Dysfunctional TRIM31 of POMC Neurons Provokes Hypothalamic Injury and Peripheral Metabolic Disorder under Long‐Term Fine Particulate Matter Exposure
Chenxu Ge, Jiamao Lin, Changsheng Yang, Chuanwang Miao, Fengxiang Li, Shuqiang Zhao, Lei Zou, Xuedong Teng, Lina Liu, Tingguang Li, Yan Sun, Qiang Li, Deshuai Lou, Linfeng Hu, Xi Liu, Gang Kuang, Jing Luo, Minxuan Xu, Minghui Chang, Jun Tan, Yanrong Ren, Bochu Wang
TL;DR
Long-term exposure to fine particulate matter causes brain and metabolic issues, and a protein called TRIM31 in specific brain cells helps prevent these effects.
Contribution
The study identifies TRIM31 in POMC+ neurons as a key regulator of PM2.5-induced hypothalamic injury and metabolic dysfunction.
Findings
TRIM31 knockout in POMC+ neurons worsened PM2.5-induced metabolic and neurological damage.
TRIM31 overexpression reduced neuronal death and inflammation caused by PM2.5.
TRIM31 activates Nrf2 signaling to protect against PM2.5 stress.
Abstract
Particulate matter ≤2.5 µm (PM2.5) elevates risks of neurological and chronic metabolic diseases, but the underlying mechanisms linking PM2.5‐induced central nervous system (CNS) injury to metabolic dysfunction remain unclear. Hypothalamic pro‐opiomelanocortin‐expressing (POMC+) neurons regulate systemic metabolic homeostasis, and tripartite motif‐containing protein 31 (TRIM31) modulates inflammation and metabolism. Here, we investigated whether TRIM31 in POMC+ neurons mediates PM2.5‐induced hypothalamic injury and peripheral metabolic disorders in mice subjected to 24‐week PM2.5 exposure. TRIM31 knockout in POMC+ neurons (POMCCre/+;TRIM31flox/flox ) exacerbated PM2.5‐;induced increases in mean blood pressure and fat weight, liver weight reduction, adipocyte hypertrophy, hepatic lipid deposition, energy expenditure abnormalities and insulin resistance. It also aggravated hypothalamic…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5
FIGURE 6
FIGURE 7
FIGURE 8
FIGURE 9
FIGURE 10- —National Natural Science Foundation of China10.13039/501100001809
- —Chongqing Research Program of Basic Research and Frontier Technology10.13039/501100013223
- —Children's Research Institute of National Center for Schooling Development Programme and Chongqing University of Education
- —Chongqing Professional Talents Plan for Innovation and Entrepreneurship Demonstration Team
- —Chongqing Natural Science Foundation Innovation Development Joint Fund
- —Basic Medicine Research Innovation Center for Cardiometabolic Diseases, Ministry of Education Open Projects Fund
- —The Scientific and Technological Research Program of Chongqing Municipal Education Commission of China
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
Topicsinterferon and immune responses · Neuroinflammation and Neurodegeneration Mechanisms · Neurogenesis and neuroplasticity mechanisms
Introduction
1
Air pollution has emerged as a critical health hazard for humanity given its extensive prevalence in our environment [1]. As a complex mixture, air pollution encompasses gases (such as carbon monoxide, nitrogen oxides, sulfur dioxide (SO_2_), and ozone) and airborne particulate components, including particulate matter (PM), which refers to inhalable airborne particles, such as dust, smoke, and chemical droplets with diameters typically ≤10 µm (PM_10_) or ≤2.5 µm (PM_2_.5). These pollutants directly degrade air quality and pose significant public health risks, and air quality ratings were calculated based on PM concentrations [2, 3]. Currently, epidemiological studies suggest that the primary air pollutant, particularly PM_2_.5, may be associated with the increasing morbidity of the global population by eliciting major adverse effects, as PM_2.5_ can penetrate into the deeper lung regions, directly access the systemic circulation, and breach the blood–brain barrier (BBB) [4, 5]. Accumulating studies, including our previous research, indicate that prolonged exposure to PM_2.5_ may give rise to respiratory and cardiovascular disorders, as well as systemic metabolic syndrome‐associated diseases, such as obesity, type 2 diabetes (T2D), and dyslipidemia‐related nonalcoholic fatty liver disease (NAFLD) [6, 7]. Mounting epidemiological evidence has established a link between PM_2.5_ inhalation and central nervous system (CNS) injuries, such as cerebral lesions, cognitive decline, behavioral abnormalities, Alzheimer's disease (AD), and Parkinson's disease (PD), through the induction of neuronal death, neuroinflammation, and oxidative stress [8, 9, 10]. Nevertheless, the implications of long‐term PM_2.5_ inhalation on the etiology and pathogenesis of the brain remain elusive.
PM_2.5_ exposure can induce insulin resistance, which is generally regarded as evolving due to inflammatory signaling in peripheral metabolic tissues such as adipose tissue, liver, and kidney [11, 12, 13]. The CNS, particularly the hypothalamus, is responsible for maintaining the homeostasis of energy and metabolism by integrating neural, nutritional, and hormonal signals [14]. Under metabolic stresses like a high‐fat diet (HFD), early hypothalamic inflammation participates in central resistance to fuel‐sensing hormones, disruption of homeostatic control of food intake, and alterations in energy expenditure, which have been focused on in both humans and rodent animals [15, 16]. In a time‐course experiment, peripheral inflammation occurs and develops after initiating a HFD; however, inflammation in the hypothalamus rapidly emerges within 24 h, preceding body weight gain and the peripheral inflammatory response [17]. Targeting hypothalamic inflammation and suppressing it exert protective effects against tissue injury, insulin resistance, and metabolic dysfunction, [18] suggesting that hypothalamic impairment is likely to precede metabolic stress‐related diseases. PM_2.5_ exposure can result in hypothalamic inflammation and hypothalamus‐pituitary‐adrenal (HPA) axis dysregulation in rodent animals, accompanied by abnormal hormone releases [19, 20]. Our previous research revealed that PM_2.5_ inhalation triggers glial activation and intense inflammation in the hypothalamus of mice by activating the nuclear factor‐κB (NF‐κB) p65 subunit signaling pathway [21, 22]. Moreover, PM_2.5_‐induced mice with liver injury and insulin resistance also exhibit more severe hypothalamic damage [23]. Hypothalamic POMC^+^ neurons play a crucial role in regulating feeding, body weight balance, energy expenditure, lipid, and glucose metabolism [24]. POMC is a protein that is expressed and secreted by POMC^+^ neurons and cleaved by prohormone convertases to generate α‐melanocyte‐stimulating hormone (α‐MSH). As a peptide released from POMC^+^ neurons, α‐MSH binds to the melanocortin 4 receptor (MC4R) and serves as a key hub connecting the CNS to peripheral organs [25, 26]. We recently provided evidence that continuous PM_2.5_ exposure decreased hypothalamic POMC expression and induced neuroinflammation, contributing to dyslipidemic function in liver and adipose tissues [21]. Nevertheless, the underlying pathogenesis of hypothalamic injury and associated metabolic imbalance triggered by long‐term PM2.5 inhalation remains largely unknown.
The tripartite motif (TRIM) family consists of a RING domain, 1‐2 B‐box domain and a coiled‐coil domain, and participates in various biological processes, such as inflammatory response, innate immunity, cell proliferation and death [27]. TRIM31, as a member of TRIM family, mediates viral infection, inflammatory response and tumor growth through regulating its target substrates [28, 29, 30]. For example, TRIM31 exacerbates myocardial dysfunction by enhancing apoptosis and the NF‐κB p65 subunit inflammatory pathway in lipopolysaccharide (LPS)‐induced sepsis [31]. TRIM31 also alleviates the activation of the NOD‐like receptor protein 3 (NLRP3) inflammasome by reinforcing the proteasomal degradation of NLRP3 [29]. A recent study by our group revealed that promoting TRIM31 significantly mitigated carbon tetrachloride (CCl_4_)‐induced hepatic inflammation, ROS production, and fibrosis [32]. Furthermore, targeting TRIM31 significantly improves HFD‐triggered fatty liver by suppressing hepatic inflammation and lipid accumulation through inhibiting rhomboid 5 homolog 2 (Rhbdf2) and mitogen‐activated protein kinase kinase kinase 7 (MAP3K7) [33, 34]. Deletion of specific Rhbdf2 in POMC^+^ neurons is found to reduce PM_2.5_‐triggered oxidative and inflammatory damage in the hypothalamus of mice [21]. These previous findings suggest a potential functional involvement of TRIM31 in PM_2.5_‐induced hypothalamic damage and metabolic dysfunction. Unfortunately, there is limited supporting data available for this hypothesis. Here in our present work, mice with specific TRIM31 knockout (*POMC^Cre/+^;TRIM31^flox/flox^ *) and overexpression mediated by AAV (AAV‐TRIM31) in hypothalamic POMC^+^ neurons were generated to investigate the biological effects of TRIM31 on PM_2.5_‐triggered hypothalamic lesions and associated pathological processes, thereby providing a promising target for the alleviation of metabolic diseases under air pollution conditions.
Results
2
TRIM31 Deficiency in POMC+ Neurons Accelerates Peripheral Metabolism Disorder in Mice after Long‐Term PM2.5 Inhalation
2.1
Mounting evidence has demonstrated the adverse impacts of long‐term PM_2.5_ inhalation on peripheral metabolic and CNS [5, 6, 7]. To further explore the pathological process and underlying mechanisms caused by air pollution, mice were exposed to PM_2.5_ (171 ± 2.5 µg/m^3^) for 24 weeks, which is equivalent to the air quality rating‐4 level in China and thereby mimics a real‐world ambient PM_2.5_ environment for humans [22]. Prior to conducting the experiments, composition of the collected PM_2.5_ was analyzed. As listed in Table S1 (Supporting Information), higher levels of sulfur, arsenic, and vanadium were detected than that of the filtered air (FA) components. Among all these elements, sulfur was identified as the predominant one. This analysis potentially suggested that the aberrant composition of the air could contribute to CNS injury and peripheral organ damage. TRIM31 plays a crucial role in mediating inflammatory response, lipid metabolic disorder, and multiple organ impairments [29, 31, 33]. Considering the potential effects of PM_2.5_ on neuroinflammation and metabolic disturbance, we hypothesized that TRIM31 expression changes might be involved in PM_2.5_‐triggered metabolic abnormalities and central lesion. Initially, our preliminary experiments demonstrated that PM_2.5_ inhalation significantly downregulated hypothalamic TRIM31 gene and protein expression levels in a time‐dependent manner (Figure S1A,B, Supporting Information). Consistently, PM_2.5_ treatment significantly decreased TRIM31 expression levels in both the hypothalamic neuron cell line GT1‐7 and the microglial cell line BV2 compared to the Con group (Figure S1C–F, Supporting Information). To further identify and confirm the types of cells in which TRIM31 was expressed, immunofluorescence (IF) staining was used to detect the specific expression of TRIM31 in hypothalamic tissues. As expected, colocalization of TRIM31 with NeuN was detected in hypothalamic tissue sections, revealing that neuronal cells expressed TRIM31 (Figure S1G, Supporting Information). Similarly, colocalization of TRIM31 with CD11b, a marker for microglial cells, was also observed in hypothalamus (Figure S1H, Supporting Information), suggesting the expression of TRIM31 in microglial cells. However, less TRIM31 was detected in astrocytes (Figure S1I, Supporting Information). These findings initially demonstrated the potential effects of TRIM31, particularly in neurons, on air pollution‐induced hypothalamic damage.
Considering the key role of POMC^+^ neurons in mediating feeding behavior and energy metabolism, we then generated POMC‐Cre‐dependent TRIM31 knockout mice to further investigate the effects of TRIM31 on PM_2.5_‐caused hypothalamic lesion and metabolic disturbance (Figure S2A, Supporting Information). Abundant positive staining for TRIM31 was observed in POMC^+^ neurons in the hypothalamus of *TRIM31^f/f^
- control mice. Nevertheless, TRIM31 staining in POMC^+^ neurons was almost undetectable in hypothalamus of *POMC^Cre/+^;TRIM31^flox/flox^
- mice, indicated by the few expressions of TRIM31 in arcuate (ARC) areas with α‐MSH staining (Figure S2B, Supporting Information), but was evident in the cortex tissues from both the *TRIM31^f/f^
- and *POMC^Cre/+^;TRIM31^flox/flox^
- groups (Figure S2C, Supporting Information). After FA and PM_2.5_ inhalation for 24 weeks, no significant differences in body weights were detected between the FA/*TRIM31^f/f^
- and PM_2.5_/*TRIM31^f/f^
- groups. Additionally, TRIM31 deletion in POMC^+^ neurons showed no significant effect on body weight changes compared to the PM_2.5_/*TRIM31^f/f^
- group; however, markedly decreased body weights were detected in *POMC^Cre/+^;TRIM31^flox/flox^
- mice compared to the FA/*TRIM31^f/f^
- mice after PM_2.5_ challenge (Figure 1A). Compared with FA‐treated mice, mean blood pressure (MBP) was strongly upregulated in PM_2.5_‐exposed mice, which was markedly aggravated upon hypothalamic TRIM31 loss in POMC^+^ neurons (Figure 1B). PM_2.5_ exposure had no significant influence on liver weight changes compared with FA group, but was enhanced in mice without TRIM31 expression. In PM_2.5_‐challenged mice, although TRIM31 loss in hypothalamic neurons decreased the liver weights, no significant difference was detected (Figure 1C). Except for the eWAT weights, PM_2.5_ did not show significant effect on the changes in fat weights, iWAT weights and BAT weights compared with the mice receiving FA. Notably, TRIM31 deletion in hypothalamic POMC^+^ neurons increased the fat weights and eWAT weights after PM_2.5_ treatment for 24 weeks but did not affect the alterations in iWAT weights and BAT weights (Figure 1D–G). Histological analysis showed that PM_2.5_ inhalation caused an enlargement of adipocyte area size in eWAT tissue sections, more severe inflammation scores, and lipid deposition in hepatic samples. Of note, the adipocyte size, inflammation score, and fat accumulation triggered by PM_2.5_ were strongly accelerated in *POMC^Cre/+^;TRIM31^flox/flox^
- mice (Figure 1H–K). Consistently, in response to a 24 week PM_2.5_ exposure, TRIM31 deletion remarkably facilitated hepatic lipid accumulation compared with the PM_2.5_/*TRIM31^f/f^
- group, as evidenced by the increased TG contents in liver (Figure 1L).
TRIM31 deficiency in POMC+ neurons accelerates peripheral metabolism disorder in mice after long‐term PM2.5 exposure. (A) Body weights of mice were measured. (B) MBP of each group of mice was recorded. Measurements of (C) liver weights, (D) fat weight, (E) eWAT weight, (F) iWAT weight, and (G) BAT weight from each group of mice were shown (n = 15 mice per group). (H) Images for H&E staining of WAT and liver tissues, and Oil red O staining of liver samples from all groups of mice (n = 6 mice per group; For each mouse, we randomly recorded 6 fields spanning the entire tissue; Scale bars: 50 µm). (I) Quantification for adipocyte area size following H&E staining (n = 6 mice per group). (J) Inflammation score of liver tissues was quantified (n = 6 mice per group). (K) Quantification of fat drop in liver sections of all groups of mice (n = 6 mice per group). (L) Hepatic TG contents were examined (n = 15 mice per group). Data are presented as means ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001; ns, no significant difference; Student's t test and one‐way ANOVAs statistics analysis for (A) to (G), and (I) to (L).
Next, we aimed to explore whether elevated TRIM31 levels would mitigate PM_2.5_‐induced events. To address this hypothesis, POMC‐Cre mice with specific TRIM31 overexpression in hypothalamic POMC^+^ neurons were generated by stereotaxic injections of AAV‐TRIM31 to the ARC nucleus of the hypothalamus (Figure S3A–C, Supporting Information). Western blotting analysis confirmed TRIM31 overexpression in hypothalamus of mice receiving AAV‐TRIM31 (Figure S3D, Supporting Information). No significant variation was detected in the changes of body weights, food intake, MBP and fat weights between the AAV‐Ctrl and AAV‐TRIM31 groups of mice (Figure S3E–H, Supporting Information), indicating that TRIM31 overexpression in hypothalamic POMC^+^ neurons did not affect the metabolic phenotypes of mice in the absence of any stresses. As depicted in Figure S4A–C (Supporting Information), following a 24 week PM_2.5_ inhalation, TRIM31 overexpression in hypothalamic POMC^+^ neurons did not exert any effect on the alterations in body and liver weights of mice; however, it significantly mitigated the increase in MBP. Decreased fat weights and eWAT weights were observed in mice with TRIM31 overexpression in hypothalamic POMC^+^ neurons, while no significant differences were identified in the alterations of iWAT and BAT weights (Figure S4D–G, Supporting Information). Importantly, PM_2.5_‐induced adipocyte size enlargement, hepatic inflammation and lipid accumulation were obviously ameliorated in mice with TRIM31 overexpression (Figure S4H–K, Supporting Information), along with reduced hepatic TG contents (Figure S4L, Supporting Information). Further studies by real time‐quantitative polymerase chain reaction (RT‐qPCR) indicated that compared with the FA/*TRIM31^f/f^
- mice, the expression levels of lipid metabolism‐associated signals in liver and eWAT were highly disturbed following PM_2.5_ inhalation, but were further accelerated upon TRIM31 ablation (Figure S5A,B, Supporting Information). In contrast, mice that received AAV‐TRIM31 exhibited opposite expression changes of these lipid metabolism‐associated genes in liver and eWAT samples (Figure S6A,B, Supporting Information). Together, these data demonstrated that PM_2.5_ indeed resulted in peripheral metabolism disorder, and this event could be exacerbated by TRIM31 ablation but ameliorated by TRIM31 overexpression in hypothalamic POMC^+^ neurons, elucidating the potential role of hypothalamic TRIM31 in PM_2.5_‐caused pathological phenotypes of metabolic disturbance.
TRIM31 Knockout in POMC+ Neurons Promotes Abnormity of Energy Expenditure and Glucose Metabolism in PM2.5‐Treated Mice
2.2
To evaluate the impact of TRIM31 expression variations on energy homeostasis, we monitored the metabolic parameters of PM_2.5_‐exposed mice using indirect calorimetry. As displayed in Figure 2A,B, a 24 week PM_2.5_ exposure significantly reduced VO_2_ compared with FA mice in both the light and dark phases. Additionally, mice challenged with PM_2.5_ exhibited significantly lower VCO_2_ compared to the FA‐treated mice (Figure 2C,D). Subsequently, we assessed respiratory exchange ratios (RER) based on VO_2_ and VCO_2_. PM_2.5_‐treated *POMC^Cre/+^;TRIM31^flox/flox^
- mice showed lower RER compared with mice expressing TRIM31 in the light and dark phases, indicating greater fat utilization (Figure 2E). In contrast, mice injected with AAV‐TRIM31 exhibited increased VO_2_, VCO_2_, and RER after long‐term PM_2.5_ treatment (Figure 2F–J). Moreover, in response to PM_2.5_ inhalation, energy expenditure (EE) was also markedly reduced in *POMC^Cre/+^;TRIM31^flox/flox^
- mice during the light and dark phases compared with the *TRIM31^f/f^
- group (Figure 2K,L), but was highly rescued in PM_2.5_/AAV‐TRIM31 mice (Figure 2M,N). Food intake did not show significant differences between the FA and PM_2.5_ groups regardless of hypothalamic TRIM31 expression (Figure 2O), which was similarly observed in the PM_2.5_/AAV‐Ctrl and PM_2.5_/AAV‐TRIM31 groups of mice (Figure 2P). We also noted higher levels of blood glucose and insulin in PM_2.5_‐treated mice than in the FA mice, which were strongly exacerbated upon TRIM31 deficiency (Figure 2Q), but were improved in mice receiving AAV‐TRIM31 injections (Figure 2R). Glucose tolerance test (GTT) and insulin tolerance test (ITT) tests revealed that compared to FA‐exposed mice, 24 weeks of PM_2.5_ treatment induced more severe glucose intolerance and insulin resistance, which were exacerbated in mice with TRIM31 deficiency, as evidenced by the elevated AUC (Figure 2S,T). Consistently, *POMC^Cre/+^;TRIM31^flox/flox^
- mice also exhibited strikingly higher levels of serum leptin after long‐term PM_2.5_ challenge than those in the PM_2.5_/*TRIM31^f/f^
- group, while the opposite variation trend was detected in serum adiponectin (Figure 2U). Reversely, promoting TRIM31 expression downregulated serum leptin levels and improved adiponectin contents in mice with PM_2.5_ exposure, accompanied by attenuated glucose intolerance and insulin resistance (Figure 2V–X). Together, these findings suggested that TRIM31 deletion might be implicated in abnormal energy expenditure and glucose metabolism processes caused by PM_2.5_.
Effects of TRIM31 in POMC+ neurons on the energy expenditure and glucose metabolism in PM2.5‐exposed mice. (A) VO2 across 48 h monitoring in TRIM31f/f and POMCCre/+;TRIM31flox/flox mice (n = 8 mice for FA‐treated groups; n = 10 mice for PM2.5‐treated groups). (B) Average quantification of VO2 in light and dark cycles of TRIM31f/f and POMCCre/+;TRIM31flox/flox mice (n = 8 mice for FA‐treated groups; n = 10 mice for PM2.5‐treated groups). (C) VCO2 production across 48 h monitoring in TRIM31f/f and POMCCre/+;TRIM31flox/flox mice (n = 8 mice for FA‐treated groups; n = 10 mice for PM2.5‐treated groups). (D) Averages of VCO2 production in light and dark cycles of TRIM31f/f and POMCCre/+;TRIM31flox/flox mice (n = 8 mice for FA‐treated groups; n = 10 mice for PM2.5‐treated groups). (E) Average RER of each group of mice indicated in (A) during light and dark cycles (n = 8 mice for FA‐treated groups; n = 10 mice for PM2.5‐treated groups). (F) VO2 across 48 h monitoring (n = 11 mice for AAV‐Ctrl group; n = 12 mice for AAV‐TRIM31 group). (G) Average quantification of VO2 in light and dark cycles (n = 11 mice for AAV‐Ctrl group; n = 12 mice for AAV‐TRIM31 group). (H) VCO2 production across 48 h monitoring (n = 11 mice for AAV‐Ctrl group; n = 12 mice for AAV‐TRIM31 group). (I) Averages of VCO2 production in light and dark cycles (n = 11 mice for AAV‐Ctrl group; n = 12 mice for AAV‐TRIM31 group). (J) Average RER of each group of mice indicated in (F) during light and dark cycles (n = 11 mice for AAV‐Ctrl group; n = 12 mice for AAV‐TRIM31 group). (K) EE of all groups of mice demonstrated in A) over a 48 h period (n = 8 mice for FA‐treated groups; n = 10 mice for PM2.5‐treated TRIM31f/f and POMCCre/+;TRIM31flox/flox mice). (L) Averages EE in light/dark cycles (n = 8 mice for FA‐treated groups; n = 10 mice for PM2.5‐treated groups). (M,N) EE of all groups of mice over a 48 h period (n = 11 mice for AAV‐Ctrl group; n = 12 mice for AAV‐TRIM31 group). (O) Records of food intake were shown (left panel), and averages of food intake were quantified (right panel) (n = 8 mice for FA‐treated groups; n = 10 mice for PM2.5‐treated groups). (P) Records of food intake were shown (n = 11 mice for AAV‐Ctrl group; n = 12 mice for AAV‐TRIM31 group). (Q) Blood glucose levels of all groups of mice were measured at the shown time points (left panel); and the final fasting blood insulin contents were calculated after PM2.5 exposure for 24 weeks (right panel) in TRIM31f/f and POMCCre/+;TRIM31flox/flox mice (n = 15 mice per group). (R) Blood glucose levels of all groups of mice were measured (right panel); and fasting blood insulin contents were examined after PM2.5 exposure for 24 weeks (right panel) (n = 11 mice for AAV‐Ctrl group; n = 12 mice for AAV‐TRIM31 group). (S) GTT and (T) ITT were conducted in TRIM31f/f and POMCCre/+;TRIM31flox/flox mice after a total of 24 week PM2.5 exposure. AUC for GTT and ITT was quantified (n = 15 mice per group). (U) Examination of serum leptin (left panel) and adiponectin (right panel) levels in TRIM31f/f and POMCCre/+;TRIM31flox/flox mice after FA or PM2.5 exposure (n = 15 mice per group). (V) Serum leptin and adiponectin contents of AAV‐Ctrl and AAV‐TRIM31 mice were calculated after PM2.5 exposure for 24 weeks (n = 8 mice for each group). (W) GTT and (X) ITT were conducted in AAV‐Ctrl and AAV‐TRIM31 mice after a total of 24 week PM2.5 exposure. AUC of each group mice was quantified (n = 7 or 8 mice per group). Data are presented as means ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001; ns, no significant difference; Student's t test and one‐way ANOVAs statistics analysis for (A) to (E), (L), (O), (Q), and (S) to (U); Student's t test analysis for (F) to (J), (N), (P), (R), and (V) to (X).
TRIM31 Deletion in POMC+ Neurons Facilitates Neuronal Injury in Hypothalamus of PM2.5‐Exposed Mice
2.3
As the center for regulating energy homeostasis, the hypothalamus was subsequently examined in mice under PM_2.5_ exposure. Using IF staining, we found a lower number of NeuN‐positive cells in hypothalamic ventromedial (VMH) and ARC areas in PM_2.5_‐challenged mice compared to the FA group; however, TRIM31 deletion in POMC neurons exerted suppressive effects on NeuN expression, particularly in ARC areas, which was comparable with the PM_2.5_/*TRIM31^f/f^
- group (Figure 3A,B), revealing the more serious injury of neurons. POMC, MC4R, NPY, and AgRP are key factors in mediating energy homeostasis [25, 26, 35]. As depicted in Figure 3C,D, PM_2.5_ exposure led to apparent reductions of POMC‐positive cells in the hypothalamus of mice, and notably, this outcome was strikingly accelerated when TRIM31 was deleted. Likewise, RT‐qPCR further confirmed that *POMC^Cre/+^;TRIM31^flox/flox^
- mice exhibited significant decrease of POMC mRNA expression levels compared to the *TRIM31^f/f^
- group of mice following PM_2.5_ exposure (Figure 3E). Additionally, enzyme‐linked immunosorbent assay (ELISA) analysis showed that PM_2.5_ inhalation markedly decreased the release of hypothalamic α‐MSH compared with the FA/*TRIM31^f/f^
- mice, which was, however, further downregulated in PM_2.5_‐challenged mice with TRIM31 deficiency in POMC^+^ neurons (Figure 3F). Immunoblotting assays ascertained that PM_2.5_‐induced reduction in protein expression levels of POMC was further suppressed in hypothalamus of *POMC^Cre/+^;TRIM31^flox/flox^
- mice (Figure 3G). POMC^+^ neurons project from the ARC to paraventricular nucleus of the hypothalamus (PVH) to release α‐MSH, which is involved in the regulation of food appetite and energy metabolism, in part, through its agonist activity on the MC4R melanocortin receptor expressed in the PVH [36, 37]. Subsequently, we found that PM_2.5_ exposure markedly reduced MC4R expression levels in hypothalamic sections, whereas being further weakened in PM_2.5_‐exposed *POMC^Cre/+^;TRIM31^flox/flox^
- mice (Figure S7A,B, Supporting Information). RT‐qPCR and western blotting assays confirmed that TRIM31 deficiency in POMC^+^ neurons significantly reduced MC4R mRNA and protein expression levels in hypothalamus samples, respectively (Figure S7C,D, Supporting Information). However, the reduced NeuN expression levels due to PM_2.5_ exposure were reversed in mice injected with AAV‐TRIM31 (Figure S8A,B, Supporting Information). Additionally, mice overexpressing TRIM31 in hypothalamic POMC^+^ neurons displayed higher expression of POMC following PM_2.5_ exposure by IF (Figure S8C,D, Supporting Information). Rescued POMC mRNA, α‐MSH contents, and POMC protein levels were detected in the hypothalamus of AAV‐TRIM31 mice after PM_2.5_ treatment (Figure S8E–G, Supporting Information). As expected, promoting TRIM31 robustly restored MC4R^+^ positive expression levels in the hypothalamus (Figure S8H,I, Supporting Information), which were confirmed by RT‐qPCR and western blot assays (Figure S8J,K, Supporting Information). These aforementioned results further delineated the potential protective function of TRIM31 in improving POMC^+^ and MC4R^+^ neurons, as manifested by the augmented hypothalamic POMC, α‐MSH, and MC4R levels, thereby contributing to the improvement of energy expenditure and peripheral metabolism. Furthermore, hypothalamic NPY gene and protein expression levels were strongly induced in PM_2.5_‐exposed mice, while remaining unaffected in mice with hypothalamic TRIM31 expression loss. Meanwhile, AgRP expression did not change, regardless of PM_2.5_ inhalation or TRIM31 knockout (Figure S9A,B, Supporting Information). Collectively, these aforementioned findings revealed that TRIM31 might be crucial for the amelioration of neuron injury and the subsequent mitigation of associated energy metabolism disorders caused by air pollution.
TRIM31 deletion in POMC+ neurons facilitates neuronal injury in hypothalamus of PM2.5‐exposed mice. (A) IF staining of NeuN in hypothalamus sections from mice after PM2.5 exposure for 24 weeks (n = 6 mice per group; for each mouse, we recorded 2 fields spanning the entire hypothalamus; scale bars: 100 µm). (B) Quantification for NeuN‐positive expression in VMH (up panel) and ARC (down panel) areas, respectively, was exhibited (n = 6 mice per group). (C) IF staining of hypothalamic POMC from all groups of mice (n = 6 mice per group; For each mouse, we recorded 2 fields spanning the entire hypothalamus; Scale bars: 100 µm). (D) POMC positive expression were quantified (n = 6 mice per group). (E) POMC gene expression levels in hypothalamus tissues were evaluated by RT‐qPCR (n = 4 mice per group). (F) ELISA results for hypothalamic ɑ‐MSH levels (n = 5 mice per group). (G) POMC protein expression levels in hypothalamus tissues by western blot assays (n = 4 mice per group). Data are presented as means ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001; ns, no significant difference; Student's t test and one‐way ANOVAs statistics analysis for (B), (D), (E) to (G).
TRIM31 Deficiency in POMC+ Neurons Enhances Neuroinflammation in Hypothalamus of PM2.5‐Challenged Mice
2.4
The inflammatory response may be involved in hypothalamic dysfunction under metabolic stress, [38] and was therefore examined in this study. As shown in Figure 4A, PM_2.5_ long‐term inhalation markedly increased the releases of IL‐1β, IL‐18, IL‐6, TNF‐ɑ, and MCP1 in serum, indicating a severe inflammatory response compared to FA mice. However, TRIM31 deficiency greatly facilitated the systematic releases of these mentioned inflammatory factors induced by PM_2.5_. GFAP and CD11b are markers of glial cells, whose activation is closely associated with neuroinflammation [39]. IF staining and western blot results showed that 24 weeks of PM_2.5_ inhalation significantly upregulated GFAP and CD11b positive expression levels in the hypothalamus compared to the FA mice (Figure 4B,C and Figure S9C–E, Supporting Information). Notably, significantly decreased serum IL‐1β, IL‐18, IL‐6, and TNF‐ɑ contents were detected in PM_2.5_/AAV‐TRIM31 mice (Figure S10A, Supporting Information). However, these expression levels were markedly intensified in mice with TRIM31 ablation and significantly attenuated in mice overexpressing TRIM31 in POMC^+^ neurons (Figure S10B–F, Supporting Information), indicating that the activation of astrocytes and microglial cells could be regulated by TRIM31 expression changes. As expected, *POMC^Cre/+^;TRIM31^flox/flox^
- mice after long‐term PM_2.5_ treatment exhibited significantly higher mRNA expression levels of inflammatory genes such as IL‐1β, IL‐18, IL‐6, TNF‐ɑ, MCP1, iNOS, CXCL2, and ICAM1 in hypothalamus than those of the PM_2.5_/*TRIM31^f/f^
- mice (Figure 4D). In contrast, PM_2.5_/AAV‐TRIM31 mice exhibited opposite expression changes in these inflammatory factors (Figure S10G, Supporting Information). The IκBα/NF‐κB p65 pathway, detected via phosphorylation of IκBα (p‐IκBα) and NF‐κB p65 (p‐NF‐κB p65), is a critical mediator of hypothalamic inflammation. Its activation by stressors or immune challenges triggers the production of proinflammatory cytokines, promotes neuronal damage, and underlies metabolic pathologies associated with hypothalamic dysfunction [22, 32, 40, 41]. Western blot results indicated that the expression levels of hypothalamic p‐IκBα and p‐NF‐κB p65 subunit were strongly elevated in PM_2.5_‐challenged mice, and were further exacerbated in TRIM31‐deficient animals (Figures 4E,F). In contrast, significantly ameliorated results were observed in mice injected with AAV‐TRIM31 (Figure S10H, Supporting Information). The data presented above clarified that the loss of TRIM31 contributed to hypothalamic inflammation induced by air pollution.
TRIM31 deficiency in POMC+ neurons enhances hypothalamic inflammation in PM2.5‐exposed mice. (A) Serum IL‐1β, IL‐18, IL‐6, TNF‐ɑ, and MCP1 levels were assessed by ELISA analysis (n = 15 mice per group). (B) IF staining for hypothalamic CD11b expression (n = 6 mice per group; for each mouse, we recorded 2 fields spanning the entire hypothalamus; scale bars: 50 µm). (C) Cells with CD11b positive expression were quantified (n = 6 mice per group). (D) RT‐qPCR results of inflammatory molecules including IL‐1β, IL‐18, IL‐6, TNF‐ɑ, MCP1, iNOS, CXCL2, and ICAM1 in hypothalamus tissues were shown (n = 3 mice in each). (E,F) Western blot results for p‐IκBɑ and p‐NF‐κB p65 subunit in hypothalamus were displayed (n = 4 mice per group). Data are presented as means ± SD. ** p < 0.01, *** p < 0.001, and **** p < 0.0001; Student's t test and one‐way ANOVAs statistics analysis for (C) and (F); two‐way ANOVAs statistics for (A) and (D).
TRIM31 Knockout in POMC+ Neurons Aggravates Hypothalamic Oxidative Stress After PM2.5 Inhalation
2.5
Oxidative stress is a crucial factor involved in PM_2.5_‐induced pathogenesis and tissue injury [3, 12, 40, 42]. As anticipated, serum and hypothalamic malondialdehyde (MDA) contents were significantly induced in mice challenged with PM_2.5_, while superoxide dismutase (SOD) and glutathione (GSH), as antioxidants, were strongly downregulated both in serum and hypothalamus. Notably, these events resulting from PM_2.5_ inhalation were further exacerbated in the absence of TRIM31 (Figure 5A–F), but were strongly attenuated in AAV/TRIM31 mice (Figure S11A–D, Supporting Information). Consistently, PM_2.5_ treatment caused a significant increase in hypothalamic ROS production, and *POMC^Cre/+^;TRIM31^flox/flox^
- mice exhibited higher ROS levels by dihydroethidium (DHE) staining (Figure 5G,H). In contrast, promoting TRIM31 significantly reduced ROS in hypothalamus of PM_2.5_‐treated mice (Figure S11E,F, Supporting Information). RT‐qPCR analysis indicated that hypothalamic expression levels of HO‐1, NQO‐1, GCLM, SOD1, and SOD2 genes were markedly restrained by PM_2.5_, but were further accelerated in mice with TRIM31 ablation (Figure 5I). Opposite results were monitored in the changes of NOX2, NOX4, and XO upon TRIM31 loss (Figure 5J). Of note, the expression levels of these oxidative stress‐related signals were significantly reversed by AAV‐TRIM31 in PM_2.5_‐exposed mice (Figure S11G,H, Supporting Information). Immunoblot analysis showed that PM_2.5_ resulted in significantly lower expression of nuclear Nrf2 but higher cytoplasmic Nrf2 in the hypothalamus compared to the FA group; however, these results were exacerbated when TRIM31 was knocked out (Figure 5K). Restored nuclear Nrf2 expression was detected in the hypothalamus of PM_2.5_/AAV‐TRIM31 mice (Figure S11I, Supporting Information). Collectively, these findings revealed that TRIM31 deletion‐accelerated hypothalamus lesion might be partially attributed to oxidative stress with Nrf2 signaling blockade.
TRIM31 knockout in POMC+ neurons aggravates hypothalamic oxidative stress after PM2.5 exposure. (A) MDA, (B) SOD, and (C) GSH in serum were evaluated (n = 15 mice per group). Hypothalamic (D) MDA, (E) SOD, and (F) GSH levels were assessed (n = 15 mice in each). (G) Representative images of DHE staining in hypothalamus were shown (n = 6 mice per group; For each mouse, we recorded 2 fields spanning the entire hypothalamus; Scale bars: 50 µm). (H) ROS production by DHE staining was quantified (n = 6 mice per group). (I,J) RT‐qPCR analysis for the hypothalamic gene expression levels of molecules associated with oxidative stress as displayed (n = 3 mice per group). (K) Western blot for cytoplastic and nuclear Nrf2 protein expression in hypothalamus (n = 4 mice in each). Data are presented as means ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001; Student's t test and one‐way ANOVAs statistics analysis for (A) to (F), (H), and (K); two‐way ANOVAs statistics for (I) and (J).
TRIM31 Knockdown Promotes Neuron Death and ROS Production upon PM2.5 Stimulation In Vitro
2.6
The regulatory role of TRIM31 in PM_2.5_‐triggered hypothalamus injury and metabolic disorder was further explored using the mouse hypothalamic neuron cell line GT1‐7 and microglial cell line BV2. GT1‐7 cells, a type of GnRH neuronal cell line, are specifically derived from the hypothalamus, and plays a crucial role in mediating energy consumption and metabolism. GnRH neurons can interact with other hypothalamic circuits involved in appetite regulation and influence the balance between hunger and satiety, thereby modulating food intake and energy expenditure. Additionally, GnRH signal may impact metabolic rate through regulating the activity of various endocrine and neural pathways that control energy homeostasis. Some studies suggest that GnRH may play a role in adipocyte function and lipid metabolism. Moreover, the hypothalamic GnRH neuronal cell line is sensitive to changes in energy reserves and thus can adjust its activity accordingly. GnRH may also affect insulin signaling and glucose metabolism, contributing to overall energy balance [43, 44, 45, 46]. Given to these reasons, GT1‐7 was used to further explore the effects of TRIM31 on neuronal impairments and dysfunctions in response to PM_2.5_. First, TRIM31 expression was silenced in GT1‐7 cells through transfection with si‐TRIM31, and si‐NC was used as a control. The successful TRIM31 knockdown in GT1‐7 cells was confirmed by RT‐qPCR and western blot assays (Figure S12A,B, Supporting Information). Cell Counting Kit‐8 (CCK‐8) analysis in Figures 6A,B indicated that PM_2.5_ treatment dose‐ and time‐dependently decreased the cell viability of GT1‐7 cells, and TRIM31 knockdown further aggravated cell death. Similarly, LDH releases were strongly triggered in PM_2.5_‐incubated cells compared to the Con group, and cells with TRIM31 deletion exerted much higher LDH than the si‐NC/PM_2.5_ group (Figure 6C). ROS induced by PM_2.5_ exposure was further accelerated when TRIM31 was deleted, as indicated by 2′,7′‐dichlorofluorescein‐diacetate (DCF‐DA) staining (Figures 6D,E), accompanied with enhanced MDA contents (Figure 6F). In addition, TRIM31 knockdown markedly downregulated SOD and catalase (CAT) activity in GT1‐7 cells upon PM_2.5_ incubation (Figures 6G,H). PM_2.5_‐reduced expression of HO‐1, NQO‐1, GCLM, SOD1, and SOD2 was strongly restrained after siTRIM31 transfection (Figure 6I), and opposite expression changes were detected in NOX2, NOX4, and XO (Figure 6J). Furthermore, nuclear Nrf2 protein expression levels were statistically repressed in PM_2.5_‐stimulated GT1‐7 cells, whereas being further weakened after TRIM31 silence. Meanwhile, much higher cytoplastic Nrf2 expression was observed in PM_2.5_‐treated cells lacking TRIM31 expression than that of the PM_2.5_/si‐NC group (Figure 6K). To verify the regulatory effects of TRIM31 on oxidative stress, TRIM31 was then overexpressed in GT1‐7 cells by infection with the constructed Ad‐TRIM31, and Ad‐GFP was used as the control (Figure S12C,D, Supporting Information). CCK‐8 results showed that promoting TRIM31 significantly rescued the GT1‐7 cell viability influenced by PM_2.5_ (Figure S13A, Supporting Information). However, PM_2.5_‐triggered LDH releases was strongly abolished upon TRIM31 overexpression (Figure S13B, Supporting Information). Ad‐TRIM31 infection remarkably suppressed ROS production in PM_2.5_‐exposed GT1‐7 cells, accompanied by decreased MDA levels and increased SOD and CAT activities (Figure S13C–G, Supporting Information). Furthermore, the expression levels of HO‐1, NQO‐1, GCLM, SOD1, and SOD2 restrained by PM_2.5_ were highly restored by TRIM31 overexpression (Figure S13H, Supporting Information), and opposite findings were detected in the expression changes of NOX2, NOX4, and XO (Figure S13I, Supporting Information). As expected, PM_2.5_‐induced increase in cytoplastic Nrf2 was abrogated in GT1‐7 cells overexpressing TRIM31; however, nuclear Nrf2 was evidently rescued in Ad‐TRIM31 cells after PM_2.5_ treatment (Figure S13J, Supporting Information). Collectively, these in vitro findings confirmed that TRIM31 might ameliorate oxidative stress, contributing to the improvement of PM_2.5_‐caused neuron injury.
TRIM31 knockdown promotes neuron death and ROS production upon PM2.5 stimulation in vitro. (A) After transfection with si‐TRIM31 for 24 h, GT1‐7 cells were exposed to PM2.5 incubation for 24 h at the shown concentrations, and were then collected for cell viability evaluation by CCK‐8 (n = 6 per group). (B) GT1‐7 cells were transfected with si‐TRIM31 for 24 h, followed by PM2.5 (50 µg/ml) treatment for the indicated time (0 to 48 h). Then, all cells were used for CCK‐8 analysis (n = 6 per group). (C–J) GT1‐7 cells were subjected to si‐TRIM31 transfection. After 24 h, GT1‐7 cells were cultured with PM2.5 (50 µg/mL) for another 24 h. Subsequently, all cells were harvested for studies as follows. (C) Calculation of LDH releases (n = 6 per group). (D) ROS production evaluation using DCF‐DA staining; n = 6 per group; For each sample, we randomly recorded 8 fields spanning the entire well; Scale bars: 15 µm). (E) Quantification of ROS generation was shown. (F) MDA, (G) SOD, and (H) CAT were assessed (n = 6 in each). (I,J) RT‐qPCR results for genes associated with oxidative stress as listed (n = 3 per group). (K) Western blot analysis for Nrf2 expression in cytoplasm and nucleus, respectively (n = 4 per group). Data are presented as means ± SD. ** p < 0.01, *** p < 0.001, and **** p < 0.0001; Student's t test and one‐way ANOVAs statistics analysis for (A) to (C), (E) to (H), and (K); two‐way ANOVAs statistics for (I) and (J).
TRIM31 Upregulation Protects against Inflammatory Response in PM2.5‐Exposed Microglial Cells
2.7
Given the crucial role of microglial cells in regulating neuroinflammation, the mouse microglial cell line BV2 was subsequently employed and subjected to si‐TRIM31 transfection and Ad‐TRIM31 infection for its knockdown and overexpression, respectively (Figure S12E–H, Supporting Information). Iba1, serving as a marker of microglial cells, was highly induced by PM_2.5_ treatment, as indicated by the elevated Iba1 gene expression and fluorescence intensity, which were further accelerated upon TRIM31 ablation (Figure S14A–C, Supporting Information), revealing the promoted microglial cell activation. On the contrary, TRIM31 overexpression remarkably mitigated PM_2.5_‐activated microglial cells, proved by the decreased Iba1 (Figure S14D–F, Supporting Information). ELISA results demonstrated that PM_2.5_ caused significant increases in the proinflammatory cytokines releases, including IL‐1β, IL‐6, and TNF‐ɑ, whereas being further accelerated in GT1‐7 cells with TRIM31 knockdown (Figure 7A). Intriguingly, overexpressing TRIM31 markedly reduced IL‐1β, IL‐6, and TNF‐ɑ secretion post PM_2.5_ incubation (Figure 7B). As expected, si‐TRIM31 strongly strengthened the gene expression levels of IL‐1β, IL‐18, IL‐6, TNF‐ɑ, and MCP1 induced by PM_2.5_ (Figure 7C). Notably, PM_2.5_‐triggered upregulation of these inflammatory genes was significantly abrogated upon TRIM31 overexpression (Figure 7D). Nuclear NF‐κB p65 subunit transition was apparently induced by PM_2.5_, and was further promoted upon TRIM31 deletion; however, TRIM31 overexpression diminished PM_2.5_‐enhanced NF‐κB p65 subunit expression in the nucleus of BV2 cells (Figures 7E,F). Western blotting analysis then indicated that the expression of the p‐NF‐κB p65 subunit was markedly elevated in BV2 cells after PM_2.5_ treatment compared with the Con group, which were aggravated by si‐TRIM31 but was diminished by Ad‐TRIM31 (Figure 7G). Together, these findings elucidated that TRIM31 might have a negative influence on microglial cell activation and inflammatory response induced by PM_2.5_.
TRIM31 upregulation protects against inflammatory response in PM2.5‐stimulated microglial cells. BV2 cells were subjected to si‐TRIM31 transfection or Ad‐TRIM31 infection for TRIM31 knockdown and overexpression, respectively. After 24 h, BV2 cells were stimulated by PM2.5 (50 µg/mL) for an additional 24 h. Next, all cells and culture medium were harvested for subsequent experiments. (A,B) ELISA analysis for IL‐1β, IL‐6, and TNF‐ɑ secretion in the collected supernatants (n = 6 per group). (C,D) RT‐qPCR for inflammatory factors, including IL‐1β, IL‐18, IL‐6, TNF‐ɑ, and MCP1 (n = 3 per group). (E,F) IF staining for NF‐κB p65 subunit expression in PM2.5‐treated cells with TRIM31 knockdown or overexpression. Nuclear NF‐κB p65 subunit expression was quantified (n = 4 per group; for each sample, we randomly recorded 8 fields spanning the entire well; scale bars: 20 µm). (G) Western blot analysis for p‐NF‐κB p65 subunit protein expression levels in cells (n = 4 in each). Data are presented as means ± SD. ** p < 0.01, *** p < 0.001, and **** p < 0.0001; Student's t test and one‐way ANOVAs statistics analysis for (A), (F), and (G); one‐way ANOVAs statistics for (B); two‐way ANOVAs statistics for (C) and (D).
TRIM31 Increase Ameliorates Oxidative Stress in Microglial Cells After PM2.5 Culture
2.8
Oxidative stress in microglial cells contributes to its activation, and thus was explored in our study. PM_2.5_ incubation evidently led to ROS generation in BV2 cells, and si‐TRIM31 cells showed higher ROS in comparison to the si‐NC/PM_2.5_ group (Figure S15A, Supporting Information). Moreover, BV2 cells with TRIM31 overexpression exerted lower ROS after PM_2.5_ exposure, as evidenced by the weaker fluorescence of DCF‐DA staining (Figure S15B, Supporting Information). In line with ROS alterations, si‐TRIM31 cells showed higher MDA levels after PM_2.5_ incubation than the si‐NC/PM_2.5_ group (Figure S15C, Supporting Information). Ad‐TRIM31 significantly reduced PM_2.5_‐caused MDA levels compared with Ad‐GFP/PM_2.5_ group (Figure S15D, Supporting Information). Reversely, SOD activity decreased by PM_2.5_ was further downregulated upon TRIM31 deletion, but was highly rescued when TRIM31 was overexpressed (Figure S15E,F, Supporting Information). As expected, PM_2.5_‐restrained expression levels of HO‐1 and NQO‐1 were further inhibited in BV2 cells with TRIM31 silence, while being strongly restored by Ad‐TRIM31; however, opposite phenotype was observed in NOX2 and NOX4 expression changes (Figure S15G,H, Supporting Information). These in vitro results elucidated that PM_2.5_ could induce oxidative stress in microglial cells, and might be attenuated by TRIM31.
TRIM31 Upregulation in Microglial Cells Improves Neuronal Survival under PM2.5 Stress
2.9
Activation of microglial cells and the resultant inflammatory response contribute to neuronal injury under various stressors [47]. To further disclose the underlying molecular mechanisms modulated by TRIM31, we collected culture medium from PM_2.5_‐exposed BV2 cells with or without TRIM31 expression, which was then mixed with fresh medium at a 1:3 ratio to strike a critical balance between inducing measurable neuronal injury, preserving nutrient availability, and ensuring physiological relevance based on previous literatures and results of our pre‐experiments (data were not shown in the study) [48, 49]. Subsequently, the conditioned medium (CM) was collected and used for GT1‐7 cell incubation (Figure 8A). Of note, the CCK‐8 results demonstrated that CM derived from PM_2.5_‐stimulated BV2 cells resulted in a significant reduction in the viability of GT1‐7 cells. However, this was facilitated by CM derived from BV2 cells with TRIM31 deletion and rescued by CM obtained from microglial cells with TRIM31 overexpression (Figure 8B). Consistently, LDH releases were highly elicited by PM_2.5_‐treated BV2 cell CM, and were aggravated in GT1‐7 cells cultured with CM from TRIM31‐knockdown BV2 cells. After exposure to CM from Ad‐TRIM31 infected BV2 cells, GT1‐7 cells exerted lower LDH releases than those in the CM/Ad‐GFP/PM_2.5_ group (Figure 8C). As expected, CM from PM_2.5_‐incubated BV2 cells markedly increased ROS generation and MDA levels in GT1‐7 cells. These events were accelerated upon TRIM31 ablation in BV2 cells, but were almost diminished upon TRIM31 overexpression in BV2 cells (Figure 8D–F), accompanied by seemingly restored SOD and CAT activities (Figure 8G,H). Furthermore, the expression levels of HO‐1, NQO‐1, GCLM, SOD1, SOD2, and nuclear Nrf2 were remarkably downregulated in GT1‐7 cells exposed to CM derived from PM_2.5_‐stimulated BV2 cells, along with upregulated NOX2, NOX4, XO, and cytoplastic Nrf2 expression levels through RT‐qPCR and western blot analysis; however, these results were further exacerbated in GT1‐7 cells exposed to CM harvested from BV2 cells with si‐TRIM31 transfection, while being strongly ameliorated by CM obtained from BV2 cells overexpressing TRIM31 (Figure 8I–K). Together, these in vitro findings newly suggested that TRIM31‐alleviated microglial cell activation and inflammatory response might contribute to the survival and oxidative stress suppression of GT1‐7 cells.
TRIM31 upregulation in microglial cells improves neuronal survival under PM2.5 stress. (A) Experimental protocols to explore the effects of TRIM31 expression changes in BV2 cells on neuronal survival with or without PM2.5 treatment. (B–I) BV2 cells with TRIM31 knockdown or overexpression were treated with PM2.5 (50 µg/mL) for 24 h. Subsequently, BV2 cell culture medium was collected and mixed with fresh culture medium at 1:3 ratio. The mixed conditional medium was then used for GT1‐7 cell culture for 24 h. Next, all GT1‐7 cells were collected for studies as follows. (B) Measurement of viable cells using CCK‐8 analysis (n = 6 per group). (C) LDH releases were tested (n = 6 per group). (D,E) DCF‐DA staining was performed to examine ROS production (n = 6 per group; for each sample, we randomly recorded 8 fields spanning the entire well; scale bars: 15 µm). (F) MDA levels, (G) SOD activity, and (H) CAT activation were examined (n = 6 per group). (I,J) Expression of oxidative stress‐related molecules were evaluated by RT‐qPCR (n = 3 per group). (K) Nrf2 protein expression in cytoplasm and nucleus was evaluated via western blot (n = 4 per group). Data are presented as means ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001; Student's t test and one‐way ANOVAs statistics analysis for (B), (C), (D) to (H), and (K); two‐way ANOVAs statistics for (I) and (J).
TRIM31 Directly Interacts with Nrf2 and Induces Its K63‐linked Polyubiquitination
2.10
We previously identified a potential interaction between TRIM31 and Nrf2, [32] and found a deeper reduction of Nrf2 activation in the hypothalamus of *POMC^Cre/+^;TRIM31^flox/flox^
- mice after PM_2.5_ treatment (Figure 5K), but restoration in mice overexpressing TRIM31 in hypothalamic POMC^+^ neurons (Figure S11I, Supporting Information). Based on these observations, we speculated that there might be a physical interaction between TRIM31 and Nrf2 during PM_2.5_‐initialized hypothalamic injury and metabolic disturbances. To explore the hypothesis, co‐immunoprecipitation (Co‐IP) assays were then performed. As shown in Figure 9A,B, a specific interaction between TRIM31 and Nrf2 in GT1‐7 and BV2 cells was identified by Co‐IP analysis. Glutathione S‐transferase (GST) pull‐down analysis further confirmed the direct interaction between TRIM31 and Nrf2 (Figure 9C,D). The interaction between TRIM31 and Nrf2 was subsequently verified in the hypothalamus of mice challenged with a 24 week PM_2.5_ (Figure 9E). Moreover, the upregulation of Nrf2 was dose‐dependently induced by TRIM31 (Figure 9F). Double IF staining displayed a colocalization of TRIM31 and Nrf2 in GT1‐7 and BV2 cells (Figure 9G,H). We further found that Nrf2 showed a series of high‐molecular weight species in cells treated with the proteasome inhibitor MG‐132, indicating that Nrf2 could be ubiquitinated and verifying that the stability of the target protein Nrf2was regulated by the proteasome (Figure 9I). Transfection of the TRIM31 plasmid into GT1‐7 and BV2 cells revealed that TRIM31 enhanced the polyubiquitination of Nrf2 (Figure 9J). More studies indicated that TRIM31 reduced the evident K48‐linked polyubiquitination but enhanced the K63‐linked polyubiquitination of Nrf2 under PM_2.5_ stresses (Figure 9K,L). In contrast, TRIM31 silence further promoted K48‐linked polyubiquitination and restrained K63‐linked polyubiquitination of Nrf2 in PM_2.5_‐treated GT1‐7 and BV2 cells (Figure 9M,N). Given that MG132, a protease inhibitor, can assist in enhancing the ubiquitination effect, we subsequently conducted experiments to determine the impact of TRIM31 on the ubiquitination status of Nrf2 in cells without MG132 treatment to circumvent potential interference resulting from MG132 addition. Consistently, we also found that in the absence of MG132, TRIM31 overexpression could indeed induce Nrf2 ubiquitination (Figure S16A, Supporting Information). As expected, TRIM31 mitigated the K48‐linked polyubiquitination while promoting the K63‐linked Nrf2 polyubiquitination upon PM_2.5_ stimulation (Figure S16B,C, Supporting Information). However, TRIM31 deficiency facilitated K48‐linked polyubiquitination and suppressed K63‐linked polyubiquitination of Nrf2 in PM_2.5_‐incubated GT1‐7 and BV2 cells (Figure S16D,E, Supporting Information). These findings revealed that TRIM31‐induced K63‐linked Nrf2 ubiquitination was independent of MG132 intervention. Together, all these results identified a direct interaction between TRIM31 and Nrf2, and TRIM31 could induce a K63‐linked polyubiquitination of Nrf2 for its activation induction.
TRIM31 directly interacts with Nrf2 and induces its K63‐linked polyubiquitination. (A,B) Co‐IP of TRIM31 and Nrf2. (A) GT1‐7 and (B) BV2 cells were transfected with Flag‐tagged TRIM31 and HA‐tagged Nrf2. The lysates were immunoprecipitated with anti‐Flag or anti‐HA and analyzed using immunoblotting with the indicated anti‐Flag or anti‐HA antibodies (three independent experiments). GST precipitation and immunoblotting analysis indicating the direct interaction of TRIM31 with Nrf2 using purified GST‐Nrf2 and His‐tagged TRIM31 (left) or purified GST‐TRIM31 and His‐Nrf2 (right) in (C) GT1‐7 and (D) BV2 cells. GST was served as a control (three independent experiments). (E) Immunoprecipitation and western blot analysis showing the binding of TRIM31 to Nrf2 in hypothalamus of the wild type mice after a 24 week PM2.5 exposure. IgG was defined as a control (three independent experiments). (F) Western blotting analysis of Nrf2 expression in GT1‐7 and BV2 cells transfected with varying amounts of Flag‐tagged TRIM31 (0, 125, 250, and 500 ng) incubated with PM2.5 (50 µg/mL) for 24 h (three independent experiments). (G,H) Colocalization of TRIM31 and Nrf2 in GT1‐7 and BV2 cells (n = 6 per group; For each sample, we randomly recorded 8 fields spanning the entire well; Scale bar, 15 µm). (I) Nrf2 immunoprecipitation reaction was conducted in GT1‐7 and BV2 cells treated with or without MG‐132 (5 µm) and subjected to immunoblotting analysis to detect Nrf2 ubiquitination (Ub) (three independent experiments). (J) GT1‐7 and BV2 cells were transfected with TRIM31 plasmid 48 h to detect the effects of TRIM31 on Nrf2 ubiquitination in the presence of MG‐132 (5 µm) (three independent experiments). (K,L) TRIM31 reversed the 24 h of PM2.5 (50 µg/mL)‐induced increased K48‐linked polyubiquitination and decreased K63‐linked polyubiquitination of Nrf2 in GT1‐7 and BV2 cells with MG‐132 (5 µm) treatment (three independent experiments). (M,N) TRIM31 deletion increased the 24 h of PM2.5 (50 µg/mL)‐induced up‐regulation of K48‐linked polyubiquitination and down‐regulation of K63‐linked polyubiquitination of Nrf2 in GT1‐7 and BV2 cells in the presence of MG‐132 (5 µm) (three independent experiments).
Nrf2 was Indispensable for TRIM31 to Play Its Neuroprotective, Anti‐inflammatory and Antioxidant Effects Caused by PM2.5
2.11
To further explore whether Nrf2 was necessary for TRIM31 to ameliorate neuron death, inflammation and oxidative stress, Nrf2 was then knocked down in GT1‐7 and BV2 cells to determine the potential of TRIM31/Nrf2 axis in air pollution‐induced hypothalamic injury. RT‐qPCR and immunoblot assays indicated that Nrf2 gene and protein expression levels were efficiently silenced in GT1‐7 and BV2 cells after transfection with si‐Nrf2, particularly si‐Nrf2‐2# (Figure S17A–D, Supporting Information), which was subsequently utilized in the following in vitro studies. CCK‐8 analysis in Figure S18A (Supporting Information) indicated that TRIM31‐improved cell viability in GT1‐7 cells was significantly abolished when Nrf2 expression was depleted, while LDH releases were strongly restored under PM_2.5_ exposure (Figure S18B, Supporting Information). Furthermore, the suppression of ROS and MDA levels by Ad‐TRIM31 was markedly reversed by si‐Nrf2 after PM_2.5_ incubation (Figure S18C–E, Supporting Information); however, we detected opposite findings in the alterations of SOD and CAT activities (Figure S18F,G, Supporting Information). Additionally, the restored expression levels of HO‐1, NQO‐1, GCLM, SOD1, and SOD2 by TRIM31 were almost abolished by Nrf2 deficiency after PM_2.5_ incubation (Figure S18H, Supporting Information), while the expression levels of NOX2, NOX4, and XO were recovered (Figure S18I, Supporting Information). These data revealed that Nrf2 was indispensable for TRIM31 to restrain PM_2.5_‐caused ROS generation.
Likewise, the inhibition of ROS by Ad‐TRIM31 was reversed in Nrf2 knockdown BV2 cells following PM_2.5_ treatment, along with increased MDA levels (Figure S19A,B, Supporting Information). In contrast, the enhancement of SOD, NQO‐1, and HO‐1 by TRIM31 overexpression were strongly diminished in PM_2.5_‐stimulated Nrf2‐deleted BV2 cells (Figure S19C,D, Supporting Information). Additionally, the inflammatory molecules, such as IL‐1β, IL‐6, TNF‐α, IL‐18, and MCP1, which were suppressed by Ad‐TRIM31, were significantly restored in Nrf2‐deleted BV2 cells after PM_2.5_ exposure, along with higher nuclear NF‐κB p65 subunit transition (Figure S19E–G, Supporting Information). Collectively, these results demonstrated that the TRIM31‐mediated repression of the inflammatory response induced by air pollution was Nrf2‐dependent.
Discussion
3
The major components of air pollution PM_2.5_ have the potential to give rise to metabolic dysfunctions, such as atherosclerosis, insulin resistance, lipid and glucose metabolism disorder, as well as cardiovascular disorders [5, 6, 7, 12, 19]. The pathogenesis of PM_2.5_‐induced tissue injury, including liver, kidney, heart, and brain, has been widely studied through irritating metabolic syndrome‐like physiological and pathological phenomena, dyslipidemia, inflammation, and oxidative stress mediated by diverse signaling pathways [12, 13, 22, 35]. Air pollution could potentially permeate directly through the BBB, stimulate the activation of glial cells, and subsequently engender an augmented production of free radicals within the peripheral regions, ultimately contributing to CNS injury and an imbalance in homeostasis [4, 5, 19, 21, 22]. The hypothalamus is accountable for numerous metabolic regulations and its related diseases, which are etiologically connected with the deregulations of hypothalamus neurons that are vulnerable to oxidative stress and inflammatory responses [15, 16, 18, 21]. For example, leptin receptor‐expressing neurons (LEPR) in the hypothalamic nuclei participates can regulate food intake and EE. The hypothalamus is also involved in insulin signaling and glucose metabolism. Disruptions in hypothalamic function can lead to insulin resistance, resulting in elevated blood glucose levels [50, 51]. Furthermore, the hypothalamus can affect lipid metabolism in the liver. Dysregulated hypothalamic activity may contribute to an increase in hepatic lipid deposition, which is associated with metabolic disorders such as fatty liver disease [52]. Increasing studies have identified that PM_2.5_ inhalation can induce hypothalamic injury, contributing to the detected incidences of peripheral metabolism disorders [19, 20, 21]. Nevertheless, the underlying molecular mechanisms governing these occurrences remain elusive.
Sulfate is one of the major constituents of PM_2.5_ in China. Generally, SO_2_ from fossil fuel combustion is a primary precursor of sulfate aerosols in PM_2_.5. These emissions dominate urban and industrial regions, where oxidation reactions convert gaseous SO_2_ into particulate sulfate [53]. Additionally, sulfur compounds exhibit high stability under varying atmospheric conditions. Sulfur‐bearing molecules form persistent complexes in particulate matter due to low volatility and hygroscopic properties, facilitating their accumulation in PM_2_.5. Moreover, meteorological factors accelerate the gas‐to‐particle conversion of sulfur species. For instance, SO_2_ oxidizes to sulfate via photochemical reactions or cloud processing, significantly amplifying PM_2_.5 mass despite lower initial emission volumes compared to other constituents. Besides, sulfur‐derived aerosols resist decomposition, leading to their relative enrichment [53, 54, 55]. In our present work, sulfur emerged as the predominant constituent among all chemical elements present in ambient PM_2.5_, having been recognized as a key component implicated in PM2.5‐induced illnesses [56]. The PM_2.5_ concentration of 171 ± 2.5 µg/m^3^ was equivalent to the air quality rating‐4 level in China standard, and was utilized for in vivo animal experiments to simulate the real‐world ambient PM_2.5_ surrounding for human. TRIM31, a member of the TRIM family, fulfills a crucial role in diverse biological processes, including cell proliferation and death, inflammatory responses, and oxidative stress, [28, 29, 30, 32] all of which are closely associated with PM_2.5_‐induced neuronal disorders [8, 9, 10, 21, 22, 23]. These prior discoveries position TRIM31 as a potentially significant therapeutic target, which might be efficient for alleviating long‐term PM_2.5_ exposure‐induced hypothalamic damage and the metabolic disturbances it governs.
Here, our in vivo and in vitro results initially uncovered that TRIM31 was significantly decreased in hypothalamus, hypothalamic neuron cell line GT1‐7 and microglial BV2 cells under PM_2.5_ treatment. Western blotting and IF staining results further demonstrated that both neurons and microglial cells expressed TRIM31, particularly in neurons. Neuronal injury can result in the activation of microglia, which subsequently release proinflammatory cytokines, chemokines, and reactive oxygen species, contributing to the inflammatory response and tissue damage [57]. Hence, the activation of microglial under PM_2.5_ may be a secondary effect closely related to the expression changes of TRIM31 in POMC^+^ neurons. Moreover, astrocytes could be activated upon prolonged exposure to PM_2.5_. Although we discovered that TRIM31 expression alterations in POMC neurons exerted a significant impact on astrocyte activation under PM_2.5_ inhalation, there was less TRIM31 expression in astrocytes. Astrocytes can be activated in response to neuronal damage, and the activated astrocytes undergo morphological and functional modifications [58]. In addition, astrocytes can interact with microglial cells, which are the immune cells of the CNS. Activation of microglial cells can affect astrocytes and consequently the overall immune response in the brain [59]. Therefore, we hypothesized that neuronal injury and activated microglial cells might contribute to the activation of astrocytes under PM_2.5_ stimulation, which was significantly mediated by the expression changes of TRIM31 in POMC neurons. In conclusion, these data unveiled the potential of TRIM31 downregulation in PM_2.5_‐induced hypothalamic dysfunction. Mice with TRIM31 knockout or overexpression in hypothalamic POMC^+^ neurons were established, followed by PM_2.5_ inhalation for a total of 24 weeks. We found that long‐period PM_2.5_ inhalation indeed caused metabolic disorders in mice, indicated by the increased MBP, fat weight, adipocyte area enlargement and hepatic lipid deposition, which were strongly accelerated in mice with TRIM31 loss, whereas being ameliorated in mice receiving AAV‐TRIM31 injection. Moreover, energy expenditure disturbance, blood glucose elevation and insulin resistance were detected in PM_2.5_‐challenged mice, accompanied with enhanced serum insulin and leptin levels but decreased adiponectin contents; however, all these events caused by PM_2.5_ were further exacerbated by TRIM31 deficiency, but were attenuated in mice overexpressing TRIM31. These in vivo findings uncovered that TRIM31 downregulation was implicated in PM_2.5_‐induced peripheral metabolism disorder. We then confirmed that a 24 week PM_2.5_ exposure significantly reduced the overall number of hypothalamic neurons in mice and caused hypothalamic injury, proved by the decreased NeuN expression levels. Additionally, energy expenditure‐related signals POMC and MC4R were highly downregulated in the hypothalamus, whereas NPY expression levels were upregulated. Intriguingly, TRIM31 loss aggravated the reduced expression of POMC and MC4R, but showed no significant influence on NPY. These results manifested that PM_2.5_ exposure directly influenced the expression of POMC and α‐MSH by acting on associated neurons to subsequently mediate energy consumption and metabolic processes, rather than simply causing a decrease in the total number of neurons. In addition to these abnormalities, *POMC^Cre/+^;TRIM31^flox/flox^
- mice were more sensitive to the chronic effects of PM_2.5_ on hypothalamic injury, proved by greater glial activation, neuroinflammation and oxidative stress. At the cellular level, PM_2.5_ stimulation contributed to neuronal cytotoxicity and ROS production, inflammatory response and oxidative stress in microglial cells. These cellular results were markedly accelerated in cells with TRIM31 knockdown, but were dramatically ameliorated by TRIM31 overexpression. Intriguingly, we showed that CM collected from PM_2.5_‐incubated microglial cells led to neuronal death and ROS generation, and this event was facilitated in hypothalamic neurons cultured in CM derived from PM_2.5_‐exposed BV2 cells with TRIM31 knockdown. These results elucidated that TRIM31‐mediated microglial activation was implicated in neuronal survival under PM_2.5_ stress. Mechanistically, we discovered a direct interaction between TRIM31 and Nrf2. TRIM31 enhanced K63‐linked and reduced K48‐linked Nrf2 polyubiquitination in neurons and microglial cells, contributing to Nrf2 activation. More intriguingly, TRIM31‐ameliorated ROS generation and inflammatory response elicited by PM_2.5_ were almost abrogated in Nrf2‐deficient GT1‐7 and BV2 cells, disclosing that Nrf2 was required for TRIM31 to play its neuroprotective, anti‐inflammatory and anti‐oxidant functions. Collectively, all our data provided solid evidence that the TRIM31/Nrf2 axis was implicated in hypothalamus damage caused by PM_2.5_, and no doubt further explaining the homeostatic control of peripheral metabolism (Figure 10).
Schematic working model showing the molecular mechanism underlying TRIM31‐mediated activation of Nrf2 signaling in the progression of PM2.5‐induced hypothalamus injury and peripheral metabolic disorder. Long‐term PM2.5 exposure led to neuronal death, inflammatory response and oxidative stress in hypothalamus, contributing to peripheral metabolic disturbance, which was closely associated with the decrease of TRIM31. Severe inflammation in microglial cells participated in neuronal death under PM2.5 stress. TRIM31 overexpression inhibited neuronal damage, inflammation and ROS generation mainly through improving the activation of Nrf2. TRIM31 directly interacted with Nrf2, increased its K63‐linked polyubiquitination and decreased its K48‐linked polyubiquitination, leading to the enhanced Nrf2 activation, which was needed for TRIM31 to perform its neuroprotective, anti‐inflammatory and antioxidant effects against PM2.5‐induced hypothalamic injury and metabolic disorder.
PM_2.5_ inhalation leads to a range of biological effects and adverse effects in metabolic organs and tissues, such as liver, kidney, muscle, adipose tissue, and brain [11, 12, 13, 21, 22]. Changes in the liver triggered by PM_2.5_ include hepatocyte injury, inflammatory cell infiltration, impairment of liver structure, lipid accumulation, and even fibrosis [11, 12, 34, 60]. Adipose tissue is also a critical modulator of metabolic homeostasis, and its disfunction is closely related to the deteriorated health status in models undergoing PM exposure [11, 61]. Exploring insulin resistance, lipid and glucose metabolism imbalances, energy expenditure, and their association with PM_2.5_ may help explain the causal relationships between air pollution and cardiovascular risk progression [2, 3, 4, 12, 23]. Particulate matter has been recognized as a key factor triggering inflammatory diseases. The passage of PM_2.5_ through the nasal route is regarded as the entry portal. Through this route, the development of local and systemic inflammation occurs, including the secretion of proinflammatory cytokines and chemokines such as IL‐1β, IL‐6, and MCP1, which contribute to neurodegenerative diseases [62, 63]. Furthermore, another way to incorporate the particulate matter, particularly the ultrafine fractions, is through neurons in the olfactory epithelium, promoting the inflammatory response and potentially disrupting the BBB [64]. Systemic inflammation‐induced increased permeability of the BBB enhances the migration of antigens and proinflammatory molecules into the neuroglial environment [65]. In the hypothalamus, PM_2.5_ induces membrane peroxidation and activates the microglia response. PM_2.5_ can also stimulate the expression of TNF‐α and IKK‐β in the hypothalamus and NF‐κB in the paraventricular nucleus [66, 67]. In our study, we supposed that the main entry of PM_2.5_ should be via neurons in the olfactory epithelium, directly accessing the CNS, which was supported by the significant upregulation of inflammatory markers in the hypothalamus after long‐term exposure to PM_2.5_. Inflammatory response and oxidative stress caused by PM_2.5_ may elicit maladaptive responses, which, in turn, adversely influence organ function [21, 22, 23, 40, 44]. Hypothalamic POMC^+^ neurons can sense the energy status of the organism to integrate peripheral signals, including leptin, glucose and insulin [24, 68, 69]. POMC^+^ neurons, in turn, coordinate responses to control energy balance and glucose homeostasis [70]. Leptin signaling through POMC^+^ neurons promote a striking anorexigenic effect while increasing EE. Using genetic engineering technology, removing elements of insulin, leptin or glucose signaling cascades in hypothalamic POMC^+^ neurons impair glucose tolerance and insulin sensitivity in murine animals [71, 72]. Although the effects of PM_2.5_‐triggered hypothalamic damage and metabolic dysfunctions have been reported, it remains unclear whether hypothalamic POMC^+^ neurons can be directly involved. We previously identified that conditional knockout of Rhbdf2 in POMC^+^ neurons attenuated PM_2.5_ exposure‐caused neurological impairment and neuronal loss by confronting oxidative stress and inflammation in mice [21]. TRIM31 has been recognized to play pivotal roles in inflammation and ROS generation under various pathological conditions [32]. For instance, we recently realized that TRIM31 expression gradually decreased in the liver of mice fed a HFD, and its absence markedly accelerated overnutrition‐induced hepatic lipid deposition and inflammatory response via suppressing Rhbdf2 expression, accompanied with facilitated insulin resistance. TRIM31 can directly bind to Rhbdf2 and facilitate its proteasomal degradation and expression suppression, ultimately mitigating fatty liver and associated metabolic disorders [33]. These previous findings revealed the potential of TRIM31 in hypothalamic POMC^+^ neurons during PM_2.5_‐induced hypothalamus injury and metabolic disturbance. In our study, we first confirmed that long‐term PM_2.5_ exposure resulted in metabolic syndrome‐like phenotypes, including higher blood glucose, insulin and leptin levels, and elevated insulin resistance. Furthermore, fat weight, WAT weight, adipocyte size and hepatic lipid accumulation were strongly induced in PM_2.5_‐challenged mice. It should be noted that after PM_2.5_ inhalation, mice with specific TRIM31 knockout in POMC^+^ neurons displayed more severe blood pressure, fat weight gain and hepatic lipid accumulation than the *TRIM31^f/f^
- mice, accompanied by stronger deterioration of insulin and leptin sensitivity. Moreover, PM_2.5_‐decreased energy expenditure was prolonged in *POMC^Cre/+^;TRIM31^flox/flox^
- mice. However, TRIM31 deletion in POMC^+^ neurons did not affect the changes in body weights and food intake. Similarly, POMC‐specific deficiency of TRIM31 led to an accelerated inflammatory response and oxidative stress in the hypothalamus of PM_2.5_‐challenged mice, thereby aggravating hypothalamic injury. Conversely, TRIM31 overexpression in hypothalamic POMC^+^ neurons via stereotactic injections significantly ameliorated PM_2.5_‐induced hypothalamic neuron damage, neuroinflammation and oxidative stress, along with improved peripheral metabolic dysfunctions. The neuroprotective, anti‐inflammatory and antioxidant functions of TRIM31 were further confirmed in PM_2.5_‐treated GT1‐7 and microglial cells. Therefore, we concluded that the loss of TRIM31 in hypothalamic POMC^+^ neurons might be a key pathological factor and sufficient to recapitulate the physiological consequences of PM_2.5_ exposure through exacerbating hypothalamic inflammation and oxidative damage. Furthermore, the abnormalities in energy expenditure, insulin resistance, and metabolic syndrome‐like phenotypes controlled by TRIM31 in POMC^+^ neurons were independent of changes in body weight or food intake.
The CNS, especially the hypothalamus, plays a crucial role in controlling glucolipid metabolism‐associated hormone secretion, food intake and basal EE [14, 15, 16]. As previously noted, severe inflammation and astrocyte activation were observed in hypothalamus of mice after long‐term PM_2.5_ inhalation via activating NF‐κB signaling [21, 22]. In this study, PM_2.5_ also strongly reduced energy expenditure in both the light and dark cycles, as indicated by the reduced VO_2_ and VCO_2_ production with lower RER and/or EE, which might be involved in the upregulation of WAT weight. Consistent with peripheral metabolism disorders and energy expenditure imbalances, the hypothalamus was significantly damaged in mice after 24 week PM_2.5_ inhalation, as evidenced by the reduction in NeuN‐positive cells. The neuron population coexpressing NPY and AgRP can be stimulated by negative energy balance. Upon activation, hypothalamic NPY/AgRP neurons suppress POMC^+^ neurons and oppose central MC4R signaling through the release of AgRP, ultimately promoting food intake and reducing energy expenditure [24, 35, 68, 69]. Although the dysregulation of metabolic phenotypes via the aberrant expression of hypothalamic neuropeptides caused by PM_2.5_ exposure has been reported, there are few reports on the inhibitory effects of air pollution on MC4R [19, 73]. POMC^+^ neurons project from the ARC to PVH to release α‐MSH, which participates in the modulation of feeding, energy expenditure, and metabolism through its agonist activity on MC4R^+^ expressed in the PVH [40, 41]. Mice overexpressing α‐MSH display increased EE, and the effects of MC4R agonists on energy homeostasis have been evaluated in animal studies [74, 75]. Additionally, the binding of α‐MSH with melanocortin receptors activates the cyclic AMP (cAMP) signaling pathway, resulting in various cellular responses. In the liver, activation of cAMP signaling inhibits the expression of genes involved in fatty acid synthesis, reducing lipid accumulation [76]. POMC^+^ neurons can also stimulate lipolysis in WAT by enhancing the breakdown of triglycerides into free fatty acids [77]. POMC^+^ neurons receive and integrate peripheral signals, such as leptin and insulin, to fine‐tune lipid homeostasis [70, 71, 72]. These metabolic adaptations can affect systemic energy balance, redirecting energy substrate toward storage. In our study, we found that after 24 weeks of PM_2.5_ treatment, the expression levels of POMC and MC4R in the hypothalamus of mice were significantly weakened, along with a decrease in hypothalamic α‐MSH content. Based on previous reports and our results, we supposed that TRIM31 knockout in POMC^+^ neurons led to a reduction in hypothalamic α‐MSH release, ultimately exacerbating energy expenditure and peripheral metabolism, as partially evidenced by fat weight gain and hepatic lipid storage. Therefore, we suggested that PM_2.5_‐induced metabolic disorders might be largely attributed to neuronal population changes in the hypothalamus. However, different with the downregulation of POMC and α‐MSH caused by PM_2.5_, no significant differences in body weight and food intake changes were detected between the FA and PM_2.5_ groups of mice, which was contrary to the expected outcome based on previous studies that POMC/α‐MSH, which acts on melanocortin receptors in the hypothalamus, can reduce appetite and food intake. As for this, we supposed that PM_2.5_ exposure may have compensatory effects on other neural circuits or molecules that regulate feeding behavior. In brief, PM_2.5_ exposure may activate alternative pathways that counteract the inhibitory effects of reduced POMC and α‐MSH levels on appetite. PM_2.5_ exposure may also have different effects on feeding behavior depending on the physiological state of the animal, such as during fasting or after a high‐fat diet. To further explore these potential explanations, future studies need to measure feeding behavior and related metabolites in other tissues to identify the effects of PM_2.5_ exposure on neural circuits and molecules involved in feeding regulation, as well as its context‐dependent effects on POMC/α‐MSH levels and feeding behavior. Together, despite these drawbacks, our study provided crucial insights that PM_2.5_‐triggered abnormal energy expenditure and peripheral metabolism dysfunctions such as glucolipid metabolism associated with POMC/α‐MSH might be independent of body weight gain and food consumption. Surprisingly, TRIM31 deletion in hypothalamic POMC^+^ neurons further accelerated NPY/AgRP upregulation and POMC/α‐MSH/MC4R downregulation in the hypothalamus of PM_2.5_‐exposed mice, contributing to the further reduction of energy expenditure. On the contrary, PM_2.5_‐caused abnormal alterations of NPY/AgRP and POMC/α‐MSH/MC4R were remarkably ameliorated in mice receiving AAV‐TRIM31, accompanied by improved energy expenditure. Although glucolipid metabolism and energy balance impairments are tightly associated with body weight changes and TRIM31 loss as we indicated, we could not establish a correlation between TRIM31 in mice and body weight status. However, considering the significant effects of TRIM31 expression changes in POMC^+^ neurons on MC4R in hypothalamic PVH areas, there remains a limitation in our current work regarding whether TRIM31 absence in PVN MC4R^+^ neurons could directly impact hypothalamic functions, feeding, energy expenditure, and metabolic disturbances caused by air pollution. Regarding this, more studies are indeed required in the future by utilizing MC4R‐Cre strains with conditional TRIM31 knockout.
PM_2.5_ can enter organs through the blood circulation and cause injury. Inflammatory response and oxidative stress are two critical pathological factors implicated in PM_2.5_‐triggered organ and tissue injury, both central and peripheral [3, 8, 12, 13, 21, 22, 23, 40, 42]. Human and animal studies confirm that PM_2.5_ exposure can trigger systemic inflammation that is involved in metabolic disorders [6, 7, 12, 13]. In our study, we consistently identified that long‐term inhalation of PM_2.5_ led to higher contents of serum inflammatory factors, such as IL‐1β, IL‐6, IL‐18, TNF‐α, and MCP1. Excessive activation of astrocytes and microglial cells is associated with neuroinflammation under various pathological conditions, such as overnutrition and air pollution [21, 22, 38]. GFAP and CD11b, as hallmarks of glial cells, were found to be highly induced by PM_2.5_ in the hypothalamus, manifesting excessive astrocyte and microglia activation. Similarly, the facilitated expression levels of numerous inflammatory genes (IL‐1β, IL‐6, IL‐18, TNF‐α, MCP1, iNOS, CXCL2, and ICAM1) were monitored in the hypothalamus of PM_2.5_‐exposed mice. These mentioned inflammatory factors can be modulated by various pathways in inflammatory diseases, particularly the IκB/NF‐κB signaling pathway [21, 22, 31, 41]. Previous studies, including those from our group, confirmed that PM_2.5_ can enhance the expression of NF‐κB family genes and their downstream inflammatory signals [22, 78]. In addition to neuroinflammation, oxidative stress and ROS production were evidently provoked by PM_2.5_ in the serum and hypothalamus of mice. Oxidative stress is a signal that is tightly associated with neuronal damage and death [8, 9, 10]. MDA and the Nrf2/HO‐1 signaling pathway are crucial elements involved in oxidative stress and antioxidant balance. As a transcription factor, Nrf2 mediates the expression of various antioxidants and cytoprotective proteins and enzymes, thereby exerting protective effects against oxidative damage [32, 79]. In our study, higher MDA contents in the hypothalamus of PM_2.5_‐treated mice were detected, accompanied by lower SOD and GSH activities. Moreover, decreased expression of antioxidants HO‐1, NQO‐1, GCLM, SOD1, and SOD2 was noted in the PM_2.5_‐intervened hypothalamus, while genes (NOX2, NOX4, and XO) that induce ROS generation were greatly upregulated. The excessive ROS generation caused by PM_2.5_ was at least in part due to the reduction of nuclear Nrf2 expression, contributing to hypothalamic oxidative damage. Neuronal TRIM31 regulates the degradation process of TIGAR, resulting in mitochondria malfunction and promoting ROS generation, which contributes to cerebral ischemic injury [80]. TRIM31 inhibits NLRP3 inflammasome activation by inducing proteasomal NLRP3 degradation [29]. Recently, our group found that genetic or pharmacological intervention to promote TRIM31 expression significantly reduces CCl_4_‐induced hepatic lesion by suppressing the inflammatory response and ROS generation through modulating the NF‐κB and Nrf2 signaling pathways, respectively [32]. In agreement with recent investigations and studies, we confirmed here that *POMC^Cre/+^;TRIM31^flox/flox^
- mice exhibited more severe hypothalamic inflammation and oxidative stress than wild‐type mice after PM_2.5_ exposure by facilitating the NF‐κB p65 subunit activation and Nrf2 blockade, while AAV‐TRIM31 mice showed reduced neuroinflammation and oxidative stress. Our in vitro experiments verified the function of TRIM31 loss in accelerating the inflammatory response in BV2 cells by promoting the nuclear transition of the NF‐κB p65 subunit after PM_2.5_ incubation. Meanwhile, the hypothalamic cell line GT1‐7 with TRIM31 knockdown showed more severe cell death and ROS generation by decreasing nuclear Nrf2 expression. Taken together, these findings enhanced our understanding of PM_2.5_‐induced neurological diseases, and the novel biological function of TRIM31 in the hypothalamus under air pollution stress is newly detected.
Microglial cells, as the resident macrophages in the CNS, are responsive to peripheral signals and responsible for monitoring and maintaining the homeostatic environment through surveillance. However, if this physiological protective function is not properly regulated, it can lead to aberrant microglial activation, which partially contributes to neurotoxicity through promoting the release of inflammatory cytokines or chemokines mediated by the NF‐κB signaling pathway [47, 78, 81, 82]. Therefore, microglia–neuron junctions provide a communication channel through which microglia can constantly monitor neuronal activation status and health, and respond accordingly when disruptions occur. The exacerbated metabolic disorders and energy metabolism abnormalities due to hypothalamic POMC neuronal damage and irregular regulation are closely related to the increased proinflammatory factors under pathological conditions, accompanied by detectable hyperactivation of microglial cells [21, 83]. HFD‐induced hypothalamic inflammatory response synchronizes with microglial activation, characterized by rapid microglial morphology changes and inflammatory molecules. The process can be detected prior to evident body weight changes, [84] revealing that microglial activation is properly an upstream process to the systemic metabolic alterations. Besides overnutrition, other stimuli including air pollutants can induce microglial cells to undergo dynamic changes in shape and function, [85] which attribute to the incidence and progression of stroke, cognitive function impair, depressive‐like symptom and other neurological diseases [86]. Importantly, microglia can affect neuronal function through releasing various different factors, including multiple cytokines (IL‐1β, IL‐6, TNF‐α) in response to different irritants [87]. In our present study, we noticed higher secretion of inflammatory cytokines IL‐1β, IL‐6, and TNF‐ɑ in the culture supernatants derived from PM_2.5_‐treated BV2 cells, which coincided with increased microglial activation labeled by Iba1; however, BV2 cells with TRIM31 deletion showed significantly higher releases of these inflammatory factors after PM_2.5_ incubation, whereas TRIM31 overexpression exerted suppressive effects on inflammation and BV2 cell activation. Consistent with previous reports, we further demonstrated here that the CM harvested from PM_2.5_‐stimulated BV2 cells indeed inhibited the survival of hypothalamic neuronal GT1‐7 cells, but caused greater LDH releases and oxidative stress. Notably, these adverse events were exacerbated in GT1‐7 cells cultured in CM from PM_2.5_‐incubated BV2 cells with TRIM31 ablation, while being significantly alleviated in GT1‐7 cells incubated with CM from PM_2.5_‐exposed BV2 cells overexpressing TRIM31. Collectively, these data suggested that there might be a mutual link between neuronal injury and microglial activation mediated by TRIM31 in response to PM_2.5_ stimuli, leading to severe oxidative stress and neuroinflammation in the hypothalamus, which consequently resulted in hypothalamic dysfunction and related energy and peripheral metabolic disorders. Moreover, previous studies have also reported that long‐term exposure to PM_2.5_ can induce the aberrant activation of astrocytes in brain [21, 88]. Here, we consistently identified similar effects of PM_2.5_ on astrocytes, as evidenced by the higher GFAP expression in hypothalamic sections. Notably, although TRIM31 was less expressed in astrocytes, it exerts suppressive effects on the activation of astrocytes in mice after long‐term air pollution challenge. Thus, we speculated that TRIM31 might regulate astrocyte activity in an indirect manner under PM_2.5_ associated with neuronal injury and aberrant activation of microglial cells, [57, 58, 59] further contributing to hypothalamic injury and associated metabolism disturbance. From a therapeutic perspective, it would be interesting to investigate the underlying mechanism by which TRIM31 controls the activation of astrocytes during PM_2.5_‐induced hypothalamic damage in the near future, given the known communication pathways between microglia, neurons, and astrocytes [89].
We previously identified a potential correlation between TRIM31 and Nrf2 in hepatocytes and hepatic stellate cells, which is involved in the pathology of hepatic fibrosis [32]. Here, our mechanistic studies by Co‐IP and GST assays consistently revealed a direct interaction between TRIM31 and Nrf2 in GT1‐7 and BV2 cells, which was confirmed by double IF staining. Ubiquitination is mediated by ubiquitin enzymes and deubiquitinating enzymes. As a crucial molecular mechanism, it regulates numerous cellular and biological processes [90]. The exposed lysines at positions 29, 48, and 63 of ubiquitin serve as modification sites, which can form polyubiquitin chains on target proteins. K48‐linked ubiquitin chains control target proteins for proteasome degradation, while K63‐linked ubiquitin chains regulate various nondegradable actions, such as signal transduction and protein–protein interactions [91]. A previous study reported that IκB kinase facilitates the ubiquitination and degradation of Nrf2 by phosphorylating the deubiquitinating enzyme cylindromatosis (CYLD), thereby accelerating oxidative stress damage in obesity‐associated nephropathy [92]. TRIM16 interacts with Nrf2 and induces K63‐linked Nrf2 ubiquitination, thereby stabilizing and activating Nrf2 [93]. Additionally, the deubiquitinase‐3 (DUB3) promotes Nrf2 stability and transcriptional activity by reducing its K48‐linked ubiquitination [94]. In our study, we confirmed that Nrf2 could be ubiquitinated. Importantly, TRIM31 significantly increased K63‐linked ubiquitination of Nrf2 and decreased its K48‐linked ubiquitination under PM_2.5_ stresses, thereby contributing to Nrf2 transcriptional activity to exert its antioxidant functions. In addition to the regulation of oxidative stress, the role of Nrf2 in mediating acute and chronic neuroinflammation has also been demonstrated in previous studies [95]. Targeting Nrf2 to enhance its activation can suppress microglial activity and brain inflammation, thereby alleviating CNS injury [96]. Given the impact of TRIM31 on Nrf2 activation, we further knocked down Nrf2 in GT1‐7 and BV2 cells to deeply explore the underlying mechanism. Interestingly, Nrf2 silencing almost abolished the biological functions of TRIM31 to improve neuronal conditions, inhibit ROS production, and suppress the inflammatory response caused by PM_2.5_ in vitro. These data notably suggested that Nrf2 signaling was essential for TRIM31 to exert its neuroprotective, antioxidant, and anti‐inflammatory effects, thereby alleviating PM_2.5_‐induced hypothalamic injury and metabolic dysfunctions.
In summary, we provided a previously unrecognized biological information of TRIM31 on how PM_2.5_ affected the hypothalamus and associated metabolism disorders. Results from our work uncovered TRIM31‐regulated POMC^+^ neuronal loss, inflammatory response and oxidative stress during the progression of hypothalamic injury through mediating Nrf2 polyubiquitination, which subsequently contributed to the pathogenesis of peripheral metabolic disturbance induced by long‐term PM_2.5_ inhalation. Thus, TRIM31/Nrf2 axis may be a promising avenue for the treatment of air pollution‐induced neurological and metabolic diseases. Regrettably, our present work only provided initial evidence about the underlying mechanism. The possibly precise upstream pathways leading to the reduction of TRIM31 under air pollution status and the exact molecular instigators of TRIM31 suppression still remain to be explored. In addition, more studies are warranted to explore the relevance of these mechanisms to the human condition.
Experimental Section
4
PM2.5 Preparation and Analysis
4.1
The methods for PM_2.5_ sampling collection and constituent analysis were carried out based on previous researches with some modification [12, 13, 21, 22, 40]. In brief, we used a 10 × 8 cm quartz filter (2500QAT‐UP, Pallflex Products, Putnam, USA) to collect PM_2.5_ particles. Subsequently, the PM_2.5_ filters were stored at −80°C for experimental use. Absolute ethyl alcohol was then added to PM_2.5_ and dissolved in sterile water, followed by sonication in an ultrasonic bath. After 48 h, the PM_2.5_ particles was subjected to vacuum freeze‐drying. The dried particles were then dissolved in sterile water, followed by centrifugation at 6,000 revolutions per minute (rpm) to remove insoluble contaminants and large aggregates, ensure a uniform particle size distribution, concentrate bioactive PM_2.5_ components, and minimize microbial/endotoxin interference. Those undissolved materials were suspended in phosphate buffer (#ST447; Beyotime Biotechnology, Shanghai, China). To analyze the constituents of particles, the mass was placed to 1% HNO_3_ and absolute ethanol, and subsequently dissolved in sterile water. The collections were continuously sonicated for another 48 h and were then concentrated via vacuum freeze‐drying. Next, the freezed‐dried extracts were resuspended in sterilized water for centrifugation at 5,000 rpm to remove the water‐soluble matters. Finally, the water insoluble extract was suspended in D‐Hanks buffer (#H1045, Solarbio, Beijing, China GIBCO Corporation, Gaithers‐burg, MD, USA) for further experimental use. The constituents of PM_2.5_ samples were analyzed using an inductively coupled plasma‐mass spectrometry (ICP‐MS) platform (ELEMENT2; Thermo Finnigan, San Jose, CA, USA) (Table S1, Supporting Information).
Animal Studies
4.2
All experimental protocols including animals used in our work were performed in accordance with National Institutes of Health Guide for the Care and Use of Laboratory Animals issued by the National Institutes of Health (NIH Publication No. 85‐23, revised 1996) and approved by the Institutional Animal Care and Use Committee in Chongqing University of Education (Chongqing, China). The protocols for animals were also conducted following the Regulations of Experimental Animal Administration issued by the Ministry of Science and Technology of the People's Republic of China (http://www.most.gov.cn). In this study, every effort was made to minimize the suffering and number of mice.
Animal Study #1
4.3
The wild type (WT) male C57BL/6 mice (6–7 weeks old with initial body weight of 20–22 g) were obtained from Beijing Vital River Laboratory Animal Technology Co. Ltd (Beijing, China).
Animal Study #2
4.4
To establish mice with a conditional knockout TRIM31 allele (cKO), the TRIM31^flox/flox^ (*TRIM31^f/f^ *) mice with C57BL/6 background were generated by the use of CRISPR/Cas9‐mediated genome engineering system as previously described [33]. Briefly, the TRIM31 exon 4/5 was used as a conditional knockout region, which were flanked by two loxP sites. Two single guide RNAs (gRNA‐1# and *gRNA‐*2#) that target TRIM31 introns were designed. The obtained mice that had exon 4/5 flanked by two loxP sites on one allele were utilized to produce *TRIM31^f/f^
- mice. POMC^+^ neuron‐specific TRIM31 knockout mice (*POMC^Cre/+^;TRIM31^flox/flox^ *) were generated through mating *TRIM31^f/f^
- mice with POMC‐Cre mice (Jackson Laboratory, Bar Harbor, Maine, USA). Schematic workflow of the construction of POMC^+^ neuron‐specific TRIM31 knockout mouse strain was displayed in Figure S2A (Supporting Information). Meanwhile, *TRIM31^flox/flox^
- mouse was defined as the control group (*TRIM31^f/f^ *). Then, all *TRIM31^f/f^
- and *POMC^Cre/+^;TRIM31^flox/flox^
- mice were separated into two subgroups by randomization: the PM_2.5_/*TRIM31^f/f^
- group and PM_2.5_/*POMC^Cre/+^;TRIM31^flox/flox^
- group.
Animal Study #3
4.5
To overexpress TRIM31 in hypothalamic POMC^+^ neurons, POMC‐Cre mice were anesthetized with 40 mg/kg of sodium pentobarbital via intraperitoneal (i.p.) injection, placed on a stereotaxic instrument, and then bilaterally injected with a Cre‐dependent AAV vector containing TRIM31 (1.2 × 10^12^ PFU/mL) in the opposite orientation flanked by two inverted loxP sites (AAV9‐Syn‐DIO‐TRIM31‐mCherry, HANBIO, Shanghai, China) at a volume of 200 nL/side into the arcuate nucleus (1.5 mm posterior to the bregma, ± 0.27 mm lateral to the midline, and 5.7 mm under the surface of the skull) or a same volume of AAV vector only containing mCherry in the opposite orientation flanked by two inverted loxP sites (AAV9‐Syn‐DIO‐mCherry), which served as a control group [97]. Fluorescence microscopy and western blot assays were performed to identify the transfection efficiency in vivo. After AAV‐TRIM31 or AAV‐Ctrl injection for 7 to 14 days, mice were subjected to long‐period PM_2.5_ inhalation.
Animal Protocols
4.6
Before animal studies, all mice were allowed to environment orientation for 7 days, and fed with a standard diet containing essential nutrients and ad libitum access to water. The mice were housed in a controlled environment with a temperature of (25 ± 1°C), humidity of (50 ± 5%), and a 12 h light/dark cycle. Mice were exposed to the quantified PM_2.5_ mass (70 L/min flow with 171 ± 2.5 ug/m^3^; equivalent to the fourth level of air quality rating in China standards) for 6 h/day and 5 times weekly (from Monday to Friday) in a movable exposure system‐S‐5002 automatic nose/mouth inhalation exposure system (Yuyan Instruments, Shanghai, China) for 24 weeks continue in succession. Mice exposed to FA served as the control group. The stability of the exposure protocols was verified and monitored using the maintenance system. During the PM_2.5_ exposure treatment, the body weights and food intake of mice were recorded weekly. After a 24 week of PM_2.5_ inhalation, all mice were killed and blood samples were collected through the cardiac puncture, followed by centrifugation at 3,000 rpm for 10 min at 4°C. Then, the supernatants were collected and stored at −80°C for further biochemical determination. The liver tissues and adipose tissues were harvested, weighed or fixed in 4% paraformaldehyde for further histological examination. The hypothalamus samples were immediately removed and frozen in liquid nitrogen and stored at −80°C for further use.
GTT and ITT
4.7
Insulin resistance was assessed using GTT and ITT assays starting at 9 AM. Following a 12 h fast, all mice were i.p. injected with 2 g/kg body weight of glucose (#50‐99‐7, Solarbio). Immediately after glucose administration, blood samples were collected from the tail vein at 0, 15, 30, 60, and 120 min. Subsequently, the commercial blood glucose test strips (SanNuo, ChangSha, China) were used to determine blood glucose levels. For the ITT test, all mice after a 12 h fasting received 1 U/kg body weight of insulin (#P3375, Beyotime Biotechnology) via i.p. injection. Blood glucose levels were measured with SanNuo blood glucose test strips at 0, 15, 30, 60, and 120 min after insulin injection.
Indirect Calorimetry
4.8
At week 23 of PM_2.5_ exposure, mice were placed in an Oxymax/Comprehensive Lab Animal Monitoring System (Columbus Instruments, Ohio, USA) for 3 consecutive days, including a 24 h acclimation period, to examine oxygen consumption (VO_2_) and carbon dioxide production (VCO_2_) under a constant ambient temperature (22°C) and a 12 h light/dark cycle (07:00–19:00 h; 19:00–07:00 h) with free access to food and drinking water. RER were determined by the ratio of VCO_2_/VO_2_. EE was calculated using the following formula: EE = VO_2_ × 1.44 × [(3.815 + (1.232 × RER)] [98]. The average value for each light and dark period was analyzed to calculate statistical significance.
Measurements of Blood Pressure
4.9
After PM_2.5_ exposure (week 24), blood pressure of mice was measured using a noninvasive automated sphygmomanometer BP‐2010A (Softron, Japan). The mice were adapted to the protocol for three consecutive days, and data were collected over four successive study days. Systolic blood pressure (SBP), diastolic blood pressure, and accordingly MBP were measured for 10 min at a fixed time. The ambient temperature was maintained at 28–30°C throughout the duration of the experiments.
Cells and Treatments
4.10
The immortalized murine microglial cell line BV2 is commonly employed as a valid surrogate for primary microglia. BV2 cells were obtained from the Cell Culture Center at the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences (Peking, China). The immortalized mouse hypothalamic neuron cell line GT1‐7 was purchased from BLUEFBIO (Shanghai, China). All cells were cultured in DMEM (Gibco, USA) supplemented with 10% FBS (Gibco) and 1% penicillin/streptomycin (Beyotime Biotechnology) in a humidified incubator at 37°C with 5% CO_2_. MG‐132 (#S1748), a proteasome inhibitor, was obtained from Beyotime Biotechnology.
TRIM31 Knockdown or Overexpression In Vitro
4.11
For in vitro knockdown transfection, TRIM31 si‐RNAs (si‐TRIM31), Nrf2 si‐RNAs (si‐Nrf2) and their corresponding scramble controls (si‐NC) were synthesized by Generay Biotechnology (Shanghai, China). The mouse full‐length TRIM31 and Nrf2 expression plasmids were generated through PCR‐based cDNA amplification and subsequently cloned into Flag‐tagged pcDNA3.1 vector, HA‐tagged pcDNA3.1 vector or His‐tagged pcDNA3.1 vector (Invitrogen). Expression plasmids encoding ubiquitin (Ub)‐K63 or ‐K48 were purchased from Genechem (Shanghai, China). Lipofectamine 3000 reagent (#L3000015, Invitrogen, USA) was used for in vitro transfection of si‐TRIM31, si‐Nrf2 or plasmid vectors following the provider's protocols. To achieve TRIM31 overexpression in cells, the Mus musculus TRIM31 gene containing the entire coding region was inserted into a replication defective adenoviral (Ad) vector under control of the cytomegalovirus promoter. Adenoviral vectors encoding the green fluorescent protein GFP gene (Ad‐GFP) served as a control. The resulting recombinant adenoviruses were purified and titrated to 5 × 10^10^ plaque‐forming units (PFU) with Vivapure AdenoPACK (#VS‐AVPQ022, Sartorius, Shanghai, China) following the products’ instructions. Subsequently, cells were infected with the constructed adenovirus at a multiplicity of infection (MOI) of 50 for 24 h.
RNA Extraction and RT‐qPCR
4.12
Total RNA was extracted from cells or hypothalamic tissues using TRIzol reagent (#15596026, Invitrogen) following the manufacturer's instructions. Briefly, RNA extraction (1 µg) was reverse transcribed into cDNA with a M‐MLV‐RT system (#28025021, Invitrogen). Subsequently, the procedures were carried out for 60 min at 42°C and halted by enzyme inactivation at 70°C for 10 min. A SYBR Green Master Mix kit (YEASEN, Shanghai, China) was used for PCR performance on a Real‐Time QPCR detection system (Agilent Mx3000P/Mx3005P, Agilent Technologies Inc., Beijing, China). Primer sequences for each target gene were generated by Sangon Biotech Co., Ltd. (Shanghai, China) and are presented in Table S2 (Supporting Information). Relative gene expression was analyzed using the 2^−ΔΔCt^ value. The expression of target genes was normalized to the value of GAPDH.
Calculation for Biochemical Indexes
4.13
Commercial assay kits including MDA (#A003‐1‐2), CAT (#A007‐1‐1), SOD (#BC0175) and GSH (#BC1175) were obtained from Nanjing Jiancheng Bioengineering Institute (Nanjing, China) or Solarbio, and the determination of these parameters in cells or tissues was performed following the respective manufacturers’ instructions. The releases of cytokines in sera, hypothalamic samples, and culture supernatants were examined using commercially available ELISA kits, including the mouse IL‐1β (#PI301, Beyotime Biotechnology), the mouse TNF‐α (#PT512, Beyotime Biotechnology), the mouse MCP‐1 (#PC125, Beyotime Biotechnology), and the mouse IL‐6 (#M6000B, R&D System, USA) according to the manufacturers’ recommendations. Mouse α‐MSH ELISA Kit (#orb408908) was purchased from Biorbyt (Cambridge, England) to examine hypothalamic α‐MSH levels following the manufacturers’ instructions.
Western Blot Assay
4.14
Cells or hypothalamic samples were lysed and homogenized in RIPA Lysis Buffer (#P0013B, Beyotime Biotechnology) to obtain a homogenate. Nuclear and cytoplasmic proteins were extracted using a Nuclear and Cytoplasmic Protein Extraction Kit (#P0028, Beyotime Biotechnology) following the supplier's protocols. The homogenates were then centrifuged at 13,000 × g for 20 min at 4°C. The supernatants were discarded, and the protein samples were collected. The protein concentration was determined using an Enhanced BCA Protein Assay Kit (P0009, Beyotime Biotechnology). Subsequently, 25–50 µg of proteins were separated on 10%–12% sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes (#10600023, GE Healthcare Life Science, Germany). Next, 5% nonfat‐dried milk (#LP0033B, Biosharp Life Science, Beijing, China) containing 0.1% Tween‐20 (#ST825, Beyotime Biotechnology) was used for membrane blocking for 1 h at room temperature. The membranes were then incubated with the indicated primary antibodies (Table S3, Supporting Information) in the prepared TBST buffer overnight at 4°C, followed by washing and incubation with appropriate HRP‐conjugated secondary antibodies (Table S3, Supporting Information) diluted in TBST buffer for 1.5 h at room temperature. The blotting bands were finally visualized by enhanced chemiluminescence reagents (BeyoECL Moon, #P0018FM, Beyotime Biotechnology) and exposed to Kodak X‐ray film (XBT‐1, Eastman Kodak Company, USA) for 30 s to 10 min. Quantification of protein bands was calculated as gray‐scale value (Version 1.52 g, Image J, National Institutes of Health, Bethesda, MD, USA). Band expression intensity was normalized to GAPDH. Fold expression changes for each protein were quantified and presented.
Histological Assay
4.15
The liver and eWAT tissues collected from all groups of mice were fixed in 4% paraformaldehyde for 48 h, embedded in paraffin, and sectioned at a thickness of 5 µm each. The tissue sections were subsequently subjected to histological examination using hematoxylin and eosin (H&E) and Oil Red O (ORO) staining.
IF Staining
4.16
Mouse brain tissues were mildly and carefully removed, collected and placed in 4% paraformaldehyde for 24 h after perfusion with 4% cold paraformaldehyde. Subsequently, the brain tissues were dehydrated in 20%–40% sucrose (#ST1670, Beyotime Biotechnology) for 24 h. Thereafter, the samples were frozen in Sakura Tissue‐Tek O.C.T. Compound (SAKURA, USA) and cut into 20–30 µm thick slices. These sections were subjected to IF staining analysis. Briefly, after washing with phosphate‐buffered saline (PBS), the frozen brain slides were treated with 0.1% Triton X‐100 (#ST797, Beyotime Biotechnology) for 5 min to permeabilize, followed by blocking in 10% goat serum (#C0265, Beyotime Biotechnology) for 1 h at room temperature. Subsequently, appropriate primary antibodies were added to the sections for incubation overnight at 4°C. After washing, the sections were labeled with the Alexa‐Flour 488‐, or 647‐conjugated secondary antibodies at room temperature for 1 h in a humidified chamber. BV2 cells after treatments were collected, washed and fixed for 15 min with precooled 4% paraformaldehyde. Next, the cells were permeabilized for 10 min using 0.1% Triton X‐100 (#ST797, Beyotime Biotechnology). After blocking in 10% goat serum (#C0265, Beyotime Biotechnology) for 1 h at room temperature, primary antibodies were subjected to samples for incubation overnight at 4°C. After washing with PBS, the cells were incubated with secondary antibody at 37°C for 1 h. DAPI (#C1006, Beyotime Biotechnology) was used for nuclear staining. Positive‐stained cells and areas were quantified using ImageJ software (NIH), calculated by positively expressed cells relative to DAPI‐stained cells. The number of cells with NF‐κB p65 subunit nuclear translocation in four to eight random fields was counted in a masked manner and expressed as a percentage of the number of translocated cells compared to that of total cells. As for colocalization, the targeting proteins were double‐labeled by immunofluorescence Confocal Microscopy, and Z‐stack scanning was used to eliminate optical errors to ensure the accuracy of spatial positioning. The degree of colocalization was quantified using Mander's overlap coefficient (MOC) and Pearson correlation coefficient (PCC). A threshold of ≥0.6 was set to determine significant colocalization. The overlapping areas of fluorescence signals were automatically analyzed through the ImageJ software (NIH). For quantification, the sample size of each group is ≥4, and the number of selected fields of view is specifically described in the legend. Fold changes relative to the control group were quantified. Detailed information for primary and secondary antibodies was displayed in Table S3 (Supporting Information).
ROS Production Calculation
4.17
ROS production in the hypothalamus was measured by employing the fluorescent probe DHE (#D23107, Invitrogen). Mice after PM_2.5_ exposure were killed and brain tissues were collected for DHE experiment. In brief, 20 µm thick cryosections of frozen brain tissues were obtained and incubated with 10 µm of DHE at 37°C for 30 min in a humidified chamber, followed by counterstaining with DAPI (#C1006, Beyotime Biotechnology). As for cellular ROS generation, dichlorofluorescin diacetate (DCFH‐DA) dye (#S0033M, Beyotime Biotechnology), a widely used cell‐permeable fluorescent probe for detecting intracellular, was used. DCF‐DA passively diffuses into cells, where intracellular esterases hydrolyze it to DCFH (nonfluorescent form). In the presence of ROS, DCFH is oxidized to DCF (2′,7′‐dichlorofluorescein), emitting green fluorescence (Ex/Em = 488/525 nm). This fluorescence intensity quantitatively reflects ROS levels. Briefly, the viable cells after treatments were subjected to DCFH‐DA (10 µm) incubation for 20 min at 37°C, followed by washes to remove excess probe. Elevated DCF fluorescence indicates ROS accumulation. Fluorescent images were obtained under a confocal FV3000 (Olympus Co., Japan) or fluorescence microscope (Olympus Co., Japan).
Examination for Cell Viability and LDH Releases
4.18
For cell viability assessment, a Cell Counting Kit‐8 assay kit (CCK‐8; #C0039, Beyotime Biotechnology) that follows a colorimetric reaction that reduces a tetrazolium salt (WST‐8) by cellular dehydrogenases (present in viable cells) to produce a measurable formazan dye was utilized following the manufacturer's protocols. Briefly, cells (200 µL/well) were cultured on 96‐well plates. After each treatment, CCK‐8 reagent (10 µL) was added to each well for 4 h incubation at 37°C. Finally, the absorbance at 450 nm was read on a microplate reader (SpectraMax iD3, Molecular Devices, USA). To examine LDH releases from cells, a LDH Cytotoxicity Assay Kit (#C0017, Beyotime Biotechnology) was employed according to the manufacturer's instructions. Culture medium was collected, and 60 µL of working fluid was added, followed by a 30 min incubation at room temperature in the dark. LDH release rates were quantified by measuring the absorbance at 490 nm on a microplate reader (SpectraMax iD3).
Co‐IP Assay
4.19
Co‐IP detection in our study was conducted as previously reported [33, 34, 99]. Briefly, cells or hypothalamic tissues were homogenized in Lysis Buffer (20 mm Tris (pH7.5), 150 mm NaCl, 1% Triton X‐100 and 1 mm EDTA) with Protease Inhibitor Cocktail (#P2179M, Beyotime Biotechnology) at 4°C to obtain an IP‐specific lysis solution, followed by centrifugation as described in the immunoblotting analysis protocols. After being precleared with immunoglobulin G (IgG) and protein A/G‐agarose beads (#P2179M, Beyotime Biotechnology), the lysates were collected and incubated with the shown primary antibodies and Protein A/G‐agarose (#P2179M, Beyotime Biotechnology) at 4°C overnight with gentle shaking. Next, the immunoprecipitated proteins were rinsed 3 times with lysis buffer and boiled with 100 µL 1 × SDS sample loading buffer for 5 min, which were then subjected to western blot assay using the respective primary antibodies and appropriate secondary antibodies as listed in Table S3 (Supporting Information).
Ubiquitination Detection
4.20
The binding ubiquitination analysis were operated according to our previous protocols with minor modification [33, 34]. Briefly, the transfected cells were lysed in 100 µL SDS lysis buffer (#P0013G, Beyotime Biotechnology), followed by denaturation via heating. The supernatants were then diluted using immunoprecipitation buffer and subjected to immunoprecipitation with the indicated antibodies. Subsequently, immunoblot analysis was performed on the obtained immunoprecipitants using anti‐HA or anti‐ubiquitin (Ub) antibodies.
GST Pull‐Down Detection
4.21
GST pull‐down analysis was carried out to investigate the direct interaction between TRIM31 and Nrf2 in cells using the methods as previously demonstrated [33, 34, 98]. In brief, pGEx‐4T‐1/GST‐TRIM31 or pGEx‐4T‐1/GST‐Nrf2 were generated and transformed using the Rosetta (DE3) competent cell system, followed by expression induction through adding Isopropyl β‐D‐1‐thiogalactopyranoside (IPTG) (0.5 mm, #ST098, Beyotime Biotechnology). The fused proteins GST and GST‐Nrf2 or GST‐TRIM31 were then purified using glutathione‐sepharose 4B beads (GE Healthcare, USA). The GST‐bound, GST‐Nrf2 or GST‐TRIM31‐bound glutathione‐sepharose 4B beads were subsequently incubated with His‐TRIM31 or His‐Nrf2 for 4 h at 4°C. Immunocomplexes were washed and eluted for immunoblot analysis.
Statistical Analysis
4.22
All data were presented as means ± standard deviation (SD) and analyzed using GraphPad PRISM (version 9.4; GraphPad Software, USA). The Student's t‐test was used to analyze differences between two groups. One‐way or two‐way analysis of variance (ANOVA) with Tukey's post hoc tests was conducted for the significance of differences among three or more groups. The p value < 0.05 was considered statistically significant. Data were independently repeated at least 3 times unless indicated. Animal studies were performed in a blinded manner, and a randomization process was conducted in mice grouping. All images were analyzed in a blinded manner.
Author Contributions
Conceptualization & Methodology: Chenxu Ge, Yan Sun, Bochu Wang, Minxuan Xu. Investigation: Chenxu Ge, Yan Sun, Qiang Li, Deshuai Lou, Linfeng Hu, Xi Liu, Gang Kuang, Jing Luo, Jun Tan, Bochu Wang, Minxuan Xu, Fengxiang Li, Chuanwang Miao, Shuqiang Zhao, Lei Zou, Xuedong Teng, Lina Liu, Tingguang Li, Jiamao Lin, Changsheng Yang. Data analysis: Chenxu Ge, Yan Sun, Jiamao Lin, Changsheng Yang. Funding acquisition, Project administration & Supervision: Chenxu Ge, Jun Tan, Bochu Wang, Minxuan Xu, Yanrong Ren. Writing – original draft: Minxuan Xu, Jun Tan, Chenxu Ge, Bochu Wang, Minghui Chang. Writing – review & editing: Minxuan Xu, Bochu Wang, Minghui Chang, Jun Tan.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supporting Files 1: advs73616‐sup‐0001‐SuppMat.doc.
Supporting Files 2: advs73616‐sup‐0002‐SuppMat.docx.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1H. Boogaard , K. Walker , and A. J. Cohen , “Air Pollution: The Emergence of a Major Global Health Risk Factor,” International Health 11, no. 6 (2019): 417–421, 10.1093/inthealth/ihz 078.31613318 · doi ↗ · pubmed ↗
- 2R. D. Brook , S. Rajagopalan , C. A. Pope III , et al., “Particulate Matter Air Pollution and Cardiovascular Disease: An Update to the Scientific Statement From the,” American Heart Association Circulation 121, no. 21 (2010): 2331–2378.20458016 10.1161/CIR.0b 013e 3181 dbece 1 · doi ↗ · pubmed ↗
- 3T. Marchini , A. Zirlik , and D. Wolf , “Pathogenic Role of Air Pollution Particulate Matter in Cardiometabolic Disease: Evidence From Mice and Humans,” Antioxidants & Redox Signaling 33, no. 4 (2020): 263–279, 10.1089/ars.2020.8096.32403947 · doi ↗ · pubmed ↗
- 4R. Sinkemani , A. Sinkemani , X. Li , and R. Chen , “Risk of Cardiovascular Disease Associated with the Exposure of Particulate Matter (PM 2.5): Review,” Journal of Environmental Protection 09, no. 6 (2018): 607–618, 10.4236/jep.2018.96038. · doi ↗
- 5P. Fu , X. Guo , F. M. H. Cheung , and K. K. L. Yung , “The Association Between PM 2.5 Exposure and Neurological Disorders: A Systematic Review and Meta‐analysis,” Science of the Total Environment 655 (2019): 1240–1248, 10.1016/j.scitotenv.2018.11.218.30577116 · doi ↗ · pubmed ↗
- 6M. Mazidi and J. R. Speakman , “Ambient Particulate Air Pollution (PM 2.5) Is Associated With the Ratio of Type 2 Diabetes to Obesity,” Scientific Reports 7, no. 1 (2017): 1–8, 10.1038/s 41598-017-08287-1.28831041 PMC 5567252 · doi ↗ · pubmed ↗
- 7J. Chen , L. Wu , G. Yang , et al., “The Influence of PM 2.5 Exposure on Non‐alcoholic Fatty Liver Disease,” Life Sciences 270 (2021): 119135, 10.1016/j.lfs.2021.119135.33513397 · doi ↗ · pubmed ↗
- 8Y. Shou , X. Zhu , D. Zhu , et al., “Ambient PM 2.5 Chronic Exposure Leads to Cognitive Decline in Mice: From Pulmonary to Neuronal Inflammation,” Toxicology Letters 331 (2020): 208–217, 10.1016/j.toxlet.2020.06.014.32569800 · doi ↗ · pubmed ↗