PREreview de A CK2α–G3BP1 signaling axis regulates local translation in developing neurons and is disrupted in OCNDS

The review is the result of a collaborative live review discussion held on October 20th, 2026, organized and hosted by the PhD school in Biomedical Sciences of the University of Padova as part of the Journal Club activity. The discussion was followed by additional asynchronous work by students and the JC committee, who jointly contributed to the composition and refinement of this report.

Background and aim

This manuscript investigates the contribution of the CK2α-G3BP1 axis in the pathogenesis of Okur-Chung neurodevelopmental syndrome (OCNDS). It asks whether disease-associated CK2α mutations disrupt downstream mechanisms required for neuronal development. The authors build on their previous evidence that CK2α phosphorylation of G3BP1 at serine 149 regulates G3BP1 granule dynamics and the release of axonal mRNAs for local translation. However, whether this mechanism is involved in neuronal development and is disrupted by OCNDS-associated variants had not previously been investigated. This is particularly relevant in the context of the OCNDS pathology since it is still unclear how mutation in CK2α, a pleiotropic kinase ubiquitously expressed across cells, results in neurodevelopmental phenotypes. The choice of the authors to focus on CK2α-G3BP1 pathway is particularly coherent as neurons are highly polarized cells and rely heavily on spatially and temporally regulated protein synthesis for correct function and development, therefore OCNDS-mutations of CSNK2A1 could affect G3BP1 granules dynamics and local translation, providing a crucial mechanistic link in the pathology. Finally, the study aims to identify a potentially targetable pathway and to test whether targeting G3BP1 rescues the disease-associated phenotypes.

Main findings

The research question is clear and this study provides the first evidence linking several OCNDS-associated CK2α variants to reduced G3BP1-S149 phosphorylation, altered G3BP1 granule dynamics and suppressed neuronal translation. The main results obtained by the authors are that OCNDS-associated CK2α variants reduce G3BP1 phosphorylation at S149 and are associated with the accumulation of larger and more persistent G3BP1-positive granules. These alterations modify the association of G3BP1 with selected neurodevelopmentally relevant mRNAs and are accompanied by reduced protein synthesis in neuronal compartments, as well as allele-specific defects in neuronal morphology, synapse number and network activity. In primary cortical neurons expressing the CK2α variants, G3BP1 knockdown can rescue several translational, morphological and synaptic phenotypes.

Major concerns

1. Choice of OCNDS variants. The choice to investigate multiple OCDNS variants shows that the observations are not restricted to a single allele and also highlights the phenotypic heterogeneity of OCNDS-associated mutations. However, the rationale for selecting these specific variants should be explained more clearly. K198R is described as the most prevalent OCNDS-associated variant, but it would be useful to know how the other variants were selected and if they result in different structural and biochemical properties of the kinase and in different clinical phenotypes. Moreover, the in vivo validation experiments focus exclusively on the K198R variant, and even there only a subset of readouts (pS149 and granule) is assessed. The authors should therefore discuss whether the same mechanism applies equally to all variants studied in the overexpression experiments, otherwise they should highlight this point as a limitation of the present study.

2. Ectopic expression of CK2α constructs. Much of the mechanistic work relies on ectopic expression of CK2α-WT or mutant constructs, so the authors should show that these proteins are expressed at comparable levels and have similar stability and subcellular localization, and discuss the contribution of endogenous CK2α. Otherwise, unequal expression, altered stability or mislocalization could produce apparent allele-specific effects, and overexpressing a kinase with many substrates may itself cause non-physiological phenotypes.

3. The conclusion that OCNDS-associated mutants produce “persistent, less dynamic SGs with defective SG cores” is broader than the supporting data. FRAP were measured only for K198R, only in COS-7 cells and only after arsenite stress. Either FRAP should be extended to other mutations, and ideally to neuronal axonal and dendritic granules, or dynamics claim should be explicitly restricted to K198R in COS-7 cells. In addition to that, the co-IP experiment data leave some uncertainty on the effect of CK2 mutations in reducing the interaction of G3BP1 with G3BP2 and CAPRIN1. Fig. 1E does not represent what is in Fig. 1F, according to the IP bands, the intensity of G3BP2 bands in CK2 mutants IP does not differ markedly from that of controls, especially if one considers the relative amount of G3BP2 in the input. Did the authors normalize the G3BP2 amount for the protein signal in the input lane? If not, how was the normalization performed? This information should be included in the Methods section.

4. Compartment-specific analysis. It is a major strength of the present work because it reveals allele- and compartment-specific effects that would be obscured by whole-cell measurements. However, this approach is not applied consistently across all experiments and model systems, and some findings, particularly the divergent allele-specific neuron phenotypes across different compartments, are not fully discussed. Although the authors discuss selected unexpected and compartment-specific results, a more systematic integration of these observations would strengthen the proposed mechanistic model.

5. Fig. 3 presents the quantification of immunolabeled phosphorylated and total G3BP1 in axon and dendrites. It is quite puzzling the fact that the phosphorylated G3BP1 signal is strong in GFP and CK2wt samples (panel A) but the very same samples are almost not labeled by the total protein antibody (panel E). Could the authors explain this evident inconsistency?

6. Interpretation of the CK2α-G3BP1 model and the G3BP1 knockdown rescue. Although the CK2α-G3BP1 axis is biologically well motivated, several aspects of the proposed model require further justification and discussion. The introduction establishes G3BP1 as a plausible neuronal CK2α substrate but does not explain why it should be a functionally dominant effector among the roughly 500 CK2 substrates, and the authors should discuss the physiological role of G3BP1 in developing neurons and possible compensation by G3BP2. In addition, G3BP1 knockdown increases protein synthesis and alters neuronal morphology even in GFP- and CK2αWT-expressing neurons, so these experiments support G3BP1 as a functional modifier of the mutant phenotypes rather than a selective or complete rescue of the disease mechanism, and they do not show that the rescue is specifically mediated through G3BP1-S149 phosphorylation. Experiments with phosphomimetic and phospho-deficient G3BP1 mutants would test this mechanism more directly. Given the broad functions of G3BP1 in stress responses, RNA metabolism and basal translational control, its therapeutic targeting may also have pleiotropic consequences, so the therapeutic implications should be presented more cautiously and the associated limitations explicitly discussed.

7. Terminology of G3BP1-positive granule. The terminology used to describe G3BP1-positive structures also requires clarification. In unstressed neurons, G3BP1-positive puncta may represent constitutive neuronal RNA granules rather than canonical stress granules, and the authors should define the molecular and biophysical criteria used for this classification.

Minor comments

● The human neuron experiments compare a single control line (CC1A) with a single patient line (SV0002578) that are not genetically matched, and no gene-corrected control is described, so the phenotypes could reflect genetic background or other individual-specific variability. Validation with an isogenic corrected line or additional independent lines is needed, or this limitation should be explicitly acknowledged and the conclusions qualified.

● The study is restricted to neurons, although glial cells also express CK2α, so the authors should discuss possible contributions of them.

● The inclusion of K198R patient-derived iNeurons is an important strength of the study. However, although these cells show increased G3BP1 granule number and size, G3BP1-S149 phosphorylation does not appear to have been directly measured in this model. The statement in the Discussion that reduced G3BP1-S149 phosphorylation is conserved across rat cortical neurons, the K198R knock-in mouse and patient-derived iNeurons is therefore not fully supported by the data presented. The authors should directly quantify G3BP1-S149 phosphorylation in the patient-derived iNeurons or revise this conclusion to state that only G3BP1 granule accumulation and morphological alterations were validated in the human model. In addition, the Methods refer to the quantification of G3BP1-S149 in iNeurons, whereas the corresponding description, antibody list and figures report only total G3BP1. This apparent inconsistency should be clarified.

● The figure panels and legends should consistently identify whether total G3BP1 or G3BP1-pS149 is being detected (Fig. 3). Where pS149 levels are quantified by immunofluorescence, the authors should clarify whether the signal is normalized to total G3BP1, as performed in the immunoblot analysis or only to the corresponding compartment marker. In addition, authors should add a control for the phospho-specific S149 antibody from Sigma-Aldrich (G8046) particularly because its phospho-specificity has previously been questioned [1].

● The definition of the controls (for example, why GFP was used as the control in each experiment) should likewise be stated clearly. In addition, the use of non-transfected cells in addition to GFP-expressing cells would have been an important additional control to include in the experimental setup.

● Regarding the compartment analysis, the authors should include markers to exclude possible contamination from other compartments in the axonal fractions used for the G3BP1-RNA co-IP, for example somatic and dendritic markers alongside axonal ones, in both the input and the immunoprecipitates. The soma is also missing from some experiments, such as the axonal RIP, which limits comparison across compartments and makes it harder to interpret the opposite changes seen in whole neurons and axons.

● The authors should state how neuronal density was controlled in the branching complexity and neurite-growth experiments (for example: equal plating density, per-neuron measurements or cell counts per field), particularly where results are given per image or per culture area, as well as in the G3BP1 knockdown experiments.

● The manuscript should clarify the apparent discrepancy between the increase in total synapse number observed with K198R (Fig. 2G-H) and the absence of a significant change in the individual numbers of PSD95 and synaptophysin puncta (Fig. S2A-B).

● As for the figures, the labels that indicate the different fluorescence channels (for example, “Nuclei (Blue)” and “G3BP1 (Green)”) are not shown in the colors of the channels they refer to. Printing each label in its corresponding color would make the figures easier to read and let readers identify the channels immediately, without having to match the text labels to the images. In addition, large brackets marking non-significant comparisons make the figures hard to read without adding information, so it would be preferable to show only the significant or most informative comparisons and report the rest in a table or legend. Also, the number of biological replicates, animals, independent cultures, cells, images and neurites should be reported clearly for every experiment, and the statistical unit (cell, neurite, granule, puncta, electrode or animal) should be defined, since many analyses appear to treat granules, neurites or electrodes as independent samples rather than biological replicates.

● G3BP1 binds fewer of the target mRNAs in whole neurons for R47G and K198R but more of them in axons, which needs further analysis and discussion.

● Fig. 1 B reports the quantification of cells displaying SGs as a heat map, which might be not the most appropriate to represent this kind of data. A histogram would be preferred.

● in Fig. S1 we suggest to indicate the statistical significance of each condition with respect to its own control rather than referring to the same not treated sample for all the samples.

● Movie 1 is cited in the Supplementary Materials, but it appears to be missing. Could the authors please provide it?

● Why do the authors decide to use the Dunn statistical test after Friedman analysis? Usually the Wilcoxon test is preferred after Friedman while the Dunn one is used after Kruskal-Wallis. Can the authors justify their choice?

● The raster plots of Fig. S2D are not representative of the data presented in the graph of panel G and H of the same Figure, please select different images that better correlate with the global result data.

Overall evaluation

Overall, this manuscript provides a relevant and methodologically solid investigation into the role of the CK2α-G3BP1 axis in OCNDS, supported by complementary models and compartment-specific analyses. Addressing the points outlined above would further strengthen the mechanistic and translational impact of the study.

References

1. Panas, M. D., Kedersha, N., Schulte, T., Branca, R. M., Ivanov, P., & Anderson, P. (2019). Phosphorylation of G3BP1-S149 does not influence stress granule assembly. The Journal of cell biology, 218(7), 2425.

Competing interests

The authors declare that they have no competing interests.

Use of Artificial Intelligence (AI)

The authors declare that they did not use generative AI to come up with new ideas for their review.