Sepsis is the leading cause of acute kidney injury (AKI) and leads to increased morbidity and mortality in intensive care units. Current treatments for septic AKI are largely supportive and are not targeted towards its pathophysiology. Sepsis is commonly characterized by systemic inflammation and increased production of reactive oxygen species (ROS), particularly superoxide. Concomitantly released nitric oxide (NO) then reacts with superoxide, leading to the formation of reactive nitrogen species (RNS), predominantly peroxynitrite. Sepsis-induced ROS and RNS can reduce the bioavailability of NO, mediating renal microcirculatory abnormalities, localized tissue hypoxia and mitochondrial dysfunction, thereby initiating a propagating cycle of cellular injury culminating in AKI.
Endothelial dysfunction and microvascular rarefaction have been described as common pathophysiological features of AKI and are postulated to be critical factors mediating progression to CKD following recovery from AKI [30,31,32][30][31][32]. The NO system is an important regulator of vascular tone within the renal microcirculation, but it can be deleteriously affected in sepsis (Figure 2). In the healthy state, the biosynthesis of NO by vascular endothelial cells is dependent on the coupling state of endothelial nitric oxide synthase (eNOS) and the bioavailability of the co-factor tetrahydrobiopterin (BH4) [33]. When endogenous levels of the co-factor BH4 are sufficient, L-arginine is coupled with the reduction of oxygen, leading to the production of the potent vasodilator NO [33,34][33][34]. However, when BH4 levels are low, eNOS is uncoupled and superoxide is produced instead. Furthermore, BH4 is highly susceptible to oxidization to BH2 when levels of superoxide are high, further depleting the pool of the rate limiting co-factor BH4 [35,36][35][36]. Uncoupling of eNOS is reported to contribute to the pathophysiology of a myriad of kidney diseases arising from diabetes, hypertension and ischemia-reperfusion injury [37,38][37][38]. Intravenous supplementation with BH4 in an ovine model of sepsis improved microvascular dysfunction via increasing the number of perfused vessels, the proportion of perfused small vessels and the microvascular index within the sublingual circulation [39]. In an ovine model of severe septic AKI induced by intravenous infusion of live Escherichia Coli for 48 h, eNOS gene expression was selectively down regulated in the renal medulla, but not the renal cortex [11]. However, whether an uncoupling of eNOS contributes to the early onset of microcirculatory abnormalities reported within the renal medulla in ovine septic AKI [13] warrants further investigation.
In sepsis, excessive superoxide generation and accumulation, in tandem with inflammation, also results in direct structural damage to the vasculature, resulting in vascular leakage and tissue edema (Figure 2) [40].
The damaged endothelium also attracts leukocytes to the site of injury, as part of the innate immune response facilitated by the exposed intercellular and vascular cell adhesion molecules. This homing of pro-inflammatory cells, in conjunction with compromised gap junctions, leads to extravasation of the pro-inflammatory cells from the endothelium into the surrounding tissue, contributing to persistent inflammation [48][41]. Notably, inflammatory cells can generate ROS themselves and so reduce NO bioavailability [48][41], thereby contributing to the extensive pool of superoxide, essentially setting up a vicious cycle of oxidative stress, inflammation and vascular injury [19] (Figure 1). Sepsis-induced microvascular injury can also release microparticles into the systemic circulation.
Renal medullary hypoxia is emerging as a common pathophysiological feature of AKI arising from sepsis [13[13][42],50], cardiopulmonary bypass [51,52][43][44] and radiocontrast-induced nephropathy [53][45]. Furthermore, renal medullary hypoxia has been implicated as an important driver in the transition and/or propensity for progression from AKI to CKD [54,55][46][47]. The relatively high metabolic requirements of the tubular elements in the renal medulla, coupled with the topography of vascular and tubular architecture within the medulla, result in a steep oxygen gradient between the capillaries (vasa recta) and both the thick and thin ascending limbs of the loop of Henle and the collecting ducts [56][48]. There is also the potential for diffusive oxygen shunting in the renal medullary microcirculation (from descending to ascending vasa recta), which could further compromise renal medullary oxygen delivery [57][49]. In healthy sheep, graded occlusion of the renal artery and thus progressive reductions in renal blood flow resulted in proportionally greater degrees of renal medullary ischemia and hypoxia compared with a renal cortex indicative of an intrinsic deficit in the autoregulatory capacity of the renal medullary microcirculation [58][50]. Accordingly, under pathophysiological settings such as sepsis, renal medullary microcirculatory perturbations leading to even modest reductions in medullary oxygen delivery or increases in oxygen consumption can have adverse consequences for medullary tissue oxygenation.
Renal medullary hypoxia can be a major driver of a cascade of events leading to cellular injury, vascular injury and tubular dysfunction [59][51]. Acute renal insults, including endotoxemia, can both increase renal tissue oxygen consumption and reduce tissue oxygen delivery. For example, these changes can result in tubular injury and obstruction, and mislocalization of Na/K-ATPase and transport proteins within renal tubular epithelial cells, thereby reducing the efficiency of oxygen utilization for sodium reabsorption [60][52].
Mitochondrial dysfunction is proposed to be both a cause and consequence of renal hypoxia in the pathogenesis of septic AKI [72,73][53][54]. Mitochondria are the main consumers of oxygen within the kidneys. Thus, the production of physiological levels of mitochondrial ROS in the mitochondrial matrix is important because ROS serve as signals and regulators for a myriad of biological processes.
However, as cells experience prolonged periods of hypoxia, there is a change in metabolism and poor utilization of the available oxygen for ATP production in the mitochondrial electron transport chain, resulting in increased leakage of electrons and elevated production of free radicals/ROS [77,78][55][56]. Mitochondria use oxygen as the final acceptor of the respiratory chain, but its incomplete reduction can also produce ROS, especially superoxide [79][57]. Complex III of the electron transport chain is the inherent oxygen sensor during acute hypoxia, and it regulates the production of superoxide inversely with oxygen availability [80][58]. The transition of complex I from the active to “de-active” form was also reported to have the capacity to produce ROS outbursts during acute hypoxia [81][59]. Patients with septic AKI have elevated levels of receptor-interacting protein kinase-3 (RIPK3) in urine and plasma [82][60]. RIPK3 promotes oxidative stress and mitochondrial dysfunction in kidney tubular epithelial cells by increasing the expression and mitochondrial translocation of NADPH oxidase 4 and inhibition of mitochondrial complexes I and III [82,83,84][60][61][62]. It is therefore not surprising that mitochondrial injury has been commonly related to multi-organ dysfunction in patients with sepsis [6,85][6][63].
There are pre-clinical and clinical studies demonstrating that the adaptive processes of mitochondrial fission are downregulated in sepsis, which likely contributes to the loss of mitochondrial mass, thereby propagating ROS-induced damage during septic AKI. Sepsis is associated with considerable morphological changes in mitochondria. These changes include reduced numbers of cristae due to swelling of the inter-cristae space and the mitochondrial matrix, and vacuolation within the mitochondria space [82,86][60][64]. Depending on the severity of mitochondrial damage, the removal of mitochondria can be carried out by two pathways: mitophagy and apoptosis (Figure 3).