3.1 Compound D30 improved cognitive function in mice impaired by fAβ-injection.
Intracerebroventricular injection (i.c.v.) of Aβ is commonly used to investigate the function of Aβ in the pathogenesis of AD. Injection of Aβ induces AD-like pathological changes in the mouse hippocampus by increasing Aβ deposition, oxidative stress, and neuroinflammation, and impairing memory and cognitive functions [23-26]. Here, i.c.v. of fAβ provided a sophisticated in vivo condition for testing the effects of D30 and its potential as a therapeutic compound for AD.
We first designed an experimental protocol and prepared an animal model by icv of fAβ (Fig. 1A-C; see Methods section 2.6, Experiment 1). The mice were randomly divided into Control, Sham, Fibril (fAβ-injected), and low-dose (10 mg/kg), middle-dose (20 mg/kg), high-dose (40 mg/kg) D30-treated groups. During the experimental period, the body weight of mice in all groups displayed upward trends, which indicated that the animals were healthy (Fig. 1D).
We evaluated the long-term memory of the animals by measuring their escape latency in the Morris Water Maze test. During the training period, mice in different groups showed decreasing day-to-day escape latency, which indicated the effectiveness of the training (Fig. 1E). As shown in the representative trajectories of the fifth day of training, D30-treated mice appeared more frequently in the quadrant where the platform was located compared with (untreated) fAβ-mice, which suggested that the D30-treated mice remembered the location of the platform better (Fig. 1I). On the test day, we recorded the number of times the animals crossed the region that used to be the hidden platform. Compared with the Control and Sham group, the fAβ group (Fibril alone) recorded fewer crossings, which indicated that fAβ indeed caused cognitive impairment. Interestingly, all three of the D30-treated Fibril groups performed significantly better than the Fibril alone group (p< 0.05). In fact, the platform crossing performance of the D30- treated Fibril groups were equivalent to that of the Sham group (Fig. 1F).
We conducted Open Field test experiments to investigate spatial exploration behavior of fAβ-injected mice. In contrast to the Fibril alone model group, D30-treated groups showed significant differences in the distance traveled in the center region, which indicated that D30 improved spatial exploration behavior of fAβ-induced mice (Fig. 1G and Fig. 1J).
Likewise, New Object Recognition experiments were used to assess the memory loss and cognitive impairment. Injection of fAβ impaired the ability to remember and differentiate novel objects, and this impairment was nearly completely prevented by D30 treatment (Fig. 1H and Fig. 1K); thus, D30 attenuated fAβ-induced visuospatial memory impairment. Collectively, the behavioral experiments indicated that D30 counteracted fAβ-induced memory loss and cognitive impairment.
3.2 D30 reduced Aβ deposition and neuroinflammation in fAβ-injected mice.
To better understand the effect of D30 on cognitive improvement, we investigated the effect of D30 on pathological indicators associated with AD progression, particularly the Aβ deposition, oxidative-stress, and neuroinflammation. Immunohistochemical results revealed that lateral ventricular injection of fAβ led to Aβ deposition in the hippocampus and cortex. The D30 treatment groups showed reductions of fAβ deposition (Fig. 2A). Additionally, Enzyme-linked immunosorbent assay (ELISA) results demonstrated that intraventricular fAβ injection increased cortical Aβ deposition, which was reduced by D30 treatment, consistent with the immunohistochemical results (Fig. 2B). In addition, ELISA assays indicated that fAβ injection increased reactive oxygen species (ROS) levels and decreased superoxide dismutase (SOD) expression in the cortex, and D30 intervention restored ROS and SOD to normal levels (Fig. 2C-D).
Activated microglia and astrocytes secrete inflammatory cytokines, including TNF-α and interleukin-1β (IL-1β), and increase the expression of inducible nitric oxide synthase (iNOS); these factors promote neuroinflammation and neuronal death [27]. Activated microglia upregulate the expression of ionizing calcium adaptor binding protein (Iba1), a commonly used marker for microglia, and activated astrocytes increase expression of glial fibrillary acidic protein (GFAP) [28]. Considering the activity of neuroinflammation in AD progression, we focused on assessing inflammatory indicators. ELISA assays demonstrated that fAβ injection led to upregulation of cortical levels of IL-1, IL-6, Iba1, and GFAP. However, D30 intervention counteracted these inflammatory responses induced by fAβ (Fig. 2E-H). In parallel, immunoblotting confirmed that fAβ injection increased the expression of Iba1, GFAP, and iNOS and decreased neuronal marker NeuN in the hippocampus. These indexes were significantly reversed in all the D30 treated groups, which indicated a neuroprotective activity of D30 against inflammatory damage in fAβ-injected mice (Fig. 2I-M). We then chose 20 mg/kg dose for further in vivo experiments employing D30.
3.3 D30 reduced Aβ deposition and microglia activation in fAβ-injected mice.
In general, neuroinflammation due to Aβ deposition and glial cell activation are closely associated with AD. Therefore, inhibiting Aβ deposition and maintaining glial cell homeostasis is an important strategy for treating AD [29]. To investigate the effect of D30 on fAβ-induced neuroinflammation and glial cell activation, D30 was applied once daily for one week from the day after fAb injection. The mice were then sacrificed for the 1-week fAb paradigm or administered an additional fAb injection followed by a second week of D30 for a 2-week paradigm (Fig. 3A).
As shown in Fig. 3B-C, a single fAβ injection resulted in Aβ deposition, and a second fAβ injection led to an outbreak of Aβ in the hippocampus; D30 intervention effectively reduced Aβ deposition induced with one and two injections of fAβ. Injection of fAβ into the lateral ventricle led to the activation of microglia, as evidenced by typical morphological changes such as enlargement of the cell body and branch retraction in the hippocampal dentate gyrus regions. In contrast, D30 application suppressed the fAβ-induced microglial morphological activation. Similar changes were observed in other hippocampal regions (Fig. S1).
Notably, with two injections of fibril, accompanied by a large amount of Aβ deposition, the morphology of microglia underwent more significant changes. We confirmed that the activation state of microglia was positively correlated with the amount of Aβ deposition. Quantitative analysis showed that D30 effectively reduced fAβ deposition in both the 1-week (p<0.05, Fig. 3D) and 2-week (p<0.05, Fig. 3G) paradigms. Immunoblotting confirmed that D30 effectively reduced the deposition of Aβ in the hippocampus after two fAβ injections (p<0.01, Fig. 3J-K). In addition, D30 suppressed the fAβ-induced morphological changes of microglia, namely reduction in the numbers and length of the branches, at 1-week (p<0.05, Fig. 3E-F) and 2-week paradigm (p<0.05, Fig. 3H and p<0.01, Fig. 3I) quantified by Skeleton and Sholl analysis (Image J).
3.4 D30 suppressed microglia M1 polarization in fAβ-injected mice.
Microglia are known to be the primary resident innate immune cells that can be activated and release inflammatory cytokines causing inflammatory injury. Activated proinflammatory microglia are commonly referred to as M1 microglia. However, alternatively activated M2 microglia can protect neighboring cells and promote tissue repair by releasing anti-inflammatory cytokines [30]. Inhibiting microglia overactivation and inflammatory response can effectively reduce neuroinflammation and mitigate AD progression. Therefore, microglia polarization depends on the activation status, and maintaining the microglia polarization balance is a foreground therapeutic option for AD. We found by immunofluorescence staining that, under the stimulation of fAβ, microglia underwent changes in morphology and were transformed into neurotoxic M1 microglia, accompanied by a significant increase in CD68 in the hippocampus (Fig. 4A and Fig. S2A). D30 treatment remodeled microglia morphology and counteracted the burst of CD68 expression. Fluorescence quantification showed that D30 reduced the expression of CD68 (p<0.05, Fig. 4C) and the co-localization of CD68 and Iba1 induced by a single fAβ injection (p<0.05, Fig. 4D-E).
Second injection of fAβ resulted in a surge in the number of microglia and drastic change in their morphology, accompanied by a burst of CD68. D30 counteracted the activation of microglia and the concomitant burst of CD68 and reshaped the morphology of microglia in the hippocampal dentate gyrus (Fig. 4B) and other hippocampal regions (Fig. S2B). Quantitative fluorescence analysis confirmed that double fAβ stimulations was much more effective in inducing CD68 expression and microglia activation. Intervention with D30 again suppressed the double fAβ-induced microglia activation (p<0.01, Fig. 4F-H). Immunoblot analysis demonstrated that two injections of fAβ resulted in an explosion of the inflammatory markers CD68 (p<0.001, Fig. 4I) and TNF-α (p<0.05, Fig. 4K), which were effectively blocked by D30 intervention (Fig. 4J).
Collectively, Aβ application induced dose-dependent changes in CD68 production and microglia M1 polarization that could be suppressed by D30, suggesting a pivotal role of D30 in regulating microglial polarization.
3.5 D30 regulation of Gal3 and PI3K/AKT/mTOR in fAβ-injected mice.
Gal-3, a key upstream regulator of the microglial immune response, impairs the degradation and clearance of fAβ in AD [31]. Boza-Serrano et al. demonstrated that Gal-3 promoted proinflammatory activation of primary microglia cells in response to fAβ [32]. However, still unknown was whether injected fAβ affected Gal-3 expression and the association between Gal-3 and fAβ-mediated neuroinflammatory responses.
To investigate the aforesaid unknown, we measured Gal-3 expression after fAβ injection without and with D30 treatment. Immunofluorescence staining (Fig. 5A) showed that a single fAβ injection did not lead to Gal-3 expression in the dentate gyrus of mice (Fig. 5C). Interestingly, a second fAβ injection was more effective in increasing Gal-3 level, which suggested that the induction of Gal-3 by fAβ was dose-dependent. Not surprisingly, D30 effectively counteracted fAβ-induced Gal-3 expression (Fig. 5B). Fluorescence quantification showed that the expression of Gal-3 was elevated in the presence of fAβ compared with the sham group, whereas the level of Gal-3 was suppressed by D30 (Fig. 5D).
In addition, immunoblot experiments confirmed the sharp increases in APP and Gal-3 with two injections of fAβ, and these increases were reduced by D30 intervention (p<0.05, Fig. 5E-G). In primary microglia (co-cultured with fAβ for 4 h in the presence or absence of 20 μM D30), fAβ increased Gal-3 transcription; however, acute application of D30 did not significantly affect Gal-3 mRNA level regardless of fAβ stimulation (Fig. 5L).
Network pharmacology is a novel approach based on the theory of systems biology and network analysis to systematically study drugs and diseases. We screened 123 targets associated with D30, and 6273 targets associated with AD disease. By creating Venn maps, we obtained 105 targets in common (Fig. S3A). Based on the drug-disease target network, we constructed a protein-protein interaction network (Fig. S3D), and we performed MCODE cluster (Fig. S3B) and topology analyses (Fig. S3C).
To further analyze the drug-disease-target relationship, we performed the drug-disease-target analysis, Gene Ontology (GO) enrichment analysis (Fig. 5H), and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis (Fig. 5I). On this basis, we constructed a comprehensive drug-disease-pathway-target network (Fig. S3E). Our network pharmacological studies revealed that AKT was the highest-ranking target of D30 intervention in AD, and the mechanism of D30 intervention in AD was highly related to phosphorylation. In addition, PI3K/Akt was the most significant signaling pathway for D30 intervention in AD.
In conjunction with the predictions of network pharmacology, we measured the mRNA levels of PI3K, Akt, and mTOR in fAβ-induced microglia. There was no significant change in the mRNA of PI3K and mTOR after treatment with D30 under fibril stimulation. However, D30 reversed fibril - induced AKT elevation (p<0.01, Fig. 5L). Immunoblotting experiments confirmed that two fibril injections resulted in reduction of PI3K, p-AKT, and mTOR, and D30 intervention counteracted these reductions (p<0.05, Fig. 5J-K).
The phosphatidylinositol 3-kinase/protein kinase B (PI3K/AKT) pathway regulates cell proliferation, motility, growth, survival, and metabolic functions [33]. Recent reports have shown that a variety of natural products can inhibit inflammation and exert neuroprotective and cognitive-improving effects in animal models by upregulating the PI3K/AKT/mTOR pathway [34]. We confirmed that fAβ induced reduction of PI3K, p-AKT, and mTOR, and D30 counteracted PI3K, p-AKT, and mTOR induction in fAβ-injected mice. These results suggested that D30 regulation of PI3K/AKT/mTOR had an important role in fAβ-induced Aβ deposition, microglia activation, and inflammation.
3.6 D30 improved Aβ disposal by reprograming p62/Nrf2/HO-1 signaling in fAβ-induced microglia.
Ubiquitin-binding protein p62 (Chelate 1/SQSTM1) regulates various physiological processes and has an important activity in oxidative stress defense. The p62 protein can sequester Keap1, which uncouples Nrf2 from the ubiquitin-proteasome system, leading to its stabilization and translocation to the nucleus, followed by activation of antioxidant target genes, such as glutamate cysteine ligase and heme oxygenase 1 (HO-1) [35,36]. Our previous study showed that D30 possessed anti-inflammatory and antioxidant effects by upregulating Nrf2-HO-1. Therefore, we investigated the activity of D30 on fibril-induced inflammation and oxidative stress in primary microglia.
Immunofluorescent study confirmed that D30 increased the expression of p62 (p<0.05, Fig. 6 A and D) and its downstream factor HO-1 (p<0.05, Fig. 6C and F). Immunoblotting confirmed that D30 elevated expression of p62 and HO-1 (p<0.001, Fig. 6G-I).
In parallel, immunoblotting experiments confirmed that D30 increased fAβ uptake in fAβ-stimulated primary microglia (p<0.05, Fig. 6 J- K). Nrf2 staining of microglia confirmed that D30 increased nuclear translocation of Nrf2 (p<0.05, Fig 6. B and E), which suggested that D30 may facilitate fAβ disposal by upregulating the p62/Nrf2/HO-1 signaling pathway. Interestingly, D30 increased p62 transcription, and fAβ stimulated p62 translation, likely as a response to fAβ challenge. D30 exerted a synergistic effect on top of fAβ by further stimulating p62 transcription, which suggested that D30 stimulated microglia to better handle the fAβ challenge (p<0.001, Fig. 6L).
Collectively, we confirmed that the disposition of fAβ by D30 was highly correlated with the upregulation of p62/Nrf2/HO-1 in primary microglia. This correlation suggested that p62/Nrf2/HO-1 was an essential mechanism by which D30 promoted fAβ disposal by microglia and the D30 counteracting Aβ-induced oxidative stress and inflammation.
3.7 D30 reduced inflammation and activation of astrocytes in fAβ-injected mice.
Like microglia, aberrant activation of astrocytes is strongly associated with the progression of AD. In response to Aβ stimulation, glial cell activation is accompanied by elevated iNOS, which is involved in plaque formation and leads to faster plaque aggregation. Thus, glial cell activation creates a vicious circle that accelerates Aβ-mediated inflammatory response. The application of fAβ led to the elevation of iNOS in the hippocampal dentate gyrus (Fig. 7A) and other regions (Fig. S4 A-B). In addition, the morphology of astrocytes was changed by Aβ application, as reflected in the thickening of astrocyte GFAP skeleton, and D30 alleviated the activation of astrocytes. Fluorescence quantification showed that D30 suppressed the fAβ-induced iNOS elevation. (p<0.05, Fig. 7B).
We observed a robust increase of iNOS accompanied by more pronounced astrocyte activation in response to second fAβ injection. D30 again near completely inhibited the fAβ-induced iNOS burst. Meanwhile, the morphology of astrocytes in the presence of D30 appeared well resembled in the sham group despite of two fAβ injections, with much less GFAP skeleton thickening than those received two fAβ injections alone (p<0.001, Fig. 7C-D). Consistently, other hippocampal regions followed the same pattern (Fig. S4 C-D).
Besides GFAP skeleton changes, fAβ-injection also induced C3 expression in astrocytes that could be suppressed by D30 application (Fig. 8A and Fig. S5A). Fluorescence quantification demonstrated that D30 intervention reversed the fAβ-induced increase of C3, and the co-localization of GFAP and C3 (p<0.05, Fig. 8B -D), which suggested that D30 had exerted a strong effect in preventing proinflammatory astrocyte activation. With two fAβ applications, D30 remained effective at suppressing astrocyte activation and C3 increase in mice received repeated fAβ application (Fig. 8E-F). The Manders’ Colocalization Coefficients for co-localization of GFAP and C3 were significantly reduced by D30 (p<0.05, Fig. 8G), while there was no significant difference in Pearson’s R value (Fig. 8H). Consistently, other hippocampal regions followed the same pattern (Fig. S5B).
Collectively, Aβ application increase iNOS production, induced astrocyte morphological changes with elevated C3 expression. Interestingly, all these characteristics of astrocyte activation were all near completely suppressed by D30.
3.8 The protective effect of D30 on neurons
Above all,D30 reversed the fAβ-induced activation of microglia and astrocytes. Particularly, with D30, the fAβ elevated expression of CD68 and C3, marker of M1 microglia and A1 astrocytes, were suppressed to near sham levels. M1 microglia and A1 astrocytes have been proved to be neurotoxic and destructive to synapses [37,38]. Thus, D30 is expected to be neuroprotective against fAβ. To investigate this, we assessed the activity of D30 toward fAβ-induced neuronal damage. Thy1.1 GFP mice were selected to investigate neuronal processes and dendritic spines. Lateral ventricle injections of fAβ led to neuronal damage as reflected by the reduction of dendritic branches and spines; D30 maintained the number of neuronal branches and spines (Fig. 9A-C). By quantitative analysis, D30 reversed the reduction of branches and spines induced by fAβ, implying a neuroprotective role of D30 (p<0.05, Fig. 9D-H).
We further investigate the effect of D30 on complement and and synaptic protein levels, including PSD95, VGAT (vesicular GABA transporter), and synaptophysin, in fAβ-induced cognitively impaired ICR mice. Immunoblot assay confirmed that application of fAβ increased C3 expression, and D30 reduced C3 production. In parallel, fAβ application led to a decrease in PSD95, VGAT, and synaptophysin, whereas D30 upregulated PSD95, VGAT, and synaptophysin, suggesting a neuroprotective effect (p<0.05, Fig. 9I-K). Given that GABA (γ-aminobutyric acid) is a major neurotransmitter in the CNS, and VGAT is a marker for GABA transporter [39,40], while PSD95 is linked to glutamatergic neurotransmission, D30 reversed the fAβ-induced reduction of VGAT and PSD95, suggesting the protection of both glutamatergic and GABAergic synapses.
In conclusion, our study confirmed that D30 increased the number of dendritic spines and the expression of synaptic proteins, exerting a neuroprotective effect that may have been due to the remodeling of glial cells by D30 and mitigation of the inflammatory environment.