The phosphodiesterase 5 inhibitor, KJH‐1002, reverses a mouse model of amnesia by activating a cGMP/cAMP response element binding protein pathway and decreasing oxidative damage

Abstract

Background and Purpose

Inhibition of PDE5 improves synaptic plasticity and memory via enhancing cGMP expression, thus activating the cGMP/cAMP response element binding protein (CREB) signalling pathway. This study investigated the effects of a PDE5 inhibitor on scopolamine‐induced cognitive dysfunction, using memory‐related behavioural tests and biochemical assays.

Experimental Approach

In mice were pretreated with PDE5 inhibitor, amnesia was induced by scopolamine. The learning and memory abilities of mice were tested using the Morris water maze test, the Y‐maze test, the passive avoidance test and the novel object recognition test in sequence. Expression of memory‐related bio‐molecules and oxidative stress parameters in brain tissue were measured using Western blot and spectrophotometry respectively.

Key Results

KJH‐1002, a novel and potent inhibitor of PDE5 (IC₅₀ 0.059 ± 0.04 nmol·L⁻¹), was synthesized. In the behavioural tests, it markedly improved the memory performance impaired by scopolamine, indicating a restoration of cognitive function in the mice. Moreover, KJH‐1002 increased cGMP levels in the cortex and the scopolamine‐reduced expression of phosphorylated CREB, Levels of ERK 1/2, Akt and brain‐derived neurotrophic factor in the cortex and hippocampus were restored by KJH‐1002 treatment. In addition, KJH‐1002 administration increased the activities of SOD, glutathione peroxidase and glutathione reductase, and decreased the level of malondialdehyde.

Conclusion and Implications

KJH‐1002 restored cognitive function in scopolamine‐induced amnesia mice by activating the cGMP/CREB signalling pathway and attenuating oxidative stress. The beneficial effects of KJH‐1002 on cognition indicate its potential as a therapeutic candidate for Alzheimer's disease.

Abbreviations

BDNF

brain‐derived neurotrophic factor

CREB

cAMP response element binding protein

GPx

glutathione peroxidase

GR

glutathione reductase

MDA

malondialdehyde

MWM

Morris water maze

NOR

novel object recognition

Introduction

The cyclic nucleotides, http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2352 and http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2347 are second messengers that can modulate signal transduction by activating the phosphorylation of other enzymes or transcription factors in a variety of biological systems. The phosphodiesterases (PDEs) comprise 11 subtypes and hydrolyse cAMP and/or cGMP and thereby alter the cAMP/cGMP signalling pathways (Hutson et al., 2011; Bollen et al., 2014). http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1304&familyId=260&familyType=ENZYME, the enzyme that selectively breaks down cGMP, is expressed in a variety of brain regions related to cognitive function (Van Staveren et al., 2003). Numerous studies have demonstrated that cyclic nucleotides are involved in learning and memory and cGMP plays a key role in early memory consolidation, while the late phase memory consolidation process is mainly attributed to cAMP (Bollen et al., 2014). Inhibition of PDE5 raises levels of cGMP which that activates the phosphorylation of the cAMP response element binding protein (CREB), thereby enhancing synaptic plasticity and cognition (Puzzo et al., 2009; Cuadrado‐Tejedor et al., 2011; Garcia‐Osta et al., 2012).

Alzheimer's disease (AD), defined as the progressive and irreversible decline of cognitive function, is the most prevalent form of dementia. Deficits in memory and learning are typical features of AD patients, and cognitive impairment is attributed to a marked loss of cholinergic neurons and decreased levels of http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=294, via http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=2465&familyId=765&familyType=ENZYME (Nedaei et al., 2016). Oxidative stress is another crucial neurodegenerative disorder factor in AD patients (Gella and Durany, 2009; Bonda et al., 2010). http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=330, a non‐selective http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=285 antagonist, disturbs learning and memory via blocking the effects of the neurotransmitter ACh (Gautam et al., 2016; Kulshreshtha and Piplani, 2016). It also induces neurotoxicity by altering the levels or activities of antioxidant enzymes and abnormal expression of proteins (Li et al., 2016; Venkatesan et al., 2016).

Accumulating evidence has demonstrated administration of PDE5 inhibitors, such as http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=4743 and http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=7320 enhance cognitive function by increasing the cGMP levels in the brains of various animal models (Prickaerts et al., 2002; Rutten et al., 2005; Puzzo et al., 2009). Sildenafil improves memory consolidation in object recognition tasks in unimpaired rats (Prickaerts et al., 2005) and mice (Rutten et al., 2005). Another study (Cuadrado‐Tejedor et al., 2011) reported that sildenafil ameliorates memory in the Morris water maze (MWM) task in an AD model in Tg2576 transgenic mice. Moreover, scopolamine‐impaired learning and memory in 14‐unit T‐maze task was attenuated by sildenafil administration (Devan et al., 2004). Cognitive function is regulated by complicated signalling pathways, and http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=4872, a downstream target protein of CREB, plays a key role in memory plasticity and formation and facilitates hippocampal cell proliferation (Erickson et al., 2010). Moosavi et al. (2012) suggested that the hippocampal kinases, http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=514 and activated http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=285 play important roles in regulating proteins in synaptic plasticity and different kinds of learning and memory. Furthermore, Akt and CREB are described as enhancers of neuronal survival, as they exert anti‐apoptotic effects on neuronal cells, and they can be regulated by cGMP (Takada‐Takatori et al., 2006; Calabrese et al., 2007).

New derivatives of existing bioactive compounds can be synthesized by the modification of heterocyclic nucleus, rigid coplanarity or variation of bioisosteric factors. To develop new potent and selective PDE5 inhibitors, we synthesized a novel compound by bioisosteric transformation based on the structure of sildenafil and have tested its inhibition of PDE5. We then utilized a mouse model of amnesia, induced by scopolamine, to mimic the cognitive dysfunction symptoms in AD patients. To evaluate the enhancing effects of this synthesized compound on memory and learning, the mice were subjected to a series of behavioural tests. Furthermore, to elucidate the mechanisms underlying the improvement of cognitive function by this compound, biochemical analyses of brain tissues were carried out.

Methods

Animals

All animal care and experimental protocols in this study complied with the Korea Institute of Science and Technology Animal Care Committee guidelines. Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny et al., 2010; McGrath and Lilley, 2015). We took all effort to minimize the number of mice used and to relieve any suffering.

Imprinting control region mice (8‐week‐old males) were purchased from Orient Bio Inc. (Seongnam, Korea). Four mice were housed in each cage (30 × 18.5 × 13 cm), with wood shaving bedding and nesting material, and they were allowed unlimited access to food and water. The animals were housed under constant conditions (12 h light/dark cycle, lights on from 6:00 to 18:00; temperature 23 ± 1°C: humidity 50 ± 10%). To habituate the mice to the ambient environment prior to the experiments, they remained under these conditions for 1 week after arrival.

Grouping and pretreatment

Forty mice were randomly divided into five groups (eight mice per group), and all the mice in each group were subjected to a 7‐day pretreatment period. Mice in the normal control group (Control) and in the scopolamine control group (Vehicle) were given saline (p.o.). Mice in the positive control group (DON) were treated with http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=6599 (4 mg·kg⁻¹, p.o.), and mice in sample‐treated groups (KJH‐1002‐10 and KJH‐1002‐20) were given 10 or 20 mg·kg⁻¹ KJH‐1002 (p.o.) respectively. All treatments were given by gavage, once daily in a volume of 10 mL·kg⁻¹. The group size for each experiment was determined by referring to previous studies (Moosavi et al., 2012; Li et al., 2016; Nedaei et al., 2016) using similar protocols to those of the present study. At least five independent samples are needed to detect significant differences between groups. Therefore, we decided to use six or eight samples for each experimental group (in vitro or in vivo).

Preparation of KJH‐1002

The synthetic pathway is summarized in Figure 1. Numbers in bold refer to the compounds in this pathway, as shown in the Figure.

Figure 1.

Synthesis of the potent PDE5 inhibitor KJH‐1002. (A) P₄S₁₀, toluene, reflux, 2 h; (B) (i) ClSO₃H, SO₂Cl₂, room temperature, 13 h; (ii) N‐methylpiperazine, ethanol, room temperature, 15 h.

Synthesis of 4‐(2‐ethoxyphenyl)‐1‐methyl‐3‐N‐propyl‐1,6‐dihydro‐7H‐pyrazolo[4,3‐d]pyrimidine‐7‐thione (2)

5‐(2‐Ethoxyphenyl)‐1‐methyl‐3‐N‐propyl‐1,6‐dihydro‐pyrazolo[4,3‐d]pyrimidine‐7‐one (1) 593 mg (1.186 mmol) and Lawesson's reagent (P₄S₁₀) 400 mg (0.953 mmol) in toluene (7 mL) were refluxed for 2 h. The solvent was concentrated by a rotary evaporator and the residual solution was purified using column chromatography (eluent was chloroform: ethyl acetate = 10:1) to yield the required compound as a yellow solid (547 mg, 89%).

¹H‐NMR (300 MHz, CDCl₃): δ 12.64 (bs, 1H), 8.46 (d, J = 7.9 Hz, 1H), 7.45 (t, J = 8.4 Hz, 1H), 7.13 (t, J = 7.5 Hz, 1H), 7.04 (d, J = 8.4 Hz, 1H), 4.52 (s, 3H), 4.18 (t, J = 6.5 Hz, 2H), 2.94 (t, J = 7.4 Hz, 2H), 2.6 (m, 2H), 1.85 (m, 2H), 1.18 (t, J = 7.4 Hz, 3H), 1.02 (t, J = 7.4 Hz, 3H); ¹³C‐NMR (75 MHz, CDCl₃): δ 171.90, 157.17, 148.34, 146.52, 134.74, 133.14, 132.63, 131.14, 122.18, 119.33, 113.06, 71.80, 39.71, 28.05, 22.90, 22.70, 14.45, 11.27; mp: 110–111°C.

Synthesis of 5‐[2‐ethoxy‐5‐(4‐methyl‐piperazine‐1‐sulfonyl)‐phenyl]‐1‐methyl‐3‐propyl‐1,6‐dihydro‐pyrazolo[4,3‐d]pyrimidine‐7‐thione (3)

4‐(2‐Ethoxyphenyl)‐1‐methyl‐3‐N‐propyl‐1,6‐dihydro‐7H‐pyrazolo[4,3‐d]pyrimidine‐7 thione (4.26 g, 0.0130 mol) was added to chlorosulfonic acid (8.63 mL, 0.123 mol) in a portionwise manner in an ice bath, followed by the addition of thionyl chloride (1.42 mL, 0.0195 mol). The mixture was stirred at room temperature overnight and then poured on ice (40 g) dropwise. The yellow solid was filtered and washed with water, and the wet solid was suspended in EtOH (55 mL), prior to 1‐methylpiperazine (5.09 mL, 0.0454 mol) being added in an ice bath. The mixture was stirred at room temperature overnight, and the solvent was removed under reduced pressure. The residue was added to a 6N aqueous solution of NaOH (15 mL) and extracted with CH₂Cl₂ (40 mL). The organic layer was washed with brine (10 mL), dried with anhydrous MgSO₄, filtered and evaporated to dryness. The crude product was purified by crystallization from EtOH (40 mL) to yield the required compound as a yellow solid (5.75 g, 90.3%).

¹H‐NMR (300 MHz, CDCl₃): δ 12.41 (bs, 1H), 8.82 (d, J = 2.3 Hz, 1H), 7.82 (dd, J = 8.7 Hz, J' = 2.3 Hz, 1H), 7.19 (d, J = 8.8 Hz, 1H), 4.51 (s, 3H), 4.41 (q, J = 6.9 Hz, 2H), 3.10 (bs, 4H) 2.95 (t, J = 7.5 Hz, 2H), 2.50 (bs, 4H), 2.27 (s, 3H), 1.86 (m, 2H), 1.72 (t, J = 6.9 Hz, 3H), 1.02 (t, J = 7.4 Hz, 3H). ¹³C‐NMR (75 MHz, CDCl₃): δ 172.09, 159.83, 146.81, 146.43, 134.23, 132.64, 132.35, 131.03, 129.29, 120.09, 113.56, 66.81, 54.38, 46.31, 46.09, 39.74, 27.97, 22.53, 15.05, 14.41. IR (KBr pellet, cm⁻¹) 3278, 2940, 2790, 1538; mp: 181–181.5°C.

Synthesis of 5‐[2‐ethoxy‐5‐(4‐methyl‐piperazine‐1‐sulfonyl)‐phenyl]‐1‐methyl‐3‐propyl‐1,6‐dihydro‐pyrazolo[4,3‐d]pyrimidine‐7‐thione salt with 3‐carboxy‐3‐hydroxy‐pentanedioic acid (KJH‐1002)

Citric acid was added to 26.4 mL of an acetone/ water solution (10:1), and the undissolved material was removed by filtration. Compound 5‐[2‐ethoxy‐5‐(4‐methyl‐piperazine‐1‐sulfonyl)‐phenyl]‐1‐methyl‐3‐propyl‐1,6‐dihydro‐pyrazolo[4,3‐d]pyrimidine‐7‐thione (5.74 g, 0.0117 mol) was dissolved in refluxing acetone (72.5 mL) and thionyl chloride (1.42 mL, 0.0195 mol), and undissolved material was filtered. These two solutions were mixed and refluxed for 1.5 h and then gradually cooled to room temperature. The resulting solid was filtered and vacuum‐dried to yield the required compound as a yellow‐white solid (7.75 g, 97.0%, KJH‐1002 (C₂₂H₃₆N₆O₃S₂)).

¹H‐NMR (300 MHz, DMSO‐d₆): δ 13.55 (bs, 1H), 7.94 (d, J = 2.4 Hz, 1H), 7.87 (dd, J = 8.8 Hz, J' = 2.4 Hz, 1H), 7.40 (d, J = 8.9 Hz, 1H), 4.45 (s, 3H), 4.24 (q, J = 6.9 Hz, 2H), 2.97 (bs, 4H) 2.82 (t, J = 7.4 Hz, 2H), 2.67 (ABq, J = 15.5 Hz, 4H), 2.55 (bs, 4H), 2.27 (s, 3H), 1.76 (m, 2H), 1.35 (t, J = 6.9 Hz, 3H), 0.94 (t, J = 7.4 Hz, 3H). ¹³C‐NMR (75 MHz, DMSO‐d₆): δ 175.87, 172.86, 172.13, 161.11, 148.53, 145.63, 134.28, 132.84, 132.68, 130.82, 127.13, 123.56, 114.28, 73.04, 65.97, 54.02, 46.11, 45.55, 43.86, 41.21, 40.02, 27.81, 22.42, 15.16, 14.65, IR (KBr pellet, cm⁻¹) 3282, 2958, 1722, 1572; mp: 179–180°C.

Assay to measure PDE5 activity inhibition by KJH‐1002

Inhibition of the activity of PDEs 1, 3, 4, 5, 6 was determined using a modified version of the method described by Thompson and Appleman (1971). Briefly, the reaction mixtures (total volume of 100 μL: PDE enzyme 3–10 μL, 10 nM CaCl₂, 20 μM calmodulin, 1 μCi·μL⁻¹ [³H] cGMP, 1 μCi·μL⁻¹ [³H] cAMP, 0.01 nM–1μM KJH‐1002, 50 mM Tris–HCl buffer (pH 7.4), 15 mM MgCl₂ and distilled water) were incubated in a water bath at 30°C for 30 min. After boiling for 2 min, ice‐cold water (0.5 mL) was added to the reaction mixture and the resulting solution was applied to a DEAE‐Sephacel A‐25 anion exchange column (approximately 1mL). Aqueous scintillation fluid (20 mL) was added to each eluate (30 μL) and the radioactivity (as [³H] guanosine) was measured by a β counter. The KJH‐1002 compound was used by diluting a stock solution in DMSO to a final concentration of 2% (v/v). The IC₅₀ values were calculated by plotting the remaining enzyme activity (% of control) against the log (compound concentration) and the point at which 50% activity was attained was determined by linear regression.

Morris water maze (MWM) test

After pretreating the mice for 7 days, spatial memory and learning was evaluated by the MWM test. The test was performed as described by Morris (1984). In brief, the apparatus consisted of a circular pool (120 cm in diameter, 60 cm in height) with a featureless inner surface and an automatic tracking system (Ethovision System, Noldus, Wageningen, The Netherlands). The circular pool was equally divided into four quadrants and filled with water (45 cm deep, 23 ± 1°C), and a white, non‐toxic dye (Blick Premium Grade Tempera White Gallon, 00011‐1009, BLICK, USA) was added to the pool to make the water opaque. A circular platform (8 cm in diameter) was placed in the centre of the target quadrant, and the platform was submerged 1 cm below the water surface to be invisible to the mice. A black curtain was arranged around the pool to isolate interference from irrelevant visual cues, and four cues of assorted colours and shapes were fixed on the curtain opposite each quadrant as the only way for the mice to discriminate their position.

The MWM test consisted of an acquisition section and a retention section, and the test was carried out for six consecutive days; the acquisition section was performed for the first 5 days. The mice were allowed to swim in the water while searching for the hidden platform for 90 s. If the mice found the hidden platform successfully, they were allowed to remain on the platform for at least 5 s. Mice that failed to locate the platform in 90 s were placed on the platform for at least 5 s and then returned to their cage by the operator. Each mouse was subjected to three trials per day at intervals of 30 min. For each trial, the mice began swimming from different position and faced the inner wall of the pool. The time the mice spent searching for the platform was recorded as the escape latency. For the mice unable to find the platform, a latency of 90 s was recorded. On the day after the acquisition section, all mice were subjected to a retention section in which the hidden platform was removed. The mice were allowed to swim in the pool for 120 s, and the time spent in the target quadrant where the platform had been fixed during the acquisition test was recorded. The mice were treated with saline (Control and Vehicle groups), donepezil (4 mg·kg⁻¹, p.o.) 1 h before the trial, and scopolamine was injected i.p., 30 min before the trial. For the KJH‐1002‐10 and KJH‐1002‐20 groups, 10 or 20 mg·kg⁻¹ KJH‐1002 was given p.o., 5 min before the trial.

Y‐maze test

The Y‐maze was conducted to test the spontaneous alternation behaviour as an estimation of short‐term memory. Briefly, the apparatus consisted of a three‐arm maze with equal angles. The identical arms were made of polyvinyl plastic, and the floors and walls were 30 cm in length, 5 cm in width and 12 cm in height. The mice were allowed to begin the test facing the terminal wall of one arm, and the mouse movements in the maze were recorded by a camera fixed 1.5 m above the apparatus for duration of 5 min. Arm entry was defined as a mouse placing all four paws in any arm completely, and the total number of arm entries and sequence of arm entry (e.g. ABC, BCA or CAB) were counted manually. The actual alternation behaviours were defined as entries in which the mice went into three consecutive arms in sequence, such as ABC, BCA or CAB, but not ABA, BCB or CAC. An ethanol solution (10%, v/v) was used to remove the residual odors between mice. One hour prior to the Y‐maze test, mice in Control, Vehicle and DON groups were given p.o., saline, saline or donepezil (4 mg·kg⁻¹) respectively. After 30 min, all mice except for those in Control group were injected i.p. with scopolamine (1 mg·kg⁻¹). Mice in the KJH‐1002‐10 and KJH‐1002‐20 groups were treated with 10 or 20 mg·kg⁻¹ KJH‐1002 (p.o.), respectively, 25 min after the scopolamine injection. Thirty minutes after the scopolamine injection, the Y‐maze test was carried out. The ratio of actual alternation behaviour, calculated as the % alternation = [(number of actual alternations)/(number of total arm entries‐2)] × 100, was used as an index to estimate short‐term memory. The total number of mouse arm entries was analysed to assess the locomotor activity.

Passive avoidance test

The passive avoidance test was conducted in an apparatus with two identical compartments (20 × 24 × 30 cm). One compartment was bright (illuminated by an 8 W lamp), and the other compartment was dark, and equipped with 2 mm (diameter) stainless steel rods delivering an electrical shock to the mice. The two compartments were separated by a wall with a guillotine door (5 × 5 cm). The passive avoidance test was performed during two consecutive days. And the acquisition trial took place on day one. The mice were placed in the illuminated compartment, and after exploring for 30 s, the guillotine was opened and the mice were allowed to freely enter the dark compartment. The door was closed once the mice entered the dark compartment with no hesitation, and an electrical shock (0.3 mA, 3 s) was immediately delivered to the mice via the steel rods. After the electrical stimulation, the mice were returned to their cages, and the apparatus was cleaned using a 10% ethanol solution to remove residual odours. The time from when the door opened to when the mice entered the dark compartment was recorded in seconds as the step‐through latency. One hour before the acquisition trial, mice in the Control, Vehicle and DON groups were given saline, saline or donepezil (4 mg·kg⁻¹ ) respectively (p.o.). Except for those in Control group, all mice were injected i.p. with scopolamine (1 mg·kg⁻¹) after 30 min. Mice in the KJH‐1002‐10 and KJH‐1002‐20 groups were treated with 10 or 20 mg·kg⁻¹ KJH‐1002 (p.o.), respectively, 25 min after the scopolamine injection. Thirty minutes after the scopolamine injection, the mice were allowed to explore in the passive avoidance task. Twenty‐four hours after the acquisition trial, the retention trial was carried out to measure the long‐term avoidance memory. The mice were placed in the illuminated compartment, the guillotine door was opened after 30 s, and the mice were then allowed to freely enter the dark compartment. The step‐through latencies were recorded with a maximum of time 180 s and used as an indicator of avoidance memory.

Novel object recognition (NOR) test

The NOR test is based on whether rodents are more interested in novel rather than familiar objects (Francis et al., 2012). The test was conducted during four consecutive days, divided into three phases: habituation (days one and two), introduction (day three) and recognition (day four). Briefly, a grey, polyvinyl plastic open‐field arena (50 × 50 × 50 cm) was fixed in a soundproof room. During the habituation phase, mice were placed in the arena in the absence of any objects and allowed to explore the environment for 5 min each day. During the introductory phase, mice were placed in the arena for an exploration period in the absence of any object. After 2 min, two identical objects (the same in texture, colour and size) were placed in two opposite corners of the arena 8 cm from the wall. The mice were allowed to explore the objects for 3 min. One hour before the NOR test, mice in the Control, Vehicle and DON groups were given p.o. saline, saline or donepezil (4 mg·kg⁻¹) respectively. Except for those in the Control group, all mice were injected i.p. with scopolamine (1 mg·kg⁻¹) after 30 min. Mice in the KJH‐1002‐10 and KJH‐1002‐20 groups were treated with 10 or 20 mg·kg⁻¹ KJH‐1002 (p.o.), respectively, 25 min after the scopolamine injection. The introductory phase trial was carried out 30 min after scopolamine. Twenty‐four hours after the introductory phase, one of the two old identical objects was replaced by a novel object (different in texture, colour and size) during the recognition phase. Exploratory behaviour was recorded when a mouse sniffed or touched the objects with its nose or paws. Time spent exploring the objects (both old and novel objects) was recorded using a stopwatch. To remove the residual odours, a 10% ethanol solution was used to clean the arena and the objects after each mouse. A discrimination index was calculated by the following formula: DI = [T_n/(T_n + T_f)] × 100 where T_n is the time spent exploring the novel object and T_f is the time spent exploring the familiar object.

Tissue preparation

The day after the behavioural tests, mice in the Control, Vehicle and DON 4 groups were given saline, saline or donepezil (4 mg·kg⁻¹), respectively (p.o.), and all mice except for those in the Control group were injected i.p. with scopolamine (1 mg·kg⁻¹) 30 min later. Mice in the KJH‐1002‐10 and KJH‐1002‐20 groups were treated with 10 or 20 mg·kg⁻¹ KJH‐1002 (p.o.), respectively, 25 min after the scopolamine injection. Thirty minutes after the scopolamine administration, mice were killed by cervical dislocation. The skulls were dissected, and the brain cortices and hippocampi were carefully collected and stored at −80°C before use. Next, the brain tissue (part of the cortex and all of the hippocampus) were homogenized in PRO‐PREPTM lysis buffer (iNtRON, Gyeonggi, Korea) supplemented with Phosphatase Inhibitor Cocktail Set I (Sigma‐Aldrich, MO, USA). The homogenates were stored at 4°C for 30 min and then centrifuged at 13 000 × g at 4°C for 30 min. The supernatants were collected, and the protein concentrations were assayed by the Bradford method with a standard curve constructed using BSA. A portion of the supernatants was diluted by adding loading buffer to a concentration of 1.5 mg·mL⁻¹ and then denatured by heating at 99°C for 5 min. The denatured and non‐denatured parts of the supernatants were stored at −80°C for later Western blot and enzyme activity analyses.

Determination of cGMP level in the cortex

An immunoassay kit (BioVision, Milpitas, CA, USA) was used to analyse the cGMP levels in the cortex, as reported by Fiorito et al. (2013). The remainder of the cortex was weighed and homogenized with five volumes of 0.1 M HCl. The homogenate was centrifuged at 13 000 × g for 5 min and the supernatant was collected. After the protein concentration was determined, the supernatants were used in the immunoassay, which was conducted according to the manufacturer's instructions. Each sample was assayed in duplicate and the OD at 450nm was inversely proportional to the concentration of cGMP, which is expressed as pmol·mg⁻¹ protein.

Western blot analysis

To measure the expression of marker proteins in the cortex and hippocampus, brain samples (20 μg) were separated using 12 or 8% SDS‐PAGE. The proteins were then transferred onto PVDF membranes using a Trans‐Blot Turbo transfer system (Bio‐Rad, USA) in transfer buffer. After transfer, the membranes were further incubated in blocking buffer (5% skim milk in Tris‐buffered saline with 0.1% Tween 20) for 1 h at room temperature and then incubated in rabbit monoclonal primary antibodies (anti‐BDNF antibody, anti‐p‐CREB antibody, anti‐CREB, anti‐p‐ERK, anti‐ERK, anti‐p‐Akt, anti‐Akt and anti‐GAPDH diluted in blocking buffer at a ratio of 1:1000) at 4°C overnight. After incubation with the primary antibodies, the membranes were washed three times with wash buffer (Tris‐buffered saline with 0.1% Tween 20) for 30 min and then incubated with HRP‐conjugated goat anti‐rabbit secondary antibodies (ratio of 1:3000) in blocking buffer for 1 h at room temperature. After being washing three times for 30 min, the protein bands were developed using an ECL detection kit according to standard procedure and visualized by the LAS‐4000 mini system (Fujifilm, Japan). The intensities of the protein bands were analysed using Multi Gauge V3.0 software (Fujifilm, Japan) and normalized according to the GAPDH intensity.

Analysis of AChE activity

AChE activity in the cortex and hippocampus were measured using the protocol described by Ellman et al. (1961). In brief, 40 μL of PBS (50 mM, pH 7.0) and 10 μL of supernatants (1.5 mg·kg⁻¹) were added to the wells of a 96‐well cell culture plate and mixed by gentle shaking. The reaction was started by adding a 50 μL mixture of acetylthiocholine iodide (AChI, 10 mM in 50 mM PBS, pH 7.0) and 5, 5′‐dithiobis (2‐nitrobenzoic acid) (DTNB, 10 mM in 50 mM PBS, pH 8.0). The absorbance at 412 nm was immediately measured using a microplate spectrophotometer (BioTek, VT, USA) with a kinetic mode (at an interval of 1 min) for 20 min. The absorbance value at 5 min and a molar extinction coefficient of 14.15 mM⁻¹·cm⁻¹ was employed to calculate the final AChE activity, expressed as nmol·min⁻¹·mg⁻¹ protein.

Measurement of the malondialdehyde (MDA) levels

MDA is a product of lipid peroxidation, and the MDA levels in the cortex and hippocampus were measured by the method provided by Esterbauer and Cheeseeman (1990) with slight modifications. Briefly, 240 μL of freshly prepared 0.5% phosphoric acid (pH 2.0), 80 μL of 0.6% thiobarbituric acid and 40 μL of prepared cortex or hippocampus homogenates were mixed in a 1.5 mL micro‐centrifuge tube. The mixture was heated at 95°C for 1 h, and after cooling, an additional 300 μL of n‐butanol was added and shaken vigorously. The mixture was centrifuged at 13 000 × g for 5 min, 200 μL of the n‐butanol layer (pink colour) was collected, and the absorbance at 532 nm was measured using a microplate spectrophotometer (BioTek) with an endpoint mode. A molar extinction coefficient of 156 mM⁻¹·cm⁻¹ was used to calculate the MDA level, which is expressed as nmol·mg⁻¹ protein.

Measurement of SOD activity

Measurement of SOD activity was mainly performed according to the method described by Marklund and Marklund (1974) with minor modifications. The method was based on the inhibitory effects of SOD on the autoxidation of pyrogallol in alkaline solution. Briefly, 700 μL of tris‐cacodylic solution (pH 8.5) containing 20 mM EDTA (St. Louis, MO, USA), 20 mM diethylenetriaminepentaacetic acid (St. Louis, MO, USA) and 200 μg·mL⁻¹ catalase was mixed with 200 μL of the cortex or hippocampus supernatant in the well of 48‐well cell culture plate. Autoxidation was initiated by adding 100 μL of 3 mM freshly prepared pyrogallol (in 10 mM HCl), the absorbance at 420 nm was recorded from 1.5 to 3.5 min within a kinetic mode (interval of 30 s). A standard SOD substance (1 mg equal to 5763 units, St. Louis, MO, USA) was employed to construct a standard curve. SOD activity in the cortex or the hippocampus was estimated from the standard curve and was expressed as U·mg⁻¹ protein.

Estimation of glutathione peroxidase (GPx) activity

GPx activity in the cortex and hippocampus was quantitatively analysed, according to the method described by Ellman (1959) and Hancianu et al. (2013) with slight modifications. Briefly, 20 μL of brain tissue homogenate was mixed with 20 μL of 2 mM GSH in PBS solution (0.2 M, pH 7.0) containing 2.5 mM NaN₃ and 0.2 mM EDTA. After incubating at 37°C for 5 min, 10 μL of 0.6 mM H₂O₂ was added and incubated at 37°C for 5 min. Then, an additional 100 μL of 5% trichloroacetic acid (in ultrapure water) was added and incubated for 10 min at room temperature. After centrifugation at 3000 × g for 10 min, the reaction was initiated by mixing 100 μL of supernatant and 35 μL of 4% DTNB (in ultrapure water). Absorbance of the yellow colour was immediately measured at 412 nm by a microplate spectrophotometer (BioTek) with an endpoint mode. A reaction system in which the homogenate was replaced with PBS solution (0.2 M, pH 7.0) served as a control. The GPx activity was calculated using a molar extinction coefficient of 14.15 mM⁻¹·cm⁻¹ and expressed as nmol·min⁻¹·mg⁻¹ protein.

Estimation of glutathione reductase (GR) activity

We determined the GR activity in the cortex and hippocampus according to the method provided by Ellman (1959). The reaction system contained 20 μL of assay buffer (0.2 M PBS containing 2.5 mM NaN₃ and 0.2 mM EDTA pH 7.0), 100 μL of GSSG (2 mM in assay buffer), 20 μL of brain tissue homogenate and 50 of μL DTNB (3 mM in assay buffer). The reaction was immediately initiated when 10 μL of NADPH (2 mM in assay buffer) was added. After 5 min, the absorbance at 412 nm was measured by microplate spectrophotometer. A molar extinction coefficient of 14.15 mM⁻¹·cm⁻¹ was used to calculate the GR activity, which is expressed as nmol·min⁻¹·mg⁻¹ protein.

Data and statistical analysis

The data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology (Curtis et al., 2018). All data are presented as mean ± SEM and were analysed using GraphPad Prism 7 software (GraphPad Software, CA, USA). The escape latency data from MWM test were analysed by two‐way ANOVA followed by a post hoc Tukey's multiple comparison test. All other data were analysed by one‐way ANOVA followed by post hoc Dunnett's multiple comparison tests. Post hoc tests were run only when F value achieved P < 0.05. Statistical significance was considered when a value of P < 0.05 was confirmed. The operator and data analysis were blinded.

Materials

Scopolamine hydrobromide, donepezil, 5,5′‐dithiobis (2‐nitrobenzoic acid), AChI, thiobarbituric acid, phosphoric acid, pyrogallol, catalase, SOD, GSH, oxidized GSH, NADP⁺ and anti‐BDNF antibody were purchased from Sigma‐Aldrich (St. Louis, MO). [³H] cAMP and [³H] cGMP were supplied by Amersham Life Science Inc. (Arlington Heights, IL). The anti‐CREB, anti‐phosphorylated CREB (p‐CREB), anti‐ERK, anti‐phosphorylated ERK (p‐ERK), anti‐Akt, anti‐phosphorylated Akt (p‐Akt) and anti‐GAPDH antibodies were purchased from Cell Signaling Technology (Danvers, MA,). All other reagents were of the highest grade and were purchased from commercial resources.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018) and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 (Alexander et al., 2017a, 2017b).

Results

Inhibition of PDE5 by KJH‐1002

The KJH‐1002 compound exhibited excellent inhibitory activity (Table 1) against PDE5 with an IC₅₀ value of about 60pM. It also inhibited PDE1, PDE3, PDE4 and PDE6 but with much lower potencies (Table 1). Based on these data, KJH‐1002 was considered to be a specific PDE5 inhibitor. Meanwhile, due to the limitation of the test, we cannot rule out the possibility that KJH‐1002 exhibits more potent inhibitory activity against other subtypes of PDE, not tested at this time.

Table 1.

Inhibitory potencies of KJH‐1002 against several PDE isoforms

Compound	IC50 values (nM)
	PDE1	PDE3	PDE4	PDE5	PDE6
KJH‐1002	2039 ± 598	1177 ± 99	920 ± 68	0.059 ± 0.04	47 ± 28

Effects of KJH‐1002 on spatial learning and memory in the MWM test

The effect of KJH‐1002 on scopolamine‐impaired long‐term spatial learning and memory was assessed by the MWM test. As shown in Figure 2A, mice in the control group rapidly learned and memorized the location of the platform in the target quadrant and exhibited the shortest escape latencies during the acquisition phase. The scopolamine‐treated (1 mg·kg⁻¹) group showed longer average escape latency, and significant differences were observed compared with the control group on day four and day five, indicating that an amnesia model was successfully induced by scopolamine injection. KJH‐1002 (20 mg·kg⁻¹) and donepezil (4 mg·kg⁻¹) significantly decreased the scopolamine‐prolonged escape latency from day four onward. The day after the acquisition phase, all mice were subjected to a probe trial by removing the platform. Mice in the scopolamine‐treated group spent a markedly shorter time swimming in the target quadrant than mice in the control group (Figure 2B). The shortened swimming time in the target quadrant was increased by donepezil or KJH‐1002 administration.

Figure 2.

Enhancing effect of KJH‐1002 on scopolamine‐induced cognitive dysfunction in different behaviour tests. (A) Spatial memory was analysed using an escape latency to find a platform for five consecutive days in the acquisition section and (B) time spent in the target quadrant on the day following the acquisition section (probe section with the absence of platform) in the MWM test. KJH‐1002 reversed the long‐term avoidance memory loss after scopolamine (SCOP), as determined by the passive avoidance test (C). Scopolamine‐impaired short‐term working memory was improved by KJH‐1002, as shown in (D) KJH‐1002 ameliorated scopolamine‐impaired spontaneous alternation behaviour, spontaneous alternation (%) = [(number of actual alternations)/(number of total arm entries‐2)] × 100, and (E) KJH‐1002 improved the locomotor activity by reducing the number of arm entries increased by scopolamine in the Y‐maze test. (F) KJH‐1002‐treated mice spent more time exploring the novel object [recognition index (%) = (exploring time on novel object/total exploring time for both objects) × 100], revealing that scopolamine‐impaired recognition memory was reversed by KJH‐1002. The results are expressed as the mean ± SEM (n = 8 for each group). ^* P < 0.05, significantly different from the control group; ^# P < 0.05, significantly different from the scopolamine‐treated group. DON, donezepil.

Effect of KJH‐1002 on scopolamine‐impaired spontaneous alternation behaviour in the Y‐maze test

The effect of KJH‐1002 on the impairment of short‐term spatial cognition induced by scopolamine was evaluated using the Y‐maze test. The ratio of spontaneous alternation behaviour in the Y‐maze test was used as an important index to assess the learning and memory abilities of the mice. Scopolamine injection resulted in significantly decreased ratio of spontaneous alternation behaviour, compared with that of the control group (Figure 2D). This scopolamine‐induced decrease was reversed by treatment with donepezil or KJH‐1002 (10 and 20 mg·kg⁻¹). Moreover, scopolamine injection also resulted in significant increase in the number of arm entries in comparison with control group (Figure 2E). Interestingly, KJH‐1002 (20 mg·kg⁻¹) markedly reversed the scopolamine‐increased number of arm entries, indicating improved locomotor activity.

Effect of KJH‐1002 on step‐through latency in the passive avoidance test

The passive avoidance test was conducted to assess the enhancing effect of KJH‐1002 on scopolamine‐impaired step‐through latency. As shown in Figure 2C, no significant differences in step‐through latency were found among any of the groups in the acquisition trial. However, in the retention trial, a significant decrease in step‐through latency was induced by scopolamine administration compared with that of the control group, which revealing that long‐term memory was impaired by scopolamine. Interestingly, the decrease in step‐through latency was also significantly improved by KJH‐1002 (20 mg·kg⁻¹) or donepezil (4 mg·kg⁻¹).

Effect of KJH‐1002 on scopolamine‐induced memory impairment in the NOR test

Recognition memory was evaluated using the NOR test in which the time mice spent exploring novel and familiar objects was monitored, and object recognition memory was assessed by a defined discrimination index. As shown in Figure 2F, mice in the control group clearly preferred the novel object to the familiar one, while scopolamine significantly reduced the time mice spent exploring the novel object, as a marked decline in the discrimination index was observed. Pretreatment with donepezil or KJH‐1002 (at both 10 and 20 mg·kg⁻¹) significantly improved the discrimination index, demonstrating that scopolamine‐impaired recognition memory was reversed by KJH‐1002 administration.

The effect of KJH‐1002 on cGMP levels in the cortex

The effect of KJH‐1002 on the cGMP level in the cortex was measured using a direct competitive immunoassay. As shown in Figure 3, scopolamine administration did not affect cGMP levels in the cortex, compared with those in the control group. Similarly, no significant increase in cGMP levels was found in the donepezil‐treated group in comparison with the scopolamine or control groups. Interestingly, KJH‐1002 markedly increased the cGMP levels in the cortex, and the high dose (20 mg·kg⁻¹ ), especially, significantly increased cGMP levels.

Figure 3.

KJH‐1002 increased the cGMP level in the cortex. No significant difference in the cGMP levels was found between the control and scopolamine (SCOP)‐treated groups. Donepezil (DON) failed to increase the cGMP level. KJH‐1002 (20 mg·kg⁻¹) significantly increased cGMP levels in the cortex. The results are expressed as the mean ± SEM (n = 6 for each group). *P < 0.05, significantly different from control group; ^# P < 0.05, significantly different from scopolamine‐treated group.

Effect of KJH‐1002 on the expression of memory‐related proteins in the cortex and hippocampus

To elucidate the molecular mechanism underlying how KJH‐1002 exerts cognition‐enhancing effects, we investigated the phosphorylation levels of CREB, ERK 1/2 and Akt, as well as BDNF expression in the cortex and hippocampus. Figure 4A shows the representative immunoblots in the cortex, and Figure 4C shows that scopolamine significantly reduced the cortical phosphorylation levels of CREB, ERK 1/2 and Akt, and the BDNF level compared with those in the control group. Pretreatment with donepezil, the positive control, significantly reversed the phosphorylation levels of CREB and Akt, as well as BDNF expression in the cortex. KJH‐1002 (20 mg·kg⁻¹) administration significantly enhanced the phosphorylation of CREB, ERK and the expression of BDNF in the cortex. Interestingly, KJH‐1002 at doses of 10 and 20 mg·kg⁻¹ significantly enhanced the cortical phosphorylation levels of Akt. Similarly, Figure 4B, D shows that the phosphorylation levels of CREB, ERK 1/2 and Akt, as well as the expression of BDNF in the hippocampus were significantly decreased by scopolamine administration in comparison with the control group. The decreased phosphorylation of CREB, ERK 1/2 and Akt, as well as the expression of BDNF were significantly reversed by donepezil administration. Both 10 and 20 mg·kg⁻¹ of KJH‐1002 significantly reversed the phosphorylation levels of CREB and ERK 1/2 in the hippocampus. Significant enhancing effects on Akt phosphorylation and BDNF expression in the hippocampus were observed in the group pretreated with 20 mg·kg⁻¹ KJH‐1002.

Figure 4.

KJH‐1002 modulated the expression of memory‐related proteins in the cortex and hippocampus. Representative Western blot bands of p‐CREB, BDNF, p‐ERK and p‐Akt in the cortex and in hippocampus are shown in (A) and (B) respectively. (C) and (D) exhibit the relative density of p‐CREB, BDNF, p‐ERK and p‐Akt protein expression in the cortex and hippocampus respectively (n = 6 for each group). The results are expressed as the mean ± SEM. *P < 0.05, significantly different from control group; ^# P < 0.05, significantly different from the scopolamine‐treated group. DON, donezepil; SCOP, scopolamine.

Effect of KJH‐1002 on inhibiting the activity of AChE

Scopolamine is a competitive muscarinic receptor antagonist, which can block the neurotransmitter ACh from binding the muscarinic receptor. To investigate whether KJH‐1002 exerted its cognition‐enhancing effects by inhibiting AChE activity, we measured the AChE activity in the brain tissues. The cortical and hippocampal AChE activities were increased by scopolamine administration in comparison with the activity in the control group (Figure 5), and the scopolamine‐increased AChE activities in the cortex and hippocampus were markedly inhibited by donepezil administration. However, KJH‐1002 administration did not reduce the AChE activity increased by scopolamine injection in the cortex or hippocampus, suggesting that KJH‐1002 did not exert its protective effect on cognitive function by inhibiting AChE activity.

Figure 5.

Effect of KJH‐1002 on scopolamine‐induced increase in AChE activity in the cortex and hippocampus. KJH‐1002 failed to reduce AChE activity increased by scopolamine. The results are expressed as the mean ± SEM (n = 6 for each group). *P < 0.05, significantly different from control group; ^# P < 0.05, significantly different from scopolamine‐treated group. DON, donezepil; SCOP, scopolamine.

Effect of KJH‐1002 on oxidative stress

Lovell et al. (1995) and Markesbery (1997) emphasized that increased oxidative stress is clearly involved in neuron degeneration and death in AD. Our experiments assessed the levels of MDA, an important marker of lipid peroxidation, and the activity levels of SOD, GPx and GR in the cortex and hippocampus. The cortical and hippocampal MDA levels were significantly increased by scopolamine, compared with the levels in the control group (Figure 6A). Administration of donepezil (4 mg·kg⁻¹) and KJH‐1002 at doses of 10 and 20 mg·kg⁻¹ significantly decreased the MDA level in the cortex. In the hippocampus, donepezil and KJH‐1002 (20 mg·kg⁻¹) significantly affected the decreased MDA level induced by scopolamine.

Figure 6.

The effect of KJH‐1002 on scopolamine‐induced oxidative stress in the cortex and hippocampus. KJH‐1002 decreased the level of MDA (A) and increased the activities of SOD (B), GPx (C) and GR (D) affected by scopolamine in the cortex and hippocampus respectively. The results are expressed as the mean ± SEM (n = 6 for each group). *P < 0.05, significantly different from control group; ^# P < 0.05, significantly different from scopolamine‐treated group. DON, donezepil; SCOP, scopolamine.

As shown in Figure 6B, SOD activity was decreased by scopolamine, compared with the levels in the control group in both the cortex and hippocampus. Additionally, this decreased SOD activity was restored by administration of donepezil (4 mg·kg⁻¹) or KJH‐1002 at either dose. Similarly the GPx activity in the cortex and hippocampus was decreased by scopolamine, compared with the control group (Figure 6C). This effect of scopolamine was reversed by treatment with donepezil (4 mg·kg⁻¹) or KJH‐1002 (10 and 20 mg·kg⁻¹).

The effects of scopolamine and KJH‐1002 on the GR activity in cortex and hippocampus are shown in Figure 6D. Significant decreases in cortical and hippocampal GR activity levels were found in the scopolamine‐treated group compared with the levels in the control group. Donepezil enhanced the GR activity reduced by scopolamine in both the cortex and the hippocampus. Moreover, the high dose of KJH‐1002 (20 mg·kg⁻¹) increased GR activity in the cortex, while KJH‐1002 at doses of 10 and 20 mg·kg⁻¹ increased the GR activity in the hippocampus, compared with the activity in the scopolamine‐treated group.

Discussion

The cGMP specific enzyme, PDE5 was identified as a potential target for the treatment of neurological diseases (Garcia‐Osta et al., 2012) and PDE5 has been found in human brain Teich et al. (2016). In the present study, we synthesized an analogue of sildenafil, KJH‐1002, that was a highly potent inhibitor of PDE5 (IC₅₀ about 60 pmol·L⁻¹) and a much less potent inhibitor of PDE1, PDE3, PDE4 and PDE6 (IC₅₀ values from about 50 – 2000 nmol·L⁻¹). Because the PDE5 inhibitor sildenafil enhanced cognitive function in a variety of animal models, we tested the cognition‐enhancing effect of KJH‐1002 in a mouse model of amnesia induced by scopolamine and found it to reverse several aspects of learning and memory, in a range of behavioural tests. Biochemical analysis indicated that KJH‐1002 may improve learning and memory by activating the cGMP/CREB signalling pathway and inhibiting oxidative stress, without influencing AChE activity. Another isoform, PDE2, hydrolyses both cAMP and cGMP, PDE9 is a cGMP‐specific hydrolyser, and inhibition of these two subtypes of PDE leads to the accumulation of cGMP and improves the memory performance in a variety of animal model (Boess et al., 2004; Kleiman et al., 2012). Due to the limitation of the present study, the inhibitory activity of KJH‐1002 against PDE2 and PDE9 remains unknown. We cannot rule out the possibility that the beneficial effects of KJH‐1002 in our scopolamine‐impaired animal model could reflect also its inhibitory activity against these two PDE subtypes.

Scopolamine is known to impair cognitive function in healthy humans and animals via its effects on cholinergic neurotransmission. Administration of scopolamine affects a variety of learning and memory, and locomotor activity, thus induces abnormal behavioural processes (Klinkenberg and Blokland, 2010). After pretreatment with KJH‐1002, all mice were subjected to the MWM test. Scopolamine markedly increased the escape latency for finding the hidden platform, and the higher dose of KJH‐1002 (20 mg·kg⁻¹ ) and donezepil (4 mg·kg⁻¹) reversed the spatial learning and memory impaired by scopolamine to normal, control levels. Our results are consistent with those of Cuadrado‐Tejedor et al. (2011), who reported a cognition‐enhancing effect of sildenafil in an AD model of transgenic mouse, in the MWM test. Additionally, sildenafil attenuated scopolamine‐induced learning and memory impairment in a 14‐unit T‐maze test (Devan et al., 2004). By contrast, Prickaerts et al. (2004) reported that a different PDE5 inhibitor, zaprinast (10 mg·kg⁻¹), failed to enhance the performance of adult rats in MWM test. This difference may be attributed to the animal model, as Prickaerts et al. (2004) used an unimpaired rat model with optimal memory performance, and it is difficult to enhance memory performance at a high level. In contrast, in our mouse model, cognitive functions were significantly impaired by scopolamine and KJH‐1002 restored cognitive functions from a low level. KJH‐1002 may also exhibit some inhibitory effects against PDE4 (IC₅₀ = 920nmol·L⁻¹) which could be relevant, as another PDE4 inhibitor, http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=5260 (3 mg·kg⁻¹ ) reportedly improved learning and memory in a microsphere embolism‐induced cerebral ischaemia rat in the MWM test (Nagakura et al., 2002).

The Y‐maze test is used to assess short‐term memory by analysing spontaneous alternation behaviour (Kwon et al., 2013). Scopolamine significantly damaged the spontaneous alternation behavior, compared with that of the control group. KJH‐1002 (10 and 20 mg·kg⁻¹) or donepezil (4 mg·kg⁻¹) restored this effect of scopolamine to the normal level. Moreover, KJH‐1002 decreased the number of arm entries increased by scopolamine, revealing that KJH‐1002 reduces the scopolamine‐induced increase of locomotor activity (Klinkenberg and Blokland, 2010). Numerous studies have shown that PDE5 inhibitors can enhance long‐term memory in the passive avoidance test. Sildenafil was reported to enhance memory in unimpaired young and aged mouse (Patil et al., 2004), as well as reversing memory performance in streptozotocin‐induced diabetic rats and electroconvulsive shock‐stimulated rats in the passive avoidance test (Patil et al., 2006). Our study showed that the PDE5 inhibitor KJH‐1002 also protected against scopolamine‐induced cognitive dysfunction in passive avoidance. Mice treated with KJH‐1002 remembered the electric shock delivered during the acquisition trial and exhibited longer step‐through latency in the retention phase. This finding is consistent with the studies by Van der Staay et al. (2008) showing that the inhibitor of PDE9 (also a cGMP‐specific hydrolyser), BAY 73‐6691, ameliorated long‐term memory impairment induced by scopolamine in the passive avoidance test. In the NOR test of the present study, KJH‐1002‐treated mice preferred the novel object, as they spent more time exploring the new object compared with the mice administered only scopolamine, demonstrating that KJH‐1002 also improves the object recognition memory deficit induced by scopolamine. This finding is consistent with results of van der Staay et al. (2008) and Hutson et al. (2011), who found that the PDE9 inhibitors, BAY 73‐6691 and PF‐04447943, significantly enhanced the memory performance in unimpaired and scopolamine‐impaired rats in the NOR test respectively.

Scopolamine is well‐ known to impair learning and memory in various types of behaviour tests by not only blocking the cholinergic system but also reducing expression of CREB and BDNF (Shi et al., 2013; Hong et al., 2014; Jeon et al., 2017). The transcription factor CREB serves as a molecular switch in learning and memory. CREB phosphorylation plays an important role in activating transcription, thus leads to the expression of downstream genes, such as BDNF, which is an important protein for synaptic plasticity and memory performance (Huang and Reichardt, 2001; Scott Bitner, 2012). Several studies have suggested that CREB and BDNF have therapeutic potential in AD patients (Bejar et al., 1999; Nagahara et al., 2009). PDE5 inhibition activates CREB, leading to transcriptional promotion of the target BDNF gene and thereby promoting synaptic plasticity and memory formation (Cuadrado‐Tejedor et al., 2011; Garcia‐Osta et al., 2012). We found a significant decrease in p‐CREB and BDNF expression in the cortex and hippocampus after scopolamine administration, and subsequent KJH‐1002 administration could rescue expression of these proteins. Thus, the enhancing effect of KJH‐1002 on cognitive function may be attributed to the recovery of p‐CREB and BDNF. Moreover, BDNF is the upstream molecular effector of phosphorylated ERK 1/2 responsible for establishing synaptic activity and enhancing different types of memory (Revest et al., 2014). Ota et al. (2008) suggested an involvement of the lateral amygdala ERK/MAPK in the NO‐cGMP‐PKG signalling pathway in modulating memory consolidation. We observed that KJH‐1002 could also significantly reverse the cortical and hippocampal expression of p‐ERK after the recovery of p‐CREB and BDNF, which suggests an involvement of activated ERK in the promotion of cognition function. Akt, an enhancer of neuronal survival, serves as a signal transducer of neurotrophin‐regulated survival and is important for the prevention of various neurodegenerative challenges (Calabrese et al., 2007). Interestingly, we found that KJH‐1002 counteracted the effect of scopolamine and induced a marked increase in Akt phosphorylation supporting an important role of Akt in reducing scopolamine‐induced cognitive dysfunction. Western blot analysis showed that activation of the CREB signalling pathway by KJH‐1002 plays a key role in reversing cognitive impairments induced by scopolamine.

Several studies have suggested that PDE5 inhibitors elevate cGMP levels in various brain regions, such as the cortex, hippocampus and cerebellum, in transgenic AD mice or unimpaired rats, and cGMP elevation promotes CREB phosphorylation, thus improving memory performance in different types of behaviour tests (Marte et al., 2008; Puzzo et al., 2009; Jin et al., 2014). In this study, an ELISA kit was used to determine cGMP level in the cortex. Scopolamine and donepezil did not affect cGMP levels in the cortex, while the PDE5 inhibitor KJH‐1002 increased the cortical cGMP levels to a very high level. Thus, the effects of KJH‐1002 on cognitive dysfunction is dependent upon the enhancement of cGMP levels.

Scopolamine‐induced amnesia is also characterized by an increased activity of AChE, which interferes with learning and memory by hydrolysing the neurotransmitter ACh into choline and acetic acid (Gautam et al., 2016; Kulshreshtha and Piplani, 2016; Nedaei et al., 2016). In the present study, KJH‐1002 did not inhibit the increased AChE activity caused by scopolamine, indicating that cognitive function enhancement by KJH‐1002 is not dependent on inhibition of AChE activity.

Much evidence has demonstrated that oxidative stress is involved in AD (Gella and Durany, 2009). A study by Sikandaner et al. (2017) demonstrated that sildenafil administration decreased oxidative stress and enhanced memory performance in mice impaired by chronic noise exposure. In the present study, KJH‐1002 exerted anti‐oxidative effects on scopolamine‐impaired mice by significantly decreasing the MDA level and increasing enzyme activities of SOD, GPx and GR in the cortex and hippocampus. Thus, some part of the effects of KJH‐1002 on cognition enhancement could be attributed to its potency in preventing oxidative stress.

In conclusion, both the AChE inhibitor donepezil and the PDE5 inhibitor KJH‐1002 ameliorated the decline of learning and memory induced by scopolamine . It appears that both of them enhance cognitive function via modulating memory‐related proteins and attenuating oxidative stress damage, in a similar manner. However, the enhancing effect of donepezil on scopolamine‐impaired cognitive function depends on the inhibition of AChE, while the PDE5 inhibitor KJH‐1002 exerts its enhancing effects, through increasing cGMP levels.

Author contributions

G.N. and J.H.S. designed and synthesized the KJH‐1002 compound. H.O.Y. designed the in vivo experiments, and L.Z. performed the animal behaviour tests. H.L. analysed the behaviour results. L.Z. performed the assays for measurements of cGMP, MDA level and enzymes activity, H.L. analysis these results. H.L. carried out the Western blot assay and L.Z. analysed these results. L.Z. wrote the manuscript.

Conflicts of interest

The authors declare no conflicts of interest.

Declaration of transparency and scientific rigour

This http://onlinelibrary.wiley.com/doi/10.1111/bph.13405/abstract acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research recommended by funding agencies, publishers and other organisations engaged with supporting research.

Zhang, L. , Seo, J. H. , Li, H. , Nam, G. , and Yang, H. O. (2018) The phosphodiesterase 5 inhibitor, KJH‐1002, reverses a mouse model of amnesia by activating a cGMP/cAMP response element binding protein pathway and decreasing oxidative damage. British Journal of Pharmacology, 175: 3347–3360. 10.1111/bph.14377.