Connectivity mapping of glomerular proteins identifies dimethylaminoparthenolide as a new inhibitor of diabetic kidney disease

Abstract

While blocking the renin angiotensin aldosterone system (RAAS) has been the main therapeutic strategy to control diabetic kidney disease (DKD) for many years, 25-30% of diabetic patients still develop the disease. In the present work we adopted a systems biology strategy to analyze glomerular protein signatures to identify drugs with potential therapeutic properties in DKD acting through a RAAS-independent mechanism. Glomeruli were isolated from wild type and type 1 diabetic (Ins2Akita) mice treated or not with the angiotensin-converting enzyme inhibitor (ACEi) ramipril. Ramipril efficiently reduced the urinary albumin/creatine ratio (ACR) of Ins2Akita mice without modifying DKD-associated renal-injuries. Large scale quantitative proteomics was used to identify the DKD-associated glomerular proteins (DKD-GPs) that were ramipril-insensitive (RI-DKD-GPs). The raw data are publicly available via ProteomeXchange with identifier PXD018728. We then applied an in silico drug repurposing approach using a pattern-matching algorithm (Connectivity Mapping) to compare the RI-DKD-GPs's signature with a collection of thousands of transcriptional signatures of bioactive compounds. The sesquiterpene lactone parthenolide was identified as one of the top compounds predicted to reverse the RI-DKD-GPs's signature. Oral treatment of 2 months old Ins2Akita mice with dimethylaminoparthenolide (DMAPT, a water-soluble analogue of parthenolide) for two months at 10 mg/kg/d by gavage significantly reduced urinary ACR. However, in contrast to ramipril, DMAPT also significantly reduced glomerulosclerosis and tubulointerstitial fibrosis. Using a system biology approach, we identified DMAPT, as a compound with a potential add-on value to standard-of-care ACEi-treatment in DKD.

Figure 1

Influence of Ramipril-treatment on Ins2Akita mice. Urinary ACR (A), glycemia (B) and body weight (C) were measured in 4 months diabetic Ins2Akita that had been treated with (DKD + R) or without ramipril (DKD) for 2 months before sacrifice. Wild type (WT) mice of the same age were analyzed in parallel as a non-diabetic control. Values are mean ± SEM and One-way ANOVA test for multiple comparisons. Comparison with WT: **P < 0.01; ****P < 0.0001. Comparison between DKD and DKD + R: ^#P < 0.05; ^##P < 0.01.

Figure 2

Selection of ramipril-insensitive DKD-associated glomerular proteins (RI-DKD-GPs). (A–C) Volcano-plot representation of the differential abundance of the glomerular proteins (GPs) in Set#1 (DKD vs WT) (A), Set#2 (DKD + R vs DKD) (B) and Set#3 (WT + R vs WT) (C) comparisons (colored circles: proteins representing a significant difference between the indicated comparison (P < 0.05); black circles: non-significant proteins). The Set#1 comparison led to the identification of 666 DKD-GPs out of which RI-DKD-GPs were selected according to their behavior in Set#2 and Set#3 comparisons. (D) Flowchart of RI-DKD-GPs selection (drawn using Microsoft PowerPoint for Mac, Version 16.16.19).

Figure 3

Comparative influence of DMAPT- and Ramipril-treatment on Ins2Akita mice. (A) representative glomerular injury (PAS staining) and interstitial fibrosis (Masson trichrome) in kidneys from 4 month old diabetic Ins2Akita (DKD) that had been treated or not with Ramipril (DKD + R) or DMAPT (DKD + D) for 2 months before sacrifice (scale bar = 50 µm). (B–G) Quantification of ACR (B), glomerular injury (C), glomerular area (D), fibrosis (E), glycemia (F), and body weight (G) in WT (n = 10), DKD (n = 9), DKD + R (n = 9) and DKD + D (n = 9). Values are mean ± SEM and One-way ANOVA test for multiple comparisons. Comparison to wild type mice (WT): *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Comparison of DKD + R or DKD + D to DKD: ^#P < 0.05; ^##P < 0.01; ^###P < 0.001; ^####P < 0.0001. Comparison of DKD + R to DKD + D: ^@@P < 0.01; ^@@@@P < 0.0001.