*Editorial* **Molecular Mechanisms of Synaptic Plasticity 2.0: Dynamic Changes in Neurons Functions, Physiological and Pathological Process**

**Giuseppina Martella 1,2**


Due to the success of the first Special Issue on synaptic plasticity, I endeavored to promote a new Special Issue with an emphasis on dynamic changes in neuronal functions and physiological and pathological processes.

I have endorsed this Special Issue with the aim of collecting scientific research capable of intensifying readers' interest in phenomena related to synaptic plasticity.

Synaptic plasticity is a crucial molecular mechanism whose actions are carried out from the developmental period to old age. Described briefly, our brain is subject to structural and functional modifications in response to external stimuli. Synaptic plasticity phenomena include macroscopic changes and cortical remapping in response to injury, microscopic changes, and spine pruning [1,2]. The emergency resulting from the increase in neuropsychiatric and neurological disorders over the last few years has stressed the urgency of understanding the aberrant processes connected to synaptic plasticity failure [3–5].

The authors invited to contribute to this Special Issue have provided important contributions consisting of the translational and clinical studies that I am pleased to promote in these few lines. This Special Issue contains two original clinical articles, two literature reviews, and five original translation articles. I hope that I have provided you with an extensive thematic collection and that you enjoy your reading.

Patnaik and co-workers examined 29 patients with certified bipolar disorder, 32 patients affected by cerebellar neurodegenerative pathologies, and 37 healthy subjects using the 3T-MRI technique in order to determine the similarities and differences in cerebellar grey matter loss. They found a pattern of grey matter cerebellar alterations in both the bipolar and cerebellar groups that involved the anterior and posterior cerebellar regions, demonstrating the involvement of the cerebellum in the synaptic plasticity of patients with bipolar disorder [6].

In multiple sclerosis, inflammation can modify synaptic transmission and plasticity. Professor Centonze's group explored the influence of proinflammatory cytokines on associative Hebbian synaptic plasticity. In their cohort, they found that IL-1β levels were associated with synaptic hyperexcitability and inversely related to LTP-like synaptic plasticity.

These findings support the evidence that anti-IL-1β drugs represent a new potential therapeutic target. Considering the identification of IL-1β as a marker of inflammatory synaptopathy, antagonistic drugs may also represent a specific target during the different phases of the progression of multiple sclerosis [7].

In their systematic review, Eltokhi and coworkers underline the correlation between neuropsychiatric disorders and deficits in the glutamatergic system, and they also consider the psychiatric features of neurodevelopmental disorder as well as autism. Alterations in synaptic plasticity, accompanied by structural modifications of excitatory synapses, were observed in schizophrenia and autism spectrum disorders using EM-imaging methods. In addition, it was revealed that the expression of glutamatergic receptors is differentially

**Citation:** Martella, G. Molecular Mechanisms of Synaptic Plasticity 2.0: Dynamic Changes in Neurons Functions, Physiological and Pathological Process. *Int. J. Mol. Sci.* **2023**, *24*, 12685. https://doi.org/ 10.3390/ijms241612685

Received: 2 August 2023 Accepted: 7 August 2023 Published: 11 August 2023

**Copyright:** © 2023 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

affected in various brain regions, thus revealing an undeniable link between altered synaptic plasticity and psychiatric illness [8].

D'Angelo and colleagues investigated the A2A receptor's role in different areas inside of a DYT1 mouse model of dystonia. They showed that A2A receptors are significantly improved in the striatal and globus pallidus nuclei and reduced in the entopeduncular nucleus. These opposite modifications may suggest that the pathophysiology of dystonia is correlated with an imbalance of the direct or indirect pathway [9].

The motor thalamus (MTh) is involved in the basal ganglia cortical loop and acts on the codifying of motor information. Di Giovanni and his co-workers showed for the first time that acute dopamine depletion caused by tetrodotoxin (TTX) results in an increase in GABA concentration in the MTh along with unchanged glutamate levels. Chronic dopamine denervation via 6-hydroxydopamine (6-OHDA) in anesthetized rats affects the coupling of MTh cortical activity in relation to the TTX-induced acute depletion state. The authors' findings demonstrate that the MTh, among other areas in the basal ganglia, is influenced by DA chronic deprivation and makes alterations in the basal ganglia network in relation to counterbalancing the profound alteration arising after the onset of the acute DA depletion state [10].

La Recchiuta aimed to emphasize the role of the amygdala and medial prefrontal cortex in functional plasticity and synaptic wiring in conditions of fear extinction. Her results demonstrated that the optogenetic activation of pyramidal neurons in mice conditioned by induced fear extinction deficits causes an increase in cellular excitability, excitatory neurotransmission, and spinogenesis and is also associated with modifications of the transcriptome of amygdala pyramidal neurons [11].

The brain-derived neurotrophic factor (BDNF) drives brain development and maturation. Altered BDNF levels have been observed in many neurological diseases to such an extent that new therapeutic strategies are being developed to increase the level of BDNF.

Fingolimod-phosphate (FTY720-P) can modulate BDNF levels. However, the mechanisms by which the FTY720-P operates are still unclear. Patnaik and colleagues have shown that FTY720-P can regulate dendritic architecture, increase dendritic spine density, and modify the morphology of mature primary hippocampal neuron cultures. This study confirms that BDNF-dependent therapies may represent a new goal for many neurological diseases [12].

The Ghiglieri group, on the other hand, carried out scientific work demonstrating that food restriction can improve the lifespan of different species.

According to their analyses, food is a natural reward, and where this reformulation is restricted, some neuroadaptive responses are necessary to maintain physiological homeostasis.

They examined the AMPA receptor subunit composition in the dorsal striatal neurons of mice that had been acutely food deprived and showed that even moderate food deprivation in experimental animal models reflects a series of neuroadaptations and the remodeling of striatal synaptic plasticity [13].

Stroke is a great enemy of modern medicine. Usually, deficiencies of the neurovascular unit caused by reperfusion lesions, or inflammatory processes, constitute the main field of study. In their literature review, De Luca and collaborators provided a road map that could help to improve both therapy and rehabilitation through the knowledge of translational studies. They also suggest that further research should involve the cellular capacity to avoid neuroinflammatory phenomena and the capacity of cells during reperfusion to actively reshape the matrix [14].

**Conflicts of Interest:** The author declares no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

## *Article* **miRNA-132/212 Deficiency Disrupts Selective Corticosterone Modulation of Dorsal vs. Ventral Hippocampal Metaplasticity**

**Shima Kouhnavardi <sup>1</sup> , Maureen Cabatic 1,†, M. Carmen Mañas-Padilla 2,†, Marife-Astrid Malabanan <sup>1</sup> , Tarik Smani <sup>3</sup> , Ana Cicvaric 1,4 , Edison Alejandro Muñoz Aranzalez <sup>1</sup> , Xaver Koenig <sup>1</sup> , Ernst Urban <sup>5</sup> , Gert Lubec <sup>6</sup> , Estela Castilla-Ortega <sup>2</sup> and Francisco J. Monje 1,\***


**Abstract:** Cortisol is a potent human steroid hormone that plays key roles in the central nervous system, influencing processes such as brain neuronal synaptic plasticity and regulating the expression of emotional and behavioral responses. The relevance of cortisol stands out in the disease, as its dysregulation is associated with debilitating conditions such as Alzheimer's Disease, chronic stress, anxiety and depression. Among other brain regions, cortisol importantly influences the function of the hippocampus, a structure central for memory and emotional information processing. The mechanisms fine-tuning the different synaptic responses of the hippocampus to steroid hormone signaling remain, however, poorly understood. Using ex vivo electrophysiology and wild type (WT) and miR-132/miR-212 microRNAs knockout (miRNA-132/212−/−) mice, we examined the effects of corticosterone (the rodent's equivalent to cortisol in humans) on the synaptic properties of the dorsal and ventral hippocampus. In WT mice, corticosterone predominantly inhibited metaplasticity in the dorsal WT hippocampi, whereas it significantly dysregulated both synaptic transmission and metaplasticity at dorsal and ventral regions of miR–132/212−/<sup>−</sup> hippocampi. Western blotting further revealed significantly augmented levels of endogenous CREB and a significant CREB reduction in response to corticosterone only in miR–132/212−/<sup>−</sup> hippocampi. Sirt1 levels were also endogenously enhanced in the miR–132/212−/<sup>−</sup> hippocampi but unaltered by corticosterone, whereas the levels of phospo-MSK1 were only reduced by corticosterone in WT, not in miR–132/212−/<sup>−</sup> hippocampi. In behavioral studies using the elevated plus maze, miRNA-132/212−/<sup>−</sup> mice further showed reduced anxiety-like behavior. These observations propose miRNA-132/212 as potential region-selective regulators of the effects of steroid hormones on hippocampal functions, thus likely fine-tuning hippocampus-dependent memory and emotional processing.

**Keywords:** microRNA; miR-132/212; synaptic plasticity; dorsal hippocampus; ventral hippocampus; corticosterone; emotional behavior; anxiety-like behavior

### **1. Introduction**

The steroid hormone cortisol (also known as hydrocortisone) is a highly potent human hormone produced by the adrenal glands and whose secretion into the blood stream is triggered by the corticotropin-releasing hormone originating from the hypothalamus in the brain. Cortisol is central for the regulation of many biological functions, including the modulation of the body responses to stress, the reduction of inflammatory processes, the

**Citation:** Kouhnavardi, S.; Cabatic, M.; Mañas-Padilla, M.C.; Malabanan, M.-A.; Smani, T.; Cicvaric, A.; Muñoz Aranzalez, E.A.; Koenig, X.; Urban, E.; Lubec, G.; et al. miRNA-132/212 Deficiency Disrupts Selective Corticosterone Modulation of Dorsal vs. Ventral Hippocampal Metaplasticity. *Int. J. Mol. Sci.* **2023**, *24*, 9565. https://doi.org/10.3390/ ijms24119565

Academic Editor: Giuseppina Martella

Received: 8 May 2023 Revised: 21 May 2023 Accepted: 25 May 2023 Published: 31 May 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

regulation of blood pressure and metabolic functions, and the influences on the immune system and circadian sleep-wake cycles [1,2]. Cortisol is also a powerful regulator of brain function, influencing learning, memory and emotional behaviors in both humans and animals [2–5]. Moreover, dysregulated cortisol levels have been identified during the onset and progression of a variety of conditions associated with memory decline and emotional dysregulation, including post-traumatic stress disorder, depression, stress and anxiety [6–15]. Several studies have also shown that altered levels of corticosterone (the molecular counterpart of cortisol in mice and rats) result in profound changes in neuronal synaptic plasticity as well as altered memory-related emotional behaviors [16–22]. The molecular mechanisms by which cortisol can distinctly affect the different regions of the brain to influence memory and emotion-related behaviors remain, however, poorly understood.

The hippocampus, a brain structure critical for learning and memory processing [23–26], is involved in the regulation of emotional memories [27–29] and is importantly associated with the modulation of the hypothalamic–pituitary–adrenal axis function in health and disease [30–33]. Interestingly, exposure to early-life stress has been shown to correlate with a reduction in the anatomical features of hippocampal structures as well as alterations in cortisol levels [30,33]. Moreover, while the hippocampus is the brain region with perhaps the highest concentration of receptor target sites for adrenocortical steroids [29], and whereas several studies have shown that corticosterone can directly influence hippocampal synaptic potentiation [34,35], little continues to be known about how cortisol differentially affects the different regions of the hippocampus to influence memory-related emotional behaviors [28]. Here, we hypothesized that microRNAs (miRNAs) are key molecular elements participating in the modulation of the region-specific effects of glucocorticoid hormones on hippocampal synaptic transmission and metaplasticity (the newly generated synaptic changes that happen after synaptic plasticity has occurred [36–38]).

miRNAs, are non-encoding short transcript RNAs that participate in the posttranscriptional modulation of gene expression [39,40]; that serve in a variety of pivotal functions of the central nervous systems, including the modulation of synaptic activity [41–46]; and which are importantly involved in the brain neuropathology [47–51]. Elevated levels of some miRNAs (e.g., miR-455-3p), have been found in serum samples from Alzheimer's disease patients relative to those obtained from healthy controls, thus suggesting the possibility that miR-455-3p and other miRNAs [49] could even be used as biomarkers in the contexts of very severe neuropathologies, as is the case of Alzheimer's disease [48]. In line with this, specific studies of the levels of the miRNA miR-132 have been conducted using quantitative real time PCR in postmortem brain tissue samples obtained from deceased patients that had being diagnosed with Alzheimer's disease and with mild cognitive-decline. These studies showed significantly enhanced levels of miR-132 in the samples from both Alzheimer's disease and cognitive-decline groups as compared with the data obtained from their normal (control) samples, suggesting that miR-132 might be a critical player in the pathogenesis of Alzheimer's disease [47]. Interestingly, other authors have described not elevated but rather markedly reduced levels of miR-132 at middle and advanced stages of Alzheimer's disease [50], and that the reduction in this specific miRNA aggravates amyloid and TAU pathological features [52]. For many years, miR-132 has been jointly examined together with miR-212 in a variety of different multidisciplinary studies (see for example [53–58]), as both these molecules are described to belong to the same highly conserved cluster family of miRNAs derived from a shared phylogenetic ancestor; further having their gene chromosomic organization arrayed as a tandem [59]. Similarly, in experiments using genetically modified mice, the combined genetic deletion of the genes encoding for miR-132 and miR-212 has been shown to result in an enhancement in the levels of amyloid beta peptides as well as in augmented amyloid-related plaque formation [51]. All these observations had clearly identified the microRNA212/132 cluster family as highly interesting candidates in studies about the brain neuronal function in health and disease.

Several groups, including ours, have previously described some of the functional effects of the double-deletion of the genes encoding for the miRNAs 132 and 212 in the regulation of plasticity-related functions in the mouse hippocampus [43,57,60,61]. However, although both steroid-hormones, including cortisol and corticosterone, as well as the miRNAs 132 and 212, have been independently implicated in the regulation of synaptic plasticity and emotion-related functions ([62,63]), the potential functional crosslink between steroid hormones; brain region-specific regulation of synaptic plasticity; and the miRNAs 132 and 212, remained -to our knowledge- unexplored. Therefore, the main objective of this work was to use a previously described double miR-132 and miR-212 miRNAs knockout mouse line [56] (here referred as miRNA-132/212−/−, or KO), in order to electrophysiologically study the synaptic responses to corticosterone stimulation in hippocampal slices, and to examine whether these responses are functionally homogeneous or not when measurements at the dorsal and ventral hippocampi are compared. We have obtained experimental data suggesting a potential role for the miR-132/12 in hippocampal neuroendocrine signaling and emotion-related behaviors.

#### **2. Results**

### *2.1. Comparable Basal Synaptic Transmission in Dorsal and Ventral Hippocampi of WT and miR–132/212*−*/*<sup>−</sup> *Mice*

In order to examine the physiological relevance of the microRNAs 132 and 212 in mediating the hippocampal synaptic responses to corticosterone stimulation, we explored the effect of 1 µM corticosterone on basal synaptic transmission and memory-related synaptic plasticity in the dorsal and ventral hippocampus of wild type and miR–132/212−/<sup>−</sup> mice. To this aim, we conducted recordings of extracellular field potentials ex vivo in acutely-dissociated hippocampal slices following standardized electrophysiological protocols previously described by our group and others [43,60,61,64–70].

As illustrated in Figure 1A, for these experiments, we extracted the hippocampus and split it into two main pieces lengthwise the septo-temporal axis, relative to the morphological transverse middle, and then slices for electrophysiological recordings were obtained from about 30–40% of the section comprising the defined center towards either the dorsal or the ventral ending regions (longitudinally). These areas are known to exhibit classical LTP responses as induced by electrical stimulation [71,72]. Subsequently, field potential recordings were conducted at the CA1 region upon delivering electrical stimulation at the Schaffer's collaterals from the CA3 region for both dorsal and ventral regions (Figure 1B). For plasticity experiments, recordings were conducted in the presence and absence of corticosterone in slices from both WT and miR–132/212−/<sup>−</sup> mice. Corticosterone was applied in the bath for the time specified in the figures and then removed using gravity-based-perfusion and peristaltic pump-driven solution exchange. Before examining the effect of 1 µM corticosterone on the properties of synaptic plasticity, we verified the functional integrity of the synaptic circuits by measuring input/output (I/O) field responses (see Materials and Methods) in untreated dorsal and ventral hippocampi of WT and miR–132/212−/<sup>−</sup> mice. As shown in Figure 1CD, no statistically significant differences were found in basal synaptic transmission when comparing the data obtained in slices derived from WT and miR–132/212−/<sup>−</sup> mice. Two-way RM-ANOVA with Bonferroni's and Geisser–Greenhouse's corrections, and alpha set to 0.05 (*n* = 21 animals per group), showed no significance for the effects of the 1 µM corticosterone treatment for WT vs. miR–132/212−/<sup>−</sup> at dorsal (*p* = 0.7450; F (1, 40) = 0.1073) or ventral (*p* = 0.1748; F (1, 40) = 1.908) hippocampi.

**Figure 1.** Description of the electrophysiological approach and study of basal synaptic transmission. (**A**) The cartoon on the left represents a sagittal section of the mouse brain with the relative localization of one of the two hippocampi and Anterior–Posterior and Dorsal–Ventral references indicated. To the right, the black arrow points towards the extracted hippocampus, which was divided into two parts using its anatomical middle as the nominal center. Slices were subsequently obtained from the ending dorsal (represented by a left-90°-rotated "V" shape) and ventral (represented by a horizontally flipped "C" shape) regions that comprise circa 30–40% (blue dotted lines) of the portion located longitudinally from the center towards each respective end. (**B**) Two representative microphotographs of hippocampal slices from the ventral (**left**) and dorsal (**right**) regions were used to hereby illustrate the synaptic regions examined electrophysiologically and the positioning of the recording and stimulating electrodes. The tossed "V"- and "C"-like black lines partially surrounding the dentate gyrus (gray line inside the slices) illustrate the morphology typically observed ventrally and dorsally under the microscope. The CA3 and CA1 regions are indicated. In the large, boxed inset below (**left**): Schematic representation of the protocol used to generate input/output (I/O) curves used to assess basal synaptic transmission. The large, red-filled arrow at the top left points towards the delivered electrical stimulation pulses, which consisted of 10 discrete 200 µs voltage steps (in red) from 0–9 V delivered with 15 s intervals. Below, the large purple-filled arrow on the left points towards representative traces of the 10 elicited field potential recordings. The small blue arrow preceded by an asterisk indicates a field recording with an amplitude of approximately **Figure 1.** Description of the electrophysiological approach and study of basal synaptic transmission. (**A**) The cartoon on the left represents a sagittal section of the mouse brain with the relative localization of one of the two hippocampi and Anterior–Posterior and Dorsal–Ventral references indicated. To the right, the black arrow points towards the extracted hippocampus, which was divided into two parts using its anatomical middle as the nominal center. Slices were subsequently obtained from the ending dorsal (represented by a left-90◦ -rotated "V" shape) and ventral (represented by a horizontally flipped "C" shape) regions that comprise circa 30–40% (blue dotted lines) of the portion located longitudinally from the center towards each respective end. (**B**) Two representative microphotographs of hippocampal slices from the ventral (**left**) and dorsal (**right**) regions were used to hereby illustrate the synaptic regions examined electrophysiologically and the positioning of the recording and stimulating electrodes. The tossed "V"- and "C"-like black lines partially surrounding the dentate gyrus (gray line inside the slices) illustrate the morphology typically observed ventrally and dorsally under the microscope. The CA3 and CA1 regions are indicated. In the large, boxed inset below (**left**): Schematic representation of the protocol used to generate input/output (I/O) curves used to assess basal synaptic transmission. The large, red-filled arrow at the top left points towards the delivered electrical stimulation pulses, which consisted of 10 discrete 200 µs voltage steps (in red) from 0–9 V delivered with 15 s intervals. Below, the large purple-filled arrow on the left points towards representative traces of the 10 elicited field potential recordings. The small blue arrow

50% of the maximal achievable amplitude, which, in this diagram, and as an illustration, would

preceded by an asterisk indicates a field recording with an amplitude of approximately 50% of the maximal achievable amplitude, which, in this diagram, and as an illustration, would have been generated by the step of voltage stimulation of 3 V, as indicated in the upper panel by a small blue arrow preceded by an asterisk. In the large, boxed inset below (**right**): Schematic representation of the high-frequency stimulation (HFS) protocol used to generate long-term potentiation (LTP). In the top panel, the large, red-filled arrow on the left points towards the 5 delivered bursts of electrical stimulation pulses, each burst consisting of 10 biphasic voltage steps (100 µs/phase, 100 Hz, with 500 ms intervals) given at voltage intensities eliciting about 50% of the maximal inducible amplitude. In the lower panel, the large, purple-filled arrow on the left points towards representative traces illustrating elicited field potential recordings during the baseline recordings (before HFS) and after having delivered the LTP-inducing protocol (after HFS). During experiments, the "before" and "after" field recordings are obtained upon delivering one biphasic pule (100 µs/phase) with inter-stimulus intervals of 30 s, using the same voltage stimulation that elicited about 50% of the maximal inducible amplitude. (**C**,**D**) The line charts represent the data from the changes in the field slopes versus the different values of voltage delivered (normalized to maximal slope) for slices from the dorsal and ventral regions, respectively, obtained from WT and miRNA-132/212−/<sup>−</sup> mice (*n* = 21 animals per group). No statistically significant differences were observed between the different groups (details in the main text). Data are shown as mean ± SEM.

#### *2.2. miRNA-132/212 Gene-Depletion Differentially Affects Short-Term Plasticity in Ventral and Dorsal Hippocampus*

In order to further characterize the functional activity at the dorsal and ventral hippocampi of WT and miR–132/212−/<sup>−</sup> mice, we sought to investigate the properties of presynaptic-dependent short-term plasticity. To this aim, we implemented electrophysiological protocols of Paired-Pulse-induced Facilitation (PPF) in brain slices following methods previously described by our group and others [60,61,73,74] (see also Materials and Methods). Through PPF protocols, it is possible to experimentally evoke a form of short-term plasticity that has been functionally associated with the modulation of exocytosis [75] and which is further proposed to influence the properties of long-term forms of synaptic plasticity [76].

We therefore examined PPF in dorsal and ventral hippocampi in slices from WT and KO mice. As depicted in Figure 2, analysis of the changes in the relative amplitudes (Figure 2A) and field slope ratios (Figure 2B) measured at all the different interpulse intervals examined (see also Material and Methods) showed significant differences between the groups indicating a marked impact of the miRNA-132/212 gene deletion on the properties of dorsal vs. ventral responses during short-term synaptic plasticity. Mixed-effects model three-way ANOVA with Tukey multiple comparisons correction showed significant (\*\*) effect of the interpulse interval (*p* = 0.0012), as well as highly significant (\*\*\*\*) difference for WT dorsal–WT ventral vs. KO dorsal–KO ventral (*p* < 0.0001) and high significance (\*\*\*\*) for WT dorsal–KO dorsal vs. WT ventral–KO ventral (*p* < 0.0001) for the analyses of the data from the raw amplitudes (with Chi-square = 20.98; df = 1; *p* < 0.0001; (\*\*\*\*)) (Figure 2A). Accordingly, mixed-effects model three-way ANOVA with Tukey multiple comparisons correction showed significant (\*\*) effect of the interpulse interval (*p* = 0.0033), as well as both significant (\*) difference for WT dorsal–WT ventral vs. KO dorsal–KO ventral (*p* = 0.0327) and high significance (\*\*\*\*) for WT dorsal–KO dorsal vs. WT ventral–KO ventral (*p* = 0.0001) for the analyses of the slopes (Chi-square = 14.71; df = 1; *p* = 0.0001; (\*\*\*)) (Figure 2B).

**Figure 2.** miRNA-132/212 gene-deletion influences presynaptic-dependent short-term facilitation in the mouse hippocampus. The cartoon in the upper inset illustrates the paired-pulse facilitation protocol used to induce short-term facilitation. Two consecutive pulses of electrical stimulation (indicated by black filled arrows) were delivered with an initial 20 ms interpulse interval (t), followed by consecutive increments in 20 ms (Δt) of duration until reaching 160 ms. The red-filled arrow pointing towards the right indicates the advance in time. Ratios for the values of raw amplitude or initial decay slope of the field potential responses (EPSP2/EPSP1) were used to quantify the power of paired-pulse-induced facilitation. No major differences are detected in the PPF amplitude (**A**) and field-slope ratios (**B**) when recordings from the dorsal and ventral hippocampi are examined in response to the different interpulse time intervals in slices from WT animals. Conversely, a marked difference is observed in the properties of the PPF amplitude, and field-slope ratios are examined in slices derived from miR–132/212<sup>−</sup>/<sup>−</sup> mice. *p* < 0.05 was considered significant. \* *p* < 0.05, \*\*\*\* *p* < 0.0001. Data are shown as mean ± SEM. A total of 21–22 animals per group were examined. Statistical values are described in the main text. *2.3. miRNAs-132/212 Regulate the Region-Specific Effects of Corticosterone on Metaplasticity in*  **Figure 2.** miRNA-132/212 gene-deletion influences presynaptic-dependent short-term facilitation in the mouse hippocampus. The cartoon in the upper inset illustrates the paired-pulse facilitation protocol used to induce short-term facilitation. Two consecutive pulses of electrical stimulation (indicated by black filled arrows) were delivered with an initial 20 ms interpulse interval (t), followed by consecutive increments in 20 ms (∆t) of duration until reaching 160 ms. The red-filled arrow pointing towards the right indicates the advance in time. Ratios for the values of raw amplitude or initial decay slope of the field potential responses (EPSP2/EPSP<sup>1</sup> ) were used to quantify the power of paired-pulse-induced facilitation. No major differences are detected in the PPF amplitude (**A**) and field-slope ratios (**B**) when recordings from the dorsal and ventral hippocampi are examinedin response to the different interpulse time intervals in slices from WT animals. Conversely, a marked difference is observed in the properties of the PPF amplitude, and field-slope ratios are examined in slices derived from miR–132/212−/<sup>−</sup> mice. *p* < 0.05 was considered significant. \* *p* < 0.05, \*\*\*\* *p* < 0.0001. Data are shown as mean ± SEM. A total of 21–22 animals per group were examined. Statistical values are described in the main text.

#### *the Mouse Hippocampus*  We next analyzed the effect of the 1 µM corticosterone treatment on dorsal hippo-*2.3. miRNAs-132/212 Regulate the Region-Specific Effects of Corticosterone on Metaplasticity in the Mouse Hippocampus*

campal metaplasticity in slices obtained from WT animals. To this aim, two consecutive LTP-inducing stimulation protocols known to induce metaplasticity [36,77] were delivered before and after corticosterone treatment, which generated the transient post-tetanic potentiation peaks PTP1 and PTP2 (see also Materials and Methods). Figure 3A shows the development in time of the averaged field-slope responses (normalized to the initial baseline). No differences between the untreated and the corticosterone-treated group were apparent for PTP1, and in the treated group, corticosterone treatment did not result in salient differences in the subsequent development of the potentiated field responses compared to the untreated group. However, the PTP2 response obtained after corticosterone treatment presented a statistically significant reduction in its initial amplitude, whereas the subsequent field slopes developed in time in a pattern that did resemble that of the untreated group (Figure 3A). Two-way RM-ANOVA with Bonferroni´s and Geisser– We next analyzed the effect of the 1 µM corticosterone treatment on dorsal hippocampal metaplasticity in slices obtained from WT animals. To this aim, two consecutive LTP-inducing stimulation protocols known to induce metaplasticity [36,77] were delivered before and after corticosterone treatment, which generated the transient post-tetanic potentiation peaks PTP1 and PTP2 (see also Materials and Methods). Figure 3A shows the development in time of the averaged field-slope responses (normalized to the initial baseline). No differences between the untreated and the corticosterone-treated group were apparent for PTP1, and in the treated group, corticosterone treatment did not result in salient differences in the subsequent development of the potentiated field responses compared to the untreated group. However, the PTP2 response obtained after corticosterone treatment presented a statistically significant reduction in its initial amplitude, whereas the subsequent field slopes developed in time in a pattern that did resemble that

18.25).

of the untreated group (Figure 3A). Two-way RM-ANOVA with Bonferroni's and Geisser– Greenhouse's correction, and alpha set to 0.05 (with *n* = 10–11 animals per group), showed no significance for the effect of the treatment (F (1, 19) = 0.01610; *p* = 0.9004) but a significant (\*\*\*\*) time × treatment interaction (F (159, 3021) = 2.482; *p* < 0.0001). Unpaired *t*-test (two-tailed) was also used to determine significance for three individual times: 10.5 min, corresponding to the first PTP response (Figure 3B); 25 min, corresponding to the early effect of corticosterone on the synaptic responses (Figure 3C); and 50.5 min, corresponding to the second PTP response (Figure 3D). Significant differences were found only for the PTP2 response (50.5 min: *p* = 0.0078; Welch-correction t = 3.149, df = 12.80; 10.5 min: *p* = 0.4654; Welch-correction t = 0.7529, df = 12.50; 25 min: *p* = 0.7769, Welch-correction t = 0.2876, df = 18.25). (with *n =* 10–11 per group), showed no significance for the effect of the treatment (F (1, 19) = 2.375; *p* = 0.1398) but a significant (\*\*\*\*) time x treatment interaction (F (159, 3021) = 2.884; *p* < 0.0001). As conducted for dorsal hippocampal slices, unpaired t-test (two-tailed) was also used to examine differences in the ventral hippocampus for 10.5 min, 25 min and 50.5 min. As illustrated in Figure 3F–H, no significant differences were found in any of the three examined time points. (10.5 min: *p* = 0.2748; Welch-correction t = 1.124, df = 19; 25 min: *p* = 0.0814, Welch-correction t = 1.840, df = 19; 50.5 min: *p* = 0.0703; Welch-correction t = 1.917, df = 19), but a mixed-effects model (REML) without sphericity assumption, used to independently verify the behavior of the data right after the PTP2 (min 52–79), found a highly significant time x treatment interaction (*p* < 0.0001 (\*\*\*\*), F (54, 1026) = 1.959), a phenomenon not observed in dorsal hippocampal slices (see large down-pointing gray-filled arrows after PTP2 in Figure 3A,E).

An equivalent examination was conducted in order to study the effects of corticosterone on ventral hippocampal metaplasticity using the double LTP-inducing protocol in the WT group. As depicted in Figure 3E, the PTP1 responses showed no differences when the untreated and corticosterone-treated groups were compared. Two-way RM-ANOVA with Bonferroni´s and Geisser–Greenhouse´s correction, and alpha set to 0.05

Greenhouse´s correction, and alpha set to 0.05 (with *n =* 10–11 animals per group), showed no significance for the effect of the treatment (F (1, 19) = 0.01610; *p* = 0.9004) but a significant (\*\*\*\*) time x treatment interaction (F (159, 3021) = 2.482; *p* < 0.0001). Unpaired t-test (two-tailed) was also used to determine significance for three individual times: 10.5 min, corresponding to the first PTP response (Figure 3B); 25 min, corresponding to the early effect of corticosterone on the synaptic responses (Figure 3C); and 50.5 min, corresponding to the second PTP response (Figure 3D). Significant differences were found only for the PTP2 response (50.5 min: *p* = 0.0078; Welch-correction t = 3.149, df = 12.80; 10.5 min: *p* = 0.4654; Welch-correction t = 0.7529, df = 12.50; 25 min: *p* = 0.7769, Welch-correction t = 0.2876, df =

*Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 7 of 30

**Figure 3.** Corticosterone affects hippocampal metaplasticity in a region-specific manner. (**A**) Double high-frequency stimulation protocols, known to induce LTP (see also Materials and Methods) and **Figure 3.** Corticosterone affects hippocampal metaplasticity in a region-specific manner. (**A**) Double high-frequency stimulation protocols, known to induce LTP (see also Materials and Methods) and generate robust peaks of post-synaptic-potentiation (PTP) responses (PTP1 and PTP2), were implemented in slices in order to examine the effect of 1 µM corticosterone on facilitated synaptic transmission and plasticity in the dorsal hippocampus of WT animals. No salient differences were apparent for the PTP1 response or the succeeding development of synaptic transmission during the exposure to corticosterone, as compared to the untreated control group. However, a very pronounced, statistically significant reduction of the PTP2 response was observed in the corticosterone-treated group, with the following field responses developing analogously to those of the untreated control group. Independent statistical analyses (details in the main text) were conducted for the PTP1 response measured at 10.5 min (**B**), which showed no significant differences; as well as for the development of synaptic transmission in the presence of corticosterone at 25 min (**C**), also showing no differences; and for the PTP2 response at 50.5 min (**D**), which showed significant (\*\*) differences (statistical values in the main text). Corresponding analyses were done for the ventral hippocampus, which showed no differences for PTP1 and PTP2 (**E**), and also no significant differences in the field responses of 10.5 min (**F**), 25 min (**G**) and 50.5 min (**H**). Gray-filled large arrows show effects on metaplasticity. ns = not significant. *p* < 0.05 was considered significant. \*\* *p* < 0.01. Data are shown as mean ± SEM. Statistical values and the number of subjects are described in the main text.

An equivalent examination was conducted in order to study the effects of corticosterone on ventral hippocampal metaplasticity using the double LTP-inducing protocol in the WT group. As depicted in Figure 3E, the PTP1 responses showed no differences when the untreated and corticosterone-treated groups were compared. Two-way RM-ANOVA with Bonferroni's and Geisser–Greenhouse's correction, and alpha set to 0.05 (with *n* = 10–11 per group), showed no significance for the effect of the treatment (F (1, 19) = 2.375; *p* = 0.1398) but a significant (\*\*\*\*) time × treatment interaction (F (159, 3021) = 2.884; *p* < 0.0001). As conducted for dorsal hippocampal slices, unpaired *t*-test (two-tailed) was also used to examine differences in the ventral hippocampus for 10.5 min, 25 min and 50.5 min. As illustrated in Figure 3F–H, no significant differences were found in any of the three examined time points. (10.5 min: *p* = 0.2748; Welch-correction t = 1.124, df = 19; 25 min: *p* = 0.0814, Welch-correction t = 1.840, df = 19; 50.5 min: *p* = 0.0703; Welch-correction t = 1.917, df = 19), but a mixed-effects model (REML) without sphericity assumption, used to independently verify the behavior of the data right after the PTP2 (min 52–79), found a highly significant time × treatment interaction (*p* < 0.0001 (\*\*\*\*), F (54, 1026) = 1.959), a phenomenon not observed in dorsal hippocampal slices (see large down-pointing gray-filled arrows after PTP2 in Figure 3A,E).

#### *2.4. miRNA–132/212 Gene Deletion Disrupts the Region-Specific Effect of Corticosterone on Hippocampal Metaplasticity*

In order to assess the relevance of the microRNAs 132 and 212 as regulators of the effects of corticosterone in hippocampal synaptic transmission and metaplasticity, we next investigated the acute impact of corticosterone stimulation on the dorsal and ventral hippocampus using slices from miR–132/212−/<sup>−</sup> mice. We first examined the effects of corticosterone on metaplasticity using the paired LTP-inducing protocol in the dorsal hippocampus. As shown in Figure 4A, slices from miR–132/212−/<sup>−</sup> mice presented with metaplasticity responses are indistinguishable from those observed in slices from their related WT counterparts after corticosterone treatment. That is, corticosterone induced a reduction in the amplitude of the PTP2 response in miR–132/212−/<sup>−</sup> slices but did not affect the temporal development of the subsequent field responses relative to the untreated miR–132/212−/<sup>−</sup> control group. Two-way RM-ANOVA with Bonferroni's and Geisser– Greenhouse's correction, and alpha set to 0.05 (with *n* = 10–11 per group), showed no significance for the effect of the treatment (F (1, 19) = 0.002677; *p* = 0.9593) but a significant (\*\*\*\*) time × treatment interaction (F (159, 3021) = 5.025; *p* < 0.0001). Unpaired *t*-test (twotailed), used to examine differences at 10.5, 25 and 50.5 min, returned non-significant values of *p* = 0.2252 and Welch-correction t = 1.280 with df = 11.75 for 10.5 min (Figure 4B); *p* = 0.0549 and Welch-correction t = 2.048 with df = 18.62 for 25 min (Figure 4C); and significant values of *p* = 0.0083 (\*\*\*\*) with Welch-correction t = 3.011 and df = 16.01 only for 50.5 min (Figure 4C).

We subsequently analyzed the impact of corticosterone on metaplasticity in the ventral hippocampus of slices derived from miR–132/212−/<sup>−</sup> mice. Figure 4E shows comparable metaplasticity-related LTP responses between the untreated and the corticosterone-treated groups for the PTP1 response and no statistically significant differences at 10.5 min timepoint (Figure 4F). However, corticosterone produced a statistically significant increase in the synaptic responses in the presence of corticosterone as recorded at 25 min, which is indicated by a large gray-filled down-pointing arrow in Figure 4E, a phenotype also revealed at the 25 min timepoint in Figure 4G (\*\*). Moreover, corticosterone also produced a significant reduction in the PTP2 response in ventral hippocampal slices from miR–132/212−/<sup>−</sup> mice (Figure 4H), an effect not detected in corticosterone-treated ventral hippocampal slices derived from WT mice (Figure 3H). Two-way RM ANOVA with stacked matching and no sphericity assumption (done with Geisser–Greenhouse's and Bonferroni's corrections and alpha set to 0.05) for data depicted in Figure 4E reported no significance for the effect of treatment (*p* = 0.0991) and a highly statistically significant (\*\*\*\*) time × treatment interaction (*p* < 0.0001; F (159, 3180) = 3.899). Independent two-tailed unpaired *t*-test yielded

values of *p* = 0.6237 with Welch-corrections of t = 0.4998 and df = 16.76 for datapoints at 10.5 min (Figure 4F); *p* = 0.0064 (\*\*) with Welch-corrections of t = 3.232 and df = 13.26 for 25 min (Figure 4G); and *p* = 0.0492 (\*) with Welch-corrections of t = 2.107 and df = 18.27 for 50.5 min (Figure 4H). *Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 9 of 30

**Figure 4.** miRNAs–132/212 gene deletion impairs the region-selective effect of corticosterone on hippocampal synaptic plasticity. The properties of synaptic transmission and plasticity examined in the dorsal hippocampus of untreated and corticosterone-treated slices derived from miR–132/212<sup>−</sup>/<sup>−</sup> mice were comparable with the WT group described above. That is, no major differences were observed for PTP1, but a significant reduction in the PTP2 response (**A**) was apparent, also with no detectable differences for 10.5 min (**B**) and 25 min (**C**), and a statistically significant (\*\*) difference between untreated and treated groups for 50.5 min (**D**). However, for the ventral hippocampus, while slices from miR–132/212−/− mice exhibited similar PTP1 responses for the untreated and corticosterone-treated groups (**E**), the facilitated postsynaptic responses measured in the presence of corticosterone presented an enhanced amplitude (large gray-filled arrow) as well as a significant reduction of the PTP2 response, to a degree that was not observed in the corticosterone-treated WT group. Statistical analyses conducted for responses measured at 10.5 min showed no significant differences (**F**); whereas both recordings at 25 min (**G**) and 50.5 min (**H**) showed statistically significant differences (\*\* and \*, respectively). ns = not significant. \* *p* < 0.05, \*\* *p* < 0.01. Data are shown as mean ± SEM. Statistical values and number of subjects are described in the main text. We subsequently analyzed the impact of corticosterone on metaplasticity in the ven-**Figure 4.** miRNAs–132/212 gene deletion impairs the region-selective effect of corticosterone on hippocampal synaptic plasticity. The properties of synaptic transmission and plasticity examined in the dorsal hippocampus of untreated and corticosterone-treated slices derived from miR–132/212−/<sup>−</sup> mice were comparable with the WT group described above. That is, no major differences were observed for PTP1, but a significant reduction in the PTP2 response (**A**) was apparent, also with no detectable differences for 10.5 min (**B**) and 25 min (**C**), and a statistically significant (\*\*) difference between untreated and treated groups for 50.5 min (**D**). However, for the ventral hippocampus, while slices from miR–132/212−/<sup>−</sup> mice exhibited similar PTP1 responses for the untreated and corticosterone-treated groups (**E**), the facilitated postsynaptic responses measured in the presence of corticosterone presented an enhanced amplitude (large gray-filled arrow) as well as a significant reduction of the PTP2 response, to a degree that was not observed in the corticosterone-treated WT group. Statistical analyses conducted for responses measured at 10.5 min showed no significant differences (**F**); whereas both recordings at 25 min (**G**) and 50.5 min (**H**) showed statistically significant differences (\*\* and \*, respectively). ns = not significant. \* *p* < 0.05, \*\* *p* < 0.01. Data are shown as mean ± SEM. Statistical values and number of subjects are described in the main text.

tral hippocampus of slices derived from miR–132/212−/− mice. Figure 4E shows comparable metaplasticity-related LTP responses between the untreated and the corticosteronetreated groups for the PTP1 response and no statistically significant differences at 10.5 min timepoint (Figure 4F). However, corticosterone produced a statistically significant increase in the synaptic responses in the presence of corticosterone as recorded at 25 min, which is indicated by a large gray-filled down-pointing arrow in Figure 4E, a phenotype also revealed at the 25 min timepoint in Figure 4G (\*\*). Moreover, corticosterone also produced a significant reduction in the PTP2 response in ventral hippocampal slices from miR–132/212−/− mice (Figure 4H), an effect not detected in corticosterone-treated ventral hippocampal slices derived from WT mice (Figure 3H). Two-way RM ANOVA with Taken together, all these observations point towards a possible involvement of the miRNA 132/212 cluster family in the region-specific regulation of the effects of glucocorticoid hormones on hippocampal synaptic metaplasticity functions and memory-related emotional behaviors. These data further suggested that molecular elements known to mediate in the regulation of both synaptic functions and glucocorticoid hormone activity could be differentially affected by corticosterone in WT and miR–132/212−/<sup>−</sup> mice hippocampi. In an effort to launch a first experimental verification of this hypothesis, we used the technique of western blot (Materials and Methods) and examined the expression levels of CREB, Sirt1, MSK1, CDK5 and PTEN in the hippocampi of WT and miR–132/212−/<sup>−</sup> mice.

stacked matching and no sphericity assumption (done with Geisser–Greenhouse´s and Bonferroni´s corrections and alpha set to 0.05) for data depicted in Figure 4E reported no significance for the effect of treatment (*p* = 0.0991) and a highly statistically significant

unpaired t-test yielded values of *p* = 0.6237 with Welch-corrections of t = 0.4998 and df =

#### *2.5. Deletion of miR–132/212*−*/*<sup>−</sup> *Prevents Enhanced Expression of CREB in the Hippocampi of Corticosterone Treated Slices 2.5. Deletion of miR–132/212−/− Prevents Enhanced Expression of CREB in the Hippocampi of Corticosterone Treated Slices*  The participation of CREB in the regulation of synaptic plasticity and hippocampus-

16.76 for datapoints at 10.5 min (Figure 4F); *p* = 0.0064 (\*\*) with Welch-corrections of t = 3.232 and df = 13.26 for 25 min (Figure 4G); and *p* = 0.0492 (\*) with Welch-corrections of t

Taken together, all these observations point towards a possible involvement of the miRNA 132/212 cluster family in the region-specific regulation of the effects of glucocorticoid hormones on hippocampal synaptic metaplasticity functions and memory-related emotional behaviors. These data further suggested that molecular elements known to mediate in the regulation of both synaptic functions and glucocorticoid hormone activity could be differentially affected by corticosterone in WT and miR–132/212−/− mice hippocampi. In an effort to launch a first experimental verification of this hypothesis, we used the technique of western blot (Materials and Methods) and examined the expression levels of CREB, Sirt1, MSK1, CDK5 and PTEN in the hippocampi of WT and miR–132/212−/− mice.

*Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 10 of 30

= 2.107 and df = 18.27 for 50.5 min (Figure 4H).

The participation of CREB in the regulation of synaptic plasticity and hippocampusdependent learning and memory functions has been extensively described, and functional crosslinks between microRNA-mediated regulation (including the 132/212 cluster) and CREB expression have also been proposed [44,78–83]. Moreover, our group has recently described significantly enhanced levels of CREB in the hippocampus of miR–132/212−/<sup>−</sup> mice [61]. However, to the best of our knowledge, the acute effects of corticosterone on the levels of CREB has not been examined in hippocampi lacking miR-132/212. We therefore prepared acutely-dissociated hippocampal slices from both WT and miR–132/212−/<sup>−</sup> mice and subsequently stimulated them for 1 h with corticosterone. Two-way ANOVA (Alpha 0.05) reported a significant effect of the genotype (F (1, 12) = 6.504; *p* = 0.0255), high significance for the effects of the treatment (F (1, 12) = 45.16; *p* < 0.0001), and a significant interaction between factors (F (1, 12) = 80.05; *p* < 0.0001). dependent learning and memory functions has been extensively described, and functional crosslinks between microRNA-mediated regulation (including the 132/212 cluster) and CREB expression have also been proposed [44,78–83]. Moreover, our group has recently described significantly enhanced levels of CREB in the hippocampus of miR–132/212−/<sup>−</sup> mice [61]. However, to the best of our knowledge, the acute effects of corticosterone on the levels of CREB has not been examined in hippocampi lacking miR-132/212. We therefore prepared acutely-dissociated hippocampal slices from both WT and miR–132/212−/<sup>−</sup> mice and subsequently stimulated them for 1 h with corticosterone. Two-way ANOVA (Alpha 0.05) reported a significant effect of the genotype (F (1, 12) = 6.504; *p* = 0.0255), high significance for the effects of the treatment (F (1, 12) = 45.16; *p* < 0.0001), and a significant interaction between factors (F (1, 12) = 80.05; *p* < 0.0001).

As shown in Figure 5A,C, further post hoc analyses revealed that untreated hippocampi derived from miR–132/212−/<sup>−</sup> mice presented enhanced expression levels of CREB compared to untreated WT controls, thus corroborating our previously reported observations [61]. However, both the levels of CREB (Figure 5B,C) were considerably reduced in the corticosterone-treated hippocampi of miR–132/212−/<sup>−</sup> mice when compared to both untreated KO miR–132/212−/<sup>−</sup> slices and corticosterone-treated hippocampi from the WT animals. Phospho-CREB was also downregulated in KO compared to WT slices when treated with corticosterone (Figure 5B,D), as shown by unpaired *t*-test (*p* < 0.0001; t = 8.834, df = 6). As shown in Figure 5AC, further post hoc analyses revealed that untreated hippocampi derived from miR–132/212−/− mice presented enhanced expression levels of CREB compared to untreated WT controls, thus corroborating our previously reported observations [61]. However, both the levels of CREB (Figure 5BC) were considerably reduced in the corticosterone-treated hippocampi of miR–132/212−/− mice when compared to both untreated KO miR–132/212−/− slices and corticosterone-treated hippocampi from the WT animals. Phospho-CREB was also downregulated in KO compared to WT slices when treated with corticosterone (Figure 5BD), as shown by unpaired t-test (*p* < 0.0001; t = 8.834, df = 6).

**Figure 5.** Corticosterone induces reduction in the levels of CREB in miR–132/212−/− mice hippocampus. (**A,B**) Representative photographs of blotting membranes containing transferred proteins **Figure 5.** Corticosterone induces reduction in the levels of CREB in miR–132/212−/<sup>−</sup> mice hippocampus. (**A**,**B**) Representative photographs of blotting membranes containing transferred proteins obtained from untreated and corticosterone-treated hippocampal tissue from WT and miRNA-132/212−/<sup>−</sup> (KO) mice, respectively. The membranes were incubated with antibodies for CREB, pCREB and GAPDH. (**C**) The chart shows the averaged data for the levels of CREB normalized to those of the GAPDH as derived from densitometric analysis of the blots. The levels of CREB are significantly augmented in the hippocampus of the miRNA-132/212 KO mice and significantly reduced in response to corticosterone treatment. (**D**) Charts of the blots averaged densitometry data for the levels of pCREB relative to GAPDH, as examined using corticosterone-treated hippocampal tissue from WT and miRNA-132/212−/<sup>−</sup> mice. Note the significant reduction in the levels of the two forms of CREB in response to corticosterone. The results in all charts are shown in each case as a fold change relative to the detected levels of the enzyme GAPDH. \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001. Data are shown as mean ± SEM (*n* = 4 animals per group).

#### *2.6. Sirt1 Protein Levels Are Enhanced and Insensitive to Corticosterone Stimulation in the miR–132/212*−*/*<sup>−</sup> *Mice Hippocampi* 0.0001. Data are shown as mean ± SEM (*n =* 4 animals per group). *2.6. Sirt1 Protein Levels Are Enhanced and Insensitive to Corticosterone Stimulation in the*

obtained from untreated and corticosterone-treated hippocampal tissue from WT and miRNA-132/212<sup>−</sup>/<sup>−</sup> (KO) mice, respectively. The membranes were incubated with antibodies for CREB, pCREB and GAPDH. (**C**) The chart shows the averaged data for the levels of CREB normalized to those of the GAPDH as derived from densitometric analysis of the blots. The levels of CREB are significantly augmented in the hippocampus of the miRNA-132/212 KO mice and significantly reduced in response to corticosterone treatment. (**D**) Charts of the blots averaged densitometry data for the levels of pCREB relative to GAPDH, as examined using corticosterone-treated hippocampal tissue from WT and miRNA-132/212<sup>−</sup>/<sup>−</sup> mice. Note the significant reduction in the levels of the two forms of CREB in response to corticosterone. The results in all charts are shown in each case as a fold change relative to the detected levels of the enzyme GAPDH. \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* <

*Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 11 of 30

A functional link has been very recently established between the brain neuronal microRNAs miR-132 and miR-212 and their target-regulated protein Sirtuin 1 (Silent Information Regulator 1; in short, Sirt1) in the context of Alzheimer's Disease [84]. Sirt1 has indeed been shown to regulate endocrine activity and participates in the regulation of memory-related neuronal functions through its capability to induce axonal and dendritic morphological rearrangements [84,85], see also [86]. Moreover, corticosterone (the brain stress-signaling molecule in rodents equivalent to cortisol in humans) has also been shown to dose-dependently regulate the levels of Sirt1 in cellular stress models [87]. While a crosslink between Sirt1 and the miRNAs 132/-212 had been previously examined [88], the acute impact of corticosterone on the levels of hippocampal Sirt1 and its relation to the levels of the miRNAs 132 and 212 remained, to our knowledge, uncharacterized. We, therefore, next sought out to examine the acute effects of corticosterone on the levels of Sirt1 in the hippocampi of both WT and miR–132/212−/<sup>−</sup> mice (see also Materials and Methods). Two-way ANOVA (Alpha 0.05) reported statistically significant differences (\*\*\*) for genotype (F (1, 12) = 31.52; *p* = 0.0001); also, significance (\*) for treatment (F (1, 12) = 5.293; *p* = 0.0401) with no differences for interactions (F (1, 12) = 2.287; *p* = 0.1564). Subsequent post hoc analysis showed that the hippocampi of miR–132/212−/<sup>−</sup> mice presented a significant enhancement in the levels of Sirt1, while the corticosterone treatment did not significantly revert this phenotype (Figure 6). *miR–132/212−/− Mice Hippocampi*  A functional link has been very recently established between the brain neuronal microRNAs miR-132 and miR-212 and their target-regulated protein Sirtuin 1 (Silent Information Regulator 1; in short, Sirt1) in the context of Alzheimer's Disease [84]. Sirt1 has indeed been shown to regulate endocrine activity and participates in the regulation of memory-related neuronal functions through its capability to induce axonal and dendritic morphological rearrangements [84,85], see also [86]. Moreover, corticosterone (the brain stress-signaling molecule in rodents equivalent to cortisol in humans) has also been shown to dose-dependently regulate the levels of Sirt1 in cellular stress models [87]. While a crosslink between Sirt1 and the miRNAs 132/-212 had been previously examined [88], the acute impact of corticosterone on the levels of hippocampal Sirt1 and its relation to the levels of the miRNAs 132 and 212 remained, to our knowledge, uncharacterized. We, therefore, next sought out to examine the acute effects of corticosterone on the levels of Sirt1 in the hippocampi of both WT and miR–132/212−/− mice (see also Materials and Methods). Two-way ANOVA (Alpha 0.05) reported statistically significant differences (\*\*\*) for genotype (F (1, 12) = 31.52; *p* = 0.0001); also, significance (\*) for treatment (F (1, 12) = 5.293; *p* = 0.0401) with no differences for interactions (F (1, 12) = 2.287; *p* = 0.1564). Subsequent post hoc analysis showed that the hippocampi of miR–132/212−/− mice presented a significant enhancement in the levels of Sirt1, while the corticosterone treatment did not significantly revert this phenotype (Figure 6).

**Figure 6.** Corticosterone does not affect the otherwise enhanced levels of Sirt1 detected in the miR– 132/212−/<sup>−</sup> mice hippocampus. (**A**) Representative photographs of blot membranes derived from WB studies using untreated (upper set) and corticosterone-treated (lower set) hippocampal tissue from WT and miRNA-132/212−/<sup>−</sup> (KO) mice. The membranes were incubated with antibodies for Sirt1 and GAPDH. (**B**) Averaged levels of Sirt1 under untreated and corticosterone-treated conditions, normalized to GAPDH values, as derived from densitometric analysis of blots. The levels of Sirt1 appeared significantly enhanced in the untreated hippocampi from miRNA-132/212−/<sup>−</sup> (KO) mice. The levels of hippocampal Sirt1 continued to be significantly enhanced upon corticosterone-treatment in miRNA-132/212−/<sup>−</sup> (KO) mice compared to those of the WT controls. Results in charts represent fold changes relative to GAPDH. \*\* *p* < 0.01, \*\*\* *p* < 0.001. Data are shown as mean ± SEM (*n* = 4 animals per group).

#### *2.7. Corticosterone Effects on MSK1 Levels in WT and miR–132/212*−*/*<sup>−</sup> *Mice Hippocampi* fold changes relative to GAPDH. \*\* *p* < 0.01, \*\*\* *p* < 0.001. Data are shown as mean ± SEM (*n =* 4 animals per group).

**Figure 6.** Corticosterone does not affect the otherwise enhanced levels of Sirt1 detected in the miR– 132/212<sup>−</sup>/<sup>−</sup> mice hippocampus. (**A**) Representative photographs of blot membranes derived from WB studies using untreated (upper set) and corticosterone-treated (lower set) hippocampal tissue from WT and miRNA-132/212<sup>−</sup>/<sup>−</sup> (KO) mice. The membranes were incubated with antibodies for Sirt1 and GAPDH. (**B**) Averaged levels of Sirt1 under untreated and corticosterone-treated conditions, normalized to GAPDH values, as derived from densitometric analysis of blots. The levels of Sirt1 appeared significantly enhanced in the untreated hippocampi from miRNA-132/212<sup>−</sup>/<sup>−</sup> (KO) mice. The levels of hippocampal Sirt1 continued to be significantly enhanced upon corticosterone-treatment in miRNA-132/212<sup>−</sup>/<sup>−</sup> (KO) mice compared to those of the WT controls. Results in charts represent

*Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 12 of 30

The Mitogen- and Stress-activated Kinase 1 (MSK1) enzyme is a nuclear serine/ threonine protein kinase whose in vivo activity becomes triggered upon activation of either the Mitogen-Activated Protein Kinases (MAPK) ERK1/2 or p38, which participate in the structural/functional modifications of the histones during the regulation of gene transcription that is associated with emotional processing and hippocampus-dependent learning and memory functions [89–93]. Moreover, it has also been shown that MSK1 is associated with the regulation of transcription of miR-132/212 [56]. However, how glucocorticoid stimulation could impact the levels of MSK1 in a miRNA-132/212−/−-dependent manner remained unknown. We therefore investigated the effects of acute corticosterone stimulation on the hippocampal levels of MSK1 in WT and miR–132/212−/<sup>−</sup> mice. For total MSK1 analyses, two-way ANOVA (Alpha 0.05) reported no differences for genotype (F (1, 12) = 0.1503; *p* = 0.7050), significant (\*\*\*\*) differences for the treatment (F (1, 12) = 63.25; *p* < 0.0001) and significant (\*\*) differences for the interaction (F (1, 12) = 9.618; *p* = 0.0092). For pMSK1 analyses, two-way ANOVA (Alpha 0.05) reported no significant differences for genotype (F (1, 12) = 3.486; *p* = 0.0865), significant (\*\*\*) differences for the treatment (F (1, 12) = 30.38; *p* = 0.0001) and no significant differences for the interaction (F (1, 12) = 2.473; *p* = 0.1418). Further post hoc analysis revealed that the untreated hippocampi from miR–132/212−/<sup>−</sup> and WT showed comparable levels of MSK1 and phospho-MSK1 (pMSK1) (Figure 7A–D). Additionally, the corticosterone treatment resulted in an overall pronounced elevation in the detected levels of MSK1 in the hippocampi of both WT and miR–132/212−/<sup>−</sup> animals, indicating that corticosterone has the ability to modulate the levels of total MSK1 in a manner independent of the levels of miRNAs 132/212 (Figure 7C). However, an inhibitory effect of corticosterone was observed for the levels of p-MSK1 in WT hippocampi that was not present in the hippocampi from miR–132/212−/<sup>−</sup> mice (Figure 7D), indicating the possible role of pMSK1 in the corticosterone response that is missing in the KO mouse. *2.7. Corticosterone Effects on MSK1 Levels in WT and miR–132/212−/− Mice Hippocampi*  The Mitogen- and Stress-activated Kinase 1 (MSK1) enzyme is a nuclear serine/threonine protein kinase whose in vivo activity becomes triggered upon activation of either the Mitogen-Activated Protein Kinases (MAPK) ERK1/2 or p38, which participate in the structural/functional modifications of the histones during the regulation of gene transcription that is associated with emotional processing and hippocampus-dependent learning and memory functions [89–93]. Moreover, it has also been shown that MSK1 is associated with the regulation of transcription of miR-132/212 [56]. However, how glucocorticoid stimulation could impact the levels of MSK1 in a miRNA-132/212−/−-dependent manner remained unknown. We therefore investigated the effects of acute corticosterone stimulation on the hippocampal levels of MSK1 in WT and miR–132/212−/− mice. For total MSK1 analyses, two-way ANOVA (Alpha 0.05) reported no differences for genotype (F (1, 12) = 0.1503; *p* = 0.7050), significant (\*\*\*\*) differences for the treatment (F (1, 12) = 63.25; *p* < 0.0001) and significant (\*\*) differences for the interaction (F (1, 12) = 9.618; *p* = 0.0092). For pMSK1 analyses, two-way ANOVA (Alpha 0.05) reported no significant differences for genotype (F (1, 12) = 3.486; *p* = 0.0865), significant (\*\*\*) differences for the treatment (F (1, 12) = 30.38; *p* = 0.0001) and no significant differences for the interaction (F (1, 12) = 2.473; *p* = 0.1418). Further post hoc analysis revealed that the untreated hippocampi from miR–132/212−/− and WT showed comparable levels of MSK1 and phospho-MSK1 (pMSK1) (Figure 7A–D). Additionally, the corticosterone treatment resulted in an overall pronounced elevation in the detected levels of MSK1 in the hippocampi of both WT and miR–132/212−/− animals, indicating that corticosterone has the ability to modulate the levels of total MSK1 in a manner independent of the levels of miRNAs 132/212 (Figure 7C). However, an inhibitory effect of corticosterone was observed for the levels of p-MSK1 in WT hippocampi that was not present in the hippocampi from miR–132/212−/− mice (Figure 7D), indicating the possible role of pMSK1 in the corticosterone response that is missing in the KO mouse.

**Figure 7.** The levels of hippocampal MSK1 in WT and miR–132/212−/<sup>−</sup> are comparable and enhanced by corticosterone. (**A**,**B**) Representative blots for untreated and corticosterone-treated hippocampal tissue from WT and miRNA-132/212−/<sup>−</sup> (KO, in the figure) mice in membranes incubated with antibodies for MSK1, phospho-MSK1 (pMSK1) and GAPDH. (**C**) Averaged MSK1 levels, relative to GAPDH, as from densitometric analysis of blots. WT and miRNA-132/212−/<sup>−</sup> presented with enhanced MSK1 levels in the corticosterone-treated groups. (**D**) Averaged pMSK1 levels, relative to GAPDH, as from densitometric analysis of blots. While the levels of pMSK1 were comparable between miRNA-132/212−/<sup>−</sup> and WT hippocampi, corticosterone significantly reduced the levels of pMSK1 in the WT group. *p* < 0.05 was considered significant. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001. Results represent fold changes relative to GAPDH. Data are shown as mean ± SEM (*n* = 4 animals per group).

#### *2.8. Unaltered Levels of CDK5 and PTEN in the Hippocampi of miRNA-132/212*−*/*<sup>−</sup> *Mice 2.8. Unaltered Levels of CDK5 and PTEN in the Hippocampi of miRNA-132/212−/− Mice*

**Figure 7.** The levels of hippocampal MSK1 in WT and miR–132/212<sup>−</sup>/<sup>−</sup> are comparable and enhanced by corticosterone. (**AB**) Representative blots for untreated and corticosterone-treated hippocampal tissue from WT and miRNA-132/212−/− (KO, in the figure) mice in membranes incubated with antibodies for MSK1, phospho-MSK1 (pMSK1) and GAPDH. (**C**) Averaged MSK1 levels, relative to GAPDH, as from densitometric analysis of blots. WT and miRNA-132/212−/− presented with enhanced MSK1 levels in the corticosterone-treated groups. (**D**) Averaged pMSK1 levels, relative to GAPDH, as from densitometric analysis of blots. While the levels of pMSK1 were comparable between miRNA-132/212<sup>−</sup>/<sup>−</sup> and WT hippocampi, corticosterone significantly reduced the levels of pMSK1 in the WT group. *p* < 0.05 was considered significant. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001. Results represent fold changes relative to GAPDH. Data are shown as mean ± SEM (*n =*

*Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 13 of 30

4 animals per group).

Both CDK5 and PTEN are brain neuronal proteins expressed in the hippocampus that are involved in the regulation of memory-related synaptic functions [94–103]. Additionally, both CDK5 and PTEN have been also associated with the stress-related effects of corticosterone and the hippocampal function (see also [104–106] and [94,107,108], respectively). Similarly, both CDK5 and PTEN have been shown to be regulated by microRNAs (see e.g., [109,110] and [111,112], respectively). However, the impact of corticosterone on the hippocampal levels of CDK5 and PTEN and its relation to miRNAs 132 and 212 remained, to our knowledge, uncharacterized. We therefore examined the impact of corticosterone on the levels of both CDK5 and PTEN in the hippocampus of WT and miRNA-132/212−/−. Figure 8 shows data from WB experiments indicating that neither the deletion of the genes encoding for the miRNAs 132 and 212 nor the treatment with corticosterone affects the expression levels of CDK5 or PTEN in either WT or miRNA-132/212−/<sup>−</sup> mice hippocampi. For CDK5 analyses, two-way ANOVA (Alpha 0.05) reported no significant differences for genotype (F (1, 12) = 1.303; *p* = 0.2759), interaction (F (1, 12) = 0.2628; *p* = 0.6175) or treatment (F (1, 12) = 0.08639; *p* = 0.7738). For PTEN analyses, two-way ANOVA (Alpha 0.05) reported no significant differences for genotype (F (1, 12) = 0.2911; *p* = 0.5994), no differences for the treatment (F (1, 12) = 3.248; *p* = 0.0967) and no differences for the interaction (F (1, 12) = 0.2734; *p* = 0.6106). Both CDK5 and PTEN are brain neuronal proteins expressed in the hippocampus that are involved in the regulation of memory-related synaptic functions [94–103]. Additionally, both CDK5 and PTEN have been also associated with the stress-related effects of corticosterone and the hippocampal function (see also [104–106] and [94,107,108], respectively). Similarly, both CDK5 and PTEN have been shown to be regulated by microRNAs (see e.g., [109,110] and [111,112], respectively). However, the impact of corticosterone on the hippocampal levels of CDK5 and PTEN and its relation to miRNAs 132 and 212 remained, to our knowledge, uncharacterized. We therefore examined the impact of corticosterone on the levels of both CDK5 and PTEN in the hippocampus of WT and miRNA-132/212−/−. Figure 8 shows data from WB experiments indicating that neither the deletion of the genes encoding for the miRNAs 132 and 212 nor the treatment with corticosterone affects the expression levels of CDK5 or PTEN in either WT or miRNA-132/212−/− mice hippocampi. For CDK5 analyses, two-way ANOVA (Alpha 0.05) reported no significant differences for genotype (F (1, 12) = 1.303; *p* = 0.2759), interaction (F (1, 12) = 0.2628; *p* = 0.6175) or treatment (F (1, 12) = 0.08639; *p* = 0.7738). For PTEN analyses, two-way ANOVA (Alpha 0.05) reported no significant differences for genotype (F (1, 12) = 0.2911; *p* = 0.5994), no differences for the treatment (F (1, 12) = 3.248; *p* = 0.0967) and no differences for the interaction (F (1, 12) = 0.2734; *p* = 0.6106).

**Figure 8.** Corticosterone does not affect the otherwise comparable levels of CDK5 or PTEN in WT and miRNA-132/212<sup>−</sup>/<sup>−</sup> mice hippocampi. (**A,B**) Blots for untreated and corticosterone-treated (treated) hippocampal tissue from WT and miRNA-132/212<sup>−</sup>/<sup>−</sup> (KO) mice in membranes incubated with antibodies for CDK5, PTEN and GAPDH. (**C,D**) Averaged CDK5 and PTEN levels, respectively, relative to GAPDH, as from densitometric analysis of blots for untreated and corticosteronetreated tissue. No significant differences in the levels of CDK5 or PTEN were found between the **Figure 8.** Corticosterone does not affect the otherwise comparable levels of CDK5 or PTEN in WT and miRNA-132/212−/<sup>−</sup> mice hippocampi. (**A**,**B**) Blots for untreated and corticosterone-treated (treated) hippocampal tissue from WT and miRNA-132/212−/<sup>−</sup> (KO) mice in membranes incubated with antibodies for CDK5, PTEN and GAPDH. (**C**,**D**) Averaged CDK5 and PTEN levels, respectively, relative to GAPDH, as from densitometric analysis of blots for untreated and corticosterone-treated tissue. No significant differences in the levels of CDK5 or PTEN were found between the untreated or corticosterone-treated hippocampi of WT and miRNA-132/212−/<sup>−</sup> mice. Results represent fold changes relative to GAPDH. Data are shown as mean ± SEM (*n* = 4 animals per group).

### *2.9. Reduced Anxiety-Like Behavior in miRNA-132/212*−*/*<sup>−</sup> *Mice*

Our data from the pharmacological treatment with corticosterone in hippocampal electrophysiology, as well as the western blot analyses of changes in the levels of molecules involved in the physiological responses to stress described here in response to corticosterone treatment, suggested that the miRNA 132/212 cluster was implicated in the regulation of emotional behaviors. Consequently, changes in anxiety-like behaviors were predicted to be detectable in miRNA-132/212−/<sup>−</sup> mice. In order to test this hypothesis, we conducted behavioral experiments using the open field test (OFT) and the elevated plus maze [EPM

(Materials and Methods)]. As shown in Figure 9, no significant differences in the behavior of WT and miRNA-132/212−/<sup>−</sup> mice were observed in the OFT. In the OFT, unpaired two-tailed *t*-test yielded values of *p* = 0.064 (t = 1.949, df = 22) for total distance traveled comparisons (Figure 9A) with no between-group differences found in the cumulative duration in the center of the open field (Figure 9B). In the EPM, however, statistically significant differences in features associated to anxiety-like behaviours became apparent in miRNA-132/212−/<sup>−</sup> subjects when compared to WT mice. Unpaired two-tailed *t*-test yielded values of *p* = 0.0094 (t = 2.874, df = 20) for the absolute time spent in the open arm (Figure 9C), with no differences detected for the latency to the first access to the open arm (*p* = 0.4491; t = 0.7721, df = 20) in the EPM (Figure 9D). Moreover, when the time spent in the open arms was analysed relative to the time spent in the closed arms (here referred to as analysis factor score), miRNA-132/212−/<sup>−</sup> animals exhibited significantly (\*\*) enhanced time spent in the open arms compared to their relative WT control counterparts (*p* = 0.0094; t = 2.872, df = 20) (Figure 9E). Taken together, all these observations encourage further research in order to independently verify the herein proposed potential involvement of the 132 and 212 miRNA cluster in the modulation of stress-related emotional responses mediated by steroid hormones belonging to the pituitary adrenocortical axis. unpaired two-tailed t-test yielded values of *p* = 0.064 (t = 1.949, df = 22) for total distance traveled comparisons (Figure 9A) with no between-group differences found in the cumulative duration in the center of the open field (Figure 9B). In the EPM, however, statistically significant differences in features associated to anxiety-like behaviours became apparent in miRNA-132/212−/− subjects when compared to WT mice. Unpaired two-tailed ttest yielded values of *p* = 0.0094 (t = 2.874, df = 20) for the absolute time spent in the open arm (Figure 9C), with no differences detected for the latency to the first access to the open arm (*p* = 0.4491; t = 0.7721, df = 20) in the EPM (Figure 9D). Moreover, when the time spent in the open arms was analysed relative to the time spent in the closed arms (here referred to as analysis factor score), miRNA-132/212−/− animals exhibited significantly (\*\*) enhanced time spent in the open arms compared to their relative WT control counterparts (*p* = 0.0094; t = 2.872, df = 20) (Figure 9E). Taken together, all these observations encourage further research in order to independently verify the herein proposed potential involvement of the 132 and 212 miRNA cluster in the modulation of stress-related emotional responses mediated by steroid hormones belonging to the pituitary adrenocortical axis.

untreated or corticosterone-treated hippocampi of WT and miRNA-132/212<sup>−</sup>/<sup>−</sup> mice. Results represent fold changes relative to GAPDH. Data are shown as mean ± SEM (*n =* 4 animals per group).

Our data from the pharmacological treatment with corticosterone in hippocampal electrophysiology, as well as the western blot analyses of changes in the levels of molecules involved in the physiological responses to stress described here in response to corticosterone treatment, suggested that the miRNA 132/212 cluster was implicated in the regulation of emotional behaviors. Consequently, changes in anxiety-like behaviors were predicted to be detectable in miRNA-132/212−/− mice. In order to test this hypothesis, we conducted behavioral experiments using the open field test (OFT) and the elevated plus maze [EPM (Materials and Methods)]. As shown in Figure 9, no significant differences in the behavior of WT and miRNA-132/212−/− mice were observed in the OFT. In the OFT,

*Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 14 of 30

*2.9. Reduced Anxiety-like Behavior in miRNA-132/212−/− Mice* 

**Figure 9.** Performance of WT and miRNA-132/212<sup>−</sup>/<sup>−</sup> mice in the open field (OF) and EPM. In the OF, no differences in the behavior of the animals were apparent after examinations of the total distance **Figure 9.** Performance of WT and miRNA-132/212−/<sup>−</sup> mice in the open field (OF) and EPM. In the OF, no differences in the behavior of the animals were apparent after examinations of the total distance travelled (**A**) and cumulative time duration that the animals spent in the center (**B**). An *n* = 12 animals per group was used for the OF studies. In the EPM, miRNA-132/212−/<sup>−</sup> (KO) mice presented significantly increased time spent in the open arms (**C**), whereas no differences were detectable in the latency to the first enter to the open arms (**D**) compared to their WT littermates used as controls. (**E**) Analysis of the time spent in the open arms relative to closed arms showed enhanced time in the open arms for miRNA-132/212−/<sup>−</sup> mice compared to WT controls. An *n* = 10–12 animals per group was used for EPM studies, ns = not significantly different. *p* < 0.05 was considered significant. \*\* *p* < 0.01. Data are expressed as mean ± SEM.

#### **3. Discussion**

Steroid-hormones, including cortisol, have been proposed as modulators of the hippocampal functions in the context of mood-related disorders in humans [113,114]. Corticosterone is a very potent steroid-hormone influencing several neuronal functions (including neuronal morphology and synaptic plasticity in rodents), and also affecting learning and memory and emotional behaviors in a manner analogue to that of human cortisol [16–19]. In humans, the physiological relevance of neurohormonal regulators becomes even more apparent for its relation to conditions such as stress, depression and anxiety; and for its crosslink with other maladies and psychiatric disorders, some of which have been widely studied in experimental animal models [6–14]. In the amygdala, a brain region critical for the regulation of learning, memory, and emotional behaviors [115,116], regulation by steroid-hormone-mediated signaling has been previously described [63,117–119].

The hippocampus had also been shown to play pivotal roles not only in spatial learning and memory, but also in the processing of emotion-related memory storage [23–29,120]. Circa 123 years ago, Ramon y Cajal had indeed provided some of the first anatomical, pictographic renditions highlighting some of the distinctive morphological differences between the dorsal and the ventral hippocampus (see also [121]). However, experimental/unequivocal demonstrations about the relevance of specific regions of the hippocampus as selected areas important for the regulation of emotional behaviors started to emerge only about 15-20 years ago [121–127], and, notwithstanding of this, little continue to be known about the molecular mechanisms determining the region-specific effects of steroidhormones on the hippocampal function [30–33].

#### *3.1. miRNAs 132/212 in the Brain Neuronal Function*

Using a mouse model of learned safety, Ronovsky et. al., proposed that, in the amygdala, miR-132/212 regulate fear-inhibitory mechanisms as well as emotional-responses and plasticity-related synaptic functions [62]. Ronovsky et. al., also identified candidate proteins in the amygdala as potential gene targets for miRNA-132/212 which could participate in the miR-132/212-mediated regulation of mood-related behaviors [62]. We here expanded that line of research, by addressing the impact of miRNA-132/212 gene deletion on hippocampal synaptic functions and the effects to emotion-related steroid-hormone stimulation. Our data proposes the first functional description, to the best of our knowledge, for an involvement of miR-132/212 in the region-specific regulation of the effects of steroid-hormones on synaptic transmission and plasticity in the hippocampus.

Interestingly, other authors have found a long-term enhancement of corticosterone as well as impaired hippocampal synaptic plasticity in trauma-susceptible mice [128]. In line with these studies, and in agreement with our data, other reports have also shown a dysregulation of miRNAs in the hippocampus of trauma-susceptible mice [129]. Our behavioral examinations showing altered anxiety-like behavior in miRNA-132/212−/<sup>−</sup> animals provide further support to the possible involvement of the miRNAs 132/212 in mood-related behaviors, as they reveal the existence of basal differences in behavioral responses that are known to be influenced by neuroendocrine signaling. Discrepancies between data from different tests examining anxiety-like behaviors (e.g., using the open field) have precedents in the scientific literature. These differences can be due to factors such as the size of the experimental arenas (many different laboratories use open fields of different dimensions), differences in the different intensities of illumination used in different laboratories, or the use of partially reflective or even transparent materials in some arenas (see e.g., [130]). All these different subtle factors are known to influence the outcome in different experimental settings depending on different animal strains; depending on the type of pharmacological treatments; or depending on the genetic modifications introduced into the experimental subjects (see also [131–137]). Consequently, additional examinations (implementing other behavioral tests (e.g., Novelty suppressed feeding; Light/dark box, or Social Interaction Tests), are therefore encouraged.

#### *3.2. Proteins Associated with Mood and Steroid Hormone Signaling*

We here provide a basic biochemical analysis of proteins implicated in the regulation of steroid-hormone signaling, which propose the potential participation of the miRNAs 132/212 in the regulation of hippocampal steroid hormone signaling and emotionrelated behaviors. Our results are in line with previous observations from our group and others [53,112], which pointed towards a possible involvement of miRNA-132/212 in the processing of emotion-related functions. We had previously described that miRNA-132/212 deletion influences the expression levels of alpha7-nAChR hippocampal receptors [43], which, in the basolateral amygdala, are proposed to mediate emotional behaviors [138,139]. We had also described that miRNA-132/212 deletion influenced the effects of nicotine (a modulator of mood and anxiety [140]) on hippocampal synaptic plasticity [61]. Previous reports had also associated the function of SIRT1 with miR-132- and miR-212-mediated regulation in the context of aging and Alzheimer's Disease [84]. Moreover, the levels of miR-132 and miR-212 have been shown to become primarily augmented in the hippocampus in response to short-term (5 h) but not to long-lasting (15 days) exposure to stress, and depletion of the miR-132 has been shown to result in enhanced anxiety-like behaviors [112]. In this last report ([112]), the authors used single miR-132 and combined miRNA-132/212 depletion and showed, in both cases, altered expression levels for hippocampal SIRT1 and PTEN-2; two proteins known as targets of miR-132 that are involved in the modulation of anxiety-like behaviors. Interestingly, while the pioneering results from Aten, et.al., 2019 [112] already showed a marked, yet not significant, trend towards enhanced levels of SIRT1 in miRNA-132/212−/<sup>−</sup> mice hippocampi, their results are perfectly in line with our observations here showing (with different methods) significantly enhanced levels of SIRT1 in the hippocampus of miRNA-132/212−/<sup>−</sup> mice. Taken together, data derived from the pioneering studies from the group of Dr. Obrietan [112], together with our data here describing enhanced levels of SIRT1 in the hippocampi of miRNA-132/212−/<sup>−</sup> mice (as compared to their related WT counterparts), and also the absence of significant effects of corticosterone in the two groups showed by us, indicate that whereas the presence of miRNAs 132/212 might negatively regulate the levels of SIRT1, this action is not affected by corticosterone treatment.

Additionally, previous reports have proposed a role for the proteins CDK5 and PTEN in the regulation of brain neuronal synaptic functions. For example, PTEN has been associated to the regulation of the cognitive function (including social behaviors) likely via its capability to regulate brain growth and to modulate neuronal circuit formation and synaptic functions, as examined in cortico-amygdala synapses [94]. Other groups have established PTEN is involved in the regulation of emotion-related fear memories as well as spatial memory, and that PTEN is also a critical regulator of hippocampal LTP via its functional association with the protein CaMKII [95] (see also [96]). CDK5, on the other hand, has been functionally involved in the regulation of hippocampal dendrite morphology [101], in the regulation of both hippocampal neurotransmitter release and amplitude of hippocampal field EPSP slope [102], and CDK5 has been also proposed as a possible molecular regulator of amyloid beta production and in the mechanisms associated to the pathogenesis of Alzheimer's disease [98,103].

Our collaborative groups had also described PTEN as a potential target for miRNAs 132 and 212 in the amygdala [62]. Moreover, previous reports have described that the enhancement in the levels of miR-132 can induce augmented expression of CDK5 [141]. However, our findings here show comparable levels of PTEN and of CDK5 in WT and miRNA-132/212−/<sup>−</sup> mice hippocampi without detectable effects in both cases of corticosterone. These observations thus suggest that miRNA-132/212−/<sup>−</sup> might not play major roles in the regulation of the levels of PTEN and CDK5 in the mouse hippocampus. Nevertheless, given that total hippocampal tissue was examined here, further studies need to be conducted examining, separately, whether differences in the levels of these proteins or their transcripts might still exist when dorsal vs. ventral hippocampal regions are examined and compared using higher resolution techniques.

Here, we also show that whereas MSK1 shows comparable levels in WT and miRNA-132/212−/<sup>−</sup> hippocampi (and in both cases, their levels become strikingly enhanced by corticosterone), the levels of phospho-MSK1, on the contrary, become significantly reduced only in WT hippocampi. Interestingly, and in line with our results here, previous observations have proposed a BDNF-related regulation in the levels of the miRNA-132/212 clusters via MSK activity [56], all together thus accumulating experimental evidence in support for the existence of a possible functional hippocampal crosslink between miRNA-132/212−/<sup>−</sup> and MSK1 likely associated to the regulation of the effects of steroid hormones and, possibly, associated emotional and behavioral effects.

Additionally, we observed significantly enhanced levels of hippocampal CREB in miRNA-132/212−/<sup>−</sup> mice, together with a significantly stronger reduction of CREB levels in response to the corticosterone treatment (an effect of corticosterone also observed for pCREB) compared to WT mice hippocampi. Future research addressing the levels of pCREB under untreated condition might provide a larger picture of the impact of the miRNA 132/212 gene deletion on the hippocampal response to corticosterone and how this, comparatively, affects the levels of CREB phosphorylation (see also Figure 10). Abundant literature has linked CREB to hippocampal function in the context of stress ([92,119,142–145]) and anxiety ([80,146–149]). The link between CREB and members of the miRNA-132/-212 gene cluster has also been described before (see for example [150,151]). Similarly, an involvement of miR-132/212 in stress/anxiety-related behaviors as well as in the regulation of the levels of Sirt1 and PTEN have been also recently described [112]. Our data here showed that miRNA-132/212−/<sup>−</sup> deletion significantly changes the effects of corticosterone on the levels of CREB, which is a critical regulator of memory [152], but not of other proteins involved in the responses to stress (e.g., PTEN).

Our work thus provides unprecedent data suggesting that miRNA-132/212 might participate, in vivo, in the regulation of synaptic plasticity and mood-related behaviors by fine-tuning the effects of steroid hormones via regulation of the levels of specific geneproduct targets in an inter- and intra-brain-region selective manner (Figure 10). Further experiments are thus required in order to elucidate whether the levels of proteins that mediate in steroid hormone regulation are influenced by miRNA-132/212 in a subregion-specific manner. In the light of recent reports linking the hippocampus to learned safety [153,154], our findings also propose miRNA-132/212 as a potential modulator of learned safety through their capability to influence the levels of molecular targets responsive to neuroendocrine signals in both the hippocampus and in the amygdala (see also [62,155,156]). These observations encourage further verification in female subjects, as the existence of gender-specific effects of steroid-hormones on behavior, cognition, as well as on neuronal morphology and plasticity-related functions, have been established [157–165], and also brain microRNAs are regulated in a sex-dependent manner (see also [166–168]).

**Figure 10.** A possible role of the miRNA-132/212 family in the regulation of the effects of corticosterone on hippocampal synaptic plasticity. The hippocampus (**A,B**) is not a homogenous structure, as it presents several anatomical and functional differences between its dorsal and ventral regions. For example, the dorsal region has been shown to be primarily involved in spatial learning and memory functions, whereas the ventral region has been predominantly implicated in emotionrelated processing. Here, we postulate that changes in the levels of miRNA-132/212 (e.g., its downregulation, as illustrated by red arrows in (**C**), could result in changes in the properties of both shortterm and long-term forms of synaptic plasticity as well as in modulated metaplasticity (**C**). Alterations in the levels of miRNA-132/212 could also influence the synaptic responses to neuromodulators, such as corticosterone (**D**), which could, in a loop-like manner, impact the levels of miRNA-132/212 via activation of CREB, thus regulating the expression of proteins known to be downstream components of the glucocorticoid receptor (such as Sirt1 and MSK1). This work therefore proposes that miRNA-132/212 might contribute to increasing the in vivo mechanisms responsible for the functional heterogenicity between hippocampal regions, which might be critical to differentially distribute the effects of corticosterone on brain areas important for different cognitive functions. **4. Materials and Methods Figure 10.** A possible role of the miRNA-132/212 family in the regulation of the effects of corticosterone on hippocampal synaptic plasticity. The hippocampus (**A**,**B**) is not a homogenous structure, as it presents several anatomical and functional differences between its dorsal and ventral regions. For example, the dorsal region has been shown to be primarily involved in spatial learning and memory functions, whereas the ventral region has been predominantly implicated in emotion-related processing. Here, we postulate that changes in the levels of miRNA-132/212 (e.g., its downregulation, as illustrated by red arrows in (**C**), could result in changes in the properties of both short-term and long-term forms of synaptic plasticity as well as in modulated metaplasticity (**C**). Alterations in the levels of miRNA-132/212 could also influence the synaptic responses to neuromodulators, such as corticosterone (**D**), which could, in a loop-like manner, impact the levels of miRNA-132/212 via activation of CREB, thus regulating the expression of proteins known to be downstream components of the glucocorticoid receptor (such as Sirt1 and MSK1). This work therefore proposes that miRNA-132/212 might contribute to increasing the in vivo mechanisms responsible for the functional heterogenicity between hippocampal regions, which might be critical to differentially distribute the effects of corticosterone on brain areas important for different cognitive functions.

#### *4.1. Animals*  **4. Materials and Methods**

#### *4.1. Animals*

All the experiments reported here were performed using male adult (8–10 weeks old) wild type (WT) C57Bl/6 (substrain N) mice as well as knockout (KO) miRNA-132/212 (miRNA-132/212−/−) mice. These knockout mice, originally produced and described by the group of Dr. Pankratov [57], were engineered via the insertion of LoxP sites at non-coding regions (1st intron and exon 2) of the RNA gene encoding for the miRNAs 132 and 212, and generated in a C57Bl/6 background [57]. Here, as controls, only WT littermate animals All the experiments reported here were performed using male adult (8–10 weeks old) wild type (WT) C57Bl/6 (substrain N) mice as well as knockout (KO) miRNA-132/212 (miRNA-132/212−/−) mice. These knockout mice, originally produced and described by the group of Dr. Pankratov [57], were engineered via the insertion of LoxP sites at non-coding regions (1st intron and exon 2) of the RNA gene encoding for the miRNAs 132

and 212, and generated in a C57Bl/6 background [57]. Here, as controls, only WT littermate animals (as verified by PCR-based genotyping), were used. All experiments followed ethic directives of the Bundesministerium für Wissenschaft und Forschung of Austria (BMWF-66.009/0200-WF/V/3b/2016). ARRIVE and U.K. Animals usage guidelines (Scientific Procedures Act, 1986 and associated guidelines, EU Directive 2010/63/EU for animal experiments) were implemented. Animals were kept in standard Thoren Plexiglas cages housed in a colony-room with a temperature of (22 ± 2) ◦C. Since animal grouping has been reported to reduce stress and aggression [169], in agreement with reports examining the importance of cage size for animal wellbeing [170], 3–5 mice were grouped together per Thoren Plexiglass cage (with cages of ≈ 22 × 31 × 16 cm). Cages were located in a Thoren mouse vent rack with maximizer (Thoren, Hazleton, PA, USA). The housing room had an automatically controlled illumination system programmed on a 12 h light/dark cycle (with light switched on at 6:00 a.m. to deliver (200 ± 20) lux). All cages were provided with aspen wood bedding (ABEDD-LAB & VET Service, Vienna, Austria), and every cage was equipped with 2 layers of fragrance-free TORK Advance Soft paper (Tork AT, Essity Austria GmbH, Storchengasse 1, Vienna (1150) Austria) which animals consistently shredded and used as nesting material. Both bedding and nesting materials were entirely changed once a week. All animals had both food and water available ad libitum. All the animals were handled by expert technical personnel, and an expert veterinarian supervised animal management. While in this work we are only including male animals, our group is also currently conducting parallel experiments using female experimental subjects in order to examine the possible influence of sex as a critical factor mediating the effects of both corticosterone treatment and miRNA-132/212 gene deletions on the biophysical properties of hippocampal circuits (see also Discussion).

#### *4.2. Animal Grouping and Work Plan*

Experimental animals were randomly assigned to one of the different experimental groups studied, which were organized as follows: 2 groups were used for untreated controls (untreated WT vs. untreated KO), and 2 groups were used for corticosterone-treated (corticosterone-treated WT vs. corticosterone-treated KO). For the electrophysiological studies (see below) and all the four groups referred, the experiments were organized by sub-dividing the recordings conducted in the hippocampus into recordings obtained from the dorsal and ventral regions. For the western blot analysis, the experimental groups studied were organized as follows: 2 groups were used as untreated controls (untreated WT vs. untreated KO), and 2 groups were used for corticosterone-treated (corticosteronetreated WT vs. corticosterone-treated KO). It must be noted that for the WB experiments, the entire hippocampus was used; that is, the hippocampal tissue was not sub-divided into dorsal or ventral regions (see also the sections Discussion and Limitations of this work). For behavioral studies, animals were separated into two major groups (WT vs. KO), and they were not subjected to any pharmacological treatments. The number of subjects used in each experiment has been noted in the main text and/or in Figures Legends.

#### *4.3. Slice Electrophysiology*

For the preparation of hippocampal slices, the animals were subjected to mild sedation with low CO<sup>2</sup> inhalation followed by quick cervical dislocation and swift decapitation using a sharp-blade guillotining (DCAP-M, World Precision Instruments, Inc., Sarasota, FL, USA). Brains were extracted and placed on an ice-cold artificial Cerebrospinal Fluid (aCSF) solution containing (in mM): 125 NaCl, 2.5 CaCl2, 2.5 KCl, 1 MgCl2, 20 NaHCO3, 25 D-glucose, 1 NaH2PO<sup>4</sup> (pH 7.4). Hippocampi were transversally sliced using a McIlwain tissue chopper (TC752, Campden Instruments Ltd., Loughborough, UK) into 400 µm thick sections. Slices were transferred into an aCSF-filled recovery chamber submerged in a bath filled with water at 30 ◦C, where they recovered for at least 1 h before electrophysiology measurements. All the aCSF solutions were continuously supplied with a carbogen gas mixture (95% medical O<sup>2</sup> + 5% medical CO2). The slice electrophysiology measurements

were conducted with slices placed in a submerged recording chamber that was continuously supplied with 3–4 mL/min of a pre-carbogenated aCSF that had been pre-warmed at (30 ± 2) ◦C. Field excitatory postsynaptic potentials (fEPSPs) were obtained through heatpulled glass capillary pipettes (capillary glass from Harvard Apparatus, GmbH; Hugo Sachs Elektronik, Germany). Capillary tubes were pulled using a horizontal P-87 puller from Sutter Instrument (Model P-87, Novato, CA (94949), USA). Heat-pulled capillary pipettes were then back-filled with aCSF, yielding series resistances of (3 ± 1) MΩ. fEPSPs were registered from the hippocampal CA1 region, with recording electrodes positioned at the stratum–radiatum layer. fEPSPs were evoked by electrically stimulating the Schaffer's collateral projections that originated from the CA3 region, as described before [43,68]. To induce the fEPSPs, biphasic-square pulses of voltage were delivered through bipolar electrodes made of tungsten wire isolated to the tip with a Teflon coating layer (~50 µm diameter tips). The voltage pulses were generated from an ISO-STIM 01D stimulator (NPI Electronics, Tamm, Germany).

To examine the properties of basal synaptic transmission, input/output (I/O) curves were generated by plotting the raw amplitudes (and first-decaying phase slopes (normalized to maximum)) of the recorded output fEPSPs against the different values of delivered input voltages, consisting of increasing voltage steps of 200 µs, with pulses of 0–9 V delivered (in 1 V increments) with interpulse intervals of 15 s (see Figure 1).

In order to induce long-term potentiation (LTP), 5 separate sets of electrical stimulation were applied (500 ms apart), each comprising a total of 10 biphasic voltage pulses (100 µs/phase) delivered at 100 Hz (Figure 1; see also [66–68,171,172]). For LTP experiments, 10 min baseline field recordings were followed by the delivery of the LTP-inducing protocol; subsequently, an additional 40 min of field recordings were obtained, and a second LTP-inducing electrical stimulation protocol was delivered, followed by 30 min of field recordings. This protocol, thus, comprises an experimental way to induce hippocampal metaplasticity (see also [36,77]). Changes across time in the values of the slopes of the initial field decay (obtained by offline linear fittings of the recorded traces and normalized to baseline) were used as a measure of synaptic plasticity. LTP measurements were averaged from the values obtained from all slices within each animal, thus generating a single value per subject. For the electrophysiological recordings, the corticosterone (1 µM) treatment was conducted only for LTP experiments as follows: 10 min of baseline recordings > 1st high-frequency stimulation step > 10 min of recordings in ACSF solution > 30 min recordings in either ACSF solution (for untreated controls) or ACSF solution + 1 µM corticosterone applied in the bath (for the treatment groups) > 2nd high-frequency stimulation step and immediately start completely washing out the bath solution using only ACSF solution without corticosterone for the rest of the recordings. Recordings were obtained using an AxoClamp-2B amplifier, digitalized using the Digidata-1440 interface, and acquired and analyzed using the pClamp-10 (version 11.1) software (all from Axon Instruments, Molecular Devices, 660-665 Eskdale Rd, Winnersh, Triangle, Wokingham RG41 5TS, UK). Paired-pulse-induced plasticity examinations were also conducted by delivering voltages evoking ~50% of the maximum inducible field amplitude, as described before [43].

#### *4.4. Western Blotting*

In order to conduct the Western blot (WB) examinations, hippocampi from both WT and miR–132/212−/<sup>−</sup> mice (8–10 weeks old) were extracted and hippocampal slices prepared, allowed to recover for 1 h, and subsequently stimulated with 1 µM corticosterone or vehicle (Cat.Nr.27840, Sigma-Aldrich, Zimbagasse 5, Vienna (1140) Austria) for 1 h. Immediately after, the tissue was carefully transferred into Eppendorf® tubes and snapfrozen in liquid nitrogen. For WB preparations, the tissue was treated with a freshly prepared homogenizing protein lysis buffer with the following composition (in mM): 150 NaCl, 1 EDTA, 10 Tris-HCl, 10 NaF, 10 Na3VO4, 5 Na4P7O2, 0.5% Triton ×100, 1% SDS and the protease inhibitor cocktail cOmplete™ (Roche Diagnostics, Mannheim, Germany). Samples of protein-containing tissue extracts were subsequently weight-separated by 10%

SDS-PAGE gel electrophoresis. Separated protein contents were then transferred from the gels into polyvinylidene fluoride (PVDF) membranes, which were then subjected to 1 h of blocking treatment (5% BSA in TBST) at room temperature. The membranes were then exposed to an overnight bath treatment at 4 ◦C with the respective primary antibodies. The next day, the membranes were washed out and exposed to a 1 h bath treatment (room temperature) with the corresponding secondary antibodies. Indirect detection of the hippocampal protein levels was determined by chemiluminescence-reporter labeling of secondary antibodies and fluorescent image acquisition using a FluorChem HD2 system (Alpha Innotech, San Leandro, CA, USA). Obtained images were examined using the open-source image analysis software ImageJ (Version 1.53t) [173]. Protein levels from detected bands were quantified by densitometrical analysis with data normalized to values of GAPDH levels. The antibodies used were: CREB (Cell Signaling Technology (Danvers, MA, USA); USA, Cat. Nr. 9197); p-CREB (Cell Signaling Technology; USA, Cat. Nr. 9196s); Sirt1(Cell Signaling Technology; USA, Cat. Nr. 2028); MSK1 (Cell Signaling Technology; USA, Cat. Nr. 3489); p-MSK1 (Cell Signaling Technology; USA, Cat. Nr. 9595); CDK5 (Cell Signaling Technology; USA, Cat. Nr. 2506); Pten (Abcam (Biomedical Campus, Discovery Dr, Trumpington, Cambridge CB2 0AX, United Kingdom); UK, Cat. Nr. ab154812); GAPDH (Thermo Fisher Scientific. 168 Third Avenue. Waltham, MA USA 02451); USA, Cat. Nr. MA5-15738).

#### *4.5. Behavioral Testing*

All the behavioral experiments described here were conducted at morning hours, throughout the light phase of the imposed light/dark cycle, in a noise-isolated room. Before the start of the experiments, all the animals went through a period of daily handling for 5 min, conducted by the experimenter and implemented to familiarize the animals with the experimenter and the handling, thus reducing stress of the animals derived from the manipulations required for the experiments. On the testing days and before the beginning of the experiments, the animals were allowed to rest for 1 h in their home cages so that they could become habituated to the testing room.

#### *4.6. Open Field Test*

The Open Field test (OFT) is commonly used in experiments used with rodents in order to monitor basic locomotor activity as well as some aspects of anxiety-related behaviors [174]. Experiments in the OFT were conducted following protocols previously described by our group [65–67,175,176]. In brief, mice were placed in the center of an open field arena, illuminated at ~300 lux, consisting of a four white-mate plastic-walled box (30 × 30 × 30 cm). Animals were allowed to freely occupy the arena for 10 min, and were then returned to their home cages. The OFT box was then thoroughly cleaned using 70% ethanol and allowed to dry for several minutes before reuse. In order to examine the animals' behavior in the open field, the cumulative time that animals spent in the center of the arena as well as the total distance traveled were considered (Figure 9; see also [177]). All the OFT routines were digitalized using a zenithal high-definition digital video camera. Spatial/Temporal behavioral factors were evaluated using the XT12 version of the Ethovision software package (Noldus, Wageningen, The Netherlands).

#### *4.7. Elevated Plus Maze*

Behavioral measurements using the Elevated Plus Maze (EPM) (see Figure 9) were conducted as previously described [65,178]. This test continues to be widely used for the examination of anxiety-like behaviors in experimental animal models, including mice and rats, and relies on the instinctive fear response that animals display to open, elevated zones. In the case of mice and rats with healthy/natural self-preservation instincts, they will initially have the generalized tendence to avoid entering to open/elevated potentially dangerous areas, and would remain in or move towards enclosed zone in search for safety [179]. The experiments were conducted using a customized plus-shaped platform made of white-mate Plexiglas (elevated 50 cm above ground) having two of its opposing arms enclosed by 40 cm high walls. Time in the open arms as well as open arm latency to first were examined in the EPM. All arms had 50 cm in length and were 10 cm wide. The open arms were exposed to an illumination of ~110 lux, whereas ~15 lux illuminated the enclosed arms. The testing sessions (5 min each) begin immediately after having placed the animal at the central square area, with animals positioned facing towards one of the open arms, and animals were allowed to freely move across the maze settings. The EPM behavioral performance was digitalized using a zenithal high-definition digital video camera. Spatial/Temporal behavioral factors were evaluated using the XT12 version of the Ethovision software package (Noldus, Wageningen, The Netherlands). Behavioral parameters in the EPM were examined following protocols previously described [180].

#### *4.8. Statistical Analysis*

All data analyses were conducted using the GraphPad-Prism-9 software package (version 9.5.1(733), GraphPad Software, 225 Franklin Street. Fl. 26, Boston, MA 02110, USA). In order to comply with the 3Rs regulation and reduce as much as possible the number of animals, for the electrophysiological and behavioral analyses, statistical calculations were based on power analysis to estimate the minimum sample size required, using the G\*Power software (version 3.1.9.7). For biochemical studies, since the normalized data obtained from examinations using western blot are relative/semi-quantitative, we used a small number of animals in agreement with related published studies by other groups [181]. Unpaired two-tailed *t*-tests (with confidence level set to 95%) was used to examine differences of means between groups in behavioral experiments. Two-way ANOVA (Alpha 0.05) with Tukey's multiple comparisons test was used to examine data from WB experiments. Threeand two-way repeated measures (RM)-ANOVA with Tukey multiple comparisons, and/or Bonferroni's and Geisser–Greenhouse's corrections (and alpha set to 0.05), were used for the data derived from the electrophysiological analyses as in each case indicated in the main text. For bias control, data were examined by researchers who were blind to the experimental groups.

**Author Contributions:** S.K. conducted the electrophysiological recordings, analyzed data and contributed to the preparations of the final figures. M.C. conducted biochemical experiments and handled animals with assistance from M.-A.M. and M.C.M.-P. and E.C.-O. led behavioral experiments. T.S. and A.C. supervised data analyses of biochemical studies. E.A.M.A. helped with slice preparations. X.K., E.U. and G.L. provided crucial scientific advice. F.J.M. conceived, directed and funded the project and supervised the work of S.K., M.C.M.-P. and E.A.M.A. contributed to data analyses and wrote the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** F.J.M., S.K. and E.A.M.A. received support from the Austrian Science Fund/FWF (Project Nr. P\_31004 to F.J.M.).

**Institutional Review Board Statement:** Animal studies followed directives from the Bundesministerium für Wissenschaft und Forschung of Austria (BMWF-66.009/0200-WF/V/3b/2016).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** All the data derived from this work are included in the article.

**Conflicts of Interest:** Authors are not aware of any competing financial interests or other associations influencing this work.

#### **Abbreviations**

aCSF = artificial cerebrospinal fluid; EMP = elevated plus maze; fEPSPs = field excitatory postsynaptic potentials; HFS = high frequency stimulation; HPA = hypothalamic–pituitary–adrenal; KO = knockout; LTP = long-term potentiation; PTP = post-tetanic potentiation; I/O = input/output; miRNA = microRNA; OFT = open field test; WB = western blot; WT = wild type.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

## *Article* **Loss of CDKL5 Causes Synaptic GABAergic Defects That Can Be Restored with the Neuroactive Steroid Pregnenolone-Methyl-Ether**

**Roberta De Rosa <sup>1</sup> , Serena Valastro <sup>1</sup> , Clara Cambria <sup>2</sup> , Isabella Barbiero <sup>1</sup> , Carolina Puricelli <sup>1</sup> , Marco Tramarin <sup>1</sup> , Silvia Randi <sup>1</sup> , Massimiliano Bianchi 3,4, Flavia Antonucci <sup>2</sup> and Charlotte Kilstrup-Nielsen 1,\***


**Abstract:** CDKL5 deficiency disorder (CDD) is an X-linked neurodevelopmental disorder characterised by early-onset drug-resistant epilepsy and impaired cognitive and motor skills. CDD is caused by mutations in cyclin-dependent kinase-like 5 (CDKL5), which plays a well-known role in regulating excitatory neurotransmission, while its effect on neuronal inhibition has been poorly investigated. We explored the potential role of CDKL5 in the inhibitory compartment in *Cdkl5*-KO male mice and primary hippocampal neurons and found that CDKL5 interacts with gephyrin and collybistin, two crucial organisers of the inhibitory postsynaptic sites. Through molecular and electrophysiological approaches, we demonstrated that CDKL5 loss causes a reduced number of gephyrin puncta and surface exposed γ<sup>2</sup> subunit-containing GABA<sup>A</sup> receptors, impacting the frequency of miniature inhibitory postsynaptic currents, which we ascribe to a postsynaptic function of CDKL5. In line with previous data showing that CDKL5 loss impacts microtubule (MT) dynamics, we showed that treatment with pregnenolone-methyl-ether (PME), which promotes MT dynamics, rescues the above defects. The impact of CDKL5 deficiency on inhibitory neurotransmission might explain the presence of drug-resistant epilepsy and cognitive defects in CDD patients. Moreover, our results may pave the way for drug-based therapies that could bypass the need for CDKL5 and provide effective therapeutic strategies for CDD patients.

**Keywords:** CDKL5; inhibitory synapse; gephyrin; collybistin; GABAA receptor; pregnenolonemethyl-ether

### **1. Introduction**

Mutations in the X-linked cyclin-dependent kinase-like 5 (*CDKL5*) gene cause a severe neurodevelopmental disorder (Early Infantile Epileptic Encephalopahty, OMIM 300672), commonly referred to as CDKL5 deficiency disorder (CDD). CDD patients are characterised by intellectual disability, autistic features and drug-resistant epilepsy that normally manifest within the first three months of age [1]. *CDKL5* is mutated in approximately 1:50.000 live births, making CDD one of the most frequent causes of genetic epilepsy.

CDKL5 is a serine-threonine kinase that is highly abundant in the brain and whose expression peaks in the first postnatal weeks [2]. The constitutive loss of CDKL5 in CDD mouse models causes impaired learning and memory, altered locomotion and autistic-like features [3–6]. Spontaneous epilepsy has only recently been observed in aged heterozygous female mice [7,8] and in male mice harbouring the conditional knock-out (KO) of *Cdkl5* in glutamatergic forebrain neurons [9].

Various studies of *Cdkl5* mouse models and CDKL5 deficient primary neurons converge on the role of CDKL5 in regulating excitatory neurotransmission [6]. Indeed, CDKL5

**Citation:** De Rosa, R.; Valastro, S.; Cambria, C.; Barbiero, I.; Puricelli, C.; Tramarin, M.; Randi, S.; Bianchi, M.; Antonucci, F.; Kilstrup-Nielsen, C. Loss of CDKL5 Causes Synaptic GABAergic Defects That Can Be Restored with the Neuroactive Steroid Pregnenolone-Methyl-Ether. *Int. J. Mol. Sci.* **2023**, *24*, 68. https: //doi.org/10.3390/ijms24010068

Academic Editor: Giuseppina Martella

Received: 2 November 2022 Revised: 9 December 2022 Accepted: 16 December 2022 Published: 21 December 2022

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

interacts with the scaffolding protein PSD95, and CDKL5 deficient neurons are characterised by morphological and molecular alterations linked to excitatory synapses [10–12]. These defects depend in part on altered microtubule (MT) dynamics leading to the impaired invasion of MTs into dendritic spines and reduced spine maturation. Indeed, we previously demonstrated that CDKL5 regulates the MT-binding of the plus-end tracking protein CLIP170, thus impacting MT dynamics [13–15]. Importantly, the modulation of MT dynamics in vitro and in vivo, mediated by treatment with the pregnenolone analogue pregnenolone-methyl-ether (PME), which promotes CLIP170 functioning, can restore several CDKL5-related defects, including spine maturation, expression of synaptic proteins and hippocampal-dependent learning and memory [13–15].

Notwithstanding the importance of maintaining a proper excitation/inhibition balance, the possible role of CDKL5 in regulating inhibitory neurotransmission has 'til now been rather neglected. Mice with the conditional inactivation of *Cdkl5* in their glutamatergic forebrain neurons display altered miniature inhibitory synaptic currents (mIPSCs) [16], but a molecular basis for such a defect is still unknown.

Here, we report that CDKL5 interacts with the inhibitory postsynaptic scaffolding complex containing gephyrin and collybistin (CB) and show that CDKL5 loss leads to reduced levels and functioning of the γ<sup>2</sup> subunit-containing γ-aminobutyric acid type A receptors (GABAAR) in vitro and in vivo. GABAARs are heteropentameric Cl− permeable channels, generally composed of two α- and two β-subunits and a single γ- or δ-subunit [17]. Typically, synaptic GABAARs, which mediate phasic inhibition upon presynaptic GABA release, contain a γ-subunit; contrariwise, the extrasynaptic receptor complexes that mediate tonic inhibition as a response to low ambient GABA levels contain the δ-subunit [18]. Surface levels of synaptic GABAARs depend on a complex and dynamic regulation including transport, recycling and stabilisation, of which the latter relies on a scaffolding complex containing gephyrin and CB [19].

Alterations in any of the components linked to inhibitory neurotransmission are associated with various neurodevelopmental disorders characterised by cognitive defects, autism-like features and epilepsy [20]. The pharmacological targeting of the GABAergic system, based both on the allosteric modulation of GABAAR subtypes and on the functional control of GABAAR-associated proteins, represents an important interventional strategy against epilepsy and other clinical manifestations linked to the altered GABAAR levels or functioning [21].

We here show that treatment with PME can restore GABAAR γ<sup>2</sup> expression and functioning in primary *Cdkl5*-KO neurons. Importantly, we also find that the synaptic GABAAR subunit γ<sup>2</sup> is significantly reduced in hippocampi of *Cdkl5*-KO mice, but that its levels are normalised to those of WT animals upon treatment with PME. Altogether, these data suggest that CDKL5, through its interaction with the inhibitory scaffolding complex, regulates synaptic GABAAR levels and, importantly, that altered membrane insertion of GABAARs can be restored by the targeting of MT dynamics.

#### **2. Results**

#### *2.1. CDKL5 Interacts with the Gephyrin-Collybistin Complex and Regulates the Number of Postsynaptic Gephyrin Clusters*

To investigate the possible role of CDKL5 at the inhibitory synapses, we examined its interaction with gephyrin and CB, two proteins playing a fundamental role in the organisation of the postsynaptic sites of these synapses. To this aim, brain lysates of young mice at postnatal day 20–30 (PND20-30) were used to immunoprecipitate CDKL5 using IgGs as negative control; through the subsequent western blotting (WB), both gephyrin and CB could be detected in the CDKL5 immunocomplexes (Figure 1A,B). We further confirmed the interaction in a heterologous system expressing Flag-tagged CDKL5 in HEK293T cells together with the Myc-tagged CB2 isoform (Figure 1C). Upon immunoprecipitation of CDKL5, allowing the precipitation of both exogenous and endogenous CDKL5, overexpressed CB2 could easily be detected through WB analysis. Interestingly, the

weakly expressed endogenous CB could also be detected as interacting with endogenous CDKL5 that was visible upon a higher exposure of the membrane. These results confirm a recent report by Uezu et al. [22] in which CB was identified as a direct interactor of CDKL5 through a chemical-genetic proximity-labelling approach.

Gephyrin is a key protein in the organisation of inhibitory synapses through the formation of submembranous clusters that regulate the accumulation of GABAARs at the postsynaptic sites [23]. The accumulation of gephyrin under the cell membrane depends on the guanine nucleotide exchange factor CB that can recruit gephyrin to the postsynaptic sites. CB exists in a closed inactive conformation, which depends on its N-terminal SH3 domain [24]. When GFP-tagged gephyrin is expressed together with the SH3-containing CB2 derivative (CB2-SH3<sup>+</sup> ) in a heterologous system, it therefore accumulates in cytoplasmic deposits (Figure 1D), whereas deletion of the SH3 domain (CB-∆SH3) renders CB constitutively active, leading to the translocation of gephyrin to submembranous microaggregates. Interestingly, the exogenous expression of Flag-tagged CDKL5 in cells expressing CB2-SH3<sup>+</sup> causes GFP-gephyrin to accumulate under the cell membrane, suggesting that CDKL5 is capable of activating CB through the interaction of the two proteins (Figure 1E).

We further evaluated whether CDKL5 loss influences the total levels of gephyrin or CB in neurons. WB analyses on neuronal lysates from primary hippocampal cultures of *Cdkl5*-WT and -KO neurons, cultured for 14 days in vitro (DIV), did not reveal any changes in the expression of either gephyrin or CB (Figure 1F–I). We next examined whether CDKL5 loss affects the capacity of gephyrin to form the typical submembranous clusters in primary hippocampal neurons. Interestingly, we observed a significantly reduced number of gephyrin puncta in the dendritic segments of *Cdkl5*-KO neurons (Figure 1J,K). The interaction of CDKL5 with the gephyrin-CB complex, together with the reduced number of gephyrin clusters in *Cdkl5*-KO neurons, indicate that CDKL5 might play a hitherto undescribed role in the organisation of the postsynaptic sites of inhibitory synapses.

#### *2.2. CDKL5 Loss Affects the Membrane Levels of γ<sup>2</sup> Subunit-Containing GABAARs and Impairs mIPSCs*

Gephyrin deficiency impacts the surface accumulation of γ2-containing GABAARs [25–27]. We therefore speculated that the reduced number of gephyrin clusters in *Cdkl5*-KO neurons might influence the cell-surface expression of the GABAAR subunit γ2. To address this, we performed a cell-surface biotinylation assay. Briefly, upon biotinylation of hippocampal neurons, labelled proteins were affinity purified, and the amount of the synaptic GABAAR subunit γ<sup>2</sup> was analysed through WB in parallel with a fraction of the total cell extract (Figure 2A). The proportion of the total receptor pool that resides in the neuronal surface was determined by quantifying the ratio of the biotinylated fraction (surface) over the amount in the total lysate (total). As control of the biotinylation procedure, we verified that the cytoplasmic protein GAPDH was barely present in the pool of affinity-purified proteins. As illustrated in the graph in Figure 2B, we found that *Cdkl5*-KO neurons displayed reduced surface levels of the γ<sup>2</sup> subunit, whereas its total levels were unaltered (Figure 2C,D). We further examined the surface expression of the γ<sup>2</sup> subunit through the immunofluorescence staining of hippocampal *Cdkl5*-WT and -KO neurons performed under non-permeabilising conditions. In accordance with the biotinylation experiment, *Cdkl5*-KO neurons displayed significantly reduced surface levels of the γ<sup>2</sup> subunit (Figure 2E,F).

CDKL5 through a chemical-genetic proximity-labelling approach.

mice at postnatal day 20–30 (PND20-30) were used to immunoprecipitate CDKL5 using IgGs as negative control; through the subsequent western blotting (WB), both gephyrin and CB could be detected in the CDKL5 immunocomplexes (Figure 1A,B). We further confirmed the interaction in a heterologous system expressing Flag-tagged CDKL5 in HEK293T cells together with the Myc-tagged CB2 isoform. (Figure 1C). Upon immunoprecipitation of CDKL5, allowing the precipitation of both exogenous and endogenous CDKL5, overexpressed CB2 could easily be detected through WB analysis. Interestingly, the weakly expressed endogenous CB could also be detected as interacting with endogenous CDKL5 that was visible upon a higher exposure of the membrane. These results confirm a recent report by Uezu et al. [22] in which CB was identified as a direct interactor of

**Figure 1.** CDKL5 interacts with the gephyrin-collybistin complex. (**A**,**B**) Representative WBs showing the coimmuniprecipitation of CDKL5 and gephyrin (GPHN; A) or collybistin (CB; B) from **Figure 1.** CDKL5 interacts with the gephyrin-collybistin complex. (**A**,**B**) Representative WBs showing the coimmuniprecipitation of CDKL5 and gephyrin (GPHN; A) or collybistin (CB; B) from PND20-30 mouse brain lysates. Unrelated IgGs were used as negative control. Whole brain lysate (input; 2.5%) and immunocomplexes were analysed through WB with antibodies against CDKL5, GPHN (**A**) and CB (**B**). n = 3. (**C**) Flag-CDKL5 was expressed in HEK293T cells together with Myc-CB2 and immunoprecipitated with a monoclonal anti-CDKL5 antibody. Whole cell lysates (input; 3%) and immunoprecipitated proteins were analysed through WB using antibodies against CDKL5, CB and, as loading control, GAPDH. n = 3. (**D**,**E**) Representative confocal images of COS7 cells expressing GFPtagged gephyrin (GFP-GPHN, green) and Myc-CB2-SH3<sup>+</sup> or Myc-CB2-∆SH3 (blue) together with Flag-CDKL5 (red). Scale bar: 5 µm. (**F**,**H**) Representative WB of whole cell lysates of *Cdkl5*-WT/KO primary hippocampal neurons at DIV14. Antibodies against GPHN and CB were used together with anti-CDKL5 and, as loading control, GAPDH. (**G**,**I**) The graphs show the quantification of normalised GPHN (**G**) and CB (**I**) levels. n = 8 biological replicates. Mean ± SEM. Not significant (ns), *p* > 0.05. Unpaired Student's *t*-test. (**J**) Representative images of DIV14 *Cdkl5*-WT/KO hippocampal neurons stained with antibodies against GPHN (red) and the dendritic marker MAP2 (blue). Scale bar, 5 µm. (**K**) Graph showing the quantification of GPHN puncta along 30 µm long segments. n = 7 biological replicates, N = 37/35 WT/KO neurons. Mean ± SEM. \*\*\*\* *p* < 0.0001. Unpaired Student's *t*-test.

unit (Figure 2E,F).

through the immunofluorescence staining of hippocampal *Cdkl5*-WT and -KO neurons performed under non-permeabilising conditions. In accordance with the biotinylation experiment, *Cdkl5*-KO neurons displayed significantly reduced surface levels of the γ<sup>2</sup> sub-

**Figure 2.** CDKL5 loss affects surface expression of synaptic GABAARs in primary hippocampal neurons. (**A**) Representative WB of a biotinylation experiment on *Cdkl5*-WT/KO primary hippocampal neurons at DIV14. Levels of GABAAR subunit γ<sup>2</sup> were analysed together with GAPDH in the surface fraction, obtained from 500 µg of lysate, and in 30 µg of whole cell lysate. (**B**) Graph showing the ratio of surface/total levels of GABAAR γ2. N = 5 biological replicates. Mean ± SEM \* *p* < 0.05. Unpaired Student's *t*-test. (**C**) Representative WB analysis of whole cell lysates of *Cdkl5*-WT/KO primary hippocampal neurons at DIV14. GAPDH was used as loading control. (**D**) Graph showing the quantification of normalised GABAAR γ<sup>2</sup> levels. N = 8 biological replicates. Mean ± SEM ns, *p* > 0.05. Unpaired Student's *t*-test. (**E**) Surface exposed GABAAR γ<sup>2</sup> (green, sGABAAR γ2) was detected through immunostaining of *Cdkl5*-WT/KO primary hippocampal neurons at DIV14 under non-permeabilising conditions. MAP2 (blue) was used as dendritic marker. Scale bar, 5 µm. (**F**) Graph showing the quantification of the fluorescence intensity of sGABAAR γ<sup>2</sup> staining of 30 µm long dendritic segments. n = 6 biological replicates, N = 33/42 WT/KO neurons. Mean ± SEM \*\*\*\* *p* < 0.0001. Unpaired Student's *t*-test. **Figure 2.** CDKL5 loss affects surface expression of synaptic GABAARs in primary hippocampal neurons. (**A**) Representative WB of a biotinylation experiment on *Cdkl5*-WT/KO primary hippocampal neurons at DIV14. Levels of GABAAR subunit γ<sup>2</sup> were analysed together with GAPDH in the surface fraction, obtained from 500 µg of lysate, and in 30 µg of whole cell lysate. (**B**) Graph showing the ratio of surface/total levels of GABAAR γ<sup>2</sup> . N = 5 biological replicates. Mean ± SEM \* *p* < 0.05. Unpaired Student's *t*-test. (**C**) Representative WB analysis of whole cell lysates of *Cdkl5*-WT/KO primary hippocampal neurons at DIV14. GAPDH was used as loading control. (**D**) Graph showing the quantification of normalised GABAAR γ<sup>2</sup> levels. N = 8 biological replicates. Mean ± SEM ns, *p* > 0.05. Unpaired Student's *t*-test. (**E**) Surface exposed GABAAR γ<sup>2</sup> (green, sGABAAR γ<sup>2</sup> ) was detected through immunostaining of *Cdkl5*-WT/KO primary hippocampal neurons at DIV14 under non-permeabilising conditions. MAP2 (blue) was used as dendritic marker. Scale bar, 5 µm. (**F**) Graph showing the quantification of the fluorescence intensity of sGABAAR γ<sup>2</sup> staining of 30 µm long dendritic segments. n = 6 biological replicates, N = 33/42 WT/KO neurons. Mean ± SEM \*\*\*\* *p* < 0.0001. Unpaired Student's *t*-test.

To evaluate the functional consequences of the reduced levels of the γ2-containing GABAARs on the neuronal surface, we measured mIPSCs from *Cdkl5*-WT and -KO hippocampal neurons at DIV14. Notably, in *Cdkl5*-KO neurons, mIPSCs were less frequent, while the amplitude was unaltered (Figure 3A–C). The reduced frequency of mIPSCs points to possible presynaptic deficits in the *Cdkl5*-KO neurons, and we therefore analysed the presynaptic markers bassoon and vesicular GABA transporter (VGAT). Bassoon is localised at the active zone of both excitatory and inhibitory presynaptic terminals, while VGAT is specific for the inhibitory presynaptic sites. Immunofluorescence staining To evaluate the functional consequences of the reduced levels of the γ2-containing GABAARs on the neuronal surface, we measured mIPSCs from *Cdkl5*-WT and -KO hippocampal neurons at DIV14. Notably, in *Cdkl5*-KO neurons, mIPSCs were less frequent, while the amplitude was unaltered (Figure 3A–C). The reduced frequency of mIPSCs points to possible presynaptic deficits in the *Cdkl5*-KO neurons, and we therefore analysed the presynaptic markers bassoon and vesicular GABA transporter (VGAT). Bassoon is localised at the active zone of both excitatory and inhibitory presynaptic terminals, while VGAT is specific for the inhibitory presynaptic sites. Immunofluorescence staining showed reduced numbers of both bassoon and VGAT puncta in *Cdkl5*-KO hippocampal neurons (Figure 3D–G).

neurons (Figure 3D–G).

showed reduced numbers of both bassoon and VGAT puncta in *Cdkl5*-KO hippocampal

**Figure 3.** CDKL5 loss leads to a reduction in the frequency of mIPSCs and in presynaptic inhibitory markers. (**A**) Representative traces of mIPSCs recorded from *Cdkl5*-WT/KO primary hippocampal neurons at DIV14 patching at +10 mV. (**B**,**C**) Graphs showing mIPSC frequency (**B**) and amplitude (**C**) in *Cdkl5*-WT/KO cultures. n = 3 biological replicates; frequency: N = 16/11 WT/KO neurons. Mean ± SEM \* *p* < 0.05, not significant (ns), *p* > 0.05. mIPSC frequency: Mann Whitney U test; mIPSC amplitude: Unpaired Student's *t*-test. (**D**,**F**) Representative images of *Cdkl5-*WT/KO neurons at DIV14 stained for Bassoon (red, D), VGAT (green, F) and MAP2 (blue). Scale bar, 5 μm. (**E**,**G**) Graphs showing the quantification of bassoon and VGAT puncta along 30 μm long dendritic segments. Bassoon: n = 7 biological replicates, N = 40/44 neurons. VGAT: n = 3 biological replicates, N = 30/26 neurons. Mean ± SEM \*\* *p* < 0.01, \*\*\*\* *p* < 0.0001. Unpaired Student's *t*-test. (**H**) Representative traces of mIPSCs recorded from CDKL5-silenced neurons (shCDKL5) and controls (shLacZ). Neurons were transfected with GFP-expressing shRNA vectors at DIV11 and mIPSCs were recorded at DIV14 patching at +10 mV. (**I**,**J**) Graphs showing mIPSC frequency (**I**) and amplitude (**J**) upon acute CDKL5-silencing. n = 3 biological replicates; N = 11/16 shLacZ/shCDKL5 neurons. Mean ± SEM \* *p* < 0.05; ns, *p* > 0.05. Unpaired Student's *t*-test. **Figure 3.** CDKL5 loss leads to a reduction in the frequency of mIPSCs and in presynaptic inhibitory markers. (**A**) Representative traces of mIPSCs recorded from *Cdkl5*-WT/KO primary hippocampal neurons at DIV14 patching at +10 mV. (**B**,**C**) Graphs showing mIPSC frequency (**B**) and amplitude (**C**) in *Cdkl5*-WT/KO cultures. n = 3 biological replicates; frequency: N = 16/11 WT/KO neurons. Mean ± SEM \* *p* < 0.05, not significant (ns), *p* > 0.05. mIPSC frequency: Mann Whitney U test; mIPSC amplitude: Unpaired Student's *t*-test. (**D**,**F**) Representative images of *Cdkl5*-WT/KO neurons at DIV14 stained for Bassoon (red, D), VGAT (green, F) and MAP2 (blue). Scale bar, 5 µm. (**E**,**G**) Graphs showing the quantification of bassoon and VGAT puncta along 30 µm long dendritic segments. Bassoon: n = 7 biological replicates, N = 40/44 neurons. VGAT: n = 3 biological replicates, N = 30/26 neurons. Mean ± SEM \*\* *p* < 0.01, \*\*\*\* *p* < 0.0001. Unpaired Student's *t*-test. (**H**) Representative traces of mIPSCs recorded from CDKL5-silenced neurons (shCDKL5) and controls (shLacZ). Neurons were transfected with GFP-expressing shRNA vectors at DIV11 and mIPSCs were recorded at DIV14 patching at +10 mV. (**I**,**J**) Graphs showing mIPSC frequency (**I**) and amplitude (**J**) upon acute CDKL5-silencing. n = 3 biological replicates; N = 11/16 shLacZ/shCDKL5 neurons. Mean ± SEM \* *p* < 0.05; ns, *p* > 0.05. Unpaired Student's *t*-test.

The interaction of CDKL5 with the postsynaptic scaffolding proteins led us to test whether the ablation of CDKL5 at the postsynaptic site would be sufficient to generate the observed reduction in mIPSC frequency. We therefore recorded the mIPSCs in neurons transfected with a construct expressing a short-hairpin RNA specific for CDKL5 (shCDKL5) or, as control, against LacZ (shLacZ). Transfected neurons are easily detectable thanks to the concomitant expression of GFP from these vectors; moreover, due to the low transfection efficiency (below 5%), the synaptic input to transfected neurons is generated from non-silenced cells. Interestingly, the mIPSC frequency was reduced in shCDKL5 neurons as compared to the shLacZ controls (Figure 3H–J), thus supporting a postsynaptic CDKL5-dependent effect on presynaptic inputs. The interaction of CDKL5 with the postsynaptic scaffolding proteins led us to test whether the ablation of CDKL5 at the postsynaptic site would be sufficient to generate the observed reduction in mIPSC frequency. We therefore recorded the mIPSCs in neurons transfected with a construct expressing a short-hairpin RNA specific for CDKL5 (shCDKL5) or, as control, against LacZ (shLacZ). Transfected neurons are easily detectable thanks to the concomitant expression of GFP from these vectors; moreover, due to the low transfection efficiency (below 5%), the synaptic input to transfected neurons is generated from nonsilenced cells. Interestingly, the mIPSC frequency was reduced in shCDKL5 neurons as compared to the shLacZ controls (Figure 3H–J), thus supporting a postsynaptic CDKL5 dependent effect on presynaptic inputs.

#### *2.3. GABAergic Defects in Cdkl5-KO Neurons Are Normalised upon Treatment with PME*

The presence of GABAARs on the neuronal membrane is determined by the dynamic balance of delivery, recycling and degradation, which altogether depends on complex regulatory mechanisms involving the interaction of various proteins, including gephyrin, with MTs [28,29]. We previously showed that various neuronal defects linked to CDKL5 deficiency in vitro and in vivo can be restored through treatment with PME, which promotes MT dynamics [13–15]. We therefore found it intriguing to analyse whether the treatment of *Cdkl5*-KO neurons with PME could restore the observed defects at the inhibitory synapse. *Cdkl5*-WT and -KO neurons were treated with 0.3 and 1 µM PME at DIV11 and stained at DIV14 for gephyrin, GABAAR γ2, bassoon or VGAT (Figure 4A–H). Whereas treatment with PME affected neither of the pre- and postsynaptic markers in WT neurons, we observed a significant effect already with 0.3 µM of PME in *Cdkl5*-KO neurons. Indeed, at the postsynaptic site, both the gephyrin and the GABAAR γ<sup>2</sup> levels were restored to those in WT neurons (Figure 4A–D); a similar positive effect was observed also with the presynaptic markers bassoon and VGAT (Figure 4E–H). In line with this, treatment with PME could also normalise the frequency of mIPSCs in *Cdkl5*-KO neurons, whereas no effect was observed on the amplitude (Figure 4I–K).

By treating symptomatic *Cdkl5*-KO mice with PME, we previously found that CDKL5 related hippocampal-dependent behavioural and excitatory synaptic defects benefit from increased MT dynamics in vivo [15]. Intrigued by the positive effect of PME on inhibitory synapses in vitro, we proceeded to evaluate its effect in vivo, also. With this aim, we subjected hippocampal sections of *Cdkl5*-WT and -KO mice treated with 10 mg/kg of PME for 7 days (starting at PND60) to immunofluorescence staining against GABAAR γ2. As shown in Figure 5 we observed a dramatic decrease in the density of GABAAR γ<sup>2</sup> clusters in the dentate gyrus of the vehicle-treated *Cdkl5*-KO mice with respect to the WT mice; interestingly, upon treatment with PME, the number of GABAAR γ<sup>2</sup> clusters was similar between the two genotypes. *Int. J. Mol. Sci.* **2023**, *24*, 68 8 of 17

tude: ns, *p* > 0.05. One-way ANOVA, followed by Tukey's post-hoc.

**Figure 4.** Treatment with PME normalises CDKL5-dependent inhibitory synaptic defects. (**A**,**C**,**E**,**G**) Representative images of *Cdkl5*-WT/KO primary hippocampal neurons stained for MAP2 together with gephyrin (GPHN, red; A), GABAAR γ<sup>2</sup> (green, C), bassoon (red, E) or VGAT (green, G) upon treatment with 0.3 or 1 µM of PME or vehicle (0.1% DMSO) at DIV11 for 72 h. Scale bar, 5 µm. (**B**,**D**,**F**,**H**) Graphs showing the quantification of GPHN puncta (**B**), sGABAAR γ2 fluorescence intensity (**D**), bassoon puncta (**F**), VGAT puncta (**H**) along 30 µm long dendritic segments. GPHN: n = 6 biological replicates, N ≥ 19 neurons; sGABAAR γ2: n ≥ 5 biological replicates, N ≥ 32 neurons; bassoon: n = 3 biological replicates, N ≥ 20 neurons; VGAT: n = 3 biological replicates, N ≥ 25 neurons. Mean ± SEM \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001; \*\*\*\* *p* < 0.0001. Two-way ANOVA, followed by Tukey's post-hoc. (**I**) Representative traces of mIPSCs in *Cdkl5*-WT and -KO primary hippocampal neurons treated as indicated with either vehicle (0.1% DMSO) or 0.3 µM PME for 72 h starting at DIV11. (**J**,**K**) Graphs showing mIPSC frequency (**J**) and amplitude (**K**) of *Cdkl5*-WT and -KO neurons treated as indicated. n = 3 biological replicates, N ≥ 22. Mean ± SEM. mIPSC frequency: \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001. Kruskal-Wallis test, followed by Dunn's multiple comparisons test. mIPSC ampli-

**Figure 4.** Treatment with PME normalises CDKL5-dependent inhibitory synaptic defects. (**A**,**C**,**E**,**G**) Representative images of *Cdkl5*-WT/KO primary hippocampal neurons stained for MAP2 together with gephyrin (GPHN, red; A), GABAAR γ<sup>2</sup> (green, C), bassoon (red, E) or VGAT (green, G) upon treatment with 0.3 or 1 µM of PME or vehicle (0.1% DMSO) at DIV11 for 72 h. Scale bar, 5 µm. (**B**,**D**,**F**,**H**) Graphs showing the quantification of GPHN puncta (**B**), sGABAAR γ2 fluorescence intensity (**D**), bassoon puncta (**F**), VGAT puncta (**H**) along 30 µm long dendritic segments. GPHN: n = 6 biological replicates, N ≥ 19 neurons; sGABAAR γ2: n ≥ 5 biological replicates, N ≥ 32 neurons; bassoon: n = 3 biological replicates, N ≥ 20 neurons; VGAT: n = 3 biological replicates, N ≥ 25 neurons. Mean ± SEM \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001; \*\*\*\* *p* < 0.0001. Two-way ANOVA, followed by Tukey's post-hoc. (**I**) Representative traces of mIPSCs in *Cdkl5*-WT and -KO primary hippocampal neurons treated as indicated with either vehicle (0.1% DMSO) or 0.3 µM PME for 72 h starting at DIV11. (**J**,**K**) Graphs showing mIPSC frequency (**J**) and amplitude (**K**) of *Cdkl5*-WT and -KO neurons treated as indicated. n = 3 biological replicates, N ≥ 22. Mean ± SEM. mIPSC frequency: \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001. Kruskal-Wallis test, followed by Dunn's multiple comparisons test. mIPSC amplitude: ns, *p* > 0.05. One-way ANOVA, followed by Tukey's post-hoc. **Figure 4.** Treatment with PME normalises CDKL5-dependent inhibitory synaptic defects. (**A**,**C**,**E**,**G**) Representative images of *Cdkl5*-WT/KO primary hippocampal neurons stained for MAP2 together with gephyrin (GPHN, red; A), GABAAR γ<sup>2</sup> (green, C), bassoon (red, E) or VGAT (green, G) upon treatment with 0.3 or 1 µM of PME or vehicle (0.1% DMSO) at DIV11 for 72 h. Scale bar, 5 µm. (**B**,**D**,**F**,**H**) Graphs showing the quantification of GPHN puncta (**B**), sGABAAR γ<sup>2</sup> fluorescence intensity (**D**), bassoon puncta (**F**), VGAT puncta (**H**) along 30 µm long dendritic segments. GPHN: n = 6 biological replicates, N ≥ 19 neurons; sGABAAR γ<sup>2</sup> : n ≥ 5 biological replicates, N ≥ 32 neurons; bassoon: n = 3 biological replicates, N ≥ 20 neurons; VGAT: n = 3 biological replicates, N ≥ 25 neurons. Mean ± SEM \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001; \*\*\*\* *p* < 0.0001. Two-way ANOVA, followed by Tukey's post-hoc. (**I**) Representative traces of mIPSCs in *Cdkl5*-WT and -KO primary hippocampal neurons treated as indicated with either vehicle (0.1% DMSO) or 0.3 µM PME for 72 h starting at DIV11. (**J**,**K**) Graphs showing mIPSC frequency (**J**) and amplitude (**K**) of *Cdkl5*-WT and -KO neurons treated as indicated. n = 3 biological replicates, N ≥ 22. Mean ± SEM. mIPSC frequency: \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001. Kruskal-Wallis test, followed by Dunn's multiple comparisons test. mIPSC amplitude: ns, *p* > 0.05. One-way ANOVA, followed by Tukey's post-hoc. *Int. J. Mol. Sci.* **2023**, *24*, 68 9 of 17 By treating symptomatic *Cdkl5*-KO mice with PME, we previously found that CDKL5-related hippocampal-dependent behavioural and excitatory synaptic defects benefit from increased MT dynamics in vivo [15]. Intrigued by the positive effect of PME on inhibitory synapses in vitro*,* we proceeded to evaluate its effect in vivo, also. With this aim, we subjected hippocampal sections of *Cdkl5*-WT and -KO mice treated with 10 mg/kg of PME for 7 days (starting at PND60) to immunofluorescence staining against GABAAR γ2. As shown in Figure 5 we observed a dramatic decrease in the density of GABAAR γ<sup>2</sup> clusters in the dentate gyrus of the vehicle-treated *Cdkl5*-KO mice with respect to the WT mice; interestingly, upon treatment with PME, the number of GABAAR γ<sup>2</sup> clusters was similar between the two genotypes.

**Figure 5.** Reduced levels of synaptic GABAAR γ2 in hippocampi of *Cdkl5*-KO mice are normalised upon treatment with PME. (**A**) Hippocampal slices (dentate gyrus) from *Cdkl5*-WT/KO male mice treated with PME (10 mg/kg, s.c. for seven days starting from PND60) or sesame oil (vehicle) were stained for GABAAR γ2. Scale bar, 5 µm. (**B**) Representative image of a hippocampal section stained with the nuclear dye DAPI. Images from the molecular layer (MOL) of the dentate gyrus (DG) were used for the analyses shown in panels A and C. (**C**) Graph showing the quantification of GABAAR **Figure 5.** Reduced levels of synaptic GABAAR γ<sup>2</sup> in hippocampi of *Cdkl5*-KO mice are normalised upon treatment with PME. (**A**) Hippocampal slices (dentate gyrus) from *Cdkl5*-WT/KO male mice treated with PME (10 mg/kg, s.c. for seven days starting from PND60) or sesame oil (vehicle) were stained for GABAAR γ<sup>2</sup> . Scale bar, 5 µm. (**B**) Representative image of a hippocampal section stained with the nuclear dye DAPI. Images from the molecular layer (MOL) of the dentate gyrus (DG) were used for the analyses shown in panels A and C. (**C**) Graph showing the quantification of GABAAR γ<sup>2</sup> clusters/100 µm<sup>2</sup> . n ≥ 3. Mean±SEM, \*\* *p* < 0.01, \*\*\* *p* < 0.001. Two-way ANOVA followed by Tukey's post-hoc.

study reveal a hitherto undescribed function of CDKL5 in the inhibitory compartment. Our results suggest that CDKL5, through its interaction with the inhibitory scaffolding complex containing gephyrin and CB, regulates membrane levels of synaptic γ2-containing GABAARs. Deranged GABAAR levels are frequently linked to cognitive deficits and epilepsy, and we speculate that our results may help in explaining the seizure phenotype

. n ≥ 3. Mean±SEM, \*\* *p* < 0.01, \*\*\* *p* < 0.001. Two-way ANOVA followed by

γ2 clusters/100 µm<sup>2</sup>

observed in CDD patients.

Tukey's post-hoc.

**3. Discussion**

#### **3. Discussion**

In this study, we investigated the possible role of CDKL5 in the inhibitory synapse. Beyond the well-established role of CDKL5 at glutamatergic synapses, the results of this study reveal a hitherto undescribed function of CDKL5 in the inhibitory compartment. Our results suggest that CDKL5, through its interaction with the inhibitory scaffolding complex containing gephyrin and CB, regulates membrane levels of synaptic γ2-containing GABAARs. Deranged GABAAR levels are frequently linked to cognitive deficits and epilepsy, and we speculate that our results may help in explaining the seizure phenotype observed in CDD patients.

#### *3.1. CDKL5 Deficiency Leads to Dysfunctions in the Inhibitory Synapse*

GABAARs are ion channels permeable to chloride and bicarbonate ions and are, in the mammalian CNS, localised at postsynaptic inhibitory specialisations or at extrasynaptic sites, where they mediate inhibitory neurotransmission [29,30].

Herein, we analysed the synaptic GABAARs, focusing our attention on the γ<sup>2</sup> subunit, which is the most abundant GABAAR subunit in the rat brain [17]. Our immunofluorescence analyses showed a significant reduction of GABAAR γ<sup>2</sup> surface expression in both the hippocampal neurons and hippocampal slices of *Cdkl5*-KO mice. This defect was corroborated by biochemical approaches. Indeed, through the highly sensitive cell-surface biotinylation assay, we showed that the absence of CDKL5 influenced surface levels of γ2-containing GABAARs in primary cultures of *Cdkl5*-KO neurons. Since the absence of CDKL5 did not affect the total levels of the GABAAR γ<sup>2</sup> subunit, the reduced surface levels might be a consequence of an altered transport or recycling.

The cell membrane distribution of synaptic GABAARs is dynamically regulated through various mechanisms, including the subsynaptic scaffolding factor gephyrin, which binds and clusters synaptic GABAARs at sites directly opposite to GABA-releasing axon terminals [23]. Through in vitro and ex vivo immunoprecipitation experiments, we found that CDKL5 forms a complex with both CB and gephyrin. Furthermore, the reduction of surface expressed synaptic GABAARs was accompanied by a reduction in the number of gephyrin-positive puncta in the *Cdkl5*-KO primary cultures.

The molecular interaction between CDKL5 and the cytoplasmic CB-gephyrin complex is likely a key mechanism through which CDKL5 exerts its control on synaptic GABAARs. In particular, gephyrin takes part in the aggregation, but not in the surface insertion or stabilisation, of α<sup>2</sup> and γ<sup>2</sup> subunit-containing GABAARs [31]. CB is a brain-specific GDP/GTP-exchange factor, which interacts with gephyrin and regulates its recruitment from intracellular deposits to postsynaptic membranes [32]. The loss of CB leads to a strong reduction in gephyrin and synaptic GABAAR clusters in several regions of the forebrain, including the hippocampus, amygdala and cerebellum [33,34]. Normally, CB is present in an auto-inhibited conformation and depends on other neuronal factors such as neuroligin 2 (NL2), GABAAR subunit α<sup>2</sup> or the Rho-like GTPase TC-10 for its activation [35,36]. Our results, showing that exogenous CDKL5 expression is sufficient for localising gephyrin under the cell membrane when coexpressed with full-length CB, suggest that CDKL5 may also be capable of relieving CB from the inhibited conformation. Our data also support a previous study by Uezu et al. [22] that identified CDKL5 as a direct interactor of CB. The presence of gephyrin in the immunocomplexes is likely caused by an indirect interaction mediated by CB. Since CB is required for inhibitory receptor clustering and function, via the recruitment of gephyrin [37], we speculate that CDKL5 plays a direct role in the stabilisation of the key components of the inhibitory synapse by the above interaction. Future studies will be performed to address how CDKL5 loss affects the gephyrin-CB complex and which GABAAR subunits are affected.

#### *3.2. CDKL5 Deficiency Impairs the Functional GABAergic Synapse*

The postsynaptic inhibitory defects were accompanied by a reduction in the frequency of mIPSCs in *Cdkl5*-KO neurons. Various mechanisms can underlie the reduced frequency,

such as reduced number of synapses at the pre- and postsynaptic levels, a reduced neurotransmitter release and a reduced number of presynaptic vesicles. Even if we cannot rule out a direct role for CDKL5 at the presynaptic level, we find it relevant to consider that a similar result was observed when CDKL5 expression was acutely silenced through the transfection of a CDKL5-silencing construct also expressing GFP. Our experimental settings allowed the recording of the mIPSCs of CDKL5-silenced neurons that were innervated by non-silenced cells. Therefore, the altered inhibitory neurotransmission could be ascribed to a direct role for CDKL5 at the postsynaptic site, given that the presynaptic compartment is normal. The major phenotype of CDKL5 deficiency in both *Cdkl5*-KO and silenced neurons was a remarkable decrease in mIPSC frequency, hence reflecting a strong reduction in the number of functional GABAergic synapses. We speculate that the loss of CDKL5 at the postsynaptic site influences GABAergic innervation in *Cdkl5*-KO cultures similar to what has been reported for γ2-containing GABAAR clusters in cortical neurons silenced for GODZ, which is implicated in trafficking and postsynaptic accumulation of γ<sup>2</sup> subunit-containing GABAARs [38]. In support of our hypothesis, we observed a significantly reduced number of both bassoon- and VGAT-positive puncta in the primary cultures of *Cdkl5*-KO neurons, which indicates a presynaptic defect.

#### *3.3. PME Treatment Ameliorates CDKL5-Related Defects*

Various recent data have shown that CDKL5 is involved in regulating MT dynamics [13,39–41]. Interestingly, the possibility of targeting MT interacting proteins, the function of which is impaired in the absence of CDKL5, seems to represent an interesting disease modifying therapeutic strategy for CDD [14,15,42].

Here we show that treatment in vitro with the neuroactive synthetic steroid PME restores CDKL5-dependent GABAAR defects both molecularly and functionally. Indeed, the decreased frequency of mIPSCs in primary *Cdkl5*-KO cultures was normalised, together with the surface exposure of synaptic GABAARs and the number of gephyrin, bassoon and VGAT puncta. Intriguingly, we also found that the number of γ<sup>2</sup> subunit-containing GABAAR clusters was restored in the hippocampal slices of *Cdkl5*-KO mice treated with subcutaneous injections of PME (10 mg/kg) for seven consecutive days starting from PND60. Of relevance, this treatment schedule restored hippocampal-dependent learning and memory defects in *Cdkl5*-KO mice [15].

At present, we can only speculate about the precise mechanism through which PME can promote synaptic GABAAR accumulation. Our previous data showed that PME activates the +TIP CLIP170 by inducing its open conformation, thus promoting MT dynamics [15]. CLIP170 was found to be involved in the efficient loading of dynein-bound cargoes for retrograde transport in axons [43]. Dynein is central for minus-end directed transport of various cargoes and, by activating CLIP170, PME may therefore promote dynein-dependent transport. Indeed, gephyrin interacts directly with dynein and transports glycine receptor-containing vesicles in dendrites [28]. Moreover, forward trafficking of γ2-containing GABAARs is mediated by GABARAP-containing complexes, which interact with dynein [44]. Future studies will allow us to reveal the effect of PME on dyneindependent transport.

In conclusion, we have demonstrated for the first time that CDKL5 plays a direct role in the expression of functional GABAARs at synaptic sites, in part through its interaction with the cytoplasmic CB-gephyrin complex. Importantly, treatment with the synthetic neuroactive steroid PME can bypass the need for CDKL5 and restore GABAAR expression and GABAAR functioning. Currently, there are no approved therapies for CDD and any pharmacological strategies that reduce the frequency, duration or severity of seizures may positively impact the quality of life for CDD patients. PME might represent an important breakthrough in the CDD field, as the restoration of GABAAR expression might be beneficial also for the cognitive defects and autistic-like features present in CDD patients.

#### **4. Materials and Methods**

#### *4.1. Mice*

Protocols and use of animals were approved by the Animal Ethics Committee of the University of Insubria and in accordance with the guidelines released by the Italian Ministry of Health (D.L. 2014/26) and the European Community directives regulating animal research (2010/63/EU). Adult mice were euthanised by cervical dislocation, while neonates were sacrificed by exposure to CO<sup>2</sup> followed by decapitation.

Male *Cdkl5*-KO mice [4] on the genetic background CD1 were used. Littermate controls were used for all experiments. The day of birth was designated as postnatal day (PND) zero. After weaning, three to five animals belonging to the same litter were housed in activity enriched cages on a 12 h light/dark cycle in a temperature-controlled environment with food and water provided *ad libitum* and checked daily for general health conditions. Genotyping was performed through PCR on genomic DNA from tail biopsies using the Rapid Extract PCR Kit (PCRBIO, [15]).

#### *4.2. Plasmids*

pEGFP-GPHN, pRK5 Myc-CB2-SH3<sup>+</sup> , and pRK5 Myc-CB2-∆SH3 were kindly provided by Dr. Theophilos Papadopoulos's laboratory (Max-Planck Institute for Brain Research, Göttingen) and were generated as described previously [45]. pFlag-CDKL5 encoding the 107 kDa splice variant was generated as described elsewhere [46].

#### *4.3. Cell Cultures and Transfections*

African green monkey kidney cells (COS7) and human embryonic kidney 293T (HEK293T) cells were maintained in DMEM (Euroclone) supplemented with 10% fetal bovine serum (Euroclone), 2 mM L-glutamine (Euroclone) and penicillin/streptomycin (100 units/mL and 100 µg/mL respectively, Euroclone) at 37 ◦C with 5% CO2. Cells were transfected with LipofectamineTM 3000 (Life Technologies Incorporated).

Primary hippocampal cultures were prepared from embryonic day 17 (E17) mouse embryos considering the day of the vaginal plug as E0, as described previously [47], and plated on poly-L-lysine (Sigma-Aldrich, Sant Louis, MO 63103, USA) coated plates.

CDKL5 expression was abruptly silenced in primary hippocampal neurons at DIV11 through the transfection of a shCDKL5 targeting the sequence GCAGAGTCGGCACAGC-TAT, using as negative control a shRNA against LacZ (shLacZ). Plasmid transfection was performed with LipofectamineTM 2000 transfection reagent (Life Technologies Incorporated).

#### *4.4. Pharmacological Treatment*

For treatment in vitro, 0.3 or 1 µM of PME or 0.1% DMSO (vehicle) was added to the primary cultures at DIV11, and neurons were harvested after 72 h. For treatment in vivo, mice received daily subcutaneous injections from PND60 to PND66 of 10 mg/kg of PME or sesame oil (vehicle) between 9:00–11:00 as described in Barbiero et al. [15]. For neuroanatomical experiments, mice were sacrificed 24 h after ended treatment and processed as described below. PME was suspended in sesame oil and freshly prepared each day.

#### *4.5. Antibodies*

The following primary antibodies were used in immunofluorescence and western blotting experiments: mouse anti-bassoon (Santa Cruz, sc-58509), rabbit anti-CDKL5 (Sigma, HPA002847), mouse anti-CDKL5 (Santa Cruz, sc-376314), rabbit anti-collybistin (SYSY, 261003), rabbit anti-GABAAR γ<sup>2</sup> (SYSY, 224003), goat anti-GABAAR γ<sup>2</sup> (Invitrogen, PA5- 19299), rabbit anti-GAPDH (Sigma, G9545), mouse anti-gephyrin (Santa Cruz, sc-25311), anti-GFP (chicken, Molecular Probes, A10262), chicken anti-MAP2 (SYSY, 188006), mouse anti-c-myc (clone 9E10), rabbit anti-VGAT (SYSY, 131003). Secondary Alexa Fluor antirabbit, -mouse and -chicken were purchased from Abcam or Invitrogen. HRP-conjugated

goat anti-mouse, goat anti-rabbit and donkey anti-goat secondary antibodies for western blottings were purchased from Jackson Immunoresearch.

#### *4.6. Immunofluorescence*

COS7 cells and hippocampal neurons: neurons were grown on poly-L-lysine (1 mg/mL; Sigma-Aldrich) coated coverslips (300 cells/mm<sup>2</sup> ) until DIV14. After fixation in 4% formaldehyde (PierceTM) with 4% sucrose (Sigma-Aldrich, Sant Louis, MO 63103, USA), cells were blocked in 1X PBS, 5% goat serum (Euroclone) and 0.2% Triton X-100. Surface exposed γ<sup>2</sup> subunit-containing GABAAR were immunostained under nonpermeabilising conditions with blocking in 1X PBS, 5% goat serum. Incubation with the primary antibody was performed overnight at 4 ◦C and with the secondary antibody for 1 h at room temperature. Slides were mounted with ProLong Gold antifade reagent (Life Technologies).

To quantify gephyrin, bassoon and VGAT puncta as well as surface expressed GABAAR γ<sup>2</sup> along MAP2-positive dendrites, images were captured with a 60X objective coupled to an Olympus BX51 fluorescence microscope equipped with Retiga R1 (QImaging) CCD camera. The number of gephyrin, bassoon and VGAT puncta, along with the fluorescence intensity of GABAAR γ<sup>2</sup> staining, were measured along 30 µm long segments (proximal part of secondary branches) using the software Fiji ImageJ (function: analyse particles or measure). Primary antibodies were used: bassoon, 1:50; CDKL5, 1:50; GABAAR γ2, 1:150; gephyrin, 1:50; GFP, 1:100; MAP2 chicken, 1:500; MAP2 rabbit and mouse, 1:1000; Myc, 1:200; VGAT, 1:1000.

Hippocampal slices: 24 h after treatment, mice were decapitated (upon dislocation) and brain hemispheres were rapidly excised and frozen in liquid nitrogen. Cryosections (30 µm) were cut and mounted onto coated slides (SuperFrost® Plus, Thermo Scientific, 38116 Braunschweig, Germany) and stored at −80 ◦C. The samples were fixed in 2% paraformaldehyde (4 ◦C) for 90 s, rinsed thrice with 1X PBS and blocked for 1 h in blocking solution (0.05% goat serum, 3% Triton X-100 in 1X PBS). Upon incubation with anti-GABAAR γ<sup>2</sup> antibody (1:1000) in a humid chamber overnight at 4 ◦C, the slices were incubated with the secondary antibody (Alexa Fluor goat anti-rabbit 488 nm) in blocking solution for 1 h at room temperature, rinsed thrice with 1X PBS and mounted with ProLong Gold antifade reagent (Life Technologies). As negative control, a sample was incubated with only the secondary antibody.

Images from the molecular layer of the dentate gyrus were acquired with a LEICA TCS SL confocal microscope (LEITZ; Leica Mycrosystems, Wetzlar, Germany) with objective 63X (NA 1.32; zoom factor = 8) and the pinhole set at 1 Airy unit. Four slices per animal were analysed, and the number of GABAAR γ<sup>2</sup> clusters was quantified using the software Fiji ImageJ (plugin: analyse particles). Optimised threshold values and size filters were applied for all the images to identify GABAAR γ<sup>2</sup> clusters. The number of GABAAR γ<sup>2</sup> puncta was calculated in four identical sections for each slice (to have a mean of four separate zones of the dentate gyrus per single slice) and expressed per µm<sup>2</sup> .

#### *4.7. Western Blotting and Immunoprecipitation*

Primary hippocampal neurons were lysed in 3X Laemmli buffer, and samples were separated by 10% SDS-PAGE, transferred to nitrocellulose membranes and blocked in 5% non-fat milk in TBS-T (20 mM of Tris-HCl pH 7.4, 150 mM of NaCl, 0.2% Tween-20). Blots were incubated with primary antibodies overnight at 4 ◦C, washed in TBS-T and incubated with appropriate secondary antibodies for 1 h at room temperature. Blots were developed with protein detection system-ECL (Genespin) coupled to G:BOX Chemi Imaging System (Syngene). Densitometric expression analyses were performed using ImageJ software.

CDKL5 was immunoprecipitated from 400 µg of HEK293T cells or 1 mg of a mouse brain extract (PND20-30) lysed in lysis buffer [mM: 50 Tris-HCl pH 7.4, 150 NaCl, 1 EDTA, 1 EGTA, 1% Triton X-100, 1X protease inhibitor cocktail (PIC, Sigma-Aldrich, Sant Louis, MO 63103, USA) and 1X PhosSTOP (Roche)] and incubated overnight at 4 ◦C with 1 µg of

anti-CDKL5 or unrelated IgGs as control. The immunocomplexes were precipitated with protein-G agarose (Life Technologies), washed several times with lysis buffer and analysed by SDS-PAGE and western blotting.

#### *4.8. Biotinylation Assays*

Biotinylation assays were performed according to previously described protocols [48,49] with slight modifications. Primary hippocampal neurons were plated in 6-well plates coated with 0.5 mg/mL poly-L-lysine (Sigma-Aldrich, 400,000 neurons/well) and used at DIV14. Neurons were washed twice with HBSS/Ca2+/Mg2+, followed by incubation with 0.5 mg/mL of Sulfo-NHS-SS-Biotin (Cyanagen). Quenching was performed using HBSS/Ca2+/Mg2+ supplemented with glycine (50 mM) and BSA (0.5%), after which cells were lysed with standard radio-immunoprecipitation assay (RIPA) buffer (mM: 50 Tris-HCl pH 7.4, 150 NaCl, 1 EDTA, 2 EGTA, 1% NP40, 0.1% SDS, 0.5% SDC, 1X PhosSTOP, 1X PIC). After correction for protein content using a BCA protein assay kit (PierceTM), biotinylated proteins were purified on StreptAvidin UltraLink Resin (PierceTM) and resolved by SDS-PAGE and western blotting. Surface expression was evaluated as the ratio between the biotinylated fraction (surface) and the total cell lysate normalised with GAPDH.

#### *4.9. In Vitro Electrophysiological Recordings*

GABA-mediated inhibitory postsynaptic currents in miniature (mIPSCs) were recorded using the patch-clamp technique in the whole cell voltage-clamp configuration in the presence of 1 µM tetrodotoxin (TTX, Tocris) to block the generation of action potentials. Recordings were obtained from primary hippocampal neurons at DIV14 (plated on poly-Llysine coated coverslips at the density of 234 cells/mm<sup>2</sup> ) using the Axopatch 200B amplifier and the pClamp-10 software (Axon Instruments). Recordings were performed in Krebs'- Ringer's-HEPES (KRH) external solution (mM: 125 NaCl, 5 KCl, 1.2 MgSO4, 1.2 KH2PO4, 2 CaCl2, 6 glucose, 25 HEPES-NaOH pH 7.4). Recording pipettes were pulled from glass capillary (World Precision Instrument) using a two-stage puller (Narishige) and had tip resistances of 3–5 MΩ when filled with the intracellular solution (mM: 130 Cs-gluconate, 8 CsCl, 2 NaCl, 10 HEPES, 4 EGTA, 4 MgATP, 0.3 GTP pH 7.3). Voltage-clamp recordings were performed at holding potentials of +10 mV. The recorded traces were analysed using Clapfit-pClamp 10 software, after choosing an appropriate threshold. In particular, events that exceeded at least twice the standard deviation of the baseline noise were considered as mIPSCs and included in our analyses.

#### *4.10. Statistical Analyses*

All experiments and analyses were performed knowing the respective genotypes of the animals. Data were analysed with Prism software (GraphPad), and all values were expressed as the mean ± SEM. Data that were identified as significant outliers by the software were removed from the datasets. The significance of the western blotting and immunofluorescence results was evaluated using Unpaired Student's *t*-test, One-way ANOVA followed by Tukey's multiple comparisons test and Two-way ANOVA followed by Tukey's multiple comparisons test. The significance of in vitro electrophysiological studies was evaluated by Mann Whitney U test, Kruskal-Wallis test followed by Dunn's multiple comparisons test, One-way ANOVA followed by Tukey's multiple comparisons test and Unpaired Student's *t*-test. Probability values of *p* < 0.05 were considered as statistically significant.

**Author Contributions:** Conceptualization, R.D.R., M.T. and C.K.-N.; methodology, R.D.R., F.A. and C.K.-N.; investigation, R.D.R., S.V., C.C., I.B., C.P., M.T. and S.R.; resources, M.B.; formal analysis, R.D.R., C.P., M.T., F.A. and I.B.; writing—original draft preparation, R.D.R.; writing—review and editing, C.K.-N.; supervision, C.K.-N.; project administration, R.D.R. and C.K.-N.; funding acquisition, C.K.-N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Fondazione Telethon [GGP20024], AIRETT onlus and the Italian parents' association Albero di Greta (to CKN).

**Institutional Review Board Statement:** The animal study protocol was approved by the Ethics Committee of the University of Insubria and by the Italian Council on Animal Care, the Italian Government decree No. 28/2019.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We thank Theophilos Papadopoulos (Max-Planck Institute for Brain Research, Göttingen) for the CB and gephyrin constructs.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

## *Communication* **Pattern of Mitochondrial Respiration in Peripheral Blood Cells of Patients with Parkinson's Disease**

**Tommaso Schirinzi 1,\* ,†, Illari Salvatori 2,†, Henri Zenuni <sup>1</sup> , Piergiorgio Grillo <sup>1</sup> , Cristiana Valle 2,3 , Giuseppina Martella 2,\* , Nicola Biagio Mercuri 1,2,‡ and Alberto Ferri 2,3,‡**


**Abstract:** Mitochondria are central in the pathogenesis of Parkinson's disease (PD), as they are involved in oxidative stress, synaptopathy, and other immunometabolic pathways. Accordingly, they are emerging as a potential neuroprotection target, although further human-based evidence is needed for therapeutic advancements. This study aims to shape the pattern of mitochondrial respiration in the blood leukocytes of PD patients in relation to both clinical features and the profile of cerebrospinal fluid (CSF) biomarkers of neurodegeneration. Mitochondrial respirometry on the peripheral blood mononucleate cells (PBMCs) of 16 PD patients and 14 controls was conducted using Seahorse Bioscience technology. Bioenergetic parameters were correlated either with standard clinical scores for motor and non-motor disturbances or with CSF levels of α-synuclein, amyloid-β peptides, and tau proteins. In PD, PBMC mitochondrial basal respiration was normal; maximal and spare respiratory capacities were both increased; and ATP production was higher, although not significantly. Maximal and spare respiratory capacity was directly correlated with disease duration, MDS-UPDRS part III and Hoehn and Yahr motor scores; spare respiratory capacity was correlated with the CSF amyloid-β-42 to amyloid-β-42/40 ratio. We provided preliminary evidence showing that mitochondrial respiratory activity increases in the PBMCs of PD patients, probably following the compensatory adaptations to disease progression, in contrast to the bases of the neuropathological substrate.

**Keywords:** Parkinson's disease; PBMCs; mitochondria; Seahorse; immunometabolic pathway; neuroinflammation; biomarkers; synaptopathy

#### **1. Introduction**

Parkinson's disease (PD) is a common neurodegenerative disorder, responsible for both motor and non-motor disturbances, whose neuropathological hallmarks are the loss of dopaminergic nigral cells and the brain accumulation of α-synuclein (α-syn)-positive Lewy bodies [1,2].

Mitochondria are highly plastic and dynamic organelles, and are critical for cellular bioenergetics and neuronal homeostasis [3]. Mitochondrial dysfunction and PD pathogenesis have been historically connected. Solid evidence from genetics and experimental models [4,5] shows how dysfunctional mitochondria lead to oxidative stress and synaptopathy, namely the earliest neurodegeneration steps in PD [6,7]. Mitochondrial impairment also precipitates lysosomes activity, protein turn-over, and α-syn metabolism, triggering or fostering neurodegeneration at every stage of PD [8]. Moreover, mitochondria are central in neuroinflammation and immune response [9], other critical determinants of PD pathology [10,11]. Accordingly, mitochondria-related immunometabolic pathways could emerge as alternative targets for disease-modifying treatments in PD [12], useful for the

**Citation:** Schirinzi, T.; Salvatori, I.; Zenuni, H.; Grillo, P.; Valle, C.; Martella, G.; Mercuri, N.B.; Ferri, A. Pattern of Mitochondrial Respiration in Peripheral Blood Cells of Patients with Parkinson's Disease. *Int. J. Mol. Sci.* **2022**, *23*, 10863. https://doi.org/ 10.3390/ijms231810863

Academic Editor: Hari Shanker Sharma

Received: 18 August 2022 Accepted: 13 September 2022 Published: 17 September 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

substitution or integration of anti-α-syn drugs, which failed to halt disease progression alone [8,13,14].

From this perspective, the role of mitochondria in PD must be further elucidated. In particular, we have to overtake current knowledge, mostly relying on preclinical studies [15], and acquire evidence directly from patients.

Peripheral blood mononucleate cells (PBMCs) exhibit typical PD-neuropathology signatures [1,16]. In addition, immune cells directly participate in the pathogenic cascade of PD, either at the central or peripheral level [11], thus representing the circulating leukocytes as an ideal tissue to analyze in vivo molecular events underlying PD.

In this study, we shaped the pattern of mitochondrial bioenergetics in vivo in PD. We thus assessed mitochondrial respiration in the PBMCs of PD patients and examined the respective correlations with clinical features and levels of cerebrospinal fluid (CSF) neurodegeneration biomarkers (α-synuclein, amyloid-β peptides, tau proteins and lactate), which are considered as an indication of pathological changes in the brain [17].

Indeed, mitochondrial respiration is a key function in cellular homeostasis, an epicenter of many metabolic pathways that are crucial for the clinical–pathological progression of PD [18], whose deeper knowledge is fundamental to design novel therapeutic intervention strategies in PD.

#### **2. Results**

#### *2.1. Study Population*

Table 1 summarizes the clinical and demographic data of the study population. Groups were homogeneous regarding age and sex distribution.


**Table 1. Demographics, clinical parameters and CSF biomarkers of the study population.** Age and disease duration are expressed in years; biomarkers in pg/mL. F = female; M = male; ns = not significant; other abbreviations are spelled out in the text.

#### *2.2. Bioenergetic Parameters*

PBMCs obtained from PD patients and controls had a similar baseline oxygen consumption rate (OCR) (mean ± st.dev. in pmol/min) (PD = 22.4 ± 10.6; controls = 20.8 ± 8.8). ATP-linked respiration was higher in PD (40.2 ± 23.4) than in the controls (26.3 ± 9.4),

although it was not statistically significant (*p* = 0.07). Maximal respiration was significantly higher in PD (155.6 ± 115.0) than in the controls (79.3 ± 29.9, *p* = 0.038). The spare (or reserve) respiratory capacity was significantly higher in PD (134.8 ± 108.2) than in the controls (58.5 ± 28.04, *p* = 0.02) (Figure 1a–e). There were no significant gender differences in either of the groups. Sample normalization was optimized through Western Blot analysis (Supplementary Figure S1). reserve) respiratory capacity was significantly higher in PD (134.8 ± 108.2) than in the controls (58.5 ± 28.04, *p* = 0.02) (Figure 1a–e). There were no significant gender differences in either of the groups. Sample normalization was optimized through Western Blot analysis (Supplementary Figure S1).

PBMCs obtained from PD patients and controls had a similar baseline oxygen consumption rate (OCR) (mean ± st.dev. in pmol/min) (PD= 22.4 ± 10.6; controls = 20.8 ± 8.8).

although it was not statistically significant (*p* = 0.07). Maximal respiration was significantly higher in PD (155.6 ± 115.0) than in the controls (79.3 ± 29.9, *p* = 0.038). The spare (or

*Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 3 of 8

*2.2. Bioenergetic Parameters* 

**Figure 1. PBMC respirometry by Seahorse Bioscience.** (**a**–**d**) Bar graphs showing differences in bioenergetic parameters between PD and controls (values are expressed as means ± S.E.M; \* = *p* < 0.05). (**e**) Representative time course of OCR during an experimental session (one patient and one **Figure 1. PBMC respirometry by Seahorse Bioscience.** (**a**–**d**) Bar graphs showing differences in bioenergetic parameters between PD and controls (values are expressed as means ± S.E.M; \* = *p* < 0.05). (**e**) Representative time course of OCR during an experimental session (one patient and one control).

#### *2.3. Correlation Analysis*

control).

*2.3. Correlation Analysis*  Maximal respiration was directly correlated with disease duration (R = 0.5, *p* = 0.04), Maximal respiration was directly correlated with disease duration (R = 0.5, *p* = 0.04), Hoehn and Yahr scale (H & Y) stage (R = 0.7, *p* = 0.003), and MDS-UPDRS part III (R = 0.62,

Hoehn and Yahr scale (H & Y) stage (R = 0.7, *p* = 0.003), and MDS-UPDRS part III (R =

*p* = 0.01). Spare respiratory capacity was directly correlated with disease duration (R = 0.5, *p* = 0.04), H & Y stage (R = 0.7, *p* = 0.004), and MDS-UPDRS part III (R = 0.63, *p* = 0.009). No further correlations emerged between bioenergetic parameters and clinical features (age, Mini-Mental State Examination (MMSE), Montreal Cognitive Assessment (MOCA) and Non-Motor Symptoms Scale (NMSS) scores, and levodopa equivalent daily dose (LEDD)).

Spare respiratory capacity was directly correlated with both amyloid-β-42 (Aβ42) and amyloid-β-42/amyloid-β-40 (Aβ42/Aβ40) ratio (R = 0.68, *p* = 0.02 and R = 0.66, *p* = 0.04, respectively). No further correlations between bioenergetic parameters and other CSF biomarkers (α-syn, total-tau (t-tau) and phosphorylated-tau (p-tau), lactate, and the CSF/blood albumin ratio) were found.

#### **3. Discussion**

PBMCs offer the opportunity to track the in vivo molecular events underlying PD, serving as a reliable model for central neuropathology [1,16]. Here, we shaped the pattern of mitochondrial respiration in PD by applying Seahorse Bioscience technology to the PBMCs of both PD patients and healthy controls, and examining the correlations of bioenergetic parameters with clinical features and levels of CSF biomarkers of neurodegeneration.

We found that, in PD, PBMCs had a peculiar pattern of mitochondrial respiration, with normal basal respiration, a significant increase in both maximal respiratory capacity and spare respiratory capacity, and a higher (but not statistically significant) ATP production. The increase in maximal respiratory capacity and spare respiratory capacity was directly associated with disease duration and severity of motor impairment, suggesting that mitochondrial respiration capability could rise in parallel with the clinical–pathological progression of PD. In addition, the spare respiratory capacity was greater in patients with higher CSF levels of Aβ42 and Aβ42/Aβ40 ratio, which identifies individuals with a lower burden of cerebral amyloidopathy [17].

In contrast with theoretical expectances, which would have predicted a reduction in bioenergetic activity in the mitochondria of PD patients [8,15,19], we discovered an increase in respiratory capacity. Albeit surprising, these data are consistent with previous findings from other experimental settings. In fact, lymphoblasts of patients with PD, namely immortalized cells derived from peripheral blood lymphocytes, also showed a dramatic increase in mitochondrial respiration, ATP synthesis and maximum OCR. The greater energy production found by Annesley et al. was assumed to satisfy an increased requirement due to the higher mitochondrial turn-over induced by pathological α-syn accumulation, or, in general, to an abnormal metabolic state with high energy consumption [20]. However, the changes in respiratory activity could follow modifications in the lipid composition of different cell compartments or in lipid trafficking overall, which markedly affect the mitochondrial architecture, the transport of proteins into mitochondria, and the functioning of respiratory proteins [21]. Indeed, mitochondrial membrane potential could be altered in the PBMCs of PD patients, suggesting a certain degree of stress in those organelles [22]. Alternatively, the growing neuroinflammatory state or the greater immune activation against progressive neurodegeneration might account for the increased bioenergetics of leukocytes in PD [20]. In addition to PBMCs, fibroblasts from PD patients also displayed higher mitochondrial respiratory rates compared to healthy controls. This has been observed either in Parkin mutant carriers [23] or in subjects with the idiopathic form [24], and interpreted as a compensatory bioenergetics activation or as a consequence of abnormal mitochondrial functioning.

In line with the hypothesis that the implementation of PBMC bioenergetic activity could help cellular compensations to the progression of neuropathology and neuroinflammation, we noticed that maximal and spare respiratory capacities tended to increase with disease evolution, either in terms of disease duration or clinical impairment.

Nuclear factor erythroid 2-related factor 2 (Nrf2) is a transcription factor regulating the cellular defense against oxidative stress, inflammation and neurodegeneration, by promoting overall mitochondrial respiration and metabolism [25]. We recently demonstrated that Nrf2 was highly expressed and its pathway activated in the PBMCs of PD patients, especially in those with a longer disease duration [1]. It is thus reasonable that the progressive increase in respiratory capacity over the clinical–pathological course of PD may follow an Nrf2-mediated mitochondrial activation aimed at sustaining a systemic defensive response. Of interest, similar findings were observed in the lymphoblasts of patients with Amyotrophic Lateral Sclerosis, where the respiratory activity was increased together with the upregulation of the Nrf2 pathway, suggesting this axis as a common compensation mechanism among different sporadic neurodegenerative diseases [26].

Then, we found that, in PD patients, the respiratory capacity was directly correlated with CSF levels of the Aβ42 to Aβ42/Aβ40 ratio. Experimental models of Alzheimer's disease showed that neurons exhibit mitochondrial dysfunction and respiration impairment in the presence of Aβ42 peptides [27]. A reduction in CSF Aβ42 corresponds to the greater brain accumulation of amyloid-β plaques even in PD [17]. Accordingly, we could hypothesize that the poorer respiratory activity in patients with lower CSF Aβ42 (and worse amyloidopathy in the CNS) reflects a detrimental association between mitochondrial performances and pathological amyloid peptides, as well as at a peripheral level. In fact, PD patients with lower CSF Aβ42 present a more malignant and frailer phenotype [28], which is consistent with the weakening of defensive or compensatory mechanisms.

In contrast, other CSF biomarkers were not correlated with bioenergetic parameters, preventing further suppositions on the interactions between central neuropathology and systemic reactions.

Although limited by the sample size, this study demonstrated that PBMC mitochondria in PD patients had a peculiar pattern of respiration, with increased maximal and spare respiratory capacities. Respiratory changes probably reflect the increased energetic requirement due to the clinical–pathological progression of the disease and the subsequent compensatory adaptations. Such changes vary depending on the disease stage and the neuropathological substrate; they are more contained in patients with biochemical signatures of frailty.

We could interpret our results in two ways. Accepting PBMCs as a model of central neuropathology, they allow for a better understanding of the role of mitochondria in PD, profiling those dynamics that may underlie PD pathology throughout its course. On the other hand, they provide additional evidence on the presence of metabolic abnormalities in systemic immune cells, which could be useful in developing novel therapeutic strategies. Indeed, mitochondria-related metabolic pathways could be specifically targeted in immune cells to modulate their activity either at the central or systemic level, counteracting neuroinflammation and neurodegeneration. Further investigations on larger replication cohorts are now necessary to confirm and extend our findings.

#### **4. Methods**

#### *4.1. Study Population and Biosampling*

This study involved sixteen PD patients and fourteen sex/age-matched healthy controls enrolled at Tor Vergata University Hospital (Rome, Italy) in 2021–2022. PD was diagnosed according to MDS 2015 Postuma's criteria. Controls were healthy volunteers without a history or clinical signs of neurological diseases. Subjects with main acute/chronic infectious/inflammatory/internal diseases and/or abnormal blood cell count were excluded [1].

For each subject, demographics, anthropometrics, and medical history were collected. PD patients were assessed using the Hoehn and Yahr scale (H & Y), MDS-UPDRS part III, Non-Motor Symptoms Scale (NMSS), Mini-Mental State Examination (MMSE) adjusted for age and educational level, Montreal Cognitive Assessment (MOCA), and the levodopa equivalent daily dose (LEDD) calculation.

All participants underwent venous blood sampling (20 mL), in the morning, after overnight fasting (morning drugs allowed). Blood was immediately processed to separate PBMCs through density gradient centrifugation with Ficoll-Hypaque (GE Healthcare Life Sciences), according to standard procedures. PBMCs were carefully frozen in cryoSFM medium, stored in liquid nitrogen and subsequently thawed for bioenergetics analyses [29].

Ten PD patients also underwent cerebrospinal fluid (CSF) analysis for neurodegenerationrelated biomarker measurement. CSF was obtained through lumbar puncture, which was performed following standard procedures. The CSF levels of total α-synuclein (α-syn), amyloid-β-42 (Aβ42) and amyloid-β-40 (Aβ40), total-tau (t-tau) and phosphorylatedtau (p-tau), lactate, and the CSF/blood albumin ratio were quantified as previously described [2,30]. The Aβ42/Aβ40 and the Aβ42/p-tau ratios were then calculated.

The study was approved by the local EC (protocol n◦ 16.21), following the principles of the declaration of Helsinki. All participants signed informed consent.

#### *4.2. Bioenergetics Analysis*

Mitochondrial function was determined using a Seahorse XF96e Analyzer (Seahorse Bioscience—Agilent, Santa Clara, CA, USA) [31]. The PBMCs were plated at the density of 15 <sup>×</sup> <sup>10</sup><sup>4</sup> cells/well. An equal number of cells, from each sample, was processed for a quantitative Western Blot assay (as described below). A mitochondrial stress test was performed according to Agilent's recommendations. In brief, growth medium was replaced with XF test medium (Eagle's modified Dulbecco's medium, 0 mM glucose, pH = 7.4; Agilent Seahorse) supplemented with 1 mM pyruvate, 10 mM glucose and 2 mM L-glutamine. Before the assay, the PBMCs were incubated in a 37 ◦C incubator without CO<sup>2</sup> for 45 min to allow the pre-equilibration of the assay medium. The test was performed by first measuring the baseline oxygen consumption rate (OCR), followed by sequential OCR measurements after the injection of oligomycin (1.5 µM), carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (1 µM) (FCCP) and Rotenone (0.5 µM) + Antimycin A (0.5 µM). This allowed the measurement of the key parameters of the mitochondrial function, including the basal respiration, the ATP-linked respiration (obtained through the oligomycin-induced inhibition of ATP synthase with a subsequent decrease in electron flow through the electron transport chain, ETC), the maximal respiration (obtained through the uncoupling effect of FCCP, which induces ETC to operate at maximal capacity), the spare respiratory capacity (the difference between maximal respiration and basal respiration), and the non-mitochondrial ATP production (obtained through the rotenone/antimycin A-induced inhibition of both complex I and III).

#### *4.3. Western Blot Analysis*

Western Blot analysis was performed on protein extracts according to Scaricamazza et al. 2022 [32]. Specifically, 15 <sup>×</sup> <sup>10</sup><sup>4</sup> PBMCs from each sample were lysed in 100 <sup>µ</sup>L RIPA buffer (50 mM Tris–HCl pH 7.4, 0.5% Triton X-100, 0.25% Na-deoxycholate, 0.1% SDS, 250 mM NaCl, 1 mM EDTA and 5 mM MgCl2), 25 µL of which was loaded on SDS– polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes (Perkin Elmer, Cat# NBA085B). Membranes were probed using β-Actin antibody (Santa Cruz Biotechnology Inc., Cat# sc-47778, WB 1:5000).

#### *4.4. Statistical Analysis*

The distribution of variables was preliminarily examined using the Shapiro–Wilk test, and the non-normally distributed variables were Log10+1 transformed for analysis when necessary. Categorical variables were compared using a chi-square test. Group analysis was conducted via parametric (one-way ANOVA or Student's *t*-test) or non-parametric tests, as appropriate; correlations were evaluated using Spearman's test. Statistical significance was set at *p* < 0.05. A blind analysis was run using IBM-SPSS-23. Data are available from the authors upon reasonable request.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ijms231810863/s1

**Author Contributions:** T.S., I.S. and A.F. conceived the study, analyzed the data and wrote the manuscript. H.Z., P.G. and C.V. collected the data and performed the experimental session. N.B.M. and G.M. provided supervision and edited the final manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of Fondazione PTV—Policlinico Tor Vergata (Rome, Italy) (protocol n◦ 16.21).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** Data are available from the authors upon reasonable request.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

