Compromised DNA repair is responsible for diabetes-associated fibrosis.

Abstract Diabetes‐associated organ fibrosis, marked by elevated cellular senescence, is a growing health concern. Intriguingly, the mechanism underlying this association remained unknown. Moreover, insulin alone can neither reverse organ fibrosis nor the associated secretory phenotype, favoring the exciting notion that thus far unknown mechanisms must be operative. Here, we show that experimental type 1 and type 2 diabetes impairs DNA repair, leading to senescence, inflammatory phenotypes, and ultimately fibrosis. Carbohydrates were found to trigger this cascade by decreasing the NAD +/NADH ratio and NHEJ‐repair in vitro and in diabetes mouse models. Restoring DNA repair by nuclear over‐expression of phosphomimetic RAGE reduces DNA damage, inflammation, and fibrosis, thereby restoring organ function. Our study provides a novel conceptual framework for understanding diabetic fibrosis on the basis of persistent DNA damage signaling and points to unprecedented approaches to restore DNA repair capacity for resolution of fibrosis in patients with diabetes.


Appendix Figures
• Appendix Figure S1

Appendix Tables
• Appendix Table S1 • Appendix Table S2 • Appendix Table S3 • Appendix Table S4 • Appendix Table S5 (A) Representative images of lungs from postnatal pups at the indicated age (in days), stained for the peroxide linked oxidative stress marker DHR-123. The lungs studied in this Figure and in Figure 1c are identical (Scale 10µm).

A B
Appendix Figure S2: Legend Appendix Figure S2: (A) Mean percentage of DSB positive nuclei in camptothecin treated (1μM) murine lung fibroblasts as evidenced by γH2AX positivity after culture in the presence of different reducing carbohydrates at 0, 3 or 24 hours. More than 400 cells were analyzed for each bar (mean ± SD, *: p<0.05, **: p<0.01).
(B) Mean percentage of DSB positive nuclei in etoposide treated (5μM) murine lung fibroblasts as evidenced by γH2AX positivity after culture in the presence of different reducing carbohydrates at 0, 3 or 24 hours. More than 400 cells were analyzed for each bar (mean ± SD, **: p<0.01, ***: p<0.001).
(C) Representative immunoblots from lysates of HEK293 cells cultured in the presence of the reducing sugars as indicated and treated with etoposide (5μM) were probed for the DNA damage marker γH2AX. Histone-H3 used as a loading control.
(D) Representative immunoblots from the lysates of murine podocytes, cultured in the presence of reducing sugars indicated, then treated with etoposide (5μM) and probed for the DNA damage marker γH2AX. Histone-H3 used as a loading control.  The cell extract was immunoprecipitated using anti-PARP, or a non-specific species control antibody (lane 4 and 9) PARP, or its interacting partner DBC-1, was detected using a PARP or a DBC-1 specific antibody. The incubation time with fructose and ribose had to be reduced to 3 days only due to toxicity.
(B) A549 cells cultured in either low (5.5mM) or high glucose (30mM) for 5 days, or with fructose (40mM), or ribose (20mM) for 3 days (graph bar 4). The cell extract was immunoprecipitated using anti-PARP with, or without the presence of NAD+ (500µM) in the wash buffer or a non-specific species control antibody (lane 4 and 9). PARP or its interacting partner DBC-1 was then detected using a PARP, or DBC-1 specific antibody. The experiments presented in this Figure and in Figure 2F were performed simultaneously. The incubation time with fructose and ribose had to be reduced to 3 days only due to toxicity. Representative images of A549 cells cultured in the presence of the indicated glucose concentrations for 5 days alone or with other reducing sugars (here F indicates for fructose at concentration 1mM, and R indicates for Ribose at concentration 100µM) along with NAC (N-Acetyl cysteine at 2mM) and treated with etoposide (5μM for 60 minutes), washed and cultivated in the sugars as indicated for 10days and stained for β-galactosidase, as indicated in methods. Eosin used as a morphology stain (Scale 40µm).
(B) Quantitative analysis of the IL-6 secreted in the culture supernatant of A549 cells treated as described in appendix figure S5A. (mean ± SD, *: p<0.05).
(C) Representative images of murine podocytes cultured in the presence of the indicated glucose concentrations for 5 days alone or with other reducing sugars (here F indicates for fructose at concentration 1mM, and R indicates for Ribose at concentration 100µM) along with NAC (N-Acetyl cysteine at 2mM) and treated with etoposide (5μM for 60 minutes), washed and cultivated in the sugars as indicated for 10days and stained for β-galactosidase as indicated in methods. Eosin used as a morphology stain (Scale 40µm).
(A) Representative immunoblots of lungs harvested from 3-or 6-months non-diabetic control or STZ diabetic mice and probed for γH2AX. Total-H2AX used as loading control.
(B) Representative immunoblots of lungs harvested from 3-or 6-months non-diabetic control or STZ diabetic mice and probed for the DNA damage sensor kinase pATM. Total ATM used as a loading control.
A B Appendix Figure S7: Figure S7: (A) Representative immunoblots of kidneys harvested from 6-months non-diabetic control or STZ-diabetic mice and probed for γH2AX. Histone-H3 used as a loading control.
(B) Representative images of kidneys for nuclei positive for the DNA-DSB marker γH2AX, versus DAPI, in age-matched 3-, or 6-months control, versus STZ induced diabetic mice, as determined by immunofluorescence analysis (scale 10μm).
(C) Quantitative analysis of the peroxide linked oxidative stress marker DHR-123 in age-matched 3-, or 6-month control lungs, or lungs from STZ-diabetic mice. DHR-123 was determined by mean fluorescence intensity of respective lungs.
(D) Representative images of kidneys from age-matched control versus 3-or 6-month STZ-diabetic mice, stainedfor the peroxide linked oxidative stress marker DHR-123, as determined by mean fluorescence intensity of respective kidneys stained with DHR-123 in red. Blue nuclear staining represents DAPI (Scale 10µm).
(E) Representative images of kidneys from age-matched control versus 3-or 6-month STZ-diabetic mice, stained for the SASP marker IL-6, as determined by mean fluorescence intensity of respective kidneys stained with anti-IL-6 in red. Blue nuclear staining represents DAPI (Scale 10µm).
Appendix Figure S8: Legend Appendix Figure S8: (A) Representative images of H&E stained lung sections from age-matched 3-or 6-months control, or STZ-induced diabetic mice, as described in Methods. H&E staining was visualized by bright field and polarized light (Scale 40µm).
(B) Representative images of H&E stained kidney sections from age-matched 3-or 6-months control, or STZ induced diabetic mice, as described in Methods. H&E staining was visualized by bright field and polarized light (Scale 40µm).
(C) Representative images of kidneys from age-matched 3-or 6-month control versus STZ-induced diabetic mice, stained for Masson's Trichrome stain, as described in Methods and visualized by bright field and polarized light, ECM is recognized by its blue staining (Scale 40µm).
A B C Appendix Figure S9: Legend Appendix Figure S9: (A) Representative images of H&E stained lung or kidney sections from 4-months old lean controls (+/db), or age-matched db/db diabetic mice, as described in Methods. H&E staining was visualized by bright field and polarized light (Scale 40µm).
(B) Representative images of lungs and kidneys from 4-months old lean controls (+/db), or age-matched db/db diabetic mice stained with Masson's Trichrome stain, as described in Methods and visualized by bright field and polarized light, ECM is recognized by its blue staining (Scale 40µm).
Correlation between γH2AX with clinical scores of urinary albumin-creatinine-ratio (ACR) and diffusion capacity (DLco). Urinary ACR is given on the right y-axis, DLco is given on the left y-axis. Urinary ACR was log-transformed to achieve normal distribution. Correlation analysis was calculated via Spearman rho and R2-values and p-values were given in the figure.
Appendix Figure S11: (B) Representative images of nuclei positive for the DNA-DSB marker γH2AX in kidneys from 6-month STZ-diabetic mice and control mice, transduced with the indicated AAV2/8 virions as described in Methods. The kidneys were harvested 6-weeks after viral transduction. No primary antibody control served as a negative control of the staining. Mean ± SD of 6 animals per group is shown (Scale 10µm).
(C) Representative images of Masson's Trichrome stain for respective RAGE virions transduced kidneys as described in Methods and visualized by bright field and polarized light, where the accumulated ECM or senescent areas are recognized by its blue staining (Scale 40µm). The dotted lines represent the respective zoomed window.
(D) Quantitative analysis of transduction effects of RAGE (AA or EE) or RFP on pro-inflammatory, fibrotic and SASP gene expression in kidney tissue. The mRNA of pro-inflammatory, fibrotic and SASP cytokines were significantly suppressed in RAGE-EE transduction group. The data was normalized against control value (mean ± SD, *: p<0.05, **: p<0.01).
(E) Quantitative analysis of the urinary creatinine in urine from 6-month STZ-induced diabetic mice, transduced with AAV2/8 as described in Figure EV5A (mean ± SD, *: p<0.05, **: p<0.01, N=8). Data presented in this Figure, as well as in Figure 5 and EV 5, are from the same mice group.
(F) Quantitative analysis of the total urinary volume excreted in 24 hours from 6-month STZ induced diabetic mice, transduced with AAV2/8 as described in Figure EV5A (mean ± SD, **: p<0.01, N=8). Data presented in this Figure, as well as in Figure 5 and EV Figure 5, are from the same mice group.