Metformin inhibits chronic kidney disease‐induced DNA damage and senescence of mesenchymal stem cells

Mesenchymal stem cells (MSCs) are promising source of cell‐based regenerative therapy. In consideration of the risk of allosensitization, autologous MSC‐based therapy is preferred over allogenic transplantation in patients with chronic kidney disease (CKD). However, it remains uncertain whether adequate cell functionality is maintained under uremic conditions. As chronic inflammation and oxidative stress in CKD may lead to the accumulation of senescent cells, we investigated cellular senescence of CKD MSCs and determined the effects of metformin on CKD‐associated cellular senescence in bone marrow MSCs from sham‐operated and subtotal nephrectomized mice and further explored in adipose tissue‐derived MSCs from healthy kidney donors and patients with CKD. CKD MSCs showed reduced proliferation, accelerated senescence, and increased DNA damage as compared to control MSCs. These changes were significantly attenuated following metformin treatment. Lipopolysaccharide and transforming growth factor β1‐treated HK2 cells showed lower tubular expression of proinflammatory and fibrogenesis markers upon co‐culture with metformin‐treated CKD MSCs than with untreated CKD MSCs, suggestive of enhanced paracrine action of CKD MSCs mediated by metformin. In unilateral ureteral obstruction kidneys, metformin‐treated CKD MSCs more effectively attenuated inflammation and fibrosis as compared to untreated CKD MSCs. Thus, metformin preconditioning may exhibit a therapeutic benefit by targeting accelerated senescence of CKD MSCs.

and easily expanded in vitro. MSCs are multipotent and can differentiate into adipocytes, osteocytes, and chondrocytes (Dominici et al., 2006). Further, MSCs may accelerate tissue repair via a paracrine mechanism leading to proangiogenic, anti-inflammatory, and antifibrotic effects than directly differentiating into tissue-specific cells (Gnecchi et al., 2005). Preclinical and clinical trials are in progress to evaluate the beneficial effects of MSC therapy in various kidney disease models (Fazekas & Griffin, 2020;Perico et al., 2018). However, it is imperative to consider that autologous cell transplantation is preferred over allogenic transplantation in patients with CKD who may need kidney transplantation in the future, as the latter may increase the risk of allosensitization. Therefore, adequate functionality of endogenous MSCs is a critical factor for the success of cell therapies in patients with CKD.
We and others have shown that MSCs obtained from patients or rodents with CKD exhibit decreased viability and functional abnormalities (Klinkhammer et al., 2014;Noh et al., 2012Noh et al., , 2014. Moreover, under uremic milieu, the angiogenic potential of transplanted MSCs was significantly reduced at ischemic hind limb (Han et al., 2019;Noh et al., 2014). In the present study, we investigated the effects of metformin on the functional incompetence of CKD MSCs, particularly by focusing on its renoprotective potential. Metformin is a firstline drug for the treatment of type 2 diabetes (Davies et al., 2018).
Aside from its anti-diabetic effect, metformin is known to exert pleiotropic actions, including beneficial effects on the kidney and cardiovascular system and by possibly lowering cancer risk (Foretz et al., 2014). The renal protective effects of metformin have been demonstrated in multiple disease models such as acute kidney injury (AKI) induced by gentamicin (Morales et al., 2010) and CKD using 5/6 nephrectomy (Satriano et al., 2013), unilateral ureteral obstruction (UUO) (Cavaglieri et al., 2015), and adenine diet (Neven et al., 2018). Despite the potential benefits observed in animal studies, the clinical use of metformin is not recommended in patients with severe kidney dysfunction and absolutely contraindicated in those with an estimated glomerular filtration rate (eGFR) <30 ml/min/1.73 m 2 because of the risk of lactic acidosis (Davies et al., 2018). Herein, we assessed the therapeutic effects of metformin on CKD MSCs and the applicability of autologous cell transplantation of metformin-treated CKD MSCs to avoid the risk of adverse effects related to systemic metformin administration.

| Metformin inhibits senescence of CKD MSCs
We isolated MSCs from the bone marrow of sham-operated or subtotal nephrectomized mice. These cells were previously characterized using surface markers (Noh et al., 2012), and their adipogenic or osteogenic differentiation capacity was confirmed as shown in Figure   S1. CKD or metformin did not affect the differentiation capacity.
CKD MSCs more frequently showed flat and enlarged morphologies as compared with spindle-shaped control cells (Figure 1a), and their proliferation was significantly lower, as evident from growth curves, bromodeoxyuridine (BrdU) incorporation assay, and proliferating cell nuclear antigen (PCNA) expression (Figure 1b-d). As chronic inflammation and oxidative stress in CKD may lead to the accumulation of senescent cells, we measured the level of oxidation of guanosine residues by 8-oxo-2′-deoxyguanosine (8-oxo-dG) immunofluorescence staining. As shown in Figure 1e, CKD MSCs had significantly higher expression level of 8-oxo-dG than control MSCs, suggestive of oxidative DNA damage. Other senescence markers such as higher expression of senescence-associated β-galactosidase (SAβ-gal) and p16 Ink4a and lower expression of cyclin D1 and cyclin-dependent kinase 4 (CDK4) also confirmed cellular senescence in CKD MSCs  Figure S2).
We next investigated nuclear factor-kappa B (NF-κB) activation and senescence-associated secretory phenotype (SASP). CKD MSCs showed activation of NF-κB and increased secretion of multiple proinflammatory cytokines (Figure 2a,b, Figure S3a). CKDinduced NF-κB activation, increased expression of proinflammatory SASP factors such as interleukin (IL)-1β, IL-6, tumor necrosis factor (TNF)α, and chemokine (C-X-C motif) ligand 1 (CXCL1), and a profibrotic phenotype with increased expression of collagen I and fibronectin were significantly attenuated by metformin treatment as shown in Figure 2a,c.
To investigate the involvement of the adenosine monophosphateactivated protein kinase (AMPK) pathway in mediating the effect of metformin on NF-κB activation and SASP, we determined the effect of metformin following small-interfering RNA (siRNA)-mediated AMPKα subunit expression knockdown, which abolished the effects of metformin (Figure 2d-f). Of note, AMPK pathway appears to be involved in suppression of NF-κB activity and SASP even in basal state since siRNA-mediated AMPKα knockdown enhanced those factors in the absence of metformin ( Figure S3b,c).

| Metformin treatment improves the renoprotective effects of CKD MSCs in UUO kidneys
To understand the therapeutic potential of metformin-treated CKD MSCs in alleviating structural and functional damage as compared to control or CKD MSCs, we used a mouse model of kidney fibrosis induced by UUO. Transplanted MSCs were occasionally found in injured mouse renal tissues ( Figure S4). The kidneys of C57Bl/6J mice subjected to UUO showed higher mRNA expression of proinflammatory cytokines such as TNFα, MCP-1, and nitric oxide synthase 2 (NOS2) and profibrotic genes such as those encoding fibronectin and collagen I. The transcript levels of these genes were significantly decreased in groups injected with control MSCs or metformin-treated CKD MSCs (Figure 4a). The changes in fibronectin and collagen I/ IV expression and the effect of control and metformin-treated CKD MSCs as compared to that of CKD MSCs were further confirmed by Western blotting. E-cadherin expression showed an opposite trend ( Figure 4b). Histological analysis indicated that the transplantation of control MSCs significantly attenuated tubular atrophy, tubular F I G U R E 2 Metformin inhibits chronic kidney disease (CKD)-induced NF-κB activation and senescence-associated secretory phenotype in adenosine monophosphate-activated protein kinase (AMPK)-dependent manner. Control or CKD MSCs isolated from sham-operated or subtotal nephrectomized mice were treated with 10 μM metformin for 24 h. (a, e) NF-κB activity was measured in nuclear protein extracts, n = 6. (b) Cytokine expression was evaluated in the supernatants of control or CKD MSCs, n = 3. (c, f) Real-time RT-PCR was performed to measure the mRNA levels of interleukin (IL)-1β, IL-6, tumor necrosis factor (TNF)α, chemokine (C-X-C motif) ligand 1 (CXCL1), collagen I, and fibronectin, n = 4-5. (d) Quantitative RT-PCR analysis shows a decreased AMPKα mRNA level following siRNA transfection, n = 8. *p < 0.05, **p < 0.01, and ***p < 0.001 dilation, interstitial fibrosis, and infiltration of CD68-positive cells and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL)-positive apoptotic cells as compared to that of vehicletreated UUO kidneys ( Figure 4c). Transplantation of CKD MSCs, however, showed less significant effects, which were rescued by metformin pretreatment. Deterioration of renal function, as evident from elevated blood urea nitrogen (BUN) and urinary neutrophil gelatinaseassociated lipocalin (NGAL) to creatinine ratio, was alleviated following transplantation of control MSCs but not CKD MSCs; metformin pretreatment showed a tendency toward better preservation of renal function as compared with CKD MSC treatment ( Figure 4d).

| Metformin inhibits CKD-induced senescence and DNA damage in human retroperitoneal adipose tissue-derived MSCs (ADMSCs)
The clinical significance of CKD-associated senescence in MSCs was examined using ADMSCs from healthy kidney donors and patients with CKD. The clinical demographics of the subjects are shown in Table 1. The expression of SAβ-gal and cyclin-dependent kinase inhibitor 2A (CDKN2A) gene encoding p16 Ink4a was higher in the ADMSCs from patients with CKD than in those from control subjects, confirming the CKD-associated cellular senescence in human MSCs (Figure 5a,b). Furthermore, we found a significant correlation between NF-κB activity and creatinine or cystatin C-based eGFR. Thus, reduced renal function correlated with the activation of NF-κB in ADMSCs, as observed in the MSCs from the bone marrow of rodents. In contrast, NF-κB activity was not correlated with age, body mass index, hemoglobin A1c, or C-reactive protein

| DISCUSS ION
In the present study, we demonstrate the increased senescence of CKD MSCs that may contribute to poor regenerative potential and suggest metformin preconditioning as an effective strategy to overcome the senescence-associated barrier to autologous patientderived MSC therapy in CKD. The actions of metformin against senescence were confirmed by multiple observations, including inhibition of SAβ-gal activity, p16 Ink4a expression, and activation of p53 and NF-κB, resulting in a decreased expression of several genes related to SASP. Furthermore, we show that metformin decreased CKD-induced prelamin A accumulation and DNA damage signaling, which may lead to cellular senescence. The renoprotective properties of CKD MSCs were significantly enhanced following metformin pretreatment in co-culture experiments and in UUO model. MSC senescence is associated with impairment in their regenerative potential. Senescent MSCs show limited proliferation, decreased differentiation potential, and impaired migratory and homing ability, and affect neighboring cell and tissue functions by producing factors related to SASP (Gnani et al., 2019;Hickson et al., 2016). As accumulation of oxidative stress is the key mechanism underlying cellular senescence (He & Sharpless, 2017;McHugh & Gil, 2017), it is likely that CKD MSCs exhibit increased senescence. Our data using mouse bone marrow MSCs revealed enhanced oxidative DNA damage and cellular senescence associated with CKD. Furthermore, as in rodents, accelerated senescence was confirmed in ADMSCs isolated from patients with CKD as compared to those from healthy kidney  (Sosa et al., 2018), which have been suggested as potential mechanisms of cardio-renal cross talk (Hénaut et al., 2018;Kaesler et al., 2020) or sarcopenia associated with CKD (Sosa et al., 2018).
Growing evidence has implicated nuclear lamina dysfunction in vascular senescence (Liu et al., 2013;Mattout et al., 2006;Ragnauth et al., 2010). The nuclear lamina is a protein meshwork that underlies the nuclear envelope, plays an important role in maintaining nuclear integrity, and is involved in diverse cellular functions. Lamins are the major components of the nuclear lamina, and mutations in LMNA gene encoding lamins have been implicated in various premature aging disorders such as Hutchinson-Gilford progeria syndrome (HGPS) (Eriksson et al., 2003). Prelamin A, a precursor protein of lamin A, undergoes stepwise post-translational modifications, including farnesylation and proteolytic cleavage mediated by zinc metalloproteinase STE24 homologue, ZMPSTE24, to produce lamin A (Sinensky et al., 1994). Previous studies have shown that the accumulation of prelamin A in the vasculature results in the induction of vascular calcification and senescence (Liu et al., 2013) and that cardiac-specific prelamin A accumulation induces inflammatory car- In line with these observations, our data indicate that CKD MSCs had higher NF-κB activity, and this effect was rescued by metformin treatment, resulting in a decrease in SASP.
Metformin is the most widely used anti-diabetic drug and currently recommended as the first-choice treatment in patients with type 2 diabetes. Aside from anti-hyperglycemic effects, the role of metformin in metabolic and cellular processes such as inflammation, autophagy, oxidative damage, apoptosis, and senescence is now well documented (Foretz et al., 2014;Kulkarni et al., 2020). Although several studies have highlighted the renal protective effects of metformin (Cavaglieri et al., 2015;Kwon et al., 2020;Morales et al., 2010;Neven et al., 2018;Satriano et al., 2013), there is a concern that metformin use is associated with the risk of lactic acidosis in patients with CKD. Therefore, current clinical guidelines (Davies et al., 2018) suggest that metformin therapy should not be initiated in patients with eGFR <45 ml/min/1.73 m 2 and should not be used in patients with eGFR <30 ml/min/1.73 m 2 . Our approach of preconditioning CKD MSCs with metformin before cell transplantation would be a clinically applicable method to enhance the regenerative potential of patient-derived dysfunctional MSCs without the concern for lactic acidosis. The metformin concentration used in this experiment was similar to that detected in the serum of patients with type 2 diabetes treated with metformin (Foretz et al., 2014;Frid et al., 2010). To the best of our knowledge, this is the first study to reveal the role of metformin preconditioning in the functional improvement of CKD MSCs. An important question of whether metformin suppresses the development of senescent cells or reverses senescence should be examined in another independent studies.
While the molecular mechanism of action of metformin remains poorly understood, its primary action is to activate AMPK (Foretz et al., 2014). Here we demonstrated the involvement of AMPK in mediating the effects of metformin, as AMPKα suppression using siRNA significantly blocked the effects of metformin on the activation of NF-κB and upregulation of SASP genes. The mechanisms driving prelamin A accumulation in CKD ADMSCs, which is rescued by metformin treatment, remain undetermined. Deposition of prelamin A has been observed in prematurely aged vessels of children on dialysis   (c) NF-κB activity was measured in nuclear protein extracts. NF-κB activity of ADMSCs correlates with creatinine or cystatin C-based estimated glomerular filtration rate (eGFR) but does not with age, body mass index (BMI), hemoglobin A1c (HbA1c), or C-reactive protein (CRP). (d) Representative immunofluorescence staining for phosphorylation of the histone H2A variant H2AX (γH2AX) and p53-binding protein 1 (53BP1). Scale bar, 20 μm. (f) Representative Western blots show protein levels of prelamin A and phosphor ataxia telangiectasia mutated (ATM)/ATM-and Rad3-related (ATR). n = 11 biologically independent samples for control ADMSCs and n = 8 for CKD ADMSCs. *p < 0.05, **p < 0.01, and ***p < 0.001 warranted to study the mechanism underlying metformin-mediated attenuation of prelamin A accumulation in ADMSCs, including regulatory factors responsible for prelamin A accumulation, such as ZMPSTE24.
A limitation of our study is that we could only test ADMSCs from patients with CKD stage 5 since it was technically difficult to obtain adipose tissues from patients with early stages of CKD.
Future experiments will need to investigate whether ADMSCs from patients with early stages of CKD would be more responsive to metformin. Another limitation is that possible differences in the properties of bone marrow MSCs and ADMSCs should be taken into consideration. Although a previous study reported that no significant differences were observed for growth kinetics, multi-lineage differentiation capacity, and gene transduction efficiency as well as cellular senescence according to their sources (De Ugarte et al., 2003), the effect of CKD on these characteristics has not been directly compared.
In conclusion, our study suggests that metformin preconditioning may exhibit a therapeutic benefit to target accelerated senescence of CKD MSCs. This can be applied to achieve adequate cell functionality for developing patient-derived autologous MSC-based therapeutics in patients with CKD. Further studies to assess the efficacy and feasibility of this approach are important for future investigations.

| Animals
All animal studies were conducted following approval from the

| Isolation of bone marrow MSCs
Primary MSCs were isolated from the pooled bone marrow from three mice and cultured as previously described. In brief, and CD11b (Noh et al., 2012). For adipogenic differentiation, the cells were incubated in α-MEM containing 10% fetal calf serum, 10% horse serum, 100 U/ml penicillin, 100 μg/ml streptomycin, 12 mM l-glutamine, 5 μg/ml insulin, 1 μM dexamethasone, and 0.5 μM 3-isobutyl-1-methylxanthine (Sigma-Aldrich) for 3 weeks. For osteogenic differentiation, the cells were grown in the osteogenic medium (Life Technologies, Grand Island, NY) for 1 week. Adipocytes were stained with Oil Red O (Sigma-Aldrich), and osteocytes were stained with Alizarin Red S (Sigma-Aldrich). Three biologically independent MSC lines per group were obtained, and the cells at equivalent passages (between passages 6 and 11) were used for experiments.

| Isolation of ADMSCs from human subjects
The study protocol was reviewed and approved by the Institutional  Figure S5. Their adipogenic, osteogenic, or chondrogenic differentiation under specific medium conditions was confirmed as described earlier (Yoon et al., 2020). ADMSCs between passages 2 and 4 were used for experiments.

| Growth curve
Mouse bone marrow MSCs were seeded into six-well culture plates at 5 × 10 4 cells/well in complete media (10% FBS in DMEM). After growth for 24, 48, and 72 h, the number of cells was counted.

| BrdU incorporation assay
BrdU incorporation was measured using a cell proliferation enzymelinked immunosorbent assay (ELISA) for BrdU (Roche, Mannheim, Germany) according to the manufacturer's instructions.

| SAβ-gal assay
SAβ-gal assay was performed using a senescence detection kit

| Blood and urine chemistry
BUN and urine creatinine levels were measured using a colorimetric method on a Cobas 8000 analyzer (Roche). Urinary NGAL levels were measured using an ELISA kit (R&D Systems, Minneapolis, MN).

| Co-culture experiment
HK2 cells (American Type Culture Collection, Manassas, VA) were seeded at a density of 1 × 10 6 cells in the lower chamber of a transwell system (Corning Inc., Kennebunk, ME). Control or CKD MSCs with or without metformin (10 μM) were seeded onto transwell inserts at a concentration of 2.5 × 10 5 cells. One day later, the medium was replaced with serum-free media supplemented with or without metformin. After 6 h, HK2 cells were treated with LPS (50 ng/ml, Sigma-Aldrich) or TGF-β1 (2 ng/ml, R&D Systems) for 48 h.

| Statistical analyses
Data are presented as the mean ± standard error (SE) unless otherwise specified. Normality was assessed using GraphPad Prism 5 (GraphPad Software, San Diego, CA) Kolmogorov-Smirnov test.
For experiments with more than 2 conditions, differences between groups were evaluated using one-way analysis of variance followed by Bonferroni post hoc tests. When only 2 groups were compared, either an unpaired two-tailed t test or Mann-Whitney test was used. A value of p < 0.05 was considered statistically significant.

ACK N OWLED G M ENTS
We

CO N FLI C T O F I NTE R E S T
None.

AUTH O R CO NTR I B UTI O N S
H Kim analyzed the data, provided critical discussion, and edited the manuscript. MR Yu performed the experiments. H Lee, SH Kwon,

JS Jeon, and DC Han provided human samples and related data. H
Noh designed the study, analyzed the data, and wrote and revised the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.