C-reactive protein promotes diabetic kidney disease via Smad3-mediated NLRP3 inflammasome activation

Abstract

Diabetic kidney disease (DKD) is the leading cause of end-stage kidney diseases resulting in enormous socio-economic burden. Accumulated evidence has indicated that C-reactive protein (CRP) exacerbates DKD by enhancing renal inflammation and fibrosis through TGF-β/Smad3 signaling. NLRP3 inflammasome is the key sensor contributing to renal inflammation. However, whether CRP enhances inflammation in DKD via NLRP3 inflammasome-related pathway remains unknown. In this study, we demonstrate that CRP promotes DKD via Smad3-mediated NLRP3 inflammasome activation as mice overexpressing human CRP gene exhibits accelerated renal inflammation in diabetic kidneys, which is associated with the activation of Smad3 and NLRP3 inflammasomes. In contrast, blockade of CPR signaling with a neutralizing anti-CD32 antibody attenuates CRP-induced activation of Smad3 and NLRP3 in vitro. Importantly, genetic deletion or pharmacological inhibition of Smad3 also mitigates CRP-induced activation of NLRP3 in diabetic kidneys or in high glucose-treated cells. Mechanistically, we reveal that Smad3 binds to the NLRP3 gene promoter, which is enhanced by CRP. Taken together, we conclude that CRP induces renal inflammation in DKD via a Smad3-NLRP3 inflammasome-dependent mechanism. Thus, targeting CRP or Smad3-NLRP3 pathways may be a new therapeutic potential for DKD.

Graphical abstract

Chen, Lan, and colleagues demonstrated C-reactive protein (CRP) induces renal inflammation in diabetic kidney disease (DKD) via a Smad3-NLRP3 inflammasome-dependent mechanism. Blockage of Smad3 or CRP receptor CD32 alleviates CRP-induced NLRP3 inflammasome activation in DKD. P-Smad3 can bind to the promoter area of NLRP3 gene to stimulate the transcription.

Introduction

Diabetic kidney disease (DKD) is one of the major complications of diabetes. There are globally 537 million diabetic patients in 2021 with an increase of 246 million predicted by 2045.¹ The global expenditure of diabetes grew from USD 232 billion to 966 billion in the past two decades.² About 30%–40% of diabetic patients develop DKD, where type 2 diabetes is the largest contributor. The number of new DKD patients has increased significantly from 1.4 to 2.4 million in the past three decades.³ DKD has caused an immense cost of global healthcare expenditure, with substantial burden on both society and patients.

We and others have demonstrated that renal inflammation and fibrosis are the key pathogenic features in DKD.^4,5 Studies have shown that renal inflammation and fibrosis in DKD are mediated by multiple mediators, where Smad3 plays a vital role. We previously found that activation of Smad3 exacerbates DKD, which is mediated by Smad3-dependent noncoding RNAs including miR-29b and Erbb4-IR.⁵ A recent study also showed that DKD increases the susceptibility of patients to acute kidney injury (AKI) induced by COVID-19 via a Smad3-dependent mechanism.⁶ Silencing Smad3-dependent LNA9884 attenuates diabetic kidney injury in the db/db mouse model via downregulation of MCP-1-induced renal inflammation.⁷ Furthermore, dipeptidyl peptidase-4 evoked by C-reactive protein (CRP) is a druggable target for DKD.⁸

Inflammasomes play a crucial role in the development of acute and chronic inflammation and metabolic dysfunction.^9,10 Pattern recognition receptors, such as intracellular nucleotide-binding oligomerization domain (NOD)-like receptors, can detect internal and external pathogen-associated molecular patterns and danger-associated molecular patterns to activate inflammasomes.¹¹ The NLRP3 (NOD-, LRR-, and pyrin domain-containing protein 3) inflammasome is one of most prominent inflammasomes associated with kidney diseases.¹⁰ Activation of the NLRP3 inflammasome leads to the release of inflammatory cytokines of IL-1β and IL-18, which contribute to the progression of various kidney diseases, including unilateral ureteral obstructive (UUO),^12,13 nephrocalcinosis-related chronic kidney disease (CKD),^14,15,16 DKD,¹⁷ lupus nephritis,¹⁸ IgA nephropathy,^19,20 hypertensive nephrosclerosis,²¹ hyperhomocysteinemia-induced glomerulopathy, and western diet nephropathy.^22,23 Renal inflammasome activation is linked to the development of DKD,¹⁷ which can be triggered by mitochondrial ROS and result in insulin resistance and obesity.^17,24,25 The knockout of the NLRP3 gene or caspase-1 gene in the streptozotocin-treated mice model could ameliorate DKD.¹⁷ In DKD, hyperuricemia is a potent inducer of NLRP3 expression.¹⁷ Oligomerization of NLRP3 can activate caspase-1 and induce expression of IL-1β and IL-18.²⁶ Absence of NLRP3 alleviates DKD,¹⁷ providing direct evidence for the pathological role of NLRP3 in the development of DKD.

CRP is primarily regarded as an acute-phase protein marker with a rapid response to inflammation.²⁷ Thus, CRP has been extensively used clinically as a marker for inflammation in kidney diseases.^28,29 CRP gene polymorphisms have been found to regulate its expression level in serum and is associated with 1-year mortality in peritoneal dialysis patients and DKD progression.^30,31,32 Intriguingly, CRP is associated in the age-related decline of glomerular filtration rate (GFR), even in non-diabetic population.³³ Emerging evidence indicates that CRP not only serves as an inflammatory biomarker but also plays a pathogenetic role in kidney diseases, such as AKI, DKD, and UUO nephropathy.^34,35,36 We previously revealed that overexpression of CRP in mice aggravates renal function in both acute and chronic kidney diseases via TGF-β/Smad3 signaling.^{27,34,35,36,37} In addition, our previous study revealed that overexpression of CRP in the kidneys of db/db mice leads to more significant kidney damage, as evidenced by PAS and Masson staining.³⁷ However, mechanisms whereby CRP enhances inflammation in DKD remain unclear.

Smad3 plays as a pivotal role in CRP-related renal diseases, especially in DKD, as a mediator in the TGF/β signaling pathway, which is involved in both renal fibrosis and inflammation.³⁸ CRP has been shown to exacerbate AKI by impairing the G1/S cell cycle and inhibiting the CDK2/cyclin E pathway through Smad3-dependent mechanisms.^27,39 Deletion of Smad3 has been found to protect against CRP-induced renal inflammation in UUO models.⁴⁰ Meanwhile, CRP exacerbates DKD through the CD32b-smad3-mTOR signaling pathway.³⁷ Deficiency of Smad3 in db/db mice inhibits renal inflammation and crosstalk with the NF-κB signaling pathway, leading to significantly reduced TNF-α, ICAM-1, MCP-1, and macrophage infiltration in Smad3 knockout db/db mice.⁵ These findings suggest that Smad3 may be a potential therapeutic target for alleviating CRP-induced activation of the NLRP3 inflammasome.

In this study, we tested a hypothesis that CRP may mediate renal inflammation in DKD via an Smad3-NLRP3-dependent mechanism. This was examined in mouse models of DKD induced in human CRP transgenic gene mice with or without deletion of Smad3.

Results

CRP was negatively associated with renal function in DKD patients

To investigate the relationship between CRP levels and kidney function in patients with DKD, we conducted an association analysis involving 2,257 DKD patients (1,159 males, 1,098 females, mean age: 66.34 ± 12.34 years, mean estimated GFR [eGFR]: 61.89 ± 18.15 mL/min/1.73 m², mean CRP level: 0.71 ± 1.07 mg/dL). Given that the normality tests indicated non-normal distributions for CRP and eGFR values (p < 0.05), we utilized the Spearman rank correlation coefficient for the analysis. The findings demonstrated a significant negative linear correlation between CRP levels and eGFR in DKD patients (Spearman’s ρ = −0.05, p = 0.01) (see Figure 1A).

Figure 1.

CRP-related renal function in DKD patients and NLRP3 genes was significantly differentiated in RNA sequencing between the CRPwt-db/db and CRPtg-db/db groups

Correlation analysis showed that CRP was negatively associated with renal function in DKD patients (A). Heatmap analysis indicated that there was a significant difference between the CRPwt-db/db (DB) and CRPtg-db/db (CDB) groups (B). KEGG enrichment analysis indicated that the NOD-like receptor signaling pathway (KO04621) had the largest number of significantly increased genes (five genes) in the CDB group compared with the DB group, and three of five genes were the upstream of the NLRP3 signaling pathways (C). Volcano analysis indicated that the NLRP3 gene is significantly increased in the CDB group compared with the DB group, with positive Log2 fold change (D).

NLRP3 expression in the diabetic kidney is significantly increased in CRPtg-db/db mice

To further clarify the role of CRP, we performed RNA sequencing analysis between kidney RNA expression of CRPtg-db/db mice and CRPwt-db/db mice. Differentially expressed genes in kidneys of CRPtg and CRPwt mice is shown in Figure 1B and Table S2. The top 15 pathways by the Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis revealed that NOD-like receptor signaling pathway has the largest number of upregulated genes in the classification of organismal systems, as shown in Figure 1C and Table S3. Among them, oligoadenylate synthetases (Oas1a), Oas2, and Oas3 are the upstream genes of the NLRP3-related signaling pathway.⁴¹ The Oas family proteins are a group of proteins involved in the innate immune response against viral infections. Studies have shown that Oas proteins can induce the activation of the NLRP3 inflammasome, a key component of the innate immune system that triggers inflammation in response to infection or cellular damage. Studies suggest that Oas stimulates the assembly of the NLRP3 complex with the DExD/H-box helicase DHX33 and that the mitochondrial adaptor protein MAVS is essential for the efficient activation of the NLRP3 inflammasome.⁴² Indeed, we found that NLRP3 was one of the differential expressed genes in CRPtg-db/db mice (Figure 1D). These results indicated that NLRP3 may participate in CRP-induced DKD.

CRP-enhanced diabetic kidney injury is associated with activation of NLRP3 inflammasome and Smad3 signaling in vivo and in vitro

Overexpression of CRP in db/db mice significantly increased the kidney injury, activation of the NLRP3 pathways (NLRP3, NEK7, and cleaved caspase-1), and inflammatory markers (IL-1β and TNF-α) in kidneys of CRPtg-db/db mice compared with CRPwt-db/db mice (Figure 2A). Immunofluorescence also showed that NLRP3 was highly expressed in both glomeruli and tubules in the kidneys of the CRPtg-db/db group compared with the CRPwt-db/db mice (Figure 2B). Consistently, the mRNA expressions of NLRP3, NEK7, IL-1β, and TNF-α were also upregulated in the CRPtg-db/db group compared with controls (Figure 2C).

Figure 2.

NLRP3 inflammasome activation and phosphorylation of Smad3 significantly increased in CRPtg-db/db mice

Western blot and quantitative analysis indicated that NLRP3, IL-1β, NEK7 and C-caspase-1 protein expression levels increased significantly in the CRPtg-db/db mice group (A). Expression level of NLRP3 was increased in the CRPtg-db/db mice group measured by immunofluorescence (B). Scale bars, 20 μm. qPCR and quantitative analysis exhibited NLRP3, NEK7, IL-1β, and TNF-α mRNA expression levels raised significantly in the CRPtg-db/db mice group (C). Phosphorylation of Smad3 was enhanced significantly in the CRPtg-db/db mice group (D). Each bar represents the mean SEM from groups of six mice. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 versus the indicated group.

In vitro, we found that the treatment of CRP resulted in the sustained activation of NLRP3 pathway from 6 to 12 h (Figures 3A and 3B), which was positively associated with levels of NLRP3 and related indexes, including IL-1β, NEK7, and cleaved caspase-1 in high glucose-treated human kidney epithelial 2 (HK-2) cells (Figure 3C) and mouse mesangial cells (Figure 4A). Real-time PCR analysis detected that mRNA level of NLRP3, NEK7, IL-1β, and TNF-α were elevated prominently in both cell lines under high glucose conditions (Figures 3D and 4B). Afterward, consistent results were also obtained from immunofluorescence that the treatment of CRP elevated the level of NLRP3 in high glucose-treated HK-2 cells and mesangial cells (Figures 3E and 4C). Significantly, we found that NLRP3 activation were associated with the phosphorylation of Smad3 in vivo and in vitro (Figures 2D, 3F, and 4D).

Figure 3.

CRP-induced NLRP3 inflammasome activation and phosphorylation of Smad3 in HK-2 tubular epithelial cells under high-glucose conditions

Western blot exhibited CRP-induced NLRP3 expression in HK-2 cells in a time-dependent manner, being significant as early as 6 h and peaking at 12 h (A). qPCR demonstrated that CRP induces NLRP3, IL-1β, and TNF-α mRNA expression level in a time-dependent manner, being significant at around 6 h (B). High-glucose (HG)+CRP conditions significantly enhanced NLRP3 inflammasome-related indexes of protein expression level, including NEK7, C-caspase-1, and IL-1β (C). qPCR showed that HG+CRP conditions significantly enhanced the mRNA expression level of NLRP3, NEK7, IL-1β, and TNF-α (D). Expression levels of NLRP3 and caspase-1 were increased in the HG+CRP group measured by immunofluorescence (E). Scale bars, 20 μm. Phosphorylation of Smad3 significantly increased in the HG+CRP group (F). Each bar represents the mean SEM from three independent experiments in vitro. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 versus the indicated group.

Figure 4.

CRP-induced NLRP3 inflammasome activation and phosphorylation of Smad3 in mouse mesangial cells under HG conditions in vitro

Western blot showed that HG+CRP condition significantly enhanced NLRP3 inflammasome-related indexes of protein expression level, including NEK7, C-caspase-1 and IL-1β (A). qPCR showed that HG+CRP conditions significantly enhanced the mRNA expression level of NLRP3, NEK7, IL-1β and TNF-α (B). Expression levels of NLRP3 and caspase-1 were downregulated in the HG+CRP group measured by immunofluorescence (C). Scale bars, 20 μm. Phosphorylation of Smad3 significantly increased in the HG+CRP group (D). Each bar represents the mean SEM from three independent experiments in vitro. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 versus the indicated group.

Blockage of CRP with the CD32 (Fcγ RII) antibody inhibits CRP-induced NLRP3 inflammasome and Smad3 signaling in vitro

Next, we found that blockage of CRP with a neutralizing antibody of CD32 significantly decreased the NLRP3 activation (NLRP3, NEK7, cleaved caspase-1, and IL-1β) in CRP-treated HK-2 cells and mouse mesangial cells (Figures 5A and 5E). Consistently, the expression levels of NLRP3, NEK7, IL-1β, and TNF-α were downregulated by the block of CRP in both cell lines (Figures 5B and 5F). The phosphorylation of Smad3 was also significantly attenuated by the neutralizing antibody (Figures 5D and 5H). Immunofluorescence assay also showed that blockage of CRP receptor inhibited the expression levels of NLRP3 and caspase-1 (Figures 5C and 5G). Our findings highly suggested that Smad3 and NLRP3 were highly associated. It was unknown if Smad3 had directly interaction with NLRP3.

Figure 5.

CRP-induced NLRP3 inflammasome activation was blocked by CRP receptor CD32 antibody

Western blot showed that the addition of CD32 antibody (CD32Ab) significantly inhibited NLRP3 inflammasome-related indexes of protein expression level, including NEK7, C-caspase-1, and IL-1β in HK-2 cells (A) and mouse mesangial cells (E). qPCR demonstrated that CRP-induced NLRP3, NEK7, IL-1β, and TNF-α mRNA expression levels in HK-2 cells were significantly inhibited with the addition of CD32Ab in HK-2 cells (B) and mouse mesangial cells (F). Expression levels of NLRP3 and caspase-1 were decreased with the addition of CD32Ab, measured by immunofluorescence in HK-2 cells (C) and mouse mesangial cells (G). Scale bars, 20 μm. Phosphorylation of Smad3 significantly decreased in the HG+CRP+CD32Ab group in HK-2 cells (D) and mouse mesangial cells (H). Each bar represents the mean SEM from three independent experiments in vitro. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 versus the indicated group.

Deletion of Smad3 inhibits CRP-induced NLRP3 inflammasome activation in db/db mice and in vitro under diabetic conditions

Our research group previously found no significant difference in phenotype between Smad3 WT-db/m and Smad3 KO-db/m mice.⁴³ However, in db/db mice, the knockout of Smad3 (Smad3 KO-db/db mice) offered protection against histological damage, as evidenced by reduced mesangial matrix expansion and glomerular basement membrane alterations.⁵ Herein, we tested our hypothesis that CRP may induce NLRP3 activation via a Smad3-dependent mechanism, which was examined in Smad3 KO-db/db, Smad3 WT-db/db, Smad3 KO-db/m, and Smad3 WT-db/m mice. Western blot, real-time PCR, and immunofluorescence found that deletion of Smad3 significantly downregulated NLRP3 and related inflammasome indexes, including IL-1β, TNF-α, NEK7, and cleaved caspase-1 in Smad3 KO-db/db mice compared with Smad3 WT-db/db mice (Figure 6).

Figure 6.

NLRP3 inflammasome activation and phosphorylation of Smad3 were significantly inhibited in Smad3 knockout mice

Western blot and quantitative analysis indicated that NLRP3, IL-1β, NEK7, and C-caspase-1 protein expression level decreased significantly in the Smad3 knockout (KO)-db/db mice group (A). Expression level of NLRP3 and decreased significantly in the Smad3 KO-db/db mice group measured by immunofluorescence (B). Scale bars, 20 μm. qPCR and quantitative analysis showed that NLRP3, NEK7, IL-1β, and TNF-α mRNA expression levels were downregulated in the Smad3 KO-db/db mice group (C). Phosphorylation of Smad3 significantly decreased in the Smad3 KO group (D). Each bar represents the mean SEM from groups of 6 mice. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 versus the indicated group.

In vitro, we also found that CRP was able to time-dependently induce phosphorylation of Smad3 in HK-2 cells (Figure 7A). Western blot, real-time PCR, and immunofluorescence assay also showed that blockage of Smad3 by a Smad3 inhibitor SIS3 significantly attenuated the expression levels of NLRP3 and its related inflammasome indexes including IL-1β, NEK7, and cleaved caspase-1 in both HK-2 cells and mouse mesangial cells (Figure 7). Overall, these results suggested that the exacerbating effect of CRP on DKD is Smad3 dependent.

Figure 7.

CRP-induced NLRP3 inflammasome activation was blocked by Smad3 inhibitor SIS3

Western blot showed that CRP induced phosphorylation of Smad3 in HK-2 cells in a time-dependent manner, being significant as early as 6 h and peaking at 12 h (A). Western blotting demonstrated that CRP-induced NLRP3 inflammasome activation was significantly inhibited with the addition of SIS3 under high-glucose conditions in HK-2 cells (B) and in mouse mesangial cells (F). qPCR demonstrated that the addition of SIS3 significantly inhibited CRP-induced NLRP3, NEK7, IL-1β, and TNF-α mRNA expression levels in HK-2 cells (D) and mouse mesangial cells (H). Expression levels of NLRP3 and caspase-1 were downregulated with the addition of SIS3 measured by immunofluorescence in HK-2 cells (C) and mesangial cells (G). Scale bars, 20 μm. Phosphorylation of Smad3 significantly decreased in the HG+CRP+SIS3 group in HK-2 cells (E) and mouse mesangial cells (I). Each bar represents the mean SEM from three independent experiments in vitro. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 versus the indicated group.

Smad3 promotes expression of NLRP3 by binding to the promoter of NLRP3

Next, we determined whether Smad3 could bind to the NLRP3 gene promoter region to regulate NLRP3 activation. Here, chromatin immunoprecipitation-PCR (ChIP-PCR) assay was applied. TGF-β1 was used to enrich the phosphorylation of Smad3, which served as a positive control. After pull-down by P-Smad3 antibody, the enrichment of NLRP3 was significantly shown, which suggested the binding relationship between P-Smad3 and NLRP3 promoter (Figure 8A). The ChIP-PCR assay consistently suggested that the P-Smad3 induced by CRP could bind to the promoter region of NLRP3 (Figure 8A).

Figure 8.

CRP-induced NLRP3 inflammasome activation was through the binding between phosphorylated Smad3 and the NLRP3 promoter area

Chromatin immunoprecipitation (ChIP) assay indicated that P-Smad3 could bind to the promoter area of the NLRP3 gene in HK-2 cells at two binding sites (A). Dual-luciferase reporter assay indicated that P-Smad3 blinded to the NLRP3 promoter area at the −1138 to −1129 sequence (B). Each bar represents the mean SEM from three independent experiments in vitro. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 versus the indicated group.

To further explore the specific location of the binding site, the dual-luciferase reporter assay was used. The dual-luciferase reporter assay demonstrated that P-Smad3 could bind to the NLRP3 promoter at the location of −1138 to −1129 at the promoter region of NLRP3 transcription but no other two predicted regions (Figure 8B).

Discussion

In this study, we found that CRP promotes DKD via Smad3-mediated NLRP3 inflammasome activation. This was supported by the findings that mice overexpressing the human CRP gene largely promoted activation of Smad3 and NLRP3 inflammasome signaling, resulting in accelerated renal inflammation in diabetic kidneys of db/db mice, which was blocked by addition of a neutralizing anti-CD32 antibody. Importantly, we also found that genetic deletion or pharmacological inhibition of Smad3 was able to block CRP-induced activation of NLRP3 in diabetic kidneys of db/db mice or in high-glucose and CRP-treated tubular cells and mesangial cells in vitro.

The CRP mechanism promotes DKD via Smad3-mediated NLRP3 inflammasome-dependent mechanism, as evidenced by the findings that, with the stimulation of Smad3 signaling pathway inducer TGF-β or CRP, after pull-down by P-Smad3 antibody, the enrichment of the NLRP3 promoter area gene was significantly enhanced. Dual-luciferase reporter assay indicated the specific binding location between P-Smad3 and NLRP3 promoter areas. This finding may well explain the promoter role of CRP in DKD or other inflammatory diseases.

Our study found that CRP induces activation of NLRP3 inflammasomes, providing a novel trigger for renal inflammation in DKD. As CRP is commonly used as a clinical marker for diabetic patients, these findings provide experimental evidence of the importance of controlling CRP in DKD.⁴⁴ In addition to kidney inflammation, CRP expression levels have been associated with other diseases, including cardiovascular disease,^45,46 rheumatoid arthritis,⁴⁷ atrial fibrillation, and metabolic syndrome.^48,49,50 It is known that CRP enhances atherosclerosis by upregulating VEGF expression through activation of hypoxia-inducible factor-1α,⁵¹ while the mechanisms for its involvement in other diseases remain unclear. Therefore, reducing serum CRP levels may be a target for inflammation treatment.

We demonstrated that knockout of Smad3 alleviates renal inflammation in CRP-induced DKD through NLRP3 signaling. Importantly, Smad3, a downstream mediator of the TGF-β signaling pathway, plays critical roles in renal fibrosis. Previous studies have showed that deletion of Smad3 or inhibition significantly reduce renal fibrosis in animal models of DKD,⁴³ UUO,^40,52,53 lipopolysaccharide-induced,⁵⁴ and ischemia reperfusion-induced AKI.^27,55 Targeting on Smad3 by chemical compounds, miRNAs, and lncRNAs may have a therapeutic potential.^38,53,56,57 Hereby, we illustrate that CRP directly induced the binding between P-Smad3 and NLRP3 promoter genes. We previously demonstrated that CRP promotes TGF-β/Smad3-mediated renal fibrosis.⁴⁰ Therefore, blockage of CRP can synergistically reduce diabetic kidney injury.

Several approaches had been applied to lower serum CRP levels, including antisense oligonucleotides (ASO),⁵⁸ selective CRP apheresis, and small-molecule CRP inhibitors.^59,60,61 However, the use of ASO treatment is controversial as one type of ASO has been found to increase IL-6 and CRP levels in healthy volunteers.⁶² In addition, since CRP is involved in host defense, simply inhibiting serum CRP levels may lead to immune system dysfunction.⁴⁸ Previous studies have demonstrated that lowering serum CRP levels through selective apheresis can be used as an anti-inflammatory therapy, and there are currently several ongoing clinical trials in this area.⁴⁸ Therefore, hemodialysis may be a more suitable and controllable method for reducing serum CRP levels.

In conclusion, although multiple canonical and non-canonical inflammasome pathways were identified in the previous studies, our data demonstrate a direct signaling mechanism that CRP would activate NLRP3 inflammasome in DKD. Our finding provides a potential therapeutic target for clinical treatment of diabetic patients. Blockage of CRP and inhibition of the Smad3 signaling pathway may be a useful alternative therapeutic target to suppress renal inflammation in DKD.

Materials and methods

Correlation analysis of CRP and renal function in DKD patients

The study utilized data from the National Health and Nutrition Examination Survey (NHANES) spanning 1999 to 2010. NHANES is a cross-sectional survey series employing a complex, multistage probabilistic design to obtain a representative sample of civilian, noninstitutionalized populations across all 50 states and the District of Columbia, thus encompassing the entire US population.⁶³ The Centers for Disease Control and Prevention/National Center for Health Statistics Ethics Review Board reviewed and approved all ethical aspects of the study, and each participant provided written informed consent. Given the public and definitive nature of the data, the analysis is exempt from institutional review board approval. Participants included in the analysis had laboratory-tested blood results for CRP and hemoglobin A1c (HbA1c), along with urine results for eGFR. Diagnostic criteria for DKD were defined as HbA1c levels ≥6.5% and eGFR < 90 mL/min/1.73 m². The Pearson correlation coefficient was utilized to analyze the relationship between CRP and eGFR for data adhering to normal distribution. For data not meeting normality assumptions, the Spearman rank correlation coefficient was applied.

Cell culture

HK-2 cells and mouse mesangial cells were cultured in Dulbecco’s modified Eagle’s medium F12 (DMEM F12) (cat. no. 11320033, Invitrogen) supplemented with 10% fetal bovine serum (cat. no. 10270106, Invitrogen) and 1% penicillin-streptomycin (cat. no. 15140122, Invitrogen) as described previously.^39,64 For high-glucose conditions, the glucose concentration was adjusted to 5.5 mM by adding glucose solution (cat. no. A2494001, Invitrogen) to DMEM low-glucose medium (cat. no. 11885084, Invitrogen), supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. As control, DMEM low glucose medium, supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin was applied for low-glucose conditions. D-Mannitol (35 mmol/L) was applied as a control for osmolality.⁶⁵ Human CRP (cat. no. 1707-CR-200/CF, R&D Systems), human Fc gamma RII/CD32 antibody (cat. no. AF1330, R&D Systems), and SIS3 (cat. no. 566405, Sigma-Aldrich) were applied for cell stimulation.

Animals

Male db/db mice overexpressing human CRP (CRPtg-db/db) and littermates (CRPtg-db/m, WT-db/m, and WT-db/db) were used in this study as generated previously.³⁷ The phenotype of CRP transgenic mice was consistent as reported previously.³⁷ Groups of six male mice were harvested at the age of 36 weeks for further experiments. To investigate the regulatory role of Smad3 in the development of DKD, male db/db mice with Smad3 knockout (Smad3 KO, db/db) and control littermates (Smad3 WT db/m; Smad3 KO db/m; Smad3 WT db/db) were used in this study as described previously.⁵ Phenotype of Smad3 knockout mice were consistent as our study previously reported.⁵ Groups of six male mice were harvested at the age of 32 weeks for further experiments. All experimental methods were approval by the Animal Experimentation Ethics Committee, The Chinese University of Hong Kong. We used the ARRIVE1 reporting guidelines for animal experiments.⁶⁶

RNA extraction and quantitative real-time PCR

Total RNA samples from whole-kidney samples or cultured cells were isolated by TRIzol reagent (cat. no. 15596026, Invitrogen). Template cDNA was prepared using an RNA reverse transcription kit (TAKARA). Expression levels of NLRP3, IL-1β, and TNF-α were measured with SYBR green (cat. no. 1725122, Bio-Rad). Primer sequences are shown in Table S1. β-Actin was applied for internal control. The expression levels of specific genes were calculated by the ratio to β-actin.

Western blot analysis

The western blot analysis was conducted as described previously.⁶⁷ In brief, the total proteins from mouse kidneys or cell samples were extracted by RIPA lysis buffer (cat. no. 89900, Thermo Scientific). Concentrations of protein were measured using the BCA protein assay kit (cat. no. 23225, Thermo Scientific). The cell lysates were boiled for 5 min at 99°C. Equal volumes of protein were separated by PAGE gel (cat. no. PG112, EpiZyme Biotechnology) at 10% concentration. Afterward, separated proteins were transferred to 0.45 μm polyvinylidene difluoride membranes at 4°C (cat. no. IPVH00010, Immobilon) and blocked by blocking buffer (5% fetal bovine serum in Tris-buffered saline with 0.1% Tween 20) overnight at 4°C. Primary antibodies including NLRP3 (cat. no. WL02635, Wanlei), P-Smad3 (cat. no. 9520S, Cell Signaling Technology), Smad3 (cat. no. 9513S, Cell Signaling Technology), NEK7 (cat. no. ab133514, abcam), caspase-1 (cat. no. sc-56036, Santa Cruz), IL-1β (cat. no. ab9722, abcam), and β-actin (cat. no. 69879, Santa Cruz) were incubated at 4°C overnight. Horseradish peroxidase-conjugated anti-rabbit (cat. no. 7074S, Cell Signaling Technology) and anti-mouse (cat. no. 7076S, Cell Signaling Technology) secondary antibodies were incubated for 1 h at room temperature. The signal was detected with western blotting detection reagent (cat. no. RPN2235, Amersham) using an imaging system (ChemiDoc XRS+ System, Bio-Rad). β-Actin was applied for internal control. The expression levels of specific genes were calculated by the ratio to β-actin.

Immunofluorescence analysis of kidney samples

Immunofluorescence was performed on 5-μm paraffin tissue sections. The mouse kidney tissues were formalin fixed and paraffin embedded. After paraffin removal and blocking, a microwave-based antigen retrieval process was applied. Tissue sections were immersed in citric acid antigen retrieval buffer (cat. no. G1202, Servicebio) in a microwave oven, with medium heat, for 8 min until boiling, with cessation of heat for 8 min, and then to medium-low heat for 7 min. After natural cooling, sections were washed in phosphate-buffered saline (pH 7.4) 3 times for 5 min each. Then, sections were incubated with primary antibody, including NLRP3 and caspase-1 overnight at 4°C. Afterward, sections were incubated with Alexa-Fluor 488 (cat. no. A11001, Invitrogen) and Alexa-Fluor 594 (cat. no. A11012, Invitrogen) secondary antibody for 1.5 h at room temperature. Nuclei were stained with DAPI (cat. no. C1006, Beyotime) for 10 min. Sections were mounted with mountain medium (cat. no. S3023, Dako). Inverted laser scanning confocal microscopy (LSM980, Carl Zeiss) was applied for image capturing (Imaging and Flow Cytometry Core, The University of Hong Kong). ZEN image analysis software (ZEN 3.0) was applied for image processing.

Immunofluorescence analysis of cell samples

Cells were seeded on glass slides and cultured to 70% density, followed by fixation with 4% paraformaldehyde for 5 min and blocked with 10% goat serum for 1 h. Cells were incubated with primary antibodies including NLRP3 and caspase-1 overnight at 4°C, followed by incubation with secondary antibodies, Alexa-Fluor 488 and Alexa-Fluor 594 for 1.5 h at room temperature. Cell nuclei were stained with DAPI for 10 min. Sections were mounted with mountain medium and observed using an inverted laser scanning confocal microscope (LSM980, Carl Zeiss). ZEN image analysis software (ZEN 3.0) was employed for image process.

ChIP-PCR assay

ChIP assay was performed as described previously.⁵³ HK-2 cells were stimulated with 10 μg/mL CRP for 12 h before being harvested. HK-2 cells stimulated with TGF-β1 (cat. no. 246-LP, R&D Systems) for 12 h before harvest were used as the positive control group. ChIP assay was conducted using a Chromatin IP Kit (cat. no. 9005, Cell Signaling Technology) according to the manufacturer’s protocol. P-Smad3 antibody and normal IgG antibody were used to immunoprecipitate crosslinked protein-DNA complexes. Purified immunoprecipitated DNA samples were applied for subsequent PCR analysis. Two pairs of primers were designed specific according to predicted binding sites located on the promoter area of the NLRP3 gene. The primers were designed using the Basic Local Alignment Search Tool (BLAST) of NCBI according to the promoter sequence of NLRP3 gene. The sequence of primer 1 was 5′-AGC CCC GAA TGA CAC CTT TC-3′ and 5′-ACA GCA GAC CTG AGT CTC CA-3ʹ. Primer 2 was 5′- TTT CAC TCA CCC AGA GGC TG-3′ and 5′- CAC TCA CGG AAT GGC TGG TA-3ʹ.

Cell transfection and dual-luciferase reporter assay

Luciferase reporter plasmids were then constructed according to ChIP assay results, containing either the wild-type sequence or one of three mutant sequences, at the top 3 selected binding sites. The binding sites for Smad3 on the NLRP3 promoter region (2 kb) were predicted using the transcription factor affinity prediction (TRAP) web tools (Computational Molecular Biology Department at the Max Planck Institute for Molecular Genetics in Berlin, Germany). A total of six binding sites on the NLRP3 promoter region were predicted, and the top 3 binding sites were selected based on their TRAP weight score. To increase the level of P-Smad3, in addition to culture with TGF-β1, we constructed P-Smad3 overexpression plasmid according to the phosphorylation site (Ser 423/425) of Smad3 antibody we used in ChIP assay. The transfection process was conducted as previously reported,⁶⁸ using Lipofectamine 3000 (cat. no. L3000001, Invitrogen) according to the manufacturer’s protocol. HEK293T cells were transfected with P-Smad3 plasmid and Renilla plasmid first, and then transfected plasmids of NLRP3 wild-type or three NLRP3 mutations (Nanjing Genebay). Afterward, the cells were stimulated by TGF-β1 for 12 h before being harvested. A dual-luciferase reporter assay kit (cat. no. DL101, Vazyme) was applied for measurement of the luciferase density.

mRNA library constructing and sequencing

The mRNA sequencing was conducted at GENEWIZ (Shanghai, China). Isolation of total RNA was described previously. A total of 1 μg of RNA was utilized for library preparation. Library construction and sequencing were conducted according to the manufacturer’s standard process. The KEGG enrichment figure was created by chiplot (https://www.chiplot.online/).

Statistical analysis

All experiments were repeated at least 3 times. Data from studies are expressed as means ± standard error of the mean (SEM). Statistical analyses were performed by one-way ANOVA analysis of variance followed by Newman-Keuls multiple comparisons test using Prism 6.0 (GraphPad Software).

Data and code availability

The data of RNA sequencing supporting the findings of this study are openly available in repository at Gene Expression Omnibus repository under accession number GSE271870.

Acknowledgments

This study was supported by grants from the Hong Kong Research Grants Council (17109019, 17125323, 17125524, and 17113416), the Shenzhen Science and Innovation Fund (JCYJ20210324114604013), and the HKU Seed Funds (202011159210, 202111159235, and 109000219). We extend our profound gratitude to the investigators of the original NHANES study and the USA National Center for Health Statistics for generously sharing the data. The graphical abstract was created with BioRender.com.

Author contributions

Y.W., Y.Y.-K., and J.G. contributed equally to this work. H.-Y.C. and H.-Y.L. conceived, designed, and supervised the study. Y.W. and Y.Y.-K. performed most of the experiments. J.G. conducted part of data analysis. H.L., J.W., and B.S. performed part of cell studies. Y.W. and H.-Y.C. wrote the manuscript. H.-Y.L. and X.M. revised the manuscript. All authors read and approved the final version of the manuscript.

Declaration of interests

The authors declare no competing interests.