The sequential behavioral approach used in
The sequential behavioral approach used in the present study reinforced the involvement of the BDNF/TRKB system in the effect of losartan. In mice with dampened BDNF expression, losartan was no longer able to exert antidepressant-like effects. Similar to what was observed after losartan treatment, these mice are not responsive to classical antidepressants, such as imipramine and fluoxetine (Castrén and Antila, 2017; Karpova et al., 2011; Saarelainen et al., 2003). To corroborate the behavioral experiments and to further explore the mechanisms involved in losartan effects, we used primary cultures of embryonic LAQ824 to evaluate the interaction between RAS and BDNF/TRKB signaling. First, we observed that K252a and PD123319, but not losartan, prevented the ANG2 effect of increasing pTRK levels, thus suggesting a putative action on AGTR2. This possibility is strengthened by the administration of AGTR2 agonist CGP42112, that also induced an increase in TRK activation, and this effect was counterbalanced by prior incubation with the AGTR2 antagonist PD123319. Since the soluble form of TRKB (TRKB.Fc) did not prevent the ANG2 effect on pTRK, it is plausible to consider that AGTR2-induced activation of TRKB would not require increment in BDNF release. In this sense, transactivation of TRKB or a facilitatory effect of basal levels of BDNF on its receptor are reasonable scenarios. Previous studies have described that both the GPCR ligands adenosine and pituitary adenylate cyclase-activating polypeptide can transactivate TRK (Rajagopal et al., 2004). In addition, TRK transactivation by an adenosine 2A receptor agonist was blocked by PP1, suggesting involvement of the SRC family tyrosine kinase (Lee and Chao, 2001). Thus, FYN and other SRC-family kinases are responsible for TRK transactivation (Huang and McNamara, 2010; Rajagopal and Chao, 2006), and lipid raft localization of TRKB is regulated by FYN (Pereira and Chao, 2007). Consistent with this evidence, we observed that ANG2 was also able to increase levels of TRK/FYN coupling in cortical cultures. Therefore, we propose that FYN acts as an intermediary molecule capable of inducing TRKB transactivation when ANG2 acts on AGTR2. Moreover, BDNF itself promotes TRKB/FYN coupling (Iwasaki et al., 1998). Corroborating that prospect, our data also showed that BDNF increased TRKB/FYN coupling. Therefore, both ANG2 and BDNF, which are able to increase pTRK levels, also induce TRK/FYN coupling. As expected, PD123319 blocked TRKB/FYN coupling from ANG2 action, but unexpectedly PD123319 also prevented such coupling from BDNF action. Also unexpected was the data indicating that PD123319 prevents the BDNF effect of increasing pTRKB levels. A generalized interaction of AGTR2 with other TRK members is unlikely, as PD123319 did not prevent NFG action of increasing pTRKA levels. These unforeseen interactions can be explained by the observation that PD123319 was able to reduce surface expression of TRKB, whereas ANG2 led to an increase, thereby suggesting a putative displacement of TRK to the surface upon AGTR2 signaling and a decrease of BDNF effectiveness with prior PD123319. Indeed, modulation of TRK surface trafficking is important considering the two possible scenarios mentioned above for the activation of TRKB, which may occur on cell membrane (Rajagopal et al., 2004). In addition, the MG87.TRKB cell line, which overexpresses TRKB, allowed us to observe co-immunoprecipitation of GFP-tagged AGTR2 and TRKB, suggesting AGTR2/TRKB dimerization. This approach was chosen for the following two reasons: first, as analyzed by the group of Juan Saavedra, commercially available antibodies against AGTRs are far from ideal (Hafko et al., 2013); and second, the cell line used expresses exclusively TRKB, thus being an ideal tool for our purpose. A preliminary analysis showed that Agtr2 mRNA levels were 5-fold higher than Agtr1a in our primary cultures, and this ratio was inverted to 2.5-fold more Agtr1a after incubation with glutamate. This later feature seems to allow a cooperative effect of losartan and ANG2. Using a model of retinal ischemia, it was observed that increased expression of Agtr1a mRNA peaked 12 h after reperfusion, while treatment with candesartan was able to prevent ischemia-induced glutamate release (Fujita et al., 2012). Taken together, these data indicate a possible positive feedback between AGTR1 signaling and glutamatergic transmission. Moreover, consistent with our in vitro observations, Agtr1a levels increased while Agtr2 levels decreased in the medulla of stress-induced hypertensive rats (Du et al., 2013). However, an opposite effect of glutamate on Agtr2 mRNA has also been described (Makino et al., 1998). In this study, glutamate insult led to an increase of Agtr2 mRNA. The precise mechanism where glutamate release might reduce the levels of Agtr2 are still not understood and these apparent discrepancies could rest on methodological differences. For example, the culture method of Makino and colleagues rely on cortical cells cultivated for 14 days, supplemented with calf serum and mitosis inhibitors; our cultures were serum-free (substituted by B27) and cultivated for 8 days without any drugs to prevent cell proliferation.