1. Introduction
Postmenopausal women face a heightened risk of cognitive impairment, with estrogen depletion identified as a key contributing factor [
1], [
2]. Compounding this endocrine vulnerability, longitudinal epidemiological evidence indicates that midlife obesity further increases the risk of later-life cognitive decline in women [
3]. Chronic hypercaloric nutrition, particularly a high-fat diet (HFD), is a primary driver of obesity. Experimental models have corroborated the neurotoxic synergy of an HFD with estrogen deprivation, with ovariectomized (OVX) HFD-fed (OXHF) mice developing accelerated cognitive deterioration [
4], [
5]. Given the accelerating global obesity pandemic [
6], identifying the potential molecular mechanisms that underlie menopausal obesity-induced cognitive impairment and corresponding effective treatment options may greatly benefit the cognitive health of obese middle-aged women.
Traditional Chinese medicine (TCM) is a network-targeting therapeutic paradigm for managing chronic multisystem disorders. Bazi Bushen capsule (BZBS), a Chinese herbal compound approved by the National Medical Products Administration (NMPA) of the People's Republic of China (No. B20020585), is used clinically to treat fatigue and forgetfulness [
7]. Animal studies showed that BZBS protected against altered testicular morphology and cognitive decline in
D-gal- and NaNO
2-induced aging mice by upregulating sirtuin 6 (SIRT6) [
8], [
9]. Moreover, BZBS alleviated atherosclerosis through G-protein coupled estrogen receptor 1 (GPER1)-dependent anti-apoptotic and anti-inflammatory effects in OVX/apolipoprotein E-deficient (
Apoe−/−) mice [
10]. These anti-aging, anti-cognitive decline, and blood lipid-regulating effects position BZBS as a promising therapeutic candidate for treating obesity-exacerbated cognitive decline in estrogen-depleted women. However, the therapeutic potential of BZBS for treating the dual effects of chronic HFD exposure and hypoestrogenism and its underlying mechanisms are unclear.
Emerging evidence has revealed the dual regulatory role of the aryl hydrocarbon receptor (AhR) in metabolic and neurodegenerative pathologies. The obesity-associated adipokine kynurenine (Kyn) exacerbates metabolic dysfunction by hyperactivating AhR [
11]. Interestingly, AhR is sensitive to estrogen depletion [
12]. Our previous study revealed a critical neuropathological association: In both aged and 5-month-old female OVX estrogen-deficient mouse models, elevated accumulation of intranuclear AhR in the hippocampal formation mechanistically drove memory impairment [
13]. However, how AhR regulates cognitive decline in HFD-fed estrogen-deficient mice and the role of BZBS are unknown.
According to the
Huangdi Neijing (
Yellow Emperor's Canon of Internal Medicine), Qi is a vital force sustaining life, and its deficiency is the fundamental driver of aging. A contemporary bioenergetic study revealed striking parallels between the traditional concept of Qi and function of mitochondria, the cellular powerhouse that governs energy metabolism [
14]. Mitochondria-derived adenosine triphosphate (ATP) production is a quantifiable manifestation of Qi in biomedical terms, as it supports systemic homeostasis and critically influences aging trajectories. Notably, mitochondrial dysfunction, characterized by low ATP generation and oxidative stress, has emerged as a central pathophysiological driver of age-related neurodegeneration and cardiovascular disorders [
15]. Intriguingly, AhR has been shown to be a critical disruptor of mitochondrial homeostasis, with AhR activation exacerbating lipid peroxidation and cellular senescence by suppressing SIRT1 and SIRT3 [
16], [
17], thereby impairing ATP synthesis and the redox balance. The AhR/SIRT signaling axis may constitute a mechanistic nexus linking an HFD and estrogen deficiency-associated mitochondrial dysfunction with cognitive decline. On the basis of the theory of the
Yellow Emperor's Canon of Internal Medicine, BZBS is formulated to maintain a balance of Qi, blood, Yin, and Yang in the five viscera, and therefore elucidating the role of BZBS in mitochondrial function and clarifying the relationship between BZBS and both AhR and SIRTs warrant further investigation.
In this study, we investigated whether menopausal obesity disrupts mitochondrial bioenergetics through the AhR-miR-23a axis and whether BZBS mitigates this process by restoring nicotinamide phosphoribosyltransferase (NAMPT)-nicotinamide adenine dinucleotide (NAD+)-SIRT1 and proliferator-activated receptor gamma coactivator 1 alpha (PGC-1α)-SIRT3 pathway activity. This study provides a mechanistic validation of the Qi-replenishing bioactivity of BZBS. By integrating bioinformatic, molecular, and neurophysiological methods to decode BZBS mechanisms at the molecular level, we established a foundational pathway by which Chinese herbal medicine may achieve global acceptance.
2. Materials and methods
2.1. Chemicals and reagents
BZBS powder, comprising 16 characterized Chinese herbal components (
Table 1), was provided by Hebei Shijiazhuang Yiling Pharmaceutical Co., Ltd. (Lot 2104001; China).
A BZBS suspension was prepared by homogenizing the powder in 0.5% sodium carboxymethylcellulose (CMC-NA; Cat. No. 419281; Sigma-Aldrich, USA) through sequential magnetic stirring and ultrasonication steps for 60 min. The mixture was stored at 4 °C. Fresh suspensions were prepared weekly for all in vivo administrations to maintain the consistency and bioactivity of the formulation.
Lyophilized BZBS powder was prepared by solubilizing BZBS powder (1.0 g) in a 50% (v/v) dimethyl sulfoxide (DMSO) aqueous solution, followed by sequential vortex mixing (3000 r∙min-1, 5 min) and ultrasonication (40 kHz, 45 min). The mixture was centrifuged at 3200g for 10 min (4 °C) to remove insoluble particulates. The clarified supernatant was flash-frozen at −70 °C for at least 24 h and lyophilized in a Christ Alpha 2-4 LSCplus freeze dryer (−50 °C under 1 Pa vacuum). The lyophilized powder was chemically stable when stored at −20 °C for 24 months. This specific preparation was exclusively used for in vitro cell culture experiments.
Normal chow (Cat. No. H10010) and high-fat chow (Cat. No. H10060) were acquired from Beijing Huafukang Biotechnology Co., Ltd. (China). The compositions of the normal and high-fat chow are shown in Tables S1 and S2 in Appendix A. Adeno-associated virus (AAV)-pre-miR-23a-3p and AAV-AhR-nNLS/mNLS viral vectors were constructed and packaged by GenePharma (China).
Chlorogenic acid (CGA; Cat. No. HY-N0055), osthole (OST; Cat. No. HY-N0054), schisandrin A (Sch A; Cat. No. HY-N0693), schisandrin B (Sch B; Cat. No. HY-N0089), hyperoside (HYP; Cat. No. HY-N0452), imperatorin (IMP; Cat. No. HY-N0285), and Kyn (Cat. No. HY-104026) were purchased from MedChemExpress (MCE, China). These were all dissolved in DMSO to specific concentrations.
2.2. Human blood sample preparation
Plasma samples were collected from overweight or obese postmenopausal women (body mass index (BMI) ≥ 24.0 kg∙m-2) who underwent community health examinations. Detailed protocol is provided in Section S1.1 in Appendix A.
2.3. Animals and drug treatment
Female C57BL/6 mice (8 weeks old, (20 ± 2) g) were procured from Liaoning Changsheng Biotechnology Co., Ltd. (China) and acclimatized under controlled environmental conditions ((23 ± 1) °C, 55% ± 5% humidity, 12 h/12 h light/dark cycle) with
ad libitum access to autoclaved water and standard chow. Mice were randomly allocated to experimental groups using a computer-generated random number table that ensured balanced baseline body weights among the groups. The following groups were used for experiments: the sham control group fed normal chow and the OXHF group. A BZBS suspension in 0.5% (w/v) CMC-Na (vehicle) was administered daily via intragastric gavage (i.g.) for 60 consecutive days, and an equivalent volume of vehicle was administered to the control group. On the basis of previous studies [
18], the dosages of BZBS were 0.7, 1.4, and 2.8 g∙kg
-1∙d
-1. Oral donepezil (DON; 1 mg∙kg
-1; Cat. No. D6821; Sigma-Aldrich) was used as a positive control (Table S3 in Appendix A). All behavioral tests, tissue collection, and subsequent molecular analyses were conducted by investigators blinded to the group assignments. Behavioral assessments were conducted during week 8 postintervention. Mice were euthanized via cervical dislocation under deep anesthesia (pentobarbital, 50 mg∙kg
-1 intraperitoneally (i.p.)) 24 h after the final behavioral testing, and the hippocampi were rapidly dissected on ice-cooled plates for subsequent molecular analyses.
All experimental protocols were approved by the Institutional Animal Care and Use Committee of Harbin Medical University (approval No. IRB3000722) and the Chinese Association for Laboratory Animal Sciences (approval No. A5655-01), and were conducted in strict compliance with EU Directive 2010/63/EU and ARRIVE guidelines 2.0.
2.4. Ovariectomy experiment
OVX surgery was performed as previously described [
13]. Details are provided in Section S1.2 in Appendix A.
2.5. Behavioral tests
All animals underwent systematic behavioral acclimation comprising 2 min daily handling sessions by investigators for two consecutive days and 30 min of environmental habituation in the testing arena prior to each experiment. A 24 h inter-test interval was strictly followed to mitigate carryover effects between assessments. Cognitive performance was evaluated using the Barnes maze (spatial memory) and novel object recognition (NOR) tests (episodic memory). Behavioral parameters were recorded by a XinRuan behavior analysis system (China).
The Barnes maze and NOR tests were performed as previously described [
13]. Nesting behavior was evaluated following established ethological protocols [
19] with minor adjustments. Detailed protocol is provided in Sections S1.3-S1.5 in Appendix A.
2.6. Transmission electron microscopy (TEM)
TEM was performed according to our previously published methods [
20]. Detailed protocol is provided in Section S1.6 in Appendix A
2.7. Primary hippocampal neuron culture and treatment
Hippocampal tissue was aseptically dissected from postnatal day 0 Sprague-Dawley rat pups and enzymatically dissociated using 0.25% trypsin in a shaking water bath (37 °C, 60 r∙mim-1) for 15 min with gentle trituration every 5 min using fire-polished Pasteur pipettes. Enzymatic digestion was terminated by adding Dulbecco’s modified eagle medium (DMEM; Cat. No. 11965092; Invitrogen, USA) supplemented with 10% heat-inactivated fetal bovine serum (Cat. No. 3022A; Umedium, China). The mixture was centrifuged (1000g, 5 min, 4 °C) and the pellet was resuspended in complete culture medium. The cell suspension was plated on poly-D-lysine-coated six-well plates at a density of 1 × 106 cells per well. After 4 h of incubation, the medium was replaced with serum-free neurobasal medium (Cat. No. 10888022; Gibco, USA) supplemented with 2% B27 (Cat. No. 17504044; Invitrogen), 0.5 mmol∙L-1 L-glutamine, and 1% penicillin-streptomycin. Cultures were maintained at 37 °C in a 5% CO2 humidified incubator, with half-medium changes every three days. The neurons were used 5-7 days after plating in all experiments.
Oligonucleotide synthesis was performed as previously reported [
13]. Cultures were transfected with an
miR-23a mimic, and parallel cultures were treated with 400 μg∙mL
-1 BZBS lyophilizate dissolved in neurobasal medium. After the transfected cells incubated at 37 °C for 6 h, the medium was replaced with fresh conditioned medium.
2.8. Construction of Ahr overexpression plasmids
The pcDNA3.1-3xFlag-C vector was used as the backbone to construct Ahr overexpression plasmids. Two types of plasmids were generated: one with an intact nuclear localization signal (NLS) and the other with a mutated NLS. The plasmid with the intact NLS was based on the full-length rat AhR complementary DNA (cDNA) according to National Center for Biotechnology Information (NCBI) RefSeq protein sequence NP_001295183.1 and contained the native NLS motifs 12-15 (RKRR) and 36-38 (KRH). The cDNA was amplified by polymerase chain reaction (PCR) and cloned into the pcDNA3.1-3xFlag-C vector between the appropriate restriction enzyme sites. The NLS-mutated plasmid was generated using site-directed mutagenesis to replace the arginine residues at positions 12, 14, and 37 of Ahr with alanine residues, with the wild-type Ahr overexpression plasmid as the template. All constructed plasmids were verified by DNA sequencing to ensure that the fragment was properly inserted and the mutation sites were correct.
2.9. Dual-luciferase reporter assay
A dual-luciferase reporter assay was performed as previously reported [
21]. HEK293T cells at ∼50% confluence were co-transfected with a 20 nmol∙L
-1 miR-23a mimic (
miR-23a), antisense oligonucleotide inhibitor (AMO-23a), or scrambled negative control (NC) small interfering RNA (siRNA), along with 0.5 μg of the psiCHECK-2-
Nampt 3′ untranslated region (UTR) wild-type or mutant reporter plasmid expressing firefly luciferase and 1 μg of the empty vector (pcDNA3.1). Lipofectamine 2000 (Cat. No. 11668019; Invitrogen) at a 3:1 reagent-to-DNA ratio was used for transfection. Site-directed mutagenesis of the predicted
miR-23a binding sites in the
Nampt 3′ UTR was performed by overlap extension PCR with synthesized mutagenic primers (Sangon Biotech, China). All the constructs were sequence-verified by Sanger sequencing (TSINGKE Biotechnology, China). Forty-eight hours post-transfection, the cells were lysed with passive lysis buffer (1×; Cat. No. C8011; Promega, USA), and the luciferase activity was quantified using a GloMax Navigator luminometer (Promega) with sequential measurements of Renilla (50 μL of coelenterazine substrate) and firefly luciferase (50 μL of luciferase assay reagent II) signals. Relative luciferase activity was normalized to the Renilla/firefly luminescence ratio, with empty vector-transfected samples serving as baseline controls.
2.10. Quantitative real-time PCR (qRT-PCR)
qRT-PCR was performed as previously described [
13]. Detailed protocol is provided in Section S1.7 in Appendix A.
2.11. Mitochondrial respiration measurements
Mitochondrial respiration was measured as previously reported [
22]. Primary hippocampal neurons were isolated via trypsin digestion, centrifuged (1000
g, 5 min), resuspended in DMEM, and plated at a density of 3 × 10
6-4 × 10
6 cells per well. Mitochondrial respiratory function was assessed using high-resolution respirometry (Oxygraph-2k; Oroboros Instruments, Austria) with the following sequential substrate-inhibitor titrations: ① Basal respiration was measured in assay buffer; ② oligomycin (1 μmol∙L
-1; Cat. No. HY-N6782; MCE) was added to induce a nonphosphorylating (leak) state; ③ carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP; 2 μmol∙L
-1; Cat. No. HY-N6782; MCE) was titrated to uncouple the mitochondria and thus determine the maximum electron transport system capacity; and ④ rotenone (1 μmol∙L
-1; complex I inhibitor; Cat. No. HY-B1756; MCE) and antimycin A (5 μmol∙L
-1; complex III inhibitor; Cat. No. MS0070; Shanghai Maokang Biotechnology Co., Ltd., China) were added to inhibit residual oxygen consumption. Data were normalized to cell counts and analyzed using DatLab 7.4 software (Oroboros Instruments) to determine key parameters, including basal respiration, ATP-linked respiration, maximal respiratory capacity, and spare respiratory capacity.
2.12. Stereotaxic injection of AAV
A scrambled control sequence was injected as a NC. AAV was synthesized and packaged according to our previously reported study [
13]. Detailed protocol is provided in Section S1.8 in Appendix A.
2.13. ATP determination
An ATP assay kit (Cat. No. S0026B; Beyotime Biotechnology, China) was used to determine ATP levels in both the cells and tissue. Detailed protocol is provided in Section S1.9 in Appendix A.
2.14. Glutathione (GSH) determination
A GSH assay kit (Cat. No. S00052; Beyotime Biotechnology) was used to measure GSH levels in the cells and tissue. Detailed protocol is provided in Section S1.10 in Appendix A.
2.15. Reactive oxygen species (ROS) determination
ROS levels were measured as previously described [
18]. Detailed protocol is provided in Section S1.11 in Appendix A.
2.16. Total NAD+ determination
An NAD+ assay kit (Cat. No. S0175; Beyotime Biotechnology) was used to measure the total NAD+ concentration in the cells and tissue. Detailed protocol is provided in Section S1.12 in Appendix A.
2.17. Western blot
Total protein was extracted from mouse hippocampi and cultured cells. Detailed protocol and detailed information of primary antibodies are provided in Section S1.13 and Table S4 in Appendix A.
2.18. Bioinformatic analysis
Bioactive phytochemicals in BZBS related to cognitive impairment-associated pathways were identified. A systems pharmacology method was used to characterize the bioactive constituents of BZBS potentially involved in cognitive impairment- and oxidative stress-related pathways. Targets of 14 BZBS phytochemicals (Table S5 in Appendix A) were predicted using SwissTargetPrediction (STP) [
23], which were screened in the Traditional Chinese Medicine Systems Pharmacology Database (TCMSP) for Alzheimer’s disease (AD) relevance [
24]. Compound‒target‒disease networks were constructed in Cytoscape (v3.9.1) using the NetworkAnalyzer toolkit.
Functional enrichment profiling was performed by analyzing the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databased using the R “clusterProfiler” and “ggplot2” packages. ClueGO was used to reveal the GO pathway network.
Molecular docking was conducted to investigate the binding interactions between AhR and the four bioactive compounds in BZBS: CGA, OST, Sch A, and Sch B. The crystal structure of AhR was obtained from the AlphaFold protein structure database and the three-dimensional (3D) structures of the four bioactive compounds were obtained from the NCBI PubChem database. Docking simulations were performed using AutoDock 1.5.7 with 3D conformers of the receptor and ligands. The lowest free energy docking poses were visualized and analyzed using PyMOL 2.5.0. Binding affinities ≤ -5.0 kcal∙mol
-1 were considered to indicate strong receptor-ligand interactions [
25].
2.19. Local field potential (LFP) recording
LFPs were obtained as previously reported [
4]. LFPs in the dorsal hippocampal CA1 region were recorded under anesthesia. Detailed protocol is provided in Section S1.14 in Appendix A.
2.20. Immunohistochemistry
Immunofluorescence staining was performed as previously reported [
4]. Detailed protocol is provided in Section S1.15 in Appendix A.
2.21. Chromatin immunoprecipitation (ChIP) assay
A ChIP assay was performed as previously described [
13]. Prior to performing the ChIP assay, the
miR-23a-3p promoter sequence was retrieved from the University of California, Santa Cruz (UCSC) Genome Browser and input into the JASPAR database for predicting transcription factors. Thirty-two putative binding sites were predicted between AhR and the promoter sequence of
miR-23a-3p. Samples were divided into three groups for the ChIP hippocampal tissue assay: the sham group, OXHF group, and BSBZ group. The assay was performed with an imprint ChIP assay kit (Cat. No. 26156; Thermo Fisher Scientific, USA) in accordance with the manufacturer’s instructions. The PCR primers used in the assay were as follows: NC ChIP forward primer (F)1, 5′-GGAAGGACAGCATCATGCCG-3′; NC ChIP reverse primer (R)1, 5′-ATTATTCCCAGGCCCAAAGC-3′;
miR-23a-3p F2, 5′-CCAGAAGAGGGAGTCAGATC-3′;
miR-23a-3p R2, 5′-TGTCATGCACAAAGACCAGAC-3′;
miR-23a-3p F3, 5′-GATAGCTGTGCTTGTCATGTAG-3′;
miR-23a-3p R3, 5′-GAGAAGGTATTTAGCACCCAG-3′;
miR-23a-3p F4, 5′-TGGAAGTAGAGGAGGGCTAG-3′; and
miR-23a-3p R4, 5′-TAGCACCTAAAGAGAGAAAG-3′. These primers were designed to amplify fragments 260, 218, 324, and 241 bp in size, respectively, from selected genomic regions of
miR-23a-3p. qRT-PCR of the genomic regions containing the putative AhR-binding sites was performed on each sample in the three groups in triplicate. The cycling parameters were as follows: initial denaturation at 98 °C for 5 min; 35 cycles of denaturation at 98 °C for 30 s, annealing at 55 °C for 20 s, and extension at 68 °C for 20 s; and a final extension at 72 °C for 5 min. The relative occupancy of each immunoprecipitated factor at a locus was estimated by the comparative threshold method.
2.22. Cellular thermal shift assay (CETSA)
A CETSA was performed as previously described with modifications [
26]. SH-SY5Y cells were treated with BZBS (400 μg∙mL
-1) and incubated at 37 °C for 6 h. The cells were then exposed to specific temperatures for 5 min at each temperature. Immediately afterward, the samples were snap-frozen in liquid nitrogen and subjected to three freeze-thaw cycles. Finally, the lysates were centrifuged at 12 000 r∙min
-1 for 10 min, and the supernatant was collected for Western blot analysis.
2.23. Statistical analysis
All quantitative data are presented as mean ± standard error of the mean (SEM). Data quantification and statistical analyses were performed blindly, with the analyst unaware of group identities during data processing. Between-group differences were analyzed by a two-tailed unpaired Student’s t-test with Welch’s correction for unequal variances. Comparisons among three or more groups were conducted using one-way analysis of variance (ANOVA) with post hoc multiple comparisons adjusted by Tukey’s honestly significant difference test when omnibus F-values reached significance (P < 0.05). All statistical tests were implemented in GraphPad Prism 8.0 (GraphPad Software, USA), with graphical representations generated using the software’s built-in data visualization toolkit.
3. Results
3.1. BZBS rescues cognitive deficits and hippocampal mitochondrial dysfunction in OXHF mice
To evaluate the therapeutic potential of BZBS for treating cognitive decline associated with menopausal obesity, we established an OXHF mouse model and treated the mice with BZBS (0.7, 1.4, or 2.8 g∙kg
-1∙d
-1) or DON (1.0 mg∙kg
-1∙d
-1) by gavage for eight weeks (
Fig. 1(a)). Consistent with previous reports [
10], BZBS significantly inhibited body weight gain and attenuated the increase in total cholesterol and triglyceride plasma levels of OXHF mice (Fig. S1 in Appendix A). We then performed the Barnes maze test to evaluate the spatial memory of the mice (
Fig. 1(b)). The OXHF-treated mice showed significantly longer search distances than the sham-treated control mice (
Fig. 1(c)) during the probe trial. Moreover, they exhibited extended latencies (
Fig. 1(d)) and an increased number of errors (
Fig. 1(e)). These differences were effectively ameliorated by treatment with BZBS at doses of 1.4 and 2.8 g∙kg
-1 (
Figs. 1(c)-(e)); however, no effect was observed during the training period (Fig. S2 in Appendix A). We then investigated the means by which BZBS influenced the episodic memory of OXHF-treated mice using the NOR test. We found that when we replaced object B with the novel object C on the second day, there was a significant decrease in either the exploration time or the preference for the novel object C in OXHF mice compared with sham control mice, and this change was successfully reversed by BZBS (1.4 and 2.8 g∙kg
-1) (
Figs. 1(f)-(i), Fig. S3 in Appendix A). Additionally, OXHF mice treated with BZBS (1.4 and 2.8 g∙kg
-1) had better nesting scores than OXHF mice without drug intervention (
Figs. 1(j) and
(k)). All the effects of BZBS were similar to those of DON (
Figs. 1(b)-(k)). Notably, BZBS at a dose of 0.7 g∙kg
-1 did not reduce the cognitive deficits that occurred in OXHF mice, including reduced spatial memory, episodic memory, and nest building ability (
Fig. 1). These results indicated that BZBS ameliorated cognitive impairment in OXHF mice, with an effective dose threshold of 1.4 g∙kg
-1∙d
-1.
Given the established link between mitochondrial dysfunction and neurodegenerative aging [
27], we investigated whether BZBS-mediated cognitive protection involved mitochondrial rescue. TEM of the hippocampi of OXHF mice revealed severe mitochondrial damage as evidenced by swelling and loss of cristae (
Fig. 2(a)). These changes were confirmed by increased ROS levels and reduced GSH and ATP levels in the hippocampi of OXHF mice compared with those in sham control mice (
Figs. 2(b)-(e)). BZBS (1.4 and 2.8 g∙kg
-1) effectively reversed these changes (
Figs. 2(b)-(e)). Notably, the effects of BZBS were similar to those of DON (
Figs. 2(b)-(e)), which was previously reported to inhibit ROS generation [
28]. Consistent with these
in vivo cognitive data, BZBS at a dose of 0.7 g∙kg
-1 failed to achieve these protective effects on the hippocampal mitochondria of OXHF mice. These results suggest that the protective effects of BZBS on hippocampal mitochondria are related to the ability of BZBS to prevent decreased cognition in OXHF mice.
3.2. BZBS restores mitochondrial bioenergetics through the dual modulation of miR-23a-dependent pathways
The sirtuin (SIRT1-7) family comprises key proteins that regulate ATP and ROS formation [
29]. However, it is not known whether the SIRT family participates in OXHF-induced mitochondrial dysfunction. A comprehensive profiling of hippocampal SIRT levels revealed that SIRT1 and SIRT3 were selectively downregulated in OXHF mice, whereas levels of other isoforms (SIRT2/4-7) remained unchanged (Fig. S4 in Appendix A). BZBS treatment (1.4 and 2.8 g∙kg
-1) restored the protein levels of both SIRT1 and SIRT3 to near physiological levels in the hippocampi of OXHF mice (
Figs. 3(a) and
(b)). Strikingly, although DON rescued levels of SIRT1 protein, it did not affect levels of SIRT3 (
Figs. 3(a) and
(b)). These data imply distinct mechanisms between the two therapeutics despite their comparable antioxidant and cognitive benefits.
SIRT3, a mitochondrial-localized deacetylase that is central to oxidative metabolism, is transcriptionally regulated by PGC-1α [
30], which is a key regulator of mitochondrial biogenesis and redox homeostasis [
31], [
32]. Our prior work established a mechanistic link between estrogen deficiency and mitochondrial dysfunction, in which OVX surgery induced cardiac PGC-1α deficiency through
miR-23a-mediated posttranscriptional repression [
33]. The regulatory microRNA
miR-23a was shown to be elevated in the circulation of aging females [
20]. Consistently, our analysis of human plasma samples showed that
miR-23a levels were also increased in overweight and obese postmenopausal women and were negatively associated with Montreal Cognitive Assessment (MoCA) scores (Fig. S5 in Appendix A). These results suggest that
miR-23a may be a systemic mediator in postmenopausal women and therefore a potential therapeutic target. Surprisingly, we found that the PGC-1α level in the hippocampus of OXHF mice was significantly lower than in the sham control mice (
Fig. 3(c)), whereas the
miR-23a expression was higher. These effects were reversed by BZBS treatment (1.4 and 2.8 g∙kg
-1) (
Fig. 3(d)). Neither BZBS (0.7 g∙kg
-1) nor DON altered the level of PGC-1α or the expression of
miR-23a (
Figs. 3(c) and
(d)). As predicted,
miR-23a overexpression significantly inhibited the levels of both PGC-1α and SIRT3
in vitro (Figs. S6 in Appendix A). Similarly, BZBS effectively prevented the downregulation of both PGC-1α and SIRT3 induced by
miR-23a overexpression (Figs. S6 in Appendix A). These findings established
miR-23a as an upstream epigenetic regulator of the PGC-1α-SIRT3 axis in OXHF-induced mitochondrial dysfunction.
However, an initial bioinformatic screening using established microRNA prediction platforms (TargetScan 8.0 and PicTar 4.0) failed to identify putative
miR-23a binding sites within the 3′ UTRs of either SIRT3 or SIRT1 messenger RNA. The SIRT family consists of NAD
+-dependent protein deacetylases that are crucial regulators of cellular processes [
34] and emerging evidence has suggested a positive correlation between NAD
+ availability and SIRT1 level, as NAD
+ depletion was shown to significantly downregulate SIRT1 expression [
35]. Therefore, we quantitatively analyzed the NAD
+ concentration in the hippocampi of OXHF mice. As predicted, the NAD
+ concentration was significantly lower in the hippocampi of OXHF mice than in those of age-matched sham control mice (Fig. S7 in Appendix A). Surprisingly, these effects were reversed by BZBS (1.4 and 2.8 g∙kg
-1) but not by DON (
Fig. 3(e)). To elucidate the mechanistic basis by which BZBS restored NAD
+ levels, we focused on the NAMPT-dependent salvage pathway, the predominant NAD
+ biosynthetic route in mammals [
36]. This pathway includes sequential catalysis by NAMPT (converting nicotinamide (NAM) to nicotinamide mononucleotide (NMN)) and nicotinamide mononucleotide adenylyltransferase (NMNAT) isoforms (synthesizing NAD
+ from NMN) (
Fig. 3(f)). Western blot analysis revealed that NAMPT levels were significantly reduced in OXHF mice (Fig. S8(a) in Appendix A); however, no observable changes were detected in NMNAT2 or NMNAT3 levels compared with age-matched sham control mice (Figs. S8(b) and (c) in Appendix A). Notably, BZBS treatment (1.4 and 2.8 g∙kg
-1) specifically rescued NAMPT levels in the hippocampi of OXHF mice; this effect was not detected with DON treatment (
Fig. 3(g)).
To elucidate the molecular basis of NAMPT dysregulation and BZBS-mediated protection, we again conducted a bioinformatic screening using the microRNA prediction platforms TargetScan 8.0 and PicTar 4.0 and identified a conserved
miR-23a binding motif within the 3′ UTR of the
Nampt gene (
Fig. 3(h)). Compared with the NC, functional validation based on dual-luciferase reporter assays in HEK293T cells demonstrated that transfection with
miR-23a mimics markedly inhibited luciferase signals (
Fig. 3(i)). However, the luciferase signals were diminished when the 3′ UTR of
Nampt was mutated (
Fig. 3(i)). Furthermore, we found that
miR-23a overexpression significantly inhibited NAMPT, NAD
+, and SIRT1 levels (Fig. S9 in Appendix A). Importantly, BZBS (1.4 and 2.8 g∙kg
-1) effectively prevented these changes (Fig. S9). Furthermore, ROS levels decreased (
Figs. 3(j) and
(k)), GSH levels increased (
Fig. 3(l)), and the oxygen consumption rate (OCR) increased in response to BZBS (
Figs. 3(m)-(o)). These results suggest that, unlike DON, the protective effects of BZBS on mitochondrial function are mediated in part by the
miR-23a-NAMPT-NAD
+-SIRT1 axis.
3.3. miR-23a contributes to the protection against cognitive decline exerted by BZBS in OXHF mice
To further investigate the role of
miR-23a in BZBS-mediated cognitive preservation, we established a
miR-23a-overexpressing mouse model by injecting AAV-pre-
miR-23a bilaterally into the hippocampal CA1 regions of adult normal female mice. The mice were then administered BZBS (1.4 g∙kg
-1) via gavage for two months (
Fig. 4(a)). The mice in the NC group underwent identical stereotaxic surgery and virus injection procedures as those of the experimental group but with a control virus that did not express
miR-23a. As shown in
Fig. 4(b), AAV-pre-
miR-23a injection significantly upregulated
miR-23a expression in the mouse hippocampi; this upregulation was inhibited by BZBS treatment. A Barnes maze test assessment (
Figs. 4(c)-(f)) revealed that, compared with NC mice, mice injected with AAV-pre-
miR-23a showed significantly increased search distances, prolonged latencies, and an increase in the number of errors when finding the target hole, all of which were ameliorated by BZBS treatment. Similarly,
miR-23a overexpression also impaired episodic memory and nest building in mice, but these cognitive impairments were reversed by BZBS treatment (Figs. S10 and S11 in Appendix A). As anticipated,
miR-23a overexpression in the hippocampi induced by AAV-pre-
miR-23a injection significantly downregulated the protein level of NAMPT, reduced the content of NAD
+, and inhibited SIRT1 level, all of which were reversed by BZBS treatment (
Figs. 4(g)-(i)). Concurrently, BZBS treatment also ameliorated the reduced levels of PGC-1α and SIRT3 in the hippocampi of mice caused by AAV-pre-
miR-23a injection (
Figs. 4(j) and
(k)). Functionally, BZBS protected against increased ROS levels and reduced GSH and ATP levels induced by
miR-23a overexpression (
Figs. 4(l)-(o)). These results suggest that BZBS protects against cognitive decline induced by
miR-23a overexpression.
Next, we evaluated whether the protective effect of BZBS on OXHF-induced cognitive impairment was due to its inhibition of miR-23a. To test this hypothesis, we injected additional AAV-pre-miR-23a into the hippocampi of OXHF mice to overcome the inhibitory effect of BZBS on miR-23a. We found that AAV-pre-miR-23a effectively blocked the protective effects of BZBS on spatial memory (Fig. S12 in Appendix A), episodic memory (Fig. S13 in Appendix A), and nest-building (Fig. S14 in Appendix A) in OXHF mice. In addition, the mitochondria protection exerted by BZBS through the NAMPT-NAD+-SIRT1 and PGC-1α-SIRT3 pathways was blocked by AAV-pre-miR-23a injection in OXHF mice (Fig. S15 in Appendix A). Collectively, these results confirmed that BZBS alleviated OXHF-induced cognitive decline by silencing miR-23a, thereby reactivating the NAMPT-NAD+-SIRT1 and PGC-1α-SIRT3 mitochondrial antioxidant pathways.
3.4. BZBS regulates miR-23a transcription by targeting AhR
Our study confirmed that BZBS critically regulated mitochondrial antioxidant pathways by suppressing miR-23a expression; however, the upstream regulatory mechanism was not clear. BZBS contains 16 ingredients (
Table 1) [
10], from which 14 unique components were identified by ultra-performance liquid chromatography (UPLC) analysis (Table S5) [
37], [
38]. Multidimensional target prediction using the STP and KEGG databases revealed that the potential targets of 11 of the 14 components were significantly associated with AD pathogenesis (
Fig. 5(a)). Further analysis using the TCMSP database identified that six of these, CGA, OST, Sch A, Sch B, HYP, and IMP, had targets related to AD pathways (
Figs. 5(b) and
(c)). Notably, GO enrichment analysis revealed that the target genes of these six components were significantly clustered in oxidative stress response pathways (
Fig. 5(d)), suggesting that these may mediate the antioxidant effects of BZBS.
To determine the specific BZBS components that regulate
miR-23a expression, we systematically evaluated
miR-23a levels in SH-SY5Y cells after treatment with the six candidate compounds. After 48 h of treatment, CGA (12 and 24 μmol∙L
-1), OST (40 μmol∙L
-1), Sch A (20 and 40 μmol∙L
-1), and Sch B (4 and 8 μmol∙L
-1) significantly reduced
miR-23a expression compared with the solvent (DMSO) (
Figs. 5(e)-(h)), whereas HYP and IMP had no effect (Fig. S16 in Appendix A). Building on these findings, we conducted a multidimensional analysis of the four active components (CGA, OST, Sch A, and Sch B) using the STP database, which identified 290 potential molecular targets, nine of which were transcription factors (
Fig. 5(i)). To further explore the transcriptional regulatory mechanisms of these four components, we used the JASPAR and TransmiR databases to comprehensively predict transcription factor binding sites. We ultimately identified three candidate transcription factors that could bind the
miR-23a promoter: AhR, progesterone receptor (PGR), and androgen receptor (AR) (
Fig. 5(j)). Western blotting revealed marked nuclear accumulation of the AhR protein in the hippocampi of OXHF model mice (
Fig. 5(k)), while the subcellular distribution of PGR and AR was unchanged (Fig. S17 in Appendix A). Notably, plasma AhR levels also significantly increased in overweight and obese postmenopausal women and were inversely correlated with MoCA scores (Fig. S18 in Appendix A), providing human evidence supporting the pathological relevance of AhR signaling.
To validate the direct targeting effect of BZBS on AhR, molecular docking analysis was conducted to examine the binding capacity between AhR and the BZBS active components (CGA, OST, Sch A, and Sch B). As shown in
Figs. 6(a)-(d)), CGA bound to the ligand-binding domain (LBD) of AhR at key residues, including ARG236, ASN121, GLN234, and GLU211, with a binding energy of −5.31 kcal∙mol
-1. OST also interacted with the LBD by forming hydrogen bonds with GLU211, GLN234, and ARG236, with a binding energy of −7.22 kcal∙mol
-1. Sch A was predicted to bind at the interface of the LBD and DNA-binding domain (DBD), through hydrophobic interactions with GLN118, THR264, and GLN234 and a binding energy of −8.06 kcal∙mol
-1. Sch B bound specifically to an allosteric regulatory site adjacent to the LBD through hydrogen bonds with ARG212, GLY235, and GLU211, with a binding energy of −6.75 kcal∙mol
-1. Collectively, these docking results suggest that all four active components may interact with AhR, thus providing a structural basis for the regulation of AhR by BZBS.
To further assess these interactions under physiological conditions, CETSA was performed. Compared with the DMSO control, BZBS significantly increased the thermal stability of AhR over a broad temperature range (41-53 °C) (
Fig. 6(e)). Furthermore, CGA, OST, and Sch B significantly stabilized AhR at 45 and 49 °C, while Sch A stabilized AhR at 45, 49, and 53 °C (Fig. S19 in Appendix A). These results indicated that these four active constituents of BZBS each engaged with AhR in cells, providing qualitative support for constituent-level AhR targeting.
Our previous work confirmed that the transcription factor AhR directly bound to the
miR-23a promoter in primary neurons [
13]; however,
in vivo evidence to support this has been lacking. In the present study, we performed ChIP analysis, which demonstrated that AhR bound to the −1963 to −995 bp interval of the
miR-23a promoter in hippocampal tissue (
Figs. 6(f)-(j)), consistent with our earlier
in vitro findings [
13]. Therefore, we further examined the dynamic changes in the binding of AhR to the
miR-23a promoter under pathological conditions. Binding of AhR to the
miR-23a promoter region was significantly greater in the OXHF group than in the sham group (
Figs. 6(h)-(i)). Notably, BZBS treatment reversed this OXHF-induced increased binding (
Figs. 6(h)-(i)), indicating that BZBS effectively modulated AhR and
miR-23a promoter interactions. Taken together, these
in vivo findings clearly demonstrated that BZBS regulates
miR-23a transcription by attenuating AhR activation.
3.5. BZBS alleviates oxidative damage and cognitive decline by inhibiting AhR nuclear translocation
Activated AhR translocates to the nucleus, a process required for its transcriptional regulatory activity [
39]. To determine whether AhR nuclear translocation was required for
miR-23a upregulation and for BZBS to exert its therapeutic effects, we generated AhR constructs with either an intact nuclear localization signal (AhR-nNLS-Flag) or a mutated nuclear localization signal (AhR-mNLS-Flag), as previously reported, to prevent ligand-dependent nuclear localization of AhR [
40]. The Flag tag was used to distinguish between exogenous and endogenous AhR. Western blot analysis demonstrated that both Flag-tagged constructs produced high levels of total AhR in primary hippocampal neurons (
Fig. 7(a), upper panel). AhR-nNLS-Flag was detected in both the nucleus (
Fig. 7(a), middle panel) and the cytoplasm (
Fig. 7(a), lower panel); in contrast, AhR-mNLS-Flag was detected exclusively in the cytoplasm (
Fig. 7(a)). Following stimulation with the AhR agonist Kyn [
11], AhR-nNLS-Flag efficiently translocated to the nucleus, while AhR-mNLS-Flag remained in the cytoplasm (
Fig. 7(a)). qRT-PCR analysis showed that, in the absence of Kyn, transfection with either construct alone did not alter
miR-23a expression. However, upon Kyn stimulation, neurons transfected with AhR-nNLS-Flag showed significantly increased
miR-23a expression compared with controls transfected with the empty vector. Transfection with AhR-mNLS-Flag did not enhance Kyn-induced
miR-23a expression (
Fig. 7(b)). These findings indicated that Kyn-induced
miR-23a upregulation required AhR nuclear translocation.
To validate these findings
in vivo, OXHF mice received bilateral hippocampal injections of AAV-AhR-nNLS or AAV-AhR-mNLS followed by BZBS treatment for eight weeks. Western blot analysis showed that both AAV-AhR-nNLS and AAV-AhR-mNLS injections reversed the BZBS-induced reduction in total AhR protein levels (
Figs. 7(c) and
(d)). Specifically, AAV-AhR-mNLS produced cytoplasm-limited AhR, whereas AAV-AhR-nNLS underwent normal nuclear translocation (
Figs. 7(c) and
(d)). qRT-PCR analysis showed that AAV-AhR-nNLS abolished the BZBS-mediated inhibition of hippocampal
miR-23a expression, whereas AAV-AhR-mNLS did not (
Fig. 7(e)). Consistently, BZBS failed to improve spatial memory and episodic memory in mice injected with AAV-AhR-nNLS yet further ameliorated cognitive deficits in mice injected with AAV-AhR-mNLS (
Figs. 7(f)-(l)). Together, these
in vivo data supported that AhR nuclear translocation was required for both OXHF-induced pathological signaling and the therapeutic actions of BZBS.
Next, to further verify that AhR activity is required for BZBS to exert its protective effects, we performed a pharmacological rescue experiment with Kyn. Immunofluorescence staining and Western blotting showed that BZBS reduced the nuclear accumulation of AhR in the hippocampus of OXHF mice, whereas Kyn co-treatment markedly reversed this effect (
Figs. 8(a) and
(b)). Functionally, Kyn restored
miR-23a expression, abolished the BZBS-mediated activation of the NAMPT-NAD
+-SIRT1 and PGC-1α-SIRT3 pathways (
Figs. 8(c)-(h)), and significantly weakened the antioxidant effects of BZBS in the hippocampus (
Figs. 8(i)-(l)). These findings further supported that BZBS exerted its mitochondrial antioxidant protection, at least in part, by suppressing AhR-dependent
miR-23a transcription.
After establishing that AhR nuclear translocation must be inhibited for BZBS to exert its mitochondrial antioxidant effects, we next asked whether this molecular pathway contributed to higher-order neural network function and cognition. Hippocampal theta oscillations (4-12 Hz) are neurophysiological hallmarks of spatial navigation and memory consolidation [
41]. To determine whether AhR-
miR-23a axis dysregulation was linked to network-level dysfunction, we assessed the theta rhythm integrity in OXHF mice using
in vivo electrophysiology techniques (
Figs. 9(a) and
(b)). Power spectral analysis revealed a marked deterioration of theta oscillations in OXHF mice, as indicated by a reduced peak theta power and frequency (
Figs. 9(c)-(e)). Notably, BZBS restored these theta parameters to near-baseline levels, whereas Kyn co-administration completely abolished this effect. Consistently, Kyn also reversed the beneficial effects of BZBS on spatial memory (
Figs. 9(f)-(i)), episodic memory (
Figs. 9(j)-(l)), and executive function (
Figs. 9(m) and
(n)).
4. Discussion
Our study is the first to demonstrate that menopausal obesity activates AhR-miR-23a crosstalk, dysregulating both the NAMPT-NAD+-SIRT1 and PGC-1α-SIRT3 pathways to drive mitochondrial energy failure, whereas BZBS targets the AhR protein to control miR-23a-regulated NAMPT-NAD+-SIRT1 and PGC-1α-SIRT3 pathway activity. This study provides mechanistic validation of the Qi-replenishing bioactivity of BZBS, a novel pathway modulator. The results of this study advance our understanding of the interplay between epigenetic control and neuronal bioenergetics and confirm BZBS as a promising multitarget therapeutic candidate for menopausal obesity-related cognitive impairment.
4.1. Mitochondrial dysfunction as a central pathogenic hub
The OXHF model recapitulates the key features of postmenopausal cognitive impairment, including hippocampal mitochondrial damage, oxidative stress, and bioenergetic failure. Consistent with clinical observations linking mitochondrial dysfunction to neurodegenerative aging [
27], OXHF mice had swollen mitochondria with disrupted cristae, reduced ATP production, and increased ROS levels. BZBS treatment (≥ 1.4 g∙kg
-1) effectively reversed these mitochondrial defects, along with improvements in three behavioral paradigms: spatial memory, assessed by the Barnes maze; episodic memory, evaluated by the NOR test; and executive function and motivational behavior, measured by the nesting test. To benchmark the clinical efficacy of BZBS, we selected DON, a first-line acetylcholinesterase inhibitor, as a positive control in this study. Notably, the behavioral efficacy of BZBS was comparable to that of DON, although their underlying mechanisms appeared to differ. While DON partially rescued SIRT1 (likely by suppressing ROS) [
28], BZBS uniquely restored both SIRT1 and SIRT3, suggesting it facilitated broader mitochondrial reprogramming. This finding aligned with emerging evidence that activating both SIRT1 (nuclear NAD
+ sensor) and SIRT3 (mitochondrial deacetylase) synergistically enhanced oxidative metabolism and stress resistance [
42].
4.2. miR-23a: A dual-target epigenetic regulator
A pivotal finding of this study was the identification of
miR-23a as a master regulator that linked estrogen deficiency to mitochondrial collapse.
miR-23a overexpression in OXHF hippocampi suppressed both PGC-1α (mitochondrial biogenesis) and NAMPT (NAD
+ salvage), creating a “two-hit” bioenergetic crisis in mitochondria. Mechanistically,
miR-23a directly targeted the 3′ UTRs of
Ppargc1a (PGC-1α) and
Nampt, as confirmed by luciferase assays. BZBS normalized
miR-23a levels, which reactivated PGC-1α-SIRT3- and NAMPT-NAD
+-SIRT1-mediated redox homeostasis. Importantly, AAV-mediated
miR-23a overexpression abolished the benefits of BZBS, confirming that the effects of BZBS depended on
miR-23a silencing. These findings extended prior reports that
miR-23a functioned in processes from cardiac aging [
33] to the central nervous system (CNS), and therefore is a systemic mediator of menopausal metabolic dysfunction. The microRNAs typically regulate multiple targets; our study specifically focused on validating the functional relevance of NAMPT and PGC-1α as key effectors in menopausal obesity. This finding was supported by the consistent rescue of their downstream pathways and mitochondrial phenotypes following BZBS treatment. However, we cannot exclude the possibility that additional
miR-23a targets may contribute to these phenotypes, which warrants further investigation.
4.3. AhR as a novel target of BZBS
The discovery that AhR is the transcriptional driver of
miR-23a in OXHF mice is a significant advance. AhR, a ligand-activated transcription factor, accumulates in hippocampal nuclei under estrogen deficiency where it directly binds the
miR-23a promoter [
13]. In this study, by integrating bioinformatics, molecular biology, and neurophysiology, we found that BZBS suppressed AhR aggregation in the nucleus, likely through the bioactive components of BZBS, CGA, OST, Sch A, and Sch B. This hypothesis was strengthened by evidence from a battery of experiments, including molecular docking, CETSA-based target engagement, and
in vivo ChIP assays on hippocampal tissue. Importantly, Kyn, a physiological activator of AhR, was shown to negatively correlate with serum estrogen levels in women [
43], [
44], [
45]. Kyn-mediated activation of AhR induced its pathological nuclear translocation, an essential molecular event mechanistically linked to aggravated obesity-associated metabolic dysfunction and AD pathogenesis [
11], [
46]. In the present study, we found that Kyn reversed the BZBS-induced reduction in nuclear AhR accumulation and abolished BZBS-regulated
miR-23a expression and levels of its downstream signaling molecules, including SIRT1, PGC-1α, and SIRT3, and the oxidative stress markers ROS, GSH, and ATP. Collectively, these findings provided the first evidence of the AhR-
miR-23a axis as a potential therapeutic target in menopausal obesity-related cognitive impairment. Genetic manipulation using AhR-nNLS and AhR-mNLS further confirmed that AhR nuclear translocation was required for pathological signaling and represents a critical step forward in understanding the therapeutic mechanism of BZBS, as BZBS exerted its
miR-23a-suppressive and cognitive-protective effects only when AhR nuclear entry was blocked.
4.4. Multicomponent synergy in BZBS
The pharmacological complexity of BZBS, with 14 identified bioactive compounds, is the basis of its multitarget efficacy. BZBS was formulated according to the theory of the
Yellow Emperor's Canon of Internal Medicine, a foundational TCM doctrine emphasizing a balance among Qi, blood, Yin, and Yang across the five viscera, with a focus on tonifying Qi and supporting kidney function to counteract aging and chronic diseases. Intriguingly, in TCM, the concept of “Qi deficiency” is considered the root of aging, which aligns with modern medical insights into the effects of low mitochondrial ATP production [
14], a hallmark of age-related neurodegenerative and cardiovascular disorders [
15]. This similarity between ancient theory and contemporary science was exemplified by the restoration of mitochondrial ATP synthesis mediated by BZBS, mechanistically bridging the TCM concept of “Qi” with bioenergetic regulation. Functional screening revealed that
miR-23a expression was significantly suppressed
in vitro by four key BZBS components, CGA, OST, Sch A, and Sch B (but not HYP or IMP), which aligned with their AhR-modulating and antioxidant properties. For example, Sch B was reported to enhance mitochondrial energy metabolism by increasing the protein expression of SIRT3 in H
2O
2-stimulated HT22 cells [
47], and CGA was shown to scavenge oxidized low-density lipoprotein (oxLDL)-induced endothelial oxidative damage and mitochondrial dysfunction by increasing SIRT1 expression [
48]. These findings suggest that multiple bioactive components of BZBS may collectively contribute to its overall activity, potentially exerting broad pathway regulation rather than targeting a single pathway lacking direct mitochondrial or epigenetic effects. This integration of TCM principles with molecular mechanisms underscores the unique ability of herbal formulations to address the multifactorial pathology of chronic diseases. Therefore, precisely elucidating how individual small molecules within TCM formulations interact to target intracellular molecular pathways is key to advancing the globalization of TCM. By accurately mapping the metabolic characteristics of BZBS after administration and detecting the concentrations of small-molecule compounds that enter the brain, the mechanisms by which BZBS improves cognitive decline in menopausal women with obesity can be further elucidated.
4.5. Translational relevance and limitations
The preclinical findings presented here have several implications for future clinical research. First, we demonstrated that BZBS effectively ameliorated cognitive deficits and mitochondrial dysfunction in OXHF mice by inhibiting the AhR-miR-23a axis and activating the NAMPT-NAD+-SIRT1 and PGC-1α-SIRT3 pathways at doses of 1.4 and 2.8 g∙kg-1, whereas a 0.7 g∙kg-1 dose had no significant effects. However, these findings only provide experimental evidence to inform clinical dose selection in individualized administration (e.g., in patients with elevated AhR/miR-23a expression). Moreover, the unique dual-pathway regulatory mechanism of BZBS (unlike that of DON, in which only SIRT1 is affected) suggests that BZBS may warrant further investigation in postmenopausal patients with cognitive impairment accompanied by mitochondrial dysfunction. Second, although we believe that BZBS inhibits AhR, the intensity of the inhibitory effect relative to that of BAY-218 (AhR antagonist) should be thoroughly evaluated. Additionally, although the OXHF model effectively recapitulates the interaction between estrogen deficiency and obesity, it does not reflect key features of human AD pathology, including amyloid-β deposition and tau hyperphosphorylation. Future studies using models with classical AD pathology may help determine whether the benefits of BZBS extend beyond menopausal obesity-related cognitive impairment.
In addition to mechanistic insights and disease model considerations, the optimal therapeutic regimen that will translate the preclinical promise of BZBS into clinical practice remains a critical unknown factor. Specifically, whether greater efficacy can be achieved by directly targeting AhR or
miR-23a than by using the multicomponent BZBS formulation is likely to depend on the clinical context. Although single-target interventions can precisely modulate specific pathways in menopausal obesity, they may fail to address the broader systemic perturbations associated with the disease, such as metabolic dysfunction and neuroinflammation, that are simultaneously targeted by BZBS [
49], [
50].
Although a previous clinical study reported that 12 weeks of BZBS treatment did not produce adverse changes in vital signs, physical examination findings, electrocardiogram results, or laboratory values compared with those at baseline [
7], potential long-term and dose-dependent side effects still warrant further evaluation.
Our previous study demonstrated that neither mice undergoing OVX surgery alone nor sham mice fed an HFD exhibited cognitive decline after two months. In contrast, OXHF mice showed significant cognitive impairment [
4]. Therefore, in this study, we did not include a non-OVX obese control group. Further study is needed to disentangle the respective contributions of obesity and estrogen deprivation to cognitive decline in patients with menopausal obesity.
Our clinical cohort study showed elevated plasma AhR and miR-23a levels in overweight or obese postmenopausal women, and both levels negatively correlated with cognitive function. These findings provide initial clinical support demonstrating the involvement of the AhR-miR-23a axis in cognitive impairment in this population. However, the relatively small sample size and the absence of intervention data preclude conclusions on whether BZBS can regulate this axis or improve cognitive function in clinical settings. Future studies with larger cohorts are needed to determine whether AhR and miR-23a plasma levels and cognitive performance change dynamically in response to BZBS intervention.
5. Conclusions
BZBS orchestrates cognitive recovery in menopausal obese subjects through a sequential mechanism in which AhR-mediated
miR-23a transcription is inhibited. The consequent downregulation of
miR-23a reactivates the PGC-1α-SIRT3 and NAMPT-NAD
+-SIRT1 pathways to restore the mitochondrial ATP/ROS balance (
Fig. 10). Although previous studies demonstrated the clinical potential of BZBS [
38], the present study is the first to translate the TCM concept of “replenishing Qi” mediated by BZBS into a framework of mitochondrial function. Collectively, our findings identified BZBS as a promising therapeutic candidate for menopausal obesity-related cognitive decline. With its preclinical efficacy and established clinical safety profile, BZBS merits further investigation in human studies.