Responsivity to stress and fear is governed by neural circuits that involve the PFC, BLA, CeA, and mediodorsal thalamus (MD). How these regions communicate with each other is incompletely understood, however. Pyramidal neurons in the infralimbic region of the PFC send glutamatergic projections to both the BLA and CeA (Ehrlich et al., 2009), and these glutamatergic inputs can activate GABAergic interneurons to inhibit BLA and CeA activity (Ehrlich et al., 2009). Activation of the infralimbic to BLA pathways has been shown to reduce conditioned fear responses (Sierra-Mercado et al., 2011). The BLA receives input from several thalamic nuclei but sends few or no projections back to these nuclei (Mitchell and Chakraborty, 2013). Nevertheless, both CeA and BLA send robust projections to the magnocellular MD in nonhuman primates and to the medial MD (MDm) in nonprimates (Mitchell and Chakraborty, 2013). Therefore, the BLA sends afferents strictly to the MDm and not other substructures of the MD, such as the magnocellular MD.
Neurons projecting from the BLA to the PFC contain vesicular glutamate transporter 1 (VGLUT1), whereas BLA neurons that project to the MDm have been reported to contain vesicular transporter 2 (VGLUT2) (Timbie and Barbas, 2015). However, there is evidence that most of the MDm-projecting BLA neurons are located in the basolateral nucleus, which is devoid of VGLUT2-expressing neurons (Hur and Zaborszky, 2005), suggesting that VGLUT2 projections to MDm might originate from a neighboring nucleus. Further, there have been reports that BLA neurons projecting to the MDm primarily comprise spine-sparse nonpyramidal neurons containing GABA (McDonald, 1987). Therefore, whether BLA neurons exert an excitatory or inhibitory influence on MDm neurons has been unclear. Recently, Ahmed and Paré (2023) reconciled these inconsistent findings by presenting evidence that the BLA sends both GABAergic and glutamatergic projections to the MDm.
To phenotype BLA neurons that project to the MDm, Ahmed and Paré (2023) injected the retrograde tracer fast blue (FB) into the MDm and measured its coexpression with markers of different subtypes of GABAergic neurons. They first used markers for calbindin and calretinin because 70% of GABAergic BLA neurons express these markers. Further, calbindin has been shown to be expressed in the majority of parvalbumin and somatostatin cells in the BLA. Remarkably, however, none of the FB cells were positive for calbindin or parvalbumin. Nevertheless, a large proportion of FB cells were somatostatin-positive (21 ± 4.3%) and calretinin-positive (10.9 ± 2.3%). Moreover, although many cholecystokinin cells coexpress calretinin, none of the FB cells were cholecystokinin-positive. Together, this suggests that ∼30% of BLA neurons projecting to MDm are GABAergic and express somatostatin and calretinin markers. The remainder of the FB-labeled cells were assumed to be glutamatergic. Thus, the authors conclude that there are both GABAergic and glutamatergic projections from the BLA to the MDm.
To understand how BLA neurons modulate MDm activity, Ahmed and Paré (2023) used optogenetic tools to manipulate BLA terminals in the MDm. They expressed channelrhodopsin 2 selectively in BLA neurons, then photostimulated the channelrhodopsin-expressing terminals in the MDm while recording postsynaptic neurons via patch clamp. Blue light stimuli were also applied at different frequencies to adjust the membrane potential of an MDm cell; and when there was a light-evoked response, the cell’s voltage was set to different values (−55 to −100 mV) to distinguish EPSPs from IPSPs. In the cells that responded, there were incidences of EPSPs, isolated IPSPs, and EPSPs overlapping with IPSPs. To investigate the type of receptors underlying EPSPs and IPSPs, the authors applied antagonists to MDm neurons. EPSPs were abolished by antagonists of NMDARs and AMPARs when applied concurrently. However, EPSPs continued to be evoked, albeit at lower levels, when Na2+ and K+ channels were blocked. These findings suggest that the EPSPs were mediated by NMDARs and AMPARs via monosynaptic glutamatergic connections. When neuronal responses were measured at the cell’s resting potential, light-evoked EPSPs had a low amplitude and never elicited a short latency action potential unless cells were depolarized enough to reach spike threshold (−60 mV). Furthermore, in a large portion of cases (40%), EPSPs followed light-evoked IPSPs. Several findings suggested that these IPSPs were monosynaptic. First, MDm does not contain any local GABAergic cells (Ehrlich et al., 2009). Second, when GABA receptor antagonists were applied, the amplitude of the optically evoked EPSPs was greater than the control groups. Third, a subset of light-evoked IPSPs had no coincident EPSP; and in these cases, IPSP latencies were identical to EPSP latencies, which is inconsistent with the dependence of IPSP excitation from other neurons that project to the MDm. Last, in these cells, light-evoked IPSPs persisted even after the simultaneous addition of NMDA and AMPA antagonists. Interestingly, neurons differed in how PSP amplitude was affected by stimulation frequency. Some neurons showed little to no amplitude attenuation with repetitive stimulation, whereas others showed a marked hyperpolarization with increased stimulation frequency revealing an underlying EPSP that was largely masked by an IPSP. Therefore, GABAergic activity either masked or shunted EPSPs within a subset of MDm neurons. Together, these findings suggest that some MDm neurons receive exclusively glutamatergic or GABAergic projections from the BLA, while others receive both inhibitory and excitatory inputs.
As mentioned by Ahmed and Paré (2023), there are many factors that could explain the modulatory effects of BLA afferents on MDm activity. One possibility involves converging BLA terminals onto a single MDm cell; this is supported by the fact that optogenetic stimulation often resulted in EPSPs with low amplitudes and EPSPs coinciding with IPSPs (Fig. 1A). Another possibility involves BLA GABAergic neurons providing presynaptic inhibition of excitatory input to the MDm, which would diminish EPSP amplitudes in MDm cells. Results show an increase in EPSP amplitude after the introduction of GABA receptor antagonists. This finding could be attributed to GABAergic cells acting on BLA glutamatergic terminals in the MDm, or an MDm cell receiving both glutamatergic and GABAergic input from the BLA, limiting the depolarization caused by the glutamatergic input. Therefore, blocking presynaptic GABA receptors would disinhibit glutamatergic terminals and lead to higher EPSP amplitudes in MDm cells (Fig. 1B).
Potential glutamatergic and GABAergic transmission in MDm-projecting BLA neurons. A, Converging BLA glutamatergic and GABAergic terminals in the MDm. B, Disinhibition of a glutamatergic terminal from the BLA by a long-range GABAergic neuron. C, Corelease of glutamate and GABA from single or separate vesicles in a BLA neuron projecting to the MDm. Figure made in www.Biorender.com.
Another potential mechanism that would explain the findings of Ahmed and Paré (2023, their Fig. 3B) is the corelease of glutamate and GABA from a single BLA terminal. Ahmed and Paré (2023) tried to use a Dlx enhancer that drives transgene expression exclusively in neurons that are GABAergic and a synapsin reporter to drive transgene expression in GABAergic and glutamatergic neurons. Surprisingly, though, both viruses gave a similar EPSP response pattern in MDm neurons. The authors concluded that, while the Dlx enhancer may favor transgene expression in GABAergic cells, it does not prevent expression in glutamatergic cells. However, it is also possible that BLA neurons that are GABAergic corelease glutamate. Classically, it is thought that glutamate and GABA are released from a distinct set of excitatory and inhibitory neurons within the CNS, but several recent reports describe corelease of glutamate and GABA from the same CNS axon or terminal. For example, expression of VGLUT3 in striatal GABAergic neurons can result in the packaging of GABA and glutamate within a single vesicle (Zimmermann et al., 2015). Further, neurons in the VTA and entopeduncular nucleus were recently shown to release both glutamate and GABA onto neurons in the lateral habenula, although the authors acknowledge that it is hard to determine whether the vesicles are packing the same neurotransmitters within the same vesicle (Root et al., 2014). Hence, we hypothesize that corelease of GABA and glutamate could account for some of the postsynaptic effects of optogenetic stimulation, including the shunting or masking of EPSPs and hyperpolarization in MDm neurons (Fig. 1C).
In conclusion, using a variety of techniques, Ahmed and Paré (2023) demonstrated that the BLA sends glutamatergic and GABAergic projections to MDm. This research deepens our understanding of the complex dynamics and mechanisms that regulate communication in neural circuits. The BLA is a stress-sensitive brain region that undergoes morphologic changes in response to stress exposure (Ehrlich et al., 2009), which may give rise to stress-induced psychopathologies, such as major depression and post-traumatic stress disorder. In particular, numerous studies have indicated that BLA plasticity is important for the acquisition of learned fear (Ehrlich et al., 2009). Further, in the context of social defeat stress, overexpression of cAMP-responsive binding element (CREB) within the BLA enhances subsequent defeat-induced changes in social behavior, such as increased submissive and defensive behavior (Jasnow et al., 2005). However, the downstream targets of CREB activity following the binding of cAMP response elements in the promoter region of these target genes within the BLA remain unclear, although one possibility is that CREB affects glutamatergic and GABAergic neuronal changes, which then affect neuronal output to other regions. Therefore, GABA and glutamate neurons from the BLA to other regions, such as the MDm, may prevent overexcitation of fear-related responses to ensure appropriate reactivity to fearful stimuli. In the context that these neurons project separately to the MDm, this may be more efficient in modulating neuronal activity with less cellular input, which may improve responsiveness and adaptability of cellular communication. Future research will be needed to delineate cross-talk between GABAergic and glutamatergic projections and how the BLA-MDm pathway fits within the complex tripartite circuit of the BLA, MD, and PFC.
Footnotes
Editor’s Note: These short reviews of recent JNeurosci articles, written exclusively by students or postdoctoral fellows, summarize the important findings of the paper and provide additional insight and commentary. If the authors of the highlighted article have written a response to the Journal Club, the response can be found by viewing the Journal Club at www.jneurosci.org. For more information on the format, review process, and purpose of Journal Club articles, please see http://jneurosci.org/content/jneurosci-journal-club.
C.J.W. was supported by National Academies Science, Engineering, and Medicine Ford Graduate Fellowship. We thank Dr. Matthew A. Cooper for comments on the manuscript.
The authors declare no competing financial interests.
- Correspondence should be addressed to Conner J. Whitten at cwhitte9{at}vols.utk.edu