3. Altered hippocampal glia-neuron interactions in ASD model animals
Multiple lines of evidence have demonstrated the essential roles of glial cells, particularly astrocytes and microglia, in maintaining the physiological function of the hippocampus \cite{RN63,RN64,RN65,RN66}. Astrocytes, the most abundant cell type in the human brain, coordinate synapse formation and elimination, neuronal survival, and axon guidance in the developing brain \cite{RN67,RN68,RN69,RN70}. Moreover, astrocytes form tripartite synapse, actively regulating synaptic transmission by interacting with presynaptic axon terminal and postsynaptic dendritic spines of neurons \cite{RN71,RN70}. On the other hand, microglia, as the dominant immune cells in the brain, possess phagocytic capacity, enabling them to engulf dead cells and detrimental substances in the brain \cite{RN72}. They also participate in synaptic pruning by engulfing excessive synapses through interactions with neuron \cite{RN66,RN74,RN73}. These glial cells work in concert to maintain brain homeostasis \cite{RN75,RN70}. Consequently, morphological and functional alterations of glia have been widely observed in many brain disorders, including neurodevelopmental disorders such as ASD \cite{RN76} and dysfunction in glial cells has been recognized as an etiological factor in their pathogenesis \cite{RN76}.
Due to the highly heterogeneous nature of ASD, many different transgenic mice modelling the disorder have been developed, and how hippocampal glia are altered substantially varies depending on the specific model. Nonetheless, most previous studies have commonly reported that, regardless of the model, astrocytes and microglia exhibit morphological, molecular, and/or functional alterations in the hippocampus.
In the Fmr1-KO mouse model multiple studies have reported astrocytic changes \cite{RN46,RN78,RN77}. During early postnatal development, dynamic alterations in the expression of Hevin, a protein secreted at excitatory synapses, are observed, suggesting the involvement of astrocytic alterations in abnormal synaptic development in FXS \cite{RN77}. In 2-month-old Fmr1-KO mice, astrocytes and microglia were reported to show reduced function at tripartite synapse and synaptic pruning respectively \cite{RN46}, while astrocytes showed increased GFAP expression in 3-month-old mice \cite{RN78}. Another notable study has developed astrocyte-specific Fmr1 conditional KO mice and observed that astrocytic GABA synthesis increased alongside a modest elevation in GFAP expression \cite{RN79}, demonstrating a positive correlation between astrocytic GABA synthesis and reactivity \cite{RN82,RN81,RN80}.
In the Mecp2-KO mouse model astrocytes undergo cytoskeletal atrophy during severe symptomatic time point, but not at earlier time points \cite{RN83}. The reduced ramification of astrocytes was also observed in postnatal Mecp2-conditional KO mice \cite{RN84}. Microglia in these mice also exhibit reduced branch complexity in the late-phenotypic stage, whereas no such changes occur during the pre-phenotypic period \cite{RN85}. Furthermore, CA1 astrocytes in the same mouse model have been implicated in the reduction of tonic inhibition due to decreased expression of the GABA transporter 3 (GAT3) within astrocytes \cite{RN86}, a phenotype linked to the hyperexcitability and increased seizure susceptibility in these mice. A similar finding was also observed in the astrocyte-specific Mecp2-KO mice \cite{RN86}. Collectively, these studies suggest that hippocampal glia are significantly impacted by, as well as contribute to, the pathology of Rett syndrome.
At the level of synaptic proteins, mice with C-terminal deleted Shank3 (deletion of exon 21, which includes the Homer- and Cortactin-binding domains; Shank3+/ΔC) exhibit morphological changes in astrocytes and microglia within the hippocampus, with no observable alterations in the number or size of astrocytes (GFAP+ and S100+) in most hippocampal subregions. However, there was a noteworthy reduction in the GFAP+ area specifically within the CA1 stratum radiatum. In contrast, microglia exhibited no discernible changes in their morphological characteristics \cite{RN87}. On the other hand, a separate transcriptomics study conducted with Shank3-KO mice revealed an increase in the expression of gene sets related to astrocytes, microglia, and oligodendrocytes \cite{RN88}. However, Shank2-KO (lacking exon 6 and 7) mice showed that astrocytic and microglial genes are negatively enriched \cite{RN88}. These findings implicate the Shank2 and Shank3 deletions lead to differential transcriptomic changes of glial cells in the hippocampus.
A recent study, utilizing NLG4-KO mice, revealed sex-dependent morphological and functional alterations in microglia \cite{RN89}. These microglial changes were more pronounced in males than in females. Specifically, male Nlgn4-KO mice, aged 13 weeks and 20 weeks, exhibited reduced microglial density and branching, diminished phagocytic activity, decreased expression of MHC1 and CD54, impaired response to injury, and disrupted energy metabolism in the CA3 region of the hippocampus. In contrast, in the NGL3R451C point-mutant knock-in mouse model, DG microglia showed no discernible morphological alterations, while DG astrocytes showed a shrunken morphology \cite{RN90}.
In Cntnap2-KO mice there were only subtle morphological changes observed in astrocytes in the DG molecular layer and CA1 stratum radiatum, with no alterations in the numbers or area of GFAP+ cells. Notably, in the ventral hippocampus, there was a significant reduction in the number of S100+ astrocytes, although the area of S100+ pixels remained unaltered. Similarly, microglia exhibited minimal alterations with no change in the numbers or area of Iba1+ cells. There was a slight, non-significant increase in the CD68+ area observed in the dorsal molecular layer and ventral stratum radiatum, but this was not observed in other hippocampal subregions \cite{RN87}. Intriguingly, there was a report with a mouse model that underwent a plasma exchange operation using plasma from from two male patients with CASPR2 (contactin associated protein 2, encoded by Cntnap2 gene)-positive encephalitis. In this mouse model, regardless of their age, hippocampal microglia showed no morphological alterations, while there was a significant increase in microglial numbers in the cortex \cite{RN91}. The modest change or the absence of the alterations in hippocampal glial morphology in the Cntnap2-KO model may distinguish it from many other ASD-like mouse models.
In SCN1A haplodeficient (-/+) mice, a Dravet syndrome model, one recent study reported increased GFAP expression and an increased number of Iba1-positive microglia in DG, implying increased reactivity of astrocytes and microglia \cite{RN92}. Additionally, this study reported a reduction in tonic GABA current in CA1 pyramidal neurons, which needs further investigation due to its inconsistency with existing evidence showing the positive correlation between astrocytic reactivity and tonic GABA current \cite{RN92}. In SCN2A-deficient mice, another study reported partially activated microglia, evidenced by increased cell bodies and reduced branches \cite{RN93}. These activated microglia showed excessive phagocytic pruning of synapses, which occurs during development and continues into adulthood. These changes in microglia led to a reduction in spine density and glutamatergic synaptic transmission in CA1 pyramidal neurons \cite{RN93}. Collectively, these studies suggest that glial alterations in both SCN1A and SCN2A deficiency-induced ASD-like mouse models could impact E/I balance in the hippocampus.