A single injection of pregabalin induces short- and long-term beneficial effects on fear memory and anxiety-like behavior in rats with experimental type-1 diabetes mellitus

Anxiety Disorders and Posttraumatic Stress Disorders (PTSD) associated with type-1 diabetes mellitus (T1DM) are increasingly common comorbidities and the treatment is quite challenging. In that sense, evidence indicates that the anticonvulsant pregabalin is highly effective in treating severe cases of anxiety, as well as PTSD and diabetic neuropathic pain which is also very prevalent in T1DM. Herein, the short- and long-term effects of a single injection of pregabalin on the acquisition of a fear extinction memory and parameters of anxiety in induced-T1DM animals were investigated. For that, we used the contextual fear conditioning (CFC) and elevated plus maze paradigms, respectively. A putative antioxidant activity was also evaluated. Our findings demonstrated that induced-T1DM animals presented greater expression of fear memory, difficulty in extinguishing this fear memory, associated with a more pronounced anxiety-like response. Pregabalin was able to induce a short and long-lasting effect by facilitating the acquisition of the fear extinction memory and inducing a later anxiolytic-like effect. Also, the increased lipid peroxidation levels in the hippocampus and prefrontal cortex of induced-T1DM rats were reduced after pregabalin injection, while the decreased levels of reduced glutathione were increased in the hippocampus. Despite the need for more studies to understand the mechanism of action of pregabalin under these conditions, our data demonstrate for the first time that a single injection of pregabalin in a specific time window was able to improve behavioral parameters in addition to inducing neuroprotective effect. Thus, pregabalin has potential worth exploring for the treatment of PTSD and/or Anxiety associated with T1DM.


Introduction
Accumulating evidence indicates that diabetic encephalopathy and psychopathologies share common neurobiological mechanisms (de Morais et al. 2014;Gupta et al. 2014;Prabhakar et al. 2015;da Silva Dias et al. 2016;Aswar et al. 2017;Chaves et al. 2020). In this sense, preclinical studies conducted in diabetic or non-diabetic animals show that increase in inflammation and oxidative stress relatedprocesses, along with a dysregulation in neurotransmission systems in brain areas related to the emotions, like the hippocampus (HIP) and prefrontal cortex (PFC), are present in animals with anxiety and/or depressive-like behaviors (MacQueen et al. 2005;de Morais et al. 2014;Muriach et al. 2014;Moulton et al. 2015;Bludau et al. 2016;Buchberger et al. 2016;da Silva Dias et al. 2016;Zanoveli et al. 2016;Pereira et al. 2018;Banagozar Mohammadi et al. 2019;Chaves et al. 2020).
Because these brain areas are highly involved in the processing of learning and memory, it is not surprising that clinical studies demonstrate cognitive dysfunctions in diabetic patients along with cortical and HIP atrophy and white matter lesions (Bremner et al. 1995). Here, it is important to highlight that diabetic patients are more likely than nondiabetic individuals to experience cognitive dysfunction along with anxiety (Jacobson et al. 2011;dos Santos et al. 2014;Rajput et al. 2016). Also, studies point out that posttraumatic stress disorder (PTSD), a disorder characterized by aberrant consolidated fear memories and/or failure in extinguishing fear memory, is a risk factor for the development of type-1 diabetes mellitus (T1DM) and type-2 diabetes mellitus (Weisberg et al. 2002;Zung et al. 2012;Vaccarino et al. 2014;Renna et al. 2016). Preclinical studies reinforce these clinical findings, by demonstrating that animals with experimental T1DM present learning and memory impairments (Matsunaga et al. 2016;Ahmed et al. 2019). This impairment is also extended when studies turn to fear memory processing (Ikeda et al. 2015;Gambeta et al. 2016;de Souza et al. 2019;Ribeiro et al. 2020). Thus, our group showed that induced-T1DM animals present an overconsolidation of fear memory coupled with a generalization of this fear response when they are exposed to a neutral (non-aversive) environment (Gambeta et al. 2016;de Souza et al. 2019). Furthermore, they also have difficulty in extinguishing a fear memory (Ribeiro et al. 2020).
The association of these diseases considerably impairs the quality of life of patients, in addition to facilitating the emergence of other comorbidities and causing a great economic impact on society. Also, treating PTSD associated with T1DM with first-line drugsantidepressant drugs-is a huge challenge (Lustman and Clouse 2005; for a review see Zanoveli et al. 2016). Thus, exposure psychotherapy has been applied to contribute to the development of a fear/trauma extinction memory and it has become a valuable and safer treatment. Bearing in mind that extinction memory means new learning capable of inhibiting the expression of fear memory, the search for drugs that facilitates extinction processes without impairing the glycemic control and/or body weight would be excellent candidates for the treatment of PTSD particularly in individuals with T1DM.
In this respect, it is known that the anticonvulsant pregabalin (PGB), a gabapentin derivative of gammaaminobutyric acid, has been used to treat severe forms of generalized anxiety disorder (GAD) and PTSD (Pande et al. 2003;Baniasadi et al. 2014;Greenblatt and Greenblatt 2018) and it has shown promising results as a potential therapeutic alternative. Considering particularly the diabetic condition, it is noteworthy that PGB presents a broad-spectrum efficacy in the treatment of diabetic neuropathic pain (Finnerup et al. 2015), a condition that affects more than 25% of diabetic patients (Tesfaye et al. 2013;Finnerup et al. 2015). Thus, from this perspective, treating PTSD associated with T1DM with PGB would be advantageous. The exact mechanisms of action of PGB are still not completely elucidated; however, studies in vitro demonstrate that the drug binds presynaptic to the protein subunit (α2-δ) of voltage-dependent calcium channels in the central nervous system (Rickels et al. 2005), thereby suppressing the release of excitatory neurotransmitters, including glutamate, substance P and noradrenaline (Field et al. 2001;Taylor et al. 2007;Patel and Dickenson 2016). It is important to note that PGB crosses the blood-brain barrier (BBB) and preferentially binds to the α2-δ1 auxiliary subunit, which is present in higher density amounts in brain regions involved in cortical processing, learning and memory, defensive behavior, neuroendocrine secretion, defensive behavior, general arousal (Takahashi et al. 2018;Cole et al. 2005). Additionally, studies have demonstrated that PGB induces neuroprotective effects by reducing oxidative stress and disrupting inflammatory processes in different pathological situations like sepsis, neuropathy, and retinopathy associated with diabetes (Sałat et al. 2016;Aslankoc et al. 2018;Cruz-Álvarez et al. 2018;Ali et al. 2019;Demir et al. 2021).
Hence, in the present study, we initially investigated whether non-diabetic and induced-T1DM animals exposed or not to a contextual fear conditioning session would perform distinctly, in a short-and long-term way, the behaviors related to the context-conditioned fear and anxiety. Next, we studied whether a single injection of PGB, before a fear memory extinction training, would be able to induce beneficial-short and long-term-actions on fear extinction memory and anxiety-like responses. In the last set of experiments, we investigated in vitro and ex vivo a likely antioxidant activity of PGB.

Animals
Based on previous studies from our laboratory De Morais et al. 2014;Pereira et al. 2018), we used at the beginning of the study young male Wistar rats weighing between 180-220 g (aged 8-10 weeks) from the Central Vivarium of the Federal University of Paraná were used. These animals were kept under controlled conditions, with temperature 22 ± 2 °C, 12:12 light/dark cycle, with water and food (Nuvilab CR1-Nuvital Nutrientes S/A) ad libitum. Four rats were kept per box (60 × 25x15 cm), with the shavings changed daily due to the polyuria induced by the diabetic condition. All experimental procedures were performed in accordance with the Brazilian Law for Animal Experimental Ethics and Care (11.794/8 October 2008) and was approved by the local Ethics Committee for the Use of Animals in the Biological Sciences Sector of the Federal University of Paraná (CEUA/BIO-UFPR, #1131). All efforts were made to minimize the number of rats and their suffering.

Drugs and Treatment protocol
The following drugs were used: Streptozotocin (STZ, Cayman Chemicals, USA), Pregabalin (PGB, Pfizer, Brazil). STZ (60 mg/kg) was freshly dissolved in sodium citrate buffer (10 mM; pH 4.5) and administered intraperitoneally (i.p.) to induce T1DM. Citrate buffer was administered as a control for normoglycemic (NGL) groups. PGB (0, 30, 100 and 300 mg/kg, i.p.) was suspended in sterile 0.9% saline and prepared immediately before administration, in a volume of 2 mL/kg of body weight. Dose of STZ was chosen based on previous studies from our group (de Morais et al. 2014;Gambeta et al. 2016;Pereira et al. 2018;de Souza et al. 2019;Chaves et al. 2020). Regarding PGB, route of administration and doses were taken from previous studies (Field et al. 2001;Zohar et al. 2008).The time of treatment (1 h before being exposed to the apparatus) was defined based on pharmacokinetic characteristics, i.e. PGB reaches its peak plasma concentration after 1 h of administration, bioavailability is approximately 90%, regardless of dose and frequency of administration, and half-life of the molecule is about 6 h (Buoli et al. 2017).

Experimental type-1 diabetes mellitus(T1DM) induction
T1DM was induced by a single intraperitoneal (i.p.) injection of STZ (60 mg/Kg) in overnight fasten rats (Chaves et al.2020). The diabetic condition was confirmed 72 h after the STZ injection using samples of about 5 µL of blood from the tail vein added to test strips impregnated with glucose oxidase (Accu-Check Active™, Roche). NGL animals received citrate buffer (10 mM, pH 4.5, equivalent volume). Only rats with blood glucose levels ≥ 250 mg/dL were considered diabetic and maintained in the study (about 80% of animals reached this parameter).

Behavioral tests
All behavioral tests were video recorded using camera Sony action cam 4 K for further analysis. The tests were carried out during the light phase (between 7:00 a.m. and 12:00 p.m.) in a sound-attenuated experimental room kept at 22 ± 2 °C. The rats were acclimatized to the experimental room for at least 30 min before each test. After each experimental session, the apparatus was cleaned with a 20% ethanol solution.

Contextual Fear Conditioning (CFC)Test
For the investigation of behaviors related to fear memory, the contextual fear conditioning (CFC) model was used (Ribeiro et al. 2020). The apparatus consisted of a rectangular chamber (26 × 31.5 × 21 cm; Insight, Ribeirão Preto, SP, Brazil). The bottom of the box consisted of small metal bars attached to an electrical stimulator which delivered foot shocks. The following steps were performed: -CFC session: This session was conducted by exposing the animal to the chamber (conditioned stimulus-CS) and after 30 s they received 3 electrical footshocks (US; 1 mA, lasting 3 s), with 30 s of intertrial interval. The animal remained in this chamber 30 s more before being returned to its home cage. -Extinction training session: 24 h after, the animals were placed in the same chamber (CS) used during the CFC session, but they did not receive any US presentation. For20 minutes, the freezing time of the animals was evaluated in seconds (divided into 4 blocks of 5 min each). -Extinction test 1: In the next day, the test was performed.
For that, animals remained in the context (CS) for 3 min without the presentation of US and the time (in seconds) that the animal spent in freezing was quantified. -Extinction test 2: The test was performed 7 days after the Extinction test 1, being this test exactly as the Extinction test 1.
The freezing behavior was used as an aversive conditioning index, and it was measured in seconds (s) and expressed as the percentage of total session time. Freezing behavior was considered when the animal presented a posture of complete immobility, except for breathing movements.

Open-field test (OFT)
The test was performed to evaluate the locomotor/exploratory activity according to Chaves et al. (2020). Thus, animals were placed in the center of open field apparatus that consisted of a rectangular wood arena (40 × 50 × 63 cm) with the floor divided into 9 rectangles situated in an isolated room and illuminated with an incandescent lamp (60 lx at the arena floor level). The number of crossed rectangles with the four paws were evaluated during 5 min.

Elevated plus-maze test (EPMT)
The animals were exposed to the EPMT immediately after the Extinction test 2 (see Sect. 2.4.1for details). The apparatus consisted of a wooden platform with 4 arms (2 open and 2 closed), raised 50 cm from the floor. There was a central area of 10 cm 2 making the intersection between all the arms. The test was initiated by placing a rat on the central platform of the maze, facing a closed arm and the session had a duration of 5 min.
The percentage of open arm entries (total amount of entries/ number of open-arm entries × 100) and the percent of time spent in the open arms served as the measure of anxiety. The frequency of entries in the closed arms was used as locomotor activity index (Cruz et al. 1994). Additionally, one ethological measure was added, the head dipping, which consist in the exploratory movement of head/shoulders over sides of the open arms and down towards the floor (Silva and Brandão 2000).

2-diphenyl-1-picrylhydrazyl (DPPH) assay
This assay was performed according to previous study (Pereira et al. 2018) to evaluate the potential antioxidant of PGB. It is used to evaluate the free radical scavenging activity of antioxidants. It is based on the principle that DPPH, upon accepting a hydrogen atom from a scavenger molecule (as an antioxidant), is reduced, and the purple color of the solution becomes yellow, concomitant with a decrease in absorbance (Mishra et al. 2012). The technique consisted of measuring the reactivity of PGB at different concentrations (1, 3, 10, 30, 100 μg/mL) mixed with DPPH methanolic solution (10 μg/mL). Ascorbic acid solution (50 μg/mL) was used as the positive control, and distilled water was used as the negative control. Absorbance was measured at 517 nm using a multi-modal microplate reader (BioTek Synergy HT, BioTek Instruments, Highland Park, VT, USA).

Evaluation of oxidative stress parameters
Brain samples The animals were euthanized by decapitation and the PFC and HIP were dissected. After dissection, the brain areas were frozen in liquid nitrogen, and stored at -80 °C until further analysis. The brain samples were homogenized in phosphate buffer pH 6.5 (1:10). The homogenate was used to determination of the reduced glutathione (GSH) levels. To quantify lipid peroxidation (LPO) rate, a part of homogenate was centrifuged at 9000 g in a micro-high-speed refrigerated centrifuge (VS-15000 CFNII, Vision Scientific, Daejeon, South Korea) for 20 min and the supernatant was used to measure the LPO.

Determination of the reduced glutathione (GSH) levels
The GSH levels were measured by the method described by Sedlak & Lindsay, (1968) and according to Pereira et al. (2018). Thus, 100 µL of the homogenate were mixed with trichloroacetic acid (80 µL of 12.5% purity) and centrifuged at 6000 rotations per minute (VS-15000 CFNII, Vision Scientific, Daejeon, South Korea) for 15 min at 4 °C. Then, 20 µL of the clear supernatant was mixed with 280 µL of Tris buffer (0.4 M, pH 8.9) and 5 µL of 5,5′-dithiobis-(2-nitrobenzoic acid) in methanol. The absorbance of the reaction solution was measured at 415 nm in a microplate reader (BioTek Synergy HT, BioTek Instruments, Highland Park, VT, USA). The individual values were interpolated in a standard curve of GSH to verify the linearity of the reaction. The results are expressed as μmol/g of tissue.
Determination of the lipid peroxidation (LPO)levels The LPO levels was measured by the FOX-2 method described by Jiang et al. (1991) and according to Pereira et al. (2018). For that, 100 μL of supernatant of PFC and HIP were homogenized in 100 μl of methanol, vortexed, and centrifuged at 5000 rpm (VS-15000 CFNII, Vision Scientific, Daejeon, South Korea) for 5 min at 4 °C. Then, 100 μL of the supernatant was added to 900 μL of FOX2 reagent (Wolff's reagent; 4 mM BHT, 250 mM FeSO4, 250 mM H2SO4, and 100 mM xylenol orange). The absorbance was measured at 560 nm in in a microplate reader (BioTek Synergy HT, BioTek Instruments, Highland Park, VT, USA). The results are expressed as [hydroperoxides] nmoL.min −1 / mg of tissue.

Experimental design
In all experiments, the blood glucose was assessed 72 h after the experimental induction of T1DM with STZ and at the end of the experiments, being this last measurement considered for data analysis. The weight gain considered in the analysis was the subtraction of the body weight (taken after the experiments) from that body weight taken immediately before the STZ injection.

Experiment 1
This experiment was carried out to evaluate the impact of CFC on the fear extinction memory (short and long-term) and on late anxiety-like responses in induced-T1DM (STZ) and normoglycemic (NGL) animals. The animals were divided into groups exposed or not exposed to the CFC, conditioned (Cond) and non-conditioned (N.Cond), respectively. Thus, four groups were formed: STZ/Cond (n = 10), STZ/N.Cond (n = 9), NGL/Cond (n = 10), and NGL/N.Cond (n = 10).
On the 24 th day after the induction of diabetes, part of the animals was submitted to the protocol of CFC and part of them, the N.Cond group, was exposed to the same context but they did not receive footshocks (US). On day 25, all animals were submitted to the extinction training session. Twenty-four hours after, they were evaluated in the same context (Extinction test 1). Seven days later, the Extinction test 2 was performed followed by the EPMT, on the day 33. Then, the body weight and blood glucose (BG) were checked one more time.

Experiment 2
In this experiment, we aimed to evaluate if one single injection of PGB before the fear memory extinction training would facilitate this acquisition. Moreover, if this single injection would induce a long-term effect on the extinction memory and would disrupt the late anxious-like responses of induced-T1DManimals (STZ). Hence, five groups -all subjected to CFC-treated with vehicle (VEH) or pregabalin (PGB) were formed: NGL/VEH (n = 7), STZ/ VEH (n = 7), STZ/PGB 30 mg/Kg (n = 10), STZ/PGB 100 mg/Kg (n = 10), and STZ/PGB 300 mg/Kg (n = 10).
On the 24 th day after the induction of diabetes, animals were submitted to the protocol of CFC. On day 25, one hour before the Extinction training session all animals received injection of PGB (different doses) or VEH. Twenty-four hours later, they were evaluated in the same context (Extinction test 1) and after seven days, the Extinction test 2 1was performed followed by the EPMT (day 33 th ). Then, the body weight and blood glucose (BG) were checked. Experiment 3 Bearing in mind that higher doses of PGB can cause a sedative effect, this experiment was performed to investigate whether during the extinction training session (occurred in the Experiment 2) the animals treated with PGB were under a sedative effect. Thus, five groups -all subjected to contextual conditioned fear -treated with vehicle (VEH) or pregabalin (PGB) were formed: NGL/VEH (n = 6), STZ/VEH (n = 6), STZ/PGB 30 mg/Kg (n = 6), STZ/PGB 100 mg/Kg (n = 6), and STZ/PGB 300 mg/Kg (n = 7).
On the 24 th day after the induction of diabetes, animals were submitted to the protocol of CFC. In the next day, all animals received injection of PGB (different doses) or VEH and 1 h after they were submitted to the OFT. Then, the body weight and blood glucose (BG) were checked one more time.
Experiment 4 This in vitro experiment was performed to evaluate the antioxidant potential of PGB, according to the method of Chen et al. (2004) with modifications.

Experiment 5
This study was carried out to assess whether the PGB in its lowest dose (30 mg/kg) and highest dose (300 mg/kg) would present a neuroprotective profile in induced-T1DM animals (STZ). Thus, on the 24 th day after the induction of diabetes the animals were conditioned to the context with 3 footshocks (US), and in the next day (25 th day) they received a single injection of VEH or PGB 1 h before the extinction training session. The animals were euthanized immediately after this session of extinction and had HIP and CPF dissected for analysis of the GSH and LPO levels. For comparative purposes, these analyses were also performed in the HIP and PFC from NGL animals. The sample n for each 4 groups formed for analysis in the HIP and PFC were: For HIP-NGL/VEH (n = 6), STZ/ VEH (n = 5), STZ/PGB 30 mg/Kg (n = 7), and STZ/PGB 300 mg/Kg (n = 7). For PFC-NGL/VEH (n = 6), STZ/VEH (n = 5), STZ/PGB 30 mg/Kg (n = 7), and STZ/PGB 300 mg/ Kg (n = 7).

Statistical analysis
The Shapiro-Wilk normality test was initially used to ensure that the data met the criteria for performing parametric tests.
When the criteria were met, the results were reported as the mean ± standard error of the mean (SEM). For the experiment 1, two-way analysis of variance (ANOVA) with or without repeated measures (RM) was performed. To analyze the data obtained from extinction training session we considered as independent variables the different groups and the time (blocks of time) being the time the RM. In the other experimental sessions, the condition of the animals (NGL or STZ) and the procedure (Cond and N.Cond) were considered as independent variables. Freezing time was the dependent variable. For the experiment2, two-way analysis of variance (ANOVA)with or without repeated measures (RM) was used. To analyze extinction training data, we considered the different groups and the time as independent variables, being the time a RM. The freezing time was used as the dependent variable. In the other experimental sessions and for the experiment 3 (OFT) and 4 (oxidative stress analysis), the one-way ANOVA was applied.
When appropriate, Bonferroni post-hoc test was used to perform multiple comparisons and the differences were considered statistically significant when p < 0.05. All the tests were carried out using the GraphPad Prism program (version 8, San Diego, CA, USA).

Short-and long-term effects of CFC on the fear extinction memory in induced-T1DM and NGL animals
Two-way ANOVA with RM showed a significant difference when the data from extinction training session were analyzed (Fig. 1A)-difference for blocks of time [F (1, 35) = 159.9, p < 0.05] as well as groups [F (3, 35) = 175, p < 0.05], and interaction between these variables [F (3, 35) = 47.82, p < 0.05]. The CFC procedure increased the freezing time of NGL and induced-T1DM(STZ) animals when compared with its respective N.Cond groups(p < 0.05). Also, the freezing time was reduced in both Cond groups-STZ and NGLwhen comparing the 4 th block with the 1 st block (p < 0.05). However, the freezing time was even more pronounced in STZ animals (Cond group), when compared with NGL (Cond group) at 4 th block (p < 0.05).
In the analysis of the freezing time during the extinction test1 session (Fig. 1B), two-way ANOVA showed statistically significant difference for the condition of animals Also, an increase in freezing time was observed in the STZ(Cond) group in these two sessions when compared to its respective N.Cond group (p < 0.05) and also to NGL (Cond) (p < 0.05).

Long-term effects of CFC procedure followed by fear extinction memory training on anxiety-like responses in induced-T1DM and NGL animals
In the analysis of the anxiety-like behavior ( Fig. 2A-D), the two-way ANOVA showed significant difference for the condition of animals (NGL and STZ) when time [F (1, 35) = 43.06, p < 0.05; Fig Fig. 2A].
Bonferroni post-hoc test showed that among NGL animals, compared to N.Cond group, the Cond group presented a decrease on the time spent in the open arms and number of entries in these same arms (p < 0.05), indicative of anxiogenic-like response. Comparing NGL and STZ (Cond groups), STZ animals exhibited a more pronounced decrease on time spent in the open arms and number of entries in these arms (p < 0.05), suggestive of a more expressive anxiogenic-like effect. When head dipping was evaluated, Bonferroni test showed a decrease in this behavior in STZ animals (Cond and N.Cond groups) compared to NGL (N.Cond) (p < 0.05).Also, a significant difference (p < 0.05) in the entries in the closed arms was observed in STZ animals, compared to NGL animals, independent of the procedures (p < 0.05).

Short-and long-term effects of a single injection of PGB on the acquisition of fear extinction memory in induced-T1DM animals
As can be seen in Fig. 3 when analyzed the extinction training data (Fig. 3A), two-way ANOVA with RM showed statistical difference between the groups [F (4, 36) = 14.81, p < 0.05],difference in time [F (1, 36) = 104.8 p < 0.05] and interaction between these variables [F (4, 36) = 9.737 p < 0.05].Bonferroni post-hoc analysis showed a significant reduction in freezing among all the first PGB block groups compared to STZ/VEH 1 st block group (p < 0.05).The NGL, STZ and PGB300 groups showed a significant difference in relation to their respective 1 st block group (p < 0.05).

Long-term effects of a single injection of PGB on anxiety-like responses in T1DM animals exposed previously to the CFC procedure and fear extinction training
In the EPMT, one-way ANOVA showed that the treatment was able to change the percentage of time spent in the open arms [F (4, 36) = 5.322, p < 0.05; Fig. 4A], the percentage of entries in the open arms [F (4, 36) = 11.89, p < 0.05; Fig. 4B], frequency of head dipping [F (4, 36) = 5.661, p < 0.05; Fig. 4C], but did not change the total number of entries in the closed arms [F (4, 36) = 3.283, p > 0.05; Fig. 4D]. Bonferroni post-hoc test showed that STZ animals treated with PGB 300 presented a significant reduction of the anxious-like behavior when compared to STZ/VEH during time and entries in the open arms, and frequency of head dipping (p < 0.05). Moreover, regarding to anxiety, STZ groups treated with PGB30 and/or PGB100 presented an increase in the percentage of time spent and entries in the open arms (p < 0.05).
One-way ANOVA showed a significant difference among the groups when weight gain [F (4, 36) = 119.5, p < 0.05] and blood glucose [F (4, 36) = 56.92, p < 0.05] were analyzed. Bonferroni post-hoc test showed that all STZ animals presented a significant reduction in the weight gain (p < 0.05) and an increase in the glycemia compared to NGL animals (p < 0.05) (see Table S1-supplementary material) and these parameters in STZ animals were not altered by any treatment employed in this study.

Experiment 3
Effect of a single injection of PGB on the locomotor activity ofinduced-T1DM animals submitted to the OFT As shown in Fig. 5, one-way ANOVA test showed difference between NGL and STZ groups in the number of crossings in the OFT [F (4, 26) = 30.67, p < 0.05]. The Bonferroni posthoc test showed that STZ animals treated with PGB 300 presented a significant decrease in the number of crossings compared to groups STZ/VEH and NGL/VEH (p < 0.05).

Effects of a single injection of PGB on parameters related to the oxidative stress, evaluated in the HIP and PFC ofinduced-T1DM animals
As shown in Fig. 7 test showed that STZ/VEH animals presented a significant increase of LPO (PFC and HIP) and decrease of GSH (PFC and HIP), when compared with NGL/VEH (p < 0.05). While animals treated with PGB 300 mg/kg showed a significant reduction in the LPO in PFC and HIP compared to the STZ/ VEH group (p < 0.05), a significant increase of GSH in the PFC was also observed when PGB 30 and 300 mg/kg was administered.

Discussion
The main findings of the present study are that STZ animals show a persistent impairment in extinguishing a fear memory associated with delayed fear sensitization demonstrated by the increased anxiety-like response. Interestingly, a single injection of PGB in these STZ animals just before fear extinction training was able to facilitate fear extinction memory and alleviate anxiety-like behavior. These effects were observed in short and long-term (9 days later) periods. In addition, our data indicate that the beneficial effects induced by PGB may be related, at least partially, to its antioxidant activity demonstrated in the HIP and PFC.
As already observed (Ikeda et al. 2015(Ikeda et al. , 2021de Souza et al. 2019;Ribeiro et al. 2020), our data confirmed that STZ animals present a greater fear response and a difficulty in extinguishing fear memory when compared to NGL animals ( Fig. 1), indicating an overconsolidation of this fear memory. However, this is the first study to show the persistence of this fear memory in animals with induced-T1DM when reexposed in the same context after 1 week later (Fig. 1C) indicating how strong or dysfunctional are the mechanisms associated with this type of learning/memory. In this same direction, clinical studies point out a higher prevalence of PTSD in adults with T1DM (Renna et al. 2016), along with a positive association between the trauma of war and an increase in the incidence of T1DM in children and adolescents (Zung et al. 2012). The presence of severe and moderate PTSD symptoms has also been demonstrated in children (8-18 years) diagnosed with T1DM (Auxéméry 2012).
Regarding anxiety, STZ animals (not exposed to the same context of CFC) showed a more exacerbated anxietylike response (decrease in the time spent and the number of entries in the open arms and decrease in the head dipping frequency) when compared to NGL animals-conditioned or not ( Fig. 2A-C). These findings reinforce previous reports showing that these induced-T1DM animals present a more pronounced anxious-like behavior when compared to NGL animals (de Morais et al. 2014;Gambeta et al. 2016;Rebolledo-Solleiro et al. 2016;de Souza et al. 2019). Concerning NGL animals, even after acquiring extinction memory, they exhibited a more pronounced later anxiety-like behavior than non-conditioned NGL animals ( Fig, 2A-D). field test. The scheme above graphs represents the experimental design used in this experiment. *p < 0.05 when compared to NGL (VEH) group; # p < 0.05 compared to STZ (VEH) group Fig. 6 Effect of pregabalin, ascorbic acid (AA; positive control) and water (negative control) on free radical scavenging activity in the DPPH assay. Results are expressed as mean ± SEM; n = 3, * p < 0.05 compared to VEH This result is not surprising since other evidence show that previous stress may induce a response related to a fear sensitization or even to a generalization of fear memory, which depends on several aspects, such as external factors including the type and intensity of aversive stimulation, early-life stress, as well as the saliency of particular elements in the environment (Korte and de Boer 2003;Asok et al. 2019).
The next block of experiments was designed based on two main premises: 1. Induced-T1DM rats have an overconsolidation of fear memory, in addition to demonstrating the persistence of this memory and anxiety-like behavior, and 2. the use of PGB is approved for diabetic neuropathic pain, and its use has been successfully employed for treating more severe states of anxiety disorders and also PTSD (Pande et al. 2003;Pohl et al. 2005;Garakani et al. 2020). Our results revealed that all doses of PGB were able to decrease the freezing time when submitted to the extinction tests 1 and 2 (Fig. 3B-C), indicating that the drug facilitated the acquisition of this extinction memory, being this effect of long-lasting. Interestingly, the highest dose of PGB (300 mg/ kg) was able to restore the behavior of the STZ animals, equating it with NGL animals (Fig. 3B-C). Regarding the delayed effect on the anxiety-like response, PGB (two highest doses) restored the most expressive anxious-like behavior of STZ animals, while only the highest dose (300 mg/kg) was able to induce a significant improvement in the behavior of head dipping (Fig. 4C). That is, the beneficial effect of PGB on fear memory seems not to be dissociated from its anxiolytic-like effect.
Although we found a sedative effect caused by the acute injection of PGB in the dose of 300 mg/kg (during extinction training) as demonstrated in the OFT (Fig. 5), the animals showed significant and consistent effects on fear extinction memory ( Fig. 3B-C) and anxiety-like responses (Fig. 4A-C) in the subsequent tests in which they were not under this acute effect of PGB, once the half-life of PGB is around 6 h (Buoli et al. 2017). Thus, this sedative effect during the extinction training session did not impair the acquisition of fear extinction memory and the anxiolytic-like effect. In addition, lower doses of PGB (30 mg/kg and 100 mg/kg) also showed effects in reducing freezing and improving anxiety, without inducing a sedative effect (Fig. 5).
At this point, we could mention as a limitation of the present study the fact that the best effects have been achieved after the administration of the highest dose of PGB. In this sense, it is known that the clinical efficacy dosage of PGB in GAD, for example, is 150-600 mg/day, and the frequently reported side effects (drowsiness and dizziness) are dose-dependent (Baldwin et al. 2015). In double-blind, placebo-controlled trials the adverse events during treatment with PGB were considered mild and moderate (Mann et al. 2014). However, in a recent metaanalysis study, Onakpoya et al. (2019) demonstrated that despite being effective for the treatment of diseases such Fig. 7 Effect of a single injection of pregabalin (30, 300 mg/kg) or vehicle (VEH) in normoglycemic (NGL) or STZinduced T1DM (STZ) animals on lipid peroxidation levels (LPO, panel A -prefrontal cortex and panel B-hippocampus) and reduced glutathione levels (GSH, panel C -prefrontal cortex and panel D-hippocampus). Results are expressed as mean ± SEM; n = 6-7; a p < 0.05 compared to NGL/VEH; b p < 0.05 compared to STZ/VEH as diabetic peripheral neuropathy and post-herpetic neuralgia, many patients discontinue the treatment due to adverse effects. Furthermore, some studies have reported cases of increased heart failure, peripheral edema, and also Addiction. All these effects seem to be related to the chronic use of PGB (Robert Lee Page et al. 2008;Grosshans et al. 2010;Dobrea et al. 2012;Gahr et al. 2013;Aldemir et al. 2015). According to Buoli et al. (2017) at the highest doses of PGB, there is an increase of the enzyme responsible for the synthesis of GABA (L-Glutamic acid decarboxylase), and although PGB does not directly bind to GABA-A or GABA-B receptors, these receptors may be more active due to increased GABA levels (Buoli et al. 2017). Also, at the highest dose, PGB may bind to α2δ-2 subunit, in addition to the α2δ-1 subunit, which in the brain is concentrated in the cerebellum and partially correlates with GABAergic neurons (Barclay et al. 2001;Li et al. 2011). Thus, the adverse effects observed at the highest doses of PGB may be related preferentially to the unspecific effects of PGB on other sites of action. Here, if we think in translational terms, it is important to bear in mind what would be the risk and benefit of the treatment for the patient, and this response should be linked to the type of treatment and the dosage used. For example, in the present study, despite the most effective dose being the highest dose, which could cause a series of unwanted effects in continuous treatment, the PGB application was unique and possibly not associated with the consequences of adverse effects.
The International guidelines consider PGB along with the selective serotonin reuptake inhibitors (paroxetine, sertraline, escitalopram), and serotonin noradrenaline reuptake inhibitors as first-line options for treating GAD (Pande et al. 2003;Buoli et al. 2017;Greenblatt and Greenblatt 2018). Moreover, both open-label and randomized, double-blind, placebo-controlled studies have demonstrated the efficacy of PGB, being in many of these studies superior to antidepressants and benzodiazepines depending on the parameter evaluated (for a review see Frampton 2014;Baldwin et al. 2015 andBuoli et al. 2017). Furthermore, compared to other GABA analogues such as gabapentin, PGB has several pharmacokinetic and pharmacodynamic advantages. In this sense, PGB takes greater potency than gabapentin as an α2-δ ligand (Greenblatt and Greenblatt 2018). In addition, the peak plasma concentrations of pregabalin occur after 1 h compared to 3 h with gabapentin and the oral bioavailability for PGB is over 90% compared to 30-60% for gabapentin (Bockbrader et al. 2010). Another advantage of PGB is in relation to absorption, which is linear and non-saturable, unlike gabapentin which is easily saturable, resulting in a dosedependent kinetic dose (Bockbrader et al. 2010;Chincholkar 2020;Greenblatt and Greenblatt 2018). For example, in patients with compromised renal function, the risk of toxicity is greater compared to the dose required for the desired effect (Blum et al. 1994).
Regarding treating PTSD with PGB, although studies are still limited, a retrospective clinical study of 290 burned service members revealed that PGB or gabapentin did not affect the development of PTSD (Fowler et al. 2012). However, in a randomized clinical trial, Baniasadi et al. (2014) demonstrated that PGB was able to reduce emotional symptoms of PTSD related to combat in a group of 18 male patients (non-diabetics) who received 300 mg/day of PGB for a period of 6 weeks. In preclinical studies, it is evident that depending on the experimental protocol, such as the condition of the animal (including the animal´s condition or the type of stress), doses used, and the time of treatment and its duration, the results are quite different. For example, Zohar et al. (2008) showed, in non-diabetic animals, that PGB (at the same doses used in the present study) induced beneficial short-term effects on behavioral responses (anxiety-like response) in animals exposed to traumatic stimuli (predator urine scent). However, the anxiolytic-like effect of PGB was observed only in these animals pre-exposed to predator urine scent, and not in that one's not pre-exposed. Also, these effects were not observed after 30 days of the PGB injection. Differently, Valdivieso et al. (2018) did not observe any beneficial effect after PGB treatment on anxiety-like behavior in previously stressed (restraint and tail shock) nondiabetic animals. Nevertheless, they used a smaller dose of PGB (10 mg/kg).
Of particular interest for the present study, all these changes related to an exacerbated behavioral response observed in induced-T1DM animals have been related to the diabetic encephalopathy (Gupta et al. 2014;Prabhakar et al. 2015;da Silva Dias et al. 2016;Zanoveli et al. 2016;Aswar et al. 2017;Wang et al. 2019;Chaves et al. 2020). In this sense, these studies report in induced-T1DM rats an increase in inflammatory processes and oxidative stress, together with a dysregulation of neurotransmission systems in brain areas related to emotions, such as HIP and PFC. In addition, these animals present increased anxiety and/or depressive-like behaviors (Buchberger et al. 2016;Chaves et al. 2020;De Morais et al. 2014;Moulton et al. 2015;Muriach et al. 2014;Gupta et al 2014;Zanoveli et al. 2016). Moreover, it is known that an increase in corticosterone level is up to twice as high in a TDM1 model when compared to normoglycemic animals (Kuznetsova et al. 2014;Diz-Chaves et al. 2016), increasing levels of glucocorticoids and corticotropin-releasing hormone. This hyperactivation of the HPA axis can generate metabolic changes that favor the development and/or worsening of diabetes, since glucocorticoids can increase resistance to insulin action, further aggravating the hyperglycemic condition and consequently increasing oxidative stress. The same has been observed in animal models of depression induced by exogenous corticosterone administration. For example, Gupta et al (2015) reported in mice exogenously treated with corticosterone that they exhibit depression-like behavior indicated by increased despair effects in forced swim test and anhedonia in sucrose preference test. In addition, corticosterone administration induced oxidative load in the brain with significant increase in pro-oxidant markers and a substantial decline in antioxidant defense system, indicating a direct effect of stress hormones in the induction of the brain oxidative damage.
In view of these alterations in these animals with T1DM, in the following experiments (Experiment 4 and 5) we investigated whether PGB would present an antioxidant activity per se and whether this single injection of PGB, 1 h before the fear extinction memory training, would exert an antioxidant action by improving indirect oxidative stress parameters in the HIP and PFC from STZ rats. Our data demonstrated that PGB (in all concentrations- Fig. 6) presented an antioxidant activity by showing a lower absorbance level compared to the negative control (water). Also, PGB was able to reduce the increased LPO levels of STZ animals (HIP and PFC-highest dose) and to increase the reduced GSH levels (PFC). These changes on reduced GSH and LPO levels in these brain regions of STZ animals (Fig. 7) have been previously demonstrated (Pitocco et al. 2010;Pereira et al. 2018;de Souza et al. 2019;Réus et al. 2019). Regarding LPO, it is known that high levels of LPO are attributed to the increase of reactive oxygen species (Siba et al. 2017); in addition, it affects cell integrity only when antioxidant mechanisms are no longer able to cope with the generation of free radicals (Anwer et al. 2012). The GSH is the main antioxidant in the brain (Kanazawa et al. 2016) and it protects the cellular system against the toxic effects of LPO. Thereby, according to Sharma et al. (2016), this decline in the level of GSH in induced-T1DM animals may be due to excessive free radical generation, exposure to high glucose levels. In addition, it has been proposed that antioxidant compounds may play an important role in improving the dysregulations reported here, especially in brain regions such as PFC and HIP, linked to memory/learning and emotional processing (Venturini et al. 2010;de Morais et al. 2014;Pereira et al. 2018;Banagozar Mohammadi et al. 2019;de Souza et al. 2019).
In this sense and reinforcing our data after PGB treatment, we have already observed in this animal model of T1DM that prolonged treatment with other antioxidant compounds, the vitamin E or gallic acid, improved the dysfunctional processing of fear memory and/or anxiety-like responses from these induced-T1DM animals along with an improvement on oxidative stress related-parameters in these same brain areas (de Morais et al. 2014;Pereira et al. 2018;de Souza et al. 2019). This antioxidant action of PGB has already been demonstrated in STZ animals (Sałat et al. 2016;Demir et al. 2021). For example, Sałat et al. (2016) evaluated the effects of the single injection (i.p.) of PGB in STZ mice on contextual memory (-not related to fear) and oxidative stress parameters. The authors reported that although PGB in a smaller dose (10 mg/kg) was not able to attenuate T1DMinduced memory impairments, the PGB did not aggravate learning deficits of these diabetic mice. However, this dose was able to disrupt some markers of oxidative stress. More recently, Demir et al. (2021) demonstrated that prolonged treatment with PGB (50 mg/kg/day per orally for 8 weeks) was able to reduce LPO, improve antioxidant capacity with a significant increase in superoxide dismutase and protect cells against apoptosis. Other studies conducted in non-diabetic animals also demonstrate that peripheral and central effects of PGB were able to decrease oxidative stress and inflammation (Asci et al. 2020;Yamamoto et al. 2021). Considering these data, together with the wide distribution of α2δ1 subunits and the ability of PGB to surpass the BBB (Takahashi et al. 2018;Cole et al. 2005), we cannot affirm in our study that our data is due to a central or a peripheral effect. It also could be due to a summation effect of both -central and peripheral effects. In that sense, studies indicate that the peripheral administration of PGB may also be beneficial to cognition, emotion, and memory (La Porta et al. 2016;Sałat et al. 2017).
It is important to note that in the present study, the lowest dose of PGB (30 mg/kg), which was the dose that changed more discreetly the behavioral parameters evaluated in the present study, was able to increase only the reduced level of GSH in the PFC. Thus, we can speculate that the antioxidant action of PGB may not be the only mechanism responsible for the beneficial action of PGB on fear memory and anxiety-like response in these induced-T1DM animals. In this regard, it is not surprising that PGB induces neuroprotective effects through other mechanisms, including an anti-inflammatory action by disrupting the expression of several inflammatory markers (Sałat et al. 2016;Aslankoc et al. 2018;Cruz-Álvarez et al. 2018;Ali et al. 2019). Curiously, in 2014 we had already observed in this animal model of T1DM that a prolonged treatment with insulin, vitamin E or imipramine induced antidepressant-like effect that was related to the improvement on oxidative stress parameters in the PFC and HIP (de Morais et al. 2014). However, since all treatments were not able to induce the complete reversal of the depressive-like state of T1DM animals, we suggested that oxidative stress seemed not to be the unique factor involved in the pathophysiology that reports depression to diabetes. The same can be thought based on the data from the current study. In fact, further studies are needed to be conducted to better understand the pathophysiological mechanisms that link memory and anxiety to diabetes.
Taken together, a single injection of PGB in induced-T1DM rats in a specific time window-before a fear memory extinction training session-facilitates the acquisition of fear extinction memory in the short-and long-term, being this effect additionally associated with late improvement on anxiety-like behavior. All these short-and long-term beneficial actions of PGB may be associated with neuroprotective mechanisms of PGB, including its antioxidant action in brain areas like PFC and HIP. Despite the need for further investigations, the data are quite interesting and highlight the potential of PGB for future translational investigations, taking into account the cost-benefit of PGB as a facilitator or adjuvant in the process of extinction of the traumatic memories, in addition to anxiolytic effects.