I would like to bring to the attention of the authors, which are apparently not aware of it, that a large part of their result is in contradiction to previous literature. This is notably the section related to thyroid hormone receptor alpha (Thra). Specifically, the authors claim that the elimination of this receptor in the GABAergic lineage of Nkx2.1-Thrafl/+ mice results in an increased density in interneurons (PV neurons) in the hippocampus, cortex, and striatum. There are several problems with this claim: 1) If there is no typo, Nkx2.1-Thrafl / + mice keep an intact copy of the Thra gene. So there is no elimination of the receptor in these heterozygous mice. In decades of previous studies, the elimination of a single copy of the Thra gene did not appear to have a clear consequence on the mouse phenotype. Therefore, the phenotype shown in figure 4 is very surprising. 2) Most importantly, previous studies used a knock-in mutation (ThraAMI), which has a much stronger adverse effect on thyroid hormone signaling than a KO mutation. ThraAMI/+ mice were crossed with Gad2Cre or Gad2Cre-ERT2, which are also drivers for Cre/loxP recombination in GABAergic neurons. The result was a DECREASE in PV neurons density (Richard et al. iScience. 2020 Mar 27;23(3):100899. Ren et al. Thyroid. 2025 Dec;35(12):1443-1452). The discrepancy cannot only be due to the choice of driver transgene as Nkx2.1 and Gad2 expression pattern in cortex are similar. 3) Accordingly, all previous models, hypothyroid rat, mice, humans, KO transporters, etc. converge to indicate that a defect in thyroid hormone signaling, at least in the cortex and hippocampus, causes a reduction in PV neurons density, not an increase (review in ref. 30 of the article). This defect appears when signaling defect occurs during early life (Ren et al. iScience 29, Issue 6115970 June 19, 2026).

I believe that it would be the best interest of the authors to address these paradoxes before concluding that the ASXL3 / + mouse phenotype is explained by increased expression of Dio3, causing depletion of thyroid hormone in the brain.

Here are a few suggestions to solve the paradox: To my knowledge, Thrafl/+ mice (Cat# S-CKO-17774 from Cyagen) have not been used in previous studies. It would be interesting to perform germline recombination and compare the phenotype with other Thra KO mice, which have been extensively used in the past. Nkx2.1-Thrafl /+ mice have not been sufficiently characterized. What happens to gene expression in the cortex or striatum? (see citations above for genes sensitive to thyroid hormone deficiency in PV neurons). Additional controls should be helpful. For example Nkx2.1-Thra+/+ or Thrafl/+, all treated with tamoxifen, could also be used as negative controls to ensure that the phenotype is not a side effect of tamoxifen treatment or transgene expression.