Loss of TIMP3 underlies diabetic nephropathy via FoxO1/STAT1 interplay

Abstract

ADAM17 and its inhibitor TIMP3 are involved in nephropathy, but their role in diabetic kidney disease (DKD) is unclear. Diabetic Timp3(-/-) mice showed increased albuminuria, increased membrane thickness and mesangial expansion. Microarray profiling uncovered a significant reduction of Foxo1 expression in diabetic Timp3(-/-) mice compared to WT, along with FoxO1 target genes involved in autophagy, while STAT1, a repressor of FoxO1 transcription, was increased. Re-expression of Timp3 in Timp3(-/-) mesangial cells rescued the expression of Foxo1 and its targets, and decreased STAT1 expression to control levels; abolishing STAT1 expression led to a rescue of FoxO1, evoking a role of STAT1 in linking Timp3 deficiency to FoxO1. Studies on kidney biopsies from patients with diabetic nephropathy confirmed a significant reduction in TIMP3, FoxO1 and FoxO1 target genes involved in autophagy compared to controls, while STAT1 expression was strongly increased. Our study suggests that loss of TIMP3 is a hallmark of DKD in human and mouse models and designates TIMP3 as a new possible therapeutic target for diabetic nephropathy.

Figure 1. Expression of TIMPs and ADAMs in diabetic mice

1. Levels of Timp family proteins expression were analysed by real-time PCR on mRNA extracted from kidneys of diabetic mice and normoglycemic littermates (n = 6, Student's t-test).

2. Real-time PCR analysis of Adam10, 15 and 17 expression in diabetic and healthy kidneys (n = 6).

3. Immunohistochemical staining of kidney sections from healthy and diabetic (STZ-treated) WT mice (picture magnification: 250×).

4. Representative Western blot of kidney proteins extracted from healthy and diabetic WT and Timp3^−/− mice showing TIMP3 expression levels. Tubulin was used as loading control. ns, non specific bands. Source data is available for this figure in the Supporting Information.

5. Fluorimetric measurement of ADAM17 proteolytic activity in kidneys of WT and Timp3^−/− diabetic, streptozotocin-treated mice compared to untreated control (n = 6, Student's t-test).

6. Soluble form of TNF-α was measured by ELISA on kidney lysates from normoglycemic and diabetic WT and Timp3^−/−mice (n = 6, Student's t-test).

7. Western blot assessment of membrane-bound and soluble TNF-α in kidneys from healthy and diabetic WT and Timp3^−/− mice. Actin was used as loading control.

8. PAS staining of kidney sections from diabetic and normoglycemic WT and Timp3^−/− mice. Arrow indicates arteriolar hyalinosis at the vascular pole, stars indicate tubular dilation and atrophy and open arrowheads indicate interstitial expansion with fibrosis. Picture magnification is shown on the left. Quantification of mean glomerular area and fractional mesangial area is shown. Statistical significance was evaluated by one-way Anova.

Figure 2. Analysis of kidneys from WT and Timp3^−/− diabetic mice

1. Electron microscopy of kidney sections from WT and Timp3^−/− diabetic mice. Quantification of basal membrane thickness is shown on right panel (Student's t-test).

2. Microalbuminuria was determined in WT and Timp3^−/− control and diabetic mice as ratio between urinary albumin and creatinine concentration measured by ELISA (n = 3, Student's t-test).

3. Representative western blot analysis of lysates from kidneys of healthy and diabetic WT and Timp3^−/− mice. Levels of phosphorylation of Akt (Ser473), ERK (Thr202/Tyr204) and EGFR (Tyr1068) were quantified for STZ-treated animals and expressed as fold increase in the ratio of phospho-protein to total protein (n = 6, Student's t-test). Source data is available for this figure in the Supporting Information.

4. Immunostaining of kidney sections from WT and Timp3^−/− diabetic mice. Quantification of each staining is shown on the bottom, picture magnification on the left (n = 6, Student's t-test). CML, N-carboxymethyl-lysine; NitroTyr, nitrotyrosine.

Figure 3. FoxO1 regulation in healthy and diabetic Timp3^−/− kidneys

1. Real-time PCR on kidney mRNA from healthy and diabetic WT and Timp3^−/− mice showing FoxO1 and FoxO3A levels of expression (n = 6, Student's t-test).

2. Western blot analysis of cytoplasmic, nuclear and total lysates from kidneys of WT and Timp3^−/− diabetic mice. Topoisomerase I (TOPO I) and tubulin were used to normalize levels of nuclear and cytoplasmic proteins, respectively. Densitometric analysis of results are shown on the right (n = 6, Student's t-test). Source data is available for this figure in the Supporting Information.

3. Foxo1 immunostaining on kidney sections from WT and Timp3^−/− diabetic mice. Arrows indicate nuclear staining in STZ-WT mice (left panel) or cytoplasmic staining in STZ-Timp3^−/− mice (right panel). Magnification: 400×.

4. Higher magnification of panel C (1000×) showing FoxO1 staining in renal tubules (inserts are from Supporting Information Figure S13).

5. Real-time PCR analysis of autophagy related FoxO1 target genes in WT and Timp3^−/− diabetic and normoglycemic kidney (n = 6, Student's t-test).

Figure 4. FoxO1 regulation in T3^kd MES13 cell line

1. Analysis of Timp3 mRNA (left panel) and protein (right panel) expression in T3^kd or control MES13 cells grown in basal medium (low glucose) or treated with high glucose (25 mM) for 48 h before harvesting. Mannitol treatment was included in the real time PCR experiment as osmolarity control (n = 3, Student's t-test). Source data is available for this figure in the Supporting Information.

2. Real-time PCR on T3^kd and control MES13 cells treated as in (A), showing FoxO1 and FoxO3A mRNA modulation (n = 3, Student's t-test).

3. Expression of autophagic FoxO1 target genes in T3^kd and control MES13 cells treated with high glucose or mannitol (25 mM), or serum-starved for 24 h before harvesting (n = 3, Student's t-test).

4. Western blot analysis of autophagic FoxO1 target genes on lysates from T3^kd and control MES13 cells treated as in (C). Ratio between LC3II and LC3I form is shown on the left (n = 3, Student's t-test).

5. Immunofluorescence (IF) for LC3A/B (red) in T3^kd and control MES13 cells treated as in (A). Cells were counterstained with DAPI to detect nuclei (blue). Magnification views: 60×. Scale bars: 20 µm.

Figure 5. FoxO1 subcellular distribution in T3^kd MES13 cell line

1. Representative Western blot analysis of total, nuclear and cytoplasmic lysates from T3^kd and MES13 control cells grown in low or high glucose. Topoisomerase I and tubulin were used to normalize levels of nuclear and cytoplasmic proteins, respectively. Densitometric analysis of results is shown on the right (n = 3, Student's t-test). Source data is available for this figure in the Supporting Information.

2. Nuclear FoxO1 protein was immunoprecipitated from T3^kd and control MES13 cells grown in low or high glucose and then subjected to Western blot with an anti-acetyl-lysine antibody. FoxO1 and Topoisomerase I were used as loading control. Densitometric analysis of results is shown on the right (n = 3, Student's t-test).

3. Chromatin immunoprecipitation (ChIP) on T3^kd or control MES13 cells. Primers used to amplify the FoxO1 binding site on LC3a and Atg8 promoters are described in the Materials and Methods section. IgG, negative control; H3 antibody, positive control.

Figure 6. Re-expression of TIMP3 in T3^ko primary mesangial cells rescues FoxO1 effect on autophagy genes

1. Rescue of expression of FoxO1 and its targets in T3^ko primary cells (pMes) following reintroduction of TIMP3. RNA from T3^ko or WT pMes cells infected with GFP or TIMP3 adenovirus and grown either in low or high glucose was used for real-time PCR analysis of FoxO1, Atg5, Atg8, LC3a and Beclin expression (n = 3, Student's t-test).

2. Western blot analysis of T3^ko and WT pMes cells infected and treated as in (A). Samples were run on a 4–12% gradient gel (TIMP3, FoxO1, ATG5, ATG8, Beclin and Actin) or on a 15% gel (LC3 and actin). Ratio between LC3II and LC3I form is shown on the left (n = 3, Student's t-test). Source data is available for this figure in the Supporting Information.

3. Immunofluorescence analysis of LC3A/B (red) in T3^ko pMes cells infected with AdGFP (left) and AdTimp3 (right) in HG condition. Cells were counterstained with DAPI to detect nuclei (blue). Magnification views: 60×. Scale bars: 20 µm.

Figure 7. FoxO1 and STAT1 interplay

1. Real-time PCR of STAT1 expression in WT and Timp3^−/− diabetic and normoglycemic kidney (n = 6, Student's t-test).

2. Representative western blot analysis of STAT1 expression in kidneys from healthy and diabetic WT and Timp3^−/− mice. Quantification for STZ-treated animals is shown on the right (n = 3, Student's t-test). Source data is available for this figure in the Supporting Information.

3. Immunohistochemical staining of diabetic WT and Timp3^−/− kidney sections showing STAT1 expression. Magnification: 250×.

4. Real-time PCR of STAT1 expression in T3^ko or WT pMes cells infected with GFP or TIMP3 encoding adenovirus (n = 3, Student's t-test).

5. Western blot analysis of STAT1 expression in T3^ko or WT pMes cells infected with GFP or TIMP3 encoding adenovirus (representative analysis of three independent experiments with the same results).

6. Western blot analysis on control cells transfected with a pool of control or STAT1 siRNA, confirming inhibition of STAT1 expression.

7. Real-time PCR on T3^ko or control pMes cells transfected with a pool of siRNA directed against STAT1, showing reduction of STAT1 expression (n = 3, Student's t-test).

8. Real-time PCR on T3^ko or WT pMes cells transfected with a pool of siRNA directed against STAT1, showing reciprocal regulation of STAT1 and FoxO1 expression (n = 3, Student's t-test).

Figure 8. TIMP3 and FoxO1 regulation in diabetic patients

1. Real-time PCR on RNA extracted from four controls and five patients affected by Diabetic Nephropathy (Student's t-test).

2. TIMP3 and control staining of kidney sections from healthy and diabetic subjects. Arrows indicate some TIMP3 positive cells. 40× scanning magnification, 10× zoom.

3. Higher magnification of panel (B) (20× zoom).

4. Co-immunofluorescence of control and diabetic human kidney sections showing nuclei (blue) synaptopodin (green) and TIMP3 (red). Merged images are shown on the right panels. Magnification: 63×.

5. Higher magnification of panel (D) (20× zoom). Synaptopodin immunolabelling (green) highlights podocyte foot processes and is characterized by thin linear staining along the surface of the glomerular basement membranes (GBM) of capillary loops. Timp3 immunoreactivity (red) is mainly observed in the podocyte cell body (*) and primary processes (white arrows).