Spectrums of human and molecular complexity: mitochondrial dynamics in autism aetiology

The autism spectrum: an introduction
The autism spectrum lies under the umbrella of neurodiversity, which describes a range of conditions that deviate from what we call "neurotypical". Autism Spectrum Disorder (ASD) is defined in the Diagnostic and Statistical Manual of Mental Disorders (DSM-V) as a neurodevelopmental condition marked by differences in social communication, cognition, and behaviour. Historically, autism was considered a linear spectrum defined by different levels of functioning. However, ASD is now understood as a multidimensional spectrum of different support needs. This spectrum covers a vast range of presentations and comorbidities that look very different across genders, life-stages and sociocultural contexts.

Reconceptualizing the Autism Spectrum (Image from The Art of Autism)
Molecular Autism Research
The complexity of the autism spectrum is underpinned by a molecular aetiology that is equally multifaceted. As well as a heterogeneous genetic architecture, ASD is associated with interacting epigenetic and environmental factors that contribute to different presentations. Molecular autism research explores these underlying mechanisms to better understand the interactions between social and biological factors that affect quality of life. But almost all genetic ASD research is conducted in populations from the global North. As we highlighted in a recent scoping review, expanding our knowledge of autism in understudied populations is an emerging priority for the field.
Differential Methylation in a South African Cohort
To address the absence of molecular data from African populations, our research began with an exploratory study in a South African autism cohort. We examined epigenome-wide signatures of DNA methylation in children with ASD and controls, which was one of the first null-hypothesis based molecular studies in this population. We identified close to 900 differentially methylated genes and discovered that these genes converged on 9 canonical pathways involved in mitochondrial function.
.png)
Mitochondrial Dynamics in ASD: The Spectrum of Mitotypes
After validating the differential methylation of central mitochondrial genes using two independent techniques, we explored the mitochondrial phenotype in our South African cohort more closely. To do this, we had to consider the dynamic mechanisms that govern mitochondrial function and homeostasis. This is because mitochondria do not exist on a linear continuum ranging from function to dysfunction. Mitochondria are dynamic organelles that are constantly shifting through different states – or “mitotypes” – as they adapt to their environment.
Mitotypes are derived from differences in mitochondrial DNA sequences, metabolic state, dynamics, transport, ultrastructure, networks or substrate dependency. Each mitochondria in any given cell might have a slightly different mitotype. And mitochondria do not exist in isolation – they actually form integrated communities or networks in each cell.
The health of this network is maintained by the production of new mitochondria (called mitogenesis), the destruction of damaged mitochondria (called mitophagy) and the fission or fusion of individual mitochondria within the mitochondrial network. The balance between these processes is called mitostasis, which functions as a critical physiological regulator.
.png)
In our next study, we examined some of the processes that may regulate mitostasis in our cohort. We identified six transcriptional regulators of mitogenesis, fission, and fusion that were significantly differentially methylated in our cohort. In particular, there were several differentially methylated sites in the promoter of a gene called PPARGC1A. PPARGC1A is a transcriptional "master regulator" that is not only essential for the control of mitogenesis, but is broadly implicated as a central gene in the convergent mechanisms that govern mitostasis.
To determine whether these altered methylation patterns induced functional changes in our cohort, we examined three endpoints of mitochondrial function: mitochondrial DNA (mtDNA) copy number, mtDNA deletions, and urinary metabolomics. We found significantly elevated mtDNA copy number, which correlated with both mtDNA deletions and PPARGC1A hypermethylation in ASD. We also found significant changes to urinary metabolites associated with oxidative stress. Together, these signatures indicate disruptions to mitochondrial and redox homeostasis underpinned by differential DNA methylation in our cohort.
.png)
From signatures to mechanisms: an in vitro investigation
It turns out that mitochondrial dysfunction is widely observed in clinical studies of ASD, but the mechanistic relationship remains poorly understood. More recently, mitochondria have emerged as central regulators of neuronal maturation, physiology, and function. In vitro studies have begun to characterize the relationship between mitochondrial metabolism and neurodevelopment, showing that disruptions to mitochondrial energy production have profound consequences for neuronal differentiation, branching, and connectivity. Beyond the context of neurodevelopment, mitochondria play a crucial role in neuroimmune homeostasis, neurotransmission, neuropsychiatric disorders, and neurodegeneration. So, characterizing the mitochondrial mechanisms involved in ASD could help us to understand common psychological and physiological comorbidities across the spectrum.
To explore the mechanistic relationship between mitochondria and neurodevelopment, we established an in vitro system to model the mitochondrial dysfunction observed in our cohort. One of our most significantly differentially methylated genes was PCCB, a mitochondrial enzyme involved in TCA cycle metabolism. Since genetic deletions in PCCB lead to the toxic accumulation of a metabolite called propionic acid (PPA), we used PPA to model PCCB dysfunction in neuronal-like SH-SY5Y cells.

Interestingly, PPA is widely used to model ASD-associated behaviours in vivo, and is well-characterized in vitro. PPA is known to impair TCA cycle metabolism, reduce mitochondrial membrane potential, and induce mitophagy in SH-SY5Y cells. However, the effect of PPA on mitochondrial dynamics is still not well understood.And crucially, mitochondrial morphology and dynamics are increasingly recognized to play a regulatory role in neurodevelopment.
Such dynamic processes are hard to study because both the regulators and end points of mitostasis are constantly in flux. To circumvent this challenge, we used complementary visualisation techniques to characterize end-point mitochondrial morphology and then capture and quantify dynamic mitochondrial remodeling.
First, high-resolution transmission electron microscopy showed that mitochondria became smaller and more circular with poorly defined cristae under PPA stress. Subsequent time-lapse fluorescent microscopy showed that the mitochondrial network had retained its integrity by increasing both fission and fusion events under stress. At the same time, PPA disrupted several key transcriptional regulators of mitochondrial metabolism, fusion, and biogenesis – including several of those found to be differentially regulated in our ASD cohort.
.png)
This showed that we had recapitulated some aspects of our ASD phenotype in vitro. Futhermore, our in vitro system allowed us to observe the remodeling of mitochondrial morphology and dynamics in real-time, providing some insight into downstream mechanisms that are implicated in ASD.
This model is now poised to better characterize the underlying mechanisms that mediate the PPA-induced changes to mitostasis and examine how these disruptions influence the fate of neuronal stem cells during development. Increasingly, mitochondrial dynamics are implicated in the relationship between psychosocial stress and clinical outcomes across a range of neuropsychiatric conditions, including ASD. Ultimately, this avenue of research could help us to unravel molecular mechanisms that profoundly affect well-being and quality of life in the context of autism - and beyond.
Contact: colleen.oryan@uct.ac.za
Read more about our research on our website.
Follow the Topic
-
Scientific Reports
An open access journal publishing original research from across all areas of the natural sciences, psychology, medicine and engineering.
Related Collections
With collections, you can get published faster and increase your visibility.
Artificial intelligence and medical imaging
Publishing Model: Open Access
Deadline: May 01, 2025
Additive manufacturing
Publishing Model: Open Access
Deadline: May 15, 2025
Please sign in or register for FREE
If you are a registered user on Research Communities by Springer Nature, please sign in