Introduction

The incidence of diabetes is increasing year by year. Diabetic nephropathy (DN) is the most common secondary kidney disease of diabetes, which is mainly manifested as persistent protein uria with elevated blood pressure, and decreased glomerular filtration rate. Additionally, DN increases the risk of cardiovascular events and mortality1. Many studies confirmed that immune and inflammatory responses play an essential role in the occurrence and development of DN2. Elevated glucose caused the abnormal function of the mitochondrial and endoplasmic reticulum to increase the production of reactive oxygen species (ROS) and activate the signaling pathways associated with cellular injury and stress. Under cellular stress, renal innate cells develop a pro-inflammatory response, which promotes innate immune response and recruitment of macrophages by releasing chemokines, cell adhesion molecules, and damage-associated molecules. It can promote macrophage infiltration and induce proliferation and differentiation of fibroblasts. This process accumulates excesses of extracellular matrix to damage the structure and function of the kidney, which promotes DN progression3,4. As DN progresses, the infiltration of neutrophils and macrophages in the kidney is intensified5. These immune responses promote kidney injury and decreased kidney function in DN6. Therefore, immune and inflammatory responses are involved in the occurrence and development of DN.

Systemic immune inflammation index (SII) is a blood biochemical index that reflects the local immune response and systemic inflammatory response7. SII was first defined as platelet count × neutrophil count/lymphocyte count in 20148. SII is an integrated parameter that combines peripheral platelets, neutrophils, and lymphocytes. In contrast to a single immunological or inflammatory index, SII more comprehensively evaluates the inflammatory state of the body. Studies have demonstrated that SII was applied to many diseases, including infections, tumors, cardiovascular diseases, etc.9. It has been reported that SII has better diagnostic performance than neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) for type 2 diabetes mellitus (T2DM)8. Increased SII level was independently associated with increased urinary albumin. SII has been linked to the prognosis of diseases affected by the virus. Patients with higher levels of SII were more likely to develop serious illness. SII can provide a basis for clinical prognosis stratification in patients with nasopharyngeal carcinoma. Therefore, SII can be used as one of the auxiliary indicators to assess the severity of DN progression and predict its prognosis10. This study aimed to evaluate the association between DN and SII using the comprehensive data from The National Health and Nutrition Examination Survey (NHANES). Furthermore, the molecular mechanism of SII in DN has been explored through bio-informatics analysis.

Method

Epidemiological analysis based on the NHANES database

Data and sample sources

Data were downloaded from the NHANES, a nationally representative cross-sectional survey designed and conducted by the National Center for Health Statistics (NCHS). As described before11, the NCHS Research Ethics Review Board authorized the survey, verifying that all participants provided informed consent. Detailed statistics are accessible at https://www.cdc.gov/nchs/nhanes/.

To evaluate the participants’ nutritional and physical health, standardized in-home interviews, physical examinations, and laboratory tests were carried out at mobile examination centers. A total of 29,400 participants were involved in four NHANES cycles from 2013 to 2018. We excluded 10,243 participants under the age of 18 years, 491 with missing SII, 1689 without diabetes, 8620 without CDK, and 120 with pregnancy. Eventually, 8236 participants were enrolled in the study, including 7591 healthy controls and 645 patients with DN.

Exposure variable

Lymphocyte, neutrophil, and platelet counts (expressed as × 103 cells/μl) were measured using automated hematology analysis devices. The following formula is utilized to calculate SII12:

$$ S\,II = \frac{Platelet\,count \times neutrophil\,count}{{lymphocyte\,count}} $$

Outcome variable

Diabetes was defined as (1) a previously reported diagnosis by medical professionals, (2) fasting plasma glucose ≥ 7.0 mmol/L, (3) glycosylated hemoglobin (HbA1c) ≥ 6.5 mmol/L, or (4) taking diabetes drugs. The urine albumin/creatinine ratio was used to compute the ACR. The eGFR scores were calculated using the Chronic Kidney Disease (CDK) Epidemiology Collaboration algorithm. ACR ≥ 30 mg/g and/or eGFR < 60 mL/min/1.73m2 were used to diagnose diabetic nephropathy (DN) in diabetes patients. The following formula is utilized to calculate eGFR8,13:

$$ {\text{eGFR }}\left( {{\text{male}}} \right){ = }\left( {\text{140 - age}} \right){\text{weight }}\left( {{\text{kg}}} \right){1}{\text{.23/serum creatinine }}\left( {\text{umol/L}} \right) $$
$$ {\text{eGFR }}\left( {{\text{female}}} \right){ = }\left( {\text{140 - age}} \right){\text{weight }}\left( {{\text{kg}}} \right){1}{\text{.03/serum creatinine }}\left( {\text{umol/L}} \right) $$

Mechanism exploration based on the GEO database

Data acquisition and processing

There are three microarray data associated with DN, including GSE30528, GSE30529, and GSE96804 datasets. GSE30528 and GSE30529 datasets are based on the platform of GPL571 ( [HG-U133A_2] Affymetrix Human Genome U133A 2.0 Array). GSE96804 dataset is based on the platform GPL17586 ( [HTA-2_0] Affymetrix Human Transcriptome Array 2.0 [transcript (gene) version]). The GSE30528 dataset included 13 control samples and 9 DN samples, the GSE30529 dataset included 12 healthy human kidney samples and 10 diabetic human kidney samples, and the GSE96804 dataset included 20 healthy human glomeruli samples and 41 DN patient glomeruli samples. GSE96804 dataset acted as the training set. A merged dataset consisting of GSE30528 and GSE30529 acted as a validating set to validate the expressions of hub genes. The “combat” package in R software and the empirical Bayes method14 were used to merge, probe-annotate, normalize, and batch-correct the data of the merged dataset. The “limma” R software packages were used to probe-annotate and normalize the data from the GSE96804 dataset.

Identification of differentially expressed genes

After the data standardization and normalization of GSE96804, the DEGs between DN and control groups were analyzed by using the “limma” R package. The DEGs were screened with the criteria of log|Fold Change|> 1 and p < 0.05. R software ggplot2 package and “ComplexHeatmap” depicted a volcano plot and a heat plot to show significantly changed genes, respectively.

Identification of co-expression genes

Weighted Gene Co-expression Network Analysis (WGCNA) is an algorithm to cluster genes into different modules and uncover the relationships between modules and disease traits15. To comprehensively investigate the genetic mechanisms involved in the pathogenesis of DN, a co-expression network was constructed by the “WGCNA” package in R. The co-expression network was constructed only using the genes with the top 25% variance from the GSE96804 dataset. The dynamic cutting tree method was adopted to merge modules with a threshold of 0.25. Other criteria were used to construct the co-expression network: soft threshold power (β) based on the scale-free topology criterion (an independence index of R2 = 0.85) by using the pick Soft Threshold function; minimum genes of each module = 30. Pearson correlation analysis was adopted to uncover the potential correlations between modules and DN.

Acquisition of overlapped genes, expression analysis, and KEGG analysis

Based on the significant module genes and DEGs, Venn analysis obtained 30 overlapped genes. We then analyzed the expression differences of the 30 genes between NC and DN groups. The overlapping genes were included in KEGG enrichment analysis using the “clusterProfiler” package.

Immune cell infiltration analysis

The R software’s “CIBERSORT” package16 was used to assess the level of immune cell infiltration based on 22 immune cell types of the “LM22” document (https://cibersort.stanford.edu/index.php), The results were filtered using the screening criteria: P value < 0.05. Depending on the results obtained by the immuno-infiltration assays, the differential expression levels of 22 immune infiltrating cell types in the DN and control were visualized using violin plots. The correlation analysis of 22 infiltrating immune cell types was visualized by the R software’s “corrplot” package. The degree of association between the 22 immune cell types and potential biomarkers was evaluated by “Spearman” correlation and visualized using R software’s “ggplot2” package.

Cluster analysis of overlapped genes

We have identified 30 overlapped genes associated with DN. Then, use the R package “Consensus Cluster Plus” to perform cluster analysis of 30 genes. Determine the optimal number of clusters K = 2 through the cumulative distribution curve graph. Meanwhile, the Principal Component Analysis (PCA) was used to detect the samples between the modules whether could be distinguished by these genes’ expression levels. In addition, the expression levels of 30 genes in two clusters (C1 and C2) were determined.

Identification of hub genes

A fourfold table was used to appraise the test results of cluster Analysis. This analysis can predict whether the overlapping genes are protective and harmful to DN. Next, we performed the KEGG analysis of the harmful or protective genes. Based on the random forest algorithms, the 30 genes were ranked for importance to select the top 10 important genes in the training set. Differences in the expression of 30 genes were detected using a non-parametric test in the validation set.

Prediction of hub genes-drug regulatory modules

CyTargetlinker in the Cytoscape app (https://projects.bigcat.unimaas.nl/cytargetlinker/) was utilized to predict gene-drug interactions, which were presented in a graphical manner using an extension of the network. A regulatory interaction network is a network containing regulatory interactions, often derived from online interaction databases. We drew drug-target interactions links from the database (DrugBank). The networks were visualized and analyzed using Cytoscape v3.7.1.

Molecular docking

The Protein Data Bank (PDB) (http://www.rcsb.org) and Pub Chem database (https://pubchem.ncbi.nlm.nih.gov) were searched to obtain the structural data of the target proteins and drug components17,18,19. The protein molecules were then converted to *. pdbqt format and docked using Auto Dock1.2.0 software. Molecular docking was performed by AutoDockTools, and the folding patterns and molecular interactions between target proteins and ligand molecules were examined to derive binding energies for screening.

Statistical analysis

Data analyses were performed by R studio software (version 4.2.2). Mann–Whitney-U (continuous variables) or Chi-square test (categorical variables) were utilized to compare differences in baseline characteristics between the healthy controls and DN groups. Considering the skewed distribution of SII, Log2-transformation was conducted to facilitate interpretation. The non-linear association between logSII and DN was investigated using the restricted cubic spline (RCS) analysis. Then, three generalized linear models were used to further evaluate the correlation between SII and DN. In Model 1, no adjustment. In Model 2, albumin creatinine ratio, blood urea nitrogen, creatinine, glucose, triglyceride, uric acid, WBC, red blood cell count, red cell distribution width, and EGFR were adjusted. In Model 3, albumin creatinine ratio, blood urea nitrogen, creatinine, glucose, triglyceride, uric acid, WBC, red blood cell count, red cell distribution width, EGFR, gender, race, hypertension, poverty income ratio (PIR), body mass index (BMI), white blood cell (WBC), physical activity (PA), smoking, age, PIR, BMI, and drinking were adjusted. The supposed population attributable fraction (PAF) for the DN risk was calculated. P < 0.05 was regarded as statistically significant.

Ethics approval

The Ethics Committee of General Hospital of Ningxia Medical University deemed that this research is based on open-source data, so the need for ethics approval was waived.

Results

Baseline characteristics of participants

A total of 8236 subjects were involved, with 4042 male (49.077%) and 4194 female (50.923%); 645 (7.831%) participants were categorized as having DKD in our study. Gender, race, hypertension, PA, smoking, age, albumin creatinine ratio, PIR, BMI, drinking, blood urea nitrogen, serum creatinine, glucose, triglycerides, uric acid, WBC, Log SII, red blood cell count, red cell distribution width, EGFR were significantly different between the two groups (all p < 0.05). Cholesterol did not differ between DKD patients and healthy controls. The clinical and biochemical characteristics of the participants with DKD and healthy controls are shown in Table 1.

Table 1 Basic characteristics of participants (n = 8236) in the NHANES 2013–2018.
Full size table

The association between Log SII and DN

To explore the relationship between Log SII and DN, we first conducted an RCS analysis to reveal their nonlinear association (P for nonlinear < 0.05) (Fig. 1). Then, three generalized linear regression models were used to further investigate their correlation.

Figure 1
figure 1

RCS analysis for the association between DN and logSII.

Full size image

In crude model 1, logSII was significantly correlated to DN (P < 0.05). When albumin creatinine ratio, blood urea, nitrogen, creatinine, glucose, triglyceride, uric acid, WBC, red blood cell count, red cell distribution width, and EGFR were adjusted in model 2, this significant relation still existed (P < 0.05). When all the covariates were adjusted in model 3, logSII was an independent predictor for DN (P < 0.05). According to the quartile, the subjects were divided into 4 groups. The correlation increased significantly with increasing logSII quartile (Table 2). The P for trend test confirmed the positive relationship between logSII and DN. PAF analysis showed that 23.41% (16.18%-30.64%) of DN cases could be attributed to the increase in LogSII level (Table 3). These results suggested that immune and inflammation might play an important role in the development of DN.

Table 2 Association between SII and DN.
Full size table
Table 3 PAF analysis of logSII and DN.
Full size table

Identification of DEGs

The vital relationship between SII and DN triggers us to further investigate the underlying mechanism of DN. Firstly, differential gene expression analysis was performed using the GSE96804 dataset. A total of 367 DEGs (|logFC|> 1 and P value < 0.05) were obtained, of which 259 DEGs were downregulated and 108 were upregulated. Heatmap and volcano plot of the DEGs are shown in Fig. 2A and B; the DEGs exhibited significantly different expression patterns between the DN and NC groups (p < 0.05).

Figure 2
figure 2

DEGs analysis in the GSE96804 dataset. (A) The heatmap showed the DEGs in the DN and HC groups. (B) Red and blue represent DEGs with significantly higher and lower expression levels in DN groups, respectively. DN, Diabetic Nephropathy; HC, healthy control.

Full size image

Identification of hub modules and DN-related genes

WGCNA was employed to identify the essential modules in DN using the GSE96804 dataset. First, a scale-free topology of the network was generated by setting the soft-thresholding power to 9 (Fig. 3A). Next, we plotted the heatmap of the module-group relationships to evaluate the association of each module with DN (Fig. 3B). The green module with 516 genes that had the strongest correlation with the DN was selected as the hub module (cor = 0.81, P < 0.001). Venn analysis showed 30 overlapped genes between the DEGs and green module genes (Fig. 3C). The differential expression levels of the 30 DEGs between the healthy and DN groups are shown in Fig. 3D. The expression of ANKRD36, ANKRD3B, GOLGA8A, GOLGA8B, LRP2BP, NAIP, NPIPB11, NPIPB3, NPIPB5, PCDHGA10, PER3, SPDYE1/2/3/5/6, TAS2R30, and TAS2R31 was significantly upregulated in the DN group (P < 0.05). The expression of BTG2, CLEC4E, DUSP1, EGR1, FOS, GDF15, JUN, JUNB, NR4A1, PLTP, SST, and ZFP36 was significantly downregulated in the DN group (P < 0.05).

Figure 3
figure 3

Identification of hub modules and genes. (A) β = 9 is chosen as the soft threshold based on the scale independence and average connectivity. (B) A Heatmap of the correlation between module genes and DN shows that the blue module has the highest association with DN. For each pair, the top left triangle is colored to represent the correlation coefficient; the bottom right one is colored to indicate the p-value. (C) Venn analysis of 30 overlapped genes associated with DN. (D) The expression of 30 overlapped genes in the GSE96804 dataset.

Full size image

DN-related genes might regulate the immune-inflammatory environment

A KEGG enrichment analysis was performed to determine which pathways the 30 genes were involved in (Fig. 4A). The 30 genes were enriched in the immune- and inflammatory-related pathways, such as TNF signaling pathway, MAPK signaling pathway, PD-L1 expression and PD-1 checkpoint pathway in cancer, Th1 and Th2 cell differentiation, IL-17 signaling pathway, Toll-like receptor signaling pathway, T cell receptor signaling pathway, Th17 cell differentiation. To more accurately explore the correlation between 30 genes and immune infiltration, we used the “CIBERSORT” package to analyze the correlation between 30 gene expressions and 22 immune cells. The results showed that the expression levels of immune cells such as CD8 + T cells, macrophages M1, and macrophages M2 in the DN group were higher than those in the HC group (Fig. 4B). The correlation between the expression of 30 genes and the level of 22 immune cells is shown in Fig. 5. These results showed that the 30 DN-related genes might affect DN progression by regulating the immune-inflammatory environment, which was consistent with our clinical results.

Figure 4
figure 4

KEGG and immune infiltration analysis of 30 DEGs. (A) The KEGG pathway analysis of 30 genes. (B) Immune infiltration analysis of 30 DEGs.

Full size image
Figure 5
figure 5

The correlation between the expression of 30 genes and the level of 22 immune cells.

Full size image

DN-related gene subtypes based on consensus clustering analysis

Based on the expression levels of the 30 genes, a total of 61 samples were categorized into distinct subtypes related to DN using consensus cluster analysis. The clustering heatmap and consistent cumulative distribution function (CDF) graph revealed that the CDF slope was minimized and the clustering effect was optimized when the number of clusters k equaled 2 (Fig. 6A, B). Furthermore, PCA analysis demonstrated that the clustering method effectively differentiated the samples into two subtypes (Fig. 6C). The expressions for the 30 DEGs between the C1 and C2 are shown in Fig. 6D. We found that the expression of 18 DEGs, including ANKRD36, ANKRD3B, GOLGA8A, GOLGA8B, LRP2BP, NAIP, NPIPB11, NPIPB3, NPIPB5, PCDHGA10, PER3, SPDYE1/2/3/5/6, TAS2R30, and TAS2R31, was significantly upregulated in C1 (P < 0.05) and the expression of 12 DEGs, including BTG2, CLEC4E, DUSP1, EGR1, FOS, GDF15, JUN, JUNB, NR4A1, PLTP, SST, and ZFP36 was significantly upregulated in C2 (P < 0.05).

Figure 6
figure 6

Consensus clustering analysis for 30 genes. (A) 61 DN patients in the GSE96804 dataset were grouped into two clusters according to the consensus clustering matrix (k = 2). (B) Consensus among clusters for each category number k. (C) PCA depicted the distribution for clusters. (D) The expression of 30 genes in the C1 and C2.

Full size image

NOD-like receptor signaling pathway mediated the expression of harmful and protective genes

We have redistributed the 61 samples to 2 clusters (C1 and C2). Based on the results of the fourfold table diagnostic test (Table 4), we find that 35 samples in C1 are classified as DN group and none as HC group. Six samples from C2 were classified as DN group and 20 samples as HC group. Thus, genes highly expressed in C2 are protective genes, while genes highly expressed in C1 are harmful genes. Eighteen genes are highly expressed in C1, including ANKRD36, ANKRD36B, GOLGA8A, GOLGA8B, LRP2BP, NAIP, NPIPB11, NPIPB3, NPIPB5, PCDHGA10, PER3, SPDYE1, SPDYE2. SPDYE3, SPDYE5, SPDYE6, TAS2R30, TAS2R31. Twelve genes are highly expressed in C2, including BTG2, CLEC4E, JUN, NR4A1, PLTP, SST, ZFP36, DUSP1, EGR1, FOS, GDF15, and JUNB. We performed a KEGG pathway analysis on these genes and found that there are 3 common pathways between damaged and protected genes (Fig. 7A-C). The three pathways were circadian entrainment, NOD-like receptor signaling pathway, and pathogenic Escherichia coli infection. In particular, the NOD-like receptor signaling pathway is immune-related. Taken together, NOD-like receptor signaling pathway may regulate the immune inflammatory environment in DN by regulating the expression of the harmful and protective genes.

Table 4 Fourfold Table Diagnostic Test for gene expression to Predict risk for Patients With DN.
Full size table
Figure 7
figure 7

KEGG pathway analysis of harm genes and protective genes. (A) Venn analysis of 3 overlapped pathways. KEGG pathway analysis of harm genes (B) and protective genes (C).

Full size image

Identification and validation of the hub genes

We ranked the 30 genes by impact factor importance and the top ten genes from highest to lowest were DUSP1, FOS, ZFP36, SPDYE3, NAIP, NPIPB3, LRP2BP, PER3, PLTP, GOLGA8A (Fig. 8A). The GSE137902 and GSE137301 datasets were merged to become a validation set. Twenty of the 30 genes are expressed in the merged dataset (Fig. 8B). The important genes with significant differences in expression between the two groups are DUSP1, NAIP, and ZFP36 (Fig. 8C). Therefore, DUSP1, NAIP, and ZFP36 are important genes in the development of DN.

Figure 8
figure 8

Identification of the important hub genes. (A) Ranking of the importance of impact factors of 30 genes. (B) The merged dataset with the GSE137902 and GSE137310 datasets. (C) The expression of 20 genes in the merged dataset.

Full size image

Prediction of drugs for three hub genes

Based on the CyTargetlinker plug-in in Cytoscape software, we constructed a drug-genes network. The results showed that resveratrol targets three genes (Fig. 9A). Next, the binding stability of resveratrol to 3 hub genes was assessed using molecular docking techniques to identify the optimal DN genes. Generally speaking, a binding energy less than − 5 kcal/mol indicates an excellent binding and less than − 7 kcal/mol indicates a strong binding. There were excellent binding interactions between resveratrol and DUSP1. A binding energy was − 4.43 kcal/mol (Fig. 9B).

Figure 9
figure 9

Prediction of drugs for three genes. (A) A drug-genes network based on the CyTargetlinker plug-in in Cytoscape software. (B) Schematic diagram of resveratrol-DUSP1 molecular docking.

Full size image

Discussion

The mechanism of DN is complex and related to various factors. Microalbuminuria is used as an indicator for the diagnosis of early renal impairment in diabetics, but its levels can fluctuate due to numerous factors such as menstrual period and urinary tract infection. Therefore, its use alone limits accuracy in diagnosing DN20. Inflammatory response is significantly associated with the occurrence and development of DN2. Proinflammatory cytokines promote the proliferation of glomerular mesangial cells by activating leukocytes to produce cytokines such as interleukin (IL) and tumor necrosis factor (TNF), increase the permeability of vascular endothelial cells, and cause interstitial fibrosis, which ultimately leads to increased urinary albumin excretion21. Several studies have shown that NLR and PLR also have certain early diagnostic values for DN22. However, their diagnostic value is limited.SII is a novel inflammatory marker composed of neutrophils, platelets, and lymphocytes and is associated with poor clinical outcomes in a variety of solid tumors and inflammatory diseases23. Compared to traditional inflammatory indicators, SII more fully reflects the patient’s immune and inflammatory status24,25. This study clarified that SII could be used as an indicator for the diagnosis of DN by exploring the correlation between SII and DN and demonstrated that immune inflammation played an essential role in the occurrence and development of DN.

Next, we explore the role of immune inflammation in the pathogenesis of DN. Based on the DN-related dataset GSE96804, 30 hub genes associated with immune inflammation were screened using WGCNA analysis, DEG analysis, consensus cluster analysis, and KEGG analysis. The results of the four-fold table diagnostic test showed that 18 genes were DN progression genes and 12 genes were inhibition genes. Therefore, we speculate that immune and inflammation play an important role in the DN progression, mainly attributing to the expression imbalance between harmful and protective genes. KEGG analysis has shown that these genes are related to the NOD-like receptor signaling pathway. A previous study showed that neutrophils isolated from diabetic patients and mice produce neutrophil extracellular traps (NETs)26,27. NETs are correlated with DN severity. Under high glucose conditions, NETs can activate NOD-like receptor thermal protein domain associated protein 3 (NLRP3) inflammasome and damage the function of glomerular endothelial28.

CD8 + T and macrophages were more infiltration in the DN population29. In the early stages of DN, macrophages and T lymphocytes have already infiltrated the glomeruli and renal interstitium30. Their infiltration in the renal interstitium is associated with tubule fibrosis and is involved in the process of renal injury in DN. Macrophages and T lymphocytes recruit leukocytes to infiltrate the renal interstitium, and leukocyte infiltration in the renal interstitium is involved in the initiation of renal fibrosis. It is a key event in the development of DN31.

Our results showed that the most important immune-inflammatory genes in DN are NAIP, ZFP36, and DUSP1. Based on drug-gene network construction and molecular docking analysis, DUSP1 was identified as an effective target gene for Resveratrol. The Dual-Specificity Phosphatase (DUSP) family is the largest group of protein phosphatases (PMID: 37,038,746). It can specifically regulate the activation of Mitogen-Activated Protein Kinase (MAPK) in cells. A study of the mechanism of DKD fibrosis has shown that DUSP1 exerts an anti-fibrotic effect in the HK-2 proximal tubular cell line by inhibiting the activation of the p38MAPK signaling pathway. The activation of angiotensin II-stimulated proteasome degrades DUSP1, which in turn activates signal transducers and transcriptional activators in T cells and induces differentiation of helper T cells32. In addition, inhibition of DUSP1 in T cells may be beneficial for the treatment of T cell-mediated autoimmune diseases such as multiple sclerosis33. Our results further suggest that DUSP1 is a potential target for resveratrol. Resveratrol reduces the rate of albuminuria in patients with diabetic nephropathy. In summary, DUSP1 is involved in the pathogenesis and progression of DN.

Conclusion

In this study, we analyze the correlation between SII and DN using DN data from the NHANES database and find that SII is an effective diagnostic predictor of DN. This result suggests that immune inflammation plays an important role in the occurrence and development of DN. Based on bioinformatics analysis, NAIP, ZFP36, and DUSP1 are immune-inflammation-related genes and related to the NOD-like receptor signaling pathway. The results of molecular docking analysis show that DUSP1 has a better binding energy to resveratrol (an effective drug for the treatment of DN). Thus, DUSP1 regulates the activation of the NOD-like receptor signaling pathway, affects the occurrence of immune inflammation, and then participates in the progression of DN.