GLUT1-DS Brain Organoids Exhibit Increased Sensitivity to Metabolic and Pharmacological Induction of Epileptiform Activity

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

• Corrected isogenic control line is now available at the Coriell Institute for the GLUT1-DS patient line. The use of 1 wildtype line is not appropriate to make meaningful comparisons.
• It is not clear how this paper advances the study published by this group in Frontiers in Neuroscience, Muller et al 2024.
• The levels of glucose used do not reflect physiological levels, both are high. The authors argue that 5uM is low due to the diffusion of glucose into the center of the organoid. But, the majority of organoids are composed of neural cell types on the outside, not in the center. The authors should determine the concentration at which control organoids exhibit responses to low glucose. 
• To make conclusions about the impact of GABAergic inhibition the authors should demonstrate whether wildtype and GLUT1-DS organoids have similar proportions of GABAergic neurons. 

Author Response

Comment 1: Corrected isogenic control line is now available at the Coriell Institute for the GLUT1-DS patient line. The use of 1 wildtype line is not appropriate to make meaningful comparisons.

Response 1: We agree with the reviewer that the use of isogenic lines would provide a more direct comparison and that relying on cell lines from different genetic backgrounds represents a limitation of this study. It is indeed our intent to expand the platform capabilities by (i) comparing mutant cell lines with their corresponding isogenic controls and, most importantly, (ii) including additional iPSC lines from healthy donors and GLUT1-DS patients to account for biological variability, which could allow for personalized phenotyping. In In this context, the robust difference observed between organoids from a healthy donor and a GLUT1-DS patient in terms of epileptiform activity—highly relevant to the pathology—represents a meaningful and appropriate starting point. This limitation and these considerations have been acknowledged in the Discussion (l.401-404)

 

Comment 2: It is not clear how this paper advances the study published by this group in Frontiers in Neuroscience, Muller et al 2024.

Response 2: We believe that this study advances our previous work in several important aspects. First, it highlights the heightened sensitivity of GLUT1-DS brain organoids to epileptiform activity induced by distinct stressors, including reduced glucose availability, PTZ, and KCl treatments, which was not addressed in our previous publication. This finding is biologically relevant, as it provides insight into the increased susceptibility of GLUT1-DS networks to epileptiform triggers.

Second, this study introduces a new analytical pipeline that enables standardized, automated quantification of organoid electrophysiological activity, which we have made available to the scientific community.

 

Comment 3: The levels of glucose used do not reflect physiological levels, both are high. The authors argue that 5uM is low due to the diffusion of glucose into the center of the organoid. But, the majority of organoids are composed of neural cell types on the outside, not in the center. The authors should determine the concentration at which control organoids exhibit responses to low glucose. 

Response 3. The use of supraphysiological glucose concentrations in brain organoids and in vitro models more generally is widely adopted and accepted, although we agree that this represents a functional difference compared with in vivo conditions. The brain organoid cultures used in this study are largely based on neural stem cell induction and differentiation protocols derived from iPSCs, as well as brain organoid protocols originally published by the Pasca group (Pașca et al., Nat Methods, 2015; Sloan et al., Nat Protoc, 2018), which employ media containing 25 mM glucose (standard Neurobasal medium) to 17.5–25 mM glucose (DMEM/F-12 medium). These protocols have also been used and adapted by collaborators from Dr. Stoppini’s group, with whom we developed our organoid platform (Govindan et al., Front Bioeng Biotechnol, 2021; Cosset et al., JoVE, 2019; El Harare et al., Cells, 2025), using 25 mM glucose-containing Neurobasal medium, X-VIVO 15 mixed with Neurobasal (1:1, final 12.5 mM glucose), and Neurobasal-based neuronal media containing 25 mM glucose.

Similarly, astrocyte and neuronal monolayer cultures derived from iPSCs are typically maintained in DMEM/F-12 (17.5mM glucose) or Neurobasal (25mM glucose) media. Mouse primary astrocytes and neurons are also commonly cultured under comparable glucose concentrations (e.g., DMEM with 25 mM glucose for astrocytes and Neurobasal medium for neurons).

This latter consideration aligns with the reviewer’s remark that the requirement for supraphysiological glucose concentrations in vitro is not necessarily related to limited glucose penetration into the organoid core. We also agree that MEA measurements primarily reflect activity near the surface of the organoid, although deeper tissue layers may still influence overall network connectivity. Accordingly, we have removed the speculative interpretation regarding glucose diffusion from the Results section (l. 131-132) and have explicitly highlighted this issue as a general limitation of in vitro models (l. 133-135).

 

Comment 4: To make conclusions about the impact of GABAergic inhibition the authors should demonstrate whether wildtype and GLUT1-DS organoids have similar proportions of GABAergic neurons. 

Response 4: The characterization of brain organoids in terms of glutamatergic and GABAergic neuronal markers by qPCR was previously reported in Müller et al. (Figure 3). A description of this data has been added to the text (l. 137-139), in addition to where it was already highlighted (l. 364-365).

Reviewer 2 Report

Comments and Suggestions for Authors

The authors developed a unified Python-based workflow to analyze electrical activity recorded from brain organoids using microelectrode arrays (MEAs), enabling standardized and reproducible analysis of neuronal activity. Using this system, they assessed the activity of a control organoid (from a healthy donor) and an organoid derived from a patient with Glucose Transporter 1 Deficiency Syndrome (GLUT1-DS). The organoids were exposed to media with high or low glucose concentrations, and the authors observed that glucose deprivation exacerbated hyperexcitability in GLUT1-DS organoids, while it did not alter the activity of the healthy organoid.

The organoids were then exposed to PTZ, a compound that blocks GABA receptors and induces epileptiform activity. The organoid derived from the GLUT1-DS patient was more sensitive to GABAergic perturbations, indicating increased epileptic susceptibility. Finally, the organoids were exposed to KCl to globally increase excitability. Once again, the GLUT1-DS organoid showed a stronger response to excitatory stimulation, revealing network hyperexcitability. To conclude, the organoids were treated with TTX, a sodium channel blocker, to confirm the neuronal origin of the recorded signals. In summary, GLUT1-DS is a disorder caused by impaired glucose transport in the brain. This study demonstrates that the organoid model recapitulates the functional signatures of the disease, particularly its epileptic features.

In general, the experiments are clearly presented and the results are quite convincing, even if the organoids produced, particularly GLUT1-DS, deserve to be better characterized, particularly in terms of cell type and proportion.

Author Response

Comment 1: The authors developed a unified Python-based workflow to analyze electrical activity recorded from brain organoids using microelectrode arrays (MEAs), enabling standardized and reproducible analysis of neuronal activity. Using this system, they assessed the activity of a control organoid (from a healthy donor) and an organoid derived from a patient with Glucose Transporter 1 Deficiency Syndrome (GLUT1-DS). The organoids were exposed to media with high or low glucose concentrations, and the authors observed that glucose deprivation exacerbated hyperexcitability in GLUT1-DS organoids, while it did not alter the activity of the healthy organoid.

The organoids were then exposed to PTZ, a compound that blocks GABA receptors and induces epileptiform activity. The organoid derived from the GLUT1-DS patient was more sensitive to GABAergic perturbations, indicating increased epileptic susceptibility. Finally, the organoids were exposed to KCl to globally increase excitability. Once again, the GLUT1-DS organoid showed a stronger response to excitatory stimulation, revealing network hyperexcitability. To conclude, the organoids were treated with TTX, a sodium channel blocker, to confirm the neuronal origin of the recorded signals. In summary, GLUT1-DS is a disorder caused by impaired glucose transport in the brain. This study demonstrates that the organoid model recapitulates the functional signatures of the disease, particularly its epileptic features.

In general, the experiments are clearly presented and the results are quite convincing, even if the organoids produced, particularly GLUT1-DS, deserve to be better characterized, particularly in terms of cell type and proportion.

Response 1: We thank the reviewer for their positive assessment. The characterization of the organoids in terms of MAP2- and GFAP-expressing cells by immunostaining (Müller et al., Figure 2), as well as the gene expression analysis of markers for stem cells, neural stem cells, glutamatergic neurons, GABAergic neurons, astrocytes, and oligodendrocytes quantified by qPCR was previously published (Müller et al., Figure 3). We have now added a description of this information to the manuscript to improve clarity (ll. 137–139).

Reviewer 3 Report

Comments and Suggestions for Authors

In this work, the authors developed an in vitro model for a genetic type of neurodevelopmental disorder also characterized by seizures. Seizures can be related to a defect in glucose transportation in the brain. The authors challenged their model by using two different glucose concentrations and recorded the electrical activity in organoids using multi electrodes. The authors also used an original approach to data analysis. Although this work does not provide mechanistic information, it is useful to establish an alternative model to investigate the features of Glucose Transporter 1 Deficiency Syndrome. Some minor points:

• The use of terms as "healthy" and "epilepsy" or "seizure" does not seem to be appropriate in the context of an in vitro model, in which the notion of "healthy" state is at odds with the normal use applied to a person able to communicate such a condition. Similarly, epilepsy is a condition in which seizures occur spontaneously.
• Some statements reported in results could be supported by calculation of percentages (line 231, "substantially greater").  
• The change found in theta power is interesting and in line with that reported by Costa et al. (2020) in kainic acid-treated rats during the status epilepticus. Could the reported finding in organoids be related to a sustained epileptiform activity?

Author Response

In this work, the authors developed an in vitro model for a genetic type of neurodevelopmental disorder also characterized by seizures. Seizures can be related to a defect in glucose transportation in the brain. The authors challenged their model by using two different glucose concentrations and recorded the electrical activity in organoids using multi electrodes. The authors also used an original approach to data analysis. Although this work does not provide mechanistic information, it is useful to establish an alternative model to investigate the features of Glucose Transporter 1 Deficiency Syndrome. Some minor points:

Comment 1:

The use of terms as "healthy" and "epilepsy" or "seizure" does not seem to be appropriate in the context of an in vitro model, in which the notion of "healthy" state is at odds with the normal use applied to a person able to communicate such a condition. Similarly, epilepsy is a condition in which seizures occur spontaneously.

Response 1: We agree with the reviewer and thank them for this important remark. We acknowledge that the terms “epilepsy” and “seizure” are not strictly appropriate for describing hyperactivity observed in an in vitro model. Accordingly, we now refer to these events as “seizure-like” or “epileptiform activity” when describing organoid data, and we reserve the terms “seizure” and “epilepsy” for in vivo models or human pathology.

In addition, the term “healthy organoids” has been replaced throughout the manuscript with “organoids derived from healthy donors” or “organoids of healthy origin,” as appropriate. These changes have been implemented consistently across the text to more accurately reflect the nature of the in vitro model.

 

Comment 2: Some statements reported in results could be supported by calculation of percentages (line 231, "substantially greater").  

Response 2: A calculation of percentages has now been added to support this statement, as well as to the corresponding description in Figure 4 (line 230; lines 256–257).

Comment 3: The change found in theta power is interesting and in line with that reported by Costa et al. (2020) in kainic acid-treated rats during the status epilepticus. Could the reported finding in organoids be related to a sustained epileptiform activity?

Response 3: We thank the reviewer for highlighting this relevant study. Indeed, the reported changes in theta power represent an interesting parallel with these findings in kainic acid–treated rodents, where theta-band dynamics were associated with sustained epileptiform activity and seizure progression. This comparison has now been included in the Discussion (l. 372–377).