**3. Premature Ageing in CKD**

Because the kidneys are involved in the regulation of many systemic processes, patients with CKD are at a high risk to develop multiple systemic complications, including rheologic, metabolic, immunologic, CV, and other disturbances [42]. Importantly, several of these complications show similarities with the ageing process, such as CV diseases (CVD), sarcopenia, bone disease, as well as frailty [43,44], cognitive dysfunction [44], immune deficiency [12], and finally mortality [1]. Thus, patients with CKD appear to have a highly accelerated ageing process as compared to healthy subjects or non-CKD patients. We present possible causes for the premature ageing process in CKD, focusing on uremic toxins including AGE, the endocrine phosphate-FGF-23–klotho axis, as well as

NRF2 as a master regulator of mitochondrial dysfunction/oxidative stress. Importantly, all pathways are linked to hallmarks of the ageing process [45].

### *3.1. Uremic Toxins and Premature Ageing in CKD*

When renal function declines during the progression of CKD, distinct organic compounds accumulate [46], and signs and symptoms of *uremia* occur. The European Uremic Toxins (EUTox) Work Group has provided a comprehensive overview of circulating solutes in CKD [47] describing a large number of uremic toxins with significantly differing circulating levels in end stage kidney disease (ESKD), which can be sorted into three groups: (1) free water-soluble low-molecular-weight solutes, (2) protein-bound solutes, and (3) middle molecules [47]. For a variety of these uremic toxins, associations with the ageing process have been described. Thus, the protein-bound uremic toxins indoxyl sulfate and p-cresyl sulfate induce oxidative stress by different mechanisms [48]. Using an in vitro approach, indoxyl sulfate attenuates mitochondrial activity in human renal proximal tubule epithelial cells [49] and human umbilical vein endothelial cells [50], thereby affecting the cellular metabolic capacity. Furthermore, indoxyl sulfate induces markers of senescence, including senescence-associated beta-galactosidase (SA-β-gal), through oxidative stress [51]. Indoxyl sulfate and p-cresyl sulfate also increased mitochondrial autophagy ("mitophagy") [52]. Although autophagy in general has been suggested as a compensatory and pro-survival mechanism [53], it could also have adverse effects when mitochondrial mass decreased at a level beyond cellular compensation [52]. Thus, mitochondrial dysfunction is one of the key mechanisms linking uremic toxins with hallmarks of the ageing process [45] through increased oxidative stress [54]. It should be noted that indoxyl sulfate dose-dependently decreases protein levels of klotho in human aortic VSMC in vitro by inducing DNA methylation of the klotho-coding *Kl* gene [55]. In accordance, patients with ESKD have increased promoter hypermethylation of the *Kl* gene [55]. Conversely, the inhibition of uremic toxins should have beneficial effects on ageing markers. As an example drug, the oral sorbent AST-120 adsorbs uremic toxins and their precursors within the gastrointestinal tract [56]. AST-120 dose-dependently decreases circulating indoxyl sulfate [56] and was, therefore, tested in several clinical trials [56]. Results of these trials were heterogenous depending on the prespecified primary endpoints, follow-up period, as well as baseline characteristics of the included subjects. In conclusion, there was no effect of adding AST-120 to standard therapy in patients with moderate to severe CKD on a primary composite renal end point in the EPPIC trials [57]. It should be noted, however, that on the basis of the results of a subgroup analysis in the EPPIC trials, AST-120 delayed the time of reaching the primary renal end point in patients from the United States [58].

AGE are another group of uremic toxins that accumulate in CKD [26]. Besides the well-known associations of AGE levels and CVD (mostly investigated in patients with diabetes mellitus) [59,60], AGE and AGE-detoxifying enzymes including Glo-1 also have direct effects on ageing processes. Thus, the AGE–RAGE axis induces renal cytosolic oxidative stress and mitochondrial dysfunction, as well as inflammation [26,61]. The AGE–RAGE axis further promotes premature senescence in proximal tubular epithelial cells via endoplasmatic reticulum stress and p21 activation in an in vivo and in vitro approach [62]. Moreover, AGE-induced endoplasmatic reticulum stress also increases p16 protein expression in tubular epithelial cells in vitro in a dose- and time-dependent manner through the activating transcription factor 4 [63]. Conversely, reduced renal senescence in tubule iss observed in rodents with a transgenic overexpression of the AGE-detoxifying enzyme Glo-1 in vivo and in vitro [54]. More recently, AGE have been shown to induce renal senescence in mesangial cells via the RAGE/Signal transducer and activator of transcription 5 (STAT5) pathway and inhibiting autophagy [64].

Taken together, uremic toxins, including AGE, have distinct and adverse effects on the pro- and anti-oxidative milieu, mitochondrial function, inflammation, and finally cellular senescence; all of which are hallmarks of the ageing process [45] (Figure 3).

**Figure 3.** Mechanisms by which uremic toxins and advanced glycation end products (AGE) induce pro-ageing effects. Both uremic toxins and AGE have adverse effects on mitochondrial function including formation of reactive oxygen species (ROS) but also structural disturbances. They impair anti-ageing defenses of the organism, and induce inflammation, as well as cellular senescence. All effects result in an adverse, pro-ageing milieu caused by uremic toxins and AGE. Light blue arrows: inducers; red arrows: inhibitors. ATF4: activating transcription factor 4, ROS: reactive oxygen species, SA-β-gal: senescence-associated beta-galactosidase.

### *3.2. The Endocrine Phosphate-FGF-23–Klotho Axis and Premature Ageing in CKD*

A clear negative association of serum Pi levels with life span has been reported in mammals [8]. Phosphate homeostasis is regulated by a distinct endocrine network involving vitamin D, parathyroid hormone (PTH), klotho, and FGF-23 (Figure 1). In more detail, Pi originating from dietary intake is absorbed by the smallintestine by an active transcellular transport or a paracellular pathway [65]. In hypophosphatemia, either normal or increased levels of the active metabolite 1,25(OH)2 vitamin D3 are observed in humans [66], and 1,25(OH)2 vitamin D3 increases the rate of intestinal Pi resorption [67]. Furthermore, the parathyroid glands are affected by elevated serum Pi levels and release PTH, which enhances urinary Pi excretion by PTH-induced removal of the renal sodium Pi cotransporters from the apical membrane [68]. Bone-derived FGF-23 limits hyperphosphatemia by inducing phosphaturia and suppressing vitamin D secretion [69]. FGF-23 signals through endocrine FGF receptor complexes [70] and also associates with distinct metabolic components [71]. FGF-23 requires klotho protein as a co-receptor [70]. The full length klotho protein, which is predominantly expressed in the proximal and distal tubules of the kidneys [72], is cleaved and circulates as a soluble klotho fragment serving as a surrogate marker of renal klotho expression [70]. In the circulation, calciprotein particles (CPP) "capture" calcium (Ca) and Pi with the help of the adipokine fetuin A and prevent the precipitation of Ca-Pi [70,73].

Phosphate induces EVA via different mechanisms, and higher serum Pi concentrations, even if they are within the normal range, have been found to be associated with microvascular dysfunction in a Dutch, population-based cohort study [74] (Figure 2). In human aortic VSMC, Pi treatment arrests

the cell cycle [75], mediating cellular senescence. In vitro, Pi further induces vascular calcification and production of ROS [76]. Mediated by Ca-Pi crystals and a disturbed microRNA homeostasis, Pi has additional indirect and negative effects on EVA [77]. Importantly, these in vitro data on EVA are supported by a Scottish general population study showing that Pi is linked to markers of biological age, that is, reduced telomere length, DNA methylation content, and chronological age [78].

Adverse effects of Pi on EVA are at least in part mediated by the Wnt/beta-catenin pathway [79,80]. Thus, high Pi activates the Wnt/beta-catenin pathway in VSMC in vitro, promoting VSMC calcification and osteogenic transition [79]. Conversely, shRNA-mediated knockdown of beta-catenin in CKD rats on a high-phosphate diet reduced vascular calcification [79]. Interestingly, klotho is an inhibitor of the Wnt/beta-catenin pathway [70], ameliorating high Pi-induced VSMC calcification [81], as well as inhibiting osteogenic transition of these cells [82,83]. Taken together, Pi induces EVA at least in part by the Wnt/beta-catenin pathway, and different treatment approaches potentially mitigate the adverse effects of Pi on EVA, including klotho.

Besides preventing the precipitation of Ca-Pi (Figure 2), CPP can induce pro-inflammatory effects in VSMC in vitro [84], further linking inflammation and EVA in CKD. Furthermore, CPP differ between healthy subjects and patients with ESKD and can, therefore, promote widespread calcification in CKD [85].

Because of the above-mentioned evidence of a direct association between Pi and premature ageing in CKD, non-pharmaceutical prevention strategies could potentially reduce the burden of disease, including changes in lifestyle. However, there is a lack of studies linking diets that lower Pi levels to a beneficial clinical outcome [86,87]. Furthermore, whereas low-Pi diets can reduce FGF-23 levels in CKD [88], there was no dose-dependent effect of Pi lowering on FGF-23 reduction when low- and very low phosphate diets were compared in a recent short-term, randomized, crossover trial in 35 ESKD patients [89]. Therefore, non-pharmacological lifestyle-changing prevention strategies might provide a benefit on components of the ageing process, but randomized controlled trials are needed to prove these associations.

Vitamin D has also been linked to distinct components of the ageing process (Figure 2). The negative acute phase reactant vitamin D [90] mitigates oxidative stress at least in part through increased expression of anti-oxidant regulators, including NRF2 and klotho [91–93]. Thus, calcitriol increased renal and circulating klotho expression, whereas it reduced markers of oxidative stress in uninephrectomized, spontaneously hypertensive rats [94]. These data have been confirmed in a recent randomized controlled trial showing increased circulating klotho and total antioxidant capacity after 12 weeks of cholecalciferol treatment compared to the placebo group [95]. Human ex vivo studies indicate that vitamin D exerts anti-inflammatory effects on uremic lymphocytes under renal replacement therapy (RRT) [96]. Meta-analyses demonstrating reduced circulating markers of inflammation in mostly non-CKD subjects on vitamin D supplementation [97,98], further supporting a significant role of vitamin D on the hallmarks of ageing. However, these associations were not seen in CKD patients on vitamin D supplementation [99]. Furthermore, two systematic reviews and meta-analyses did not find strong evidence for a beneficial effect of vitamin D supplementation on arterial stiffness [100] and endothelial function [101], two major components of EVA. Thus, future studies need to validate whether vitamin D has direct effects on premature ageing in CKD or whether its effects are mediated by NRF2, klotho, and others.

FGF-23 has been associated with an adverse outcome in CKD in large studies [102]. However, only a few experimental studies have investigated causal effects of FGF-23 on inflammation and premature ageing as hallmarks in the pathogenesis of CKD and its complications. Thus, FGF-23 induces hepatic inflammation in CKD [103]. Furthermore, some [104,105] but not all [106] studies sugges<sup>t</sup> that FGF-23 might induce EVA. Taken together, data for direct FGF-23 effects on *renal* inflammation and premature ageing in CKD are limited (Figure 2).

Simic et al. [107] recently showed that kidney-derived glycerol-3-phosphate in the renal vein correlated with circulating FGF-23 levels, and they further demonstrated a novel kidney–bone axis by which renal glycerol-3-phosphate increases bone-secreted FGF-23. Future studies need to investigate whether renal glycerol-3-phosphate could also be a potential target for preventing premature ageing.

With respect to premature ageing, klotho (Figure 2) is clearly the most interesting member of this endocrine axis, as Kuro-o et al. [70] described a syndrome that resembles human ageing in mice that have a defect in the klotho-coding *kl* gene. When renal function declines, klotho levels decrease in humans and in rodents with CKD [70]. Klotho-overexpressing mice with CKD have an increased phosphaturia, as well as an improved renal function and EVA phenotype [72]. In mice with an ICR-derived glomerulonephritis, overexpression of the *kl* gene reduces blood urea nitrogen levels, proteinuria, renal SA-β-gal activity, and the development of glomerular and tubulointerstitial changes [108]. Mechanistically, klotho-overexpressing mice show an improved mitochondrial function, as well as a reduced mitochondrial DNA damage and oxidative stress, in the kidneys [108].

Interestingly, both high Pi and inflammation decrease renal *Kl* mRNA expression in vitro [109]. Thus, 5/6 nephrectomized rats on a 0.4% phosphorus diet showed reduced renal klotho protein expression compared to sham-operated animals [109]. Lipopolysaccharide-induced inflammation further reduced klotho expression in all groups [109]. In vivo and ex vivo pharmacological inhibition of NF-κB and Wnt recovered reduced klotho expression after lipopolysaccharide treatment [109] further suggesting that inflammation and the Wnt/beta-catenin pathway are crucially involved in the downregulation of renal klotho in CKD. The pro-fibrotic transforming growth factor (TGF)-β, which is upregulated in CKD, is another mediator of reduced renal klotho in CKD, facilitating adverse Wnt/beta-catenin activation in the kidney [110].

Besides these in vivo data in renal tissue, klotho exerts beneficial effects on the endothelium including attenuation of inflammatory and adhesion molecules, as well as augmented NO production and vasorelaxation [111]. Moreover, klotho-deficient mice show increased vascular calcification [112,113], further supporting a beneficial role of klotho on EVA. In addition, the above-described role of uremic toxins in reducing klotho expression further strengthens these effects in CKD. Finally and most importantly, because overexpression of klotho in mice extends their lifespan [70], klotho is one of the most promising anti-ageing targets in CKD.

### *3.3. Mitochondrial Dysfunction*/*Oxidative Stress and Premature Ageing in CKD*

Patients with CKD show increased oxidative stress on the basis of an over-production of ROS, as well as decreased anti-oxidant defenses [29,114]. In general, ROS are generated from the reduction of oxygen and can induce oxidation of important macromolecules, including proteins, lipids, carbohydrates, and DNA [29,114], thereby exerting toxic effects on mitochondria [115]. Thus, sufficient anti-oxidative mechanisms are necessary to counteract the excessive ROS formation in cells. One of the most interesting anti-oxidative targets interfering with ageing is the transcription factor NRF2, which is a key regulator of anti-oxidative enzymes. When oxidative stress occurs, NRF2 translocates to the nucleus and activates >300 distinct genes that have an anti-oxidant response element in their promoter regions [26,116]. Mitochondrial dysfunction in CKD is associated with low NRF2 expression in human skeletal muscle [117]. NRF2 is also negatively associated with different other age-related lifestyle diseases [9] and rare progeroid syndromes [7]. Thus, patients with Hutchinson–Gilford progeria syndrome (HGPS) show an impaired activity of the NRF2 pathway, and reactivation of NRF2 in ex vivo cells of patients with HGPS can reverse ageing defects [118]. Mechanistically, NRF2 knockdown in wildtype fibroblasts increases oxidative stress and recapitulates cellular senescence defects of HGPS [118]. Furthermore, caveolin-induced NRF2 inhibition induces senescence in murine adipocytes in vitro and, conversely, less inhibition of NRF2-dependent anti-oxidant signaling results in decreased senescence, as assessed by SA-β-gal and p21Waf1/Cip1 [119]. Molecular markers of vascular senescence, that is, p16*INK4a* and p21Waf1/Cip1, as well as cytokines of the SASP, are increased in NRF2-deficient mice compared to control mice [120]. Klotho also exerts some of its beneficial effects on EVA by activating the NRF2 pathway [121]. Furthermore, because testosterone ameliorates age-related renal fibrosis in mice via activation of NRF2 signaling [122], the effects of sex hormones and their supplementation in CKD

also need attention. In ESKD, reduced NRF2 expression in peripheral blood mononuclear cells has been reported in addition to an up-regulation of pro-inflammatory NF-κB [27]. Furthermore, uremic toxins correlate positively with NF-κB expression and negatively with NRF2 expression in peripheral blood mononuclear cells [123]. Accordingly, rats undergoing 5/6 nephrectomy also show an impaired activation of NRF2 in the kidneys [124]. Conversely, the NRF2 agonist bardoxolone reduces tubular cell mitochondrial damage and improves redox balance and mitochondrial function in a CKD mouse model [125]. Taken together, mitochondrial dysfunction and oxidative stress promote premature ageing. Because the protective NRF2 pathway is attenuated in CKD and other burden of lifestyle diseases associated with ageing, this o ffers the potential for future NRF2 agonist-based treatment.

### **4. Secondary Premature Ageing by CKD-Causing Diseases and CKD Treatment**

As described above, many facets of CKD contribute to premature ageing. However, distinct causes of CKD contribute to premature ageing, irrespective of renal function, and are summarized below.

### *4.1. Glomerular Diseases Excluding Diabetic Kidney Disease (DKD)*

In contrast to other CKD causes, the incidence of IgA nephropathy (IgAN) does not increase with age and is, therefore, higher in children and young adults as compared to elderly subjects [126,127]. In Asian cohorts of IgAN, renal biopsies show higher expressions of p16*INK4* [128], p21 [129], and reduced klotho [128] compared to controls. Furthermore, <sup>p</sup>16*INK4*, as well as klotho, correlated with IgAN-related renal fibrosis [128]. Because other glomerular diseases, including focal segmental glomerulosclerosis and minimal change disease, are also associated with an increased renal expression of the senescence marker p16INK4A [127], glomerular diseases seem to have direct relations to premature ageing. Patients with Fabry disease have shorter leukocyte telomere length compared to controls [130]. Thus, this rare renal disease could also serve as a model of premature ageing [131].

### *4.2. Polycystic Kidney Disease as a Model of Adverse Anti-Senescence*

Patients with autosomal dominant polycystic kidney disease (ADPKD) often progress to CKD and ESKD [132]. Interestingly and in contrast to most other underlying CKD causes, the senescence marker p21Waf1/Cip1 is decreased in both human and rodent ADPKD, suggesting that an anti-senescent milieu promotes cys<sup>t</sup> formation [127]. In accordance with this hypothesis, the cyclin-dependent kinase inhibitor roscovitine reduces cys<sup>t</sup> formation in mice with PKD [127]. In contrast, nephronophthisis, another cystic kidney disease, is characterized by cellular senescence [133]. Furthermore, senolytic treatment in mice with nephronophthisis type 7 improves cystic area, inflammation, and renal fibrosis [134]. These contradictory findings indicate that senescence must be viewed as disease-specific and in the context of a spectrum ranging from beneficial to adverse e ffects.
