MicroRNA-21 in glomerular injury

Abstract

TGF-β(1) is a pleotropic growth factor that mediates glomerulosclerosis and podocyte apoptosis, hallmarks of glomerular diseases. The expression of microRNA-21 (miR-21) is regulated by TGF-β(1), and miR-21 inhibits apoptosis in cancer cells. TGF-β(1)-transgenic mice exhibit accelerated podocyte loss and glomerulosclerosis. We determined that miR-21 expression increases rapidly in cultured murine podocytes after exposure to TGF-β(1) and is higher in kidneys of TGF-β(1)-transgenic mice than wild-type mice. miR-21-deficient TGF-β(1)-transgenic mice showed increased proteinuria and glomerular extracellular matrix deposition and fewer podocytes per glomerular tuft compared with miR-21 wild-type TGF-β(1)-transgenic littermates. Similarly, miR-21 expression was increased in streptozotocin-induced diabetic mice, and loss of miR-21 in these mice was associated with increased albuminuria, podocyte depletion, and mesangial expansion. In cultured podocytes, inhibition of miR-21 was accompanied by increases in the rate of cell death, TGF-β/Smad3-signaling activity, and expression of known proapoptotic miR-21 target genes p53, Pdcd4, Smad7, Tgfbr2, and Timp3. In American-Indian patients with diabetic nephropathy (n=48), albumin-to-creatinine ratio was positively associated with miR-21 expression in glomerular fractions (r=0.6; P<0.001) but not tubulointerstitial fractions (P=0.80). These findings suggest that miR-21 ameliorates TGF-β(1) and hyperglycemia-induced glomerular injury through repression of proapoptotic signals, thereby inhibiting podocyte loss. This finding is in contrast to observations in murine models of tubulointerstitial kidney injury but consistent with findings in cancer models. The aggravation of glomerular disease in miR-21-deficient mice and the positive association with albumin-to-creatinine ratio in patients with diabetic nephropathy support miR-21 as a feedback inhibitor of TGF-β signaling and functions.

Keywords: TGF-β; cell survival; diabetic glomerulopathy; molecular biology; podocyte.

Figure 1.

miR-21 expression levels are induced by TGF-β₁ in podocytes and Tgfb1-TG mice. (A) Levels of mature miR-21 and miR-21 pricursor increased in WT mouse podocytes up to 2.5-fold after exposure to TGF-β₁, which was completely abolished in Smad2/3-deficient podocytes (n=3 independent experiments; measured by qRT-PCR). *P<0.05 compared with WT at the same time point. (B) miR-21 pricursor levels increased 4-fold in the same cell samples (measured by qRT-PCR). *P<0.05 compared with WT at the same time point. (C) In kidneys of Tgfb1-TG mice, miR-21 levels were significantly higher in kidneys of Tgfb1-TG mice with severe phenotype (TG/severe; n=8) and mild phenotype (TG/mild; n=6) compared with WT (n=5) inferred from histology score. **P<0.001 compared with WT. (D) Similarly, TGF-β₁ mRNA levels were significantly higher in TG/severe (n=3) compared with WT (n=3) in kidneys of Tgfb1-TG mice. *P<0.05.

Figure 2.

miR-21 increases cell survival in cultured podocytes. (A and B) The number of apoptotic murine podocytes after serum withdrawal (from 20% to 2.5% FBS) increased after transfection with miR-21 inhibitor compared with scrambled oligonucleotides using Annexin V-FITC and propidium iodide double labeling and flow cytometry (n=3, 21% versus 6.9%). *P<0.05. (C) Similarly, transfection with miR-21 mimic or inhibitor decreased or increased mitochondrial membrane potential in murine podocytes after exposure to TGF-β₁ (TGF-β₁: 5 ng/ml for 24 hours; n=3). Loss of mitochondrial membrane indicates activation of the apoptotic cascades. *P<0.05.

Figure 3.

Loss of miR-21 is associated with accelerated glomerular damage in Tgfb1-TG mice. (A) In TG/miR-21-KO mice, urine protein-to-creatinine ratio was strongly increased in 8 of 19 TG/miR-21-KO but 0 TG/miR-21-WT mice (n=12) at 4 weeks of age, consistent with previous reports showing heterogeneity in this mouse model.^, (B) The urine of TG/miR-21-KO mice with severe proteinuria showed strongly increased amounts of albumin (Coomassie blue stain; normalized by loading 2 μg creatinine equivalents of urine for each sample). (C) Histologic examination of kidney tissue by periodic acid–Schiff (PAS) staining showed increased deposition of PAS-positive material and decreased cellularity in glomeruli of 4-week-old TG/miR-21-KO mice compared with TG/miR-21-WT littermates. Picrosirius red staining showed increased signal intensity, with a diffuse and nodular pattern in glomeruli of TG/miR-21-KO compared with TG/miR-21-WT littermates. Consistent with increased ECM deposition detected by picrosirius red staining, collagen III deposition was increased in glomeruli of TG/miR-21-KO detected by immunohistochemistry staining. Interestingly, the tubulointerstitial area appeared normal by PAS, picrosirius red, and collagen III staining independent of genotype. (D) Quantitative image analysis of picrosirius red staining intensity showed significantly higher staining intensity in glomeruli of TG/miR-21-KO (n=7) versus TG/miR-21-WT littermates (n=9), whereas no significant difference was noted in the tubulointerstitial area (P=0.08). *P<0.05. (E) Col1a1, Col4a1, Col6a1, and Fibronectin mRNA expressions determined by qRT-PCR were strongly increased in isolated glomeruli of TG/miR 21-KO mice (n=3) compared with WT littermates (n=4). *P<0.05. (F) Quantitative image analysis of fibronectin immunohistochemistry staining showed significantly higher staining intensity in glomeruli of TG/miR-21-KO (n=3) versus TG/miR-21-WT littermates (n=3), whereas no significant difference was noted in the tubulointerstitial area (P=0.40). *P<0.05.

Figure 4.

miR-21 deficiency is associated with increased apoptosis of glomerular cells and loss of podocytes in Tgfb1-TG mice. (A) The numbers of nuclei (4′,6-diamidino-2-phenylindole [DAPI] -positive; blue) and WT1-positive nuclei (podocytes; red [pink in merge]) were not different at 2 weeks of age between TG/miR-21-WT (n=4) and TG/miR-21-KO mice (n=5). (B) At 4 weeks of age, the numbers of nuclei and WT1-positive nuclei per glomerular tuft were significantly decreased in TG/miR-21-KO mice (n=4) compared with TG/miR-21-WT littermates (n=4). Numbers of nuclei were normalized to TG/miR-21-WT mice and presented as a percentage. *P<0.05. (C) Immunohistochemistry for cleaved caspase-3 showed an increased number of positive-stained cells in glomeruli of TG/miR-21-KO mice (n=7) compared with TG/miR-21-WT mice (n=4) littermates at 2 weeks of age determined by manual counting of positive cells in 100 glomerular tufts per mouse. *P<0.05.

Figure 5.

miR-21–deficient diabetic mice exhibit increased proteinuria, decreased kidney function, and decreased podocyte number. (A) miR-21 mutant mice developed hyperglycemia within 2 weeks of STZ treatment, and blood glucose level, were not different among different genotypes (n=7 for each genotype). (B) All STZ-treated miR-21 mutant mice developed albuminuria 4 weeks after STZ treatment, but miR-21-KO mice develop significantly increased albuminuria compared with miR-21-heterozygous (HET) and miR-21-WT mice after 8 weeks of STZ treatment. Diabetic miR-21-HET mice exhibit significantly increased albuminuria 16 weeks after STZ treatment versus WT mice (n=5–8 in each genotype). The miR-21 expression is reduced by approximately 50% in kidneys of unchallenged miR-21–HET mice (data not shown). *P<0.05 compared with miR-21-WT mice. (C) Diabetic miR-21-KO mice had higher serum creatinine level versus WT mice at 20 weeks after STZ treatment (n=7 in each genotype). *P<0.05. (D) Increased deposition of periodic acid–Schiff (PAS) -positive material was detected in glomeruli of diabetic miR-21-KO mice compared with treated miR-21-WT mice 20 weeks after STZ treatment (PAS staining). The calculated mesangial index showed significant mesangial expansion in glomeruli of diabetic miR-21-KO mice compared with diabetic miR-21-WT littermates (n=5 for each genotype). *P<0.05. (E) The number of WT1-positive nuclei (podocytes) per glomerular tuft was significantly decreased in diabetic miR-21-KO mice compared with diabetic miR-21-WT littermates 20 weeks after STZ treatment (n=5; normalized to diabetic miR-21-WT mice and presented as a percentage). *P<0.05. (F) In microdissected glomeruli of American-Indian patients with DN, miR-21 levels (log₂-transformed sequence reads by RNA sequencing) were significantly higher in patients with macroalbuminuria (n=7) compared with patients with normoalbuminuria (n=19). The levels of miR-21 in patients with microalbuminuria (n=22) were between the levels in patients with normoalbuminuria and macroalbuminuria. *P<0.05.

Figure 6.

miR-21 represses the expression of multiple transcripts that regulate apoptosis. (A) Predicted target sites of miR-21 in 3′UTRs of Tgfbr2, Tgfbi, Smad7, Pdcd4, Timp3, and Col4a1 (www.targetscan.org). (B) Smad3 phosphorylation was increased in cultured podocytes transfected with miR-21 inhibitor compared with podocytes transfected with control oligonucleotide 4 and 24 hours after exposure to TGF-β₁ (5 ng/ml; n=3; podocytes were transfected with oligonucleotides 20 hours before TGF-β₁ treatment). Protein loading control: Glyceraldehyde 3-phosphate dehydrogenase (Gapdh). *P<0.05. (C) PDCD4 protein level was decreased in podocytes 24 hours after TGF-β₁ treatment compared with buffer (n=4). PDCD4 protein expression was increased in podocytes transfected with miR-21 inhibitor compared with control oligonucleotide (n=3). Exposure of podocytes transfected with miR-21 inhibitor to TGF-β₁ for 24 hours resulted in increased PDCD4 protein levels compared with podocytes transfected with control oligonucleotide and treated with TGF-β₁ (n=4). *P<0.05. (D) Smad7 and Timp3 mRNA were increased in miR-21 inhibitor-transfected podocytes 24 hours after exposure to TGF-β₁ compared with podocytes transfected with control oligonucleotide and treated with TGF-β₁. Timp3 was also increased in miR-21 inhibitor-transfected podocytes without TGF-β₁ treatment (n=3). *P<0.05. (E) TG/miR-21-KO mice (n=3) exhibit higher glomerular mRNA expression of Tgfbr2, Tgfbi, Smad7, Tp53, and Timp3 compared with TG/miR-21-WT mice (n=4) assayed by qRT-PCR. The level of Ras homolog gene family member B (RhoB), also a predicted target of miR-21, did not differ between TG/miR-21-WT and TG/miR-21-KO mice (P=0.90). *P<0.05.

Figure 7.

miR-21 targets genes expression in glomeruli of patients with DN. The expressions of TGFBR2 and TIMP3, known direct targets of miR-21, were significantly negatively correlated with miR-21 expression (TGFBR2: r=−0.4, P=0.006; TIMP3: r=−0.33, P=0.003; no. of patients=48; TGFBR2 and TIMP3 expressions determined by Affymetrix gene array).