Dysregulated fear memories can be extremely maladaptive, such as in post-traumatic stress disorder (PTSD; Kessler et al., 2017; Watson, 2019). Recent studies suggest that altering fear memories might be an effective treatment strategy (Nader, 2015; Kida, 2019; Uniyal et al., 2020). Much of this work has relied on rodent models of fear memory, such as Pavlovian fear conditioning in which an aversive unconditioned stimulus such as a foot shock is paired with a neutral conditioned stimulus such as a tone or context (Fenster et al., 2018). Several strategies have been employed to attenuate memory after fear conditioning. For example, inhibiting the conversion of a short-term memory to a long-term memory (Schafe and LeDoux, 2000; Duvarci et al., 2008) or developing an inhibitory extinction memory (Myers and Davis, 2007) reduces conditioned freezing. Others have decreased or erased memory by using retrieval to render the memory labile and inhibiting reconsolidation (Nader et al., 2000) or facilitating erasure via extinction (Monfils et al., 2009; Clem and Huganir, 2010).

A recent study found that if mice inhaled 10% carbon dioxide (CO₂) during retrieval, cued fear memory was made more labile than with retrieval alone, making it more susceptible to modification via subsequent extinction or reconditioning. CO₂ inhalation during cued retrieval facilitated insertion of Ca2 + -permeable AMPA receptors (CP-AMPARs), increased cAMP-response element binding protein (CREB) phosphorylation and re-activated a greater percentage of neurons in the memory trace as compared to retrieval alone (Du et al., 2017). In addition to being labile after a retrieval event, fear memories are labile for a period of time immediately following acquisition during which they are vulnerable to disruption (Schafe and LeDoux, 2000; Duvarci et al., 2008). Thus, we wondered if CO₂ inhalation during this labile period following acquisition might similarly alter fear memory.

Moreover, Du et al. found that the CO₂ effect on cued retrieval depended on the acid-sensing ion channel-1 (ASIC1A; Du et al., 2017). ASIC1A is a synaptic cation channel that is activated when extracellular pH drops (Waldmann et al., 1997; Wemmie et al., 2002, 2013; Zha et al., 2006). It is abundantly expressed in the brain, but is particularly abundant in fear circuit structures such as the amygdala (Wemmie et al., 2003; Coryell et al., 2007; Price et al., 2014), where it has been implicated in synaptic transmission and plasticity (Du et al., 2014; Chiang et al., 2015). Several CO₂-induced behaviors depend on ASIC1A, likely because of its ability to sense CO₂-induced acidosis, including acid- and CO₂-evoked freezing, CO₂ aversion, CO₂ conditioned place avoidance, and CO₂-enhanced center avoidance (Ziemann et al., 2009; Taugher et al., 2014). Therefore, we hypothesized that any effects of post-acquisition CO₂ inhalation would critically depend on ASIC1A.

Because inhaling 10% CO₂ during the labile period induced by cued fear retrieval further enhanced memory lability (Du et al., 2017), we hypothesized that exposing mice to 10% CO₂ 1 h after acquisition, during the time period in which new memories are thought to be labile (Mednick et al., 2011), would also alter fear memory. Thus, we trained mice in a cued fear conditioning paradigm in which we paired an auditory cue with a foot shock and 1 h later mice were placed in a different context and exposed mice to air or 10% CO₂. 24 h after acquisition, mice underwent cued testing in which they were presented with the auditory tone in a novel context in the absence of foot shocks and conditioned freezing was assessed (Figure 1A). Prior to gas exposure, the air and 10% CO₂-treated groups exhibited similar levels of freezing during cued fear acquisition (Figure 1B). Interestingly, the mice that had been exposed to 10% CO₂ displayed a marked increase in freezing during cued testing as compared to air-exposed controls (Figure 1C). This suggests that CO₂ exposure during the time period following acquisition can potentiate fear memory.

To rule out the possibility that CO₂ exposure might have an effect on freezing that was independent of cued fear conditioning, we performed a no shock control experiment. The same cued fear acquisition paradigm (Figure 1A) was used, except that foot shocks were omitted from the protocol. This resulted in minimal freezing during both acquisition and cued fear testing, with 10% CO₂ treatment having no effect on freezing (Figures 1B,C). This suggests that the effect of CO₂ on cued fear memory depends on cued fear conditioning, rather than being a non-specific effect of CO₂ exposure.

Because exposure to 10% CO₂ is aversive and may induce stress (Ziemann et al., 2009; Taugher et al., 2014; Spiacci et al., 2018), we next sought to test if exposing mice to a different stressor would recapitulate the effects of CO₂ inhalation on cued fear memory. We chose restraint stress given its widespread use as a method for inducing stress in mice (Stewart et al., 2008; Roper et al., 2010). We trained mice in a cued fear conditioning paradigm and 1 h later mice were either placed in a restraint tube for 30 min or left in the home cage (Figure 1D). Both treatment groups displayed a similar level of freezing during acquisition (Figure 1E), and restraint stress did not alter cued fear memory (Figure 1F). This suggests that the effects of 10% CO₂ exposure might not be the result of its ability to induce stress, but instead the result of its other attributes, such as its ability to induce acidosis.

Next, we sought to determine if this effect of 10% CO₂ on cued fear memory was time-dependent. First, we administered air or 10% CO₂ 4 h after cued fear conditioning (Figure 2A). Prior to gas exposure, the air and 10% CO₂-treated groups exhibited similar levels of freezing during acquisition (Figure 2B). Mice treated with CO₂ 4 h after acquisition displayed much more freezing during cued testing than air-treated controls (Figure 2C), similar to when CO₂ was given 1 h after acquisition (Figure 1C). To determine if this effect would extend to later time points, we administered air or 10% CO₂ 24 h after cued fear conditioning and tested cued fear memory 48 h after acquisition (Figure 2D). Prior to gas exposure, the air and 10% CO₂-treated groups exhibited similar levels of freezing during acquisition (Figure 2E). No effect of gas exposure was seen during cued testing (Figure 2F), suggesting that there is a critical time period after acquisition for 10% CO₂ inhalation to affect cued fear memory. Next, we sought to determine if CO₂ exposure prior to acquisition would have a similar effect on cued fear memory as CO₂ exposure after acquisition. Thus, we administered air or 10% CO₂ 4 h before cued fear conditioning and assessed cued fear memory 24 h later (Figure 2G). 10% CO₂ pre-treatment did not affect freezing during acquisition (Figure 2H) or cued testing (Figure 2I), suggesting that it might be necessary for 10% CO₂ exposure to occur after acquisition in order to potentiate fear memory.

We next sought to determine if the effect CO₂ on fear memory was specific to cued fear conditioning. Thus, we performed context fear conditioning, a form of Pavlovian conditioning in which an association is made between a neutral context and a series of foot shocks. 4 h after context fear acquisition, we administered air or 10% CO₂, and 24 h after acquisition, we returned mice to the acquisition context and assessed conditioned freezing responses (Figure 3A). Prior to gas exposure, the air and 10% CO₂-treated groups exhibited similar levels of freezing during acquisition (Figure 3B). Similar to cued fear conditioning (Figure 2C), mice treated with 10% CO₂ 4 h after context fear acquisition exhibited more freezing during context testing than air-treated controls (Figure 3C). This suggests that CO₂ following acquisition has similar effects on cued and context fear memory.

To test if CO₂ could also potentiate other types of memory, we tested its effects on NOR. Mice acquired a memory of the familiar object by exploring two identical objects. 4 h later, we exposed mice to air or 10% CO₂. 24 h after acquisition, we assessed NOR by replacing one of the familiar objects with a novel object and assessing the amount of time that mice spent interacting with the novel and familiar objects (Figure 3D). During acquisition, mice exhibited similar levels of interaction with each object, and no difference between treatment groups was observed (Figure 3E). Surprisingly, when NOR was assessed 24 h later, we found that 10% CO₂ treatment had decreased NOR as compared to air-treated controls and also caused a failure to discriminate between the novel and familiar objects (Figure 3F). This suggests that CO₂ may affect different types of memory in distinct ways.

To explore the molecular underpinnings of these effects, we tested if they depended on ASIC1A. ASIC1A is a proton-gated cation channel that is known to mediate other effects of CO₂, likely via its ability to sense acidosis induced by CO₂ inhalation (Ziemann et al., 2009; Taugher et al., 2014). Thus, we tested the effects of 10% CO₂ exposure on memory in Asic1a^–/– mice under the same conditions in which CO₂ enhanced fear memory in wild-type mice. Neither administering 10% CO₂ 1 h after (Figures 4A–C) nor 4 h after (Figures 4D–F) cued fear acquisition altered freezing during cued testing in Asic1a^–/– mice. Similarly, we found no effect of 10% CO₂ administered 4 h after context fear acquisition on freezing in Asic1a^–/– mice during context testing (Figures 4G–I). Together, these observations suggest that ASIC1A is critical for the effects of CO₂ exposure on fear memory observed in wild-type mice.

Next, we tested whether the reduction in NOR induced by CO₂ exposure similarly depended on ASIC1A (Figure 4J). Prior to gas exposure, Asic1a^–/– mice exhibited similar levels of interaction with each object (Figure 4K). When gas was administered 4 h after acquisition, novel object preference was intact in air-treated Asic1a^–/– mice, but not in those exposed to 10% CO₂ (Figure 4L). However, the difference between air versus CO₂-exposed Asic1a^–/– mice did not reach significance (p = 0.07). These results are similar to those seen in wild-type mice, and suggest that in contrast to fear memory, the effects of CO₂ exposure on NOR may be independent of ASIC1A.

Our results identify a novel effect of post-acquisition CO₂ exposure on both cued and context fear memory. We found that during a discrete time period, including 1 and 4 h after fear acquisition, but not 24 h after, 10% CO₂ inhalation increased fear memory. In contrast, there was no effect of CO₂ exposure under the same conditions in Asic1a^–/– mice, suggesting that ASIC1A is critical for this effect. Moreover, CO₂ may affect distinct types of memory in different ways, as in contrast to these enhancements in fear memory, NOR was decreased by post-training CO₂ inhalation.

These observations raise questions about where and how CO₂ and ASIC1A act to produce their effects on fear memory. The amygdala is a likely site of ASIC1A action. ASIC1A is abundantly expressed in the amygdala (Wemmie et al., 2003; Coryell et al., 2007; Price et al., 2014; Chiang et al., 2015), and manipulating ASIC1A in the amygdala alters cued and context fear memory (Coryell et al., 2008; Chiang et al., 2015), as well as CO₂-evoked freezing (Ziemann et al., 2009). Likewise, localized acidosis in the amygdala promotes freezing (Ziemann et al., 2009) and plasticity (Du et al., 2017). However, ASIC1A is also expressed in several other fear circuit structures in which it could influence fear memory, such as the bed nucleus of the stria terminalis, periaqueductal gray, hippocampus, prefrontal, and cingulate cortices (Wemmie et al., 2003; Coryell et al., 2007; Price et al., 2014). On the cellular level, these effects are likely due to ASIC1A in neurons. Neurons are a critical site of ASIC1A action in cued and context fear memory and CO₂-induced freezing (Taugher et al., 2017), though it is possible that ASIC1A in non-neuronal cells could also be involved.

The acidosis caused by CO₂ inhalation may activate ASIC1A directly (Ziemann et al., 2009). CO₂ may also influence the activation of ASIC1A by protons released during neurotransmission (Du et al., 2014; Kreple et al., 2014; Gonzalez-Inchauspe et al., 2017). Either or both of these mechanisms might contribute to the effects of CO₂ and ASIC1A observed here. Interestingly, acidosis not only activates these channels, it also inhibits them through inactivation and steady-state desensitization (Waldmann et al., 1997; Sherwood et al., 2012). These inhibitory effects of acid exposure have been demonstrated in vitro, although it is not clear to what degree they occur in vivo. Steady-state desensitization can be limited by endogenous peptides such as dynorphins and RF-amides (Sherwood and Askwith, 2008, 2009), raising the possibility that densitization in vivo may be less pronounced. It is also conceivable that the ASIC1A effects observed here may not require sustained channel activity.

The mechanisms downstream of ASIC1A are not yet clear. We speculate that the effects of CO₂ and ASIC1A on fear memory may depend on CP-AMPARs in the postsynaptic membrane of lateral amygdala neurons, and or CREB-mediated transcriptional regulation. Conversion of short term fear memory to long lasting memory (i.e., consolidation) has been previously found to require both CP-AMPAR insertion (Rumpel et al., 2005; Nedelescu et al., 2010), and CREB phosphorylation and subsequent changes in gene transcription (Alberini, 2009). Previously, it was shown that when coupled with memory reactivation, CO₂ exposure also increased both insertion of CP-AMPARs and CREB phosphorylation, which may be critical for promoting susceptibility of the fear memory to subsequent modification (Du et al., 2017). The potential role of ASIC1A in CREB phosphorylation has not been tested. However, the increase in CP-AMPAR insertion critically depended on the presence of ASIC1A (Du et al., 2017). Thus, CP-AMPAR insertion and CREB dependent transcription would seem good candidates for mediating the fear memory effects of CO₂ and ASIC1A reported here.

Besides potentiating fear memory, our results suggests that CO₂ may decrease other types of memory. 10% CO₂ administered 4 h after acquisition attenuated NOR memory (Figure 4F). This may reflect the partially overlapping circuitry of these behaviors (Kim and Jung, 2006; Warburton and Brown, 2015; Fenster et al., 2018) or distinct molecular mechanisms being at play. Interestingly, whereas 10% CO₂ had no effect on cued and context fear memory in Asic1a^–/– mice, there was a strong trend toward a decrease in NOR in Asic1a^–/– mice, suggesting that ASIC1A might mediate the fear conditioning, but not object recognition memory. Although additional work will be needed to identify the pH-sensitive molecules involved in the CO₂ effect on novel object recognition, there are numerous pH-sensitive ion channels and receptors at synapses that could potentially be involved (Wemmie, 2011). Additional investigation will be needed to explore the effects of CO₂ inhalation on other types of memory and other receptors.

Together, these observations suggest that brain pH around the time of a traumatic event may impact the associated fear memory, potentially altering the risk of developing PTSD. Consistent with this others have observed elevated levels of PTSD in patients with conditions that may result in a chronic respiratory acidosis such as asthma (O’Toole and Catts, 2008; Allgire et al., 2021), chronic obstructive pulmonary disease (Teixeira et al., 2015), obstructive sleep apnea (Colvonen et al., 2015), COVID-19 (Mertz Schou et al., 2021), and patients who are mechanically ventilated for a prolonged period of time (Wade et al., 2013). This suggests that pH alterations such as respiratory or metabolic acidosis or changes in acid-base equilibrium might contribute to PTSD risk and raises the exciting possibility that preventing acidosis or inducing alkalosis after a traumatic event could attenuate the associated fear memory.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was reviewed and approved by The University of Iowa Animal Care and Use Committee.

RT conceptualized experiments, collected and analyzed data, and wrote the manuscript. AW conceptualized experiments, collected and analyzed data, and revised the manuscript. GW collected data. AC, BD, and JW conceptualized experiments and revised the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.