Background

Chronic kidney disease (CKD) is a major public health issue that affects 15% of U.S. adults, leading to impaired quality of life [1]. Meta-analysis and systemic review of population-based studies provide strong evidence supporting CKD as an independent risk factor for cerebral small vessel disease and cognitive impairment [2,3,4]. Notably, end-stage renal disease is associated with a substantially higher risk of both ischemic and hemorrhagic strokes [3] as well as accelerated brain aging and cognitive decline [5], and consequently, higher morbidity and mortality [6]. Postmortem examination of CKD human brains showed an increased prevalence of cerebral small vessel disease and highlighted the frequency of microvascular calcification in those brains [7]. Pathways that promote cerebral small vessel disease in the uremic milieu include loss of calcium/phosphorus homeostasis, blood pressure variability, retention of vascular toxins, and chronic inflammation [8]. Investigations into the relationship between CKD and cerebral small vessel disease subtypes are needed to identify novel prevention and treatment strategies in CKD patients.

Cerebral microhemorrhages (CMH) are the pathological substrate for cerebral microbleeds, which represent focal hemosiderin/iron deposits on MRI and are linked to cognitive impairment and ischemic and hemorrhagic stroke. Given that cerebral microbleeds are strongly age-dependent [9] and are present in up to 50% of hemodialysis patients [42,43,44]. Microglial dynamics are important in maintaining brain homeostasis. Microglia are maintained in a resting state with a morphology characterized by long, ramified processes extending from the soma and terminating with bulbous endings; they can be activated when brain injuries are present, transforming into a phagocytic appearance, exhibiting large, rounded soma with no or few processes [25, 45]. Impaired microglial function has been implicated in aging [46] and Alzheimer’s disease [47]. CKD-induced neuroinflammation has been associated with increased microglia/macrophage recruitment, a shift from an anti-inflammatory M2 to a pro-inflammatory M1 phenotype [48], and the formation of NLRP3 inflammasomes [49]. Although morphological profiles of microglia have been well-characterized in rodents [50], little is known about their morphological changes in the context of CKD, as well as the neuropathological consequences. In the current study, CKD is associated with microglial activation, shown by increased Iba-1 immunoreactivity and a phenotypic switch from resting to activated microglia (Fig. 3A–C).

We used Iba-1 immunostaining which identifies both microglia and macrophages [48]. We characterized the morphology of microglia at higher magnification (40x) to distinguish microglia from macrophages, which allowed us to determine the contribution of microglia in CKD-induced neuroinflammation (Fig. 3F). Our prior mouse work showed a positive correlation between Iba-1 immunoreactivity and CMH number [14]. Thus, the causal relationship between microglial activation and CMH formation was investigated in the current study using PLX3397 diet allowing for microglial depletion (Fig. 4). CKD-induced CMH formation was significantly decreased in aged mice with microglial depletion (Fig. 4F), while serum creatinine levels were found not to be affected (Fig. 4B). These findings indicate that CKD-induced CMH formation is at least partly mediated by microglial activation.

The BBB is formed by endothelial cells lining the capillary wall, astrocyte end-feet surrounding the capillary, and pericytes embedded in the baseline membrane, thus creating a physical barrier between the peripheral circulation and the central nervous system. The tight junctions between the endothelial cells serve to restrict passage of blood-borne substances (e.g., fibrinogen) into the brain and play a crucial role in brain homeostasis [51]. Impaired endothelial tight junctions at the BBB are well-characterized in aging brains [52] and can lead to passage of iron into the brain [53]. We have previously shown disruption in BBB tight junction proteins in young CKD mice during an inflammatory state [13]. Following the previous observations, we examined the passage of fibrinogen into the brains of CKD mice. Fibrinogen enters the brain after BBB injury and can be converted into insoluble fibrin, contributing to neuroinflammation and neuronal damage in many conditions [54,55,56,57]. Fibrinogen/fibrin deposition is associated with microglial activation and increased immune cell recruitment into the brain [54, 58]. In line with this finding, we showed microglial activation and increased fibrinogen deposition in the brains of CKD mice (Fig. 3). Activated microglia modulate expression of tight junction proteins essential for BBB integrity [59], which may further exacerbate BBB disruption and CMH formation.

Our assessment of microvascular integrity in vitro has relied on TEER and tracer permeability measurements. TEER is the measurement of electrical resistance across a cellular monolayer to evaluate integrity of the endothelial monolayer. In our in vitro study with ihBMEC, TEER was elevated in the first 24 h due to the trophic factors present in the serum and reached the highest value at 48 h. Incubation with CKD serum caused injury to the monolayer, as shown by a reduction in TEER after serum treatment initiation (Fig. 5A). This is consistent with our previous findings from an in vitro study incubating mouse brain endothelial cells (bEnd.3) with CKD serum [13]. Furthermore, the decrease in TEER was accompanied by a two-fold increase in sodium fluorescein permeability across the ihBMEC monolayer (Fig. 5B). Sodium fluorescein is a small molecular weight tracer (MW: 376 Da) that more readily diffuses through the BBB than larger molecular weight tracers, and therefore has served as a common marker for altered permeability [26]. Previously, we showed an inverse correlation between sodium fluorescein permeability and TEER measurements across the ihBMEC monolayer [26]. Together, impairment of the monolayer induced by CKD-derived serum factors allows for increased passage of sodium fluorescein across the monolayer, suggesting disruption of the endothelial monolayer and consistent with BBB injury observed with fibrinogen immunostaining.

The murine gut microbiome changes with aging and is associated with alterations in microbial carbohydrate metabolism, decreased fecal short-chain fatty acids and decreased cobalamin and biotin biosynthesis [60, 61]. In the current study, we examined well-known gut-derived uremic toxins that are generated in the CKD milieu. These toxins are derived from amino acid catabolism by intestinal microbiota and are associated with systemic inflammation and vascular injury in CKD [62,63,64,65]. The exact uremic toxins that contribute to BBB disruption and whether the effects are results of an individual toxin or a combination of several toxins remain largely unknown. We have previously shown CKD serum from dialysis patients caused marked drop in TEER, and urea was one of the key uremic toxins. Exposure of bEnd.3 cells to urea at concentrations approximate to the values measured in dialysis patients reduced TEER in a dose-dependent manner [13]. In the current in vitro study, this was further investigated by exposing bEnd.3 mouse brain endothelial cells with various gut-derived uremic toxins alone and in combination, with TEER measurements every 24 h. We demonstrated that uremic toxins (i.e., IS, PCS, and TMAO) in combination with urea and LPS exerted the most deleterious effects on the endothelial barrier. TEER was significantly lower in the 3Toxins + Urea + LPS group compared with the control group, and eventually declined to a level close to the baseline values at day 4 (Fig. 5D). These findings suggest gut-derived uremic toxins aggravate urea/LPS-induced bEnd.3 endothelial barrier dysfunction, again consistent with findings of BBB injury observed with fibrinogen immunostaining.

Despite these multiple findings, this study has some limitations. The mouse study suggests a key role of microglia in mediating CMH formation in CKD animals. It should be noted that the use of CSF1R inhibitor PLX3397 eliminates microglia in the parenchyma, but also affects the number of non-parenchymal macrophages in the perivascular spaces, the choroid plexus, and the meninges [66], which are known to cause neurovascular dysfunction [67]. Our findings do not rule-out a contribution of perivascular macrophages to CMH formation [68], a subject that warrants further investigation. The binary morphological categorization of microglia into resting and activated states may be an oversimplification. Microglia intermediate between the two states may adopt various functions in immune cell recruitment and activation, cell proliferation, and phagocytosis as their morphology changes [69]. Note that the vascular source of CMH in uremic milieu remains unclear, as iron uptake into the brain is highly regulated by BBB. Mouse models and human postmortem studies of aging, hypertension, and Alzheimer’s disease have suggested a capillary source of CMH involving BBB disruption [14, 70,71,72], consistent with the findings from the in vitro CKD models we investigated. However, these findings should be interpreted with caution, because CMH may develop via a mechanism independent of capillary injury [73,74,75]. To expand our knowledge on the vascular source of CMH, we have developed a semi-automated approach to characterize microvascular network in three-dimensional (3D) imaging of mouse brains [76]; this will enable us to visualize the co-localization of fluorescently labeled microvascular network and Prussian blue-positive CMH and therefore, identify the vascular origin of CMH. In addition, the issue of cognitive decline with CKD is important, as we have emphasized [77]. However, the experiments in our manuscript were neither designed nor powered to address this issue. In terms of in vitro BBB models, recent studies show that ihBMEC have a mixed endothelial–epithelial transcriptional profile [26, 78, 79], and we therefore used two different brain endothelial cell culture systems, i.e., ihBMEC and bEnd.3 cells, for a better understanding of the mechanistic role of microvascular endothelial function in CKD-induced CMH formation.

Conclusions

Adenine-induced CKD promotes the development of CMH in aged C57BL/6J mice independent of blood pressure, likely via microglial activation and BBB disruption. Extent of CMH development is directly proportional to degree of renal insufficiency. Moreover, serum-derived factors in CKD disrupt endothelial monolayer by reducing TEER and enhancing the passage of sodium fluorescein across the monolayer. Gut-derived uremic toxins (i.e., IS, PCS, and TMAO) aggravated urea/LPS-induced endothelial barrier dysfunction by producing a marked drop in TEER, supporting the key role of uremic toxins in CKD-specific mechanisms that contribute to microvascular dysfunction. These findings indicate that CKD provokes microvascular injury leading to CMH formation in this model and suggest that CKD makes an important contribution to cerebral small vessel disease.