Activin A Antagonism with Follistatin Reduces Kidney Fibrosis, Injury, and Cellular Senescence-Associated Inflammation in Murine Diabetic Kidney Disease

Abstract

Key Points:

1. Activin A is implicated in profibrotic and prosenescent kidney injury and correlates with kidney injury markers in animals and humans.

2. Follistatin, through activin A antagonism, reduces senescence burden, macrophage infiltration, and proinflammatory pathway activation in murine diabetic kidney disease.

3. Follistatin and other antagonists of activin A signaling pathways may be promising, novel therapeutics for diabetic kidney disease.

Background: Circulating activin A, an inflammatory mediator implicated in profibrotic kidney injury and cellular senescence-induced adipose tissue dysfunction, is increased in human diabetic kidney disease (DKD) and directly correlates with kidney dysfunction. We tested the hypothesis that activin A increases kidney injury, senescent cell abundance, and macrophage infiltration in DKD and antagonism through follistatin (FS) therapy diminishes these effects.

Methods: An accelerated nephropathy type 2 diabetes (db/db) mouse model was generated by implantation of angiotensin II-loaded osmotic minipumps resulting in increased albuminuria and glomerular and tubular injury. Kidney repair effects of FS (5 µg intraperitoneal; two doses) were assessed through markers of kidney injury, fibrosis, inflammation, cellular senescence, and macrophage infiltration. In vitro studies examined antiactivin effects of FS on high glucose-exposed human monocytes, renal fibroblasts, and renal tubule epithelial cells.

Results: Activin A antagonism with FS reduced senescence (p19), proinflammatory (including senescence-associated secretory phenotype), and profibrotic markers including activin A. FS improved kidney morphology, restored podocyte markers (nephrin and Wilms tumor-1), and reduced kidney injury biomarkers, albuminuria and kidney fibrosis. FS decreased kidney macrophage and leukocyte infiltration and absent in melanoma 2 inflammasome activation. FS seemed to suppress inflammation through the toll-like receptor-4/NF kappa-light-chain-enhancer of activated B cells pathway in vivo further supported in human macrophages in vitro. In addition, FS reduced hyperglycemia-induced renal fibroblast activation and renal tubule epithelial cell senescence in vitro.

Conclusions: Activin A is a mediator of kidney injury through macrophage-associated inflammation in murine DKD. FS acts through senomorphic activities which inhibit profibrotic, proinflammatory, and prosenescence signaling by activin A. Hence, antiactivin targeting may aid in the development of a promising, novel therapeutic for DKD.

Trial registration: ClinicalTrials.gov NCT01519349 NCT02354781 NCT02262455 NCT02848131 NCT03325322.

Keywords: chronic inflammation; diabetic nephropathy; fibronectin; lymphocytes; macrophages; podocyte; renal fibrosis; renal protection; renal proximal tubule cell.

Graphical abstract

Figure 1

Activin A antagonism through FS therapy reduces diabetic kidney injury. (A) Schematic of accelerated diabetic nephropathy murine model. In 10-week-old db/db mice, osmotic minipumps were loaded with AngII (1000 ng/kg per minute) or saline for continuous delivery through end point (day 28; D28). FS (5 µg) or vehicle was intraperitoneally injected on D15 and D18. (B) Representative images and quantification of PAS staining of paraffin embedded kidney sections from control (saline minipump), db+AngII (AngII minipump), and AngII+FS mice (scales bar=60 µm). Quantification of glomerular injury score was assessed on a 0–4 scale (0=normal, 1=mild mesangial matrix expansion, 2=moderate matrix expansion with patent capillaries, 3=severe matrix expansion with segmental capillary loop consolidation, and 4=severe matrix expansion with global capillary loop consolidation). Quantification of tubular injury score was assessed on a 0–5 scale (grade 0=no morphologic deformities, grade 1=<10%, grade 2=<25%, grade 3=<50%, grade 4=<75%, and grade 5=≥75% involved). (C) Measures of 24-hour urinary albumin excretion and plasma creatinine from control, AngII, and AngII+FS mice. (D) qRT-PCR of podocyte (nephrin, WT-1), KIM-1, and kidney disease progression markers (Mac-2 and TNFSR-1). All data are shown as mean±SEM. *P < 0.05,**P < 0.01, ***P < 0.005, and ****P < 0.001. Data were analyzed with the unpaired t test or ANOVA where appropriate. Control versus AngII, AngII versus AngII+FS, control versus AngII+FS. Control (saline minipump), AngII (AngII minipump), and AngII+FS (AngII minipump+FS). AngII, angiotensin II, db/db, diabetes or diabetic mice; FS, follistatin; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; KIM-1, kidney injury molecule 1; MAC-2, galectin-3; PAS, periodic acid–Schiff; qRT-PCR, quantitative RT-PCR; TNFSR-1, soluble TNF receptor-1; WT-1, Wilms tumor-1.

Figure 2

Inhibition of activin A with FS reduces kidney fibrosis. (A) Representative images (scale bar=60 µm) of Masson Trichrome staining and activin A IHC of paraffin embedded kidney sections from control, AngII, and AngII+FS mice. Quantification of fibrosis is shown by trichrome IHC (% blue area) and activin A IHC (% brown area). (B) ELISA and qRT-PCR of kidney tissue for activin A. Activin B gene expression levels were not different between control and AngII groups; however, FS administration tended to result in lower levels (P = 0.07). (C) qRT-PCR of kidney tissue for markers of endogenous FS, extracellular matrix deposition (tenascin-C, collagen 1, fibronectin, and collagen 4). All data are shown as mean±SEM. *P < 0.05, **P < 0.01, and ***P < 0.005. Data were analyzed with an unpaired t test or ANOVA where appropriate. Control versus AngII, AngII versus AngII+FS, control versus AngII+FS. Control (saline minipump), AngII (AngII minipump), and AngII+FS (AngII minipump+FS). IHC, immunohistochemistry.

Figure 3

Inhibition of activin A with FS reduces kidney senescent biomarkers and related SASP. (A) Representative images (scale bar=60 µm) and quantification of IHC (% brown area) of paraffin embedded kidney sections for senescence marker p19 in control, AngII, and AngII+FS mice. (B) qRT-PCR of kidney tissue for markers of senescent cell burden—p19^Cdkn2d, p16^Ink4a, and p21^Cdkn1A. (C) qRT-PCR of kidney tissue for the SASP factors, TNFα, Il-1β, Il-6, and MMP-2. All data are shown as mean±SEM. *P < 0.05, **P < 0.01, and ***P < 0.005. Data were analyzed with an unpaired t test or ANOVA where appropriate. Control versus AngII, AngII versus AngII+FS, control versus AngII+FS. Control (saline minipump), AngII (AngII minipump), and AngII+FS (AngII minipump+FS). MMP-2, matrix metalloproteinase-2; SASP, senescence-associated secretory phenotype.

Figure 4

Activin A inhibition with FS abates kidney macrophage infiltration and alters inflammatory pathways. (A) Representative images (scale bar=60 µm) and quantification of IHC (% brown area) of paraffin embedded kidney sections for macrophage (Mφ) marker F4/80 in control, AngII, and AngII+FS mice. (B) qRT-PCR of kidney tissue for total Mφ burden (F4/80) and MCP-1. (C) M2 Mφ (CD206 and Egr2) and M1 Mφ (CD38 and CD86) abundance. (D) qRT-PCR of kidney tissue for infiltrating leukocytes (CD45), TLR4/NF-κB inflammation pathway and inflammasome receptor, AIM2. All data are shown as mean±SEM. *P < 0.05, **P < 0.01, and ***P < 0.005. Data were analyzed with an unpaired t test or ANOVA where appropriate. Control versus AngII, AngII versus AngII+FS, control versus AngII+FS. Control (saline minipump), AngII (AngII minipump), and AngII+FS (AngII minipump+FS). AIM2, absent in melanoma 2; MCP-1, monocyte chemoattractant protein-1; NF-κB, NF kappa-light-chain-enhancer of activated B cells; TLR4, toll-like receptor-4.

Figure 5

Inhibition of activin A with FS reduces renal fibroblasts activation in vitro. Renal fibroblasts were serum starved for 24 hours and subsequently incubated with activin A (20 ng/ml) and HG (25 mM) with or without exogenous FS (750 ng/ml) for 24 hours. (A) Representative images of immunofluorescence for α-SMA in control, HG+activin A, and HG+activin A+FS. (B) qRT-PCR of α-SMA, activin A, type 1 collagen, fibronectin, and type 4 collagen. (C) To examine the effects of reintroducing activin A, renal fibroblasts were serum starved for 24 hours then incubated with activin A (20 ng/ml), HG (25 mM) with or without exogenous FS (750 ng/ml) for 24 hours. Medium change was then performed, where HG+activin A groups were replaced with HG medium, and groups receiving FS were replaced with either HG alone or HG+activin A replenished medium. Cells were then incubated for 10 hours, and qRT-PCR was performed to test the same markers. All data are shown as mean±SEM. *P < 0.05, **P < 0.01, ***P < 0.005, and ****P < 0.001. Data were analyzed with an unpaired t test or ANOVA where appropriate. α-SMA, α-smooth muscle actin; DAPI, 4′,6-diamidino-2-phenylindole; HG, high glucose; TBP, TATA binding protein.

Figure 6

Activin A antagonism through FS therapy reduces proinflammatory markers and activin A expression in activated human macrophages. PMA-induced human macrophages (U937/PMA) were pretreated with HG (25 mM) and LPS (100 ng/ml) to induce activation for 3 or 6 hours. Cells were subsequently incubated with exogenous FS (500 or 750 ng/ml) for 48 hours. (A) Representative images (scale bar=75 µm) and quantification of immunofluorescence for NF-κB in U937/PMA controls, +HG/LPS, and +HG/LPS+FS. Bottom panel demonstrates nuclear NF-κB staining. (B) qRT-PCR of NF-κB (FS 500 ng/ml), TLR4 (FS 750 ng/ml), AIM2 (FS 500 ng/ml), Il-1β (FS 750 ng/ml), and Il-6 and activin A (FS 750 ng/ml) on U937/PMA pretreated with HG/LPS for 3 or 6 hours. All data are shown as mean±SEM. *P < 0.05, **P < 0.01, ***P < 0.005, and ****P < 0.001. Data were analyzed with an unpaired t test or ANOVA where appropriate. U937/PMA versus +HG/LPS, HG/LPS versus HG/LPS+FS. U937/PMA (U937 treated with PMA overnight), HG/LPS (U937/PMA treated with HG and LPS), HG/LPS+FS (U937/PMA treated with HG/LPS and FS). LPS, lipopolysaccharide; PMA, phorbol 12-myristate 13-acetate.