Celastrol ameliorates experimental autoimmune neuritis by shifting the polarization of M1/M2 macrophages via the hypoxic response pathway

Background GBS is an autoimmune disease characterized by inammatory inltration and demyelination of peripheral nerves. Macrophage polarization is involved in different stages of GBS. Altering the polarization of macrophages may be an effective therapeutic strategy for GBS. Celastrol was previously shown to contribute to anti-neuroinammation. However, the mechanism underlying the effect of celastrol in GBS animal model experimental autoimmune neuritis (EAN) is unclear. We hypothesized that celastrol may shift the polarization of macrophages through the NRF /HIF-1αpathway. Methods

by accumulation of T cells and macrophages, breakdown of the blood-nerve barrier (BNB), and demyelination of peripheral nerves (2,3). As the main in ltrating cells in the peripheral nervous system (PNS) during AIDP progression, macrophages play either a proin ammatory or an anti-in ammatory role at the different stages of AIDP (3,5). Importantly, a shift in macrophage polarization from the M1 to M2 phenotype may ameliorate the severity of EAN and, thus, is suggestive of a promising therapeutic strategy for treating GBS (5)(6)(7).
In the past decade, some traditional Chinese medicines have been found to be effective for treating GBS (8,9). These studies have highlighted the unique clinical e cacies of traditional medicine. However, the underlying cellular and molecular mechanisms of these medicines remain to be investigated. Celastrol is an active ingredient of the traditional Chinese medicine, Tripterygium wilfordii (Thunder god vine), which has long been used to treat in ammatory diseases. Together with artemisinin, a previous study reported the potential of celastrol as a clinical drug for the treatment of obesity, with other possible uses as well (10). Furthermore, experimental ndings have indicated that celastrol can reduce in ammation by suppressing M1 polarization (11). During the polarization of macrophages, hypoxic adaptive responses are actively regulated (5,6,11). As a transcription factor with a high antioxidant capacity, nuclear factor-E2-related factor 2 (NRF2) exerts protective effects against oxidative and in ammatory stress by regulating antioxidant response elements (AREs) and heme oxygenase 1 (HO-1) (12,13). Furthermore, hypoxia-inducible factor (HIF-1α), known as one of the most important mediators in cellular responses to hypoxia and in ammation, has been reported to contribute to shifting the polarization of macrophages (14,15). However, the role of the NRF2 signaling pathway in regulating HIF-1α remains unclear. Herein, we hypothesized that celastrol may shift the polarization of macrophages via the hypoxia response pathway.
In the present study, we investigated the effects and mechanisms of celastrol in an EAN model of GBS. Our translational ndings demonstrate that celastrol may be effective in ameliorating the clinical course of GBS by decreasing in ltration of in ammatory cells. Furthermore, our elucidated therapeutic effect of celastrol may be related to promoting macrophage polarization toward the anti-in ammatory phenotype via the hypoxia response pathway.

Animals
Male Lewis rats (6-8 weeks old, 160-180 g, Vital River, Beijing, China) were acclimated to the environment within our institution's vivarium for one week prior to the start of any experiments. All rats were housed under equal daily periods of light and darkness (i.e., 12/12-h light/dark cycle) and were provided access to food and water ad libitum. The rats were randomly assigned to the following three groups: the control, EAN, and therapeutic (i.e., celastrol + EAN) groups (n = 8 per group). All efforts were made to minimize the numbers of animals used and their suffering. All animal experimental protocols were reviewed and approved by the Animal Ethics Committee of Zhejiang University of Traditional Chinese Medicine.

Induction Of Ean And Celastrol Treatments
The inoculum preparation was as follows. The P2 peptides 53-78 (TGSPPLATGISPLLGGGPGGTTAAAA, GL, Biochem Ltd. Shanghai, China) were dissolved in phosphate-buffered saline (PBS; 2 mg/ml) and were then emulsi ed with an equal volume of complete Freund's adjuvant (CFA; Difco) containing Mycobacterium tuberculosis (strain H37RA). The nal concentration of peptides in the inoculum was 1 mg/ml. EAN was induced by immunization via subcutaneous injection of 0.3 ml of inoculum to the base of the tail. For celastrol treatments, rats were induced by intragastric administration with celastrol at a dose of 1 mg/kg (16). Celastrol was prepared as a 0.2-mg/ml solution in 1% dimethyl sulfoxide (DMSO).
Rats in both the control and EAN groups were given an equal volume of vehicle (i.e., 1% DMSO).

Histopathological Assessments
To evaluate the in ltration of in ammatory cells and demyelination in the PNS, hematoxylin-eosin (HE) and Luxol fast blue (LFB) were applied to peripheral tissues. Speci cally, sciatic nerves were harvested at the peak of disease (day-16 post-immunization [p.i.]) and immediately xed in 4% paraformaldehyde overnight at 4 °C. After dehydration and vitri cation, the harvested sciatic nerves were embedded in para n and sliced into 4-um-thick sections. The sciatic-nerve sections were then stained with HE and LFB. The in ltration of in ammatory cells was counted at a 200-x magni cation from ve elds randomly from each slide. The averaged results are expressed as the number of in ammatory cells per square millimeter. Histological scores were applied to evaluate the severity of demyelination according to a semiquantitative grading system, as follows [20]: 0 = normal; 1 = less than 25% demyelinated bers; 2 = 25-50% demyelinated bers; 3 = 50-75% demyelinated bers; and 4 = more than 75% demyelinated bers.
Immuno uorescence was performed according to the protocols provided by the manufacturers from which the employed antibodies were purchased. First, 4-um-thick sections were permeabilized in 0.3% Triton X100. After blocking the tissues, sections were incubated with the following primary antibodies at phenylindole (DAPI) to label cellular nuclei. Images were acquired using an 20 × objective with an Olympus microscope. Image-Pro Plus was applied to quantify the uorescent intensity of each image.

Western Blotting
Lysis buffer was used to extract protein from sciatic-nerve tissues. The bicinchoninic acid (BCA) protein method was used to measure protein concentrations. Samples (20 ug each) were loaded on 10% sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) gels and were then electrophoretically separated. Proteins were then transferred to polyvinylidene di uoride (PVDF) membranes. After blocking, membranes were then incubated overnight at 4 °C with primary antibodies against the following: NRF2 (1:1000, Abcam), HIF-1α (1:1000, Abcam), and GAPDH (1:5000, Multi Science). The next day, the membranes were washed with Tris-buffered saline with Tween (TBST) and were subsequently incubated with secondary antibodies-goat anti-mouse IgG (1:5000, Multi Science) and goat anti-rabbit IgG (1:5000, Multi Science)-for 2 h at room temperature. After washing the membranes, chemiluminescence reactions were carried out and immunoreactivity levels were detected by a gel imaging analyzer (Bio-Rad).

Statistical analysis
All data are presented as the mean ± standard deviation (SD). GraphPad Prism 8 was used to analyze the data. Mann-Whitney U tests were used to compare differences of neurologic scores and histological scores among the experimental groups. Student's t-tests and one-way or two-way analyses of variance (ANOVAs) were applied for paired and multiple group comparisons. Statistical signi cance was set at a level of p < 0.05.

Celastrol ameliorates the severity of EAN
To investigate the effect of celastrol in EAN, rats were treated with celastrol (1 mg/kg) daily from the day of EAN immunization. Neurological scores (Fig. 1A) were recorded every day and weight changes were recorded every three days (Fig. 1B). The EAN group of rats started to exhibit neurological symptoms on day-6 p.i., and the symptoms progressed rapidly until they peaked on day-16 p.i.. (mean pathological score, 6.80 ± 0.45) (Fig. 1C). Compared to parameters in the EAN group, the celastrol treatment group exhibited a delayed onset of neuritis, decreased mean peak neurological scores, and lower neurological scores from days 8 to 16 p.i. (p < 0.05 for each time point) (Fig. 1A). Moreover, compared to those of the control group, the mean body-weight trends of rats were decreased in the EAN group while this decrease occurred more slowly in the EAN + celastrol group relative to that of the EAN group (Fig. 1B, D). These results demonstrated that celastrol treatments not only delayed the onset of EAN but also ameliorated the severity of EAN symptoms.

Celastrol attenuates the histopathology and in ltration of in ammatory cells in EAN
We next investigated the histopathology of the sciatic nerve at day-16 p.i. from each rat. Compared to that of the EAN group (518.08 ± 61.26 cells), we found that the number of in ltrating in ammatory cells was signi cantly reduced in the celastrol treatment group, as revealed by HE staining (282.86 ± 92.45 cells) (p < 0.05) ( Fig. 2A, C). Furthermore, the incidence of demyelination was decreased in the celastrol group compared to that in the EAN group (Fig. 2B). The mean histological scores, which were used to semi-quantitatively evaluate the severity of demyelination, were markedly lower in the celastrol treatment group (1.67 ± 0.67) compared to those in the EAN group (2.89 ± 0.74, p < 0.05) in LFB staining (Fig. 2B, D).
Celastrol promotes polarization of macrophages to the M2 phenotype in EAN We further investigated whether the therapeutic e cacy of celastrol correlated with polarization of macrophages within sciatic nerves and spleens of EAN rats. Immuno uorescent staining was used to evaluate the phenotypes of macrophages at day-16 p.i. in sciatic nerves. Compared to those in the control group, we found that the levels of in ltrating macrophages were elevated (Fig. 3A) in the EAN group, while the levels of both M1 and M2 macrophages were reduced (Fig. 3A) in the celastrol treatment group.
Moreover, compared to that in EAN rats, the ratio of M2/M1 macrophages was signi cantly elevated in celastrol-treated rats (Fig. 3A). In addition, we also used ow cytometry to detect the polarization state of macrophages in spleens. The ow cytometry results showed a similar macrophage polarization (Fig. 3B). Compared to those in the EAN group, both results showed an increase in the percentages of M2 macrophages and a decrease in the percentages of M1 macrophages in the celastrol treatment group (Fig. 3B, C), indicating that celastrol promoted polarization of macrophages into the M2 phenotype in EAN.

Celastrol reduces in ammatory cytokines while boosting anti-in ammatory cytokines in EAN
We also examined whether the bene ts of celastrol treatment were related to the expression pro les of in ammation cytokines. TNF-α, IL-6, IL-4, and IL-10 levels were evaluated in the sera and spleens collected on day-16 p.i. from control, EAN, and celastrol treatment rats. Compared to those in the control group, the levels of IL-6 and TNF-α in the EAN group were elevated (Fig. 4A, B), while the levels of IL-4 and IL-10 were reduced (Fig. 4C, D). By contrast, celastrol treatment signi cantly reduced IL-6 and TNF-α proin ammatory cytokine levels, whereas IL-4, and IL-10 anti-in ammatory cytokine levels were elevated on day-16 p.i. (Fig. 4).

Celastrol Regulates The Nrf2/hif-1α Hypoxic Pathway In Ean
To further elucidate whether the NRF2/HIF-1α pathway is related to the therapeutic effects of celastrol treatment, we analyzed NRF2 and HIF-1α expression levels in sciatic-nerve tissue via immuno uorescence and Western blotting. When compared to these levels in the control group, a decreased level of NRF2 and an increased level of HIF-1α were observed (Fig. 5A) in the EAN group, whereas the celastrol treatment group showed an increased level of NRF2 and a decreased level of HIF-1α on day-16 p.i. compared to those in the EAN group (Fig. 5A). Consistently, the results of Western blotting also showed an increased level of NRF2 and a decreased level of HIF-1α compared to those in the EAN group (Fig. 5B, C).

Discussion
In the present study, we assessed the effects of celastrol treatment on EAN, which is a widely applied animal model of GBS. We found that celastrol not only delayed the onset of EAN but also attenuated its peak severity. Histopathologically, we revealed that celastrol attenuated EAN-induced in ltration of in ammatory cells and demyelination of sciatic nerves (Fig. 6).
Celastrol, as an effective ingredient of Tripterygium wilfordii (Thunder god vine), has a long history in the treatment of in ammatory and autoimmune disease. Pharmacologically, celastrol exhibits anti-cancer, anti-in ammatory, and antioxidant properties (17)(18)(19). In in ammatory disease, celastrol had been found to ameliorate myelin oligodendrocyte glycoprotein (MOG)-induced EAE development by reducing Th17 responses in both the periphery and central nervous system (CNS) (20). Celastrol also switches T-cell responses from a predominantly Th1 to Th2 type (16). Moreover, celastrol may also reduce NF-κB expression and T-cell accumulation in the CNS, indicating its anti-in ammatory properties (20,21). Based on these previous ndings, we hypothesized that celastrol might have the potential to treat EAN. To our knowledge, the present study represents the rst to investigate the effect of celastrol on EAN. Consistent with previous reports, we found that celastrol attenuated in ammatory reactions and demyelination in the PNS and improved EAN outcomes by reducing in ltration of in ammatory cells.
GBS comprises a spectrum of autoimmune disorders that induce peripheral neuropathies, during which the in ltration of in ammatory cells into the PNS occurs at different stages of the disease(2). As the major in ammatory cells in EAN, macrophages are generally divided into two phenotypes: classically activated (M1) and alternatively activated (M2) macrophages (22). Previous studies suggest that macrophage phenotypes have high plasticity and can be altered by appropriate signals at different stages of GBS (5,23,24). M1 macrophages play a pro-in ammatory role in tissue damage and disease progression, while M2 macrophages exert anti-in ammatory effects and promote disease recovery(3). In the present study, we observed that celastrol yielded a favorable outcome in EAN that was associated with a phenotypic switch toward M2 macrophages within sciatic nerves. Furthermore, splenic macrophages also showed a switch toward M2 macrophages during celastrol treatment in the face of EAN, indicating that celastrol may attenuate EAN through promoting polarization of macrophages into the M2 phenotype within multiple tissues.
In ammatory cytokines secreted by macrophages mediate immune responses. M1 macrophages are involved in the progression of EAN by releasing pro-in ammatory cytokines such as TNF-α and IL-6 (25,26). Previous studies have shown that TNF-α levels are elevated in EAN/GBS, leading to the disruption of the BNB and the demyelination of peripheral nerves (27,28). Furthermore, IL-6 has been shown to contribute to the demyelination and progression of EAN, and increased IL-6 levels have been observed in EAN and in the sera and cerebrospinal uid of GBS patients (26). In contrast, M2 macrophages exert a protective role by releasing anti-in ammatory cytokines, including IL-4 and IL-10(29-32). Importantly, anti-in ammatory cytokines have been previously found to be involved in the repair of peripheral nerves.
In accordance with previous studies, we observed increased levels of TNF-α and IL-6 and decreased levels of IL-4 and IL-10 in both the spleen and peripheral venous blood in the EAN group. Although these particular protein levels may not fully represent all cytokine levels in peripheral nerves, they nonetheless provide molecular clues as to the underling physiological mechanisms of celastrol, which are suggestive of a shift in the polarization of macrophages to the M2 phenotype to ameliorate EAN.
Hypoxia is a physiological state of the micro-environment that is involved in the development of many diseases. NRF2 is a transcription factor that plays a pivotal role in hypoxia (13,33). Activation of NRF2 exerts antioxidant, anti-in ammatory, and neuroprotective properties (34,35). Upregulation of NRF2 induces HO-1 gene transcription, which is implicated in ameliorating the severity of EAN rats (6). HIF-1α is another key transcription factor that responds to hypoxia. In obesity, HIF-1α activation is triggered by increased oxygen consumption, thus causing in ammation (10). Inhibition of HIF-1α may decrease in ammation in tumor genesis (36). Furthermore, growing evidence has suggested that NRF2 and HIF-1α can directly or indirectly regulate each other (37,38). A recent study found that NRF2 may upregulate HIF-1α via activation of thioredoxin in stem cells (39). HIF-1α can elevate NRF2 by inhibiting thioredoxin reductase (40). However, these two factors do not always reinforce each other. During andrographolide treatment, a previous study showed that the expression level of NRF2 was increased while the HIF-1α level decreased (41). Interestingly, we observed that celastrol signi cantly increased the expression of NRF2 and decreased the expression of HIF-1α, indicating that celastrol may exert anti-in ammation and neuroprotective effects by modulating hypoxia.

Conclusion
In conclusion, the present study demonstrates that celastrol may ameliorate EAN progression through promoting polarization of macrophages into the M2 phenotype, possibly by modulating hypoxic responses. Hence, our ndings suggest that celastrol may represent a potential novel therapy for GBS.      Celastrol reduces in ammatory cytokines while boosting anti-in ammatory cytokines in EAN. Cytokine levels were evaluated in sera and spleens by ELISAs at day 16 (p.i.). The levels of pro-in ammatory cytokines (e.g., TNF-α, IL-6) were decreased signi cantly in the celastrol group compared to those in the EAN group (P<0.05). The levels of anti-in ammatory cytokines (e.g., IL-4, IL-10) were increased signi cantly in the celastrol group compared to those in the EAN group (P<0.05).  Effect of celastrol on EAN is mediated by altering the polarization of macrophages and the in ammatory cytokines. Macrophages play a pivotal role in neuroin ammation on EAN. Celastrol ameliorates the severity of EAN by altering the polarization of macrophages into M2. Moreover, celastrol reduces in ammatory cytokines while boosting anti-in ammatory cytokines. The mechanism underlying celastrol on macrophage polarization may be mediated by the NRF2/ HIF-1α pathway.