Introduction

Major depressive disorder (MDD) is the most predominant chronic mental health disorder, and causes significant burden on quality of life. It is a leading cause of disability worldwide and poses huge economic burden (James et al. 2018). It is characterised by debilitating symptoms including low self-esteem, changes in mood, reduced interest in pleasure and an overall negative affect (APA 2015). One out of sixteen people worldwide experiences depression yearly (Otte et al. 2016). Furthermore, major depressive disorder comes with an elevated risk of other burdensome diseases such as stroke, cardiovascular disorders and diabetes mellitus (Whooley and Wong 2013).

Pharmacological management of MDD is mainly with antidepressants. Although significant strides have been made in the clinical management of depression, there is still an unmet need for new antidepressants. This has been due to reduced efficacies of current antidepressants which in some cases had only modest effects compared to placebos (Cipriani et al. 2018). Also limiting current therapy is the slow onset of the first-line drugs in addition to clusters of patients who are completely unresponsive to the current antidepressants (Cipriani et al. 2018; Bear et al. 2019). These gaps in the management of depression make the search from novel therapeutics still relevant.

Xylopic acid (15-(acetyloxy) kaur-16-en-18-oic acid) is a kaurene diterpene previously isolated from plants such as Xylopia frutescens, Xylopia sericeae and Xylopia aethiopica, where it has been shown to be one of the major secondary metabolites (Takahashi et al. 1995; Cavalcanti et al. 2010). Some of the plants from which it has been isolated have traditionally been used in the management of brain disorders or have compounds isolated from them with CNS effects (Biney et al. 2016). We have previously shown the anxiolytic-like effects of xylopic acid in mouse and zebrafish protocols (Biney et al. 2018) as well as its effects in the CNS core battery test (Biney et al. 2014). Anxiety and depression are usually co-morbid conditions with some antidepressants being used in managing anxiety and vice versa (Strawn et al. 2018). Thus, we hypothesise that xylopic acid may possess antidepressant effects. With a significant role of neuroinflammation in several CNS disorders including MDD (Rossi et al. 2017; Woelfer et al. 2019) and the reported anti-inflammatory property of XA (Osafo et al. 2016; Ekuadzi et al. 2018), it is possible that xylopic acid will reduce neuroinflammation and thus ameliorate depressive symptoms.

Materials and methods

Animals

Mice (ICR, male, 20–25 g) were obtained from Noguchi Memorial Institute for Medical Research (NMIMR) (Accra, Ghana), housed at the vivarium (25 °C, 12/12-h light cycle, 75% humidity) of the School of Biological Sciences, University of Cape Coast, and were used for the study after an 8-day acclimatisation period. They were grouped 10 mice in a cage and allowed ad libitum access to mice chow (Agricare, Kumasi, Ghana) and tap water. Four to six (4–6)-month-old zebrafish (Danio rerio) (adult short fin wild-type) supplied by Aqua Marshall (Accra, Ghana) and kept in 30-L glass tanks (5 fishes/L) were used. They were maintained at 25 °C, 12/12-h light cycle and 30% constantly aerated water replaced daily (pH 7–8). They were fed 2 times a day with commercial flakes (Aquafin Professional, Guangzhou, China). Experimental procedures were subject to NIH guidelines for the Care and Use of Laboratory Animals, EU Directive 2010/63, for zebrafish with ethics approval from the Institutional Review Board of UCC (UCCIRB/CoHAS/17/073). All behavioural experiments were observed by experienced experimenters blinded to the various treatment groups.

Drugs and chemicals

Xylopic acid (structure, Fig. 1a) was isolated and purified by reflux recrystallisation from unripe fruits of Xylopia aethiopica as reported earlier (Biney et al. 2018). Its purity was confirmed by LC-MS to be 99.8% w/w. It was formulated as an emulsion by emulsifying using 1% Cremophor EL before administration to mice. Fluoxetine and methysergide (Eli Lilly and Co., England), l-NAME, p-chlorophenylalanine and l-arginine (Sigma, Switzerland), Selegiline, serotonin, d-serine, tryptamine, desipramine, α-Methyl-p-tyrosine, d-cycloserine and Cremophor EL were purchased from Sigma, St. Louis, MO.

Fig. 1
figure 1

Structure of xylopic acid (a) and schedule of experiments used in assessing antidepressant effects of XA (b–e) in zebrafish and mice

Full size image

Drugs were administered per os via oral gavage and did not exceed 10 ml kg−1 (final volume 0.3–0.5 ml). Doses of drugs including XA, fluoxetine and desipramine were selected based on work published earlier.

Preliminary screening for antidepressant-like effect in zebrafish

Chronic unpredictable mild stress

We first used a zebrafish model for chronic unpredictable mild stress (CUS) to rapidly screen for potential antidepressant effect of XA using methods described by Piato et al. (2011) and Chakravarty et al. (2013) (Fig. 1b). Wild-type adult zebrafish were put through two stressors twice a day consecutively for 14 days. The stressors applied in a random manner included the following: congestion—10 fishes in 250 ml beaker containing 150 ml of water for 1 h; confinement—restraint for 15 min in narrow 5-ml tubes; dorsal body exposure—2 min exposure of the dorsal side of fishes by reducing tank water to a shallow level; social isolation—single housing for 60 min in 250-ml beakers; cold stress—reducing water temperature to 23 °C for 30 min; chasing—8 min pursuit with a harvesting net; heat stress—increasing tank water temperature to 33 °C for 30 min; repeated tank change—rapid movement of fishes from one tank to another repeatedly in 6 cycles. Twenty-four hours after the last stressor, they were placed in 8 groups (n = 9) and exposed to XA 1, 3, 10 μM or FLX 0.3, 3, 30 μM or tank water daily for 3 days by immersing them in respective drug solutions for 30 min. A naïve group did not undergo any of the stressors nor treatments. The shoaling, novel tank diving and light-dark tests were conducted 24 h after the final drug treatment as detailed below.

Novel tank test

In the novel tank test, done on day 18 of CUS paradigm, zebrafish were gently placed a new arena (15 (l) × 10 (w) × 25 (h) cm3) containing water (10 cm deep), segmented horizontally to form two 5-cm halves. The total duration in the upper half was videotaped and later computed with JWatcher™.

Scototaxis test

In this test, conducted after the NTT test, zebrafish were gentle placed water (25 °C) in a thin arena (9 (l) × 55 (w) × 10 (h) cm3), segmented vertically to form 2 zones of either white or black backgrounds. Time spent in the light region was videotaped for 5 min and later computed with JWatcher™.

Test for shoaling

In the shoaling behaviour analyses performed after the ST, three fishes were transferred into a rectangular tank (20 ×10 ×15 cm3) containing water (25–26 °C, depth = 10 cm) and allowing them to swim freely. The average area occupied by the 3 zebrafishes every 10 s was computed with ImageJ® for a total duration of 10 min.

Antidepressant effect in mice models

Forced swim test

The forced swim test by Porsolt et al. (1977) was used (Fig. 1c). Mice (n = 8) were treated p.o. with xylopic acid (XA) 10, 30 and 100 mg kg−1, fluoxetine (FLX) 3, 10 and 30 mg kg−1, desipramine (DES) 3, 10 and 30 mg kg−1 or vehicle 10 ml kg−1. Six-minute swim sessions were conducted in plastic cylinders (height = 25 cm radius = 6 cm) filled with water (23 ± 1 °C, depth = 15 cm), 2 h after XA or 1 h after FLX, DES and vehicle treatment. The total time of behaviours directed at escaping (climbing—vertically moving with forepaws on the wall of the cylinders and swimming—moving horizontally in the water) and immobility were recorded and quantified for the last 4 min of the test using JWatcher™ 1.0 and compared to the common control group (vehicle 10 ml kg−1).

Tail suspension test

Mice (n = 8) received p.o. same XA, FLX, DES and vehicle dosage regimen as in the FST. This was followed by the tail suspension test described by Steru et al. (1985). They were individually held by the tail onto a horizontal bar 52 cm high from the top of a laboratory bench with the aid of adhesive tape. Duration of immobility (no movements save those required for breathing) was videotaped for 6 min and computed with JWatcher™ and compared to the vehicle (10 ml kg−1) group.

Lipopolysaccharide-induced depression-like behaviour

The method described by Sulakhiya et al. (2016) was used to induce depressive-like behaviour using lipopolysaccharide (LPS 0111: B4, (Sigma, St. Louis, MO)) (Fig. 1d). Seven groups of mice (n = 10 per group), housed one mouse/plastic cage, were treated for 14 days with XA 10, 30, 100 mg kg−1, FLX 3, 10, 30 mg kg−1 or vehicle 10 ml kg−1 p.o. once daily. Twenty-four (24) hours after the last treatment, all mice were injected with 830 μg kg−1 LPS i.p. Subsequently, social interaction and sucrose preference tests as well as neurochemical assays of brain samples were assessed as described below.

Social interaction test

After 4 h following LPS injection, the social interaction test (Crestani et al. 1991) with minor modifications was conducted by introducing into the cages of the LPS-treated mice, a younger mouse of the same strain but which has had no prior contact with it. The total time for social exploration (pursuits, sniffing and social grooming) was videotaped for 5 min.

Sucrose preference test

A 2-bottle-model sucrose preference test described in Jangra et al. (2014) was used to evaluate anhedonia. The test includes a 7-day adaptation time in which mice have unlimited access to bottles: one containing 50-g tap water and the other 50 g of 2% sucrose solution to determine baseline sucrose consumption. The amount of each fluid drunk was computed as the change in bottles’ weight 24 h prior to and 24 h post-LPS injection on day 15. The preference for consuming sucrose was computed as percentage using the formula:

$$ \% sucrose\ preference=\frac{sucrose\ intake(g)}{total\ fluid\ intake(g)}\times 100\%. $$

Neurochemical assays

On day 16, 24 h post-LPS injection, mice were sacrificed by cervical dislocation. Whole brain samples were isolated and quickly kept at −80 °C for further assays. 10% w/v extracts were prepared by sonicating isolated brain samples in TNGT buffer (1% Triton X 100 (Sigma, St. Louis, MO), 150 mM Tris HCl (Sigma, St. Louis, MO), 10% glycerol, 150 mM NaCl, protease inhibitor cocktail (Sigma, St. Louis, MO), pH 7.4) and centrifuging the homogenate for 20 min at 6000g. Total protein of the resultant homogenate was assayed (Bradford 1976) and the levels of oxidative stress enzymes and BDNF determined in triplicates and expressed per mg protein.

Effect on reduced glutathione, catalase and superoxide dismutase

The concentration of reduced glutathione (GSH) was assayed by the method of Ellman (1959) using 100-μl supernatant of tissue extract. Absorbance of coloured product was read at 412 nm. Catalase activity was determined with a method described by Sinha (1972) in 100 μl of tissue extract. Absorbance of the final coloured product was read at read 620 nm. Superoxide dismutase (SOD) was assessed with procedures described by Sun et al. (1988) using 500 μl of tissue extract. The absorbance of final product was read at 560 nm and enzymatic activity quantified in unit of activity per weight of protein according to the formula:

$$ \% inhibition=\frac{Absorbance_{test}-{Absorbance}_{blank}}{Absorbance_{test}}\times 100\% $$
$$ \mathrm{Units}\ \mathrm{of}\ \mathrm{SOD}\ \mathrm{activity}/\mathrm{mg}\ \mathrm{protein}=\frac{\% inhibition}{50\times weight\ of\ protein\ (mg)} $$

Lipid peroxidation assessment

Malondialdehyde levels were assayed with a method described by Draper and Hadley (1990) using 1-ml tissue extract. Absorbance was read at 532 nm and 600 nm (nonspecific absorbance) and concentration of malondialdehyde then computed according to the formula:

$$ nmol\ malondialdehyde/ mg\ protein=\frac{Absorbance_{532\ nm-}{Absorbance}_{600\ nm}}{1.56\times {10}^5\times total\ protein(mg)}\times {10}^9 $$

Effect on BDNF

Whole brain samples were sonicated in cold extraction buffer after thawing to extract bound BDNF according to procedures by Kolbeck et al. (1999). The extraction buffer contained 3 portions of neutralising buffer (0.1 M Na2HPO4, 0.1 M KH2PO4 (Sigma, St. Louis, MO), final pH 7.6) and 1 portion acid-extraction buffer (1 M NaCl (BDH Poole, England), 0.1% Triton X 100, 50 mM sodium acetate (BDH Poole, England), protease inhibitor cocktail, final pH 4). The resultant mixture was centrifuged for 20 min (4 °C) at 6000g and the supernatant quantified for BDNF by ELISA as per manufacturer-prescribed protocols (Boster Biological Technology, Pleasanton, CA; Catalog # EK0309).

Open space swimming test

The repeated open space swimming test described by Stone and Lin (2011) was adapted and use to assess time-course of antidepressant effect (Fig. 1e). Each mouse (20–22 g) swam for 15 min each day for 4 consecutive days. This was carried out in rectangular plastic containers (25 × 35 × 50 cm3) filled with water (33 ± 1 °C) to a depth of 15 cm. On the 5th day, they were placed in 7 groups (n = 8) where they received p.o. daily for 14 days, either XA 3, 10, 30 mg kg−1 or FLX 3, 10, 30 mg kg−1 or vehicle 10 ml kg−1. Individual weights of mice were measured daily before to drug dosing. Swimming was repeated on the 5th, 7th, 10th, 14th and 18th days while being videotaped. Total time spent being immobile and the distance travelled were then computed with JWatcher™. A tail suspension test was repeated on day 19 to rule out if immobility had become a learned behaviour.

Effect on coat state

On the 17th day of the open space swimming test (OSST), a coat state assessment was performed as described by Yalcin et al. (2007). Each mouse was tenderly taken out of its cage for a visual examination of the dorsal and ventral coats, head, neck, genital area, front and hind paws, and tail. A clean well laid coat in each region was assigned 0 while an obviously altered, piloerected, unclean, messy or marred coat was assigned 1 by an experienced experimenter who was blinded to the various treatment groups. The cumulative score of the 8 body areas is the coat index.

Spontaneous novelty-induced grooming behaviour

Grooming behaviour was also measured on day 17 of the OSST. The mice were tenderly removed from their cages and individually put in Perspex observation arenas before grooming behaviour videotaped for 5 min. The mice were observed for the type of grooming they performed. This included passive grooming of the dorsal, ventral and genital regions as well as active grooming of the fore paws, head and neck (rostral grooming). The total time spent in fore paw grooming and head and neck washing (rostral grooming) was quantified by a blinded experienced experimenter.

Evaluation of possible mechanism(s) of action

Effects of depleting monoamines on antidepressant-like property of XA

The impact of serotonin and noradrenaline on the established antidepressant-like property of XA was evaluated by selectively inhibiting either their storage and/or synthesis (O’Leary et al. 2007). To exclusively deplete 5-HT, mice received 3 daily intraperitoneal pre-treatments of p-chlorophenylalanine (pCPA) 300 mg kg−1. Twenty-four hours post-pCPA injection, both naïve and treated mice, they were further treated with either vehicle or equipotent doses of XA (50 mg kg−1), FLX (10 mg kg−1) or DES (10 mg kg−1). One hour later, the FST was repeated.

Using a similar approach, noradrenaline and dopamine were selectively depleted with α-methyl-p-tyrosine (AMPT) 100 mg kg−1 i. p. and subsequently orally treated with equipotent doses of XA (50 mg kg−1), FLX (10 mg kg−1) or DES (10 mg kg−1), 4 h after which the FST was repeated.

Furthermore, 5–HT, NA and DA pools in both the cytoplasm and in vesicles were depleted after mice received intraperitoneal injection of 1 mg kg−1 reserpine. Eighteen hours later, animals received equipotent doses as described above, and subsequently, the FST was repeated.

Effect of XA on 5-HT receptors on a rat fundus strip

Male Sprague-Dawley rats were sacrificed to transversely cut out the stomach and isolate the fundus. From the fundus, longitudinal strips (20 × 2 mm) were removed and bathed in Krebs-Henseleit buffer: KCl (4.7), NaCl (118), MgSO4 .7H2O (1.2), CaCl2 (2.5), glucose (11.1) NaHCO3 (1.2) (mmol/L) (BDH Poole, England). The strips were isotonically mounted in 10-ml tissue baths (tension = 1 g, temperature = 37 °C, aeration = carbogen (95%)). Mounted tissues equilibrated for 1 h with regular tissue washing every 10 min. A total of 67-mM KCl was first used to generate the highest contractile response against which subsequent contractions were normalised. Cumulative concentration response tracings (CRTs) were recorded and measured with a Harvard kymograph after cumulative addition of either tryptamine, XA or 5-HT in the presence or absence of the 5HT1 and 5-HT2 non-selective antagonist methysergide (10 nM).

Interaction of XA with MAO enzyme

Rat fundus strips were isolated and mounted as described above. Initial tryptamine-provoked CRTs was recorded after cumulatively adding tryptamine (Sigma, St. Louis, MO) in ½ log increases. To examine the inhibitory actions of XA on monoamine oxidase enzyme, the CRTs of tryptamine were regenerated in the presence or absence of XA (1 μM) or selegiline (10 mM).

Effects of pre-treatment with d-cycloserine, d-serine, L-NAME and L-arginine

The role of NMDA glutamatergic modulation on XA’s effect was evaluated. Mice received either equipotent doses of XA, FLX or DES alone or in a combination with d-cycloserine (2.5 mg kg−1 i. p.) a glycineB partial agonist, or in a combination with d-serine (320 mg kg−1 i. p) a glycineB agonist. One hour later, the FST was repeated.

To evaluate the influence of nitric oxide pathways in XA’s antidepressant-like effects, mice received either an equipotent dose of XA, FLX or DES alone or in combination l-NAME 20 mg kg−1 i. p. administered 15 min earlier or in combination with l-arginine 750 mg kg−1 administered 30 min earlier. FST was repeated 1 h post-XA, FLX or DES treatment.

Data analyses

Data are reported as means ± SEM using GraphPad Prism version 6.0. Unless specified, data were compared using one-way analysis of variance with significance level of p < 0.05. F-statistic was computed using the Brown-Forsythe test. Data in the OSST and effects of pre-treatment with D-cycloserine, D-serine, L-NAME and L-arginine were compared by two-way analysis of variance with repeated measures (treatment × time). Data on swimming and climbing time in the FST was also analysed by 2-way ANOVA. When ANOVA was significant, Holm-Sidak test for significant difference was performed except in the LPS-induced depression behaviour tests where Kruskal-Wallis post hoc was used. Dose-response curves were fitted using iterative nonlinear regression (3-parameter logistic) equation:

$$ Y=\frac{a+\left(b-a\right)}{\left(1+{10}^{\left( Log{ED}_{50}-X\right)}\right)} $$

where the logarithm of dose and response are represented by X and Y, respectively, with Y assuming a sigmoid shape from a (bottom) to b (top). The fitted midpoints ED50s) were computed from the fitted midpoints of the curves and compared with the F test.

Results

Effects of XA on CUS in zebrafish

Antidepressant-like property was first screened in a CUS model in zebrafish to rapidly determine a potential antidepressant effect. The stress paradigm employed evoked depression-like behaviours after 14 days. This manifested as decreased time in the top half of the novel tank diving test, increased scototaxis and reduced shoaling in stressed zebrafish. Both XA and FLX reversed these behaviours. XA increased frequency and duration in the upper half (F 4, 25 = 35.4, p < 0.001) (Fig. 2a, b). Likewise, both XA and FLX treatments enhanced shoaling in zebrafishes (Fig. 2c). Also, the XA 10 μM increased the frequency and duration in the light region during the scototaxis test (p = 0.041) but this behaviour was not modified by FLX treatment (F 4, 20 = 1.683, p = 0.193) (Fig. 2d, e).

Fig. 2
figure 2

Effects of XA (1, 3, 10 μM) and FLX (0.3, 3, 30 μM) pre-treatments on duration and frequency in the upper segment of novel tank (a and b) shoaling behaviour (c) and scototaxis test (d and e) after 14-day CUS. Results show mean ± SEM *p < 0.05, **p < 0.01, ***p < 0.001, comparing to CUS control. 1-way ANOVA

Full size image

Forced swim test

Having exhibited an antidepressant-like potential in zebrafish, XA was further assessed in mice models. XA at 10, 30 and 100 mg kg−1 decreased immobility in a dose-dependent fashion in the FST (F3, 55 = 35.17, p < 0.001) similar to fluoxetine and desipramine (p < 0.001) (Fig. 3a). Comparing dose-response curves (DRCs), XA had similar efficacies in reducing immobility (ED50 = 8.40 ± 0.91; Emax = 76.43 ± 1.29) compared to FLX (ED50 = 2.37 ± 0.06; Emax = 78.65 ± 3.88) and DES (ED50 = 3.75 ± 0.52; Emax = 66.04 ± 2.28) (Fig. 3b). In Fig. 3c, XA and FLX increased total time spent swimming but not climbing whereas DES increased total time spent climbing.

Fig. 3
figure 3

Antidepressant-like effect of XA in the forced swim test in mice. a Effect of acute treatment of XA (10–100 mg kg−1), FLX (3–30 mg kg−1) and DES (3–30 mg kg−1) on duration of immobility during the forced swim test. Results indicate group means ±SEM (n = 8). One-way ANOVA and Holm-Sidak post hoc test. *p < 0.05, **p < 0.01 compared to ctrl. b DRCs showing % reduction of immobility by XA, FLX and DES in the FST. c Effect of XA, FLX and DES on swimming and climbing time in FST. Data represents group mean ± SEM. Two-way ANOVA, comparing within treatment groups: **p < 0.01

Full size image

Tail suspension test

In the TST, XA decreased immobility significantly at 100 mg kg−1 (p < 0.01) while DES and FLX reduced immobility at all doses tested (F3, 53 = 13.23, p < 0.001) (Fig. 4a). From the DRCs, the rank order of efficacy in deceasing immobility during the tail suspension test was FLX > XA > DES with ED50 and Emax values as follows: XA (ED50 = 19.25 ± 1.07; Emax = 71.57 ± 4.48), FLX (ED50 = 2.03 ± 0.14; Emax = 73.09 ± 1.16) and DES (ED50 = 4.64 ± 0.8; Emax = 68.04 ± 4.46) Fig. 4b.

Fig. 4
figure 4

Antidepressant-like property of XA in the TST in mice. a Effect of acute treatment of XA (10–100 mg kg−1), FLX (3–30 mg kg−1) and DES (3–30 mg kg−1) on duration of immobility during the tail suspension. Results indicate group means ±SEM (n = 8). One-way ANOVA *p < 0.05, **p < 0.01 comparing to ctrl. b DRCs showing % reduction of immobility by XA, FLX and DES in the TST

Full size image

LPS-induced depression-like behaviours

The influence of XA on the neuroinflammation hypothesis of depression was assessed using the LPS-induced depression-like behaviours model in mice. Administering 830 μg kg−1 lipopolysaccharide produced depression-like behaviours including decreased social interaction and % sucrose preference. XA 3, 10 and 30 mg kg−1 -treated animals exhibited dose-dependent sucrose preference (p < 0.01) (Fig. 4a) coupled with enhanced social interaction (p < 0.001) (Fig. 5c) as was also observed in fluoxetine-treated mice (p < 0.001). DRCs show that XA has higher efficacy than FLX in ameliorating anhedonia induced by lipopolysaccharide at the tested dose levels (Emax: XA: 78.23 ± 11.9; FLX, 61.71 ± 7.21) (Fig. 5b). A similar trend in efficacy is seen in the social interaction test (Emax: XA: 85.79 ± 6.8; FLX, 81.98 ± 4.05) (Fig. 5d). Overall, total fluid intake did not vary significantly across groups (Table 1).

Fig. 5
figure 5

Effect of XA on LPS-induced depression-like behaviours. Top panel a Impact of 14 days of treatment with XA (3, 10, 30 mg kg−1) or FLX (3, 10, 30 mg kg−1) in the anhedonia test. Data show group means ±SEM (n = 10). Kruskal-Wallis test * p < 0.05, ** p < 0.01 comparing to ctrl. b DRCs showing % decrease in anhedonia by XA and FLX. Bottom panel c Influence of XA or FLX in social interaction test. Data show group means ±SEM (n = 10). One-way ANOVA, ** p < 0.01, *** p < 0.001 comparing to ctrl. d DRCs of % increase in social interaction by XA and FLX

Full size image
Table 1 Total fluid intake of mice during the sucrose preference test
Full size table

Neurochemical assays

Reduced glutathione, superoxide dismutase and lipid peroxidation assays

Glutathione, SOD and malondialdehyde were assayed during the LPS-induced depression test to assess influence of XA on these markers of oxidative stress. Administration of LPS-reduced significantly brain glutathione levels of untreated mice. This reduction was overcame in a dose-dependent manner by both XA and FLX treatments (F7,32 = 0.75, p < 0.01) (Fig. 6a). Activity of SOD increased with XA and FLX treatments (F7,32 = 1.00, p < 0.01). Similar observations were made with respect to the effect of XA and FLX on catalase (F7,32 = 1.23, p < 0.01) (Fig. 6b, c). Lipid peroxidation levels in XA- and FLX-treated mice were reduced as indicated by reduction in brain concentrations of malondialdehyde (F7,32 = 1.01, p < 0.01) (Fig. 6d).

Fig. 6
figure 6

Influence of XA (3, 10, 30 mg kg−1) and FLX (3, 10, 30 mg kg−1) on (a) reduced glutathione (GSH), (b) catalase, (c) superoxide dismutase, (d) lipid peroxidation and (e) brain-derived neurotropic factor in LPS-challenged mice. f Dose-response curves indicating % rise in BDNF concentrations. Results mean ± SEM (n = 3). 1-way ANOVA. * p < 0.05, ** p < 0.01, *** p < 0.001, comparing to LPS control and †† p < 0.01, ††† p < 0.001, comparing to LPS-naïve mice

Full size image

Brain-derived neurotropic factor assay

Whole brain levels of BDNF measured after the LPS challenge indicated that XA 3, 10 and 30 mg kg−1 increased brain-derived neurotropic factor levels in a dose-dependent manner (F7,32 = 21.57, p < 0.001) (Fig. 6e) and was more potent and efficacious than fluoxetine at tested doses (XA ED50 = 1.72 mg kg−1, Emax = 93.92%) (FLX ED50 = 4.25 mg kg−1, Emax = 77.10%) (Fig. 6f).

Repeated open space swimming test

The OSST was conducted to establish the time-course of antidepressant effect. Repeated swim sessions caused increased total immobility time, a behaviour which remained elevated for the rest the test (Fig. 7a, c). Treatment with XA (3, 10, 30 mg kg−1) significantly (F3, 28 = 42.85, p < 0.001) reduced immobility on day 1 of treatment (i.e. day 5 of OSST) and persistently reduced to reach the pre-induction immobility time at the end of the OSST (Fig. 7a). Significant reduction in immobility was only observed from day 10 in the fluoxetine-treated mice (F 3, 28 = 14.19, p < 0.001) (Fig. 7c). Both XA and FLX produced significant reduction in overall immobility (Fig. 7b, d). The antidepressant-like behaviour exhibited in the repeated swimming test remained even 24 h post-drug treatment when measured in the TST (Fig. 7e). The various drug treatment however did not have any significant effect (p = 0.1278) on weight of mice at the end of the experiment.

Fig. 7
figure 7

Effects of xylopic acid in the open space swimming test for antidepressants. a–d Effect of XA (3–30 mg kg−1) and FLX (3–30 mg kg−1) on immobility time in the OSST. Results indicate time-course curves (TCCs) (a, c) and area under the TCCs as whisker plots (5–95th percentile) (b, d). ***p < 0.001, **p < 0.01 and *p < 0.05 comparing to saline (2-way ANOVA for TCCs and 1-way ANOVA for AUC’s). e Effect of XA (3–30 mg kg−1) and FLX (3–30 mg kg−1) in the TST 24 h post-OSST. ***p < 0.001, compared to saline. f Effect of XA (3–30 mg kg−1) and FLX (3–30 mg kg−1) on rostral grooming in the novelty-induced grooming behaviour test conducted on day 17 of OSST. Data presented as mean ± SEM. ***p < 0.001 comparing to saline. g Effects of XA (3–30 mg kg−1), and FLX (3–30 mg kg−1) on coat index on day 17 of the OSST. Data show mean scores ± SEM of eight body segments (well-mannered coat = 0, unkempt coat = 1). **p < 0.01, ***p < 0.001 comparing to saline

Full size image

Effect on novelty-induced grooming and coat index

In comparison to untreated mice, XA- and FLX-treated mice showed significant (F3, 27 = 7.125 p < 0.001) drop in total time spent in rostral grooming (grooming of head and fore paws) when in a new environment (Fig. 7f). Furthermore, the rigorous OSST protocol reduced overall physical appearance of mice. However, both XA and FLX treatments reversed the decline in physical outlook indicated by reduced coat index at 10 and 30 mg kg−1 (F3, 28 = 12.9, p < 0.001) for XA and at all doses for FLX (Fig. 7g).

Evaluation of possible mechanism(s) of action

Effects of depleting monoamines on antidepressant-like property of XA

Reversal of antidepressant-like behaviour of XA, FLX and DES was observed in reserpine pre-treated mice in the forced swim test (Fig. 8a). Selective depletion of catecholamines with AMPT did not reverse the ability of XA and FLX to reduce immobility (F4, 60 = 2.731, p = 0.072) (Fig. 8b) but was rather reversed in pCPA pre-treated mice (F1, 60 = 32.65, p < 0.001) (Fig. 8c). The opposite scenario was observed in mice that received DES.

Fig. 8
figure 8

Influence of reserpine (a), AMPT (b) and pCPA (c) on immobility time in XA (50 mg kg−1), FLX (10 mg kg−1) and DES (10 mg kg−1) treatments. Results indicate mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, vs. ctrl, 1–way ANOVA and comparison within groups ††p < 0.01, †††p < 0.001, 2-way ANOVA

Full size image

Contractile properties of XA on rat fundus strip preparation

The pharmacological properties of XA on 5-HT receptors were evaluated using the rat fundus tissue. XA evoked a partial agonist response on an isolated rat stomach fundus preparation in a concentration-dependent manner (EC50 = 45.88 ± 0.35 μM; Emax = 46.53) (Fig. 9a). Cumulative responses to XA were antagonised by 10 mM methysergide with a shift in the concentration response curve (CRC) to the right without depressing the Emax (p = 0.016) (Fig. 9b). Similar responses were observed with 5-HT (EC50 = 4.04 ± 0.16 nM) which was also abolished by methysergide (p < 0.001) (Fig. 9c). In the presence of a fixed sub-threshold concentration of XA (1 μM), the responses of 5-HT were potentiated (Fig. 9d).

Fig. 9
figure 9

Effects of XA on contractile properties of XA on rat fundus strip preparation. a XA and 5-HT evoked concentration-dependent contractions of the fundus (normalised response to 67 mM KCl). b–c Effect of methysergide (10 nM) on concentration response curves of xylopic acid and 5-HT. d Xylopic acid potentiated CRC of 5-HT. e–f Potentiation of tryptamine-induced response by selegiline (e) and not XA (f). Data points indicate mean ± SEM (n = 4–6)

Full size image

Involvement of monoamine oxidase enzyme

In the presence of a sub-threshold concentration of XA (1 μM), the contractile response of tryptamine was not potentiated (F1, 89 = 0.755, p = 0.382) (Fig. 9f). In contrast, selegiline (10 nM) significantly (F1, 91 = 13.95, p = 0.0023) potentiated responses produced by tryptamine causing a leftward shift with no significant change in Emax (Fig. 9e).

Effects of pre-treatment with d-cycloserine, d-serine, L-NAME and L-arginine

Mice that received a combination of d-cycloserine and XA or FLX exhibited potentiated antidepressant-like effects when compared to mice that received only XA or FLX. d-cycloserine pre-treatment did not affect DES-treated animals (Fig. 10a). The established antidepressant-like property of XA, FLX and DES was absent in d-serine-pre-treated animals (Fig. 10b).

Fig. 10
figure 10

Antidepressant-like property of XA is enhanced by d-cycloserine and L-NAME but not d-serine and L-arginine. XA, FLX and DES administered alone or together with (a) d-cycloserine, (b) d-serine, (c) l-NAME or (d) l-arginine on immobility time. Results show group means ±SEM. Two-way ANOVA *p < 0.05, ***p < 0.01, **p < 0.001 comparing within treatment groups

Full size image

Also, mice that received a combination of XA and l-NAME showed significant (F1,70 = 47.6, p < 0.001) potentiation in the antidepressant-like property in the FST (Fig. 10c). Similar observation was made for FLX but not in animals that received DES which rather seem to have been reversed. The antidepressant-like effects of XA and FLX and DES were all abolished in l-arginine-treated mice (Fig. 10d).

Discussion

We isolated xylopic acid, a kaurene diterpene, from Xylopia aethiopica: a plant that has antidepressant effect in rodent models (Biney et al. 2016) in addition to significant CNS depressant, analgesic and anti-inflammatory effects (Ameyaw et al. 2014; Biney et al. 2014).

A potential antidepressant effect was first explored in zebrafish using a CUS model where XA demonstrated effects akin to antidepressant agents. The zebrafish has become a model organism for neurobehavioural assays due to its high homology to the human CNS as well as ease of use in experimentation (Kalueff et al. 2014; Stewart et al. 2014; Fontana et al. 2018). It has been used to screen anxiolytics (Maximino et al. 2010; Benneh et al. 2017), anticonvulsants (Mussulini et al. 2013), antipsychotics (Kokel and Peterson 2011) and antidepressants (Meshalkina et al. 2018; Nowakowska et al. 2020) in recent times. XA-treated zebrafish, which have undergone significant stressors, showed reduced duration of bottom dwelling, a feature characteristic of depression and anxiety in this species (Maximino et al. 2010). This effect was similarly observed in fluoxetine-treated zebrafishes.

The initial antidepressant-like effect in zebrafish was also confirmed in mice by the significant reduction in immobility duration in both FST and TST even though XA does not have stimulatory effects in the spontaneous activity and open field tests (Biney et al. 2018). Generally, there were minor differences in the antidepressant-like properties of XA between FST and TST. XA was more potent and efficacious in FST than in TST. Antidepressants have been reported to produce different magnitudes of responses in both tests (Bourin et al. 2005). For example, dopaminergic antidepressants do not reduce immobility in FST in most species while noradrenaline reuptake inhibitors were devoid of effect in the tail suspension test in NMRI mice (Ripoll et al. 2003; Bourin 2019). Thus, these differences in response in the two most popular tests provide a first shot at the potential mechanism of action of xylopic acid.

Having shown an antidepressant potential, XA was further evaluated in the neuroinflammation hypothesis of depression. Undoubtedly, a significant connection exists between inflammation and major depressive disorder with depressed patients exhibiting hyper functioning of the HPA axis and elevated pro-inflammatory cytokines which has been associated with interference in glutamate and monoamine neurotransmission and neurogenesis in the hippocampus (Allison and Ditor 2014; Amodeo et al. 2017). With the knowledge that XA has anti-inflammatory effects (Osafo et al. 2018; Osafo et al. 2019), we evaluated its ability to mitigate LPS-induced depressive behaviour. Lipopolysaccharide induces depression-like behaviours through induction of pro-inflammatory cytokines that permeate the CNS and cause neuroinflammation (Bian et al. 2020). The inflammatory response evoked leads to increased catabolism of serotonin, production of free radicals and increased NMDA activity (Zhang et al. 2020; Dantzer et al. 2008). This ultimately leads to reduced extracellular serotonin and increased hippocampal atrophy all of which are associated with depression (Wang et al. 2020). XA treatment attenuated the depression-like behaviours of reduced social interaction and anhedonia, a core symptom of MDD and a very important measure in antidepressant screening of novel compounds (APA 2015). The ability of XA to reduce these behaviours adds another layer of evidence to support its antidepressant-like effect and gives credence to a potential for further drug development as an antidepressant.

To delineate the time-course of antidepressant action of XA, we evaluated XA in a chronic model for testing antidepressants: the repeated open space swimming test (OSST). Acute test such as the TST and FST may lack face and construct validity. Though they are of good predictive value, they may pick false positives such as atypical antipsychotics and calcium channel antagonists that have central stimulatory effects (Harro 2019). Chronic models for evaluating antidepressants such as the chronic unpredictable stress (CUS) test and repeated open space swimming test (OSST) have more construct validity, and hence, such false positives fail to exhibit antidepressant-like effects.

XA showed a comparatively fast-onset antidepressant effect in the OSST compared to fluoxetine. This effect was still present 24 h after halting treatment as was observed for fluoxetine. The search for fast onset of action antidepressants is very active in current antidepressant drug development (Ramaker and Dulawa 2017) which makes this finding exciting. In recent times, modulation of some serotonin receptors and glutamatergic neurotransmission have been linked to fast unremitting antidepressant actions (Wang et al. 2015; Zanos et al. 2018). Although we have not examined the direct effect of XA on 5-HT4 5-HT7 and 5-HT2c receptors, it is possible that XA may be eliciting the fast-onset antidepressant-like action by enhancing serotonergic transmission pathways involving these receptors. This is because agonists of 5-HT4 (Vidal et al. 2014) and antagonist of 5-HT7 (Mnie-Filali et al. 2011) and 5-HT2c (Opal et al. 2014) receptors have been shown to possess rapid-onset antidepressant-like activities. Indeed, our results showed a partial agonist effect on 5-HT2 receptors in the rat fundus strip, and therefore, XA can have antagonistic effect on the 5-HT2, a mechanism implicated in fast-onset antidepressant-like effects. Additionally, XA reduced other signature symptoms of MDD during open space swim test. Classical antidepressants like the SSRI exhibit a lag in antidepressant effect in the OSSTt while agents with rapid-onset antidepressant show almost immediate response (Ramaker and Dulawa 2017). XA-treated mice had well-kept hair and reduced rostral grooming in comparison to untreated animals. Analyses of grooming behaviour in depressed mice indicates high level of rostral grooming compared to naïve animals (Smolinsky et al. 2009; Bergner et al. 2016) but this was also reversed by XA treatment. Our findings indicate that XA shows fast-onset effect in the OSST, a model that has sensitivity, reliability and specificity in identifying rapid-onset antidepressant effect (Ramaker and Dulawa 2017) and has been used to identify 5-HT2c antagonists possessing fast-onset effects (Opal et al. 2014).

Having established an antidepressant-like property, the potential mechanisms of action of XA were investigated. We first assessed its effects on monoamines considering their role in depression. It is known that inhibiting the synthesis and or storage of specific a neurotransmitter will abolish the action of an antidepressant if that specific neurotransmitter contributes to antidepressant effects. For example, by selectively reducing serotonin levels, the antidepressant action of fluoxetine, a SSRI, is reversed (O’Leary et al. 2007; Ruhé et al. 2007; Adeoluwa et al. 2019). In our study, pre-treatment with the tryptophan hydroxylase inhibitor, p-chlorophenylalanine (pCPA), specifically inhibited 5-HT synthesis and reversed the established antidepressant-like effect of XA. The inhibition of synthesis or storage of catecholamines with α-methyl-p-tyrosine (AMPT) could not offset the effects of XA nor fluoxetine. According to work done by O’Leary et al. (2007), by pre-treating mice with the vesicular monoamine transporter (VMAT) reserpine, all monoamines including 5-HT are negatively affected and consequently both noradrenaline-based and 5-HT-based antidepressant will not work. Thus, not surprisingly, the antidepressant effects of XA, FLX and DES were all abolished in the presence or reserpine. These findings suggest a dominant role of 5-hydroxytryptamine in the observed antidepressant-like action of XA.

To further assess the serotonergic mechanism, we evaluated the effects of XA on serotonin receptors on the rat stomach fundus preparation. This tissue preparation has 5-HT1 and 5-HT2 receptors that evoke contractile responses when stimulated (Vane 1959). XA evoked partial agonist-like responses. 5-HT partial agonists like vortioxetine and buspirone have successfully been used to manage MDD thus making this finding another exciting one. The XA-induced responses were antagonised by the 5-HT receptor antagonist methysergide. The same preparation was also used to assess the effect of XA on monoamine oxidase (MAO) enzymes. MAO inhibitors are known to exert a potentiating effect on contractile responses produced by tryptamine an isolated rat fundus strip (Barlow 1961). However, unlike the monoamine oxidase inhibitor selegiline, XA could not significantly enhance tryptamine-evoked responses signifying a lack of monoamine oxidase inhibiting effect by XA.

Having identified a role of 5-HT in the observed antidepressant-like effects of XA, we sought to further confirm this role using the regulatory role of 5-HT on glutamate and nitric oxide pathways. The regulatory role of nitric oxide on glutamate neurotransmission in depression has been established. Subsequently, the efficacy of nitric oxide (NO) synthase inhibitors in depression is known with some antidepressants found to reduce expression and activation of NMDA receptors (Réus et al. 2016; Zhou et al. 2018). Stimulating NMDA glutamate receptors causes downstream activation of nNOS and the production of NO which is believed to be significant in the neurotoxic effects of glutamate (Zhou et al. 2018). However, 5-HT has negative modulatory effect on glutamate neurotransmission via activation of 5-HT7 and 5-HT2A receptors (Pehrson and Sanchez 2014; Li 2020). Therefore, agents that promote 5-HT transmission have synergism with negative modulators of this pathway. For example, decreasing the activity of nNOS enzyme with L-NAME will produce synergism with 5-HT-based antidepressants without affecting antidepressants whose actions are exclusively based on noradrenaline (Ostadhadi et al. 2016). We observed a similar effect in mice treated with XA or fluoxetine but not in desipramine-treated mice. L-NAME potentiated XA and fluoxetine’s effects but L-arginine reversed it. Same pattern of results was seen when the NMDA glutamate receptor was negatively modulated allosterically with d-cycloserine, an antagonist of the gylcineB co-bind site of the NMDA receptor. We observed an increase in antidepressant-like action of XA and FLX but not desipramine in mice pre-treated d-cycloserine whereas in mice pre-treated with glycineB agonist d-serine, the already established antidepressant effects were abolished (Poleszak et al. 2011; Poleszak et al. 2016). Again, this synergistic effects of XA with glutamate/nitric oxide pathway inhibitors confirm an association of serotonin neurotransmission in the observed antidepressant-like effects of XA.

Finally, we explored the effect of XA on oxidative stress and neuroprotection, both of which are implicated in depression (Czarny et al. 2018; Gross and Seroogy 2020). LPS stimulates oxidative stress in several brain regions causing dysregulation in oxidative stress makers like glutathione, malondialdehyde, CAT and SOD in the brain (Tomaz et al. 2014; Jangra et al. 2016). XA increased glutathione, SOD and CAT which constitute important antioxidant mechanisms in the first-line of attacking reactive oxygen species. Thus, XA reduced oxidative stress and protected the brain. The neuroprotective effect was further advanced by xylopic acid’s ability to increase brain level BDNF which is known to play a key role preventing depression and is suggested as a biomarker of efficacy for antidepressants (Autry and Monteggia 2012; Schröter et al. 2020). Patients diagnosed of major depression disorder are reported to present with lower hippocampal and prefrontal cortex levels of BDNF. By promoting branch formation and reducing neurodegeneration, BDNF promotes repair and survival of monoaminergic neurons (Mizui et al. 2016). The ability of XA to increase BDNF and exert this protective role is not surprising as some diterpenes have been shown to have neuroprotective effects (Adaramoye et al. 2010; Xu et al. 2011). This appreciable neuroprotective role of XA seen here supports the established antidepressant-like effects.

In conclusion, we have demonstrated that the kaurene diterpene, xylopic acid, has appreciable antidepressant-like effects that are dependent on serotonergic and neuroprotection mechanisms. It thus has a potential for further drug development as a therapeutic agent for managing MDD.