Prefrontocortical Astrocytes Regulate Dominance Hierarchy in Male Mice

Social hierarchy is a general self-organizational scheme in animal societies in which each member achieves a certain status of domination or subordination when getting access to resources, with profound consequences for survival, health, reproductive success, and multiple behaviors [1, 2]. The prefrontal cortex (PFC) was recognized as a central regulator of neuronal circuits determining the formation of social hierarchy; brain areas upstream of the PFC convey information about social status, and downstream brain regions execute dominance behavior [3]. Social deficits have been suggested to be due to an elevation of excitatory/inhibitory (E/I) balance within the mouse dorsomedial (dm)PFC [4]. A compensatory increase of inhibitory cell excitability partially rescued social deficits caused by such a disturbance of E/I balance.

In the presently highlighted article [5], the authors measure social behavior with the tube test; two mice were trained to meet in the middle of a narrow tube in which the submissive mouse gives way to the other, the dominant one [1]. Four male mice from the same cage were tested in a pairwise fashion, in order to establish a rank order of dominance between them (rank-1-4; with the highest rank of 1). Then, the intra-astrocytic Ca²⁺ activities were measured by means of the genetically encoded Ca²⁺-indicator GCaMP6s, introduced via adeno-associated virus (AAV) injection into the dmPFCs; the generated fluorescence signals were recorded by the time-correlated single-photon counting (TCSPC) technique in freely moving mice. The dominant pushing and resisting mice exhibited elevated Ca²⁺ activities, while the submissive approaching and retrieving mice had no such response.

Then, chemogenetic and optogenetic stimulation of dmPFC astrocytes was carried out to see whether such stimulations increased their Ca²⁺ activities. When an astrocyte-specific AAV carrying hM3Dq was injected into the dmPFC leading to the expression of this modified human form of an M3 muscarinic receptor, activation was initiated by its artificial ligand clozapine-N-oxide (CNO). Interestingly, CNO application increased the Ca²⁺ activity in the dmPFC and also resulted in a higher rank of dominance in mice as determined in the tube test. Thus, rank-4 mice now could be classified as belonging to rank-3, which means more pronounced pushing/resisting behavior. To strengthen this conclusion, hPMCA2w/b was expressed in the dmPFC astrocytes to inhibit Ca²⁺ activation. Simultaneously to the decreased Ca²⁺ level, the social rank of the rank-1 mice was significantly decreased. Furthermore, in accordance with the chemogenetic results, optogenetic stimulation of GFAP-ChR2 (astrocyte-specific channel rhodopsin-2) expression in rank-4 mice elevated the social rank by one or two levels in the tube test.

Following chemogenetic stimulation of the dmPFC, brain slices were prepared from this area of the brain and patch-clamp recordings were carried out from layer V pyramidal neurons. Whereas the miniature excitatory postsynaptic current (mEPSC) frequency increased on chemogenetic astrocyte stimulation with no change in amplitude, the amplitude of the miniature inhibitory postsynaptic currents (mIPSCs) decreased under the same conditions. mEPSCs were measured in the presence of tetrodotoxin to block action potential generation, and the GABA_A (γ-aminobutyrate A) receptor antagonist picrotoxin; mIPSCs were recorded in the presence of the NMDA receptor antagonist AP-5 and the AMPA receptor antagonist CNQX as well as again in added tetrodotoxin. The increased frequency of mEPSCs pointed to a presynaptic site of action, while the depressed mIPSC amplitude suggested a postsynaptic modulatory effect. Hence, upon chemogenetic astrocyte stimulation, increased E/I ratios were observed.

In order to clarify the mechanism of the Ca²⁺-induced E/I ratio increase, gliotransmitter release was measured from CNO-stimulated hM3Dq-expressing dmPFC astrocytes in acute brain slices and microdialysed mice brains. Under both conditions, the release of glutamate as well as ATP was markedly elevated by CNO. By contrast, reduced glutamate and ATP levels were observed in the brain slice media of hPMCA2w/b-injected mice after tube testing. It was concluded that gliotransmitter release from astrocytes triggered by the tube test was reduced by restricting the causally involved increase in the intracellular Ca²⁺ level. Glutamate and ATP are the two main transmitters released from both neurons and astrocytes by exocytotic and especially from the latter ones, by non-exocytotic routes [6].

It is well known that astrocyte-derived glutamate increases excitatory synaptic transmission via both pre- and postsynaptic targeting [7]. Moreover, astrocyte-derived ATP was reported to inhibit by a post-synaptic mechanism the mIPSC amplitudes, probably by activating a P2X2 receptor, known to be located at pyramidal neurons; these receptors allow Ca²⁺ entry into these neurons promoting endocytosis of GABA_A receptors by a phosphorylation reaction [8]. In accordance with the supposed postsynaptic mode of action, the mIPSC amplitude decrease in dmPFC pyramidal neurons by chemogenetic astrocyte stimulation was almost completely rescued by PPADS treatment. However, a very high concentration of PPADS (100 µM) was used, which is no longer selective for the excitatory P2X2 receptors, but in addition also blocks the P2Y1 receptor-type, which releases [Ca²⁺]_i in the pyramidal neurons via the activation of Gα_q [9], and the subsequent obligatory stimulation of the phospholipase C/inositol 1,4,5-trisphosphate (IP3) pathway. This effect may add up to the endocytosis of GABA_A receptors by P2X2 receptor-mediated phosphorylation. Therefore, we took the liberty to include P2Y1 receptors into the original scheme proposed by the authors [5], outlining the effect of neuronal changes induced by the increase of [Ca²⁺]_i in dmPFC astrocytes, eventually regulating dominance behavior in male mice (Fig. 1).

Fig. 1.

Hypothetical astrocyte-dependent and neuronally executed mechanisms underlying the establishment of behavioral dominance in the tube test. For further explanation see the text. [Ca²⁺]_i intracellular Ca²⁺ concentration; Glu Glutamate; ATP Adenosine 5’-triphosphate; P2X2 P2X2 receptor; P2Y1 P2Y1 receptor; GABA_A γ-aminobutyrate receptor; mGlu Metabotropic glutamate receptor; iGlu Ionotropic glutamate receptor, upward arrow, stimulation; downward arrow, inhibition

Taken together, the present results unequivocally demonstrate that two distinct gliotransmitters, glutamate and ATP modulate synaptic E/I balance with distinct mechanisms, thereby controlling mouse dominance behavior. The first step in the astrocyte-triggered chain of events is the increase of [Ca²⁺]_i in astrocytes of the dmPFC caused by a winning situation of dominant mice (Fig. 1). This induces the exocytotic release of glutamate onto presynaptic ionotropic and metabotropic glutamate receptors of pyramidal neurons to facilitate glutamate release from the terminals of these very neurons. Simultaneously, astrocytic ATP is released onto postsynaptic P2X2 and P2Y1 receptors of pyramidal neurons, decreasing GABA_A receptor activation by furthering the endocytosis of GABA_A receptors. As a result of these changes, glutamatergic excitation is elevated, while GABAergic inhibition is diminished.

The results of this article have also a possible clinical significance. The loss of social status is known to be a particularly prominent risk factor for major depressive disease (MDD) in humans and depressive-like behavior in laboratory rodents [10]. The depressive-like behavior develops, when the animal is not able to cope with a stressful situation and gives up the futile trials to compensate it (“learned helplessness”). Such a situation evolves in the tube test, when formerly dominant mice are forced by the experimental arrangement to give up their previous rank-1 status and subordinate mice push forward against them.

Conflict of interest

The authors declare that they have no competing interest.