1,25-Dihydroxyvitamin D Inhibits LPS-Induced High-Mobility Group Box 1 (HMGB1) Secretion via Targeting the NF-E2-Related Factor 2–Hemeoxygenase-1–HMGB1 Pathway in Macrophages

1,25-Dihydroxyvitamin D [1,25(OH)2D3] is recognized as a key mediator of inflammatory diseases, including sepsis. Clinical studies demonstrate that 1,25 (OH)2D3 protects patients from sepsis, but clinical treatment with 1,25(OH)2D3 is rare. In this study, we report that 1,25(OH)2D3 treatment has beneficial effects and improves the survival rate in LPS-induced mouse sepsis model by blocking the secretion of high-mobility group box 1 (HMGB1), a key late regulator of sepsis. LPS-induced HMGB1 secretion is attenuated by 1,25(OH)2D3 via blocking HMGB1 translocation from the nucleus to the cytoplasm in macrophages. 1,25(OH)2D3 can induce the expression of hemeoxygenase-1 (HO-1), which is essential for blocking HMBG1 nuclear translocation and its secretion. When siHO-1 or an HO-1 inhibitor are used, the effect of 1,25(OH)2D3 on inhibition of HMGB1 secretion is suppressed. Considering that HO-1 is a downstream gene of NF-E2-related factor 2 (Nrf2), we further confirm that Nrf2 activation can be activated by 1,25(OH)2D3 upon LPS exposure. Together, we provide evidence that 1,25(OH)2D3 attenuates LPS-induced HMGB1 secretion via the Nrf2/HO-1 pathway in macrophages.

Highmobility group box 1 is a DNAbinding nuclear protein that is selected actively following cytokine stimulation and passively released during cell death (3,4). It is the prototypic damageassociated molecular pattern molecule and has been implicated in several inflammatory disorders (2)(3)(4)(5). HMGB1 is released by activated monocytes and macrophages (6). Studies using neutralizing antibodies for HMGB1 have verified that increased circulating levels of HMGB1 contribute to the late lethality of endotoxemia and sepsis (7,8). HMGB1 antibodies inhibit endotoxin lethality in mice (9) and inhibit lung inflamma tion following airway LPS exposure (10). These findings suggest that HMGB1 may serve as a target to reduce mortality from sepsis and the mechanisms responsible for inducing and controlling HMGB1 release becomes significant.
Highmobility group box 1 contains two nuclear localization signals and two putative nuclear export signals, indicating that HMGB1 shuttles between the cytoplasm and nucleus through a tightly controlled mechanism (11). Hemeoxygenase1 (HO1) has been reported to suppress the translocation and secretion of HMBG1 (12,13). Decreased HMGB1 expression through increased HO1 production takes a protective role in several disease states, including arthritis and sepsis (14)(15)(16). Various regulatory elements have been identified in the promoter region of HO1, such as AP1 and NFκb, but NFE2related factor 2 (Nrf2) is of particular importance. Nrf2 is a redoxsensitive master switch to induce HO1 activation, which modulates its gene expression by binding to a recognition site in the inducible enhancers (E1) of HO1 and heterodimerizing with activating transcription factor 4 to exert its effect (17). The Nrf2/HO1 path way plays an important role in the HMGB1 secretion (18)(19)(20).
Vitamin D is a fatsoluble vitamin primarily synthesized from 7dehydrocholesterol in the skin by ultraviolet radiation. 1,25dihydroxyvitamin D [1,25(OH)2D3] is its active form. Many studies have proved vitamin D deficiency is closely related with clinical outcomes such as mortality of sepsis, duration of mechanical ventilation, and length of stay (21). A number of observational studies show a negative association between low vitamin D levels and risk of sepsis (22)(23)(24). However, the mecha nism of antiinflammatory action of Vitamin D remains poorly understood.
In this study, we provide evidence that 1,25(OH)2D3 attenuates LPSinduced HMGB1 nuclear export and secretion in macrophage by Nrf2/HO1 pathway. Upon LPS exposure, 1,25 dihydroxyvitamin D activates Nrf2 nuclear translocation and induces HO1 expression, resulting in inhibition of HMGB1 secretion. resUlTs 1,25(Oh) 2 D 3 inhibits lPs-induced hMgB1 secretion in Macrophages 1,25Dihydroxyvitamin D plays a key role in sepsis (22,25), and HMGB1 is a late mediator of endotoxin lethality released from macrophages, so we examined the effects of 1,25(OH)2D3 on HMGB1 secretion in macrophages. Data showed that LPS induced HMGB1 secretion was suppressed by 1,25(OH)2D3 time course ( Figure 1A) and various concentrations ( Figure 1B) stimuli in bone marrowderived macrophages (BMDMs). HMGB1 release was markedly attenuated by 1,25(OH)2D3 in a dosedependent manner ( Figure 1B). Here, we note that total HMGB1 expression is consistent, meaning that only the secre tion of HMGB1 is suppressed by1,25(OH)2D3. 1,25(OH)2D3 also inhibited LPSinduced HMGB1 secretion in RAW264.7 cells, a murine macrophage cell line ( Figure S1 in Supplementary Material). HMGB1 mRNA was further assayed by qPCR and showed no change in BMDMs between LPS stimulation alone and treatment of LPS plus 20 nM 1,25(OH)2D3 (Figure 1C). At the same time, TNFα as a typical inflammatory factor was also detected and not affected by 1,25(OH)2D3 ( Figure 1D). These observations suggest that 1,25(OH)2D3 modulates HMGB1 release independent on regulation of gene expression.

1,25(Oh) 2 D 3 Blocks lPs-induced hMgB1 nuclear export in Macrophages
The translocation of HMGB1 from the nucleus to the cytoplasm can be induced by LPS, which is crucial for its release (6). To address the mechanism of 1,25(OH)2D3 regulation on HMGB1 secretion, we analyzed LPSinduced HMGB1 translocation. LPSinduced HMGB1 translocation from the nucleus to the cytoplasm was observed in Figure 2A, and its nuclear export was blocked with 1,25(OH)2D3 stimuli. Then, the distribution of HMGB1 was reconfirmed by subcellular fractionation. As shown in Figure 2B, a shift of HMGB1 from the nucleus to the cyto plasm was observed in BMDMs induced by LPS. 1,25(OH)2D3 promoted the amount of HMGB1 in nuclear in a dosedependent manner, indicating 1,25(OH)2D3 blocks the HMGB1 nuclear export in macrophages.
hO-1 is required for the inhibition of 1,25(Oh) 2 D 3 on lPs-induced hMgB1 secretion Hemeoxygenase1 has been reported to control HMGB1 nuclear translocation and block HMGB1 secretion (12). The HO1 inhibitor ZnPPIX was used to examine whether HO1 is involved in the blocking of HMGB1 secretion by 1,25(OH)2D3. We found that ZnPPIX indeed rescued the inhibition of HMGB1 secretion by 1,25(OH)2D3. Consistent with this result, the sup pression of 1,25(OH)2D3 on LPSinduced HMGB1 secretion was also recovered by siRNA HO1 (Figure 3B). Together, the above data indicated that HO1 is important for the HMGB1 secretion inhibited by1,25(OH)2D3.  HO1 expression also can be directly induced by 1,25(OH)2D3 alone ( Figure 4E).

1,25(Oh) 2 D 3 enhances nrf2 activation to Promote hO-1 Transcription
Because HO1 is a proved target of Nrf2, we expected that 1,25(OH)2D3 might have an effect on Nrf2 expression. Indeed, there is no induction of Nrf2 on stimulation with LPS or LPS plus 1,25(OH)2D3 together in BMDMs (Figures 5A,B). Considering the main form of Nrf2 activation is nuclear translocation, we examined the subcellular localization of Nrf2 by immunofluo rescent assay. 1,25(OH)2D3 or LPS could induce the Nrf2 nuclear translocation individually, but costimulation with 1,25(OH)2D3 markedly enhanced the LPSinduced Nrf2 nuclear translocation ( Figure 5C). ARE1 and ARE2 are the Nrf2 cisDNA elements identified in the mouse HO1 gene promoter (26). Chromatin immunoprecipitation (ChIP) assays showed that Nrf2 binding to these sites was increased by the 1,25(OH)2D3 stimuli in dose dependent manner ( Figure 5D). Together, these data demon strate that 1,25(OH)2D3 increases Nrf2 nuclear translocation and promotes the transcript of HO1 expression.  Figure 6A). Consistent with our finding, within 24 h of    LPS treatment, the serum HMGB1 was significantly decreased in the drug group (Figure 6B), but the serum TNFα had no much change compared the drug group with control group (Figure 6C). These data indicate that the survival rate of LPS induced sepsis can be improved by 1,25(OH)2D3 via blocking HMGB1 secretion.

DiscUssiOn
Sepsis is a common indication for ICU admission and is associated with marked morbidity and mortality. A number of observational studies have shown low vitamin D levels related to the risk of sepsis (27)(28)(29). However, the effectiveness of vitamin D supplementation in sepsis treatment is contradic tory in clinical trial (29,30). In our study, the treatment with paricalcitol in mouse model of LPSinduced sepsis increased the survival rate of mice and decreased the LPSinduced HMGB1 secretion in serum (Figure 6). As a result, our study provides novel information on the role of vitamin D in sepsis via the Nrf2-HO1-HMGB1 pathway (Figure 7). Pretreatment with the noncalcemic vitamin D analog pari calcitol was previously reported by us to suppress LPSinduced inflammation via the MicroRNA155-SOCS1 pathway (31), which suggested that vitamin D has a good preventive effect on sepsis. But the therapeutic effect of 1,25(OH)2D3 for sepsis is not clear till now. In this study, we found that the concentration of   (33). Our finding also verified that vitamin D has no effect on TNFα expression, but HMGB1 as a new regulation factor by 1,25(OH)2D3 came to light. The inhibition of HMGB1 secretion by 1,25(OH)2D3 was due to block the nuclear export of HMGB1 (Figure 2). HMGB1 cytoplasmic translocation is controlled by HO1 (12). Here, we showed that the level of HO1 expression was upregulated by 1,25(OH)2D3 (Figure 4). Consistent with our finding, 1,25(OH)2D3 was reported to upregulate HO1 expression in glial cells (34), and HO1 expression was decreased in vitamin D deficiency in obese rats (35). Then, the important role of HO1 in inhibition of HMGB1 secretion by 1,25(OH)2D3 was testified by interference experiment with HO1 inhibitor or siHO1 (Figure 3). Nrf2 is a crucial transcription factor for HO1. We further found that Nrf2 can be activated by 1,25(OH)2D3 and bound to the promoter of HO1 (Figure 5) in macrophage. The Nrf2/ARE pathway can be activated by posttranscriptional activating Nrf2 via phos phorylation by signaling protein kinases (PKC, MAPKs, and/ or PI3K) (36)(37)(38)(39). Meanwhile, 1,25(OH)2D3 can regulate PKC, MAPKs, and/or PI3K pathways (40)(41)(42). These report gener ate clues for the mechanism of the regulation of 1,25(OH)2D3 on the Nrf2 activation. In short, the possible modes of vitamin D inhibiting LPSInduced HMGB1 secretion via targeting the Nrf2-HO1-HMGB1 pathway in Macrophages are represented in Figure 7. However, the precise regulation of 1,25(OH)2D3 in the Nrf2-HO1-HMGB1 pathway remains unclear, and the detailed molecular mechanism needs to be fully defined in future studies.

reagents and cytokine Quantization
ZnPPIX was bought from Sigma. TNFα and HMGB1 con centrations in the serum or culture media were determined by ELISA using commercial ELISA kits obtained from BioLegend (San Diego, CA, USA). Data were analyzed by Student's ttest.
A pvalue <0.05 was considered statistically significant.
cell culture and Treatment L929 and RAW264.7 cells were grown in DMEM supplemented with 10% FBS. BMDMs were cultured as described previously (43). In brief, mouse bone marrow cells were plated in DMEM supplemented with 10% FBS. After overnight culture, the unat tached cells were replated and differentiated into BMDMs in 30% L929 conditioned media. Cells cultures were usually treated with 100-200 ng/ml LPS with or without 1,25(OH)2D3 treatment as specified in each experiment, followed by the isolation of total RNAs, lysates, or media supernatants (SN) for various assays.

Preparation of cytoplasm and nuclear extracts
Cytoplasm and nuclear extracts were prepared using the Nuclear and Cytoplasm Protein Extraction Kit according to the manu facturer's instructions (Beyotime Institute of Biotechnology). Briefly, cells were scraped off, washed in icecold PBS, and then resuspended in 200 µl of icecold cytoplasm extraction buffer A with 1 mM PMSF, 1 mM Na4VO3, and protease inhibitor mixture. After incubation with cytoplasm extraction buffer B for 1 min in ice bath and following vortexing for 5 s, cell lysates were centrifuged at 12,000 × g for 5 min at 4°C. SN were aliquoted and stored at −80°C. Nuclear pellets were resuspended in 50 µl of nuclear extraction buffer. After 15 sets of vortexing for 15 s every 2 min at 4°C, lysates were centrifuged at 12,000 × g for 10 min at 4°C. Nuclear extracts were aliquoted and stored at −80°C until use. aUThOr cOnTriBUTiOns YC and ZR designed the research, analyzed data, and wrote the paper; NZ, NX, YP, MX, and JW provided research reagents and technical assistance; SY and HZ assisted in data analysis and manuscript preparation; YC was responsible for the overall research design, data analysis, and paper preparation.   FigUre s2 | Negative control for anti-NF-E2-related factor 2 (Nrf2) and anti-high-mobility group box 1 (HMGB1) in immunofluorescence with isotype antibody. The cells were stained with isotype antibody and second antibody (red), the nuclei were visualized with 4′,6-diamidino-2-phenylindole (DAPI) staining (blue). Scale bar = 20 μm.