Gut microbiota depletion from early adolescence alters anxiety and depression-related behaviours in male mice with Alzheimer-like disease

Abstract

The gut-microbiota-brain axis plays an important role in stress-related disorders, and dysfunction of this complex bidirectional system is associated with Alzheimer's disease. This study aimed to assess the idea that whether gut microbiota depletion from early adolescence can alter anxiety- and depression-related behaviours in adult mice with or without Alzheimer-like disease. Male C57BL/6 mice were treated with an antibiotic cocktail from weaning to adulthood. Adult mice received an intracerebroventricular injection of amyloid-beta (Aβ)1-42, and were subjected to anxiety and depression tests. We measured, brain malondialdehyde and glutathione following anxiety tests, and assessed brain oxytocin and the hypothalamic-pituitary-adrenal (HPA) axis function by measuring adrenocorticotrophic hormone (ACTH) and corticosterone following depression tests. Healthy antibiotic-treated mice displayed significant decreases in anxiety-like behaviours, whereas they did not show any alterations in depression-like behaviours and HPA axis function. Antibiotic treatment from early adolescence prevented the development of anxiety- and depression-related behaviours, oxidative stress and HPA axis dysregulation in Alzheimer-induced mice. Antibiotic treatment increased oxytocin in the brain of healthy but not Alzheimer-induced mice. Taken together, these findings suggest that gut microbiota depletion following antibiotic treatment from early adolescence might profoundly affect anxiety- and depression-related behaviours, and HPA axis function in adult mice with Alzheimer-like disease.

Figure 1

The effects of antibiotic treatment alone or in combination with Aβ1–42 treatment on anxiety-related behaviour in male C57BL6 mice. To assess anxiety-related behaviour (A inner zone time; B inner zone entries) and locomotor activity (C rearing; D total line crossings), animals were subjected to the open field test. Data are expressed as mean ± SEM (n = 10). Antibiotic treatment significantly increased the inner zone time [p = 0.022], but it did not affect the other parameters [p > 0.05] compared to the water + vehicle mice. On the other hand, Aβ 1–42 administration in water-treated mice significantly decreased the inner zone time [p = 0.024] and entries [p = 0.003] compared to the water + vehicle group. However, Aβ 1–42 injection could not affect the inner zone time and entries [p > 0.05] in antibiotic-treated mice compared to the water + vehicle or antibiotics + vehicle groups. Besides, significant differences in all behavioural parameters in the open field test were not observed between antibiotics + Aβ 1–42 mice and water + vehicle group [p > 0.05]. Moreover, there are significant differences in the inner zone time [p = 0.009] and entries [p = 0.017] between water + Aβ 1–42-treated mice and antibiotics + Aβ 1–42-treated animals. Furthermore, both antibiotic and Aβ 1–42 treatments did not change [p > 0.05] the rearing and the total line crossings in the open field test.

Figure 2

The effects of antibiotic treatment alone or in combination with Aβ1–42 treatment on anxiety-related behaviour in male C57BL6 mice. To examine anxiety-related behaviour, animals were subjected to the light–dark box (A light time; B light entries; C latency; D faecal pellets). Data are expressed as mean ± SEM (n = 10). Antibiotic treatment significantly increased the time spent [p = 0.029] and the number of entries [p = 0.003] in the light area, whereas it did not affect the latency and faecal pellets [p > 0.05] in the light–dark box, as compared to the water + vehicle group. On the other hand, Aβ 1–42 administration significantly reduced the light time [p = 0.004] and entries [p = 0.013], and increased fecal pellets [p = 0.016; D] in water-treated mice, as compared to the water + vehicle group. In addition, Aβ 1–42 treatment significantly decreased the light entries [p = 0.049], but not the other behavioral parameters [p > 0.05] in antibiotic-treated mice relative to the antibiotics + vehicle group. However, we did not detect any significant differences in all behavioural parameters in the light–dark task between antibiotics + Aβ 1–42 mice and water + vehicle group [p > 0.05]. On the other side, significant differences in the light time [p < 0.001], the light entries [p = 0.001], the latency [p = 0.036], and faecal pellets [p = 0.008] were found between water mice treated with Aβ 1–42 and antibiotics + Aβ 1–42-treated mice.

Figure 3

The effects of antibiotic treatment alone or in combination with Aβ1–42 treatment on anxiety-related behaviour in male C57BL6 mice. To evaluate anxiety-related behaviour, animals were subjected to the zero maze (A open quadrant time; B open quadrant entries; C head dips; D stretch-attend posture). Data are expressed as mean ± SEM (n = 10). Antibiotic-treated mice exhibited a significant increase in the open quadrant time [p = 0.005] and entries [p = 0.009], while no alterations [p > 0.05] were found in the head dips and stretch-attend postures following antibiotic treatment, as compared to the water + vehicle group. On the other side, Aβ 1–42 injection significantly decreased the time spent [p = 0.02] and the number of entries [p = 0.038] in the open quadrants and increased the stretch-attend postures [p = 0.014] in water-treated mice in comparison with the water + vehicle group. In addition, antibiotics mice treated with Aβ 1–42 displayed a decrease in the open quadrant time [p = 0.05] and entries [p < 0.001] as compared to the antibiotics + vehicle group. However, there was no significant difference in all parameters in the zero maze between antibiotics + Aβ 1–42 mice and water + vehicle group [p > 0.05]. Furthermore, we found significant differences in the open quadrant time [p = 0.002] and the stretch-attend postures [p = 0.003] between water + Aβ 1–42 group and antibiotics + Aβ 1–42-treated mice.

Figure 4

The effects of antibiotic treatment alone or in combination with Aβ1–42 treatment on oxidative stress and antioxidant markers (A malondialdehyde; B glutathione) in the brain of male C57BL6 mice. Data are expressed as mean ± SEM (n = 10). Antibiotic treatment alone did not affect malondialdehyde and glutathione levels in the brain of mice [p > 0.05] as compared to the water + vehicle group. On the other hand, Aβ 1–42 treatment significantly increased malondialdehyde [p = 0.002] and decreased glutathione [p = 0.007] in water-treated mice as compared to the water + vehicle group. In addition, Aβ 1–42 administration significantly reduced glutathione [p = 0.031], whereas did not affect malondialdehyde levels [p > 0.05] in the brain of antibiotic-treated mice as compared to the antibiotics + vehicle group. However, no significant differences in these markers existed between antibiotics + Aβ 1–42 and water + vehicle groups [p > 0.05]. Furthermore, there were significant differences in malondialdehyde [p < 0.001] and glutathione [p = 0.02] levels in the brain between water + Aβ 1–42-treated animals and antibiotics + Aβ 1–42 mice.

Figure 5

Pearson correlations between anxiety tests and/or oxidative stress/antioxidant markers. Correlation coefficient: r; the number of animals: n; p-value: p.

Figure 6

The effects of antibiotic treatment alone or in combination with Aβ1–42 treatment on depression-related behaviours in male C57BL6 mice. To assess depression-related behaviour, animals were subjected to sucrose preference (A), social interaction (B; total interaction time) and forced swim (C; immobility time) tests. Data are expressed as mean ± SEM (n = 10). Antibiotic treatment did not change [p > 0.05] the percentage of sucrose preference, total interaction time, and immobility time in mice relative to the water + vehicle group. We also found that Aβ 1–42 administration significantly decreased the percentage of sucrose preference [p = 0.002] and the total interaction time [p = 0.002] but increased the immobility time [p = 0.025] in water-treated mice as compared to the water + vehicle animals. In addition, there were significant differences in the total interaction time [p = 0.029] and the immobility time [p = 0.01] between water + Aβ 1–42 and antibiotics + Aβ 1–42 groups. However, antibiotic treatment could not affect the sucrose preference [p > 0.05] in antibiotic-treated mice as compared to both the antibiotics + vehicle and water + vehicle groups. We also did not find any significant differences [p > 0.05] in sucrose preference, the total interaction time and the immobility time between antibiotics mice treated with Aβ 1–42 and the antibiotics + vehicle- or water + vehicle-treated animals.

Figure 7

The effects of antibiotic treatment alone or in combination with Aβ1–42 treatment on HPA axis function (A ACTH; B corticosterone) and brain oxytocin (C) in male C57BL6 mice. Data are expressed as mean ± SEM (n = 10). Antibiotic treatment significantly increased brain oxytocin [p = 0.013] but did not alter [p > 0.05] ACTH and corticosterone levels in the serum of mice as compared to the water + vehicle group. In addition, Aβ 1–42 administration significantly increased ACTH [p = 0.012] and corticosterone levels [p = 0.049] but not brain oxytocin [p > 0.05] in water-treated mice as compared to the water + vehicle group. Furthermore, no significant differences [p > 0.05] in ACTH, corticosterone, and oxytocin levels were observed between antibiotics + Aβ 1–42 mice and antibiotics + vehicle or water + vehicle groups.

Figure 8

Pearson correlations between depression tests and HPA axis function or oxytocin. Correlation coefficient: r; the number of animals: n; p-value: p.

Figure 9

Schematic timeline of the experimental design.