective histones and limited DNA repair capacity. mtDNA oxidative damage can lead to mutations or changes in mtDNA-cn, which may finally lead to mitochondrial dysfunction with more ROS production (20–22). Changes in the mtDNA-cn could alter the expression of mitochondrial genes and cause abnormalities in mitochondrial function and energy production. The mtDNA-cn can be measured by real-time PCR as the ratio of mtDNA copies per nucleated cell in different cell types such as blood, urine, saliva, and tissue biopsies. In previous studies, altered mtDNA-cn was linked with various human diseases such as cancer, neurodegenerative diseases, and psychiatry conditions (23–25) and was suggested as a potential biomarker of these disease. Recent evidence shows that quantitative changes in mtDNA in different tissues are associated with the pathogenesis of neurodevelopmental disorders. For example, increased mtDNA-cn in the peripheral blood was reported in patients with ASD and correlated with neurological manifestations and communication (26). Conversely, a decrease in mtDNA-cn in leukocytes (27) and peripheral blood mononuclear cells was observed in ASD patients (28). On the other hand, no significant difference in mtDNA-cn in temporal lobe was found between ASD patients and controls, suggesting that mitochondrial abnormalities in ASD may be triggered by mechanisms other than mitochondrial biogenesis (29). In addition, higher mtDNA-cn was reported in the peripheral blood of patients with ADHD (30), and remained elevated after treatment (31). These studies imply a complicated role of altered mtDNA-cn in blood and other tissues in ASD and ADHD. Since the involvement of mtDNA-cn in these disorders remains ambiguous, we performed a meta-analysis to summarize the current evidence and to provide more conclusive results of the relationship between mtDNA-cn and ASD/ADHD. We also analyzed changes in mtDNA-cn in different biological samples, and in aged-matched patients and controls.

Mitochondria play a key role in energy production and cellular function. They are the only organelles that have their own DNA (mtDNA), which presents in multiple copies per cell. The mtDNA encodes only 37 genes, and 13 proteins translated by these genes are subunits of the electron transport chain (ETC), involved in the oxidative phosphorylation of ATP, providing the body with a steady stream of energy (14, 15). The mtDNA is a direct target of ROS attack due to its proximity to ETC, the source of oxidants, lack of histones protective proteins and DNA repair capacity. Oxidative stress may ultimately lead to mutations or copy number alteration that could impair the mitochondrial function, triggering a vicious cycle of ROS generation and mtDNA oxidative damage (20–22). Many factors causing mitochondrial dysfunction are directly associated with a wide range of pathological conditions. Among these factors, quantitative changes including high and low mtDNA-cn have been frequently observed in human diseases such as cancer, neurodegenerative diseases and psychiatry conditions (23–25). One of the proposed mechanisms leading to high mtDNA-cn is augmented oxidative stress. It has been shown that the mtDNA is preferentially clonally amplified to make more mitochondria and mtDNA in response to energy deficient. The increase in mtDNA content has been suggested as the early molecular mechanism in response to oxidative stress in human cells to compensate for damaged mtDNA and defective mitochondria (20, 21). Therefore, high mtDNA-cn can be an indicative biomarker of oxidative stress, which is closely associated with mitochondrial dysfunction (17, 20–22). On the other hand, low mtDNA-cn has been linked with poor oxidative capacity, which leads to abnormal oxidative phosphorylation and defective cellular energy production (20–22).

In previous studies, both increased and decreased mtDNA-cn in blood and other tissues have been linked to neurodevelopmental disorders, suggesting a diverse and complicated role of altered mtDNA-cn in these disorders. In this study, we performed a meta-analysis to summarize the current evidence and to provide more conclusive results regarding the relationship between mtDNA-cn and ASD/ADHD. We first evaluated the relationship between mtDNA-cn and ASD/ADHD in an overall meta-analysis. The results showed a significant relationship between higher mtDNA-cn and ASD/ADHD. Although there was a large heterogeneity across the studies, further evaluation showed that the data did not significantly influence the pooled results. The variations in the range of mtDNA-cn could be explained by many factors. While real time qPCR is an accurate and reliable method for mtDNA-cn evaluation, quantification can be affected by the methodological variability in the process. For instance, the techniques used for total DNA extraction and different sets of primers used to amplify the mtDNA genes may affect the range of mtDNA-cn (52). Noticeably, some studies used multiple primer sets for different mtDNA genes (26–28, 37–41) and this can impact the precision of DNA quantification. However, mtDNA-cn could be determined using several mtDNA genes and the variations in primer selection should not affect the outcome of the meta-analysis. The wide variations of mtDNA-cn ratios between studies may be also due to the sample used for DNA extraction such as blood (26–28, 30, 31, 37, 39–41, 43), brain tissues (29, 38), and oral samples (42, 44). Even with the use of blood sample, variations in mtDNA-cn quantification may occur. For instance, platelet contamination during extraction of DNA from peripheral blood mononuclear cells (PBMCs) may lead to inaccurate estimation of mtDNA-cn (53).

The mtDNA-cn varies widely among tissues and can adapt to the energy demand and physiological condition of each individual (16, 17), we evaluated the relationship between mtDNA-cn and ASD/ADHD in stratified analysis by tissue types to reflect the biological differences between mtDNA-cn in the blood vs. non-blood tissues. The analysis of blood-based mtDNA revealed no significant association between mtDNA-cn in blood samples of patients compared to controls. Whereas, a significant relationship between higher mtDNA-cn was observed in non-blood samples of patients compared to controls. However, it is difficult to make a final conclusion on the observed increase in mtDNA-cn in non-blood samples due to the small number of studies conducted on brain and oral tissues.

In the current meta-analysis, blood was the most common tissue examined for mtDNA-cn changes in ASD and ADHD. To reduce the uncertainty induced by blood compositions as confounding factors, further analysis was performed by stratifying studies based on peripheral blood samples only. No significant association was found between higher mtDNA-cn in peripheral blood samples of ASD/ADHD patients compared to controls.

Studies have shown that changes in mtDNA-cn may be influenced by the age of individuals (18, 19). Thus further stratified analyses were conducted on studies reported differences in the mtDNA-cn in blood and non-blood tissues of patients with ASD/ADHD and controls within a range age between 2 and 13 years old. We found no significant difference in mtDNA-cn in blood samples between patients and controls, whereas a significant increase in mtDNA-cn was observed in other non-blood samples between patients and controls.

Blood-derived mtDNA-cn has been proposed as a potential biomarker for many diseases (23–25). However, different tissues have different mtDNA content (16, 17), and it remains unclear how blood mtDNA-cn can reflect disease status in other tissues. In some studies, blood mtDNA-cn was found to be associated with gene expression in other human tissues (54). Conversely, changes in mtDNA-cn in tissue samples were not reflected in blood samples (55). The mtDNA can be released by mitochondrion into the cytosol or bloodstream. The release of mtDNA into the cytosol involves DNA sensing receptors that trigger innate immune signaling (56, 57). Whereas the release of mtDNA into the bloodstream as cell-free mtDNA reflects physiological status (58). It has been shown that under psychological stress, cell-free mtDNA can be produced even in the absence of injury (59, 60), implying that the level and the biological role of cell-free mtDNA differ from tissue mtDNA-cn (61). Moreover, blood mtDNA-cn does not reflect tissue mtDNA-cn or capacity of mitochondrial energy production in other tissues (61).

Variations in mtDNA-cn measured in blood can be caused by differences in blood-cell compositions due to different mtDNA content in different blood-cell types. For instance, no consistent correlation was found between mtDNA-cn in blood leukocytes and cell-free mtDNA in plasma (62). In fact, blood composition is a strong confounder of blood-based mtDNA-cn measurement. Determinant factors of blood-derived mtDNA-cn include cellular heterogenicity of leukocytes and platelet abundance, which can change with physiological and pathological conditions (61). Leukocyte and platelet counts are therefore important sources of variation in whole blood mtDNA-cn measurement (63). Particularly, with the objective of measuring mtDNA-cn in blood to develop a biomarker, precise adjustment for the abundance of the subtypes of blood cells is important to rule out the potential confounders and to improve the specificity of mtDNA-cn as a stable biomarker. More studies are required related to quantification of mtDNA-cn in homogenous or well-defined cell populations to mitigate the possibility of confounding and to elucidate the influence of blood-cell type variations on mtDNA-cn. It is worth noting that nuclear genes involved in mtDNA replication and/or maintenance as well as common genetic variants play important roles in regulating the mtDNA-cn (64, 65). Therefore, the mtDNA-cn is suggested as a complex biomarker reflecting specific mitochondrial function, mostly related to mtDNA regulation which is also under nuclear genetic control (65, 66).

ASD and ADHD are the two most common neurodevelopmental disorders observed in childhood. These disorders affect the brain function and neurological development, causing impairments in cognition, communication, behavior, and motor functioning (1, 2). The brain is a highly energy-consuming organ and approximately 96% of this energy is used by neurons (67). In the developing and mature brain, mitochondria are centre for several pathways crucial for neural development, survival, activity, and connectivity (67). In neurons, mitochondria are also involved in calcium buffering and generation of ROS (68). Therefore, high-energy demand tissues such as neurons are often the most strongly affected by mitochondrial dysfunction.

Neurons are extremely sensitive to free radicals (69, 70). Increased oxidative stress and disrupted intracellular redox status, which contribute to mitochondrial neuronal dysfunction have been previously reported in ASD (71, 72) and ADHD (13). In addition, there is a direct links between increased mtDNA-cn and enhanced oxidative stress in ASD patients (37, 39, 42).

Impaired mitochondrial function can alter brain energy metabolism and cause neurological disorders. The increase in mtDNA-cn may be a compensatory adaptive response to overcome oxidatively defective mitochondria in order to correct for energy deficit (20, 21). Considering these observations, elevated mtDNA-cn in blood and or tissues of patients with ASD and ADHD may be a consequence of dysfunctional state in mitochondria.

Mitochondrial dysfunction can be classified as either primary or secondary defects. Primary mitochondrial dysfunction are caused by defects in mtDNA and/or nuclear DNA (nDNA) genes directly involved in mitochondrial function and ATP producing (73). Secondary mitochondrial dysfunction on the other hand are due to other metabolic and genetic abnormalities or environmental factors that impair the ability of mitochondria to produce ATP (73). The pathophysiology of mitochondrial dysfunction in ASD is very complex and can be caused by defects in mtDNA and/or nDNA genes associated with mitochondrial function as well as other mitochondrial aberrations. A meta-analysis conducted by Rossignol and Frye (74) showed a higher prevalence of mtDNA deletions in patients with ASD than in the general population. Their study also suggested that primary mitochondrial disease may occur in a subgroup of patients with ASD (74). In addition, children with mitochondrial disease or abnormal mitochondrial function may be more vulnerable to external factors causing regression in developmental milestones in ASD including language, motor skills, eye contact, play skills, social interaction and receptive skills (74). Whereas abnormalities in several biochemical markers of mitochondrial dysfunction such as elevated lactate and lactate-to-pyruvate ratio and/or toxic environmental exposures can result in secondary mitochondrial dysfunction in children with ASD (74). Another study by Varga et al. (75) showed that mtDNA alterations including mtDNA deletions are more common in patients with ASD than in control individuals, and coexist either with alterations in nuclear genes associated with ASD and genes essential for mtDNA maintenance or with environmental factors. The high prevalence of mtDNA deletions in ASD patients was suggested as a secondary effect (75). Similarly, mitochondrial dysfunction has been suggested as a crucial mechanism underlying the pathophysiology of ADHD. As such, several mtDNA defects including mtDNA mutations/polymorphisms and abnormal mitochondrial-associated protein biomarkers have been described in patients with ADHD (76). Moreover, mitochondrial defect in cybrid-neurons can significantly alter the serotonergic neurotransmitter function, which contributes to ADHD pathology and/or phenotypes (13). Given that ASD individuals with mitochondrial defects exhibit symptoms of ADHD (8, 9), mitochondrial dysfunction is probably a central, common themes in ASD and ADHD. However, the exact genetic etiology of mitochondrial dysfunction in the pathophysiology of these disorders remains uncertain.

ASD and ADHD share some clinical features (6), which may be due to shared genetic factors and other influences (8, 9). To further understand the contribution of mitochondrial dysfunction in these disorders, additional stratified analysis was conducted on studies reported differences in mtDNA-cn in the blood of aged-matched patients with ASD or ADHD and controls. No significant difference in mtDNA-cn was observed between ASD patients and controls, while a significant increase in mtDNA-cn was found in ADHD patients than in controls. It should be noted that the blood samples used to measure mtDNA-cn in ASD were peripheral blood, leukocytes, and PBMCs as compared to only peripheral blood samples in ADHD studies. This further supports the earlier observations that blood-cell compositions are important confounders of blood-based mtDNA-cn measurement (61–63).

Although brain tissues are considered as the standard target tissue in studying neurodevelopmental disorder and other brain disorders, blood is minimally invasive for mtDNA-cn measurement. Nevertheless, blood collection may cause patient’s discomfort and requires a standard level of training. Saliva is less invasive, accessible biofluid that can be collected by patients following simple instructions. Salivary mtDNA is also independent of the number of cells in saliva, thus it is not confounded by cellular content (77). Recently, a number of studies have highlighted the usefulness of salivary mtDNA-cn as a possible biomarker to monitor patients with head and neck cancers (78, 79), and to study dynamic neuroendocrine changes (77). Furthermore, saliva has been used to study cell-free mtDNA-cn in children and adults with stress and trauma (80). These observations suggest several advantages of salivary mtDNA-cn over blood and other invasive tissues.

Taken together, findings from studies investigating the pathogenesis of ASD and ADHD have suggested an important role of mtDNA-cn in these conditions, making mitochondria a potential target for effective therapeutic strategies (68). Improving the biological specificity of tissues used in mtDNA-cn quantification is important for the utility of mtDNA-cn as a promising biomarker in neurodevelopmental disorder.

Certain limitations in our study should be considered. First, we only included a small number of studies, only two publication measured mtDNA-cn in brain tissues and two publications measured mtDNA-cn in oral samples out of 14 included studies, and future analyses are warranted in larger studies. Second, lack of sufficient information about behavioral and clinicopathological characteristics of patients limited our further evaluation of these factors with mtDNA-cn. In addition, lack of randomized controlled trials reduced the reliability of the pooled outcome.

To our knowledge, this is the first meta-analysis to evaluate the relationship between mtDNA-cn and ASD/ADHD. Our overall analysis of mtDNA-cn in all blood samples and non-blood samples (brain tissues and oral samples) suggest elevated mtDNA-cn in patients than in controls, further emphasizing the implication of mitochondrial dysfunction in neurodevelopmental disorders. However, our results indicate that the mtDNA-cn in blood is not reflected in other tissues in ASD/ADHD, and the true relationship of blood derived mtDNA-cn remains to be elucidate in future studies. Our results also highlight the importance blood-cell compositions as confounder of blood-based mtDNA-cn measurement and the advantages of salivary mtDNA-cn over blood and other invasive tests. Finally, the potential of mtDNA-cn as a biomarker for mitochondrial malfunction in neurodevelopmental disorders deserves further investigations.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

GA-K: conceptualization, methodology, study design, retrieving, processing and analysing data, writing the original draft, and reviewing and editing the manuscript. HJ: conceptualization, methodology, study design, retrieving, processing and analysing data, and editing and reviewing the manuscript. YA, MA and MH: supporting processing and analysing data, and editing and reviewing the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.