Aminergic neurons mediate reward and aversive signals in Pavlovian conditioning in mammals and insects. In mammals, different classes of midbrain dopamine neurons (DANs) in the ventral tegmental area are known to mediate reward and aversive signals (Berridge et al., 2009, Schultz, 2015, Gershman and Uchida, 2019). Similarly, in fruit-flies (Drosophila melanogaster), different classes of DANs that convey appetitive and aversive reinforcement signals project to different compartments of the lobes of the mushroom bodies (MBs) and form appetitive and aversive memory traces (Liu et al., 2012, Burke et al., 2012). In the MB lobes, DANs make synapses on axon terminals of MB intrinsic neurons (Kenyon cells) that mediate signals of an olfactory conditioned stimulus (CS) and alter the efficacy of synaptic transmission from Kenyon cells to MB output neurons (MBONs) (Cohn et al., 2015). After conditioning, overall activities of MBONs guide the appetitive or aversive conditioned responses (CRs) (Aso et al., 2014, Cognigni et al., 2018, Boto et al., 2020). Octopamine neurons (OANs) also play roles in mediating appetitive reinforcement signals in fruit-flies: OANs that mediate sweet taste signals of sucrose US make synapses with DANs that project to the MB lobes (Liu et al., 2012, Burke et al., 2012, but see also Kim et al., 2013, Sabandal et al., 2020).

In crickets (Gryllus bimaculatus), we reported that injection of epinastine (a specific OA receptor antagonist in insects, Roeder et al., 1998, Degen et al., 2000) impaired appetitive conditioning and execution of an appetitive CR but did not impair aversive conditioning or execution of an aversive CR (Unoki et al., 2005, Unoki et al., 2006, Mizunami et al., 2009, Matsumoto et al., 2015). In addition, we found that silencing of OA1 octopamine receptors by RNAi impaired appetitive conditioning but not aversive conditioning (Awata et al, 2016). A pharmacological study in honey bees showed that OA1 receptors are activated by a low dose of OA and inhibited by a low dose of epinastine (Beggs et al., 2011). Therefore, we consider that the OA1 receptor is likely to be the target of epinastine. In fruit-flies, it has been reported that this type of receptor (OAMB) mediates OA-ergic synaptic transmission (Watanabe et al., 2007). We thus assumed that epinastine-induced impairments of conditioning and execution of the CR were due to blockade of synaptic transmission from OANs, and we proposed neural circuit models of salivary conditioning in which (1) a class of OANs mediate appetitive reinforcement signals for conditioning and (2) after conditioning, presentation of a CS activates another class of OANs and this activates downstream neural pathways to produce the CR (For details of the models, see Terao and Mizunami, 2017, Mizunami et al., 2019, Mizunami, 2021).

Control of the CR by aminergic neurons has been reported in other insects, but it is less well established. In honey bees, Farooqui et al. (2003) reported that injection of an OA receptor antagonist, mianserin, or RNA interference of the OA1 receptor in the antennal lobes impaired execution of the CR after conditioning of an odor with sucrose US. In this species, it has been suggested that neurons mediating olfactory CS signals and those mediating sucrose US signals converge in the antennal lobe and calyces of the MB for conditioning (Hammer, 1993; Hammer and Menzel, 1988). In fruit-flies, it has been reported that some DANs projecting to the lobes modulate a CR (Krashes et al., 2009, Cohn et al., 2015, Senapati et al., 2019, Scaplen et al., 2020). However, whether CRs are, in general, regulated by activities of DANs remains unclear. Therefore, more studies are needed to investigate whether control of the CR by aminergic neurons is ubiquitous among insects.

In cockroaches (Periplaneta americana), we reported that pairing of an odor (CS) with sucrose (US) produces an increased amount of salivation in response to the odor (Watanabe and Mizunami, 2007). This conditioning can be monitored as increased activities of salivary neurons (Fig. 1A) to presentation to the conditioned odor (Watanabe and Mizunami, 2006, 2008; Matsumoto et al., 2013). Our study with local injection of mecamylamine (an acetylcholine receptor antagonist) into various brain areas suggested that the sites of association of an odor CS with sucrose US for conditioning of salivation are the calyces or lobes of the MB but not the antennal lobe or the lateral protocerebrum (Watanabe et al., 2011). We proposed a model of salivary conditioning in which presentation of a CS after conditioning activates MBONs and signals for their activations are transmitted to the subesophageal ganglion, possibly via the lateral protocerebrum, and activate salivary neurons to produce a CR (Supplemental Fig. S1) (Watanabe et al., 2011).

In this study, we investigated whether injection of epinastine into the head hemolymph impairs the conditioning of activities of salivary neurons and execution of the CR after conditioning in cockroaches. Moreover, we investigated whether local injection of epinastine into calyces, vertical lobes of the MB or antennal lobes impairs the execution of the CR, and the results suggested that OANs in the vertical lobes play critical roles in control of the execution of the CR.