2.1 Animals and cells

ApoE4 mice (B6. Cg-ApoE^tm1UncCdh18^Tg(GFAP^{−APOE i4)1Hol}/J, No. 004631) with a C57BL/6 background were obtained from The Jackson Laboratory.²⁵ Wild-type (WT.) C57BL/6J mice were purchased from Guangdong Medical Laboratory Animal Center, China. All experimental animals were 12–14 weeks old. The experimental animals were housed at the Laboratory Animal Research Center, Peking University Shenzhen Graduate School, under a 12 light–dark cycle at 18–22°C. The mice had free access to food and tap water throughout the study period. Experimental procedures were designed so as to minimize the suffering of the mice. All the experiment procedures were carried out according to the protocols approved by the Institutional Animal Care and Use Committee of the Peking University Shenzhen Graduate School (Approval Number: 11110).

Human 293T cell lines were grown in high-glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% foetal bovine serum (FBS) (Gibco, Waltham, MA, USA). The cells were maintained in a humidified incubator with 95% air and a 5% CO₂ atmosphere at 37°C. When the cells attained 70% confluence, ApoE4 plasmid transfection and drug treatments were performed; subsequently, the cells were harvested and their fluorescence was observed.

2.2 Drug administration and schedule

The present investigation comprised the following experiments:

First experiment: Group 1: Wt treated with saline (100 μL/10 g). Group 2: Wt treated with LPS (1 mg/kg, i.p.), Group 3: ApoE4 Treated with saline (100 μL/10 g). Group 4: ApoE4 was treated with LPS (1 mg/kg, i.p.) (Figure 1A). LPS were treated for 3 days, 3 doses.

Second experiment: Group 1: Wt treated with LPS. Group 2: ApoE4 treated with LPS. Group3: ApoE4 mice treated with urolithin A (UA) (200 μg/kg, i.g.) + 5-aminoimidazole-4-carboxamide 1-β-D-ribofuranoside (AICAR) (100 mg/kg, i.p.). Group 4: ApoE4 was treated with LPS+ melatonin (10 mg/kg, i.p.) once daily for 2 h before LPS administration (Figure 1F).

Third experiment: Group 1: Wt treated with melatonin (10 mg/kg, i.p.). Group 2: Wt treated with melatonin + Luzindole (5 mg/kg, i.p.). Group 3: ApoE4 treated with melatonin (10 mg/kg, i.p.), Group: ApoE4 treated with melatonin + Luzindole (Figure 9A).

Additionally, ApoE4 and WT mice were administered melatonin (10 mg/kg, i.p.) in separate experiments.

All behavioural analyses (SPT, TST and FST) were performed as reported.^{26, 27} 24 h after the final LPS administration. Finally, the animals were sacrificed, and the serum and tissues were collected for further analysis.

2.3 Animal behaviour assay

2.4 Proteomic analysis

A proteomic analysis was performed as we have done previously.³²

2.5 NO and H₂O₂ measurement and TBARS assay, ELISA, mitochondrial copy number and ATP

Details have been mentioned in the Data S1.

2.6 Western blot

Immunoblotting was performed according to a previously described laboratory protocol. The detailed protocol is described in the Data S1.

2.7 Statistical analysis

Statistically significant differences among the experimental groups were determined using GraphPad Prism 8 software. An unpaired t-test was used to compare the two groups, while one-way analysis of variance (ANOVA), followed by Tukey's post-hoc and two-way ANOVA, was performed for multi-group comparisons. Data are presented as mean ± SEM, and p < 0.05 was considered as significant.

3.1 LPS aggravated depressive-like behaviours and neuroinflammation in ApoE4 transgenic mice

ApoE4 is a risk factor for neuropsychiatric disorders, including depression^{14, 33-35}; however, its underlying mechanism has not been addressed. To explore whether humanized ApoE4 induces depression-like behaviour in mice, we analysed mice behaviour with OFT, FST, TST and SPT. As shown in Figure 1B–E, ApoE4 mice did not show any significant behavioural changes. However, after LPS administration, mice displayed aggravated depression-like behaviours, as demonstrated by increased immobility and reduced sucrose preference compared to LPS-treated control mice. These findings indicate that ApoE4 overexpression could elicit more severe depressive-like behaviours upon stimulation by stressors (like LPS). We validated this hypothesis by treating ApoE4 mice with melatonin (before LPS administration) in the presence of LPS (Figure 1F). Interestingly, melatonin treatment rescued the LPS-stimulated depression-like behaviour in ApoE4 transgenic mice (Figure 1G–J). Similar results were also observed with UA and AICAR treatment, activators of mitochondrial biogenesis and mitophagy.^{36, 37}

Neuroinflammation coincides with depression, and LPS is a well-known inflammatory agent.^38-40 Herein, we determined whether ApoE4 acts as a proinflammatory agent in LPS-induced depression. Hence, we measured neuroinflammation in the experimental mice (with or without LPS administration). No significant differences were observed in cytokine levels between Wt and ApoE4 mice (Figure 2A,B). However, LPS administration enhanced cytokine (serum IL-1β, TNF-α, hippocampal IL-1β and IL-6) production in ApoE4 mice than in Wt mice, whereas IL-10 was significantly reduced in LPS-treated ApoE4 mice as compared to the WT controls (Figure 2A,B). Proinflammatory cytokines accelerate neuroinflammatory signalling through glial cell activation.^{39, 41} We investigated whether LPS-induced hyperinflammation caused glial cell activation. LPS administration significantly enhanced ionized Ca²⁺-binding adapter-1 (Iba-1), but not glial fibrillary acidic protein (GFAP) expression, in ApoE4 as well as in the Wt mouse hippocampus compared to saline-treated mice (Figure 2C). However, melatonin and UA + AICAR treatments reduced GFAP expression compared to that in LPS-treated ApoE4 mice. In contrast, no significant decline in Iba-1 level was observed after melatonin/UA + AICAR treatment (Figure 2D). GFAP is a major intermediate astrocytic cytoskeletal protein and is considered a marker of astrogliosis, whereas Iba1 is expressed in microglia and is upregulated during the activation of these cells.^{26, 42}

Mitochondrial abnormalities are accompanied by excessive free radical generation,⁴³ which may play a critical role in the pathological process and regulation of depression.^44-46 ROS and NO levels were measured to examine whether the LPS-induced changes in ApoE4 mice were associated with oxidative and nitrosative stress. LPS treatment significantly enhanced total ROS levels (serum and hippocampus), H₂O₂ concentration (serum and hippocampus), hippocampal TBARS levels and NO levels in ApoE4 mice compared to Wt mice, which were attenuated by melatonin administration. Additionally, we examined increased levels of ROS/H₂O₂ (serum) and NO in ApoE4 mice compared to those in Wt mice. However, no significant difference in ROS/H₂O₂ levels was observed between LP-treated Wt and ApoE4 mice. In contrast, LPS treatment elevated the ROS/H₂O₂/NO levels in ApoE4 mice. Free radical elevation was further validated using TBARS measurements (Figure S1).

3.2 LPS-altered mitochondrial-associated signalling in ApoE4 mice hippocampus

Mitochondrial dysfunction is associated with depression,¹⁹ and previous studies, including ours, have demonstrated that ApoE4 overexpression causes mitochondrial dysfunction.^{32, 47} However, the underlying mechanisms and the exact contribution of ApoE4 to MDD are yet to be elucidated. We first performed proteomic profiling to elucidate mitochondria-associated gene changes in ApoE4 mice (Figure 3).³² Mitochondrial DNA content, ATP levels and mitochondria-associated protein expression were measured in the experimental mice. ApoE4 mice displayed increased mtDNA (Figure 4A) and ATP (Figure 4B) levels, which were reduced by melatonin treatment (Figure 4E). However, LPS treatment increased ATP levels in the hippocampus of ApoE4 mice (Figure 4B), but melatonin and UA + AICAR treatment did not reduce ATP levels (Figure 4G).

To further evaluate ApoE4-associated mitochondrial dysfunction, we measured the mitochondrial complex protein expression levels. LPS treatment significantly increased complexes I and V and decreased complexes II and III, but did not significantly alter complex IV expression in ApoE4 mice (Figure 4C). melatonin increased complex I expression but did not significantly change the expressions of complex II, III, IV or V, while reducing ATP levels in the hippocampus of ApoE4 mice (Figure 4E). Additionally, melatonin and UA + AICAR combination treatment did not modify mitochondrial complex protein expression in LPS-treated ApoE4 mice (Figure 4F).

Changes in mitochondrial-associated proteins, including MFN1, DRP1, OPA1, PGC-1α and TFAM, were then explored to reveal the underlying differences between Wt and ApoE4. No significant changes were observed in the MFN1, DRP1, OPA1 and TFAM expressions, except PGC-1α, which was significantly increased in the hippocampus of ApoE4 mice (Figure 5A–F). This finding was further validated by mRNA analysis. Drp1, mfn1 and tfam mRNA levels remained unchanged in ApoE4 mice, except for opa1 mRNA level, which increased in ApoE4 mice (Figure 5G–J). However, melatonin treatment did not change MFN1, DRP1, OPA1 and PGC-1α expression in ApoE4 mice (Figure 5A–E), except for increased TFAM (Figure 5A,F). Furthermore, melatonin treatment enhanced mfn1 mRNA levels but did not affect drp1, opa1 or mfn2 mRNA levels (Figure 5G–J). Moreover, LPS treatment did not change MFN1, DRP1, OPA1 and PGC-1α expression (Figure 5K–P) in the ApoE4 mice. However, melatonin treatment significantly reduced OPA1 and increased TFAM levels compared to LPS-treated ApoE4 mice (Figure 5N,P). Furthermore, we did not find significant changes in the DRP1, MFN1, OPA1 and PGC-1α expressions in the LPS-treated ApoE4 mice after UA + AICAR treatment (Figure 5K–P).

3.3 LPS-induced autophagy (mitophagy) impairment in ApoE4 mice

Mitochondrial dysfunction, including impaired mitophagy, contributes to the progression of neuropsychiatric diseases.^{48, 49} In this study, we evaluated the potential involvement of mitophagy in the response of ApoE4 mice to lipopolysaccharides (LPS). Autophagy-associated ATGs and their associated signalling pathways were also measured. ApoE4 mice displayed increased p62, Parkin, but deceased LC3B II expression levels, whereas p-AMPKα, Atg5, Beclin-1 and PINK1 remained unchanged (Figure 6A–H). Similarly, no significant changes in atg5, p62, Beclin-1, lc3b and Parkin mRNAs levels were observed (Figure 6I–N), except for reduced pink1 mRNA levels in ApoE4 mice (Figure 6M). Similarly, PGC-1α, TFAM and PINK1 were increased, and p-AMPKα and LC3B II were decreased in HEK 293T cells transfected with ApoE4, whereas Atg5, p62, Beclin-1 and Parkin were unchanged (Figure S2D,E). Interestingly, melatonin treatment significantly increased Atg5 and LC3B II expression in ApoE4 mice (Figure 6A,C,F), suggesting that ApoE4-induced impaired mitophagy could be modulated by autophagy owing to melatonin.

Next, we sought to determine whether LPS influences ApoE4-associated mitophagy impairment. Surprisingly, LPS-treated ApoE4 mice displayed decreased Atg5, Beclin-1 and Parkin levels^{50, 51} increased PINK1 and p62 levels that remained unchanged (Figure 7A). As an autophagy inducer,^{27, 52} melatonin rescued the LPS-altered expression of ATGs, including Atg5, Beclin-1 and LC3B II, as well as that of the selective autophagy adapter protein—p62 (Figure 7B). Furthermore, in contrast to PINK1, melatonin treatment significantly increased p-AMPKα and Parkin expression in ApoE4 mice compared to that in the LPS-treated ApoE4 mice (Figure 7B,C). These results were further validated using UA + AICAR as an autophagy inducer.^{53, 54} Consistent with previous reports, these findings demonstrate that melatonin can improve mitochondrial dysfunction, including mitophagy, as described.^{55, 56} Furthermore, ApoE4 was overexpressed in HEK 293T cells (Figure S2A). Similarly, no significant changes in the expression of mitochondrial complex proteins were detected in the ApoE4-transfected 293T cells (Figure S2B). Except for increased PGC-1α, TFAM and PINK1, no MFN1, OPA1 and DRP1 changes could be observed in ApoE4 transfected 293T cells (Figure S2C,D).

We checked autophagy/lysosomal pathway impairments by further measuring LC-3 expression in HEK 293T cells using the GFP-mRFP-LC3 plasmid to validate mitophagy-related protein changes. No difference in GFP-LC3 intensity was observed between the Wt and ApoE4 mice (with or without melatonin/bafilomycin A1: BafA1/Chloroquine: Cq treatment) (Figure 8A). However, LPS treatment significantly enhanced the GFP-LC3 intensity in ApoE4 mice (Figure 8B), indicating LPS-induced autophagy/lysosomal pathway impairment. Interestingly, melatonin treatment reduced GFP-LC3 intensity in HEK 293T cells, suggesting that melatonin could reduce LPS-induced autophagy/lysosomal pathway impairment. This was further validated by BafA1 and Cq treatments. Both autophagy inhibitors (Baf1A/Cq) significantly enhanced GFP-LC3 intensity (Figure 8C). Therefore, ApoE4 dysregulation aggravates autophagy (mitophagy) impairment under stress (LPS).

3.4 Melatonin receptor antagonism blocked the effects of melatonin

Next, we sought to determine whether the effects of melatonin depended on its receptors. ApoE4 mice were treated with the melatonin receptor antagonist luzindole (Figure 9A). As shown in Figure 9B, luzindole significantly enhanced complexes I and III, whereas complexes II and V remained unchanged. Similarly, luzindole treatment enhanced PGC1-α expression but not MFN1, OPA1, Drp-1 and TFAM in the hippocampus of ApoE4 mice (Figure 9C). Concurrently, luzindole treatment reduced Atg5 expression in ApoE4 mice but enhanced Parkin expression in WT mice, whereas a decline of p-AMPKα expression could be observed in both WT/ApoE4 mice. Moreover, we did not detect significant changes in p62, Beclin-1, LC3B and PINK1 expression after luzindole treatment (Figure 9D).

Altogether, these results indicated that upon LPS stress, ApoE4 mice exhibited increased neuroinflammation, accelerated free radical generation, impaired mitophagy and led to depression-like behaviours. Melatonin reversed these detrimental effects of ApoE4 via its receptors.