Central Thyroid Hormone Deprivation Disrupts Cortical Cilia and Oligodendrocyte Lineage in an Allan–Herndon–Dudley Syndrome Mouse Model
Background: Allan–Herndon–Dudley syndrome (AHDS) is an X-linked neurodevelopmental disorder caused by loss of the thyroid hormone (TH) transporter MCT8, resulting in central TH deprivation and disrupted cortical maturation, cognition, and motor control. MCT8/OATP1C1 double-knockout (dKO) mice faithfully model the human disease, recapitulating its postnatal hypomyelination, neuromotor impairment, and cortical defects. Yet, cell-type–specific pathologies underlying AHDS remain insufficiently defined. Methods: To uncover cellular perturbations by TH deprivation, we performed single-nucleus RNA sequencing on cortex and attached cerebral nuclei from P21 WT and dKO mice. Differential gene expression, trajectory, pseudotime and gene-set enrichment analyses, and NeuronChat-based cell–cell communication modeling were integrated with LC-MS/MS-based TH quantification, immunofluorescence, and RNAscope. Results: In 48 clusters identified across cortical and striatal regions, we found increased numbers of GABAergic striatal D1 and D2 neurons in dKO mice, whereas mature oligodendrocytes were reduced. Trajectory analysis uncovered a bifurcation within the oligodendrocyte lineage, separating WT and dKO maturation paths and producing a dKO branch with gene profiles reminiscent of a stress-responsive, demyelination-prone state, despite largely preserved expression of core myelination genes. Trajectory analyses revealed shifted pseudotime states and distinct gene expression profiles in glutamatergic intratelencephalic and corticothalamic lineages of dKO mice. Differential gene expression patterns showed limited correspondence to Slc16a2 or Slco1c1 transcript levels but aligned strongly with published TH deprivation datasets, validating our findings and indicating that cellular perturbations are largely established by P21. Cell–cell communication analysis revealed a network imbalance favoring GABAergic over glutamatergic signaling, accompanied by altered neurexin–neuroligin interactions. In parallel, we identified a coordinated dysregulation of cilia-related genes, together with changes in cilia length and number. Conclusions: Our findings provide the first single-cell–level cortical map of AHDS brain pathology, revealing cilia defects, excitation–inhibition imbalance, differing pseudotime trajectories in glutamatergic neuronal populations and altered oligodendrocyte maturation, with actionable candidate genes such as Lama2, Litaf, and Dcc, as promising targets for future mechanistic and therapeutic exploration in AHDS. Slc16a2 and Slco1c1 transcript abundance alone did not predict cellular vulnerability, highlighting TH availability rather than transporter expression as key determinant of cell-type sensitivity and core mechanism for cortical network homeostasis. Graphical abstract