A20 Ameliorates Intracerebral Hemorrhage-Induced Inflammatory Injury by Regulating TRAF6 Polyubiquitination [INNATE IMMUNITY AND INFLAMMATION]
Reducing excessive inflammation is beneficial for the recovery from intracerebral hemorrhage (ICH). Here, the roles and mechanisms of A20 (TNFAIP3), an important endogenous anti-inflammatory factor, are examined in ICH. A20 expression in the PBMCs of ICH patients and an ICH mouse model was detected, and the correlation between A20 expression and neurologic deficits was analyzed. A20 expression was increased in PBMCs and was negatively related to the modified Rankin Scale score. A20 expression was also increased in mouse perihematomal tissues. A20−/− and A20-overexpressing mice were generated to further analyze A20 function. Compared with wild-type (WT) mice, A20−/− and A20-overexpressing mice showed significant increases and decreases, respectively, in hematoma volume, neurologic deficit score, mortality, neuronal degeneration, and proinflammatory factors. Moreover, WT-A20−/− parabiosis was established to explore the role of A20 in peripheral blood in ICH injury. ICH-induced damage, including brain edema, neurologic deficit score, proinflammatory factors, and neuronal apoptosis, was reduced in A20−/− parabionts compared with A20−/− mice. Finally, the interactions between TRAF6 and Ubc13 and UbcH5c were increased in A20−/− mice compared with WT mice; the opposite occurred in A20-overexpressing mice. Enhanced IκBα degradation and NF-κB activation were observed in A20−/− mice, but the results were reversed in A20-overexpressing mice. These results suggested that A20 is involved in regulating ICH-induced inflammatory injury in both the central and peripheral system and that A20 reduces ICH-induced inflammation by regulating TRAF6 polyubiquitination. Targeting A20 may thus be a promising therapeutic strategy for ICH.
Intracerebral hemorrhage (ICH), which results in high mortality and morbidity, accounts for 10–15% of all stroke cases; mortality is especially high in China, but effective therapies remain limited (1–3). Previous studies have indicated that inflammation plays a key role in ICH-induced secondary injury and that microglia activation, peripheral inflammatory cell infiltration, and the release of proinflammatory factors all participate in the pathogenesis of inflammation (4, 5). Most previous studies have focused on the role of inflammatory factors. As a result, endogenous anti-inflammatory factors that can ameliorate ICH-induced inflammation have been overlooked (6–8). Thus, targeting endogenous anti-inflammatory factors could provide insights into the development of new ICH therapies.
Zinc finger protein A20 (TNFAIP3) is an endogenous anti-inflammatory factor that can reduce the expression of destructive proinflammatory factors such as IL-1β and TNF-α by inhibiting NF-κB activation (9, 10). A20 also interacts with IL-10 and TGF-β and reduces inflammation in diabetes and arthritis (11–13). Evidence has shown that A20 is essential for restricting TLR-induced TNF receptor–associated factor (TRAF) 6 ubiquitination (11, 14). Our previous studies have shown that TLR4/NF-κB signaling plays an important role in ICH-induced inflammatory injury (15, 16). Thus, we hypothesized that A20 may also ameliorate ICH-induced inflammation by negatively modulating TLR signaling.
The present study aimed to investigate two hypotheses: 1) A20 can ameliorate morphological (hematoma and brain edema) and behavioral outcomes after experimental ICH in mice; and 2) A20 can reduce NF-κB activation by regulating TRAF6 polyubiquitination.
Materials and Methods
To evaluate the relationship between A20 expression and neurologic deficits, we screened 34 patients with ICH and 34 healthy controls for variations in A20 expression. A total of 49 patients with ICH were identified between December 2013 and June 2015. The inclusion criteria were as follows: 1) identification of patients with primary ICH within 24 h of disease onset and the extraction of blood from the cubital fossa vein before the administration of any drugs; 2) hemorrhage in the basal ganglia confirmed by computerized axial tomography; and 3) diagnosis according to the criteria of the Fourth Chinese National Meeting for Cerebrovascular Disease (1996) and the European Stroke Initiative (17). The exclusion criteria were as follows: 1) patient age <18 or >80 y; 2) surgical history within the last 6 mo; 3) coma or death within 48 h after admission; 4) hematoma caused by trauma, drug abuse, brain tumor, vascular malformations, anticoagulation therapy, or coagulation abnormalities; 5) obvious inflammatory disease (e.g., infectious disease, systemic lupus erythematosus (SLE), or rheumatoid arthritis; 6) presence of a hospital-acquired infection; 7) acute and chronic hepatopathy; 8) diabetes diagnosis; and 9) lack of agreement with the study protocol or inability to undergo all of the tests required by the study. A total of 15 patients were excluded from the study, including 3 patients who died and 12 who were lost to follow-up. A total of 34 patients were enrolled in the study (for general clinical characteristics of ICH patients and the healthy control group, see Table I). Upon admission to our hospital, 5 ml cubital vein blood was extracted for A20 detection using EDTA as an anticoagulant agent. PBMCs were isolated using a PBMC kit, according to the manufacturer’s instructions (10771; Sigma), to determine A20 expression. Modified Rankin Scale (mRS) scores were used to evaluate neurologic deficits in the patients 3 mo after ICH (18). Correlation analysis was used to analyze the relationship between A20 mRNA expression and the patient mRS scores. All blood samples collected from the healthy controls and patients were used with their informed consent, and the study procedures were approved by the Ethics Committee of Xinqiao Hospital of the Third Military Medical University, China, and conducted in accordance with the Declaration of Helsinki and its amendments.
View this table:
Table I.General clinical characteristics of ICH patients and healthy control group
C57BL/6 mice (male, 8-wk-old, 18–22 g) were obtained from the Animal Center of the Third Military Medical University (Chongqing, China). A20−/− mice were purchased from Riken BioResource Center, Japan. The C57BL/6 genetic background was maintained over the course of six generations of hybridization in the A20−/− mice, and these mice were identified using PCR. All mice were raised in specific pathogen-free–grade animal rooms in a temperature- and light-controlled environment. The mice were also provided with unlimited access to food and water. All procedures were approved by the Animal Management Committee of the Third Military Medical University, China.
Parabiosis was developed as described previously (19, 20). Briefly, the mice were anesthetized by i.p. injection with Zoletil (25 mg/kg containing 0.2% Rompun). The mice were then shaved and a unilateral flank skin incision from the elbow to the knee joint was created. The skin edge of the mice was sutured with 5.0 prolene (Ethicon). The suture area was daubed with erythromycin ointment, and the partners were fixed with gauze to avoid tearing the wound. The wild-type (WT) mouse was used as the donor parabiont and the A20−/− mouse as the recipient partner. At the same time, WT-WT parabionts were constructed as controls. ICH was induced in the WT or A20−/− parabionts 20 d after parabiotic surgery.
The ICH model was induced according to our previously described procedure (16). Whole blood (20 μl) without anticoagulant or the same volume of saline was injected at 2 μl/min into the right striatum (0.8 mm anterior and 2 mm lateral of bregma and at a depth of 3.5 mm) using a syringe pump (KD Scientific, Holliston, MA). The success rate of the ICH model was 90%; failed models and dead mice were excluded from the experiment.
The number of dead mice was counted at 1, 3, 5, and 7 d after ICH, and the survival rate was calculated as follows: (the number of ICH mice per group – the number of dead mice per group)/the number of ICH mice per group. Kaplan–Meier survival plots were produced using a log-rank test in GraphPad Prism 5. Significance was set at p < 0.05.
Construction of the A20 overexpression mouse model
PCR primers for A20 (NM_001270508.1) and the lentiviral vector (pLenti6.3_MCS_IRES2-EGFP) were designed according to the Homo sapiens full open reading frame TNFAIP3 sequence. A cDNA clone of A20 was transcribed, and the product was amplified using primers with flanking NheI and ASCI restriction sites. The DNA was then inserted into pLenti6.3_MCS_IRES2-EGFP. The primer sequences are listed in Table II. The A20 recombinant vector was named LV-A20, and the titer of the virus was 1.2 × 108 TU/ml. LV-A20 virus solution (3 μl) was injected into the lateral ventricles (0.3 mm posterior and 1 mm lateral of bregma and at a depth of 2.75 mm) of the mice using a 5-μl syringe pump (KD Scientific). The needle was held in place for 20 min, and the microinjector was then drawn out slowly. The skull was sealed with bone wax, and the scalp was sutured. A20 expression was detected by quantitative real-time PCR (qRT-PCR), Western blotting, and immunofluorescence.
View this table:
Table II.Primers used for quantitative real-time PCR
Lentiviral short hairpin RNAs infection
Recombinant lentiviral (LV) short hairpin RNAs (shRNAs) targeted to mouse TRAF6 (Gene ID: 22034, named LV-TRAF6 shRNA) were designed (shRNA1: 5′-GCAAGTATGAGTGTCCCATCT-3′; shRNA2: 5′-GCTGTCCTCTGGCAAATATCA-3′; shRNA3: 5′-GGAGGACAAGGTTGCCGAAAT-3′) by Shanghai GenePharma (China). Negative-control LV particles were also designed (5′-GTCTTCGAACGTATCAAGT-3′) and packaged. The final titers of LV-TRAF6 shRNA and negative-control LV particles were 2 × 108 TU/ml. A total of 5 μl of virus solution was injected into the lateral ventricles of mice using a microinjector. TRAF6 knockdown efficiency was confirmed by Western blotting (see Supplemental Fig. 2A). LV-TRAF6 shRNA1 showed the best knockdown effect, and was therefore selected for lateral ventricle injection. Then, 7 d after LV-TRAF6 shRNA injection, the mice underwent surgery for the ICH model. Subsequently, we harvested the brain tissues and tested NF-κB expression and downstream proinflammatory cytokine release by Western blotting and ELISA, respectively.
Neurological deficit scores
The experiment was conducted according to our previous methods (16). A 28-point neurologic deficit score (NDS) assessment system was adopted. Climbing, front limb symmetry, circling behavior, and body symmetry were assessed. The scoring was completed by two blinded laboratory researchers who were unaware of the group assignments of the mice. The average score was used as the final score for each mouse.
The experiment was conducted according to our previous methods (15). Briefly, mice were anesthetized with a lethal dose of chloral hydrate, then perfused and fixed. The brains were then removed and sectioned from the frontal lobe to the occipital lobe to prepare continuous 40-μm-thick coronal sections. One out of every five sections was collected, and the fixed brain slices were arranged in order. Image-Pro Plus 5.0 image processing software (Media Cybernetics, Bethesda, MD) was used to measure the hematoma volume (microliter ) according to the equation V = t × (A1 +…+ An), where V is the hematoma volume, t is the slice thickness, and Ai is the area of bleeding. Hemoglobin content in the hematoma tissues was also measured to further quantify hematoma size (19). Blood (0, 2, 4, 8, 16, and 20 μl) was added to fresh brain homogenates, and the OD values of the samples were measured at 540 nm using a spectrophotometer (Thermo Multiskan, Pittsburgh, PA). These values were then used to plot a standard curve. The striatum of mice in each group was removed 3 d after ICH and dissolved in Drabkin’s reagent. The supernatant of the homogenates was collected, and the OD was measured using a spectrophotometer; the hematoma volume was then calculated using the standard curve.
Brain water content
As described in our previous report (16), mice were randomly selected from each group at 1, 3, and 7 d after ICH to detect the brain water content. Briefly, the cerebral tissues were removed after the mice were anesthetized with chloral hydrate, and then the samples were divided into five parts: the ipsilateral cortex, the ipsilateral basal ganglia, the contralateral cortex, the contralateral basal ganglia, and the cerebellum. The brain water content was measured using the following formula: [(wet weight – dry weight)/wet weight] × 100.
Quantitative real-time PCR
Total RNA was extracted from the tissue using TRIzol reagent (Invitrogen, Gaithersburg, MD), and qRT-PCR was performed according to the manufacturer’s instructions (Takara Biotechnology, Dalian, China). GAPDH was used as an internal control. The primers used in the study are listed in Table II. Relative levels of mRNA expression were calculated using the 2−△△CT method.
Proteins from perihematomal tissues were resolved by SDS-PAGE and transferred onto polyvinylidene fluoride membranes by electroblotting (16). The membranes were incubated overnight with a rabbit anti-mouse TNFAIP3 Ab (1:200, ab13597; Abcam) at 4°C. GAPDH was used as a loading control, and membranes were incubated with HRP-conjugated goat anti-rabbit secondary Abs (1:200; Millipore) at 25°C for 1.5 h. Bound Abs were visualized using a chemiluminescence detection system. Signals were quantified by scanning densitometry and computer-assisted image analysis. Protein levels were expressed as the ratio of the value of the detected protein band to the GAPDH band.
Following our previously described method (21), mouse brain tissues were fixed in 4% paraformaldehyde. The samples were subjected to gradient dehydration and were then embedded, frozen, and cut into 5-μm-thick sections. A20 expression was then detected using double fluorescence immunohistochemistry. The primary Abs used in this experiment are listed in Table III. Positive cells were calculated and analyzed in three different arbitrary units that can be defined as the average number of positive cells in three randomly selected fields (800×) in three slices of each experiment group.
View this table:
Table III.Abs used for immunofluorescence
Brain samples were collected at the indicated times and assayed using a mouse IL-1β, IL-6, and TNF-α ELISA Kit (Dakewe, Beijing, China).
A TUNEL detection kit (12156792910; Roche, Germany) was used to evaluate neuronal apoptosis. Cells with blue-black nuclei were considered TUNEL-positive. For each group, the number of positive cells in three randomly chosen high-power fields (400×) within the perihematomal area was counted using a light microscope.
Immunoprecipitation was performed as described previously (16). Perihematomal tissues were collected from the mice and solubilized on ice in radioimmunoprecipitation assay lysis buffer. All lysates were then subjected to SDS-PAGE, transferred to nitrocellulose membranes, and blocked with 5% skim milk. All extracted proteins were incubated with specific primary and secondary Abs or with IgG as a control (Millipore). The resulting immunoprecipitates were centrifuged and resuspended in SDS-PAGE buffer. Protein expression was measured by Western blotting. The Abs used in this experiment are listed in Table IV.
View this table:
Table IV.Abs used for Co-IP (Western blot)
At 1, 3, and 7 d after ICH, the mice from each group were sacrificed and their brains removed. Nuclear proteins were extracted from perihematomal brain tissue as described previously (22). Equal amounts of protein were incubated for 15 min at room temperature with an NF-κB–specific [32P]-labeled oligonucleotide and binding mix following the kit instructions (GS056B; Beyotime, Nantong, China; LightShift Chemiluminescent EMSA Kit; Pierce). Competition assays were conducted by adding a 100-fold excess of unlabeled NF-κB oligonucleotide to the reaction mix. For supershift experiments, Abs were preincubated with the sample (from A20−/− mice 3 d after ICH) for 10 min prior to incubation with radiolabeled probe. The images were analyzed on a Bio-Rad image analyzer.
Data are expressed as the mean ± SEM or as percentages. Statistical differences between pairs of groups were assessed using the unpaired t test and between multiple groups using two-way ANOVA. Correlation analysis was used to analyze the relationship between A20 mRNA expression and the mRS scores of the ICH patients. Differences were considered significant at p < 0.05.
A20 expression was increased in the PBMCs of ICH patients and negatively correlated with neurologic deficits
The expression of A20 in PBMCs from 34 ICH patients and 34 sex- and age-matched healthy controls was analyzed by qRT-PCR and immunofluorescence. The baseline level (e.g., the age, male sex, body temperature, history of vascular risk factors, laboratory parameters, obvious inflammatory disease within 6 mo, acute myocardial infarction, and acute or chronic liver damage) was comparable across both groups (Table I). A20 expression in the ICH patients was significantly increased compared with that in the healthy controls (p < 0.01) (Fig. 1A, 1C). Correlation analysis indicated that A20 expression was negatively correlated with patient mRS scores 3 mo after ICH (R2 = 0.6257) (Fig. 1B).
A20 expression was increased in the PBMCs of patients and negatively correlated with neurologic deficits. (A) A20 expression in patient PBMCs was analyzed by qRT-PCR. (B) The relationship between A20 mRNA expression and the mRS scores of ICH patients 3 mo after ICH was analyzed by correlation analysis (R2 = 0.6257, p < 0.01). (C) A20 expression was analyzed by immunofluorescence (horizontal lines indicate median values). Scale bars, 25 μm. **p < 0.01 compared with healthy controls, HC. Scale bars, 25 μm.
A20 expression increased after ICH in mice, especially in microglia and neurons
To further investigate the role of A20, an ICH mouse model was used to analyze A20 expression in perihematomal tissue. qRT-PCR results indicated that compared with the sham group, A20 expression increased at 12 h after ICH, peaked at 1 d and then gradually declined (Fig. 2A). Western blotting indicated that A20 expression increased at 12 h, peaked at 3 d, and then gradually declined. Seven days after ICH, A20 expression in the model was still higher than in the sham group (Fig. 2B). Because A20 expression peaked at 3 d after ICH, we chose this time point to further analyze the A20-expressing cells. The results indicated that A20 expression was more prominent in microglia and neurons than in astrocytes (Fig. 2C).
In mice, A20 expression increased after ICH, especially in neurons and microglia. A20 expression in the perihematomal area was detected by (A) qRT-PCR and (B) Western blotting (**p < 0.01 versus sham, n = 6). (C) A20 expression in perihematomal tissue was detected by immunofluorescence (arrows indicate A20-positive cells; scale bars, 25 μm). The percent of positive A20/NeuN, A20/Microglia and A20/Astrocytes in three randomly chosen fields within the perihematomal area was counted. The data are presented as the mean ± SEM. *p < 0.05 versus A20/Astrocytes, #p < 0.05 versus A20/NeuN.
A20 knockout increases inflammatory injury in perihematomal tissue after ICH
Because A20 expression in the brain was significantly increased 3 d after ICH, inflammatory injury was analyzed in the perihematomal tissue. Brain edema and NDS are important indexes for assessing ICH-injury severity. In this experiment, hematoma volume and hemoglobin content in the A20−/− mice were increased compared with those in the WT mice (Fig. 3A). Compared with WT mice, the water content in the ICH ipsilateral hemisphere of the A20−/− mice was higher at 3 d after ICH (Fig. 3B). Here, we observed NDS deterioration in mice after A20 deficiency compared with the WT group at all time points (Fig. 3C). Furthermore, we assessed changes in the expression of inflammatory factors, including IL-1β, TNF-α, and IL-6, in the perihematoma tissue of mice after ICH using the ELISA. The results showed that IL-β, IL-6, and TNF-α expression in the WT ICH mice was elevated compared with that in the WT sham group, and the expression levels in the A20−/− ICH mice were even higher than those in the WT ICH mice (Fig. 3D). The death rate was also significantly increased after ICH in the A20−/− mice (Fig. 3E). Neuronal apoptosis was also detected by TUNEL staining, and compared with WT mice, the A20−/− mice exhibited an increase in TUNEL-positive cells in the perihematomal area (Fig. 3F). Thus, A20 can protect the brain from inflammatory injury after ICH. These results indicated that A20 deficiency significantly aggravated ICH-induced inflammatory injury in a mouse model of ICH.
A20 knockout increased inflammatory injury in perihematomal tissue after ICH. (A) Representative coronal sections from WT and A20−/− mice (a1). The hematoma volume (a2) and hemoglobin content (a3) were calculated (*p < 0.05, **p < 0.01 versus WT; n = 6 each). (B) Brain water content at sham and 3 d after ICH was compared between the WT and A20−/− group. (*p < 0.05, **p < 0.01 versus WT; n = 6 each). (C) NDS at 1, 3, 5, and 7 d after ICH was compared between the WT and A20−/− groups (**p < 0.01 versus WT; n = 6, each). (D) An ELISA Kit was used to determine IL-1β, IL-6, and TNF-α expression in WT and A20−/− mice (*p < 0.05, **p < 0.01 versus WT; n = 3 each). (E) Overall survival in the A20-deficient group was significantly lower than that in the WT group (**p < 0.01 versus WT; n = 20 each). (F) Representative TUNEL staining in WT and A20−/− mice. Scale bars, 100 μm. The number of TUNEL-positive cells was calculated (**p < 0.01, n = 3 each).
A20 overexpression can reduce post-ICH inflammatory injury in mice
To further evaluate the role of A20 in ICH-induced inflammatory injury in mice and to evaluate the therapeutic effect of A20 on ICH, we developed a mouse model of A20 overexpression. The qRT-PCR and Western blotting showed that A20 expression was higher 7 d after LV-A20 injection compared with the sham and vector groups (Fig. 4A, a1, a2). Moreover, we performed immunofluorescence to examine A20 expression in different cell types in brain 7 d after LV-A20 injection. A20 was mainly expressed in microglia and neurons (Supplemental Fig. 1), which is consistent with the cell types showing A20 expression after ICH (Fig. 2C). Thus, ICH was induced 7 d after the mice were injected with 3 μl of LV-A20 virus. At 3 d after ICH, hematoma volume and hemoglobin content in the A20-overexpressing mice were significantly lower than those in the WT mice (Fig. 4B). The brain water content in the ICH ipsilateral hemisphere of the LV-A20 mice was decreased at 3 d after ICH compared with the WT and LV-GFP groups (Fig. 4C). NDS and IL-β, IL-6, and TNF-α expression were lower in the A20-overexpressing mice than in the WT mice (Fig. 4D, 4E). Unexpectedly, the death rate in the A20-overexpressing group remained unchanged compared with that in the WT group (Fig. 4F). However, TUNEL staining demonstrated that A20 overexpression prevented ICH-induced apoptosis 3 d after ICH. Compared with the WT and LV-GFP groups, the number of TUNEL-positive cells was significantly decreased in the LV-A20 group (Fig. 4G). Thus, intracranial A20 overexpression can reduce inflammatory damage after ICH.
A20 overexpression reduces inflammatory injury after ICH (A) Construction of the A20 overexpression mouse model. A20 expression in ipsilateral brain tissue following LV-A20 or LV-GFP injection was detected by (a1) qRT-PCR, (a2) Western blotting and (a3) immunofluorescence (green-EGFP, red-A20; scale bars, 100 μm). (B) Representative coronal sections from WT, LV-GFP, and LV-A20 mice (b1). The hematoma volume (b2) and hemoglobin content (b3) were calculated (*p < 0.05, **p < 0.01 versus WT.; n = 6 each). (C) Brain water content at sham and 3 d after ICH was compared between the WT, LV-GFP and LV-A20 groups (*p < 0.05, **p < 0.01 versus WT; n = 6 each). (D) NDS at 1, 3, 5, and 7 d after ICH was compared between the WT, LV-GFP and LV-A20 groups (**p < 0.01 versus WT; n = 20 each). (E) An ELISA Kit was used to determine IL-1β, IL-6 and TNF-α expression in WT, LV-GFP, and LV-A20 mice. (*p < 0.05, **p < 0.01 versus WT; n = 3 each). (F) There were no statistically significant differences in post-ICH mortality between the WT, LV-GFP, and LV-A20 groups (p > 0.05 versus WT, n = 20 each). (G) Representative TUNEL staining in WT, LV-GFP, and LV-A20 mice. Scale bars, 100 μm. The number of TUNEL-positive cells was calculated (**p < 0.01 versus WT, n = 3 each).
A20 in peripheral blood may also be involved in regulating ICH-induced inflammation
The above findings raise a critical question of whether A20 in the peripheral system has any impact on the regulation of inflammatory injury after ICH. We therefore used WT-A20−/− parabiosis to test the efficacy of peripheral A20 in alleviating ICH-induced injury in the brains of A20−/− parabionts. After parabiosis between A20−/− and WT mice, the circulation of the parabionts was connected so that A20 in the blood of the parabiotic transgenic mice could be transported into the A20−/− mice. Thus, the A20−/− mice acquired an additional peripheral system from WT mice. Immunofluorescence staining revealed that the same amount of A20-positive cells in A20−/− parabiont PBMCs 3 d after ICH compared with WT mice and WT parabionts. However, A20-positive cells were not detected in the PBMCs of A20−/− mice 3 d after ICH (Fig. 5A). This model provided a reliable approach to test the anti-inflammatory effect of A20 in the peripheral system. Interestingly, the hematoma volume, hemoglobin content, brain water content, inflammation factors expression, and neuronal apoptosis were all obviously decreased in the A20−/− parabiont group compared with the A20−/− group. These results suggest that A20 in peripheral blood is also involved in the regulation of inflammation after ICH (Fig. 5B–E). However, the hematoma volume (Fig. 5B), brain water content (Fig. 5C), and IL-1β and TNF-α levels in perihematomal tissue still were higher 3 d after ICH compared with the WT and WT parabiont groups. Thus, A20 in the peripheral system may be only partially involved in the regulation of inflammatory injury after ICH.
A20 in peripheral blood may partially be involved in the regulation of ICH-induced inflammation. (A) A20 expression in PBMCs of A20−/− parabionts at 3 d after ICH. (B) Representative coronal sections from WT, WT parabiont, A20−/− parabiont and A20−/− mice (b1). The hematoma volume (b2) and hemoglobin content (b3) were calculated 3 d after ICH (*p < 0.05, **p < 0.01 versus WT, #p < 0.05, ##p < 0.01 versus A20−/−, n = 6 each). (C) Brain water content at sham and 3 d after ICH was compared between the WT, WT parabiont, A20−/− parabiont and A20−/− groups. (*p < 0.05, **p < 0.01 versus WT, #p < 0.05 versus A20−/−, n = 6 each). (D) An ELISA Kit was used to determine IL-1β, IL-6 and TNF-α expression at sham and 3 d after ICH in perihematomal tissue of the WT, WT parabiont, A20−/− parabiont and A20−/− groups. (*p < 0.05, **p < 0.01 versus WT, #p < 0.05 versus A20−/−; n = 3 each). (E) Representative TUNEL staining in the indicated mice. Scale bars, 100 μm. The number of TUNEL-positive cells was calculated 3 d after ICH (#p < 0.05, **p < 0.01, n = 3 each).
A20 suppressed ICH-induced inflammatory injury by regulating the polyubiquitination of the E3 ubiquitin ligase TRAF6
TRAF6 is an important transduction molecule downstream of TLR, IL-1βR, and other classical inflammatory pathways whose polyubiquitination can induce NF-κB activity and the subsequent secretion of inflammatory factors. These results suggest that TRAF6 may play an important role in regulating ICH-induced inflammation. Using Western blotting and ELISA experiments, we found that TRAF6 expression was obviously increased after ICH, and the expression trend was consistent with NF-κB p65 expression (Fig. 6). Inhibiting the expression of TRAF6 using TRAF6-targeted shRNA relieved the cerebral edema and neurologic deficits, clearly downregulated NF-κB expression, and inhibited the subsequent secretion of inflammatory cytokines (Supplemental Fig. 2). These results showed that inhibiting the expression of TRAF6 could inhibit inflammatory injury after ICH.
A20 suppressed ICH-induced inflammatory injury by regulating the polyubiquitination of the E3 ubiquitin ligase TRAF6. (A) Interactions between TRAF6 and A20, Ubc13, UbcH5c, Itch, RNF11, and TAX1BP1 in the WT, A20−/−, LV-A20 and LV-GFP groups at 1, 3, and 7 d after ICH. Proteins from lysates were immunoprecipitated with a TRAF6 Ab and detected by immunoblotting with Abs for A20, Ubc13, UbcH5c, Itch, RNF11, TAX1BP1, and TRAF6. (B) Lysates were subjected to immunoblotting with Abs to A20, Ubc13, UbcH5c, Itch, RNF11, TAX1BP1, IκBα, NF-κB p65 and GAPDH. (C) The specificity of the A20, Ubc13, UbcH5c, TAX1BP1, and TRAF6 interactions in an A20+/+ mouse for the indicated times. Proteins from lysates were immunoprecipitated with TRAF6 or a control rabbit Ab and detected by immunoblotting with Abs for A20, Ubc13, UbcH5c, Itch, RNF11, TAX1BP1, and TRAF6.
A20 can inhibit NF-κB activation by suppressing its polyubiquitination and hence the activation of the E3 ubiquitin ligase TRAF6 (14). However, whether A20 inhibited TRAF6 polyubiquitination via a similar mechanism in ICH-induced inflammation remained unclear. To investigate the mechanism of TRAF6 regulation by A20, we first examined TRAF6 protein-protein interactions in ICH-induced mice using coimmunoprecipitation (Co-IP). TRAF6-Ubc13 and TRAF6-UbcH5c (Ubc13 and UbcH5c are E2 ubiquitin-conjugating enzymes that are regarded as key mediators of ubiquitin chain assembly) interactions were increased in A20−/− mice compared with WT mice. In contrast, these interactions were reduced in the A20 overexpression group compared with the WT group. We also evaluated A20-binding proteins, which act cooperatively to suppress NF-κB activation, including the E3 ubiquitin ligase Itch, as well as ring finger protein 11 (RNF11) and Tax1-binding protein (TAX1BP1). In A20−/− mice, TRAF6 did not interact with Itch, RNF11, or TAX1BP1. However, these interactions were increased in the A20-overexpressing mice (Fig. 6A). In A20−/− mice, Ubc13 and UbcH5c expression was increased, whereas RNF11 and TAX1BP1 expression was decreased. In the LV-A20 group, Ubc13 and UbcH5c expression was decreased, whereas RNF11 and TAX1BP1 expression was increased. No change in Itch expression was observed in the A20−/− or LV-A20 groups. Moreover, we detected changes in the level of ubiquitinated TRAF6 at 3 d after ICH in the A20-positive, negative, and overexpression groups using Co-IP experiments. The results showed that inhibiting A20 promoted TRAF6 ubiquitination, whereas overexpressing A20 downregulated TRAF6 ubiquitination levels (Supplemental Fig. 3). Additionally, we examined the expression of IκBα and NF-κB p65 and found that, as expected, IκBα and NF-κB p65 degradation was increased in A20−/− mice, and an increase in NF-κB activation was also detected in A20−/− group; however, the IκBα degradation, NF-κB p65, and NF-κB activity were reversed in the LV-A20 group (Fig. 6B, Supplemental Fig. 4). No binding was observed when immunoprecipitations were performed with a control rabbit Ig Ab (Fig. 6C). Our results illustrate that the inhibition of NF-κB signaling by A20 plays a key role in ICH by disrupting the E2 and E3 ubiquitin enzyme complexes.
Inflammation plays an important role in ICH-induced secondary injury, and the current study showed that the endogenous anti-inflammatory factor A20 can protect the brain from ICH-induced inflammatory injury. Briefly, A20 expression was upregulated in the PBMCs of ICH patients and was correlated with the prognosis of ICH, as indicated by mRS score. Furthermore, A20−/− and A20-overexpressing mice were created, and the results from these mice also indicated that A20 can accelerate hematoma absorption, inhibit the release of proinflammatory factors, reduce neuron degeneration, and decrease neurologic deficits. Meanwhile, WT-A20−/− parabiosis experiments indicated that A20 in PBMCs may be partially involved in the regulation of ICH-induced inflammatory injury. We also found that the anti-inflammatory effect of A20 in ICH may be mediated through the disruption of ubiquitin enzyme complexes. To our knowledge, this study provides the first demonstration that A20 has a protective effect in ICH and offers new insights for the development of ICH therapies.
A20 in PBMCs plays a key role in inflammatory disease, but A20 expression in PBMCs differs across diseases. Increased A20 expression has been demonstrated in the PBMCs of patients with acute lymphoblastic leukemia and acute-on-chronic hepatitis B liver failure (ACHBLF) (23, 24), whereas decreased A20 levels have been observed in the PBMCs of patients with chronic inflammatory diseases such as type-2 diabetes and SLE (12, 25). However, the protective role of A20 is widely accepted, and its diverse expression in different diseases does not seem to be relevant to this role. In accordance with the results of patients with acute lymphoblastic leukemia and ACHBLF, A20 expression in PBMCs increased within 24 h after ICH. The reasons for the divergence in A20 expression remain unclear, although different stages of inflammation might cause this discrepancy. In the acute stage of ACHBLF and ICH, increased A20 expression could inhibit inflammation, but in type-2 diabetes and SLE, due to the state of continuous inflammation, the consumption of A20 by sustaining and redundant cytokines surpasses its production, resulting in a decrease in A20 expression. The specific mechanism requires further study. In this study, we established a parabiosis model to investigate the role of peripheral A20 in neuroinflammation after ICH based on a previously reported method. We observed the same amount of A20-positive cells in the PBMCs of WT-A20−/− parabionts as in the WT and WT parabiont groups. However, A20-positive cells were not detected among the PBMCs in the A20−/− group. Moreover, we found that brain cytokines, neuronal necrosis, etc., were improved in WT-A20−/− parabionts than in A20−/− mice (Fig. 5). The number of A20−/− hematopoietic cells and A20-expressing hematopoietic cells were identical in both WT-A20−/− parabionts and WT-WT parabionts at 3 d after ICH, and the baseline inflammatory levels were also similar in the brain tissue in both groups; therefore, although the A20−/− hematopoietic cells may secrete inflammatory mediators to overcome the noninflammatory influence of the A20-expressing WT hematopoietic cells, the identical effects and baseline levels in both mice are unlikely to influence our conclusion regarding the role of peripheral A20 in reducing brain injury after ICH. Based on these findings, we postulate that peripheral systematic A20 may be partially involved in the regulation of inflammatory injury after ICH. In this experiment, we did not separate the parabionts after 20 d of parabiosis and then continue to construct the ICH model, and so we could not calculate the mortality and NDS of the mice after ICH; hence, we did not provide these data in this part of our manuscript.
The results of the current study indicated that A20 is primarily expressed in neurons in the normal human brain and that more A20 was expressed in the pons, striatum, hippocampus, and medulla compared with the frontal cortex. A20 mRNA was also detected in the cortex and hippocampus of WT mice (26, 27). In experimental autoimmune encephalomyelitis models and LPS-induced inflammatory models, A20 is expressed in microglia and astrocytes and suppresses inflammation (28, 29). However, our results indicated that after ICH, A20 is mainly expressed in microglia and neurons and seldom in astrocytes, consistent with the cell types that were induced to express A20 after LV-A20 injection. In the current study, we found that the NDS was significantly decreased in the A20 overexpression group compared with the WT group after ICH (Fig. 4D). In addition, although the local overexpression of A20 reduced brain edema and improved the NDS and neuron necrosis after ICH, the local overexpression of A20 may not be sufficient to reduce the mortality after ICH.
Inflammation plays an important role in ICH-induced secondary injury. We previously showed that inflammatory injury can be induced by the activation of TLR2/TLR4 dimers, which subsequently activated NF-κB through MyD88 signaling (30, 31). Recent studies have indicated that A20 can disturb ubiquitin enzyme complexes that are important for inhibiting NF-κB activation in TLR4 signaling. These studies indicated that the protective function of A20 may be mediated by the deubiquitination of A20 (14). In this report, the authors performed in vitro experiments to confirm the molecular mechanisms of A20, and they used LPS to stimulate mouse embryonic fibroblasts, which had a very quick response; therefore, stimulation was only performed for several hours. However, in our study, we performed in vivo experiments to detect the molecular mechanisms of A20. Our and others’ previous work also showed that the neuroinflammation was aseptic inflammation after ICH, and the neuroinflammation changed over several hours to several days, mostly reaching the highest levels at 3 d after ICH; therefore, in the current study, samples were analyzed several days after ICH (16, 32). Our results showed that after ICH, TRAF6-Ubc13 and TRAF6-UbcH5c interactions, as well as the level of ubiquitinated TRAF6, were enhanced in A20-deficient mice compared with WT mice; an increase in these interactions accelerates the degradation of IκBα, which then increases the activation of NF-κB. However, the interactions between TRAF6 and Ubc13 and between TRAF6 and UbcH5c, as well as the level of ubiquitinated TRAF6, were attenuated in the A20 overexpression group compared with the WT mice; a decrease in these interactions leads to the decreased degradation of IκBα and the activation of NF-κB. As expected, A20 complex components, including RNF11 and Itch, as well as TAX1BP1, were also differentially expressed in the indicated groups. Our results indicate that A20 alleviated ICH-induced inflammation by inhibiting TRAF6 polyubiquitination.
In our study, we compared the inflammatory factors in the brain tissue of the WT, WT parabiont, A20−/− parabiont and A20−/− sham groups. The results showed no significant difference in IL-β expression between the A20−/− parabiont and WT, WT parabiont, and sham groups. However, IL-6 and TNF-α levels increased in the A20−/− sham group (Fig. 5D), suggesting that A20 deficiency caused spontaneous inflammation in the mouse brain, consistent with previous research (27). In the current study, because of the serious infection and high mortality, we did not perform A20−/−–A20−/− parabiosis.
A20 can inhibit the activation of NF-κB, but the precise pathways by which A20 induces NF-κB deactivation remain under investigation. TLR/IL-1R, IL-17R, and TCR signaling may participate in this process (33–37). Our experiments mainly focused on TRAF6, which is downstream of inflammatory pathways, but the signaling that is initiated and mediated by TLRs/IL-1R, IL-17R, and the TCR requires further study. Thus, A20 can ameliorate inflammatory injury after ICH by disturbing ubiquitin enzyme complexes.
In conclusion, our findings reveal that increased expression of A20 in PBMCs negatively correlates with neurologic deficits in ICH patients and that A20 reduces ICH-induced inflammation by inhibiting TRAF6 polyubiquitination. Moreover, A20 overexpression can reduce post-ICH inflammatory injury. These results support the possibility that A20 may play a pivotal role in ICH-induced inflammation and offer a potential therapeutic tool in the treatment of ICH.