High fat feeding promotes obesity and renal inflammation and protects against post cardiopulmonary bypass acute kidney injury in swine

Acute kidney injury (AKI), defined clinically by an acute reduction in glomerular filtration rate (GFR) and rises in serum creatinine[1‐3], complicates 25 to 30% of cardiac surgery procedures, is associated with an in-hospital mortality of 10%, and increases in-hospital resource use of up to 100%[2‐4]. Our understanding of the underlying processes is poor and the prognosis for patients with AKI has remained essentially unchanged for decades.

Recently, a number of epidemiological studies have reported improved outcomes in critically ill patients who are obese when compared with normal weight individuals[5]. This phenomenon, termed the obesity paradox, includes improved outcomes in intensive care patients with acute kidney injury[6], as well as following cardiac surgery[7, 8]. A beneficial effect of obesity on inflammatory organ injury is not universally accepted, however; other authors have shown that very obese patients are at increased risk of AKI following cardiac surgery[9]. There is also no clear mechanistic explanation, and the findings of these studies have been variously attributed to statistical bias[10], confounding and reverse epidemiology[11], the anti-inflammatory effects of adipose tissue[12, 13], and the thrifty genotype[14].

We have described a novel large animal recovery model of post-cardiopulmonary bypass (post-CPB) AKI in swine[15] that shows quantitative and qualitative homology to the renal injury and dysfunction observed in cardiac surgery patients[16]. Renal inflammation, endothelial dysfunction and refractory cellular hypoxia are central features of AKI in this model[17, 18]. In a separate study we noted significant adiposity in domestic swine fed an atherogenic high-fat diet[19]. To investigate the mechanisms underlying the obesity paradox we examined the effects of obesity caused by high-fat feeding on post-CPB AKI in the swine recovery model using an index of GFR as our primary outcome and measures of endothelial homeostasis and inflammation as secondary outcomes.

Twenty-eight female Large White Landrace crossbred pigs approximately 4 months old and weighing 50 to 70 kg were used. Animals received care in accordance with and under license of the Animals (Scientific Procedures) Act 1986 and conforming with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85–23, revised 1996). The study had received local (University of Bristol) institutional review board approval, and was conducted under UK Home Office License PPL 30/2522.

Anaesthesia, recovery and reassessment were completed in all animals. Baseline characteristics are presented in Table 1. Pigs fed a high-fat diet were significantly heavier compared with those fed a normal chow diet. Weight gain was attributed in significant part to the accumulation of adipose tissue on the trunk (>2 cm), neck (>3 cm) and surrounding the viscera, which was not evident in pigs fed a normal chow diet. A high-fat diet also resulted in significant hyperlipidaemia (Table 1). There was no difference between the groups for baseline creatinine clearance. During the 2.5-hour intervention time period, CPB resulted in lower perfusion pressures and lower central venous oxygen saturation values compared with Sham pigs, with no difference between CPB and CPB + HF groups, or between Sham and Sham + HF groups for these measures (Figure 1A,B). There was no significant difference between CPB and CPB + HF pigs in terms of global oxygen delivery (Figure 1C) or oxygen delivery indexed to body mass during CPB (Figure 1D). Pigs undergoing bypass required significantly higher doses of vasopressor compared with pigs undergoing Sham procedures (Table 1); however, doses administered to CPB and CPB + HF groups were not significantly different. There were no differences between the groups with respect to periprocedural haemoglobin concentrations or gas exchange (PaO₂/FiO₂ ratio) (Figure 1E,F). Lung compliance was higher in the CPB + HF group at baseline, but no treatment effect on compliance was detected between the groups post intervention or for work of breathing (Figure 1G,H).

A 12-week high-fat diet resulted in significant obesity and hyperlipidaemia in swine. This promoted inflammatory cell signalling, vascular endothelial homeostasis, and epithelial cell survival in swine kidneys, preserved renal blood flow and GFR, and prevented CPB-mediated injury.

This study has several strengths, in particular the ability to evaluate functional, biochemical and histological changes in kidneys in response to clinical stimuli such as CPB, or obesity, in a large animal recovery model of AKI where the physiology has significant homology to that observed in humans[15‐18]. An important finding from this study is the clear association between acutely elevated serum IL-6 levels, a candidate biomarker for AKI currently being evaluated in clinical studies[20], and persistent reductions in GFR at 24 hours that are associated with ongoing inflammation and hypoxia. In contrast to IL-6, urinary NGAL was increased by a HF diet and did not discriminate between pigs that did or did not develop sustained reductions in GFR. This is consistent with false positives of this AKI biomarker in clinical studies in patients with pre-existing inflammatory states such as chronic kidney disease[21]. The acute (within 1.5 hours) elevation in IL-6 also highlights the limitations of existing clinical definitions of AKI that are based on peak serum creatinine levels detected up to 48 hours post injury[1, 2]. Our model does not show significant rises in serum creatinine at 24 hours post CPB, indicating a modest degree of injury. This contrasts with models of AKI in swine that involve periods of arterial occlusion and reperfusion (ischaemia and reperfusion), which can demonstrate acute rises in serum creatinine within hours[22, 23]. This picture is not typical of post-cardiac surgery AKI, however; clinical studies indicate that serum creatinine does not rise for up to 48 hours post surgery in patients with AKI[24]. Warm ischaemia and reperfusion for periods of up to 2 hours elicits a severe renal injury, with measured mean creatinine clearance of 7 ml/minute 4 hours following reperfusion in one study[22]. This is not consistent with measured creatinine clearances that we have documented after cardiac surgery[16]. Together, these studies underpin the value of translational models that can reflect different clinical scenarios. In the case of post-CPB AKI, our work indicates an inflammatory pathogenesis as opposed to ischaemia and reperfusion caused by periods of arterial occlusion where tubular necrosis is more marked[22, 23].

Epidemiological studies have speculated as to the causal pathway between obesity and improved clinical outcomes in the critically ill, with some reports suggesting that these observations may be attributed to attenuation of the inflammatory response[12, 13]. In contrast, our results indicate that renal protection was conferred by increased renal inflammation attributable to an obesogenic high-fat diet. This apparent paradox merits a careful consideration of the limitations of the current study. AKI in young healthy female swine may not reflect the multifactorial AKI observed in older cardiac surgery patients. For example, atherosclerosis, a common finding in the older cardiac patient, has been shown to alter inflammation and repair mechanisms in swine models of renal ischaemia and reperfusion[22, 23]. Weight gain in young swine may also be attributed to increase in muscle as well as adipose tissue, although adiposity was marked in these animals. Serum lipid levels were higher than those commonly recorded in humans. The principal limitation of this study is that we cannot differentiate the effects of obesity as distinct from the effects of hyperlipidaemia on outcomes. Previous work has demonstrated that hypercholesterolaemic homozygous apolipoprotein-E-deficient mice fed a high-fat diet are resistant to ischaemic AKI[25] in the absence of obesity. Similarly, pigs fed a high-fat diet that develop hyperlipidaemia in the absence of significant weight gain are protected against myocardial ischaemia reperfusion injury[26]. In that study, similar to our own findings, the protective effects of high-fat feeding and hyperlipidaemia in porcine myocardium were attributed to preserved endothelial homeostasis and increased microvascular density in the presence of inflammation and oxidative stress[26]. This suggests that it is the chronic renal inflammation consequent to hyperlipidaemia which protects against post-CPB AKI. Obesity may in fact represent an epiphenomenon; a bystander also attributable to high-fat feeding that is more readily measureable in epidemiological analyses, and this is reflected by the assessment of important additional confounders such as lipoprotein levels in more recent observational analyses of the obesity paradox[27]. We emphasise that our study, however, cannot discount an additional and potentially important neurohormonal and metabolic role for adipose tissue in our observations. This merits further study.

A significant finding in the current study was the association between intrarenal inflammation and both pro-survival and pro-apoptotic signalling. Previous studies have shown that the changes in renal vascular and tubular homeostasis that we have observed in response to a high-fat diet in swine are dependent on proinflammatory redox signalling[28, 29]. A key signalling node in this respect is NF-κB[29‐31], a redox-sensitive proinflammatory transcription factor the expression of which was increased by a high-fat diet in the current study, as was the expression of proteins encoded by several NF-κB-dependent genes, iNOS, VEGF and Bcl2[32]. In contrast, in the CPB pigs fed a normal diet, acute increases in NF-κB as well as iNOS and Bcl2 expression were associated with loss of endothelial homeostasis as well as increases in cleaved caspase 3, a marker of apoptosis. This was associated with more severe inflammation, as determined by ET-1 expression and serum IL-6 levels, and ATP depletion observed following CPB. These are potent drivers of tubular epithelial apoptosis and endothelial injury[33‐35] that may modify NF-κB signalling, as evinced in this case by the suppression of VEGF[36], and may promote cell death. It is also noteworthy that apoptosis in CPB pigs fed a normal diet was associated with activation of the key pro-survival node AKt. As with NF-κB it may be that the severity of the inflammation and ATP depletion caused by the CPB injury exceeded the protective effects of Akt signalling. In complete contrast, the preconditioning against AKI by a high-fat diet was independent of the phosphorylation of Akt, which is recognised as an important signalling node in acute ischaemic preconditioning[37]. Akt phosphorylation was in fact reduced in CPB + HF pigs relative to other groups, as was the expression of proteins in several related pro-survival signalling pathways including p70S6K (mTOR), Bcl2 (anti-apoptotic) and HIF-1α (angiogenesis). This is not entirely counterintuitive; decreased mammalian target of rapamycin signalling increases autophagy, an important protective mechanism in AKI[38]. Moreover, reduced mammalian target of rapamycin signalling alongside reduced HIF-1α expression promotes the polarisation of macrophages from the classically activated (M1) proinflammatory phenotype (thereby reducing IL-6 release) towards the alternatively activated reparative (M2) phenotype[39] that can promote renal recovery.

Of immediate translational relevance was our finding that inflammation was associated with both pro-survival and pro-apoptotic processes. More specifically, we observed pleiotropic effects of NF-κB signalling in the current study that have also been demonstrated in other models of renal vascular endothelial and tubular epithelial injury[40, 41]. This suggests that inflammatory redox signalling can have a protective role in AKI, either as a form of preconditioning or perhaps as an important process in cell repair. Consequently, blanket inhibition of inflammation or oxidative stress may not confer an overall benefit to patients at risk of AKI, and this is supported by the failure of anti-inflammatory interventions such as statins, corticosteroids, and antioxidants to reduce post-cardiac surgery AKI severity in randomised trials[42, 43]. Here we speculate that it is perhaps the dysregulation of inflammatory responses, typical of risk factors such as diabetes and chronic kidney disease[44, 45], or prolonged CPB[3] that predisposes to clinical AKI.

Our study demonstrates that high-fat feeding in swine results in obesity, hyperlipidaemia and renal inflammation. This protects against post-CPB AKI. Moreover, proinflammatory signalling was associated with both renal injury and protection. These observations demonstrate the complexity and redundancy of proinflammatory and pro-survival signalling pathways in the setting of post-CPB AKI, cast light on the failure of historical interventions to improve outcomes in patients at risk, and highlight the value of complex large animal recovery models in translational research.