Moxibustion Improves Hypothalamus Aqp4 Polarization in APP/PS1 Mice: Evidence from Spatial Transcriptomics

Aquaporin-4 (AQP4) is highly polarized to perivascular astrocytic endfeet. Loss of AQP4 polarization is associated with many diseases. In Alzheimer's disease (AD), it is found that AQP4 loos its normal location and thus reduce the clearance of amyloid-ß plaques and Tau protein. Clinical and experimental studies show that moxibustion can improve the learning and memory abilities of AD. In order to explore whether moxibustion can affect the polarization of AQP4 around blood brain barrier (BBB), we used spatial transcriptomics (ST) to analyze the expression and polarization of Aqp4 in wild type mice, APP/PS1 mice and APP/PS1 mice intervened by moxibustion. The results showed that moxibustion improved the loss of abnormal polarization of AQP4 in APP/PS1 mice, especially in the hypothalamic BBB. Besides, there are other 31 genes with Aqp4 as the core have the similar depolarization in APP/PS1 mice, most of which are also membrane proteins. The majority of them have been reversed by moxibustion. At the same time, we employed the cerebrospinal fluid circulation gene set, which was found being on a higher level in the group of APP/PS1 mice with moxibustion treatment. Finally, in order to further explore its mechanism, we analyzed the mitochondrial respiratory chain complex enzymes closely related to energy metabolism, and found that moxibustion can significantly increase the expression of mitochondrial respiratory chain enzymes such as Cox6a2 in the hypothalamus, which could provide energy for mRNA transport. Our research shows that increasing the polarization of hypothalamic Aqp4 through mitochondrial energy supply may be an important target for moxibustion to improve APP/PS1 mice's cognitive impairment. Overall design: The WT mice were male C57BL/6J mice, while AD model mice were male APP/PS1 double transgenic mice. All the mice were screened by Morris water maze test after a week of adaptation. Then APP/PS1 mice were randomly divided into AD model group (APP/PS1) and moxibustion group (APP/PS1+MOX), C57BL/6J wild-type mice were used as control group (WT group). Spatial Transcriptomics were generated by deep sequencing, using Illumina NovaSeq 6000.