- Cyld knockout impairs amygdala-dependent tone-cued fear memory in mice.

- Cyld^–/– mice display aberrant neuronal activation in BLA in response to tone-cued fear test.

- Cyld^–/– mice show decreased excitability of BLA principal neurons.

- CYLD is critical for both excitatory and inhibitory synaptic neurotransmission in mouse BLA.

Learning to identify and respond to threats in the environment is essential for animal survival. Abnormal fear memory is a hallmark of many neuropsychiatric disorders, including Alzheimer’s disease (Varinthra et al., 2021), obsessive-compulsive disorder (Nielen et al., 2009), posttraumatic stress disorder (Bowers and Ressler, 2015), and drug addiction (Taylor and Torregrossa, 2015). In mammals, it is well established that the basolateral amygdala (BLA) plays a central role in the formation and retention of fear memory (Davis and Whalen, 2001; Janak and Tye, 2015; Marek et al., 2019). This form of learning can be achieved with classical auditory fear conditioning by pairing a neutral auditory stimulus conditioned stimulus (CS) with an aversive footshock unconditioned stimulus (US) (Duvarci and Pare, 2014). Broadly speaking, the BLA receives two streams of auditory innervation from the auditory thalamus and auditory cortex and sends projections to the central amygdala (CeA) and eventually to the brainstem and hypothalamus to elicit defensive behavior (freezing) and autonomic and endocrine responses (Johansen et al., 2011).

The BLA forms a cortical-like temporal lobe structure that contains around 85–90% excitatory pyramidal neurons (principal neurons) and 10–15% GABAergic interneurons (Shao et al., 2019). A subset of excitatory neurons can be activated by increased expression of the immediate early gene, c-Fos (Tronson et al., 2012). Defective neuronal activation in the BLA of mice subjected to behavioral stimulation is associated with impaired fear memory (Huang et al., 2014). In addition, numerous studies have revealed a crucial role for the excitation of BLA principal neurons in fear memory (Yiu et al., 2014; Polepalli et al., 2020). However, the underlying mechanisms by which neuronal activation and excitability in the BLA are regulated remain elusive.

Cylindromatosis is a tumor-suppressor protein whose gene is mutated in familial cylindromatosis; it serves as a negative regulator of NF-κB signaling (Ma et al., 2017) and is involved in the regulation of important physiological processes such as the immune response, cell cycle, and cell migration (Li et al., 2019). The BioGPS database¹ shows that Cyld (NCBI gene ID: 74256) is highly expressed in the striatum followed by the nucleus accumbens (NAc), amygdala, hypothalamus, prefrontal cortex (PFC), and hippocampus (Probeset numbers: 1429201 and 1429618). Interestingly, according to the in situ hybridization data of the Allen Brain Atlas database (experiment 73992938), the protein CYLD mainly distributes in the posterior portion of the BLA². Recent studies have identified CYLD as a protein component of purified rodent postsynaptic density (PSD) (Ma et al., 2017) and a regulator of neuronal cell death (Ganjam et al., 2018), neuronal dendrite morphogenesis, and spine formation (Li et al., 2019). CYLD in the PSD is recruited and further activated in a Ca²⁺-dependent manner by Ca²⁺/calmodulin-dependent protein kinase II to the PSD following high K⁺ or N-methyl-D-aspartic acid (NMDA) stimulation (Zajicek and Yao, 2021). According to recent reports, Cyld is a causative gene for frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (Dobson-Stone et al., 2020), and patients carrying Cyld variants show severe memory loss (Tábuas-Pereira et al., 2020). Several studies have shed light on the role of deubiquitinases (DUBs), such as ubiquitin-specific protease (USP) and ubiquitin carboxyl-terminal hydrolase L1 (UCHL1), in the modulation of synaptic plasticity and memory processes (Imai et al., 2013; Jarome et al., 2013; Zhang et al., 2014; Zeng et al., 2019; Srikanta et al., 2021). Nevertheless, the functions of CYLD in neuronal activity and neural circuits remain largely unknown.

In this study, by using a combination of electrophysiological, immunohistochemistry, and behavioral measurements, we reveal that Cyld^–/– mice exhibit abnormal amygdala-dependent tone-cued fear memory. In addition, Cyld^–/– mice also display aberrant neuronal activation and a decreased excitability of excitatory pyramidal neurons in BLA. What is more, Cyld^–/– mice exhibit a decrease in both the frequency of spontaneous excitatory postsynaptic currents (sEPSCs) and the amplitude of miniature excitatory postsynaptic currents (mEPSCs) of principal neurons, but an increase in the frequency of miniature inhibitory postsynaptic currents (mIPSCs) and the number of parvalbumin⁺ (PV⁺) interneurons in the BLA. Collectively, these findings suggest that CYLD plays an important role in the regulation of fear memory and acts as a regulator of neuronal activation, neuronal excitability, synaptic transmission and the number of PV⁺ interneuron in the BLA of mice.

Here, for the first time, we show that the K63-deubiquitinase (DUB), CYLD, plays an important role in modulating fear memory. The gene that encodes CYLD is highly expressed in the striatum followed by the NAc, amygdala, hypothalamus, PFC, and hippocampus (see text footnote 1). CYLD has been studied extensively for its role in negatively regulating NF-κB signaling (Lork et al., 2017), but its function in mammalian brain remains largely unknown. Previous proteomic analyses using mouse or rat brain samples indicated the presence of CYLD in purified PSD (Li et al., 2019). Other studies revealed that CYLD regulates dendritic growth and postsynaptic spine maturation (Ma et al., 2017; Li et al., 2019). What these findings imply is that CYLD may regulate learning and memory of mammalian brain.

Appropriate fear memory in response to potential danger in the environment is crucial for animal survival, while abnormal fear memory is implicated in several mental disorders (Nielen et al., 2009; Bowers and Ressler, 2015; Taylor and Torregrossa, 2015; Varinthra et al., 2021). The BLA is an obvious region of interest in view of its central role in the fear circuitry, mediating both fear learning and its expression (Luyck et al., 2020). Our TFC test revealed that Cyld knockout impairs fear memory in mice. Recent studies in which researchers performed whole-exome sequencing in FTD patients also showed that patients carrying Cyld mutations suffer severe memory impairment (Dobson-Stone et al., 2020; Tábuas-Pereira et al., 2020). In addition, numerous studies have revealed the role of other DUBs (USP2, USP6, USP14, USP46, and UCHL1) in promoting learning and memory (recognition memory, fear memory, and spatial memory) in rodents (Imai et al., 2013; Jarome et al., 2013; Zhang et al., 2014; Zeng et al., 2019; Srikanta et al., 2021). These reports indicate that DUBs including CYLD are involved in fear memory modulation.

There are two necessary detection units for tone-cued fear conditioning: an auditory sensation-based cue and noxious cutaneous stimuli (Silva et al., 2016). Thus, the auditory sensation and nociception of mice are essential physiological factors to form the fear memory. In our study, all mice showed a similar response to the presentation of the tone during the training period, indicating that Cyld knockout mice did not experience tone-cued fear conditioning-related sensory encoding dysfunctions. We also consider other factors that may affect the fear behavior. A previous report showed that Cyld^–/– mice exhibited normal motor function and did not show depression-like behavior (Han et al., 2020). However, anxiety-like behavior was found in Cyld^–/– mice (Han et al., 2020). In the current study, Cyld^–/– and Cyld^+/+ mice showed a similar freezing response during the habituation and pre-tone period, thus excluding the influence of anxiety. Therefore, the low freezing level that we observed in Cyld^–/– mice during the tone-cued test could be regarded as a fear memory deficit.

The formation of fear memory comprises three phases: acquisition [short-term memory (STM)], consolidation (STM is stabilized into a persistent long-term memory), and reconsolidation (long-term memory) (Johansen et al., 2011). Interestingly, we found that Cyld knockout had no effect on the acquisition phase of fear memory. One likely reason could be that plasticity is important for immediate learning and that STM is mediated by covalent modification of existing synaptic proteins (for example, by phosphorylating glutamate receptors) (Johansen et al., 2011). In contrast, consolidation and reconsolidation of this plasticity are generally thought to occur via activation of second messengers that initiate gene transcription and translation of new proteins like activity-regulated cytoskeletal-associated proteins and early growth response protein 1 (Johansen et al., 2011). CYLD may not affect existing synaptic proteins in the short-term, but instead may play an important role in long-term modulation of consolidation- and reconsolidation-related protein synthesis.

c-Fos is extensively used as a marker of neuronal activation in the BLA during the response to external stimuli such as fear learning, fear conditioning with an auditory cue, and memory retrieval (Knapska et al., 2007). Previous reports showed increased expression of c-Fos in the BLA in response to fear conditioning (Rajbhandari et al., 2016; Sun et al., 2018; Silva et al., 2019). Here, we found that the number of c-Fos⁺ neurons in response to the tone-cued test is reduced in the BLA of Cyld^–/– mice. However, the number of c-Fos⁺ neurons of Cyld^+/+ mice was comparable to that of Cyld^–/– mice, indicating an effect of CYLD on fear memory-related neuronal activation. Interestingly, statistical differences of the number of c-Fos⁺ neurons between Cyld^+/+ and Cyld^–/– mice were not found in the tone-only test or pre-tone stage. Together with the behavioral results, our study shows that CYLD is critical for mice to associate the CS and US during the fear memory retrieval. Generally, fear memory impairment may associate with the aberrant neuronal activation in the amygdala of mice (Huang et al., 2014; Sun et al., 2018). A reduction in neuronal activation upon behavioral stimulation could be caused by decreased intrinsic excitability of neurons, depressed excitatory synaptic activity, or increased inhibitory synaptic activity (Sun et al., 2018). It is well established that excitation of principal neurons in the BLA is a critical step in auditory fear conditioning (Johansen et al., 2010; Yiu et al., 2014; Polepalli et al., 2020). Coincident with the fear memory impairment and aberrant neuronal activation we observed, there was also an increase in the threshold and rheobase of APs and a decrease in the AP firing rate of Cyld^–/– mice, which suggests reduced neuronal excitability in BLA principal neurons. Another study also showed that the cellular mechanisms underlying changes in fear memory involve modulation of intrinsic excitability of neurons (Zhang et al., 2017).

Among K63-DUBs localized in the PSD, CYLD emerges as a central player (Zajicek and Yao, 2021). The abundance of CYLD in the PSD suggests that it may be the primary DUB for multiple proteins involved in synapse development, function, and plasticity, as well as related pathologies (Ma et al., 2017). Therefore, we examined the role of CYLD in modulating synaptic transmission in the BLA to further explore the possible cellular mechanisms of aberrant neuronal activation in Cyld^–/– mice. In line with the reduced AP firing rate, we observed a significant decrease in the sEPSC frequency of principal neurons in the BLA of Cyld^–/– mice. In addition, Cyld knockout reduces the mEPSC amplitude of these principal neurons. Our results suggest that CYLD plays an important role in modulating excitatory synaptic transmission and the levels of postsynaptic α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptors (AMPARs) or NMDA receptors. K63-linked ubiquitination chains are considered the primary non-degradable linkage, and these chains can also remodel the surface of substrate proteins, akin to phosphorylation, to directly regulate protein–protein interactions (Ma et al., 2017). Therefore, CYLD may be involved in the synthesis, turnover, and trafficking of proteins in the PSD. These functions of CYLD were implied in a study on another K63-DUB, USP46, which can cleave K63-linked ubiquitination chains from AMPARs and modulate AMPAR internalization and turnover in cultured rat cortical and hippocampal neurons; in this way, USP46 is thought to regulate AMPAR-mediated excitatory synaptic transmission (Huo et al., 2015). A recent study showed that CYLD can cleave K63-linked ubiquitination chains from PSD-95 (Ma et al., 2017), which represents the first direct evidence that CYLD modulates postsynaptic scaffolding and remodeling in excitatory synapses.

In view of the impaired excitatory synaptic transmission in the BLA of Cyld^–/– mice, we also recorded mIPSCs in BLA principal neurons. It turned out that Cyld mutation increased mIPSC frequency, but had no effect on mIPSC amplitude. This result implies the possibility of a presynaptic change in the release of inhibitory neurotransmitters or the number of GABAergic interneurons. A large body of evidence suggested that the formation and expression of conditioned fear memories also entails long-lasting functional and structural plasticity of GABAergic synapses onto pyramidal neurons of the murine BLA (Kasugai et al., 2019). Available data support the contention that PV⁺ interneurons play an important role in modulating the balance of excitation and inhibition in local circuits and they are thus involved in the regulation of fear memory (Hu et al., 2014; Lucas and Clem, 2018; Wang et al., 2020). In addition, PV⁺ interneurons form synaptic contacts predominantly onto the soma as well as proximal axon and dendrites of the principal neurons in BLA which allow them to modulate the rate and timing of the principal neuron spike initiation (Woodruff and Sah, 2007; Hu et al., 2014; Lucas and Clem, 2018). Hence, we examined PV⁺ interneurons, finding that the number of BLA PV⁺ interneurons was significantly higher in Cyld^–/– mice, consistent with excessive local circuit inhibition. This result suggests that increased number of PV⁺ interneurons in BLA of Cyld^–/– mice may disrupt the balance of excitation and inhibition in local circuits which could be associated with the fear memory deficit observed in Cyld mutant mice. A previous study also showed that Cyld mutation causes an excess of striatal GABAergic inhibition (Zhang et al., 2016). In addition, loss of the DUB, USP14, reduced neurotransmitter release and increased surface expression of the GABA_AR (Jarome et al., 2013). Taken together, these data suggest that CYLD is also crucial for maintaining inhibitory synaptic transmission in the BLA.

Research on CYLD is still at an early stage. Recent studies have focused on the role of CYLD in regulating synaptic plasticity. The abundance of CYLD in the PSD suggests that it may be the primary DUB for multiple proteins (Ma et al., 2017). More studies are needed to identify and characterize these target proteins, because this could provide insights into the mechanisms of synapse function and plasticity. Our study indicates that CYLD may provide a novel target for therapeutic interventions for various fear memory-associated brain disorders. Except for being expressed in the BLA, Cyld is also expressed in several brain regions like the striatum, NAc, CeA, and PFC (see text footnote 1), which are also involved in fear memory. It has long been known that the PFC and CeA are tightly involved in tone-cued fear memory (Ferreira et al., 2008; Tronson et al., 2012; Duvarci and Pare, 2014; Marek et al., 2019). In addition, a previous study showed that the flow of information between the basal amygdala and the NAc is necessary for signaled active avoidance behavior (Ramirez et al., 2015). It is worth noting that the indirect CeA–dorsal striatum pathway mediates tone-cued fear memory (Ferreira et al., 2008). Thus, we cannot exclude the possibility that CYLD deficiency in brain regions mentioned above are the source of the impaired fear memory of Cyld^–/– mice. Specifically ablating the Cyld of the principal neurons in BLA and observing the consequent fear memory, as well as the electrophysiological outcomes, are required to establish the precise link between CYLD in BLA and tone-cued fear memory. In addition, chemogenetically operating the principal neurons in BLA of conditional Cyld knockout mice would help to further elucidate the linkage mentioned above.

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 Ethics Committee of Animal Research of South China Normal University.

H-dL designed and performed electrophysiological, immunostaining, and behavioral experiments, analyzed the data, and wrote the manuscript. D-nL performed part of the fear conditioning behavioral test, immunostaining experiments, and data analysis. LY provided suggestions on experimental design and improvement of the manuscript. CL supervised the study, designed the experiments, and made critical revision of 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.