LRNA9884, a Novel Smad3-Dependent Long Noncoding RNA, Promotes Diabetic Kidney Injury in db/db Mice via Enhancing MCP-1–Dependent Renal Inflammation

Type 2 diabetes is a major metabolic disorder that causes a profound medical and socioeconomic burden worldwide (1). Type 2 diabetes is associated with a series of complications on multiple organ systems, where diabetic nephropathy (DN) is the most severe complication of diabetes and is a leading cause of end-stage kidney disease (2). The pathological features of DN include diffuse glomerular basement membrane thickening, mesangial expansion, podocyte reduction, and renal fibrosis (3). Increasing evidence demonstrates that low-grade renal inflammation is a key mechanism leading to diabetes and diabetic complications (4).

Emerging evidence shows that long noncoding RNAs (lncRNAs), defined as RNAs >200 nucleotides that do not encode any protein, play an important role in the development of type 2 diabetic nephropathy (T2DN) (5,6). It has been shown that an lncRNA transcript (lncMGC) is increased in the glomeruli of mouse models of DN in response to endoplasmic reticulum stress and that targeting this lncRNA can protect against early DN (7). In addition, lncRNA CYP4B1-PS1-001 protects against the progression of T2DN via inhibiting proliferation and fibrosis of mesangial cells during the early stage of DN, and lncRNA metastasis associated lung adenocarcinoma transcript 1 (MALAT1) is dysregulated in DN (8,9). A recent study also found that an lncRNA taurine-upregulated gene 1 (Tug1) is reduced in diabetic kidney and that overexpression of Tug1 largely improves mitochondrial bioenergetics in the podocytes of diabetic mice via a peroxisome proliferator–activated receptor-γ coactivator-1a–dependent mechanism (10). Nevertheless, the potential roles and working mechanisms of lncRNAs in DN are still largely unknown (11).

By using high-throughput RNA sequencing, we uncovered the involvement of lncRNAs in renal diseases on both anti- glomerular basement membrane (GBM) and unilateral ureteric obstruction (UUO)-induced mouse kidney disease models and identified 21 novel Smad3-dependent lncRNAs participating in the renal pathogenesis (12). Among these lncRNAs, a novel Smad3-dependent lncRNA, Erbb4-IR, is highly upregulated in the UUO kidney and T2DN with progressive renal fibrosis, and knockdown this lncRNA can protect kidneys from both UUO and diabetic injury (13,14). In the current study, we also found that a potential lncRNA transcript LRNA9884 was the highest expressed Smad3-dependent lncRNAs in the kidneys of db/db mice compared with the db/m mice as well as the Erbb4-IR level. We therefore further elucidated its potential pathogenic role and underlying mechanism in T2DN.

Male db/db and db/m mice were purchased from the Chinese University of Hong Kong Laboratory Animal Services Centre. All mice were fed in a standard animal house with 12-h/12-h light/dark cycle and were sacrificed by intraperitoneal injection of ketamine/xylene. Smad3 knockout (KO) db/db mice were generated by crossing heterozygous db/m with heterozygous Smad3^+/− (C57BL/6 background), as previously described (14). All studies and experimental procedures were approved by the Chinese University of Hong Kong Animal Experimentation Ethics Committee, and the experimental methods were performed in accordance with the approved guidelines.

Total RNA was isolated from the cultured cells and renal cortex of kidney tissues using Trizol (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. Real-time PCR was performed by SYBR Green Supermix using the CFX96 PCR System (Bio-Rad, Hercules, CA). The primers used in this study, including mouse collagen I, fibronectin, MCP-1, and β-actin, have been previously described (15–18). Other primers include LRNA9884 forward 5′- ACCACAGCAGATACCCAAGC-3′ and reverse 5′- CAGCAAGCTCCTTTTTCCAC-3′. The relative level of the detected gene was normalized with the internal control β-actin. Relative changes in target mRNA expression were determined by using the 2^−ΔΔCT method.

The full length sequence of LRNA9884 was obtained by conducting rapid amplification of cDNA 5′ and 3′ ends (RACE) with SMARTer RACE 5′/3′ Kit (Clontech, Mountain View, CA), as previously described (13). In brief, the cDNA generated from the total RNA of diabetic injured kidney from the 20-week-old db/db mice was used as the template for RACE PCR according to the user’s manual. The full length of LRNA9884 cDNA was synthesized by PCR with the forward primer 5′- AAAACAGGAGCAGGAAGAATTGGG-3′ and reverse primer 5′- TTGAACTCTGGTCTTCAGAC-3′.

Genomic location of the LRNA9884 was identified by Basic Local Alignment Search Tool (BLAST) (https://blast.ncbi.nlm.nih.gov/Blast.cgi and http://www.genome.ucsc.edu/). Smad3 binding site on the LRNA9884 genomic sequence was identified by the Evolutionary Conserved Regions (ECR) browser (https://ecrbrowser.dcode.org/), as reported in our previous study (12). The protein-coding potential of the LRNA9884 sequence was evaluated by two widely used computational programs: coding potential calculator (CPC) (http://cpc.cbi.pku.edu.cn/programs/run_cpc.jsp) and the coding potential assessment tool (CPAT) (http://lilab.research.bcm.edu/cpat/index.php) (19,20). For evaluation in CPC, transcripts with scores higher than 1 are predicted to be “coding,” lower than −1 are “noncoding,” and between −1 and 1 are classified as “weak noncoding” (−1, 0) or “weak coding” (0, 1). For identifying the direct binding site of LRNA9884 on the MCP-1 promoter, the sequence of LRNA9884 from RACE and the promoter region of MCP-1 from the Eukaryotic Promoter Database (https://epd.vital-it.ch/index.php) were analyzed by IntaRNA tool on the Freiburg RNA Tools website using the default settings (http://rna.informatik.uni-freiburg.de/IntaRNA/Input.jsp) (21), followed by dual-luciferase reporter assay for validating their potential RNA-DNA interaction as in our previous work (13,14).

The expression level of LRNA9884 in the kidney was detected by using in situ hybridization (ISH), as previously described (14). The kidney sections (4 μm) were fixed in 4% (w/v) paraformaldehyde with 1% (v/v) DMSO. After rehydration, we use a permeabilization step by using proteinase K. Probe for ISH is carried with a locked nucleic acid–digoxigenin labeled LRNA9884 probe (5′-ACTTGAAGGGTCCAGAAGAGAT-3′) (Exiqon, Vedbaek, Denmark) or negative control scramble probe (5′-GTGTAACACGTCTATACGCCCA-3′) (Exiqon) in a humid chamber at 54°C for 2 h. The sections were washed and incubated with anti-digoxigenin antibody conjugated to alkaline phosphatase (Roche Diagnostics, Indianapolis, IN) overnight at 4°C and developed with phosphate/nitroblue tetrazolium (Sigma-Aldrich, St. Louis, MO) (22).

Fluorescence ISH (FISH) with immunofluorescence assay was performed with a Cy3-conjugated LRAN9884 probe 5′-ACTTGAAGGGTCCAGAAGAGAT-3′ and scramble control 5′-GTGTAACACGTCTATACGCCCA-3′ (GenePharma, Shanghai, P.R. China) by using a FISH kit (C10910; RiboBio, Guangzhou, P.R. China) according to the manufacturer’s instruction. Immunofluorescence assay was followed by FISH, and FITC-conjugated Pan Cytokeratin (53-9003-82; eBioscience, San Diego, CA) and nephrin (ab58968; Abcam, Cambridge, U.K.) were the primary antibodies. All stained samples were imaged under a fluorescence microscope (Axio Observer.Z1; Carl Zeiss, Oberkochen, Germany) (14).

Immortalized murine kidney proximal tubular epithelial cells (mTECs) (a gift from Dr. Jeffrey B. Kopp, National Institutes of Health) were cultured in DMEM/F-12 (Gibco, Carlsbad, CA), supplemented with 10% FBS (Gibco) and 1% antibiotic/antimycotic solution (Life Technologies, Grand Island, NY) (13,15,16). The murine SV40 transfected mesangial cell line SV40 MES 13 (mMC) (ATCC, Manassas, VA) was maintained in a 3:1 mixture of DMEM and Ham’s F-12 medium containing 5% FBS, and 1% antibiotic/antimycotic solution (Life Technologies) (14). Mouse embryonic fibroblasts (MEFs) lacking Smad3 or Smad2 were generated and characterized as described previously (23), and were cultured in DMEM/F-12 medium (Gibco), as previously described (24). Advanced glycation end products (AGE), at a dose of 50 μg/mL (Cat# A8426; Sigma-Aldrich), was added to induce responses, which was previously dissolved in endotoxin-free 1 × PBS. The composition of AGE includes protein and 95% biuret, and specific of AGE is bovine, their activity on Smad3 activation and MCP-1 expression were validated as position control. To inhibit Smad3 activity, cells were pretreated with the Smad3 inhibitor SIS3 (S0447; Sigma-Aldrich) at a dose-dependent pattern of 1–4 μmol/L for 2 h before AGE stimulation.

Chromatin immunoprecipitation (ChIP) was performed with the SimpleChIP Enzymatic Chromatin IP Kit (magnetic beads) (#9003; Cell Signaling, Danvers, MA), as previously described (15,18,25). Immunoprecipitation was performed with the antibody against Smad3 (#9513; Cell Signaling) or a normal IgG as a negative control. Precipitated DNA fragments were detected by PCR using a specific primer of promoter region of LRNA9884: forward 5′-TCGCCTCCAGTTGTCTTTCT-3′, reverse 5′-CCTGAAGGACAGCCACTCTC-3′.

Groups of eight male mice were used: db/db mice were treated with pSuper.puro vector (Oligoengine, Seattle WA) empty vector (EV) control or plasmid overexpressing shRNA sequence targeting LRNA9884 (shRNA) (5′-GATCCGACCACAGCAGATACCCAAGCttcaagagaGCTTGGGTATCTGCTGTGGTCTTTTTTGAATTCA-3′) via ultrasound-mediated gene transfer method from the age of week 12 and sacrificed at week 20. To maintain the transgene expression levels, LRNA9884 shRNA expressing plasmids were transfected at age 12, 15, and 18 weeks, as previously described (13,14). In brief, mice received the mixed solution (400 μL/mouse) containing the LRNA9884 shRNA-pSuper.puro vector or pSuper.puro EV (200 μg/mouse) and lipid microbubbles (Sonovue; Bracco, Milan, Italy) at a ratio of 1:1 (v/v) via the tail vein injection, then immediately using an ultrasound transducer (Therasonic; Electro Medical Supplies, Wantage, U.K.) directly placed on the skin of the back against the left kidney with a pulse-wave output of 1 MHz at 2 W/cm² for a total of 5 min on each side, as previously described (13–16).

Kidneys sections were fixed in 4% paraformaldehyde, and stained with periodic acid-Schiff (PAS), periodic acid-silver methenamine (PASM), and Masson trichrome staining methods. Immunohistochemistry was performed in 4-μm paraffin sections by using a microwave-based antigen-retrieval technique. The antibodies used in this study included F4/80 (MCA497; Serotec, Oxford, U.K.), interleukin-1β (IL-1β) (sc-7884; Santa Cruz Biotechnology, Dallas, TX), MCP-1 (sc-1785; Santa Cruz Biotechnology), and tumor necrosis factor-α (TNF-α) (sc-1351; Santa Cruz). After immunostaining, sections were counterstained with hematoxylin. The mesangial matrix index was quantified using Image-Pro Plus software (Media Cybernetics, Bethesda, MD) by PAS staining in 10 random fields (original magnification ×200), as previously described (13). Glomerulosclerosis was quantified by Masson trichrome staining, where 20 glomeruli were randomly selected from each section, and positive signals within the selected glomerulus were highlighted, measured, and represented as the percentage positive area of the entire glomerulus. The percentage of positive staining areas of MCP-1, IL-1β, and TNF-α expression was quantified using the same software in 10 consecutive fields, whereas positive cells for F4/80⁺ cells were counted under the power field of microscope in 10 random areas of kidney tissues using a 0.25-mm² graticule fitted in the eyepiece of the microscope and expressed as cells/mm² (13,14,26).

Mice were placed in metabolic cages for collection of 24-h urinary samples every 4 weeks from the age of 8 weeks. Urinary microalbumin was measured by competitive ELISA according to the manufacturer’s instructions (Exocell, Philadelphia, PA). Urinary and serum creatinine was measured by an enzymatic kit (Stanbio Laboratories, Boerne, TX). Urinary albumin excretion was expressed as total urinary albumin–to–creatinine ratio (μg/mg), as previously reported (26).

Blood glucose levels were measured every 4 weeks by Accu-Chek glucose meter (Roche Diagnostics, Basel, Switzerland) after the mice were fasted for 6 h, as recommended by the Animal Models of Diabetic Complications Consortium.

The mTECs were transfected with 100 nmol/L LRNA9884 siRNA (sense 5′-GACCACAGCAGAUACCCAAGC-3′, antisense 5′-GCUUGGGUAUCUGCUGUGGUC-5′) or nonsense control (NC; sense 5′-UUCUCCGAACGUGUCACGUTT-3′, antisense 3′-ACGUGACACGUUCGGAGAATT-5′) using Lipofectamine RNAiMAX reagent (Invitrogen) according to the manufacturer’s instructions as previously reported (27), then stimulated with AGE. In addition, the mTECs were also transfected with 0.5 μg/mL pcDNA3.1 EV or plasmid containing LRNA9884 full-length RNA sequence, as shown in Supplementary Fig. 1 (LRNA9884; constructed by PharmaGene, Shanghai, China) using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s manual, and the culture.

Protein from renal cortex and cultured cells was extracted using radio immunoprecipitation assay (RIPA) lysis buffer. Western blot analysis was performed as previously described (28,29). In brief, after nonspecific binding was blocked with 5% BSA, nitrocellulose membranes were incubated overnight at 4°C with primary antibodies against phosphorylated-Smad3 (Cell Signaling Technology, Danvers, MA) and Smad3 (Cell Signaling Technology), collagens I (1310-01; Southern Tech, Birmingham, AL), fibronectin (sc-8422; Santa Cruz Biotechnology), and β-actin (sc-47778; Santa Cruz Biotechnology), followed by IRDye800-conjugated secondary antibodies (Rockland Immunochemicals, Gilbertsville, PA). Signals were detected using the LI-COR/Odyssey infrared image system (LI-COR Biosciences, Lincoln, NE), followed by quantitative analysis using the Image J program (https://imagej.nih.gov/ij/?) (13,18). The ratio for the protein examined was normalized against β-actin and expressed as the mean ± SEM.

Medium from stimulated mTECs was collected to detect the cytokine production using an ELISA kit. MCP-1 was measured with Quantikine ELISA Kit (R&D Systems, Minneapolis, MN) according to the product protocols.

The pcDNA3.1⁺ plasmids containing LRNA9884 full-length sequence with (LRNA9884) or without (LRNA9884-mutant) predicted the binding site on the MCP-1 promoter; and pGL3-basic reporter plasmids containing MCP-1 promoter sequence with (pGL3–MCP-1) or without predicted LRNA9884 binding site (pGL3–MCP-1–mutant) were constructed by Biowit Technologies Ltd., China (13,14). The luciferase reporter assay was performed by Landbiology (Guangzhou, China) by using dual luciferase reporter assay kit (Promega, Madison, WI) as our previous studies (24). The empty vectors (pGL3-basic, pcDNA3.1⁺), or MCP-1 reporter plasmids (pGL3–MCP-1, pGL3–MCP-1–mutant), or LRNA9884-overexpressing pcDNA3.1⁺ plasmids (LRNA9884, LRNA9884-mutant) were cotransfected into mTECs. Luciferase activities were measured at 48 h according to the manufacturer’s instructions by GloMax-Multi Detection System (Promega). The reporter activity was represented by the ratio of the firefly luciferase activity (M1) to the Renilla luciferase activity (M2) as M1/M2, and shown as mean ± SEM fold induction of luciferase in three independent experiments.

All of the data are expressed as mean ± SEM. Statistical analyses were performed with one-way ANOVA, followed by Newman-Keuls multiple comparison from GraphPad Prism 5.0 (GraphPad Software, La Jolla, CA). In addition, a repeated-analysis ANOVA was used for albumin excretion, body weight, and fasting blood glucose analysis.

Among all previously identified Smad3-depedent lncRNAs (11), real-time PCR analysis revealed that a potential lncRNA transcript (na_9884) was highly expressed in the diabetic kidney of db/db mice (Fig. 1A). More importantly, expression of na_9884 progressively increased in the diabetic kidneys of db/db mice compared with db/m since the age of 8 weeks (Fig. 1B). We then further characterized this novel lncRNA by obtaining its full-length sequence via RACE (Supplementary Fig. 1). Our result identified that na_9884 is 508 nucleotides in full length, which locates on murine chromosome 5 (Chr5: 52973035–52992344) and covers the whole genomic sequence of an lncRNA transcript 5033403H07Rik (Fig. 1C and D). By using the CPC and CPAT analysis, we confirmed that na_9884 scores −1.01155 and 0, respectively, as an RNA without protein-coding capacity (Fig. 1E). Thus, we named this novel lncRNA as LRNA9884. By ISH, we further observed that LRNA9884 is highly expressed in the nucleus of renal residential cells in the diabetic kidney of 8-week-old db/db mice and further upregulated in 20-week-old db/db mice (Fig. 1F).

The LRNA9884 expression was further confirmed by FISH assay with immunofluorescence costaining. Owing to lack of the specific marker in mouse mesangial cells, the epithelial cell marker keratin and podocyte marker nephrin were detected in kidneys of 20-week-old db/db mice. Results demonstrated that LRNA9884 was largely expressed in the nucleus of epithelial cells and was lacking in podocytes compared with the db/m mice (Fig. 2A), suggesting that LRNA9884 was mainly expressed in epithelial cells and mesangial cells. Aberrant glucose metabolism results in hyperglycemia, and accumulation of various byproducts (e.g., AGE) is suggested to promote renal injury in diabetic kidney disease. Thus, we cultured mTECs and mMCs under high glucose or AGE conditions in vitro. We found that LRNA9884 expression was significantly upregulated in both mTECs and mMCs at 1 h by AGE (50 μg/mL) stimulation in a time- and dose-dependent manner (Fig. 2B–E), whereas its induction was significant at 3 h under high glucose (35 mmol/L) stimulation (Supplementary Fig. 2).

In our previous study, induction of LRNA9884 was associated with chronic kidney disease mouse models in a Smad3-dependent manner as revealed by high-throughput RNA sequencing (12), but its regulatory mechanism in diabetic kidney injury is still unknown. Interestingly, we observed the renal expression of LRNA9884 in 12-week-old db/db mice was largely suppressed after Smad3 KO, as shown by ISH and real-time PCR (Fig. 3A and B). Thus, we further speculated that LRNA9884 may be regulated by Smad3 under diabetic conditions. This was examined in MEFs, in which AGE-induced upregulation of LRNA9884 was also largely suppressed in MEFs, knocking out Smad3 but not Smad2 (Fig. 3C and D and Supplementary Fig. 3A and B). By using the ECR browser, we predicted a conversed Smad3 binding site on the promoter region of LRNA9884 genomic sequence (Fig. 4A). The physical binding of Smad3 protein on the promoter region of the LRNA9884 genomic sequence was dramatically enriched in mTECs by AGE stimulation as demonstrated by ChIP assay (Fig. 4B). Thus deletion (Smad3 KO) or inhibition (SIS3) of Smad3 effectively blocked the AGE-induced LRNA9884 expression in MEFs and mTECs in vitro, (Fig. 4C and D and Supplementary Fig. 3). Thus, our findings clearly demonstrated that expression of LRNA9884 is tightly regulated by Smad3 signaling and serves as a direct Smad3 target gene under diabetic condition in vitro and in vivo.

Because LRNA9884 induction was highly associated with T2DN progression (Fig. 1B and F), we further investigated its potential role in the pathogenesis of T2DN in db/db mice. EVs or plasmids expressing shRNA specifically against LRNA9884 were transfected into both kidneys of the db/db mice (100 μg/mouse per treatment) by using an ultrasound-microbubble delivery system (9). Treatment with shRNA effectively suppressed expression of renal LRNA9884 in db/db mice compared with the EV-treated 20-week-old control mice according to results of real-time PCR and ISH (Fig. 5A and B). More importantly, inhibition of renal LRNA9884 significantly suppressed renal histological injury, including glomerular matrix deposition and glomerulosclerosis (Fig. 5A, C, and D), and decreased microalbuminuria excretion and serum creatinine (Fig. 5E and F) as determined by PAS, PASM, and Masson trichrome stain and ELISA assays. However, blockade of renal LRNA9884 showed no significant therapeutic effects on the type 2 diabetes phenotype, including body weight and fasting blood glucose (Supplementary Fig. 4), suggesting LRNA9884 promotes type 2 diabetes complications via a kidney-specific mechanism.

Immunohistochemistry and real-time PCR detected that a significant reduction of inflammatory markers (i.e., MCP-1, TNF-α, and IL-1β) occurred in the kidney of 20-week-old db/db mice by shRNA-9884 treatment compared with the EV-treated and db/db controls (Fig. 6). To elucidate the detailed mechanism mediated by LRNA9884, we screened Smad3-related downstream inflammatory biomarkers after knockdown LRNA9884 by using real-time PCR. Results showed that AGE-induced inflammation and proinflammation cytokines were downregulated after silencing of LRNA9884 in vitro, among which inhibition of MCP-1 expression by shRNA treatment was the most remarkable one, suggesting that LRNA9884 may trigger MCP-1–dependent renal injury under diabetic condition (Supplementary Fig. 5).

In addition, no significant changes in the AGE-induced renal fibrotic marker expression (i.e., collagen I, fibronectin, and phosphorylated Smad3) in mTECs after effective siRNA-mediated knockdown of LRNA9884 (siLRNA9884) compared with the nonsense-treated control as shown by real-time PCR and Western blot analysis (Fig. 7 and Supplementary Figs. 5 and 6). MCP-1 has a crucial role in renal inflammation via promoting the infiltration of inflammatory leukocytes into injured kidney; therefore, the underlying mechanism of how LRNA9884 regulates MCP-1 production at the transcriptional level was further elucidated in this study, as shown below.

Mechanistically, we found that LRNA9884 was continuously increased in the nucleus of mesangial and tubular epithelial cells during T2DN progression, which was associated with a marked upregulation of MCP-1 in the diabetic injured kidney (Figs. 1B and 6B). We also found that a potential binding site of LRNA9884 is predicted on the promoter region of the MCP-1 genomic sequence by using the Freiburg RNA Tool (Fig. 8A and Supplementary Fig. 7). To investigate the functional role of LRNA9884, we performed siRNA-mediated knockdown of LRNA9884 on mTECs and found that silencing of LRNA9884 largely inhibited AGE-induced MCP-1 production in mTECs (Fig. 8B and C and Supplementary Fig. 8). Furthermore, overexpression of LRNA9884 could directly upregulate the expression of MCP-1 without AGE stimulation (Fig. 8D and Supplementary Fig. 8). More importantly, direct physical interaction between LRNA9884 and the promoter region of the MCP-1 genomic sequence was clearly demonstrated in vitro by dual-luciferase reporter assay, where overexpression of full-length LRNA9884 significantly enhanced the transcription activity of MCP-1 via direct binding on its promoter region (pGL3–MCP-1–promoter), which was effectively prevented by deletion of the predicted LRNA9884 binding site on the MCP-1 promoter (pGL3–MCP-1–mutant) or the LRNA9884 full-length sequence (pcDNA3.1-LRNA9884-mutant) (Fig. 8E). Therefore, targeting of renal LRNA9884 effectively protected against AGE-driven diabetic kidney injury by blocking renal inflammation via a MCP-1–dependent mechanism. This was further demonstrated by a significant reduction of the kidney-infiltrating macrophages in shRNA-9884–treated diabetic kidney compared with EV-treated and untreated db/db controls (Supplementary Fig. 9). Therefore, LRNA9884 may represent as an effective novel therapeutic target for AGE-driven T2DN.

It is well accepted that DN as a severe diabetic complication is mediated by TGF-β1 via both Smad3-dependent and non–Smad3-dependent signaling pathways (3,30,31). In this study, we uncovered the pathogenic role of a novel Smad3-dependent LRNA9884 in the disease progression of T2DN and identified its molecular mechanism under diabetic conditions in vivo and in vitro. We found that LRNA9884 was a Smad3-dependent lncRNA highly expressed in the diabetic kidneys and by AGE-stimulated mMCs/mTECs of Smad3 wild-type (WT) but not Smad3 KO. We also found that the expression levels of LRNA9884 were correlated with progressive diabetic kidney disease in db/db mice. Further characterization by obtaining its full-length sequence via RACE showed that LRNA9884 is 508 nucleotides in length and locates within 5033403H07Rik at murine chromosome 5 (Chr5: 52973035–52972344). Since LRNA9884 covered the whole genomic sequence of its host gene, we speculated that LRNA9884 is the completed sequence of 5033403H07Rik. Through the use of CPC and CPAT analysis, LRNA9884 was confirmed and classified as a novel noncoding RNA.

We found that expression of LRNA9884 was significantly induced by the metabolic byproduct and hyperglycemia (AGE and high glucose) under diabetic conditions but not by the profibrotic factor TGF-β1 in the cultured renal cells in vitro. Furthermore, it was markedly upregulated in the db/db mice kidneys but not in the UUO kidneys, as previously reported (13), which implied that LRNA9884 is specifically induced under diabetic conditions. However, LRNA9884 has no regulatory role in glucose metabolism, as evidenced by blocking LRNA9884 without altering the fasting blood glucose levels and body weight in db/db mice. The high disease- and tissue-specificity of LRNA9884 encouraged us to further elucidate its therapeutic potential for T2DN.

First, we identified the biological function of LRNA9884 in the development of T2DN and found that knockdown of LRNA9884 effectively inhibited the AGE-induced inflammation markers, including MCP-1, IL-1β, and TNF-α, on cultured mTECs without influencing on the expression of fibrosis markers such as collagen I and fibronectin in vitro. The inflammatory role for LRNA9884 was further confirmed in vivo, where specifically silencing of LRNA9884 in the kidney of db/db mice was capable of inhibiting renal inflammation. Indeed, renal LRNA9884 was significantly upregulated at the age of 8 weeks before the onset of detectable diabetic renal injury, such as microalbuminuria, suggesting that LRNA9884 may be pathogenic and responsible for the early development of T2DN in db/db mice. It is likely that the early upregulation of LRNA9884 may trigger the early development of low-grade renal inflammation rather than renal fibrosis, because inhibition of LRNA9884 expression suppressed renal inflammation in db/db mice and in cultured mTECs without effect on the expression of profibrotic cytokines that mainly exert a role in renal fibrosis during the late progression of T2DN (3,32–34). Thus, LRNA9884 may function to trigger the development of low-grade renal inflammation in the early development of T2DN.

Second, we revealed that LRNA9884 was tightly regulated by Smad3 under the diabetic condition. We previously reported that AGE is able to activate the Smad-signaling pathway via both TGF-β–dependent and –independent manners (30,31). In this study, we found that the promoter region of LRNA9884 contained a Smad3-binding site and that deletion of Smad3 dramatically blocked the upregulation of LRNA9884 in the diabetic kidney of db/db mice in vivo and in AGE-stimulated MEFs in vitro, suggesting a positive regulatory role of Smad3 in the induction of LRNA9884 during renal inflammation under diabetic conditions. Interestingly, although LRNA9884 was upregulated in the kidney of Smad3-WT db/db mice, it did not participate in the AGE/Smad3-mediated renal fibrosis because knockdown of LRNA9884 did not influence expression of Col-I and fibronectin in vitro. Thus, LRNA9884 is a downstream mediator of Smad3 specific for promoting renal inflammation under diabetic conditions.

Importantly, we detected that the LRNA9884 may mediate renal inflammation in db/db mice via a MCP-1–dependent mechanism. This was supported by the findings that a physical interaction exists between LRNA9884 and the MCP-1 promoter as identified by ChIP and dual-luciferase reporter assays. Moreover, we also demonstrated that silencing of LRNA9884 dramatically suppressed MCP-1 production in the diabetic kidney of db/db mice and in AGE-stimulated mTECs. The diabetic milieu is a low-grade inflammatory disease, where the kidney-infiltrating macrophages largely promote renal inflammation-driven DN (35). Silencing of LRNA9884 reduced the renal MCP-1 level, which largely suppressed the macrophage infiltration and therefore significantly reduced the renal inflammation-driven kidney injured in db/db mice. All of these findings revealed that MCP-1 should be an important downstream target gene of LRNA9884 responsible for the T2DN development. In this study, we identified the main function of LRNA9884 is promoting tissue inflammation. Nevertheless, the biological function and working mechanism of this novel Smad3-dependent lncRNA is still largely unknown, and an unbiased approach to determine whether LRNA9884 alters the gene expression of more than MCP-1 should be further investigated.

Finally, results from this study also provided evidence that targeting LRNA9884 specifically on the diabetic kidney may represent a novel therapy for T2DN. Consistent with previous reports (7,14), by using a noninvasive ultrasound-microbubble-mediated technique, we were able to effectively silence renal LRNA9884-mediated MCP-1 production in db/db mice, which largely blunted the kidney infiltration of inflammatory leukocytes (e.g., macrophages), thereby suppressing renal inflammation and resulting in improved renal function in the db/db mice with diabetic kidney diseases. It should be noted that although we tested the specificity of siLRNA9884 by BLAST analysis, the siLRNA9884 vector used in our in vivo experiments may have had effects on other genes or transcripts than those targeted.

In conclusion, the current study identifies that LRNA9884 is a Smad3-dependent lncRNA and plays a proinflammatory role in T2DN via an MCP-1–dependent mechanism. Our findings suggest that targeting renal LRNA9884 may represent a novel and precision therapy for AGE-driven diabetic kidney diseases.

Funding. This study was supported by Research Grants Council of Hong Kong (GRF 14106518, 14117418, 14117815, 14121816, 14163317, C7018-16G, TRS T12-402/13N), Health and Medical Research Fund (03140486, 14152321), Direct Grant for Research CUHK (2017.002), Lui Che Woo Institute of Innovative Medicine (CARE), and the Major State Basic Research Development Program of China (2012CB517705).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. Y.-y.Z., P.M.-K.T., P.C.-T.T., J.X., X.-R.H., C.Y., R.C.W.M., and H.-Y.L. approved the final version of the manuscript. Y.-y.Z. drafted the original manuscript. Y.-y.Z. and P.M.-K.T. analyzed the data and created the figures. Y.-y.Z., P.M.-K.T., P.C.-T.T., J.X., and X.-R.H. performed experiments. P.M.-K.T. and H.-Y.L. reviewed and edited the writing. X.-R.H., C.Y., and R.C.W.M. supported resources. R.C.W.M. and H.-Y.L. covered the funding. H.-Y.L. designed the study. H.-Y.L. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.