**4. Gut–Bone–Vascular Axis in CKD**

Acknowledging that the gu<sup>t</sup> microbiome is a key regulator of bone [62–64] and cardiovascular [65–67] health, gu<sup>t</sup> dysbiosis may be hypothesized to be involved in the pathogenesis of the bone–vascular axis. The present review discusses mechanisms by which gu<sup>t</sup> dysbiosis may contribute to vascular calcification and bone demineralization in the setting of CKD. We herein will separately discuss the role of increased protein fermentation, decreased carbohydrate fermentation, vitamin K deficiency, and gut-derived inflammation (Figure 1).

**Figure 1.** The kidney–gut–bone–vascular axis. Chronic kidney disease is associated with gu<sup>t</sup> dysbiosis, characterized by a metabolic shift towards a predominantly proteolytic fermentation pattern and a leaky gut. Gut dysbiosis may induce bone loss and vascular calcification and as such may play a pathogenic role in the bone–vascular axis in CKD. Underlying pathophysiological mechanisms include increased exposure to protein fermentation metabolites (such as p-cresyl sulfate (PCS) and indoxyl sulfate (IndS)), a leaky gu<sup>t</sup> contributing to inflammation, and deficiency of vitamin K and short-chain fatty acids (SCFAs).

### **5. Role of Increased Protein Fermentation in the Bone–Vascular Axis**

End products of protein fermentation such as phenols and indoles are largely [68] transported across the colonic epithelium via active and passive transport mechanisms [57,69] and subsequently metabolized by phase 1 and 2 reactions (e.g., towards *p*-cresyl sulfate (PCS) and indoxyl sulfate (IndS)) in the colonic epithelium and liver before entering the systemic circulation 70. Whether CKD affects transport kinetics and metabolism of protein fermentation metabolites remains to be investigated. Protein fermentation metabolites are cleared from the circulation by the kidneys, mainly by tubular secretion, since most are strongly protein-bound [70]. Plasma concentrations of PCS and IndS increase along the progression of CKD to reach levels in patients with ESKD being 10- to 50-fold higher than in healthy controls. These high levels reflect both an increased intestinal production and absorption and a decreased renal clearance [71]. At uremic concentrations, PCS and IndS may disturb several biological processes and confer direct and indirect toxicity in various cells and tissues, at least partly by generating intracellular oxidative stress [72].

Experimental studies revealed that IndS and PCS may promote vascular calcification through various mechanisms [73–75]. These mechanisms include (a) increased shedding of endothelial microparticles [76,77], (b) impaired autophagic flux in endothelial cells [78], (c) downregulation of MiR-29b [79], and (d) suppression of the nuclear factor erythroid 2-related factor 2 (NRF2), a master regulator of cellular antioxidant activity [80]. Dahl salt-sensitive hypertensive IndS-administered rats presented aortic calcification and upregulation of osteogenic genes when compared to control rats, indicating a pro-calcifying role of IndS in an in vivo animal model [81]. In a subsequent experiment by the same group, Dahl salt-sensitive hypertensive IndS-administered rats presented markers of senescence in the area of aortic calcification [82]. Recently, Opdebeeck et al. reported that both IndS and PCS independently promote vascular calcification in the adenine-induced CKD rat model. This was demonstrated in the aorta, as well as in peripheral arteries. Uremic toxin-induced vascular calcification was associated with the activation of inflammation and coagulation pathways [83].

In line with these experimental data, the circulating levels of PCS and IndS have been repeatedly associated with cardiovascular morbidity (including arterial sti ffness, vascular calcification, ischemic and thrombotic events, and atrial fibrillation) and mortality in patients with CKD across stages of the disease [84–86] Also in the general population, clear associations between PCS and IndS concentrations and cardiovascular endpoints have been reported. For example, in a population-based study in Belgium, the prevalence of hypertension increased along with PCS and IndS quartiles [87].

Evidence of the skeletal toxicity of protein fermentation metabolites is much more limited. Protein fermentation metabolites may confer direct toxicity to bone cells and disrupt bone matrix characteristics, thereby compromising bone quality and strength [88,89]. IndS promotes osteoblast apoptosis [90] and inhibits osteoclast di fferentiation [91]. The latter may occur through aryl hydrocarbon receptor signaling-dependent suppression of receptor activator of nuclear factor kappa-B ligand (RANKL) production [92]. IndS also causes the deterioration of bone mechanical properties [93,94] and bone architecture. Finally, IndS may induce skeletal resistance to parathyroid hormone (PTH) [95]. Increased protein fermentation may contribute to the high prevalence of a dynamic bone disease in patients with CKD, despite these patients often presenting with PTH levels exceeding the normal upper limit severalfold [96].

Protein fermentation metabolites may also a ffect bone and vascular health indirectly, e.g., by promoting inflammation (vide infra) and epigenetic silencing of Klotho, an anti-aging protein [97–99]. Emerging evidence indicate that Klotho deficiency is involved in the pathogenesis of vascular calcification and bone loss in CKD. *Klotho*-null mice [100,101] show extensive vascular calcification and a low-turnover osteopenia phenotype. The bone phenotype, most probably, results from systemic disturbances in mineral metabolism associated with disrupted FGF23–Klotho signaling rather than from a functional defect of Klotho in osteocytes [102,103].

### **6. Role of Decreased Carbohydrate Fermentation in the Bone–Vascular Axis**

Fermentation of complex carbohydrates results in the generation of short-chain fatty acids (SCFAs) [104]. The main SCFAs are butyrate, propionate, and acetate, which are found in the intestine in a molar ratio of 60:20:20. SCFAs are e fficiently absorbed by the gu<sup>t</sup> mucosa by poorly selective anion-transporting proteins [105]. SCFAs, not used by the colonocytes as a source of energy, enter the portal circulation and subsequently either are metabolized by the liver or enter the systemic circulation. SCFAs entering the systemic circulation have important impacts on host physiology as sources of energy, regulators of gene expression (e.g., via inhibition of HDAC), and signaling molecules that are recognized by specific receptors. Especially butyrate is a pleiotropic molecule, functioning as a ligand for certain G protein-coupled receptor (GPCR, e.g., GPCR41 and 43, also known as free-fatty acid receptor 3 and 2) and as a peroxisome proliferator-activated receptor agonist [57].

Production of both propionate and butyrate is reduced in animal CKD models [106]. Human studies, so far, yielded inconsistent findings with regard to both the overall capacity of microbiota to produce butyrate [107] and the circulating levels of SCFAs [108,109]. Chinese patients with CKD stage 5 showed a reduction in the most abundant butyrate-producing microbial species 48 and almost threefold lower plasma butyrate levels than healthy controls [108]. A comparable study in the Netherlands, however, failed to confirm these findings [107].

An increasing body of evidence implicates SCFAs in the pathogenesis of bone disease [64]. SCFAs may promote a positive bone balance by suppressing osteoclastogenesis and stimulating osteoblastogenesis. Mechanistically, propionate and butyrate induce metabolic reprogramming of osteoclasts, resulting in enhanced glycolysis at the expense of oxidative phosphorylation, thereby downregulating essential osteoclast genes [110]. Butyrate, furthermore, suppresses osteoclast di fferentiation, most probably by increasing the production of osteoprotegerin (OPG) by human osteoblasts [111,112]. Butyrate is also capable of stimulating bone formation [111,113]. The underlying mechanisms remain poorly defined. Butyrate promotes the di fferentiation of naïve CD4+ cells into regulatory T cells (Tregs). The expansion of Tregs in the bone marrow leads to increased production of Wnt10b. This Wnt ligand subsequently activates Wnt signaling in osteoblastic cells, leading to osteoblast proliferation, di fferentiation, and survival [113]. Remarkably, this anabolic e ffect is only seen in trabecular bone. It is unclear whether, and if so, to what extent, the weak inhibition of HDACs accounts for the bone anabolic e ffects of butyrate [114].

SCFAs may also protect bone indirectly, e.g., by suppressing inflammation (vide infra) and by increasing insulin-like growth factor 1 (IGF-1), a distinct bone anabolic factor [115]. Finally, the CKD-induced microbial metabolism shift away from saccharolytic fermentation and towards proteolytic fermentation creates a colonic microenvironment (e.g., a higher luminal pH) that may hamper calcium absorption [62]. The contribution of calcium absorption in the colon to the overall calcium influx is probably limited. Nevertheless impaired colonic calcium absorption may contribute to a tight, if not negative, calcium balance, commonly observed in CKD patients free of calcium supplements [116].

Studies exploring the role of SCFAs in vascular (patho)biology are limited. Butyrate activates NRF2 at the transcription level [117–120]. This e ffect is mediated by HDAC inhibition. One of the downstream e ffects of NRF2 activation is the upregulation of the glutathione/glutathione S-transferase (GST) antioxidant system resulting in a beneficial smooth muscle cell (VSMC) redox state [121]. Activation of NRF2 signaling has been shown to alleviate high phosphate-induced calcification of VSMCs [122]. SCFAs also have anti-inflammatory properties and thus may indirectly protect against vascular calcification (vide infra).

### **7. Role of Vitamin K Deficiency in the Bone–Vascular Axis**

Microbiota are capable of producing menaquinones (vitamin K2). To what extent the microbial production of menaquinones (vitamin K2) contributes to the overall vitamin K status of the host remains a matter of ongoing debate [123]. Experimental studies on the e ffect of oral and colorectal administration of vitamin K on circulating prothrombin concentration in vitamin K-deficient rats demonstrated that the bioavailability of colonic vitamin K is more than 50-fold lower than the bioavailability of oral vitamin K [123]. Conversely, data from germ-free rodents [124] and experimental and clinical studies with broad-spectrum antibiotics indicate that gu<sup>t</sup> microbial metabolism may be important to maintain adequate vitamin K stores in the mammalian host [125–127].

Recent data indicate that a large majority of patients with CKD are vitamin K-deficient [128–133]. Besides dietary restrictions, therapy with vitamin K antagonists and phosphate chelators, and impaired

vitamin K recycling, a decreased microbial production related to gu<sup>t</sup> dysbiosis may account for the high prevalence of functional vitamin K deficiency in CKD [129,130,134,135].

Vitamin K deficiency is a well-recognized risk factor of vascular calcification and arterial sti ffness, both in the general population and in CKD patients [136,137]. Accelerated vascular calcification in individuals with functional vitamin K deficiency is explained by incomplete γ-carboxylation and reduced function of matrix Gla protein (MGP) in the vasculature [138]. MGP is a 14 kDa secretory protein synthesized by chondrocytes, VSMCs, endothelial cells (ECs), and fibroblasts. γ-Carboxylated MGP inhibits vascular mineralization both directly, as a part of a complex with fetuin-A (also known as α-2-HS-glycoprotein), and indirectly, by interfering with the binding of bone morphogenetic protein-2 (BMP-2) to its receptor and thereby inhibiting BMP-2-induced osteogenic di fferentiation.

Low dietary intake of vitamin K, therapy with vitamin K antagonists, and functional vitamin K deficiency, as determined by circulating biomarkers (such as dephosphorylated–uncarboxylated MGP), are associated with low bone mineral density (BMD) and increased risk of fractures, both in the general population [90–92] and in patients with CKD [133,139]. Vitamin K-dependent γ-carboxylation of Gla-containing bone proteins such as MGP and osteocalcin (also referred to as bone Gla protein) may positively impact mineralization and bone quality. However, much remains to be learned on the role of MGP and osteocalcin in bone biology [140–142]. Vitamin K may a ffect bone health also directly by targeting the steroid and xenobiotic-sensing nuclear receptor (SXR), expressed in osteoblasts [141,143]. Finally, vitamin K deficiency may trigger micro-inflammation and thus contribute to the calcification paradox (vide infra) [133,144].

### **8. Role of Inflammation in the Bone–Vascular Axis**

CKD is well-recognized as a state of micro-inflammation [25,145]. Several factors contribute to the inflammatory status in CKD. Only in recent years, gu<sup>t</sup> dysbiosis has been recognized as another important culprit [54,146]. The pathways linking gu<sup>t</sup> dysbiosis to inflammation are manifold.

First, gu<sup>t</sup> dysbiosis is associated with a dysfunctional epithelial barrier [69,147]. The disruption of the gu<sup>t</sup> epithelial barrier enables the entry of endotoxin and other microbial components into the systemic circulation, which in turn may elicit an inflammatory response [148]. Several studies in animal models of CKD have documented superficial mucosal erosions, mucin loss, or disrupted tight junctions between intestinal epithelial cells in several parts of the gastrointestinal tract [54,55,149,150], in line with autopsy findings in patients on chronic hemodialysis, which who show subtle pathologies indicative of di ffuse gastrointestinal wall inflammation [151]. Besides gu<sup>t</sup> dysbiosis, sympathetic overactivity and intestinal congestion due to hypervolemia as a result of heart failure are hypothesized to contribute to increased intestinal permeability in CKD [38].

Second, both an increased exposure to protein fermentation metabolites and a decreased exposure to SCFAs have been hypothesized to contribute to micro-inflammation in CKD. PCS was shown to activate leucocyte free-radical production [152] and IndS-induced proinflammatory cytokines in human primary macrophages, by a mechanism involving the activation of Delta-like 4 (Dll4)–Notch signaling [153]. Other studies, conversely, failed to confirm the proinflammatory properties of PCS and IndS [154]. Moreover, clinical studies investigating the relationship between serum levels of gut-derived uremic toxins, markers of inflammation, yielded inconsistent findings [155]. The anti-inflammatory immune-regulatory properties of circulating SCFAs are well established and best characterized for butyrate. Butyrate stimulates the production of ketone bodies, including β-hydroxybutyrate, known to suppress the activation of the NACHT leucine-rich repeat and pyd domains-containing 3 (NLRP3) inflammasome [156] and suppresses nuclear factor kappa-B (NF- kappa-B) signaling in immune cells [157,158]. Butyrate may also mediate systemic anti-inflammatory e ffects by inhibition of HDACs [57,159]. However, the clinical relevance of the latter mechanism is questionable, as butyrate circulates only at micromolar levels, which is far below the IC50 for HDAC inhibition.

Finally, vitamin K deficiency is associated with inflammation [133,144]. Both causality of the relationship and its underlying molecular mechanisms remain to be defined.

Inflammation may be a common soil for bone loss and vascular calcification [24,25,160–166]. The pathophysiological mechanisms linking inflammation to vascular calcification are complex and multifaceted. Inflammatory cytokines and C-reactive protein may (a) promote endothelial-to-mesenchymal transition [161], (b) augmen<sup>t</sup> osteo–chondrogenic differentiation of vascular smooth muscle cells through activation of Msx2–Wnt/β-catenin signaling [167] and induction of oxidative stress [168], and (c) repress the production of fetuin-A, an important calcification inhibitor [169]. Vascular calcification, in turn, may elicit an inflammatory response and as such trigger a self-perpetuating vicious circle.

Experimental data indicate that inflammatory cytokines, either circulating or locally produced in the bone, such as TNF-<sup>α</sup>, IL-6, and IL-1β, may induce increased bone resorption [170–173]. These effects are mediated, in part, via cytokine-induced increases in RANKL, a key stimulator of bone resorption, expressed by osteoblasts and T cells [174]. TNF-α is also an inhibitor of bone formation [175], further tilting the balance towards bone loss [161]. In disagreement with these data, Barreto et al., reported a positive correlation between TNF-α levels and bone area [176]. These authors speculate that elevated TNF-α expression may represent a homeostatic feedback mechanism to counteract excessive bone mass gain.
