Further we examined the degradation of
Further, we examined the degradation of Cx43. Autophagy and the proteasome system have previously been implicated in Cx43 degradation (Falk et al., 2014). DEX inhibited communication between 2 types of active transport by promoting the degradation of Cx43 through autophagy in osteocytes (Gao et al., 2016). In our experiments, only MG132 (an inhibitor of the proteasome system) reversed CORT-induced downregulation of Cx43 in hippocampal astrocytes. However, both 3-MA (an inhibitor of autophagy) and MG132 inverted the level of Cx43 protein in prefrontal cortical astrocytes. In view of these results, we concluded that the proteasome system degraded Cx43 in both prefrontal cortical and hippocampal astrocytes, but autophagy participated in the degradation of Cx43 only in prefrontal cortical astrocytes. To explain the different degradation pathways of Cx43 in prefrontal cortical and hippocampal astrocytes in response to CORT treatment, we observed the morphology of Cx43-EGFP using immunofluorescence. Annular gap junctions significantly increased in prefrontal cortical astrocytes, but internal Cx43 mainly existed as small dots in hippocampal astrocytes. Annular gap junctions were double membrane vesicles containing gap junctions between two adjacent cells. Gap junctions in the annular vesicles have been reported to be the same width as the membrane (Falk et al., 2016). Previous studies reported that it was difficult for the host cell to digest annular gap junctions due to its complex structure. Recently, Carette et al. suggested that annular gap junctions can be degraded by autophagy and lysosomes (Carette et al., 2015). For these reasons, autophagy was induced in prefrontal cortical astrocytes but not in hippocampal astrocytes in response to CORT. In addition to Cx43 protein level, the function of gap junctions was regulated by Cx43 phosphorylation. Phosphorylation of Cx43 differently modulated the function of gap junctions depending on the site. In previous studies, it was shown that PKC-induced phosphorylation of Cx43 at S368 decreased GJIC (Lampe et al., 2000, Pahujaa et al., 2007, Solan and Lampe, 2014). In our experiment, phosphorylation of Cx43 at S368 was elevated in both prefrontal cortical and hippocampal astrocytes exposed to CORT. Consistent with this finding, the expression of PKC has been found to be significantly altered in adrenalectomized mice treated with CORT (Birt et al., 2001). Another study found that phosphorylation of Cx43 at S368 induced gap junction internalization and degradation (Cone et al., 2014). Taking this evidence together with our own results, we can postulate that phosphorylation of Cx43 at S368 contributed to the decrease of Cx43 in the membrane. Cx43/drebrin interaction has been found to stabilize gap junctions in the membrane, preventing the internalization and degradation of Cx43 (Butkevich et al., 2004, Majoul et al., 2007, Stout et al., 2004). Interestingly, CORT did not alter the interaction of Cx43/drebrin in hippocampal astrocytes, and even increased Cx43/drebrin interaction in prefrontal cortical astrocytes. Thus, CORT may trigger the internalization of Cx43 in astrocytes by modulating other proteins, such as ZO-1. Moreover, given the double membrane structure of annular gap junctions in prefrontal cortical astrocytes, we speculated that intensive Cx43/drebrin interaction increased the stability of Cx43 in the membrane, which may be beneficial to the formation of annular gap junctions. The interaction between Cx43 and ZO-1 was regulated by the Src kinase through tyrosine phosphorylation (Toyofuku et al., 2001). In response to GC, GR formed cytoplasm complex with HSP90 and Src kinase (Marchetti et al., 2003). Thus, Cx43/ZO-1 interaction may be regulated by the association of GR and Src kinase. Through the interaction with Cx43, ZO-1 has been found to deliver Cx43 from a lipid raft domain to gap junctional plaques, consequently enhancing gap junctional plaques in osteoblastic cells (Laing et al., 2005). According to this previous report, decreased ZO-1 may reduce the size of gap junctions. In addition, ZO-1 is a scaffolding protein that stabilizes Cx43 in the membrane. Inhibition of the interaction between Cx43 and ZO-1 has been found to result in the internalization of Cx43 (Palatinus et al., 2011). However, contrasting results have indicated that inhibition of the Cx43/ZO-1 interaction increased gap junction size (Rhett et al., 2011). Blockade of the interaction between ZO-1 and Cx43 has been found to lead to a significant increase in plaque size (Hunter et al., 2005). Moreover, the role of ZO-1 in targeting Cx43 gap junctions to the endocytic pathway (Segretain et al., 2004) may also explain this phenomenon. Other reports consistent with this result have indicated that ZO-1 appeared to be localized at the center of Cx43 annular gap junctions (Baker et al., 2008). Therefore, intensive Cx43/ZO-1 interaction may result in gap junction internalization in prefrontal cortical astrocytes. However, this is just a speculation, and should be investigated in future studies.