The initial finding of previous epidemiologic studies suggested a strong positive correlation between early-life BPA exposure and working memory deficits in children. Here, for the first time, we provide insights into the neural circuit mechanism, with a specific emphasis on the vHPC-mPFC circuit, underlying this cognitive dysfunction in BPA-exposed mouse pups. Firstly, BPA exposure during early development (E0-P21) resulted in both spatial and visual working memory deficits of mouse pups during DNMTP and object recognition tests, respectively. Through in vivo multi-channel electrophysiological recording, we observed significant reductions in θ and γ oscillation power in the mPFC, along with altered θ-γ phase-amplitude coupling within the vHPC-mPFC circuit during the DNMTP tests. These alterations in response to BPA exposure were attributable to impaired synaptic connection within the vHPC-mPFC circuit, characterized by attenuated postsynaptic regulation and synaptic plasticity in the mPFC.
During pregnancy and lactation phases, the brain exhibits a high susceptibility to environmental risks, which can lead to significant challenges in the formation and maturation of large synapses [35]. It is well-documented that early-life BPA exposure is associated with behavioral deficits in children, such as, anxiety, social avoidance, etc. [10, 11]. In our present study, BPA exposure during the pregnant and lactation phase caused noticeable impairments in both spatial and visual working memory of male mouse pups. Compared to control group, BPA-treated pups exhibited similar learning curves in the T-maze task with a 10-second delay for food rewards during the DNMTP training phase. It indicates that mice did not experience a decline in their learning ability during this training task due to BPA exposure, and the decrease in the final testing score for a 1-minute delay is not attributable to the forgetting of DMNTP tasks in BPA-exposed mice. Moreover, BPA-exposed mice spent less time on exploring the novel objects in the object recognition test, which implies deterioration in visual working memory after early-life BPA exposure. These behavioral results are consistent with previous findings that there was a positive association between urinary BPA concentrations and poor working memory scores in boy [9]. Both of these behavioral paradigms in our present study, the T-maze and novel object recognition test, require mouse to maintain short-term memory of spatial and object cues for subsequent goal-directed decision-making [36]. To some extent, the absence of spatial learning impairment in DNMTP tasks following BPA exposure suggests that mouse pups had no decline in short-term memory within 10 seconds or hippocampus-dependent spatial navigation ability required for making goal-directed decision. However, when the delay time in DNMTP tasks extended to 1 minute, BPA-exposed mice (1 mg/kg/day) performed worse compared to control mice, which aligns with previous investigation about effects of perinatal BPA treatment (0.5 mg/kg/day) on the performance of mouse pups in Morris’s water maze (MWM) test (a common test for long-term spatial memory in rodents) [37]. It is worth noting that mouse pups with BPA exposure exhibited a clear deterioration in learning of MWM tasks [37]. This different learning outcomes to the learning performances of BPA-exposed mouse pups in DNMTP tasks maybe due to the differing difficulty levels of MWM and DNMTP tasks.
The DNMTP task, comprising encoding (sample), maintenance (delay), and retrieval (choice) phases, is a well-established paradigm for investigating spatial working memory in rodents [17]. The performance in tasks dependent on working memory relies on the processing ordering information within the mPFC, involving the encoding of remembered information and the sequential order of how the information is manipulated [24]. In the case of the DNMTP task, mice rely on the mPFC to retrieval spatial cue information acquired during sample phase to locate the new food reward during the choice phase. Neurophysiologically, neural oscillation and cross-frequency coupling, especially, θ-γ coupling, has been widely recognized as markers of information order during working memory [17]. In our present study, DNMTP task were further chose to investigate the physiological mechanisms underlying BPA-induced working memory deficits in rodents. We found that there was a significant reduction of θ and γ oscillation power in the mPFC during choice phase, which corresponds to the retrieval of previous memory, the food location in the sample phase. Additionally, θ power during choice phase is not positively correlated with running speed of mice which rules out the effects of running speed on the power of theta rhythm in LFP [38]. Moreover, the mPFC is involved in the regulation of sociability and emotional learning and memory in mice through its direct-projections to the thalamus [39, 40]. The decline in the local circuit activity of the mPFC under specific situation also provide the physiological basis of social avoidance and anxiety-like behavior in male mice after paternal BPA exposure [41].
Tracing studies have shown that there is a direct and monosynaptic projection from the vHPC to mPFC [16]. The functional connectivity between the vHPC and the mPFC supports various cognitive processes, such as working memory tasks [16]. In this study, we found that θ oscillations in the vHPC can modulate γ power in the mPFC, exhibiting a higher MI value during the choice phase compared to other phases. This finding aligns with the previous research, θ-γ oscillation coupling within the vHPC-mPFC circuit subserves a successful spatial working memory performance by encoding spatial cues during DNMTP tasks [17, 18]. Following BPA exposure, the phase-dependent θ-γ oscillation coupling during choice actions was obviously reduced compared to control mice. Notably, there was no difference of coupling values during sample phase between control and BPA-exposed mice. These results imply that BPA-induced working memory deterioration is attributed to disturbances in local cortical activity within the mPFC and disruptions in cross-cortical coupling within the vHPC-mPFC circuit, which is crucial for recalling spatial cues during DNMTP tasks. Importantly, for inhibition of vHPC-mPFC circuit can directly cause anxiety behavior in mice [20], disturbance of vHPC-mPFC circuit in offspring of mice also provides the circuit base of BPA-induced anxiety [7].
LTP, a form of activity-dependent synaptic plasticity, is widely used to assess the efficacy of synaptic transmission between two brain sites [42]. Optogenetic inhibition of vHPC terminals in the mPFC can directly cause a worse spatial working memory performance in a T-maze task [17]. In the present study, LTP induced within the vHPC-mPFC circuit by HFS after BPA exposure cannot be maintained as that in the control group. This attenuated LTP after BPA exposure provided the synaptic basis of blunted θ-γ oscillation synchrony between the vHPC and the mPFC. These findings are consistent with previous reports of synaptic plasticity impairment in the BPA caused spatial memory decline [43]. For LTP enables the storage of information at the synaptic level for memory formation [44], BPA-induced LTP decreases also provide a cellular basis of spatial working memory deficits at synaptic level. Furthermore, I/O curve in control and BPA-exposed mice had a same trend with current stimuli, which suggests no changes in basal synaptic transmission within the vHPC-mPFC circuit [45]. In addition, during sample phase of DNMTP tasks, we did not find the decline in θ oscillation power in the mPFC and θ-γ coupling within the vHPC-mPFC circuit, which implies early-life BPA exposure had no effect on neural activity of hippocampus to disturb spatial working memory performance of mouse pups [17]. Several previous studies have shown that early-life BPA exposure cause spatial memory with decrease expression of excitatory receptors in the hippocampus [37, 46], but there was no detailed alteration in different brain region division, such as the vHPC, in these studies. The hippocampus possessing a three-synaptic circuit plays a crucial role in memory processing [47]. The difference in hippocampal activity related detections could be attributed to the varying neural circuits or brain regions employed by distinct behavioral paradigms [18]. Building upon our observation of different oscillation changes between the sample and choice phases, our future research aims to investigate the contributions of different subregions and projecting cell types in reshaping behaviors by the hippocampus following early-life exposure to BPA.
LTP induction is rooted in increasing release of neurotransmitters from pre-synaptic terminal or/and accumulation of glutamate receptors on postsynaptic sites [42]. Through in vivo electrophysiological recording, we found that BPA exposure only decreased the amplitude, not the frequency, of sEPSC decreases in the mPFC when compared to control group. It implies that BPA-induced synaptic plasticity deficits in the mPFC, are primarily associated with postsynaptic deterioration in the mPFC, rather than changes in presynaptic transmitter release. In addition, we did not detect any differences in the PPF in the vHPC-mPFC circuit between control and BPA-exposed pups. This finding confirms that BPA-induced synaptic plasticity declines in the vHPC-mPFC circuit are not dependent on the alterations in presynaptic neurotransmitter release probability from the vHPC [48]. Most excitatory presynaptic terminals target postsynaptic sites, specifically dendritic spines on postsynaptic pyramidal cells, to form synapses [49]. Dendritic spines provide the anatomical basis for synapse specificity for LTP, and enlargement of spine size is often accompanied with electrophysiological LTP [50]. Here, we found an obvious reduction in dendritic spine density and mushroom head size in mPFC neurons from early-life BPA exposed mice. This suggests postsynaptic mechanism underlying BPA-induced synaptic plasticity and working memory impairment. It is also consistent with previous findings that prenatal BPA disrupt neuronal morphology and synaptic plasticity [51]. In the present study, LTP induction within 2-hour window within the vHPC-mPFC circuit is a neuronal model for short-term memory [52]. Western blotting results showed that a notable decrease in the expression level of short-term memory markers, p-CamKII and p-Erk, in the mPFC after LTP induction in mice with early-life BPA exposure. This reduction is in line with the synaptic plasticity declines in BPA-exposed mouse pups. At the last step of LTP, trafficking of overexpression of GluR1 contribute to the rapid accumulation of AMPA receptors on the postsynaptic sites [53]. The expression declines of GluR1 in the mPFC provide a molecular mechanism for mPFC-dependent synaptic plasticity and function impairment following perinatal BPA exposure.
In total, our results show that perinatal BPA exposure deteriorates working memory, which is attributable to down-regulated oscillation power in the mPFC and blunted cross-oscillation coupling within the vHPC-mPFC circuit. These behavioral and physiological deterioration after BPA exposure involves a reduction in the synaptic transmission efficiency in the vHPC-mPFC circuit. Furthermore, this synaptic plasticity decline is primarily driven by post-synaptic mechanism in the mPFC. Our current findings provide the “missing” physiological and synaptic basis for working memory deficits resulting from early-life BPA exposure.