Induction of endogenous IL-10 promotes resolution and tolerance of nitric oxide in microglia

Abstract

Endogenous interleukin-10 (IL-10), a potent anti-inflammatory cytokine, is induced in a timely and coordinated manner to dampen microglia-mediated brain inflammation. However, it remains unclear how it alters the inflammatory process to shape the immune polarization of microglia. This study aimed to investigate the anti-inflammatory mechanisms of endogenous IL-10 in activated and tolerized microglia using in vitro multiple-reconstituted primary brain cell cultures and an in vivo IL-10 knockout (IL-10KO) animal model. Upon a single or repeated lipopolysaccharide (LPS) treatment regimen, the expression levels of the inflammatory factors during the neuroinflammatory/tolerance process were measured by quantitative real-time polymerase chain reaction, enzyme-linked immunosorbent assay (ELISA), and Griess reagent assay. ELISA data showed that cell-autonomous induction of endogenous IL-10 occurs in LPS-activated and LPS-tolerized microglia. Furthermore, comparing the LPS-elicited pro-inflammatory factor expressions at different neuroinflammatory stages between the wild-type and IL-10KO groups, our data revealed the failure of negative-feedback suppression of inducible nitric oxide synthesis (iNOS) during immune resolution in the IL-10KO brains. Moreover, LPS-treated IL-10KO microglia increase the supernatant level of nitrite and become overactive during late-stage inflammation, despite no changes in cell number; in contrast, LPS-tolerized IL-10KO microglia fail to program endotoxin tolerance of nitric oxide/inducible nitric oxide synthesis (iNOS). In summary, our data demonstrate that the cell-autonomous induction of endogenous IL-10 in microglia is crucial for mitigating brain immune responses, particularly in the resolution and tolerance of nitric oxide.

1. Introduction

Neuroinflammation has been recognized as a self-defensive approach to remove harmful stimuli and initiate the healing process in the central nervous system (CNS) (Chu et al., 2015; Rodriguez et al., 2022; Rosenblum et al., 2015). In response to potent immune challenges such as lipopolysaccharides (LPS) or neuronal damage, microglia, the primary innate immune cells of the brain, produce various pro-inflammatory and anti-inflammatory mediators in an orchestrated manner along with changing their morphology from a ramified appearance to that of an amoeboid shape (Chu et al., 2015). This LPS-elicited neuroinflammatory process in microglia displays three stereotyping stages: initiation, propagation, and resolution. During the immune status of initiation and propagation, LPS-activated microglia immediately release pro-inflammatory factors such as superoxide, tumor necrosis factor-alpha (TNF-α), followed by interleukin-1 beta (IL-1β), nitric oxide (NO), prostaglandin E2 (PGE₂), and interleukin-6 (IL-6) in a cascade of rigorous time frame (Chu et al., 2015; Gao et al., 2020). Among them, the generation of NO, which is a versatile mediator, is achieved by the enzyme nitric oxide synthase (NOS) through the conversion of L-arginine (Arg) to L-citrulline. In addition to numerous homeostatic functions of NO at lower physiological concentrations, higher amounts of NO derived from activated glial cells through the transcriptional regulation of the inducible NOS (iNOS) participate in a variety of neurobiological diseases (Saha and Pahan, 2006), including oligodendrocyte degeneration in demyelinating diseases (Mitrovic et al., 1994) and neuronal death during ischemia, trauma, and neurodegenerative diseases (Liu et al., 2002). High levels of iNOS expression in glia, which can lead to oxidative and nitrosative stress in the brain, cause inhibition of neuronal respiration, depolarization, and glutamate release by NO and its toxic metabolite, peroxynitrite (ONOO−), resulting in excitotoxicity and cell death of particularly vulnerable neurons (Brown, 2007; Heales et al., 1999). Therefore, targeting the production of NO/iNOS is a potential therapeutic strategy for inflammation-mediated neurodegenerative diseases (Broom et al., 2011; Tewari et al., 2021). Subsequently, at the immune resolution stage, the anti-inflammatory cytokine, such as interleukin-10 (IL-10), is secreted to eliminate excessive inflammation (Chu et al., 2015; Gao et al., 2020). Consequently, after the precise immune resolution, the activated microglia are reversibly returned to the status of immune homeostasis. However, if this neuroinflammation is uncontrolled and unresolved, active microglia transform into the reactive phenotype and become toxic to neighboring neurons (Block et al., 2007; Burguillos et al., 2011). Tightly regulating and actively terminating the neuroinflammatory response is needed to prevent neural tissue destruction. Thus, understanding the induction of endogenous IL-10 in regulating the neuroinflammatory process may be beneficial for the development of therapeutic strategies in inflammation-mediated neurodegenerative disorders (Dansokho and Heneka, 2018; Swanson et al., 2018; Trovato Salinaro et al., 2018).

Interleukin 10 (IL-10), a potent anti-inflammatory cytokine, holds significant promise in resolving inflammation and maintaining immune homeostasis in both CNS (Gao et al., 2020; Saraiva and O'Garra, 2010; Strle et al., 2001) and peripheral tissues (Kuhn et al., 1993). The main source of endogenous IL-10 production is considered as microglia among brain cells (Chu et al., 2015). However, whether regulation of microglia-derived IL-10 expression profile by other brain cells via a non-cell-autonomous manner is still unclear. In general, the activation of the intracellular signaling cascades and SP1 transcriptional factors by most pro-inflammatory mediators leads to a rapid increase in promoter activity and mRNA expression of IL-10 after the immune challenge (Foey et al., 1998; Saraiva and O'Garra, 2010; Tone et al., 2000). In addition to intracellular signaling cascades, earlier-released TNF-α and PGE₂ from LPS-treated microglia negatively regulate IL-10 induction via the EP2 receptor-mediating G-protein-independent arrestin signaling (Chu et al., 2015). These results imply the complex intracellular and extracellular interactions among different immune factors that regulate endogenous IL-10 induction in the time frame of the neuroinflammatory process. On the other hand, endogenous IL-10 induction plays a crucial role in regulating microglial phagocytosis, repressing the expression and release of pro-inflammatory mediators, and suppressing antigen presentation (Heyen et al., 2000; Kremlev and Palmer, 2005; Qian et al., 2006; Sabat et al., 2010). However, the impacts of endogenous IL-10 on suppressing the expression of pro-inflammatory mediators during the neuroinflammatory process are still controversial due to its later-onset expression pattern. For instance, the induction of endogenous IL-10 by LPS significantly suppresses the production of IL-1β (Gao et al., 2020) but does not affect early-releasing TNF-α (Kuo et al., 2023) in vitro brain cell cultures. Therefore, an in vivo model of neuroinflammation may be needed to depict further the action of endogenous IL-10 in regulating brain immune resolution.

Recently, accumulated evidence demonstrates that microglia are capable of developing innate immune memory to either enhance (trained immunity) or suppress (immune tolerance) subsequent immune responses that are involved in the regulation of neurological disease pathogenesis (Dong et al., 2024; Neher and Cunningham, 2019; Wendeln et al., 2018). For example, two consecutive LPS challenges within a short interval immediately elicit microglial cell endotoxin tolerance (ET), where the production of pro-inflammatory mediators is decreased (Chu et al., 2016). In contrast, that of anti-inflammatory mediators, such as IL-10, is enhanced in LPS-tolerized microglia. In other words, a transient unresponsiveness or reduced sensitivity to subsequent endotoxin challenges in such ET microglial cells is seen. Several studies have demonstrated that ET development in peripheral immune cells, such as macrophages, entails multiple mechanisms, including negative signal transductions, transcriptional networks, and post-translational modifications (Zhang and Cao, 2021; Zhang et al., 2022). Our previous study showed that microglia alone fail to develop ET unless in the presence of neurons and astroglia, unlike macrophages (Chu et al., 2016). The induction of ET in microglia, which display an anti-inflammatory phenotype, prevents excessive toxic damage in the neurons due to reduced secreted cytokine production (Chu et al., 2016; Kuo et al., 2023), such as TNF-α. However, how the LPS-tolerized microglia shape their polarizing phenotype via enhancement of IL-10 production is still unknown. Altogether, this study aimed to determine the impact of endogenous IL-10 induction on regulating immune initiation, propagation, resolution, and tolerance using in vivo animal brain and in vitro brain cell culture models of neuroinflammation.

2. Materials and methods

2.1. Multiple reconstituted primary neuron-glia, mixed-glia, and microglia-enriched cultures and cell treatment

According to the previous reports (Chu et al., 2015, 2016; Liu et al., 2000), the neuron-glia and mixed-glia/microglia-enriched cultures were prepared from the whole brains of E14.5-day embryos and 1-day-old pups, respectively. The culture plates were precoated with poly-D-lysine (20 μg/mL) (P7280; Merck KGaA). Briefly, brain tissue without meninges and blood vessels was dissociated with mild mechanical trituration. For the preparation of neuron-glia cultures, 5 × 10⁵ cells were seeded in each well of 24-well culture plates with culture medium containing MEM, 10 % heat-inactivated fetal bovine serum (FBS), 10 % heat-inactivated horse serum, 1 g/L glucose, 2 mM L-glutamine, 1 mM sodium pyruvate, and 0.1 mM nonessential amino acids for the growth and maintenance at 37 °C in a humidified atmosphere of 5 % CO2/95 % air. The neuron-glia cultures were added with fresh medium (0.5 ml/well) 3 days after cell seeding. Their cellular composition included 11 % microglia, 50 % astrocytes, and 39 % neurons on the 7th day after seeding, which were used for further experiments. For the preparation of mixed-glia cultures, the isolated cells (1.5 × 10⁶ cells) were seeded in 6-well plates in DMEM/F12 medium (#11330032; Thermo Fisher Scientific Inc.), 10 % fetal bovine serum, 2 mM L-glutamine (#25030149; Thermo Fisher Scientific Inc.), 1 mM sodium pyruvate (#11360-070; Thermo Fisher Scientific Inc.), 0.1 mM nonessential amino acids (#11140-035; Thermo Fisher Scientific Inc.), and 50 U/mL penicillin, and 50 μg/mL streptomycin (#10378016; Thermo Fisher Scientific Inc.) and were changed their media four days later. The mixed-glia cultures containing about 80 % astrocytes and 20 % microglia, two weeks after culturing, were used for further experiments. For the preparation of microglia-enriched cultures, the isolated cells (5 × 10⁷ cells) were seeded in 150 cm² culture flasks in DMEM/F12 medium (10 % FBS, 2 mM of L-glutamine, 1 mM of sodium pyruvate, 0.1 mM of nonessential amino acids, 50 U/mL of penicillin, and 50 μg/mL of streptomycin) and maintained at 37 °C in a humidified atmosphere of 5 % CO2/95 % air. The medium was changed 4 days later, until reaching confluence. The enriched microglia (99 % pure) were obtained by shaking the flasks for 60 min at 180 rpm after 12–14 days of cell seeding (upon reaching confluence) for further experiments.

Cell Treatment for activated or tolerized microglia is shown in Fig. 1A. Neuron–glial, mixed glial, and microglia-enriched cultures were pre-incubated with or without LPS (E. coli O111:B4, Cat# 437627, protein contaminants ≤2.0 %, nucleic acid contaminants ≤2.5 %; 15 ng/mL) from EMD Chemicals, Inc. (Darmstadt, Germany) for 6 h. After replacing the fresh media and waiting an additional 6 h, these cells were re-stimulated with LPS (15 ng/mL). According to previous investigations (Chu et al., 2016; Kuo et al., 2023), immune-activated and immune-tolerant microglia could be induced by once and repeated LPS treatment, respectively. The expressions of supernatant IL-10 were measured at 6 and 24 h in these cells by ELISA.

Fig. 1.

Cell-autonomous induction of endogenous IL-10 in LPS-activated and LPS-tolerized microglia. (A) As an illustrated experimental procedure of endotoxin treatment regimen for the development of the activated and tolerized microglia, the neuron-glia (neurons + astroglia + microglia), mixed-glia (astroglia + microglia), and microglia-enriched cultures were prepared and pre-incubated with a vehicle (activated) or LPS (tolerized) (15 ng/ml) for 6 h. After being replaced with fresh medium for 6 h, LPS (15 ng/ml) was re-added to the cells. (B-D). After receiving the endotoxin treatment regimen (activated versus tolerized) as indicated in (A), supernatant levels of IL-10 in neuron-glia (B), mixed-glia (C), and microglia-enriched (D) cultures were detected, respectively, by ELISA assay 6 and 24 h later. Data are expressed as the mean ± SEM from more than three independent experiments (n ≥ 3) in duplicate with two-way ANOVA with Tukey's multiple comparisons test (6-h group versus 24-h group: ∗∗p < 0.01, ∗∗∗p < 0.001; LPS group versus LPS/LPS group: #p < 0.05, ##p < 0.01, ###p < 0.001).

2.2. Measurement of IL-10 and nitric oxide

The IL-10 level in the supernatant of the neuron-glia, mixed-glia, and microglia-enriched cultures, and animal serum was measured with the commercial ELISA kits from R&D Systems (Minneapolis, MN). The value of IL-10 within the standard curve, along with ∼10 % inter- and ∼15 % intra-assay coefficients of variation in the single or between different assay runs, is considered acceptable for further calculation. Under physiological conditions, nitric oxide converts to its stable oxidation product, nitrite. The Griess assay is primarily used to quantify nitrite (detection limit: 0.5 μM) to reflect the accumulated supernatant levels of nitric oxide indirectly in biological and chemical research.

2.3. Animal experiments

B10.129P2(B6)-IL-10tm1Cgn/J (IL-10 knockout or IL-10KO) mice and their wild-type (WT) (C57BL/10J) mice were purchased from the Jackson Laboratory (Bar Harbor, ME). According to the previous study (Liu et al., 2008), female mice are more resistant to LPS than male mice. To avoid the gender effect, we only applied male mice in the current study. The male WT (n = 22) and IL-10KO mice (n = 22) in young adulthood (∼3 months old) were intraperitoneally injected with saline (10 WT and 10 IL-10KO mice) or LPS (L3012, Merck KGaA) (12 WT and 12 IL-10KO mice; 0.3 mg/kg) to determine the anti-inflammatory effects of endogenous IL-10 on systemic inflammation-mediated neuroinflammatory cascades in the brains. Several inflammatory factor expression profiles in the brains of these treated animals were measured 1 and 12 h after LPS treatment by real-time PCR, including TNF-α, IL-1β, COX-2, iNOS, and IL-10. In addition, the mortality of these treated animals was measured after LPS treatment to calculate their survival rate. For the animal sacrifice procedure, animals were first anesthetized with 1–3 % isoflurane for anesthesia and then subjected to transcardiac perfusion with saline, followed by 4 % paraformaldehyde for brain tissue dissection and fixation. The WT and IL-10KO mice were group housed on a 12 h/12 h light/dark cycle at a room temperature of 21 °C ± 2 °C and humidity of 55 % ± 5 % and were given free access to water and food. All animal studies received approval from the Institutional Animal Care and Use Committee at Chang Gung University of Science and Technology, aligning with ARRIVE guidelines.

2.4. Reverse transcription quantitative real time-PCR (RT-qPCR)

The whole brain tissues of these animals treated with LPS at 1, 3, 6, and 12 h (Fig. 2 and Fig. S1A and B) and the cell pellet of neuron-glia cultures with a single or repeated LPS treatment (Fig. S2) were collected for further measurement of mRNA expression of inflammatory factors by quantitative RT-PCR as described previously (Chu et al., 2015, 2016). The RNeasy Mini kit (QIAGEN, Valencia, CA) was used to extract total RNA, according to the manufacturer's instructions. A spectrophotometer measured quantification and qualification (The ratio of absorbance at 260 nm and 280 nm is approximately 2.0) of total RNA. Using MuLV reverse transcriptase (Applied Biosystems, Foster City, CA), first-strand cDNA was synthesized from 2 μg of total RNA. After that, using SYBR-Green Master mix (Applied Biosystems, Foster City, CA), quantitative real-time PCR analysis was performed to amplify cDNA under the conditions as follows: hold at 95 °C for 10 min and perform 40 cycles at 95 °C for 15 s and 60 °C for 1 min. The sequences of the primers designed with Vector NTI software (Invitrogen) were the following: mouse TNF-α forward primer 5′ GAC CCT CAC ACT CAG ATC ATC TTC T 3’; mouse TNF-α reverse primer 5′ CCT CCA CTT GGT GGT TTG CT 3’; mouse COX-2 forward primer 5′ TGA TAT GTC TTC CAG CCC ATT G 3’; mouse COX-2 reverse primer 5′ AAC GGA ACT AAG AGG AGC AGC 3’; mouse iNOS forward primer 5′ GCC ACC AAC AAT GGC AAC A 3’; mouse iNOS reverse primer 5′ CGT ACC GGA TGA GCT GTG AAT T 3’; mouse IL-1β forward primer 5′ CTG GTG TGT GAC GTT CCC ATT A 3′; mouse IL-1β reverse primer 5′ CCG ACA GCA CGA GGC TTT 3’; mouse IL-10 forward primer 5′ CCC TTT GCT ATG GTG TCC TTT C 3’; mouse IL-10 reverse primer 5′ CAA AGG ATC TCC CTG GTT TCT C 3’; mouse GAPDH forward primer 5′ TTC AAC GGC ACA GTC AAG GC 3’; mouse GAPDH reverse primer 5′ GAC TCC ACG ACA TAC TCA GCA CC 3’. Data were normalized to GAPDH expression.

Fig. 2.

Deficiency in IL-10 impairs negative-feedback suppression of inducible nitric oxide synthesis in the inflammatory brain. (A) Schematic of the experimental design for determining the role of endogenous IL-10 in LPS-induced immune resolution of pro-inflammatory factors. The C57/10J (WT) and IL-10KO mice received intraperitoneal LPS (0.3 mg/kg) and saline treatment, followed by measurement of endogenous IL-10 induction in the brain, liver, and serum, and the inflammatory mRNA expression, including TNF-α, IL-1β, COX-2, and iNOS, using ELISA and real-time PCR at different time points as indicated. Finally, the survival rate of these LPS-treated mice was calculated. (B-E) Expression of TNF-α (B), IL-1β (C), COX-2 (D), and iNOS (E) mRNA in the WT and IL-10KO mice's brains after 1, 3, 6, and 12 h of saline (10 WT and 10 IL-10KO) and LPS (12 WT and 12 IL-10KO) treatment was measured by RT-PCR. Data are presented as a fold change relative to the saline-treated control and expressed as the mean ± SEM from three independent experiments performed in triplicate with two-way ANOVA with Tukey's multiple comparisons test (saline group versus LPS group: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; WT group versus IL-10KO group: #p < 0.05, ##p < 0.01, ###p < 0.001).

2.5. Immunocytochemistry staining

Immunocytochemistry (ICC) staining with anti-ionized calcium binding adaptor molecule 1 (Iba-1; 1:400; 019–19741; FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) was performed as described in previous studies (Chu et al., 2016, 2023). Briefly, after formaldehyde fixation and incubation with blocking buffer (0.4 % Triton X-100, 1 % BSA, 4 % normal serum in PBS), the cells were treated with primary anti-Iba-1 antibody diluted in antibody diluent overnight at 4 °C. Following incubation with an appropriate biotinylated secondary antibody, the Vectastain ABC reagents and color development with 3,3′-diaminobenzidine were used to visualize the bound primary anti-Iba-1 antibody. Then, a CCD camera and the MetaMorph software (Molecular Devices, Sunnyvale, CA, USA) were applied to film Iba-1 images.

The Nikon NIS-Element BR software (Nikon Instruments Inc., Melville, NY, USA) was used to automatically circle and measure the mean intensity value in the areas of Iba-1-positive staining cells under consistent intensity criteria. Protein expression levels of Iba-1 in each staining image are calculated by multiplying the mean intensity by the Iba-1-positive-staining area. The integrated intensity data were obtained from two to six wells per treatment condition in three independent experiments and were expressed as the percentage of the WT saline-treated control (means ± SEM). The microglia number (Iba-1+ cells) in mixed-glia cocultures was manually counted by ∼ nine representative areas in each treatment group from three independent experiments performed in triplicate (20X power) and expressed as the percentage of the WT saline-treated control (means ± SEM). 2 to 3 individuals performed cell counting and quantification of immunostaining density.

2.6. Statistical analysis

All data from at least three independent experiments are expressed as the mean ± standard error of mean (SEM) and were compared between groups using the one-way analysis of variance (ANOVA) with Dunnett's multiple comparisons test or two-way ANOVA with Tukey's multiple comparisons test (Prism 10; GraphPad Software, San Diego, CA, USA). In addition, RT-PCR results were normalized to their respective internal loading controls, GAPDH. Standard curves of IL-10 ELISA were plotted, and the data were obtained within the linear range of the curve. A p-value of <0.05 was considered statistically significant. In all cases, a p-value less than 0.05 was considered statistically significant.

3. Results

3.1. Activated and tolerized microglia autonomously produce IL-10

The presence of neurons, astroglia, or other brain cells could change cellular survival (Pesti et al., 2024) and immune properties of microglia via the non-cell-autonomous mechanisms, such as the development of TNF-α (Chu et al., 2016) and PGE₂ (Kuo et al., 2022a) immune tolerance. Using multiple reconstituted primary neuron-glia (neurons, astroglia, and microglia), mixed-glia (astroglia and microglia), and microglia-enriched cultures with once or repeated LPS treatment regimens as described in Fig. 1A, we determined whether activated (LPS-treated group) or tolerized (LPS/LPS-treated group) microglia were capable of producing and enhancing IL-10 production in the presence or absence of neurons or astroglia. Consistent with a previous study (Chu et al., 2016), ELISA data showed that induction of endogenous IL-10 by a single LPS treatment was observed in the supernatants of all neuron-glia, mixed-glia, and microglia-enriched cultures at 24 h (Fig. 1B–D). However, unlike the microglial TNF-α (Chu et al., 2016) and PGE₂ (Kuo et al., 2022b) tolerance, which occurs only in the presence of neurons and/or astroglia, enhancement of LPS-induced IL-10 production in the ET-treated microglia-enriched cultures was similar to that of neuron-glia and mixed-glia cocultures. Our data indicated that activated and tolerized microglia elicited an endogenous IL-10 expression profile in a cell-autonomous manner.

3.2. Deficiency in IL-10 impairs nitrite resolution in the inflammatory brain

To determine the role of endogenous IL-10 induction in regulating brain immune initiation, propagation, and resolution status of the neuroinflammation process in vivo, C57/10J (WT) and IL-10 knockout (IL-10KO) mice in young adulthood were intraperitoneally injected with LPS at 0.3 mg/kg concentration used as a murine neuroinflammation model (Fig. 2A) (Qin et al., 2007). First, the induction of endogenous IL-10 in the brain, liver, and serum by a single intraperitoneal 0.3 mg/kg dose of LPS injection was detected. Real-time PCR analysis showed that compared to the saline-treated mice, the mice with LPS treatment had significantly increased IL-10 mRNA levels of the brain (p < 0.05, F (3, 8) = 13.64, one-way ANOVA) and liver (p < 0.05, F (3, 8) = 26.59, one-way ANOVA) by 5 and 50 times, respectively, at 1 h after treatment (Fig. S1A and B). Its expression was sustained at 3 h and slightly reduced at 6 h in the brain of LPS-treated animals, as well as in the liver. Meanwhile, ELISA results revealed that serum level of IL-10 was detected in LPS-treated WT mice at 1 h but not in IL-10KO mice (Fig. S1C, p < 0.05, F (4, 29) = 1043, two-way ANOVA).

Next, to study how endogenous IL-10 induction affected the neuroinflammatory process in the CNS, we measured the transcripts of the pro-inflammatory factors, including TNF-α (Fig. 2B), IL-1β (Fig. 2C), COX-2 (Fig. 2D), and iNOS (Fig. 2E), in the whole brain tissues of the LPS-treated WT and IL-10KO mice at 1, 3, 6, and 12 h after treatment by real-time PCR. Our data showed that after intraperitoneal LPS injection, the mRNA levels of TNF-α (treated saline vs. one and 3 h LPS treatment in WT mice: P < 0.001; F (4, 20) = 94.42, two-way ANOVA), IL-1β (treated saline vs. 1 h LPS treatment in WT mice: P < 0.01; treated saline vs. 3 h LPS treatment in WT mice: P < 0.001; F (4, 20) = 78.30, two-way ANOVA), COX-2 (treated saline vs. 1 h LPS treatment in WT mice: P < 0.01; treated saline vs. 3 h LPS treatment in WT mice: P < 0.001; F (4, 21) = 62.21, two-way ANOVA), and iNOS (treated saline vs. 3 h LPS treatment in WT mice: P < 0.001; F (4, 20) = 18.51, two-way ANOVA) were significantly increased at 1 h, reached the peak at 3 h and then approximately returned to basal level at 12 h in the WT mice's brain. The expression profile of these proinflammatory factors during this time frame reflects three typical immune statuses of an inflammatory process.

Furthermore, LPS-treated IL-10KO group had higher TNF-α (WT vs. IL-10KO mice in three or 6 h LPS treatment: P < 0.01, P < 0.01; F (1, 20) = 47.33, two-way ANOVA), IL-1β (WT vs. IL-10KO mice in three or 6 h LPS treatment: P < 0.05, P < 0.01; F (1, 20) = 15.62, two-way ANOVA), and COX-2 (WT vs. IL-10KO mice in 6 h LPS treatment: P < 0.001; F (1, 21) = 2.049, two-way ANOVA) expression than that of the WT group at 3 or 6 h, but not 1 h. These data suggested that the induction of endogenous IL-10 had no change in the initiation stage of their neuroinflammatory process but rather suppressed the immune propagation. Moreover, at 12 h of LPS treatment, WT and IL-10KO had similar levels of TNF-α, IL-1β, and COX-2 mRNA, implying that endogenous IL-10 induction did not participate in regulating their immune resolution. Interestingly, compared to the corresponding WT mice, the iNOS transcript in the brain of LPS-treated IL-10KO mice was significantly ∼3-fold increased at 6 and 12 h (WT vs. IL-10KO mice in six or 12 h LPS treatment: P < 0.001; F (1, 20) = 16.85, two-way ANOVA). Our data indicated that, besides the broader impact of IL-10 on the LPS-induced neuroinflammatory process, its endogenous induction may also mainly contribute to the immune resolution of iNOS rather than other proinflammatory factors. Thus, this study further focused on inductive endogenous IL-10-mediated iNOS resolution.

3.3. IL-10KO microglia increase LPS-induced nitrite production in a time- and dose-dependent manner

Since iNOS is an enzyme that converts l-arginine, ultimately leading to microglial nitrite production under brain immune challenges (Liu et al., 2002; Mitrovic et al., 1994; Saha and Pahan, 2006), we determined whether endogenous IL-10 induction suppressed LPS-induced nitrite production. Mixed-glia cultures containing astroglia and microglia were prepared from postnatal day 1 WT and IL-10KO pups and then treated with LPS (Fig. 3A). During the experimental period, the levels of supernatant nitrite in these cells were measured by Griess reagent assay. Our data showed that LPS treatment induced the production of supernatant nitrite both in WT and IL-10KO mixed-glia cells (Fig. 3B). Supernatant nitrite in the WT cells peaked 72 h after LPS treatment (P < 0.05, F (6, 57) = 21.66, two-way ANOVA) and began dropping later at 96 h. However, increased supernatant nitrite levels in LPS-treated IL-10KO cells were continually kept extended to 96 h (saline treatment vs. hour 48, hour 72, or hour 96 LPS treatment in IL-10KO mice: P < 0.001; F (6, 57) = 21.66, two-way ANOVA). At 72 and 96 h, IL-10KO cells produced a ∼2-fold amount of supernatant nitrite when treated with LPS compared to the corresponding WT group (WT vs. IL-10KO mice in seventy-two- or 96-h LPS treatment: P < 0.05 or P < 0.01; F (1, 57) = 18.38, two-way ANOVA). Furthermore, under treatment with LPS at 10 and 20 ng/ml (WT vs. IL-10KO mice in 72-h LPS treatment: P < 0.05; F (1, 24) = 39.36, two-way ANOVA) concentrations, a greater amount of supernatant nitrite in the IL-10KO cells was also observed at 72 h (Fig. 3C). Our data indicated that endogenous IL-10 induction suppressed LPS-induced microglial nitrite production at a late stage in a manner that depended on both time and dose.

Fig. 3.

IL-10KO microglia increase LPS-induced nitrite production in a time- and dose-dependent manner.

(A) Schematic of the experimental design to determine the effects of endogenous IL-10 induction on nitrite production in microglia under LPS challenge. (B, C) Measurement of supernatant nitrite in LPS-treated WT and IL-10KO mixed-glia cultures at different time points (B) or concentrations (C) using a Griess reagent. Data are expressed as the mean ± SEM (n ≥ 3) with two-way ANOVA with Tukey's multiple comparisons test (Saline group versus LPS group: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; WT group versus IL-10KO group: #p < 0.05, ##p < 0.01).

3.4. IL-10 deficiency enhances microglia activation at the late-stage inflammation

To determine whether the deficiency in endogenous IL-10 induction affects LPS-induced microglial activation, treatment of WT and IL-10KO mixed-glia cells with or without LPS at different time courses was performed (Fig. 4A) and then subjected to immunocytochemistry staining of Iba-1, which is one of the markers of microglia. Our data showed that the number of IL-10KO microglia (Iba-1-positive cells) was similar to WT microglia at 24, 48, and 72 h (Fig. 4B–E) in vitro incubation but increased at 96 h (Fig. 4B–F) in the absence of LPS (P < 0.05; F (1, 36) = 5.381, two-way ANOVA). Furthermore, after LPS treatment, WT and IL-10KO microglia had a similar increase in number compared to their corresponding saline-treated control, implying the increased nitrite production in LPS-treated IL-10KO microglia was not due to a change in their cell number. Moreover, both WT and IL-10KO microglia displayed ramified morphology in normal conditions and changed to amoeba-like morphology along with increased cell body size and larger cellular process arms during the LPS-elicited neuroinflammatory process (Fig. 4B). Interestingly, compared to the early stage, intensity of Iba-1expression in the IL-10KO group was increased at 96 h after LPS treatment (Fig. 4G–J) (P < 0.05; F (1, 36) = 16.37, two-way ANOVA), suggesting IL-10KO microglia became overactive at the late-stage inflammation. On the other hand, the increased Iba-1 intensity of the IL-10KO group was also found at 24 (P < 0.05; F (1, 36) = 11.73, two-way ANOVA) and 96 h in normal conditions. Together, our data implied that the induction of endogenous IL-10 may play a role in regulating microglia activation at the late stage due to failure of nitrite resolution.

Fig. 4.

IL-10 deficiency enhances microglia activation at the late-stage inflammation. (A) Schematic of the experimental design for the study of endogenous IL-10 on microglia number and morphology in normal and inflammatory conditions. (B) The expression of Iba-1, a marker for microglia, in these mixed-glia cells at 24, 48, 72, and 92 h in the presence and absence of LPS (15 ng/ml) was detected by immunocytochemistry staining. Magnification: 200×; scale bar: 100 μm. (C-F) Quantitative data of microglial number (Iba-1-positive cells) in these cells treated with LPS at 24 (C), 48 (D), 72 (E), and 96 (F) hours are shown. Data are expressed as the percentage of the WT saline group (mean ± SEM, each dot represented an individual field from three independent experiments) with two-way ANOVA with Tukey's multiple comparisons test (WT group versus IL-10KO group: ∗∗∗p < 0.001; NS indicates no significant difference). (G–J) The density of Iba-1 immunostaining in these LPS-treated mixed cells at 24 (G), 48 (H), 72 (I), and 96 (J) hours was measured and quantified. Data are shown as integrated intensity and expressed as the percentage of the WT saline group (mean ± SEM, each dot represented an individual field from three independent experiments) with two-way ANOVA with Tukey's multiple comparisons test (WT group versus IL-10KO group: ∗p < 0.05; NS indicates no significant difference).

3.5. IL-10KO microglia fail to program endotoxin tolerance of nitrite

Changes in the level of IL-10 and iNOS/nitrite have been well-recognized as the immune hallmark of function-polarized microglia (Orihuela et al., 2016). For example, endotoxin-tolerized microglia exhibit an anti-inflammatory polarization, showing increased IL-10 levels but decreased nitrite expression (Chu et al., 2016). However, the regulatory mechanism of nitrite tolerance in microglia remains unknown. Therefore, we determined whether the enhancement of inductive endogenous IL-10 played a role in nitrite reduction in tolerized microglia. Like the treatments for LPS-tolerized microglia in Fig. 1, neuron-glia cells were pre-treated with a vehicle (the LPS-treated group) or 15 ng/mL of LPS (the LPS/LPS-treated group) for 6 h and then replaced with fresh serum media. LPS (15 ng/mL) was re-added to the neuron-glia cells 6 h later. After LPS treatment, the mRNA level of IL-10 and iNOS of the treated neuron-glia cells was measured by RT-PCR (Fig. S2A). Our data showed that once LPS treatment significantly induced the expression of both iNOS (P < 0.001, F (3, 8) = 410.4, one-way ANOVA) (Fig. S2B) and IL-10 (P < 0.001, F (3, 8) = 623.9, one-way ANOVA) (Fig. S2C) at the mRNA level. After receiving the second LPS challenge, LPS-tolerized microglia displayed decreased iNOS expression (P < 0.001, F (3, 8) = 410.4, one-way ANOVA). Still, they had an increase in IL-10 (P < 0.001, F (3, 8) = 623.9, one-way ANOVA), showing the anti-inflammatory polarization in tolerized microglia.

According to the results in Fig. 4, IL-10KO microglia had significantly higher nitrite production than WT microglia at 72 h after LPS treatment. Furthermore, the WT and IL-10KO mixed-glia cells were prepared and subjected to the same LPS tolerance treatment regimen. Supernatant levels of IL-10 and nitrite in these cells were measured 72 h after treatment (Fig. 5A). ELISA data showed that the WT mixed-glia cells with the repeated LPS treatment (LPS/LPS group) enhanced IL-10 induction compared to the once LPS-treated cells (LPS group) at 72 h (Fig. 5B, LPS group vs. LPS/LPS group in WT cells: P < 0.001; F (1, 8) = 26.73, two-way ANOVA). IL-10KO cells had no IL-10 production in the single LPS or repeated LPS treatment. Also, the nitrite level at 72 h after LPS treatment was measured in WT and IL-10KO cells (Fig. 5C). Consistent with Fig. 3, our data revealed that IL-10KO cells produced more nitrite than WT cells at their supernatural level at the late stage of the LPS-induced inflammatory process (WT vs. IL-10KO cells in LPS-treated group: P < 0.01; F (1, 8) = 165.0, two-way ANOVA), suggesting that endogenous IL-10 induction participated in nitrite resolution. Additionally, in response to this LPS tolerance regimen, the WT cells with the repeated LPS treatment produced lower nitrite than those with the once LPS treatment (LPS group vs. LPS/LPS group in WT cells: P < 0.001; F (1, 8) = 25.13, two-way ANOVA), indicating that WT cells were capable of developing nitrite tolerance. Importantly, LPS-tolerized IL-10KO cells still produced more amount of supernatant nitrite compared to the corresponding WT cells (WT vs. IL-10KO cells in LPS/LPS group: 0.001; F (1, 8) = 165.0, two-way ANOVA) and had similar nitrite levels to those of treated IL-10KO cells with LPS alone, implying that IL-10KO cells failed to develop nitrite tolerance. Our data indicated that the enhancement of endogenous IL-10 induction facilitated the nitrite tolerance, contributing to anti-inflammatory polarization in LPS-tolerized microglia.

Fig. 5.

IL-10KO microglia fail to program endotoxin tolerance of nitrite.

(A) Illustration of experimental procedure for studying IL-10's role in nitrite tolerance in LPS-tolerized microglia. WT and IL-10KO mixed-glia cultures were pre-incubated with a vehicle or LPS (15 ng/ml) for 6 h. After 6 h of replacing it with fresh medium, LPS was re-added to the cultures. (B) After receiving the ET treatment regimen (LPS versus LPS/LPS) as indicated, supernatant levels of IL-10 in WT and IL-10KO mixed-glia cultures were detected 72 h later by ELISA assay. Data are expressed as the mean ± SEM from three independent experiments in duplicate with two-way ANOVA with Tukey's multiple comparisons test (LPS group versus LPS/LPS group: ∗∗∗p < 0.001). (C) Supernatant levels of nitrite in WT and IL-10KO mixed-glia cultures under an endotoxin treatment regimen were detected 72 h later by a Griess reagent. Data are expressed as the mean ± SEM from three independent experiments in duplicate with two-way ANOVA with Tukey's multiple comparisons test (LPS group versus LPS/LPS group: ∗∗∗p < 0.001; WT group versus IL-10KO group: ##p < 0.01, ###p < 0.001; NS indicates no significant difference).

4. Discussion

Upon immune stimulation, the induction of individual endogenous cytokines displays a rigid expression profile, leading to a stereotyped neuroinflammatory process consisting of initiation, propagation, and resolution stages. In reality, the interactions among these pro- and anti-inflammatory factors during a neuroinflammatory cascade complexly shape their own expression profiles along with restricted time frames. This study determined how the inductive endogenous IL-10, a potent anti-inflammatory cytokine, regulated the immune status of the neuroinflammatory process and diverse activation of microglia using an in vivo neuroinflammatory model of LPS regimens and in vitro brain cell cultures with once or repeated LPS treatment. Our data revealed the cell-autonomous induction of endogenous IL-10 by LPS-activated and LPS-tolerized microglia. Under a single intraperitoneal 0.3 mg/kg dose of LPS injection, inductive endogenous IL-10 could be detected at the level of serum, liver, and brain tissues in 1 h in WT mice but not IL-10KO mice. Further RT-PCR analysis showed that compared to the WT mice, which had endogenous IL-10 induction, IL-10KO mice had increased the mRNA levels of the LPS-elicited proinflammatory factors (TNF-α, IL-1β, COX-2, and iNOS) in the brains at the stage of immune propagation (3- and 6-h LPS), not initiation (1-h LPS) status. At the immune resolution stage (12-h LPS), the levels of LPS-induced TNF-α, IL-1β, and COX-2 were similarly dropped in both WT and IL-10KO mice's brains. Interestingly, only the LPS-increased iNOS expression was kept high in the IL-10KO group, not the WT group, indicating the deficiency in IL-10 impaired the immune resolution of iNOS, but not TNF-α, IL-1β, and COX-2. Moreover, the Griess reagent assay showed that the production of nitrite, which is generated by a critical enzyme iNOS, is significantly increased at the last stage in LPS-treated IL-10KO microglia in a time and dose-dependent manner. On the other hand, LPS-tolerized IL-10KO microglia failed to program immune tolerance of nitrite, implying that the enhancement of endogenous IL-10 induction participates in developing anti-inflammatory polarization of LPS-tolerized microglia, where nitrite level is reduced, but IL-10 is enhanced. Together, our data demonstrated that microglia-autonomous induction of endogenous IL-10 contributes to immune resolution and tolerance of nitrite.

The LPS injection animal model has been well-documented in determining endotoxemia (Mohammad and Thiemermann, 2020) and inflammatory bowel disease (Williams et al., 2013) and studying neuroinflammation-mediated neurological disorders (da Silva et al., 2024), including depression (Yin et al., 2023) and neurodegenerative diseases (Deng et al., 2020). Depending on administration routes, animal age, sex, endotoxin source, and dose, LPS-induced neuroinflammation models are heterogeneous. In this study, a single IP injection of LPS with a 0.3 mg/kg concentration into young male mice was performed to determine the role of endogenous IL-10 neuroinflammation and the immune tolerance process. Several reports investigated that such a low concentration of LPS IP injection daily for four consecutive days indirectly increased expression levels of glial markers (GFAP and Iba-1) and oxidative stress marker (3-nitrotyrosine) in the olfactory bulb, hippocampus, midbrain, and cerebellum through induced peripheral inflammation-mediated T cells infiltration into the brain. Meanwhile, the knockout TNF-α receptor terminates the peripheral inflammation-mediated microglial activation (Qin et al., 2007). Furthermore, the behavioral tests reveal that mice with this systemically administered LPS display olfactory impairment and anxiety-like behavior as the symptoms of early non-motor Parkinson's disease on days 14–18 post-treatment (Deng et al., 2021). Due to the different ages, the single systemic 0.3 mg/kg LPS dose administration is found to be highly lethal to aged mice (36 % mortality) but not in young and middle mice (Mouton et al., 2012). Our data showed that a single 0.3 mg/kg LPS IP injection could induce a period of short-term, recoverable peripheral and CNS inflammatory events in 12 h, including endogenous IL-10 induction, without any mortality. By contrast, deficiency in IL-10 resulted in the death of all mice within three days after LPS injection (Fig. S1D), emphasizing the crucial role of endogenous IL-10 in safeguarding mice from excessive inflammatory responses. However, the CNS- or microglia-specific IL-10 knockdown animal model is needed to determine further endogenous IL-10's role in the immune resolution and homeostasis of the brain.

As the first defense and inflammatory responder, microglia temporally secrete a broad spectrum and various immunoregulatory factors, including the early-released pro-inflammatory factors and later-on anti-inflammatory IL-10, to protect the neurons against invading pathogens (Hoogland et al., 2015), and terminate the immune initiation and propagation of the neuroinflammatory process for immune resolution and homeostasis (Orihuela et al., 2016). Therefore, elucidating the regulatory mechanism of endogenous IL-10 induction and its impact on glial activation and neuronal survival is crucial for developing anti-inflammatory strategies for neuroinflammation-mediated brain disorders. In addition to well-known intracellular signaling pathways for induction of endogenous IL-10, the early-release proinflammatory TNF-α-PGE2 cascades negatively regulate endogenous IL-10 production through EP2-mediated β-arrestin signaling during an inflammatory process in microglia (Chu et al., 2015). On the other hand, due to its later expression profile, microglia-secreted IL-10 has no effect on the production of LPS-induced TNF-α in the mixed-glia cell model (Kuo et al., 2023). Furthermore, through downregulating NLRP3 inflammasome activation and caspase-1-dependent IL-1β maturation, inductive endogenous IL-10 is capable of preventing IL-1β, chronic neuroinflammation, and dopaminergic neurodegeneration (Gao et al., 2020). Our data revealed that deficiency in IL-10 enhanced the mRNA level of most proinflammatory factors (TNF-α, IL-1β, COX-2, and iNOS) at the immune propagation stage in the mice's brains under LPS IP injection. Interestingly, similar to the WT group, TNF-α, IL-1β, and COX-2 expression in the IL-10KO group went down to basal levels at the immune resolution stage. However, the increased iNOS expression in the IL-10KO group, not the WT group, still remained at the immune resolution stage as high as at the immune propagation stage. Consistently, the supernatant nitrite level in IL-10KO mixed-glia cells was enhanced at the later time points after LPS treatment. Most recently, a study reports that the early-released IL-10 also plays a role in regulating the appropriate degree of neuroinflammation via the β2-adrenergic signal at the immune initiation stage (Wang et al., 2021). Accordingly, the multi-omics approaches are required for determining the spatial-temporal expression profiles of inductive IL-10 production by the multiple intracellular (Hilliard et al., 2020; Kang et al., 2019; Kwon et al., 2022; Xing et al., 2011) and extracellular (Chu et al., 2015) signaling cascades and its role in immune regulation and innervation at different immune statuses of an inflammatory process in the brain at the levels of mRNA, protein, supernatant, and cell type.

Under repeated LPS stimulations, microglia instigate (“sensitize”) the synthesis of anti-inflammatory mediators, such as IL-10 (Chu et al., 2016). In contrast, the production of pro-inflammatory mediators (TNF-α, IL-1β, PGE₂, and nitrite) is decreased (“tolerate”). LPS-tolerized microglia display an anti-inflammatory polarization (Chu et al., 2016) and become refractory to a subsequent endotoxin challenge, a neuroprotective mechanism targeted at preventing excessive toxic damage from cytokine production (Kuo et al., 2023). Our data revealed that increased IL-10 production in the tolerized microglia is critical for their anti-inflammatory polarization by inhibiting nitrite production. In other words, enhanced induction of endogenous IL-10 is required for nitrite tolerance in the tolerized microglia. In addition, a regulatory mechanism study for microglial ET indicates that the intracellular p38 MAPK signaling pathway, rather than the subsequent extracellular IL-10 autocrine cascade, contributes to the tolerance formation to early-released TNF-α (Kuo et al., 2023), suggesting IL-10 induction produced at a later stage does not effectively regulate the production of the earlier inflammatory factors. Noteworthily, the presence of other brain cells, such as neurons and astroglia, regulates the endotoxin tolerance capacity of microglia in the reduction of TNF-α and PGE₂ through M-CSF-mediated ERK signals (Chu et al., 2016) or the cell-cell contacts (Kuo et al., 2022a) to form a self-protective cycle. Whether other brain cells participate in the immune balance of IL-10 (sensitized) and nitrite (desensitized) production in LPS-tolerized microglia is still unclear. Our data showed that microglia alone display a similar expression profile of endogenous IL-10 in both single and repeated LPS treatments compared to microglia with the presence of neurons and astroglia, suggesting a microglia-autonomous induction of endogenous IL-10 in the inflammatory brains. Accordingly, IL-10 and nitrite are mainly produced from the LPS-treated microglia, suggesting microglia develop endotoxin tolerance of nitrite via an IL-10-dependent autocrine mechanism. In addition to immune tolerance, many studies have demonstrated that microglia are capable of developing trained immunity, which is another type of innate immune memory, to enhance subsequent immune responses (Neher and Cunningham, 2019; Wendeln et al., 2018). Interestingly, the trained immunity of microglia enhances the accumulation of AD hallmarks, whereas microglial immune tolerance ameliorates AD pathological hallmarks. Besides the formation of anti-inflammatory polarization in tolerized microglia, it will be interesting to understand the regulatory mechanism of microglia-trained immunity by inductive endogenous IL-10 in neuroinflammation-mediated progressive neurodegenerative diseases.

Although our results are consistent with the role of inductive endogenous IL-10 in nitrite regulation in vivo and in vitro models of LPS-induced neuroinflammation, this study has a few limitations that require awareness. First, due to variability stemming from the effect of female hormones (Liu et al., 2008), only male mice were employed for the LPS immune challenge to increase the sensitivity of the brain neuroinflammation. It would be worthwhile to investigate whether the female counterparts recapture the similar impact of endogenous IL-10 in immune resolution and tolerance. Second, in this study, the preparation of mixed glia cocultures containing 80 % astroglia and 20 % microglia to measure the LPS-induced IL-10 and nitrite production cannot avoid the possible immunoregulatory effects of astroglia or the absence of neurons. However, microglia-enriched cultures do not allow us to measure the levels of IL-10 and nitrite in the inflammatory process in the long term and study the role of IL-10 induction in nitrite tolerance since microglia alone are more susceptible to semi-activated and die one day after cell seeding (Pesti et al., 2024) and fail to develop endotoxin tolerance (Chu et al., 2016). On the other hand, our data revealed that expression patterns of IL-10 induction in LPS once or repeated LPS among neuron-glia, mixed glia, and microglia-enriched cultures were similar, suggesting that neurons and astroglia had no effects on LPS-induced IL-10 production.

The regulatory mechanisms of iNOS expression by IL-10 may include positive transcriptional control of type-2 cationic amino acid transporter (CAT-2) (Huang et al., 2002), suppressor of cytokine signaling 3 (SOCS-3) as a negative feedback regulator (Berlato et al., 2002), and heme oxygenase-1 (HO-1), which is a stress-inducible protein (Lee and Chau, 2002), and activation of STAT3-mediated transcriptional inhibition (Murray, 2005). However, the mechanism by which endogenous IL-10 induction regulates resolution and tolerance of nitric oxide by inhibiting iNOS expression during the inflammatory process in microglia remains to be elucidated. Further study using recombinant IL-10 protein, IL-10 neutralizing antibody, or inhibitors for IL-10 receptor and its downstream signaling molecules, such as STAT3 and SOCS3, is required. Finally, because IL-10 whole-body knockout mice display immediate severe mortality under a single intraperitoneal injection of LPS at such a small amount concentration (0.3 mg/kg), the establishment of brain tissue- or microglia-specific IL-10KO mice is needed to determine further the pathophysiological role of inductive endogenous IL-10 in CNS or microglia at the behavioral, cellular, and molecular levels in a future study.

5. Conclusions

To summarize, using in vitro and in vivo LPS-induced inflammation or immune tolerance models, our research demonstrates that the cell-autonomous induction of endogenous IL-10 in microglial activation is critical for dampening brain immune responses, particularly in nitric oxide resolution and tolerance, contributing to the formation of anti-inflammatory microglia. Targeting the negative feedback loop from inducible endogenous IL-10 to reduce nitrite may be a promising microglia-based therapeutic strategy for inflammation-mediated neurodevelopmental and neurodegenerative disorders.

CRediT authorship contribution statement

Hsing-Chun Kuo: Writing – original draft, Resources, Data curation. Jia-Shing Chen: Supervision, Methodology. Chun-Nun Chao: Writing – original draft, Resources, Conceptualization. Kam-Fai Lee: Methodology, Investigation, Data curation. Yi-Te Huang: Formal analysis, Data curation. Pin-Cheng Mao: Methodology, Data curation. Tzu-Chia Lin: Methodology, Data curation. Shu-Chen Chiu: Methodology. Ya-Ling Huang: Writing – original draft, Resources, Conceptualization. Chun-Hsien Chu: Writing – review & editing, Validation, Supervision, Funding acquisition, Conceptualization.

Ethics approval

The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Animal Care and Use Committee of Chang Gung University of Science and Technology.

Availability of data and materials

All data and materials, as well as software application information, are available in the manuscript or are available from the corresponding author upon reasonable request.

Funding

This work was supported by grants NSTC 111-2320-B-006-018, NSTC 112-2320-B-006-035, and NSTC 113-2320-B-006-037 from the Taiwan National Science and Technology Council. This study was supported by the grants BMRPD42, CMRPF6N0031, CMRPF6N0032, CMRPF6N0033, and ZRRPF6N0011 from Chang Gung Memorial Hospital, Chiayi, Taiwan, and Chang Gung University of Science and Technology, Chia-Yi Campus, Taiwan.

Declaration of competing interest

The authors declare no competing financial interests.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.