**2. Bone–Vascular Axis**

Mounting evidence indicates that CKD is a state of impaired bone quantity [3–9]. In clinical practice, bone quantity is most commonly assessed by dual-energy X-ray absorptiometry (DXA). A decreased bone quantity [6,10], along with an impaired bone quality [11], contributes to an excessively high fracture risk in CKD patients. Epidemiological evidence demonstrates that the fracture risk increases along with the progression of CKD, with CKD stage-5D patients showing a non-vertebral fracture risk that is up to six times higher than the fracture risk of age- and gender-matched controls [12,13]. Fractures are a major cause of morbidity and, compared to CKD patients without fractures, those with fractures experience a several-fold increased risk of mortality [14,15]. Fractures also impose a large financial burden on healthcare systems.

Vascular calcification is a condition characterized by calcium phosphate crystal deposition in the intima, media, or cardiac valves [16]. Media calcification is most common among patients with CKD, with prevalence and severity paralleling the progression of renal failure [17]. Vascular calcification is observed in more than 60% of patients with CKD stage 5D [16]. Vascular calcification is an active, cell-regulated process. Its pathophysiology varies across vascular beds and remains incompletely understood, despite major progress in the last decade [18–21]. Vascular calcification is an established independent risk factor for cardiovascular disease (CVD), the leading cause of morbidity and mortality in patients with CKD [22,23].

Many clinical studies have demonstrated an association between low bone mass and vascular calcification in patients with CKD [24–29]. The association between osteoporosis and vascular calcification is not specific to CKD. It also is commonly observed in the elderly and in patients with diabetes mellitus or chronic obstructive pulmonary disease [30–35]. Importantly, the association remains significant after adjustment for age, which suggests an age-independent relationship [26,27,29–33,36,37]. Vascular calcification and bone mineralization are both actively regulated processes showing many similarities. The co-existence of bone loss with vascular calcification should therefore be considered a paradoxical phenomenon. It is commonly referred to as the 'calcification paradox'. It most likely reflects direct bone–vascular cross-talk and/or the involvement of common pathogenic factors [2,35].

### **3. Gut Microbial Ecosystem in Health and CKD**

The human gu<sup>t</sup> harbors a complex and dynamic microbial ecosystem that is shaped by diet and host factors [38]. The human microbiome project has shown that the composition of the microbial ecosystem is quite di fferent from one individual to the other. This variability in composition is not continuous and random, but stratified. Nutrient intake patterns are associated with both the degree of diversity and certain clusters of microbial species that are often found to act in concert. The microbial ecosystem thrives on the nutritional leftovers brought to them via the digestive tract. This requires substantial metabolic flexibility, as nutrient availability is dependent on host nutrient intake and digestion. A complex web of overlapping metabolic pathways allows access to nutritional sources inaccessible to mammalian metabolism, thereby supplementing the host metabolism.

The gu<sup>t</sup> microbiota provides the host with a variety of functions including the digestion of complex dietary components, production of vitamins, maturation of the immune system, protection against pathogens, and regulation of host metabolism [39]. A compelling set of bidirectional links between the gu<sup>t</sup> microbiota and the host (patho) physiology has emerged, and metabolites produced by the microbiota are increasingly implicated as crucial executors of the microbial influence on the host. Of note, microbial metabolites do account for about 10% of circulating metabolites [40].

CKD is associated with a disturbed gu<sup>t</sup> microbiota composition and metabolism [41–43]. These disturbances reflect the aggregate consequences of CKD, more specifically, the e ffects of kidney dysfunction combined with the e ffects of therapeutic interventions and dietary modifications. Kidney dysfunction has a major impact on a number of physiological systems, including the gastrointestinal tract. More specifically, gastrointestinal assimilation and motility, both known to modify the colonic microenvironment, may be disturbed in CKD [44,45]. CKD, furthermore, causes an increased influx of urea, uric acid, and oxalate into the colon. Urea is converted to ammonia and subsequently to ammonium hydroxide, which can raise the colonic pH and result in mucosal damage. Patients with CKD, furthermore, often consume a diet low in dietary fiber to avoid hyperkaliemia. These and other dietary measures may importantly impact on gu<sup>t</sup> microbiota composition and metabolism. Finally, not only antibiotics, but also non-antibiotic drugs are increasingly recognized to extensively a ffect human gu<sup>t</sup> bacteria [46]. This is especially relevant in the setting of CKD, as pill burden in these patients is huge.

Using bacterial DNA isolated from fecal samples, Vaziri et al. showed highly significant di fferences in the abundance of over 200 bacterial operational taxonomic units between hemodialysis patients and healthy controls [41]. Additional studies demonstrated that patients with End Stage Kidney Disease (ESKD) had an increased number of bacteria that possess urease, uricase, and p-cresoland indole-forming enzymes, and a contraction of families or genera possessing butyrate-forming enzymes (e.g., *Roseburiae*, *Lactobacillaceae*, and *Prevotellaceae*) [47,48]. Metabolomics studies showed clear di fferences in the levels of fecal metabolites (including phenols, indoles, and aldehydes) between patients with CKD and healthy controls. Of interest, the di fferences in fecal metabolite profiles were greater between patients on hemodialysis and unrelated healthy individuals than between patients on hemodialysis and household members exposed to the same diet [43]. Gryp et al., conversely, failed to observe increasing levels of p-cresyl sulfate, p-cresyl glucuronide, indoxyl sulfate, indole-3-acetic acid levels, and their precursors in stool and urine samples of patients along with the progression of CKD. In addition, anaerobic culture of fecal samples showed no di fference in ex vivo p-cresol, indole, and indole-3-acetic acid generation (https://doi.org/10.1016/j.kint.2020.01.028). The use of animal models enables the e ffects of CKD to be separated from those of therapeutic interventions and diet. Studies with uremic rats confirm that renal dysfunction itself induces profound changes in the gu<sup>t</sup> microbiota composition [41] and metabolism [43]. Taken together, current evidence indicates that CKD causes a microbial metabolism shift away from saccharolytic fermentation and towards proteolytic fermentation. Given some contradictory findings, additional prospective studies are required to confirm this shift.

CKD-induced changes to the composition and function of the intestinal microbiota also impair the intestinal barrier function, a condition commonly referred to as leaky gu<sup>t</sup> [38]. A leaky gu<sup>t</sup> in CKD is evidenced by the observation of increased concentrations of bacterial components, such as endotoxin or DNA, in the circulation of patients with increasing CKD stage. The levels of bacterial components are the highest in patients with ESKD treated with dialysis [49,50]. Although circulating bacterial components in patients on dialysis might derive from external sources such as dialysate fluids, the intestinal microbiota is by far the most likely source of these components in patients with CKD not on dialysis [50]. One study showed that after a few days of feeding uremic rodents with a non-pathogenic but green fluorescent *Escherichia coli* strain, green fluorescent bacterial colonies could be cultured from mouse livers, demonstrating that CKD facilitates the translocation across the intestinal barrier not only of bacterial components but also of entire living bacteria [51,52]. Our current understanding of the e ffects of CKD on the intestinal barrier function is in line with studies from the 1990s that demonstrated that orally ingested high-molecular-mass polyethylene glycols cross the intestinal barrier and enter the circulation and urine of uremic animals and patients [53]. Some but not all studies in animal models of CKD have demonstrated superficial mucosal erosions or disrupted tight junctions between intestinal epithelial cells in several parts of the gastrointestinal tract [52,54,55], in line with autopsy findings of patients on maintenance hemodialysis, which frequently show subtle pathologies indicative of di ffuse gastrointestinal wall inflammation. Both an increased exposure to urea-derived ammonia and ammonium hydroxide [56] and a decreased generation of butyrate may contribute to a leaky gu<sup>t</sup> [57]. Butyrate maintains the barrier function by at least two not mutually exclusive mechanisms. Butyrate is the primary energy source for colonic epithelial cells and undergoes fatty-acid oxidation to such an extent that these cells are slightly hypoxic. This leads to hypoxia-inducible factor-1-mediated upregulation of tight junction genes [58]. In addition, butyrate functions as a histone deacetylase (HDAC) inhibitor, and this has been shown to upregulate tight junction genes as well as the major intestinal mucin *MUC2* [59,60] gene and to downregulate the expression of pro-inflammatory cytokines [61]. Treatment of uremic rats with the symbiont *Bifidobacterium animalis* subsp. lactis Bi-07 attenuated epithelial erosion and decreased intestinal inflammation [52].
