Background
Both acute lung injury (either direct or indirect) and mechanical ventilation (MV) are important contributing factors for the development of acute kidney injury (AKI) in patients, but how they interact is unclear [
1‐
9]. This is important since AKI is a common problem in critically ill patients and carries significant morbidity and mortality, so that identification of mechanisms or modifiable risk factors may help to understand and manage AKI for the benefit of patients [
9,
10].
The effect of injurious high tidal volume (HV
T) versus non-injurious low tidal volume (LV
T) MV may depend on the underlying condition, such as aspiration pneumonia or sepsis [
11], and this may similarly translate into differences in susceptibility to AKI [
12]. In humans, HV
T vs LV
T MV did not increase the incidence of AKI in patients without lung injury [
13], but AKI was more common after HV
T than LV
T MV in patients with acute respiratory distress syndrome [
14]. Animal models of intra-tracheal acid and lipopolysaccharide (LPS) instillation demonstrate that HV
T MV is associated with increased kidney interleukin-6, vascular endothelial growth factor levels, apoptosis and necrosis compared to LV
T MV [
15‐
18]. However, these observations have not been confirmed by others. For example, Hoag et al. did not find an effect of acid instillation followed by HV
T MV on kidney apoptosis or kidney function [
12]. In a sepsis model, O’Mahoney et al. showed increased pulmonary cytokines and pulmonary permeability and increased plasma creatinine levels and protein accumulation in collecting tubules after LV
T MV following intraperitoneal LPS injection, but not after MV or LPS alone [
19]. Currently, known effects of MV on AKI during acute lung injury remain inconclusive and, to date, no study compared the effects of MV after either direct or indirect lung injury on the development of AKI. Otherwise, a growing body of evidence suggests that apoptosis plays a key role, particularly in inflammatory conditions, in the pathogenesis of AKI, and induction highly depends on underlying conditions [
6,
20‐
22].
Therefore, in the current study we set out to investigate whether the effects of HV
T MV on kidney apoptosis and function are differentially affected by the underlying acute lung injury; i.e. following either direct or indirect lung injury. We hypothesized that increased ventilator-induced lung injury after intra-tracheal acid instillation (direct lung injury) results in increased kidney apoptosis and decreased kidney function, as compared to less severe lung injury after cecal ligation and puncture (CLP)-induced sepsis (indirect lung injury). The present study uses the animals from our previous study and expands on this study by investigating the effect of different MV strategies on the development of kidney injury, apoptosis and dysfunction [
11].
Discussion
The most important findings of our study are that the effects of injurious MV on kidney apoptosis depend on the underlying type of acute lung injury. However, in contrast to our hypothesis, minimally lung-injurious HVT MV during sepsis caused kidney apoptosis, whereas HVT MV after intratracheal acid instillation was associated with severe lung injury but less kidney apoptosis. Second, kidney apoptosis was associated with a greater than 40% decrease in creatinine clearance after HVT as compared to LVT ventilation.
We found that HV
T during sepsis, as opposed to sepsis alone, caused kidney apoptosis in the absence of relevant lung injury. For apoptosis to occur during sepsis alone, more than one hit may be required [
17]. During sepsis systemic injury and inflammation occur which may increase the sensitivity of the kidney to apoptosis induced by ventilator-induced lung injury, whereas MV after intratracheal acid instillation may only increase local pulmonary injury and inflammation.
The observed MV induced kidney apoptosis may be explained by several mechanisms proposed previously [
4]. First, MV can induce kidney apoptosis by a direct effect of the systemic release of pulmonary produced toxic mediators [
4,
7]. We observed that only systemic aPAI-1 levels correlated with kidney apoptosis. In a rat model of pneumonia, MV with high tidal volume causes pro-coagulant changes and attenuated fibrinolysis in the lungs with alterations in systemic fibrin turnover [
27]. Although aPAI-1 was increased in the lung, aPAI-1 was not measured in the systemic circulation [
27]. Several effects of aPAI-1 on the development of kidney injury have been described [
7]. In-vitro, aPAI-1 can induce apoptosis in endothelial cells [
28]. Studies in animal models have shown that aPAI-1 messenger RNA levels in the kidney were elevated after CLP in a model of sepsis-induced acute kidney injury [
29]. Also in humans, aPAI-1 levels measured on days 0, 1 and 3 during the ARDS network trial were independently associated with AKI as measured by increased serum creatinine levels compared to baseline [
30]. In contrast, a recent study found that baseline aPAI-1 levels were not predictive of the need for renal replacement therapy but this study did not report on the incidence of AKI without the need for renal replacement therapy [
30]. Although the positive correlation of aPAI-1 and renal apoptosis may suggest a pivotal role for aPAI-1 in the development of AKI during sepsis and MV several issues remain unanswered. In contrast to our hypothesis we expected that more severe ventilator-induced lung injury would be associated with higher systemic levels of mediators, and consequently more apoptosis. Since this did not happen it remains questionable if the lungs are indeed the source of aPAI-1. In this regard aPAI-1, rather than being the cause of the increased kidney apoptosis, could be produced directly by kidney tubular cell in response to the ischemic damage caused by dysregulation of kidney vasoactive mechanism induced by sepsis and worsened theoretically by mechanical ventilation [
31,
32]. The production of aPAI-1 by tubular cells is increased in hypoxic condition and aPAI-1 is known to exert direct and indirect apoptotic effect [
31,
32].
Second, kidney apoptosis can be induced through an effect on both global and regional renal blood flow. Global differences in renal blood flow can be caused by hypoxia and/or hypercapnia. Therefore, we kept PaO2 and PaCO2 within normal limits in this study to avoid effects of MV on gas exchange with subsequent effects on renal blood flow. Also, mean arterial pressure was kept above 60 mmHg and was similar between the groups. However, the apoptosis we observed was unevenly distributed in the kidney, the apoptotic index was higher in the medulla compared to the cortex. The higher apoptotic index in the medulla suggests, despite similar mean arterial blood pressures, regional differences in renal blood flow. This indicates that HVT MV during sepsis may affect local perfusion in the kidney and, as mentioned before, as a result in local production of aPAI-1.
We showed previously that MV with HV
T in a rat model of pneumonia was associated with impaired kidney endothelium-dependent vasodilatation [
24]. Whether these changes may cause regional differences in perfusion is unknown. However, the impaired vasodilatation was attenuated after administration of a poly (adenosine diphosphate-ribose) polymerase (PARP) inhibitor [
24]. Two studies by the same group showed that during sepsis vasodilatation occurs with an increase in renal blood flow but with decreased creatinine clearance [
33,
34]. These findings may be explained by effects on kidney afferent and efferent arterioles leading to decreased glomerular capillary pressure [
35]. Damaging effects of cytokines, possibly released following increased and sustained sympathetic nerve activity, have been suggested [
36]. However, the exact mechanisms of arteriolar dysfunction remain unknown, and possibly, impaired fibrinolysis by increased aPAI-1 levels leading to endothelial dysfunction plays a role [
30].
There is increasing evidence for a pivotal role of apoptosis in AKI and septic AKI in humans [
21,
35]. Recently, a post-mortem study in patients with AKI associated septic shock demonstrated increased kidney tubular apoptosis [
20], but these data have not been confirmed by others [
8]. Additionally, genetic polymorphisms in the apoptosis regulatory protein BCL-2 gene protected against developing AKI during septic shock and MV [
37]. In a murine model of septic kidney injury the level of kidney dysfunction directly correlated with apoptosis [
22]. Although apoptosis can be reliably detected by TUNEL staining, different tests to detect apoptosis are usually performed to support the evidence of apoptosis obtained by TUNEL staining [
38]. We confirmed the TUNEL data by various other methods previously [
17]. In our study, MV induced kidney apoptosis was associated with a more than 40% decrease in creatinine clearance. Although creatinine can be actively excreted by the Lewis rat kidney, creatinine clearance is commonly used to evaluate kidney function [
24,
39]. Moreover endogenous creatinine clearance is strain specific [
40]. In Wistar rats, from which Sprague Dawley rats were developed, craetinine clearance is similar to inulin clearance [
40]. Plasma creatinine was higher in rats after acid instillation compared to sepsis. This difference is explained by one rat with an plasma creatinine level, substantially higher than the average level of plasma creatinine in rats subjected to MV after acid instillation.
The degree of sepsis in our model was relatively mild and of short duration prior to MV. This explains the absence of lung injury after sepsis and HV
T. Although CLP-induced polymicrobial sepsis is one of the best and most widely used animal models for sepsis and organ injury the bacterial inoculum is unknown and severity may vary accordingly [
23]. CLP-induced sepsis can also cause lung injury, mainly targeting the pulmonary endothelium [
23]. Campos et al., reported a 24 hr mortality rate of 50% after sepsis, whereas we observed a 6% mortality rate [
41] indicating a less severe sepsis with likely less endothelial damage in our study. The more severely injured pulmonary endothelium found by Campos et al., is highlighted by the increased wet-to-dry lung weight ratio [
41]. The increased endothelial damage may have attracted more polymorphonuclear granulocyes, including neutrophils with subsequently more oxidative stress than in our study where the pulmonary endothelium was not damaged [
11,
41]. The additional effect of HV
T MV on lung injury in our study was limited, similar to others where they found that a V
T of more than 15 ml/kg was necessary to injure the lungs during sepsis [
26].
In contrast to sepsis, MV following acid instillation in our study did not cause kidney apoptosis. Previous animal studies of acid instillation induced lung injury followed by injurious MV reported conflicting findings on kidney injury and apoptosis [
12,
15,
18]. Imai et al. showed that after 8 hrs of MV in rabbits following intra-tracheal acid instillation HV
T (15–17 ml/kg) MV increased kidney epithelial cell apoptosis 6-fold compared to a LV
T MV [
15]. After 4 hrs, this was not associated with increased plasma creatinine levels, but after 8 hrs, creatinine was higher after HV
T. In contrast, Hoag et al., did not observe kidney apoptosis after 5 hrs of MV with HV
T (25 ml/kg) following acid instillation or sham treatment in dogs. Furthermore, in this study, various measurements of kidney function did not differ between the groups [
12]. Hoag et al. suggested that in the study by Imai et al. the mean arterial pressure was maintained between 55–60 mmHg. This low mean arterial pressure may account for some of the alterations observed in plasma creatinine as a consequence of reduced renal blood flow, which was not measured [
12]. Species differences and severity of lung injury may have affected the differences in outcome in these studies.
This study has some limitations. Since rat chest wall and lung mechanics differ from the human situation these results cannot be translated to the human situation directly. We used a V
T of 15 ml/kg with no PEEP as a proof of concept. These settings are not used in humans since they are associated with increased lung injury and death in patients with ARDS [
14]. However, we observed increased renal apoptosis after HV
T in the absence of functional and histological lung injury. This suggests that during sepsis without lung injury MV with settings that do not directly injure the lung there may be effects on the kidney, especially since V
T’s greater than 6 ml/kg are still used [
42].
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
JWK participated in the design of the study, performed the animal experiments and drafted the manuscript. ABJG conceived of the study, participated in the design of the study and helped to draft the manuscript. JJH participated in the design and coordination of the study and assisted with the animal experiments and with the drafting of the manuscript. LS was critical in the mediator assays and helped to draft the manuscript. MPVB was critical in the mediator assays and helped to draft the manuscript. SJ performed the histology studies and critically reviewed the manuscript. RV participated in the design and coordination and assisted with the animal experiments. FBP conceived of the study, participated in the design of the study and helped to draft the manuscript. All authors read and approved the final manuscript.