DEPDC5: Modeling of the Two-Hit Wonder in Cortical Organoids

Commentary

In 2013, two studies using linkage analysis and exome sequencing identified loss-of-function variants in DEPDC5 as a cause of familial focal epilepsy.^1,2 DEPDC5 encodes a core component of the GATOR1 complex, a negative regulator of the mechanistic target of rapamycin complex 1 (mTORC1) pathway that inhibits cell growth in the absence of nutrients.

Individuals with DEPDC5-associated focal epilepsy exhibited remarkable phenotypic heterogeneity within and across families. It was also unclear how a germline variant present in all cells causes focal seizures. These observations led to the hypothesis that a second genetic variant may arise during embryonic brain development that could explain the phenotypic variability and seizure localization observed in individuals with DEPDC5 epilepsy. This “two-hit” mechanism, first conceptualized in the cancer field, was later implicated in epilepsy following identification of second-hit variants in TSC1 and TSC2 with tuberous sclerosis complex, and the first hit is spontaneous or inherited loss-of-function of one allele. ³ Consistent with this hypothesis, second hit mutations were identified in brain epileptogenic tissue therapeutically resected from two individuals with a pathogenic germline DEPDC5 variant. These second hits were found on opposing alleles consistent with bi-allelic inactivation.^4,5

Neuropathological features associated with focal cortical dysplasia type II (FCDII), including dysmorphic neurons, are often found in resected brain tissue from individuals with DEPDC5 epilepsy. Neurons with bi-allelic inactivation of DEPDC5 have been shown to be enriched in dysmorphic neuron populations. ⁶ Rodent models have further demonstrated that spontaneous seizures are associated with bi-allelic loss of DEPDC5, whereas heterozygous loss leads to no or only modest increases in seizure susceptibility.^7,8

Despite these advances toward understanding DEPDC5 pathophysiology, there remain significant gaps in how cells with mono- and bi-allelic hits in DEPDC5 alleles interact in the developing human brain. To begin to address this, Maletic et al ⁹ established an elegant in vitro system that uses patient-derived human cortical organoids (hCOs) to model mono- and bi-allelic mosaicism in the developing human brain. They used human-induced pluripotent stem cells (hiPSCs) derived from an affected child harboring a heterozygous germline loss-of-function DEPDC5 variant, and an unaffected sibling without the loss of function variant. Using CRISPR-Cas9, they then introduced the same pathogenic allele into the unmutated (wildtype) DEPDC5 allele, along with a fluorescent reporter, in the affected child's cells. hiPSCs from the unaffected brother were used as the control. Genetically homogenous hCO were generated from the control cell line and heterozygous (DEPDC5^+/−) hiPSC lines. Mosaic organoids modeling two-hit mosaicism were generated by mixing of mono- (DEPDC5^+/−) and bi-allelic (DEPDC5^−/−) hiPSC lines to create mosaicism.

The authors demonstrate that their hCO model system recapitulated the known mechanism of action by showing that only bi-allelic loss resulted in constitutive mTOR activation under nutrient-deprived conditions. DEPDC5^–/– neurons in their model exhibited enlarged somas and neurofilament accumulation, also consistent with dysmorphic neurons found in the human condition. These findings provide additional validation of previously published work in human tissue and rodent models that bi-allelic, but not mono-allelic, inactivation drives both mTOR hyperactivation and morphological abnormalities. Furthermore, these results establish the novel organoid model system as a valuable research tool for in-depth studies of the molecular mechanisms of DEPDC5 epilepsy.

The authors employed longitudinal single-cell RNA sequencing (scRNA-seq) across three developmental stages to analyze cellular composition, gene expression, signaling pathways, morphology, and network activity. Both pure heterozygous and mosaic organoids showed premature differentiation as evidenced by an increased proportion of upper-layer excitatory neurons. Precocious differentiation was also supported by reduced neural rosette density in both the heterozygous and mosaic developing hCOs compared to control. Computational analyses evaluating the ratio of unspliced pre-mRNA to mature spliced mRNA of cell type marker levels as a way to predict the future state of individual cells suggests that DEPDC5 loss accelerates the transition from progenitors to upper-layer neurons. Differential expression analyses suggest that this may be driven by dysregulation of Notch and Wnt signaling pathways, although additional studies will be needed to establish if the transcriptional changes are the cause or effect of cellular composition changes observed.

The DEPDC5 mosaic hCO model also captured transcriptional changes in excitatory neurons that may drive the hyperexcitability phenotype. Specifically, differential expression analyses revealed that excitatory neurons from both heterozygous and mosaic mature organoids exhibited increased expression of epilepsy-related genes, including KCNQ3, SCN2A, and CACNA1A. Synaptic genes were also enriched among differentially expressed genes in mosaic organoids.

A strength of the established mosaic model is its capacity to distinguish between cell-autonomous from non-cell-autonomous effects. Applying scRNA-seq further allowed the authors to resolve cell-type-specific effects, showing that DEPDC5^−/− excitatory neurons exhibited transcriptional profiles consistent with cell-autonomous metabolic and translational dysregulation that may contribute to dysmorphic neuron formation and hyperexcitability.

Human cortical organoid models have emerged as a powerful platform for studying the cellular and molecular underpinnings of neurodevelopmental disorders. These models provide a valuable platform for studying very early human cortical differentiation, a developmental window that is difficult to assess in humans or rodent models. By establishing an organoid platform for the study of mosaic DEPDC5 epilepsy, the authors have created a foundation upon which deeper mechanistic questions can begin to be investigated. The authors provide early mechanistic insight into several key areas ripe for deeper characterization, including linking loss of DEPDC5 to specific cellular and molecular alterations that drive early brain development phenotypes.

Among the most significant weaknesses is that all work was performed in a single patient-derived cell line, raising concerns about reproducibility across independent patient lines and the generalizability of the results. Compounding this, the study employed limited sampling across developmental time points, restricting the resolution with which the temporal dynamics can be characterized and reducing power to detect disease-relevant changes. A further concern relates to the introduction of the second hit in hiPSCs which does not recapitulate the timing of second hits thought to occur during embryonic brain development. The study also attempted to link cell-type-specific transcriptional changes to the genetic variants present in the organoids. Unfortunately, the high variability of these signals makes it difficult to draw firm conclusions about the biological significance of these findings.

It is also important to acknowledge broader constraints inherent to cortical organoid systems. The hCOs differentiation procedure used in this study lacks the full complement of cell types found in the mature human brain, notably oligodendrocytes, and does not recapitulate the circuitry that is critical to human brain development. While advances in differentiation protocols that allow for earlier development of oligodendrocytes and assembloid systems that fuse multiple brain regions offer partial solutions, organoid models will never fully replicate the complexity of an intact human or rodent brain.

While the study carries methodological limitations commonly associated with early-stage organoid studies, this does not diminish its core achievement. The hCO platform faithfully recapitulates key phenotypes associated with DEPDC5 epilepsy that are observed in patients. This model provides novel early mechanistic insight in DEPDC5 pathophysiology that can be expanded on in future work to drive forward the vision of precision therapy for genetic epilepsies. Moreover, the experimental framework developed here can serve as a template for modeling other mosaic brain diseases using hCOs.