Tris DBA Ameliorates Accelerated and Severe Lupus Nephritis in Mice by Activating Regulatory T Cells and Autophagy and Inhibiting the NLRP3 Inflammasome

Tris (dibenzylideneacetone) dipalladium (Tris DBA), a small-molecule palladium complex, has been shown to inhibit cell growth and proliferation in pancreatic cancer, lymphocytic leukemia, and multiple myeloma (1–3). It has also been shown to reduce the level of the Src/NMT-1 complex in melanoma cells and inhibit downstream signaling molecules of the complex, including MAPK and phosphoinositol-3-kinase (4). However, the effects of Tris DBA on inflammatory diseases have yet to be determined. Given that systemic lupus erythematosus (SLE) is mediated in large part by B cells and that Tris DBA has been shown to be effective against B cell malignancies, we evaluated Tris DBA in the context of lupus nephritis (LN), a major renal complication of SLE. In this study, we evaluate the efficacy of this agent against an inflammatory disease.

LN is the leading cause of morbidity and mortality in SLE patients (5, 6). African American and Asian populations are disproportionately affected by LN. In the past 50 y, only one drug has been approved for the treatment of SLE (Belimumab, Benlysta), and unfortunately, the benefits of this new drug are modest. It appears to be more efficacious in the white population, which is less afflicted with LN, than in the African American population and has a modest effect on LN itself. Other therapies, such as cyclophosphamide, high-dose steroids, and other disease-modifying antirheumatic drug agents, have benefits but also have long-term side effects, including immunosuppression and other effects associated with chronic corticosteroid use. Therefore, novel agents are urgently needed for LN.

Recently, MAPK activity and NLRP3 inflammasome activation have been strongly implicated in the development and progression of LN in patients (7–9) and LN models (10–12) and have been implicated in the pathogenesis of the progression and deterioration of LN. It has been established that the NLRP3 inflammasome mediates the production of mature IL-1β and IL-18 (13, 14). In particular, the activation of the NLRP3 inflammasome involves a priming signal from pathogen-associated molecular patterns and an activation signal (e.g., ATP), which is generated from damaged cells (13–15). Notably, immune complexes can initiate the activation of the NLRP3 inflammasome in macrophages from patients with SLE and animal models (16–19), which in turn results in cell and tissue injury. Recently, we have shown the overproduction of IL-1β or IL-18 and T cell activation in mouse LN models, pointing to the NLRP3 inflammasome and related signaling pathways as a major mechanism underlying the progression and deterioration of LN (19, 20).

In the current study, we examined the therapeutic effect of Tris DBA on a severe type of mouse LN model, namely, the accelerated and severe LN (ASLN) model, to simulate the histopathological transformation to a severe form (e.g., Class IV LN) from a mild mesangial LN form (e.g., Class I or Class II LN) as a result of acute onset or acceleration of the renal condition. These results show that Tris DBA effectively ameliorates this severe type of LN through a process that involves blunting a MAPK (ERK, JNK)-mediated priming signal of the NLRP3 inflammasome and regulating the autophagy/NLRP3 inflammasome axis and T cell activation. Our study might justify Tris DBA as a drug candidate for accelerated and deteriorated status LN.

Tris DBA palladium was obtained from Sigma-Aldrich (catalog number 328774; St. Louis, MO) and dissolved in Katimin (China Chemical & Pharmaceutical, Taipei City, Taiwan) for in vivo studies or in DMSO for in vitro studies. There has been no evidence that Tris DBA has been shown to cause any detectable clotting problem (1–3).

Eight-week-old female NZB/WF1 mice (younger than the age of autoantibody production) were purchased from The Jackson Laboratory and injected with 0.6 mg/kg body weight LPS (Sigma-Aldrich) i.p. twice weekly until sacrificed at weeks 3 and 5, respectively (16, 19, 20). Tris DBA (30 mg/kg of body weight) or vehicle (Katimin) was given daily by i.p. injection 1 wk after the LPS injections were completed. The 8-wk-old female NZB/WF1 mice that received Katimin only were used as normal control.

Twelve-week-old female NZB/W F1 mice, were divided into two groups, one of which was given the vehicle (Katimin) and served as the disease control, whereas the other was given 30 mg/kg of body weight of Tris DBA by i.p. injection daily until sacrificed at 32 wk old. Eight-week-old NZB/W F1 female mice (prior to onset of autoantibody production) were used as normal control.

All the animal experiments were performed with the approval of the Institutional Animal Care and Use Committee of The National Defense Medical Center, Taiwan, and were conducted in compliance with national guidelines.

Urine samples were collected weekly on ASLN and every 2 wk on spontaneous LN (spLN) using metabolic cages, whereas serum samples were collected at the points when the animals were sacrificed and stored appropriately until analysis. For renal function testing, serum levels of blood urea nitrogen (BUN) and creatinine (Cr) were measured using BUN kits or Cr kits (both from Fuji Dry-Chem Slide, Fuji Film Medical, Tokyo, Japan) as described previously (16, 19, 20). The concentration of urine albumin was determined by ELISA (Exocell, Philadelphia, PA), and urine albumin levels were expressed relative to the urine Cr levels measured using a kit from Fuji Film Medical as described previously (16).

For histopathological assessment, renal cortical tissue samples were fixed in 10% buffered formalin and embedded in paraffin. Renal sections (2–4 μm) were prepared and stained with H&E. Briefly, the following light microcopy features were evaluated: 1) proliferation of glomerular cells, defined as ≥3 mesangial cells per mesangial area or global increase of the intrinsic cells (mesangial cells, endothelial cells, and/or podocytes) in the glomerulus; 2) crescent, defined as cellular crescentic formation in the glomerulus, with at least two layers of podocytes or parietal epithelial cells in a segmental or global pattern; 3) neutrophil, defined as neutrophil infiltration inside the glomerulus; 4) fibrinoid necrosis, defined as necrosis with fibrin-like appearance, present inside the glomerulus; 5) sclerosis, defined as the disappearance of cellular elements from the glomerular tufts, collapse of the capillary lumen, and folding of the glomerular basement membrane with entrapment of amorphous material; and 6) peri-glomerular tubulointerstitial inflammation, defined as mononuclear leukocyte infiltration in the peri-glomerular compartment of the renal interstitium. Morphologic changes (1–5) were quantitated on a scale of 0–3⁺ (0, normal; trace; 1⁺, mild; 2⁺, moderate; 3⁺, severe) for a total of 40 or more glomeruli examined. A total score was then obtained by multiplying the severity of the changes by the percentage of glomeruli with the same severity of alterations. The extent of injury (total score) for individual tissue specimens was then obtained by the addition of these scores using an equation with mild modification as described previously (21). The values ranged from 0 to a maximum of 300. The percentage of the lesion with the peri-glomerular tubulointerstitial inflammation was expressed directly as its percentage to the total of 40 or more glomeruli examined. The scores obtained by two investigators (A.C. and S.-S.Y.) were averaged. For the immunohistochemical staining for F4/80, CD3, CD4, and CD8 formalin-fixed and paraffin-embedded renal sections were incubated with primary Abs against the mouse Ag of interest, followed by incubation with biotinylated secondary Abs and avidin-biotin-peroxidase complexes (both from Dako, Glostrup, Denmark) as described previously (19, 20). An image analysis software (Pax-it; Paxcam, Villa Park, IL) was adopted, an image analysis tool that can integrate density, color, shape factors, and size filters (https://www.paxit.com) used to detect and sort objects or areas, to quantify CD3, CD4, CD8, and F4/80 expression by counting the number of positive cells in 20 randomly selected fields of the glomeruli and peri-glomerular tubulointerstitial compartments in the cortical area at a magnification of ×400 by light microscopy, and the data are expressed as cells per field (22). For immunofluorescence (IF), frozen sections of renal tissues were stained with FITC-conjugated Abs against IgG, IgM, or C3 (Cappel Laboratories, Cochranville, PA) as described previously (20, 23), and a scoring system described previously was adopted to measure total fluorescence intensity (20, 24).

Levels of reactive oxygen species (ROS) in the urine, serum, and renal tissue were detected as described previously (16, 20). The results are presented as relative luminescence units (RLU) per 15 min per milliliter (RLU/15 min/ml) for the urine and serum samples or RLU/15 min/mg dry weight for the renal tissue. In parallel, renal levels of superoxide anion were determined by dihydroethidium (DHE) staining, and the fluorescence was quantified by counting the percentage of the total nuclei that were positive per kidney cross-section, as described previously (16, 20). Mitochondrial ROS levels in bone marrow–derived macrophages (BMDMs) were measured by MitoSOX-based flow cytometry (Thermo Fisher Scientific, Waltham, MA). Levels of ROS in cultured cells were stained with H2DCFDA (Thermo Fisher Scientific) by luminescence assay according to the manufacturer’s instructions.

Splenocytes obtained from the mice were used for the detection of activation in CD4 memory T cells, CD8 memory T cells, regulatory T (Treg) cells, and B cells (all Abs obtained from BD Biosciences, San Diego, CA), using a FACS Calibur flow cytometer (BD Biosciences) as described previously (25, 26). For measuring the expression of intracellular IFN-γ or IL-4, splenocytes were cultured for 5 h in 24-well plates and stained for IFN-γ or IL-4 in the presence or absence of PMA, ionomycin, and monensin (all from Sigma-Aldrich). Then, allophycocyanin-conjugated anti-mouse CD4 Abs (BD Biosciences) were added to the reaction, and the cells were resuspended in permeabilization buffer (eBioscience, San Diego, CA), incubated with PE-conjugated Abs against IFN-γ or IL-4 (both from BD Biosciences) and assessed using a FACS Calibur Flow Cytometer (BD Biosciences) (27). For the ex vivo experiment, splenocytes isolated from 8-wk-old NZB/W F1 mice were primed with LPS in the presence or absence of Tris DBA for 24 h and then expression levels of indicated markers were analyzed using a flow cytometer to measure the activation of B cells. The cells were stained by PE-conjugated anti-CD20 and FITC-conjugated anti-CD86 (BD Biosciences) Abs. The relative levels of each marker were calculated in comparison with those of B cells cultured with vehicle (DMSO) as normal control.

Splenocytes were cultured in triplicate in wells (5 × 10⁵ cells per well) in 96-well flat-bottom microtiter plates previously coated overnight at 4°C with 0.25 μg/ml anti-mouse CD3 Ab (BD Biosciences). After 48 h, the cultured cells were pulsed with [³H] methylthymidine (Amersham International, Buckinghamshire, U.K.) and harvested 16 h later. Then, the incorporated [³H] methylthymidine was measured using a TopCount Scintillation Counter (PerkinElmer Life Sciences, Palo Alto, CA).

Treg (CD4⁺ CD25⁺ CD45R⁻ Foxp3⁺) cells and conventional T cells (CD4⁺ CD25⁻ CD45R⁺ Foxp3⁻) were isolated from the spleen and lymph nodes of mice in week 5 at the end of the experiment by FACS (FACS Calibur Flow Cytometer; BD Biosciences). These cells, in 2-fold serial dilutions, were used in suppression assays against the cultures of anti-CD3 and anti-CD28 Ab-stimulated T cells.

This was performed in mice that were given an i.p. injection of 1 μg of LPS (Sigma-Aldrich), in 0.5 ml of PBS, followed by a second i.p. injection of ATP (Sigma-Aldrich) in 0.5 ml of PBS as descried previously with mild modification (28). Tris DBA and rapamycin (Sigma-Aldrich), an autophagy agonist, were used throughout the experiment.

Cytoplasmic protein was extracted from renal tissue using RIPA lysis buffer, and target proteins were detected by SDS-PAGE electrophoresis. Immunoblotting was performed using Abs against Atg5 (Medical and Biological Laboratories, Nagoya, Japan), LC3B (GeneTex, Irvine, CA), p62 (Abcam, Bristol, U.K.), NLRP3 (GeneTex), caspase-1 (AdipoGen, Zweigniederl Liestal, Switzerland), IL-1β (R&D Systems, Minneapolis, MN), NADPH p47phox (p47phox) (Santa Cruz Biotechnology, Santa Cruz, CA), NQO1 (Abcam, Cambridge, U.K.), p-ERK (Santa Cruz Biotechnology), p-JNK (Santa Cruz Biotechnology), and p-p38 (Santa Cruz Biotechnology). HRP-conjugated IgG Abs (Dako) were used. An Ab against β-actin (Santa Cruz Biotechnology) was used as an internal control for the cytosolic proteins.

The murine macrophage cell line J774A.1 was purchased from the American Type Culture Collection (Rockville, MD). The cells were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated FCS and 2 mM l-glutamine (all from Life Technologies, Carlsbad, CA) at 37°C in a 5% CO₂ incubator. In brief, 1 × 10⁶ cells/ml were incubated for 30 min in the presence or absence of Tris DBA before being primed with LPS for 5.5 h and stimulated with ATP (5 mM) (Sigma-Aldrich) for an additional 30 min. The levels of IL-1β in the culture medium were measured by ELISA as described previously (27, 29). The expression levels of NLRP3, caspase-1, procaspase-1, mature IL-1β, and pro–IL-1β in the cell lysates or supernatants were determined by Western blot analysis as described previously (29). PD98059 (ERK inhibitor; 5, 10, and 20 μM), SP600125 (JNK inhibitor; 5, 10, and 20 μM) (Calbiochem, La Jolla, CA) and 3-methyladenine (3-MA) (autophagy inhibitor; 5 mM) (Sigma-Aldrich) were used to detect the activation of NLRP3 inflammasome as reported previously (29).

Bone marrow–derived dendritic cells (BMDCs) were isolated from 8-wk-old female NZB/WF1 mice (younger than the age of autoantibody production). In brief, BMDCs were flushed out of the cavities of the femur and tibia with RPMI 1640 medium. BMDCs were then isolated and seeded in 24-well plates in RPMI 1640 medium supplemented with 10% FBS and 10 ng/ml mouse GM-CSF for 7 d as described previously (30). The cells were then treated with LPS (1 μg/ml) and Tris DBA (1 μM), stained with anti-mouse CD11c/CD80 or CD11c/CD86 Abs combinations (BioLegend, San Diego, CA), and assessed by flow cytometry. In parallel, presentation of Ag by dendritic cells (DCs) was analyzed using CD4⁺ T cells from transgenic mice expressing OT-II TCRs and the OVA-specific T cell proliferation in vitro assay as described previously (31). Purified DCs were pulsed with OVA_323–339 peptide and incubated with LPS or Tris DBA + LPS for 16 h. CD4⁺ T cells isolated from OT-II mice were added to the BMDCs cultures at various cell ratios, and T cell proliferation was determined by [³H]thymidine incorporation. To measure IL-4 or IFN-γ production, supernatants were collected, and the amount of IL-4 or IFN-γ was determined by ELISA kits (both from eBioscience).

IL-1β, TNF-α, and IL-6 protein levels were measured by ELISA kits according to the manufacturer’s protocols (all from eBioscience). Serum levels of anti-dsDNA autoantibodies were measured using an anti-mouse dsDNA IgG autoantibody ELISA kit (Alpha Diagnostic, San Antonio, TX) according to the manufacturer’s instructions. The absorbance at 450 nm was measured using an ELISA plate reader (Bio-Tek, Winooski, VT).

The data are presented as the mean ± SEM, and comparisons between two groups were performed using Student t test. The data from in vitro and ex vivo experiments were analyzed using one-way ANOVA and subsequent Scheffe test. A p value < 0.05 was considered statistically significant for each of the experiments.

To validate the potential therapeutic effects of Tris DBA on ASLN mice, we first performed a clinical assessment and analyzed renal pathology. Tris DBA was given daily after the onset of ASLN in female NZB/WF1 mice (Tris DBA + ASLN), and ASLN mice that received vehicle only were used as the disease control (Vehicle + ASLN), whereas NZB/WF1 mice given vehicle only were used as the normal control. As shown in Fig. 1A, 1B, significantly improved renal function was observed at weeks 3 (as demonstrated by lower serum BUN levels; Fig. 1A) and 5 (as demonstrated by lower serum BUN and Cr levels; Fig. 1A, 1B) in the Tris DBA + ASLN mice compared with the Vehicle + ASLN mice. In parallel, the Tris DBA + ASLN mice showed markedly reduced albuminuria compared with the Vehicle + ASLN mice, which began developing at week 3 and increased albuminuria at week 5, at the end of the study (Fig. 1C). In addition, the Vehicle + ASLN mice developed obvious intrinsic cell proliferation, cellular crescents, neutrophil influx, fibrinoid necrosis areas in the glomeruli, peri-glomerular tubulointerstitial inflammation, and increased glomerulonephritis activity score, but the severities of these pathological changes in the kidneys were greatly reduced in the Tris DBA + ASLN mice (Fig. 1D, 1E, Supplemental Fig. 1A). In parallel, the Tris DBA + ASLN mice showed significantly reduced glomerular deposition of IgG, IgM, and C3 by IF compared with Vehicle + ASLN mice at week 3, although not at week 5, at which Tris DBA– or vehicle-treated ASLN mice showed almost equal intensity of these immune deposits in the glomerulus (Fig. 1F–K, Supplemental Fig. 1B–D). In addition, autoantibody-induced immune complex deposition in the kidneys is a major cause of LN (32–34), so we then tested the serum levels of anti-dsDNA autoantibody to assess the therapeutic effect of Tris DBA. As shown in Fig. 1L, although serum anti-dsDNA autoantibody levels in Vehicle + ASLN mice were significantly increased as early as week 3 and further increased at week 5 compared with those of the normal control, these effects were significantly inhibited in the Tris DBA + ASLN mice at week 3, although not at week 5, at which Tris DBA– or vehicle-treated ASLN mice showed almost equal intensity of these immune deposits in the glomerulus. By immunohistochemical staining, the phenotype and distribution of mononuclear leukocytes that infiltrated the kidney were detected. The results showed that F4/80⁺ monocytes/macrophages, CD3⁺, CD4⁺, and CD8⁺ T cells (Fig. 1M, 1N, Supplemental Fig. 1E, 1F) were increased in the Vehicle + ASLN mice compared with the normal control mice at weeks 3 and 5. However, the Tris DBA + ASLN mice exhibited significantly reduced infiltration of these mononuclear leukocytes in the kidneys at weeks 3 and 5 compared with Vehicle + ASLN mice, although the degree of these changes in week 5 was significantly higher than that of the normal control. All the animals survived at the end of the experiment (at week 5).

It is well established that T cell activation and differentiation are actively involved in the development and progression of SLE (35–38). First, we examined the activation of memory T cells from the splenocyte population. Although the Vehicle + ASLN mice showed increased activation of splenic T cells at week 5, as demonstrated by the total percentage of CD4⁺CD44^hiCD62^lo-hi and CD8⁺CD44^hiCD62^lo-hi memory T cells (Fig. 2A, 2B), these effects were inhibited in the Tris DBA + ASLN mice. In contrast, Treg cells have been shown to play a protective role in SLE (31, 38) and LN (39, 40). We determined the population of Foxp3⁺CD25⁺CD4⁺ Treg cells, and the results showed that a significantly increased percentage of Foxp3⁺CD25⁺CD4⁺ Treg cells was observed at week 5 in the Tris DBA + ASLN mice compared with the Vehicle + ASLN mice (Fig. 2C). In addition, the Tris DBA + ASLN mice showed markedly decreased T cell proliferation at week 5 compared with the Vehicle + ASLN mice, which showed increased levels of T cell proliferation at week 5 (Fig. 2D). The Treg (CD4⁺CD25⁺CD45R⁻Foxp3⁺) cells and conventional T cells (CD4⁺CD25⁻CD45R⁺Foxp3⁻) were isolated from the spleen and lymph nodes of the Vehicle + ASLN mice and Tris DBA + ASLN at week 5 at the end of the experiment. The results showed that Tris DBA treatment significantly inhibited the proliferation of conventional T cells by enhancing Treg cell activity as demonstrated by the T cell proliferation assay (Fig. 2E). Intracellular staining for IFN-γ and IL-4 in CD4 T cells showed no significant difference between the Tris DBA + ASLN mice and Vehicle + ASLN mice (Supplemental Fig. 1G). In addition, we performed flow cytometry with CD20 and CD86 using LPS-primed splenocytes from 8-wk-old NZB/W F1 mice, and the relative levels of the CD20⁺CD86⁺ were calculated in comparison with those of B cells cultured with vehicle as normal control. As shown in Fig. 2F, Tris DBA treatment significantly inhibited the CD20⁺CD86⁺ B cells activation, whereas increased relative levels of the CD20⁺CD86⁺ were observed in LPS-primed splenocytes compared with normal control.

DCs and the associated adaptive immunity pathways play crucial roles in human immune defense (41, 42). However, dysregulated activation of DCs is implicated in the pathogenesis of ASLN (20, 43). As shown in Fig. 2, although increased levels of IL-1β, TNF-α, and IL-6 and percentage of CD11c⁺CD86⁺ and CD11c⁺CD80⁺ cells were observed in BMDCs isolated from normal control mice, these effects were markedly inhibited in Tris DBA–treated LPS-primed (Tris DBA + LPS) BMDCs (Fig. 2G–J). The results of an OVA-specific CD4⁺ T cell proliferation assay showed that although DCs promoted T cell proliferation after LPS stimulation (LPS + Vehicle), this effect was greatly diminished in Tris DBA + LPS CD4⁺ T cells (Fig. 2K). In parallel, the expression levels of IFN-γ and IL-4, which are specific to Th1 and Th2 cells, respectively, were also determined for the supernatants of the OVA-specific CD4⁺ T cell cultures. The results showed that significantly reduced levels of IFN-γ and IL-4 were observed in the Tris DBA + LPS compared with those of Vehicle + LPS (Fig. 2L, 2M).

The NLRP3 inflammasome has been implicated in the pathogenesis of LN (8–10). To evaluate the effect of Tris DBA on the activation of the NLRP3 inflammasome as the potential mechanism of action of the therapeutic effect of Tris DBA on ASLN mice, renal NLRP3 and IL-1β protein levels were determined. As shown in Fig. 3A, 3B, significantly reduced levels of these proteins were observed in Tris DBA + ASLN mice compared with Vehicle + ASLN mice at weeks 3 and 5.

Notably, autophagy can regulate the activation of the NLRP3 inflammasome in SLE (44) and LN (45, 46). First, we determined the effect of Tris DBA on the induction of autophagy using renal tissue obtained from normal mice. As shown in Fig. 3C, 3D, renal levels of the LC3B I/II proteins were markedly elevated until day 3 after Tris DBA administration and then declined on day 7 and day 14. Next, we examined whether Tris DBA could induce autophagy in the kidneys of ASLN mice. Unexpectedly, although renal LC3B I/II and Atg5 protein levels were higher in the Vehicle + ASLN mice than in the normal control at weeks 3 and 5, this effect was inhibited in the Tris DBA + ASLN mice compared with the Vehicle + ASLN mice at weeks 3 and 5 (Fig. 3E, 3F). In this setting, the decreased autophagy induction in the Tris DBA–treated ASLN mice could represent negative feedback during the course of the improvement of ASLN mice. However, whether the reduced autophagic response could follow the improved renal condition mediated by treatment with Tris DBA deserves further investigation.

To confirm whether Tris DBA could trigger autophagy induction and whereby inhibited NLRP3 inflammasome and subsequent IL-1β expression, an acute peritonitis was induced in NZB/W F1 mice pretreated with Tris DBA or rapamycin (an autophagy agonist) followed by the treatment with LPS and ATP. The results showed that Tris DBA and autophagy agonist was able to significantly 1) increase the production of LC3B protein and 2) reduce IL-1β production in the peritoneal macrophages and fluid of the mice (Fig. 3G, 3H).

ROS and their related molecular pathways are implicated in the progression of SLE (47, 48) and LN (16, 19, 20). First, in contrast to those of the Vehicle + ASLN mice, which exhibited high levels of ROS in the serum, urine, and renal tissue, the Tris DBA + ASLN mice exhibited reduced levels of the superoxide anion at weeks 3 and 5 by a luminescence assay, and the levels in the Tris DBA–treated ASLN mice were similar to those observed in the normal control mice at the two time points (Fig. 4A). In addition, the detection of in situ ROS production was performed using DHE fluorescent staining. As shown in Fig. 4B, enhanced production of renal ROS was observed in the Vehicle + ASLN mice at both weeks 3 and 5 compared with the saline control mice, but this effect was inhibited in the Tris DBA + ASLN mice. In addition, the protein levels of renal p47phox were measured, and the results showed significantly increased renal p47phox expression in the Vehicle + ASLN mice at weeks 3 and 5 compared with the normal control mice, but this effect was inhibited in the Tris DBA + ASLN mice (Fig. 4C). In contrast, activation of antioxidant responses is a major mechanism of cellular defense against ROS (22, 23). As shown in Fig. 4, markedly increased renal NQO1 levels were observed in the Tris DBA + ASLN mice at weeks 3 and 5 compared with the Vehicle + ASLN mice (Fig. 4C). Moreover, the activity of renal glutathione peroxidase (GPx) was significantly increased in the Tris DBA + ASLN mice at weeks 3 and 5 compared with the normal control mice. Compared with the Vehicle + ASLN mice, the Tris DBA + ASLN mice showed significantly elevated renal GPx activity at week 5 (Fig. 4D).

To define cellular ROS production, we analyzed mitochondrial ROS production and total ROS in BMDMs. As shown in Fig. 4E, 4F, enhanced production of both mitochondrial and total ROS were observed in the Vehicle + LPS BMDMs compared with the control BMDMs, but only the mitochondrial ROS levels were markedly reduced in the Tris DBA + LPS BMDMs.

The potential mechanism of action for Tris DBA involving an inhibitory effect on NLRP3 inflammasome activation was further dissected in related cell models. Notably, various factors and molecules are capable of initiating the priming signals for the activation of the NLRP3 inflammasome in LPS-primed macrophages (8, 9, 12). Among them, MAPK is a potent activator of the priming signal of the NLRP3 inflammasome (49). As shown in Fig. 5A, increased expression levels of p-ERK, p-JNK, and p-p38 were observed in the Vehicle + LPS macrophages compared with the control macrophages. Compared with the Vehicle + LPS macrophages, the Tris DBA + LPS macrophages showed significantly inhibited the expression levels of p-ERK and p-JNK, although there was no difference in the p-p38 between the two groups. Next, macrophages were treated with Tris DBA followed by LPS stimulation before the cells were subjected to the detection of their expression levels of NLRP3 and pro–IL-1β. As shown in Fig. 5B, significantly reduced levels of the two proteins were observed in the Tris DBA + LPS macrophages compared with the Vehicle + LPS cells. To further confirm this observation, two MAPK inhibitors, PD98059 (to ERK) and SP600125 (to JNK), were used, and treatment with these inhibitors resulted in significantly reduced expression levels of NLRP3 and pro–IL-1β (Fig. 5C, 5D). In addition, ROS are capable of triggering NLRP3 inflammasome activation and serve as a group of effector molecules that lead to pathological processes (50). Thus, we determined the expression levels of ROS and plausibly resultant NF-κB activation using J774A.1 macrophages. Although increased ROS production and NF-κB activation were both observed in the Vehicle + LPS cells compared with the normal control, the Tris DBA + LPS treatment failed to inhibit these effects in the cells (Fig. 5E, 5F). Next, the activation signal for NLRP3 inflammasome activation was examined by determining the levels of caspase-1 and IL-1β in macrophages. As shown in Fig. 6, upon treatment with ATP (NLRP3 inflammasome activator), increased secretion levels of mature IL-1β (Fig. 6A) and caspase-1 (Fig. 6B) were observed in the LPS-primed macrophages compared with the normal control, but this effect was significantly inhibited in the Tris DBA–treated LPS-primed cells.

Notably, autophagy can inhibit the priming and activation signals of the NLRP3 inflammasome (26, 44, 45, 49). We thus evaluated whether Tris DBA could regulate the autophagy/NLRP3 inflammasome axis to render its therapeutic effects on ASLN mice. First, we hypothesized that Tris DBA could induce autophagy in macrophages without LPS stimulation. As shown in Fig. 6C, 6D, treatment with Tris DBA significantly increased the protein levels of LC3B, Atg5, and p62 at 6 h in J774A.1 macrophages in a dose-dependent manner. Moreover, both the priming and activation signals (Fig. 6A, 6B) of NLRP3 inflammasome activation were markedly reduced in the Tris DBA + LPS macrophages compared with the Vehicle + LPS macrophages, but this effect was inhibited by treatment with 3-MA, an autophagy inhibitor (Fig. 6E–G). Our data suggest that the Tris DBA could render its beneficial effects on ASLN mice by inhibiting NLRP3 inflammasome in part through autophagy induction.

In addition to the ASLN model, we also validated the potential therapeutic effects of Tris DBA in spLN model after the onset of renal disease at the age of 12 wk old, including clinical and pathological assessments. Three of the 10 mice treated with vehicle only (Vehicle + spLN) died before sacrificed at week 32 at the end of the experiment, whereas all of the mice treated with Tris DBA (Tris DBA + spLN mice) survived until the end of the experiment (Fig. 7A). As shown in Fig. 7B, 7C, renal function was significantly improved in the Tris DBA + spLN mice, compared with that of the Vehicle + spLN mice. In addition, the Tris DBA + spLN mice showed significantly decreased albuminuria compared with the Vehicle + spLN mice, which began at week 28 and persistently increased until week 32, the end of the study (Fig. 7D). By light microscopy, although the Vehicle + spLN mice developed obvious intrinsic cell proliferation, cellular crescents, neutrophil influx, fibrinoid necrosis and sclerosis in the glomerulus, and peri-glomerular tubulointerstitial inflammation as well as increased glomerulonephritis activity score, the severities of these pathological changes in the kidney were greatly reduced in the Tris DBA + spLN mice (Fig. 7E, 7F). In parallel, IF showed diffuse, intense glomerular deposition of IgG, IgM, and C3 in the Vehicle + spLN mice, but these effects were markedly inhibited in the Tris DBA–treated spLN mice (Fig. 7G–I). In addition, although serum anti-dsDNA autoantibody levels in Vehicle + spLN mice were increased compared with those of the normal control, the Tris DBA + spLN mice depicted significantly reduced serum autoantibody levels (Fig. 7J). Furthermore, we examined the activation of memory T cells obtained from the splenocyte population. As shown in Fig. 7K, 7L, although the Vehicle + spLN mice showed increased activation of T cells (as demonstrated by increased total percentage of CD4⁺CD44^hiCD62^lo-hi and CD8⁺CD44^hiCD62^lo-hi memory T cells), these effects were inhibited in the Tris DBA + spLN mice. Next, the population of Foxp3⁺CD25⁺CD4⁺ Treg cells was determined, and the results showed a significantly increased percentage of Foxp3⁺CD25⁺CD4⁺ Treg cells at week 32 in the Tris DBA + spLN mice compared with that of the Vehicle + spLN mice (Fig. 7M). Furthermore, the Tris DBA + spLN mice showed markedly decreased T cell proliferation compared with the Vehicle + spLN mice, which revealed significantly increased levels of T cell proliferation compared with normal control (Fig. 7N). Additionally, we detected the activation of splenic B cells of the mice. As shown in Fig. 7O, 7P, the percentage of CD19⁺CD69⁺ B cells or B220⁺CD69⁺ B cells was significantly increased in the Vehicle + spLN mice as compared with that of normal control mice, but Tris DBA treatment significantly inhibited this effect.

Our study justified Tris DBA as a drug candidate for the treatment of LN, and most likely, the prevention of the histopathological transformation of mild LN into higher-grade glomerulonephritis by regulating autophagy/NLRP3 inflammasome and inhibiting the phosphorylation of JNK, ERK, and p38 MAPK signaling pathways. We found that the regulation of JNK, ERK, and p38 MAPK signaling pathways and blunting of ROS-mediated inflammatory reactions are both involved in the mechanism of action by which Tris DBA renders its therapeutic effects on ASLN mice. This observation was consistent with the findings showing reduced renal ROS levels in the Tris DBA + ASLN mice compared with the Vehicle + ASLN mice and markedly suppressed mitochondrial ROS production in the Tris DBA + LPS BMDMs. Besides, Tris DBA has been shown to induce killing in chronic lymphocytic leukemia B cells, which is of clinical interest (2). Furthermore, it can reduce Src/NMT-1 complex in melanoma tumor cells (51) and multiple myeloma (3).

In the current study, we showed that Tris DBA treatment can induce autophagy in macrophages in a dose-dependent manner. Moreover, this finding was confirmed by reduced expression of the priming (Fig. 6B) and activation (Fig. 6A) signals of the NLRP3 inflammasome in LPS-primed macrophages treated with Tris DBA, but the activation of the inflammasome was greatly restored by treatment with 3-MA, an autophagy inhibitor. All the findings obtained in vitro clearly indicated a beneficial effect of using Tris DBA to treat ASLN mice, given that greatly improved renal condition was observed in the treated mice. Consistent with the in vitro experiments, we clearly revealed dramatically increased autophagy expression as early as the first day after the administration of the compound, which persisted for at least 3 d, in normal mice treated with Tris DBA as a comparative experiment to confirm the augmented induction of autophagy and resultant potential therapeutic effect on the ASLN mice that were treated with Tris DBA in the study (Fig. 3). However, we demonstrated that reduced autophagy was observed in the ASLN mice treated with Tris DBA at the two time points used in the current study. Whether a reduced autophagic response could follow an improved renal condition induced by treatment with Tris DBA has yet to be determined. Although the exact mechanism for these seemingly opposing findings requires further investigation, we infer that uncertain complicated interactions between autophagy induction, NLRP3 inflammasome activation and compensatory reactions might be occurring in the ASLN mice. This hypothesis can be supported by the following: 1) increased autophagy induction was detected in the Tris DBA–treated macrophages (Fig. 6), whereas in the macrophages treated with both Tris DBA and LPS, the use of an inhibitor of autophagy (3-MA) restored the levels of IL-1β protein secretion (Fig. 6), suggesting that the treatment with Tris DBA could induce autophagy in the cells treated with LPS; 2) increased autophagy induction was observed in renal tissue obtained from normal mice treated with Tris DBA (Fig. 3); and 3) reduced autophagy expression could also represent negative feedback during the course of the improvement of the severe type of LN in mice during the study period.

Importantly, the results ex vivo reported in T cell proliferation may not reflect the effects in vivo of Tris DBA. An immunocytochemistry using Ki-67 staining together with intracellular cytokine staining on sorted cells may be worthwhile to better assess the T cell proliferation for the ex vivo study.

In summary, as illustrated in Fig. 6H, Tris DBA was shown to effectively ameliorate the mouse ASLN model, which simulates a transformation to a high-grade pathological type of LN from a mild mesangial one as the result of acute onset or the acceleration of the renal condition; this transformation involves differential regulation of T cell functions, a MAPK (ERK, JNK)-mediated priming signal of the NLRP3 inflammasome, and differentially regulating the autophagy/NLRP3 inflammasome axis. Thus, Tris DBA may be considered a therapeutic candidate for the accelerated and deteriorated type of LN.

The authors have no financial conflicts of interest.