Nineteen SLE patients (10 Tibetan, mean age 46.1 ± 16.42 y; nine Han, mean age 35 ± 13.67 y) and 17 healthy controls (nine Tibetan, mean age 39.70 ± 3.76 y; eight Han, mean age 29.78 ± 1.37 y) matched for age, gender, and altitude were recruited ( Supplementary Table S1 ). All patients met ≥ 4 of the revised SLE criteria of the American College of Rheumatology (1997) (12). Informed written consent was obtained, and the study design was approved by the Ethics Committee of the Second Affiliated Hospital of Kunming Medical University.

Total RNA from freshly isolated PBMCs (section 2.7) or cell cultures was extracted with TRIzol Reagent (Invitrogen, San Diego, CA, USA). The cDNA was synthesized using a SureScript™ First-Strand cDNA Synthesis Kit (GeneCopoeia, Guangzhou, China). The mRNA levels were determined by RT-qPCR with BlazeTaq™ SYBR^® Green qPCR Mix 2.0 (GeneCopoeia) and GAPDH as the internal control. The primers are listed in Supplementary Table S2 . Total RNA was reverse-transcribed and quantified with an All-in-One™ miRNA RT-qPCR Detection kit (QP015, GeneCopoeia). U6 small nuclear RNA was the internal control.

Total RNA was extracted from samples before sequencing using TRIzol reagent (Invitrogen, San Diego, CA, USA). RNA integrity and quantity were evaluated by using a Nanodrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, California, USA). The cDNA was synthesized using Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA, USA) and random primers, with enriched and fragmented RNA as template. The cDNA was converted to dsDNA for library construction. The libraries were quantified using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, California, USA). Sequencing was performed using Illumina NextSeq6000 platform (Illumina, San Diego, CA, USA). Multiplex Small RNA Library Prep Set for Illumina was used to construct small RNA libraries based on miRNA sequencing, in accordance with the manufacturer’s protocol. Fragments (18-35 nt) were extracted from the total RNA, and indices and adaptors were added for PCR amplification. Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) was used for library quantification. The library was sequenced by Shanghai Personal Biotechnology Co. Ltd. (Shanghai, China) using Hiseq platform (Illumina).

The reference genome index was set up using Bowtie2 v. 2.4.1 (http://bowtie-bio.sourceforge.net/bowtie2/index.shtml). HISAT2 v. 2.1.0 (http://www.ccb.jhu.edu/software.shtml) was used to map the filtered reads to the reference genome. Reads for each gene were counted using HTSeq v. 0.11.1 (https://htseq.readthedocs.io/en/master/). Gene expression was standardized using FPKM, and differential expression was analyzed using DESeq v. 1.39.0 (http://bioconductor.org/packages/release/bioc/html/DESeq.html) according to the criteria |log2FoldChange| > 1 and P < 0.05. MiRDeep2 v. 2.0.0.8 (https://github.com/Rajewsky-lab/mirdeep2.git) was used to map the clean reads to the reference genome in the miRBase database (http://www.mirbase.org/), to annotate unique reads. Other non-coding RNAs were used for the annotation. Mireap v. 2.0 (https://sourceforge.net/projects/mireap/) was used to analyze the unannotated sequences. MiRNA reads were counted based on the number of sequences conforming with mature miRNA. The most abundant miRNAs were selected for subsequent analysis. DESeq v. 1.30.0 was used to identify differentially expressed miRNAs as those meeting the criteria |log2FoldChange| > 1 and P < 0.05.

The mRNA 3′ untranslated region (UTR) was the target sequence. MiRanda v. 3.3a (https://anaconda.org/bioconda/miranda) was used to predict the target genes of the differentially expressed miRNAs. Kyoto Encyclopedia of Genes and Genomes (KEGG, http://www.kegg.jp/) and gene ontology (GO, http://geneontology.org/) were used to analyze the enrichment of target genes associated with differentially expressed mRNA and miRNAs. All genes were mapped to KEGG and GO terms and pathways, and the latter were filtered using corrected P ≤ 0.05.

To verify the target genes of miR-4512, putative miR-4512 target sequences located in the 3′-UTR of human C-X-C motif ligand (CXCL2, TLR4, CCR1, CXCL1, and CXCL5 transcripts were synthesized and cloned into pmirGLO vector (Promega, Madison, WI, USA) downstream of the Luc2 reporter gene. Plasmids harboring mutated sequences were constructed by inserting each target sequence with various base deletions at the putative miR-4512–binding site (The sequences are listed in Supplementary Figure S1 ). Vectors were sequenced and prepared using an Endo-Free Plasmid ezFlow Mini kit (Promega). Then, HEK293T cells were co-transfected with pmirGLO-(CXCL2, TLR4, CCR1, CXCL1 and CXCL5) and pmirGLO-(CXCL2, TLR4, CCR1, CXCL1 and CXCL5)-MUT plasmids using Lipofectamine™ 2000 (Invitrogen, Carlsbad, CA, USA) and miR-4512 agomir or miR-4512 NC. After 48 h, the cells were harvested, and relative luciferase activity was measured using a Dual-Luciferase Reporter Assay System (Promega, Madison WI, USA), according to the manufacturer’s instructions. Renilla luciferase normalized firefly luciferase activity.

The levels of various cytokines [interleukin (IL)-1β, IL-2, IL-12, tumor necrosis factor (TNF)-α, IFN-γ, IL-4, IL-6, IL-10, and IL-17] in the sera of human patients and experimental mice, supernatants of monocytes or macrophages transfected with NC/miR-4512-agomir/antagomir and mouse renal lysis solution were quantified using ELISA kits (NeoBioscience, Beijing, China), according to the manufacturer’s instructions. Mouse serum antinuclear antibody (ANA), anti-dsDNA antibody, and human serum CXCL2 concentrations were measured using appropriate ELISA kits (JL17316-48T, JL12477-48T, and JL25367-96T, respectively; Jianglai Biology, Shanghai, China). Human serum myeloperoxidase (MPO) concentration was measured with an ELISA kit (0-111594; Jonln, Shanghai, China), according to the manufacturer’s instructions. Anti-TLR4 monoclonal antibody (1:1000, clone 3G9A4; Proteintech Group, Chicago, IL USA) and Anti-neutrophil elastase (NE) monoclonal antibodies (1:100; 89241S; Cell Signaling Technology, Danvers, MA, USA) were used for western blotting, with β-actin was the internal control.

Cell supernatant was treated according to the Picogreen Kit (Invitrogen), and each sample was diluted 1:9 (cell supernatant: TE working solution) with TE working solution; Add 100ul of 10-fold diluted samples to each well, set up two duplicate Wells, and then add 200ul Picogreen reagent in an aqueous working solution chamber to incubate at warm temperature for 10min; Determination: The full-function continuous wavelength scanning microplate analysis system measured all samples in a 96-well microplate at a final volume of 200ul. The standard curve was established with the concentration of DNA standard as the X axis and the fluorescence at 480nm as the Y axis. The concentration of the corresponding sample was calculated according to the fluorescence value of the sample.

PBMCs were separated from ethylene diamine tetraacetic acid (EDTA) blood by density gradient centrifugation (450 × g, 30 min) on Ficoll-Paque Plus (GE Healthcare, Little Chalfont, UK) and resuspended in RPMI 1640 medium supplemented with 100 U/mL penicillin and 100 μg/mL streptomycin (Gibco, Grand Island, NY, USA) and 10% fetal bovine serum (Gibco, Grand Island, NY, USA).

CD14⁺CD16^– monocytes were isolated from freshly isolated PBMCs by immunomagnetic negative selection using an EasySep™ Human Monocyte Isolation kit (19669; STEMCELL Technologies, Vancouver, BC, Canada), according to the manufacturer’s instructions. Human macrophages were obtained by culturing CD14⁺ monocytes for 7 d in RPMI 1640 medium containing 50 ng/mL macrophage colony-stimulating factor(M-CSF). The macrophages were identified by F4/80 expression. Neutrophils were isolated by gradient centrifugation (500 × g, 30 min) on Hypaque-Ficoll solution (LZS11131; TBD, Tianjin, China). Monocytes, neutrophils, and differentiated macrophages were cultured in complete RPMI 1640 medium.

CD14⁺ monocytes from healthy donors and macrophages were transfected with NC/miR-4512-agomir/antagomir using EndoFectin™ Max (EF013; Genecopoeia). After 12 h, the conditioned medium was collected. Fresh neutrophils isolated from healthy donors were exposed to the medium for 3 h. Cells treated with 10 μg/mL lipopolysaccharide (LPS) and 50 ng/mL CXCL2 were used as the positive controls. Culture supernatants were collected for ELISA analysis. Stimulated neutrophils were collected for western blotting and NETs staining and quantification.

The neutrophils were inoculated in the confocal dishes at a concentration of 2×10⁵/mL, fixed with 4% paraformaldehyde for 15 min, then permeabilized with 0.5% Triton X-100 at room temperature for 10 min. Then the cells were stained with 1 μm Sytox Green (Invitrogen) and Hoechst nucleic acid stain (1:1000; Invitrogen). Random photographs were taken under the condition of scattered light at 520 nm. The obtained images were analyzed by Zen 2012 software (n=4). For quantification of NETs, neutrophils were seeded in 96-well black plates (2 × 10⁵ cells/well) in the presence of 5 μM Sytox Green. After stimulation with different conditioned medium for 3 hours, the florescence intensity was quantified with excitation at 485 nm and emission at 535 nm using Infinite M200 Pro (Tecan).

Four female MRL/lpr mice (8-wk-old, 19 ± 0.5 g; Hunan Slack Jingda Experimental Animal Co. Ltd., Hunan, China) were intraperitoneally injected with mouse CXCL2 monoclonal antibody (300-39-2; PeproTech, Rocky Hill, NJ, USA), twice weekly (50 μg/50 μL) for 4 wk. The control (n = 4, female MRL/lpr mice, 8-wk-old, 18 ± 0.5 g; Hunan Slack Jingda Experimental Animal Co. Ltd.) received 50 μL of saline per treatment. Sex- and age-matched normal C57BL/6 mice (n = 4, 8-wk-old, 18 ± 0.5 g; Hunan Slack Jingda Experimental Animal Co. Ltd.) were administered saline according to the above schema. Urine was collected twice weekly from each group. Proteinuria was detected by Coomassie Brilliant Blue (CBB) staining (Bayer, Elkhart, IN, USA). Mice were sacrificed by CO₂ asphyxiation at 12 wk. The blood was collected, and serum levels of anti-dsDNA, ANA and multiple cytokines were measured using commercial kits (Jianglaibio, Shanghai, China). The kidneys were excised and dissected for immunocytochemistry (IHC) staining (as described in section 2.10) and cytokine analysis by ELISA.

The mice were randomly assigned to different groups according to their unique code, and all data were collected by technicians blinded to the treatment protocol. The experiment was conducted in a laboratory animal facility of Kunming Medical University and approved by the Animal Research Committee of Kunming Medical University (No. kmmu2021724).

Mouse kidneys were fixed in 4% (v/v) paraformaldehyde, embedded in paraffin, sectioned (6 μm) using a rotary microtome (Leica Biosystems, Wetzlar, Germany), and stained with hematoxylin and eosin (H&E). Active and chronic renal tubular changes were semi-quantitatively analyzed as previously described (13). The activity index (AI) data included glomerular endocapillary proliferation, intraglomerular leukocyte infiltration, platinum ear-pick and/or hyaline thrombus, fibrinoid necrosis and/or nuclear fragmentation, cellular crescent, and inflammatory cell infiltration in the interstitial space. A score in the range of 0–3 was assigned to each item, and all items were summed to yield a final AI score. The chronic index (CI) parameters included glomerulosclerosis, fibrous crescent, renal tubular atrophy, and renal interstitial fibrosis. A score in the range of 0-3 was assigned to each item, and all items were summed to yield a final CI score. Four items in the renal tubulointerstitial lesion (TIL) category were evaluated: tubular degeneration or necrosis, renal tubular atrophy, renal interstitial fibrous, and renal interstitial inflammatory cell infiltration. A score in the range of 0–3 was assigned to each item, and all items were summed to yield a final TIL score. Tissue sections were examined by an experienced pathologist blinded to the treatments.

For IHC staining, renal paraffin sections were dewaxed, rehydrated, and recovered in antigen recovery solution (ETDA with citric acid) in an autoclave. Endogenous peroxidase activity was blocked with 0.3% (v/v) hydrogen peroxide. The sections were blocked with 10% (v/v) normal goat serum at 37°C for 1 h and incubated with rabbit anti-CXCL2 monoclonal antibody (MAb) (1:200; GTX74085; GeneTex, Irvine, CA, USA), rabbit anti-CD45 MAb (1:1000; 20103-1-AP; Proteintech, Chicago, IL, USA), and rabbit anti-MPO MAb (1:1000; ab188211; Abcam, Cambridge, UK) at 4°C overnight. Tissues were washed with phosphate-buffered saline, incubated with horseradish peroxidase-conjugated secondary antibody, and analyzed by using a DAB kit. The nuclei were counterstained with hematoxylin. Whole-tissue images were acquired using a Pannoramic 250 flash II scanner (3D HISTECH, Budapest, Hungary) at ×20magnification. According to the intensity of immunohistochemical staining, the glomeruli were divided into four grades. Specifically, deposition <20%, 20% ≤ deposition <50%, 50% ≤ deposition <75%, and deposition ≥ 75 are defined as degrees I, II, III, and IV, respectively. Data are expressed as the percentage of glomeruli in each staining category relative to the total. Three sequential sections per animal were examined, and averages were statistically analyzed.

Data are presented as the mean ± SEM. For normal variance distribution, unpaired Student’s t-test was used to determine the differences between group pairs. One-way ANOVA and post hoc Kruskal–Wallis test were used to compare ≥ 3 groups. GraphPad Prism 5 (GraphPad Software, La Jolla, CA, USA) was used for the calculations. Statistical significance was set at P < 0.05.

We used RNA sequencing to analyze miRNA and mRNA expression profiles of PBMCs from SLE (n = 5) and healthy (n = 5) Tibetans. Overall, 469 mRNAs (DEmRNAs) and 45 miRNAs (DEmiRNAs) were significantly differentially expressed in Tibetan SLE patients. Of these, 356 DEmRNAs and 13 DEmiRNAs were upregulated, while 113 DEmRNAs and 32 DEmiRNAs were downregulated ( Figures 1A, C ).

To identify the functional differences between SLE and healthy patients, we used the 469 DEmRNAs in GO enrichment analysis. We thus obtained top 10 GO terms related to Biological Processes (BP), Molecular Functions (MF), and Cellular Components (CC) at P < 0.05 ( Figure 1B ). Most items in the BP category were related to immune and neutrophil activation and degranulation. Chemokine activity and multiple cellular secretion-related components were enriched in the MF and CC categories, respectively. Hence, GO enrichment analysis suggested that neutrophil migration and activation participate in SLE pathogenesis.

To elucidate the regulatory mechanism of DEmRNAs in SLE, we investigated their relationships with DEmiRNAs. We used RT-qPCR to validate four downregulated and four upregulated DEmiRNAs with the largest expression fold changes in PBMCs isolated from five Tibetan SLE patients, and five age- and sex-matched Tibetan controls. MiR-3942-3p, miR-3165 and miR-3614-3p were upregulated, while miR-4482-3p, miR-99a-3p, miR-4728-3p, miR-4512 and miR-4434 were downregulated in PBMCs from Tibetan SLE patients ( Figure 1D ). We predicted the target genes of these candidate miRNAs using TargetScan (http://www.targetscan.org). The analysis revealed 332 target genes of 469 DEmRNAs, miR-4512 was predicted to target the largest number of genes.

According to the KEGG analysis, the DEmRNAs targeted by hsa-miR-4512 were significantly enriched in the chemokine signaling, rheumatoid arthritis-associated, NF-kB signaling, and other immune inflammatory pathways (P < 0.05) ( Figure 1E ), while the DEmRNAs targeted by other candidate miRNAs were not enriched in any immune-related pathways ( Supplementary Figure S2 ). CCR1, CXCL1, CXCL2, CXCL5, and TLR4 mRNA were negatively correlated with miR-4512 in the immune inflammatory pathways, and we used them to construct regulatory networks ( Figure 1F ). Thus, hsa-miR-4512 might regulate neutrophil activation and chemokine-related pathways.

TargetScan and KEGG pathway analysis were used to narrow the range of miR-4512 predicted targets based on differentially expressed mRNAs associated with miR-4512 in immune-inflammatory pathways. The recognized target genes of miR-4512 were used in pathway analysis based on CapitalBio molecular annotation system v.3.0. The prediction results suggested that CCR1, CXCL1, CXCL2, CXCL5, and TLR4 might be the target of miR-4512 ( Figure 1F ). Luciferase reporter plasmid containing mRNA sequences of WT or mutated 3’-UTR target genes was co-transfected into HEK293T cells with NC and miR-4512 agomir. Only TLR4 and CXCL2 were directly targeted by miR-4512 ( Figures 2A, B ). RT-qPCR analysis showed that TLR4 and CXCL2 were also negatively regulated by miR-4512 in BALL-1, Jurkat and K562 cells, as well as in human primary monocytes and macrophages ( Figures 2C–F and Supplementary Figure S3 ). In addition, miR-4512 inhibited the expression of CXCL2 and TLR4 proteins in human primary monocytes and macrophages ( Figures 2G–I ). The elevated serum CXCL2 level and the high expressed TLR4 in PBMCs of another group of SLE patients indicate that the down-regulation of miR-4512 in SLE promotes the expression of CXCL2 and TLR4 ( Figures 2J–L ).

Considering the direct regulation of TLR4 and CXCL2 by miR-4512, one could speculate that miR-4512 mainly plays a role in cells which express sufficient TLR4, such as monocytes, granulocytes, and macrophages (14). Accordingly, we isolated PBMCs, monocytes, and neutrophils from Tibetan and Han SLE patients and their matched normal controls. RT-qPCR analysis of miR-4512, CXCL2, and TLR4 levels revealed that miR-4512 levels were decreased in monocytes and macrophages, but not neutrophils, from SLE patients ( Figures 3A–C ).

We then analyzed the impact of miR-4512 downregulation on cytokine secretion in monocytes and macrophages. Primary monocytes and macrophages pooled from healthy subjects were transfected with NC/miR-4512-agomir/antagomir, and the cytokines in their cell culture supernatants were quantified. As shown in Figure 3D , miR-4512 inhibited the secretion of pro-inflammatory cytokines IL-1β, IL-12, TNF-α, IFN-γ, and IL-17 and promoted the secretion of anti-inflammatory cytokines IL-4 and IL-10. This finding was consistent with the cytokine secretion phenotype of SLE patients ( Figure 3E ). Therefore, miR-4512 downregulation in SLE patient monocytes and macrophages may activate monocytes and macrophages and induce pro-inflammatory cytokine secretion.

Defective apoptotic debris clearance and NETs degradation might lead to the formation of autoantibodies in SLE pathogenesis (15). As miR-4512 downregulation in monocytes and macrophages promotes pro-inflammatory cytokine secretion and CXCL2 expression, miR-4512 may indirectly prime neutrophils for NETosis. As shown in Figures 4A, B , the levels of NETs-relevant indicators [MPO and cfDNA] were elevated in SLE patient sera. The levels of NE, the main NETs component, were significantly upregulated in neutrophils isolated from SLE patients ( Figure 4C ). Hence, dysregulation of NETs release and/or defective NETs clearance are involved in the pathogenesis of SLE.

NC/miR-4512-agomir/antagomir was then transfected into primary monocytes and macrophages to determine the role of miR-4512 on the formation of NETs. 12 h after transfection, the supernatants of each group were isolated and used to stimulate neutrophils isolated from the healthy control group. As shown in Figures 4D, E , the supernatants of monocytes and macrophages transfected with miR-4512-agomir significantly reduced MPO and cfDNA release by neutrophils. The effect of the supernatant from monocytes and macrophages transfected with miR-4512-antagomir on neutrophil NETosis reached a level comparable to that of the LPS stimulation group. Straining and quantification of extracellular DNA with SYTOX Green ( Figures 4F, G ) and Western-Blot analysis of NE in the neutrophils ( Figure 4H ) also indicated similar results. Remarkably, CXCL2-stimulation also promoted neutrophil NETosis, suggesting that miR-4512 present in monocytes and macrophages may modulate neutrophil NETosis through chemoattraction and activation of neutrophils by secreting CXCL2.

In order to verify the function of miR-4512 in SLE pathogenesis, we blocked the target gene CXCL2 with a neutralizing antibody in the MRL/lpr lupus model as the miRNA was not conserved in related species ( Figure 5A ). Compared with normal C57 mice, increasing proteinuria was observed in the anti-CXCL2Ab and untreated groups. However, the proteinuria was significantly reduced in the anti-CXCL2Ab group compared with the untreated group at 11 wk (P = 0.0044) ( Figure 5B ). H&E staining ( Figure 5C ) was used to assess the pathological changes in the kidney at the end of the study period. The major features analyzed included inflammatory cell infiltration, membrane proliferation, and arterial wall destruction ( Figures 5D–F ), which were all observed in the untreated group. By contrast, the occurrence of renal lesions was markedly mitigated in the anti-CXCL2Ab group. Therefore, anti-CXCL2Ab might alleviate kidney damage.

ANA is an important element in SLE patient categorization (16). Autoantibody levels are also related to lupus nephritis activity (17). Serum anti-dsDNA antibody and ANA levels were determined. The anti-dsDNA antibody and ANA levels were significantly lower in the plasma of anti-CXCL2Ab MRL/lpr mice than those in the untreated MRL/lpr mice ( Figures 5G, H ; P < 0.001). Hence, anti-CXCL2Ab treatment decreases serum immunological markers of SLE.

Next, immunohistochemistry was used to determine the deposition levels of CXCL2, CD45, and MPO in glomerular sections ( Figures 6A–F ). As we expected, the glomerular CXCL2, CD45, and MPO levels of MRL/lpr mice were significantly higher than those of normal C57 mice ( Figures 6A–E ). The staining levels of CXCL2, CD45 and MPO in the glomeruli of MRL/lpr mice were significantly reduced by CXCL2Ab treatment ( Figures 6B–F ). Specifically, compared with untreated MRL/lpr mice, the percentage of III-degree glomeruli in MRL/lpr mice was significantly decreased, while the percentage of I degree glomeruli was significantly increased by CXCL2Ab, which were confirmed by the quantitative results of CXCL2 and CD45 ( Figures 6A–D ). In the quantitative results of MPO, the percentage of II-degree glomeruli was significantly increased by CXCL2Ab ( Figures 6E, F ). This confirmed that the glomerular damage of MRL/lpr mice was significantly reduced by CXCL2Ab, which was dependent on the decreased monocytes/neutrophils infiltration and activation.

The results of ELISA assay showed that the levels of IL-1β, IL-2, IL-6, IL-12, TNF-α, IFN-γ and IL-17 in the kidney and plasma were significantly down-regulated by CXCL2Ab, while IL-4 and IL-10 levels were significantly increased ( Figure 6G ). Therefore, CXCL2Ab was confirmed to regulate the secretion of inflammatory cytokines. These results indicate that CXCL2Ab is beneficial to relieve the immune disorder of SLE.