The current studies demonstrate that DH astrocytes undergo changes in their reactivity, cellular interactions, and intracellular signaling during chronic heroin exposure and withdrawal, and that these changes may play a role in enhanced fear learning. We show that chronic heroin exposure and withdrawal is associated with alterations in astrocyte reactivity, including increased surface area and volume of GFAP immunoreactivity in the DG. However, this effect is not maintained when surface area and volume of GFP from the GfaABC1D-Lck-GFP viral construct is analyzed, indicating that astrocyte morphology did not increase concurrently with astrocyte reactivity. Along with this enhanced reactivity, we also observed an increase in astrocyte-neuron interactions, signified by increased colocalization of PSD-95 and GfaABC1D-Lck-GFP at the 24-hour withdrawal time point. Critically, we also see that manipulating astrocytic intracellular signaling has functional consequences on the maladaptive behavioral changes that occur following chronic heroin and withdrawal. Specifically, we show that stimulating Gi signaling in DH astrocytes during withdrawal blocked the development of heroin withdrawal-enhanced fear learning. These findings provide potential new evidence that heroin withdrawal-induced astrocytic modifications may play a role in the development of future enhanced fear learning and could be a potential mechanism by which opioids and opioid withdrawal can elicit fear- and arousal-related features of PTSD symptomatology.

The present findings illustrate that enhanced astrocyte reactivity is associated with withdrawal from chronic heroin administration and may also be involved in HW-EFL. Specifically, we found that exposure to withdrawal from chronic heroin administration increases the volume and surface area of GFAP reactivity in the DG using confocal microscopy to create a 3D-reconstruction of GFAP-positive staining in the astrocytes. This complements our previous finding, which showed a percent area increase in DG-GFAP immunoreactivity at the 24-hour heroin withdrawal timepoint (Parekh et al. 2020). Additionally, we chose to focus on the 24-hour timepoint as our previous findings have supported the notion that heroin withdrawal, rather than heroin administration, drives these observed increases in GFAP expression as GFAP immunoreactivity was not increased simply after the administration of heroin (Parekh et al. 2020). The increased GFAP surface area and volume observed in the present study is indicative of an activated astrocyte profile, as reactive astrocytes tend to have upregulated levels of GFAP, leading to an increased number of GFAP-positive cells (Beitner-Johnson et al. 1993).

In further support of these observed effects on astrocyte reactivity, it has been previously identified that repeated opioid administration induces a pro-inflammatory state in astrocytes, including upregulated activation markers (Hutchinson et al. 2011). One hallmark of such astrocytic activation is the increased release of pro-inflammatory cytokines, which our lab has also demonstrated to be elevated following heroin administration and withdrawal (Parekh et al. 2020, 2021). In turn, these changes in reactivity markers may be associated with enhanced fear learning by inducing pro-inflammatory cytokine release, which influences anxiety-like behaviors and synaptic plasticity (Cao et al. 2013; Fellin et al. 2004; Shigetomi et al. 2008), and plays a critical role in hippocampal-dependent learning and memory (Jones et al. 2018a). Further, our laboratory has also established that elevations of pro-inflammatory cytokines are functionally relevant to the development of enhanced fear learning following heroin administration with withdrawal (Parekh et al. 2020, 2021).

Although we did not observe membrane-dependent morphological changes using the GfaABC1D-Lck-GFP viral construct in tandem with increased GFAP volume and surface area, Lck-GFP is a structural marker that solely examines astrocyte’s structure, not reactivity. Indeed, evidence indicates that changes in astrocyte reactivity are not always associated with morphological changes, and a variety of reactivity indicators are needed to establish an inflammatory profile (Escartin et al. 2021). Our previous findings support our evaluation of an inflammatory reaction, as we have established increases in astrocyte-derived inflammatory cytokines following heroin administration and withdrawal (Parekh et al. 2020). In turn, our findings suggest that despite unchanging astrocyte morphology, astrocytes become more reactive during withdrawal.

Along with alterations in cytokine release that lead to alterations in stress responses, changes in astrocyte reactivity can also modify astrocytic interactions with adjacent neuronal elements, particularly synapses. In the present study, we did in fact observe alterations in astrocyte-neuron interactions at the 24-hour heroin withdrawal timepoint, which was indicated by increased colocalization of LCK-GFP and PSD-95 puncta. As PSD-95 serves as a marker for neural postsynaptic activity, these findings suggest that withdrawal from chronic heroin administration led to heightened astrocyte-neuron interactions, which may be related to increased fear learning through several mechanisms.

For example, increased astrocyte-neuron interactions may occur through heroin’s ability to alter astrocytic peripheral processes, especially regarding their association with neuronal synapses (Kruyer et al. 2020). Changes in these distal processes, known as leaflets, have been shown to be involved in consolidating fearful memories through neural interactions (Badia-Soteras et al. 2022; Verkhratsky et al. 2020). In turn, astrocyte-neuron interactions observed following heroin administration and withdrawal may play a role in developing the freezing behaviors in heroin withdrawal-enhanced fear learning through changes in memory encoding.

Further, heightened interactions between astrocytes and neurons may be instigated by heroin administration and withdrawal through heroin’s ability to directly activate µ-opioid receptors found on astrocytes (Nam et al. 2018). Activation of these µ-opioid receptors raises calcium levels within the cell and prompts the release of the gliotransmitter glutamate (Corkrum et al. 2019). This glutamate release initiates slow inward currents by activating neuronal N-methyl-D-aspartate (NMDA) receptors on neurons. This enhanced glutamate release and NMDA receptor activation may represent a potential mechanism for the sensitization of fear learning by heroin. This notion is reinforced by the well-established connection between PSD-95 and synaptic plasticity of glutamatergic synapses, as the NMDA receptor/PSD-95 complex exhibits one of the strongest associations within the PSD (Naisbitt et al. 1999). In turn, increased colocalization between GFAP and PSD-95 in animals exposed chronically to heroin suggests that NMDA receptors are involved in this astrocyte-neuron association.

Interestingly, we also see that we can attenuate fear learning behaviors by altering astrocyte function. We found that activation of astrocytic intracellular signaling can counteract the effects of opioid activation on astrocytes, as we observed that stimulating astroglial Gi signaling during heroin withdrawal significantly attenuates HW-EFL. This result could function through the inhibition of inflammatory astrocytic functioning that involves glutamatergic interactions between astrocytes and neurons. Although Experiment 3 solely examined the behavioral effect of stimulating astrocytic Gi signaling, we hypothesize that suppressing the increases in GFAP expression observed in Experiment 1 and 2 through activation of astrocytic Gi signaling underlies its ability to prevent heroin withdrawal-enhanced fear learning. The literature also suggests that corresponding changes in GFAP expression are present following astrocytic Gi stimulation. For example, it has been previously shown that chemogenetic stimulation of Gi signaling via administration of CNO diminished the LPS-induced increases in the number of GFAP‐positive astrocytes, as well as GFAP immunoreactivity in the hippocampus (Kim et al. 2021). These findings suggest that the astrocytic Gi pathway plays an inhibitory role in astrocyte reactivity, which supports our findings. Furthermore, stimulation of Gi signaling is shown to reduce LPS‐induced levels of IL-1β and TNF-α lL-1B, inflammatory cytokines that are highly colocalized with astrocytes and involved in HW-EFL (Parekh et al. 2020, 2021). In turn, activating astrocytic Gi signaling may inhibit elevated levels of astrocyte-derived pro-inflammatory cytokines during heroin withdrawal. Future studies should address inflammatory mediation directly by examining changes in pro-inflammatory cytokines such as IL-1β or TNF-α, which we have previously shown to be mechanistically implicated in this paradigm. Moreover, conducting additional experiments, particularly employing a distinct astrocyte construct such as a Gq-DREADD or PMCA/CalExt, could establish a direct connection between Gi signaling and the observed changes, rather than attributing them to overall astrocyte modulation.

Additionally, the observed effect of stimulating Gi signaling in astrocytes on behavior could be mediated by the effect of Gi-coupled signaling on calcium responses in astrocytes, as activation of endogenous Gi-coupled signaling is shown to produce weak calcium responses (Chai et al. 2017). Weakened calcium responses could lead to decreased gliotransmitter release, including glutamate release, from astrocytes. Thus, stimulation of Gi signaling may prevent excess glutamatergic activation of neurons, preventing sensitization of fear-learning behaviors. Further studies can examine the effect of Gi pathway activation on astrocyte morphology and astrocyte-neuron interactions. Furthermore, the interconnected nature of regions such as the amygdala, striatum, and VTA raises questions about how alterations of astrocytic activation in the dentate gyrus may extend beyond localized effects (Kahn and Shohamy 2013). Future studies should aim to examine how astrocytic activation in the DG could shape broader neural circuits involved in addictive behaviors and reward mechanisms.

Overall, the findings suggest a link between enhanced astrocyte activation and heightened astrocyte-neuron interactions. These observations underscore the intricate relationship between astrocytes and neuronal elements, shedding light on the potential mechanisms by which astrocytes contribute to neuronal function, synaptic plasticity, and enhanced fear learning. These findings offer valuable insights into potential treatment targets for the development of innovative therapeutics aimed at addressing comorbid PTSD-OUD. Our findings strongly support the notion that heroin administration and withdrawal elicit a stress response, which can lead to alterations in astrocytic hippocampal synaptic plasticity and significantly impact drug-related memory formation. To better understand the precise mechanism through which hippocampal astroglial Gi stimulation influences heightened fear learning and heroin withdrawal, further investigations are warranted. The current study acknowledges a notable limitation concerning the lack of a direct link between the collected structural data and the observed HW-EFL effect. To address this limitation, one possible future study could utilize Gi-DREADD to specifically investigate structural and synaptic interaction endpoints. These experiments would provide a robust connection between the findings of the first two experiments and the more functional aspects explored in experiment three. These future studies will contribute to a comprehensive understanding of the complex interplay between astrocytes, synaptic plasticity, and behavioral responses associated with PTSD-OUD, thereby paving the way for more effective therapeutic interventions.