µ-Opioid Receptor Control of Glutamate/GABA Coreleasing SUM and VTA Projections to the Dentate Gyrus

The dentate gyrus (DG) is a hippocampal structure responsible for filtering cortical input and conveying it to CA3, where it forms associative memories. Granule cells, the principal excitatory neurons of the DG, have extensive dendritic arbors that occupy the molecular layer of the DG. The excitability of these neurons is tightly controlled by GABAergic basket cells, which synapse onto the proximal dendrites and soma of granule cells (Amaral et al., 2007). While granule cells receive the majority of their inputs from the entorhinal cortex, other cortical and subcortical structures provide input that modulates granule cell activity. For example, DG receives input from the ventral tegmental area (VTA) and the supramammillary nucleus (SUM), midbrain structures involved in control of motivation and learning (Ntamati and Lüscher, 2016; Hashimotodani et al., 2018).

The VTA is best known for regulating reward and aversion behaviors through dopamine release, but its dopaminergic projections to the hippocampus are quite sparse (Han et al., 2020). Instead, most VTA–DG projections corelease glutamate and GABA. Activation of these projections reduces granule cell firing in anesthetized animals (Ntamati and Lüscher, 2016). Chemogenetic inactivation of these projections impairs acquisition and reinstatement of fear conditioning in mice, suggesting they aid the formation and retrieval of emotionally valent memories (Han et al., 2020). SUM–hippocampus projections also corelease glutamate and GABA, and they synapse onto both granule cells and surrounding interneurons. Recent studies show that stimulation of SUM fibers causes an overall increase in granule cell activity (Hashimotodani et al., 2018) and inactivation of these projections in vivo impairs retrieval of spatial memories (Li et al., 2020). They have also been linked to emotional memory processing.

Granule cell activity is also regulated by activation of µ-opioid receptors (MOR), which are abundantly expressed in the DG, rendering it susceptible to the effects of MOR agonists, including endogenous opioid peptides, analgesics, and several drugs of abuse. MORs are predominantly expressed by inhibitory basket cells in the DG, and because activation of opioid receptors typically reduces the activity of neurons on which they are expressed, acute application of the MOR-selective agonist DAMGO ([d-Ala²-N-Me-Phe⁴-Glyol⁵]-enkephalin) increases the excitability of granule cells through disinhibition (Mayer et al., 1995; Drake et al., 2007). Yet despite this increased excitability, when VTA terminals in the DG are selectively activated in the presence of DAMGO, glutamate- and GABA-evoked currents in granule cells are decreased (Han et al., 2020). This suggests that MORs are also expressed on VTA terminals in the hippocampus, where they regulate neurotransmitter release.

To test this hypothesis, Kim et al. (2024) investigated how MOR activation affects GABA and glutamate release from SUM/VTA projections in the DG and how this affects granule cell activity. They used an anterogradely transported viral vector to express channelrhodopsin-2 selectively in glutamatergic neurons in SUM or VTA, thus enabling selective photostimulation of SUM/VTA axons in the DG. Currents evoked in granule cells by this stimulation were recorded by whole-cell patch clamp in the presence of kynurenic acid (a glutamate receptor blocker) or picrotoxin (a GABA receptor blocker) to isolate GABAergic or glutamatergic currents, respectively. Because no significant differences were detected in currents elicited by activation of VTA or SUM projections, the currents evoked by the two forms of stimulation were analyzed together.

Kim et al. (2024) first confirmed that activation of MORs potentiated granule cell currents evoked by electrical stimulation in the molecular layer of DG, which simulates inputs from the entorhinal cortex. In the presence of a GABA receptor blocker, the currents evoked by electrical stimulation were potentiated; DAMGO application had no additional effect, consistent with the hypothesis that MOR activation increases granule cell activity by decreasing GABAergic tone (Drake et al., 2007). Importantly, DAMGO significantly reduced both glutamatergic and GABAergic currents elicited by optical stimulation of SUM/VTA projections, suggesting that MOR agonists also inhibit neurotransmitter release from these projections. To further test the hypothesis, Kim et al. (2024) measured the paired-pulse ratio (PPR)—the ratio of PSP amplitudes elicited by two spikes evoked within a short interval. The PPR at synapses between SUM/VTA projections and granule cells increased for both GABAergic and glutamatergic currents in the presence of DAMGO, consistent with DAMGO reducing presynaptic release probability. Finally, coimmunostaining for the vesicular glutamate transporter VGluT2 and MOR1 showed that MORs were expressed in SUM/VTA synaptic terminals in the granule cell layer. Together, these results confirm that MORs are located presynaptically on the SUM/VTA terminals and strongly suggest that activation of these receptors reduces the probability of neurotransmitter release at these terminals.

Kim et al. (2024) next aimed to identify the downstream effectors of MOR activation in SUM/VTA terminals. MORs are coupled to inhibitory G-proteins, which have a variety of downstream effectors that differ across neuron types. In some neurons, MOR activation reduces excitation by inhibiting presynaptic N- and P/Q-type voltage-dependent calcium channels (Reeves et al., 2022); one or both of these channels are required for neurotransmitter release at GABAergic and glutamatergic synapses. To determine whether MOR activation affects the contribution of these channels to neurotransmitter release from VTA/SUM terminals in the DG, Kim et al. (2024) optically stimulated SUM/VTA axons while blocking P/Q-type channels and/or N-type channels in the presence of DAMGO. Without MOR activation, blocking both channel types eliminated most of the optically induced glutamatergic and GABAergic currents in granule cells, confirming that synaptic release of glutamate and GABA is driven by both P/Q- and N-type channels. When DAMGO was administered concurrently with the N-type channel antagonist, evoked currents were significantly but not completely reduced. In contrast, when DAMGO was administered in the presence of the P/Q-type channel antagonist, optically evoked glutamatergic and GABAergic currents were almost completely lost. This suggests that MOR activation reduces presynaptic release from SUM/VTA axons predominantly by inhibiting N-type calcium channels.

Overall, the findings by Kim et al. (2024) confirm that MORs localized to VTA/SUM presynaptic terminals control granule cell activity, independently from their inhibitory effect on basket cells. When MORs are activated in the dentate gyrus, granule cells are disinhibited as a result of MOR-mediated suppression of basket cell spiking, and the influence of SUM/VTA projections on granule cells is diminished by MOR-mediated suppression of glutamate and GABA release from these inputs. The reduced transmitter release from the SUM/VTA projections is primarily attributable to the suppression of N-type calcium currents, with a smaller but significant contribution of P/Q-type currents, resulting in reduced calcium influx.

How does MOR-mediated suppression of SUM/VTA contribute to hippocampal network changes? The excitation of SUM–DG projections has complex effects on granule cell activity because they provide simultaneous excitation and inhibition to interneurons and granule cells (Hashimotodani et al., 2018); but the overall effect of these inputs appears to be facilitation of LTP at certain cortical inputs as a result of increased temporal precision of granule cell responses (Ajibola et al., 2021). In the presence of MOR agonists, this targeted potentiation at select synapses may be lost in favor of widespread LTP of entorhinal inputs, which results from MOR-mediated suppression of inhibitory interneurons and disinhibition of granule cells. While there is insufficient information about the effect of GABA/glutamate-releasing VTA–DG inputs to DG, Ntamati and Lüscher (2016) showed that their stimulation results in the net inhibition of granule cells in anesthetized animals. If this finding is generalized, it may be speculated that, in the presence of MOR agonists, granule cells receive reduced inhibition from the VTA. Considering the complexity of the DG network, Kim et al. (2024) suggest that the acute inhibitory effect of MOR agonists on SUM/VTA–DG projections contributes to the shift in network-wide signaling dynamics, which serves to facilitate LTP of entorhinal inputs.

Future studies could test Kim and colleagues’ hypothesis by examining the role of MOR-mediated control of GABA/glutamate-coreleasing SUM/VTA–DG projections in LTP induction. Under drug-free conditions, optogenetic activation of SUM–DG projections is reported to increase granule cell activation in response to perforant path stimulation (Ajibola et al., 2021). Despite their electrophysiological similarity with SUM-originating projections, it is unclear whether the activation of VTA neurons would cause a similar potentiation in granule cells. Future studies could repeat the LTP protocol described by Ajibola et al. (2021) while stimulating SUM/VTA–DG projections and test the hypotheses that LTP in granule cells is facilitated by the activation of SUM/VTA–DG projections and that this is lost in the presence of DAMGO. Such a study would clarify the contribution of MOR-mediated regulation of these dual transmitter-releasing terminals to hippocampal LTP.

A limitation of the Kim et al. (2024) study is that the authors used acute bath application of MOR agonists and made electrophysiological recordings after only 2–3 min of drug application. In vivo, exogenous MOR agonists remain in the brain for much longer periods (in the order of hours), and these agonists are taken repeatedly, whether for analgesia or as drugs of abuse. Endogenous MOR agonists are released by the hypothalamus into the cerebrospinal fluid, diffusing throughout the brain with a half-life of over 30 min. Previously, Mayer et al. (1995) noted that continuous bath application of DAMGO to the DG for 20 min resulted in interneuron desensitization to DAMGO and eventual loss of disinhibition in granule cells. Indeed, MORs are susceptible to lasting desensitization and internalization in at least some brain regions and neuron types (Mayer et al., 1995). It is unknown whether MORs expressed on SUM/VTA–DG projections are similarly desensitized by prolonged MOR agonism; this should be further investigated. Exploring the temporal dynamics of MOR desensitization in the SUM/VTA–DG synapse could help further decipher the role of these projections in regulating hippocampal inputs and formation of spatial and emotional memories, building on the foundational findings by Kim et al. (2024).