Determinants of hepcidin levels in sepsis-associated acute kidney injury: Impact on pAKT/PTEN pathways?

The antimicrobial β-defensin-like role of hepcidin (HEPC) has been increasingly investigated for its potential role in acute kidney injury (AKI). In sepsis-induced AKI, there is a complex interplay between positive and negative regulation of HEPC, with consequently altered distributions of iron caused by changes in HEPC levels. The aim of the current research was to assess serum HEPC levels in a cohort of septic patients with AKI and investigate the regulatory impact of hypoxia-inducing factor (HIF)-1α, erythropoietin (EPO) and inflammation on HEPC levels and related signal cascades in these patients. Baseline, higher levels of SCr (2.3-fold), blood urea nitrogen (BUN) (1.8-fold), uric acid (2.3-fold) and white blood cell (2.3-fold) were noted in septic AKI patients, along with decreased levels of albumin (15.7%), creatinine (44.7%) and BUN/creatinine ratios (23.8%), compared to in normal subjects. These hosts also had increased serum levels of TNFα (4.4-times) and TGFβ1 (3.2-times) compared to controls (p < 0.05). Further, HEPC and HIF-1α levels were also increased (8.8- and 3.6-times control levels), while EPO levels were decreased (77.8%) from control levels. After 12 weeks of antibiotic therapy, all septic AKI patients showed significant improvement of the altered markers of kidney dysfunction. In line with significant reductions in serum TNFα and TGFβ1 (25.5% and 26.2%, respectively), HEPC and HIF-1α levels were significantly decreased (31.6% and 19.3%), and EPO levels increased (1.9-fold) compared to pretreatment values. There was a significant positive correlation between HEPC levels and kidney function markers (SCr and BUN), inflammatory TNFα and TGFβ1 and serum HIF-1α and pAKT in septic AKI patients before and after treatment. Based on the results here, we conclude that HEPC, EPO and HIF-1α are involved in the pathogenesis of sepsis-induced AKI and confirm the dominating effects of inflammatory determinants over hypoxia-related complications.

Keywords: Acute kidney injury; erythropoietin; hepcidin; hypoxia-inducing factor-1α; pAKT; phosphatase and tensin homolog; sepsis

Sepsis is considered the most common cause of acute kidney injury (AKI) in critically ill patients and is associated with increased morbidity and mortality. Considerable evidence now suggests that the pathogenic mechanisms of sepsis-induced AKI are different from those seen in other causes of AKI (Moore et al. [20]). The traditional paradigm that sepsis-induced AKI arises from ischemia and consequent hypoxia has been challenged by recent evidence that the total renal blood flow is not generally impaired during sepsis, and AKI can still develop in the presence of normal or even increased renal blood flow (Zhang [34]). Animal and human studies suggest that adaptive responses of tubular epithelial cells to injurious signals are responsible for renal dysfunction. The induced renal inflammation and microcirculatory dysfunction concomitantly further augment these mechanisms. Therefore, the identification of biomarkers should not be limited to initial injury alone but should include markers of risk, injury propagation and resolution of injury. In fact, it can be argued that markers of early resolution will be equally as important as markers of initial injury (Morrell et al. [21]).

A plethora of studies has been performed to assess promising urinary [and serum markers], e.g. NGAL, FABP, KIM-1 and cystatin C; however, renal biomarkers are less useful at predicting AKI in patients with sepsis (Glassford et al. [6]). The focus is currently on new serum indicators, as hepcidin (HEPC), hypoxia-inducing factor (HIF), erythropoietin (EPO) and their respective signaling cascades (Zhang [34]). In [15], Krause et al. described a peptide hormone (later called "HEPC" based on its hepatic expression and antimicrobial activity) whose expression was low in normal kidney, heart and brain. HEPC is considered a central systemic regulator of iron homeostasis, as it mediates intracellular iron sequestration by binding to the cellular iron export channel ferroportin receptors on hepatocytes, enterocytes and macrophages, leading to ferroportin endocytosis and degradation, and thereby decreases iron efflux from iron-exporting tissues into plasma. Within the kidney, HEPC is expressed in the apical tubular epithelium of the thick ascending limb of the loop of Henle, connecting tubules and cortical collecting duct. Moderate expression is found in the collecting ducts and thick ascending limb of the medulla (den Elzen et al. [3]).

The β-defensin-like peptide HEPC was found to be a principle regulator of systemic iron homeostasis as well as a Type II acute-phase protein (Ganz [5]). In concordance with this dual function, its expression is modulated by systemic iron requirements and inflammatory stimuli, as it is induced by cytokines such as interleukin (IL)-6. Its role in the development of anemia was first suggested by van Eijk et al. ([29]). Since that report, it has been demonstrated that HEPC is a central modulator of inflammation-associated anemia, not only by controlling the expression of ferroportin on intestinal cells and macrophages but also via a direct inhibitory effect of HEPC on erythropoiesis (Liu et al. [18]).

Several factors are postulated to affect HEPC concentration in many pathological conditions. Overall, there is a complex interplay between positive and negative regulation of HEPC and the consequent distribution of iron caused by changes in HEPC concentration (Haase et al. [8]). Downregulators of hepatic HEPC production include hypoxic conditions, iron deficiencies and erythropoiesis; upregulating factors are inflammation and the increased presence of proinflammatory cytokines, as well as iron overload. Of interest, urinary and serum HEPC showed different patterns of expression in relation to injury and repair (van Eijk et al. [29]), but the assessment of their serum levels in AKI due to sepsis needs further elucidation.

Recently, the role of HIF-1α, a regulator of cellular hypoxia responses, has received significant attention in kidney diseases (Nangaku et al. [22]). Central mediators of hypoxia-induced erythropoiesis are the O2-regulated transcription factors HIF-1 and HIF-2 (Nangaku et al. [22]). In vivo studies identified HIF as the main regulator of EPO, a glycoprotein that prevents erythroid progenitor cell apoptosis and is essential for maintenance of normal erythropoiesis and increased RBC production under hypoxia (Liu et al. [18]). To what extent AKI-induced hypoxia due to sepsis could potentially regulate EPO and HEPC levels has not been addressed before.

As previous data suggested persistently increased transforming growth factor (TGF)-β1 signaling prevents re-differentiation of tubules being repaired after injury (Lan et al. [16]), the possible role of this factor in AKI was of interest to us. There are proven TGFβ1 pathways that compromise epithelial integrity by disrupting cell junctions. Among proteins modified by TGFβ1 is phosphatase and tensin homolog (PTEN, an antagonist of PI3K signaling) that has lipid/tyrosine phosphatase-dependent and -independent activities that if not physiologically regulated alter cell behavior. It is known tubule PTEN is suppressed when TGFβ1 signaling is persistently elevated. A recent report suggested the PI3K/AKT/NF-κB signaling pathway was also involved in induced inflammation and sepsis-associated organ failure. However, the role of AKT/pAKT and PTEN in AKI remains poorly understood (Lan et al. [16]).

Most clinical and experimental studies have reported significant elevations of urinary HEPC in subjects with AKI, primarily those that were cardiopulmonary bypass- or radiocontrast-induced. However, urinary HEPC concentrations may not always accurately reflect plasma HEPC concentrations, since urinary HEPC levels may also depend on glomerular filtration, tubular reabsorption and local production by tubular epithelial cells (Peters et al. [24]). Of interest, urinary HEPC showed different patterns of expression in relation to injury and repair. While reliability/utility of urinary HEPC as a marker for AKI remains an object of intense study, to date, the utility of evaluating HEPC serum levels in AKI still remains unclear. Accordingly, the aim of the current research was to assess serum HEPC levels in septic patients with AKI before and after antibiotic treatment. Beyond assessing the potential utility of serum HEPC as a marker for AKI, this study was also designed to clarify the regulatory impact of hypoxia and inflammation on HEPC levels and any potential involvement of PTEN and pAKT as signaling pathways in these pathologies.

Graph: Figure 1. Descriptive presentation of data for (A) HEPC, (B) HIF-1α, and (C) pAKT using Box and Whisker plot.

This prospective study included 40 patients with septic AKI (22 male, 18 female; age range 42–46 years) who were recruited from Nephrology Clinics at Kasr Elaini Hospital at Cairo University (Cairo, Egypt) during December 2013 to September 2014. Sepsis was first diagnosed as a combined infection and a systemic inflammatory response, according to ACCP/SCCM Guidelines (Bone et al. [1]). Criteria for diagnosis of sepsis was confirmed when two or more of the following were met as a result of systemic infection, represented as systemic inflammatory response syndrome: (1) temperature >38 °C or <36 °C; (2) heart rate ≥90/min; (3) respiratory rate ≥20/min or a pCO2 ≤ 32 mm and (4) white blood cell (WBC) counts >12,000/mm3 or <4000/mm3 or >10% immature (band) forms.

AKI was diagnosed according to specific criteria introduced by the Risk Injury Failure Loss of kidney function End stage renal disease (RIFLE) classification and Acute Kidney Injury Network (AKIN) classification, as per KDIGO guidelines (International Society of Nephrology [13]) as an acute percentage increase in serum creatinine (SCr) of 50% (1.5 mg/dl), or reduction in urine output, defined as 0.5 ml/kg/h for >6 h. After identification of eligible patients, a detailed record of the patient's history, physical examination and laboratory investigations was made to document the demographic characteristics and the pathogenic factors causing AKI. The information collected at the time of assessment is shown in Table 1. Eligible patients then received empirical antibiotic regimens; all subjects were required to have blood cultures performed before starting the study therapy. Antibiotics were dosed per body weight.

Table 1. Laboratory results in control and septic AKI patients before/after antibiotics therapy.

1 Data are presented as means ± SD. BUN/Creatinine ratio is presented as median (25th–75th percentile range). Parametric data were analyzed via ANOVA; non-parametric data were analyzed via Friedman's test.

2 aSignificantly different from the normal group; bsignificantly different from AKI patients before therapy (both at p < 0.05).

Exclusion criteria overall included any at age <18 years, history of previous dialysis, kidney transplantation, kidney disease and/or hypovolemia responsive to fluids, as well as diabetics and pregnant patients. Thirty healthy volunteers were matched to the patient population by sex (male = 16, female = 14; age range = 45–51 years). None had a history of renal disorders; all had completely normal renal, hepatic and cardiac function tests. The following were excluded from the expected causative factors for AKI in patients enrolled in our study; nephrotoxic drugs, surgery, radiographic contrast agents within the 72 h before the defined increase in SCr concentration, decreased renal perfusion or hypoperfusion with overt/orthostatic hypotension (systolic BP <80 mmHg), or clinically-apparent congestive cardiac failure. This study was approved by the Ethics Committee of the Faculty of Pharmacy at Cairo University and conformed to ethical guidelines in the 1975 Helsinki Declaration.

Written informed consent was obtained from all study participants. Eligible patients received empirical antibiotic regimens and were required to have blood cultures performed before starting therapy. Antibiotics were dosed per age and body weight. AKI individuals were re-evaluated after 12 weeks of antibiotic therapy along with normal control group.

Venous blood samples were collected on the day of admission into the study. Samples (≈2 ml/subject) were divided into two portions; the first was collected for separation of sera while the other was immediately transferred into heparinized tubes for isolation of lymphocytes and generation of cell lysis products [for determination of PTEN and pAKT levels only]. Samples of blood were also collected after 12 weeks of antibiotic treatments for the AKI patients (placebos for controls).

Separation of sera was done by centrifugation (3500× g, 10 min) and all sera were stored at −70 °C until analyses. The resultant sera samples were analyzed to determine standard renal function test values, e.g. blood urea nitrogen (BUN) and SCr, as well as uric acid, serum albumin and WBC. All measures were taken at baseline and repeated 12 weeks after the onset of antibiotic therapy of AKI.

Peripheral blood lymphocytes (PBL) were separated from heparin-anticoagulated whole blood by density gradient centrifugation over Histopaque (ρ = 1077 g/ml, Sigma, St. Louis, MO) for 10 min [400× g, 4 °C]. The lymphocytes were isolated, washed with phosphate-buffered saline (pH 7.4) and then resuspended in 500 μl ice-cold Cell Lysate Buffer (50 mm Tris [pH 7.4], 250 mm NaCl 1 mm Na3VO4, 5 mm EDTA 0.02% NaN3) supplemented with protease and phosphatase inhibitors. The lysates were generated by resuspended by repeat pipetting and incubation of the materials at 2–8 °C for 30 min. After this period, the lysed cells were centrifuged [13,000 rpm, 10 min, 2–8 °C) and the resulting supernatants were transferred to clean tubes. These lysates were either then used immediately or sub-aliquoted and stored at −70 °C until use.

Serum TNFα levels were determined using a commercial ELISA kit (RayBio Human, RayBiotech, Inc., Norcross, GA). TGFβ1 levels were measured using a quantitative sandwich immunoassay kit from Anogen (Mississauga, ON, Canada). The intra-/inter-assay coefficient of variation for serum TNFα was 10–12% and that of TGFβ1 was 5.3–8.8%. Serum HEPC, HIF-1α and EPO levels were evaluated using ELISA kits as well: HEPC via a Hepcidin Prohormone kit (IBL International GMBH, Hamburg, Germany [level of detection (LOD) = <3.6 ng/ml]); HIF-1α via a HIF-1α kit (Cayman Chemical Co., Ann Arbor, MI [LOD = 0.63 ng/ml]) and EPO via a Quantikine IVD human EPO kit (R&D Systems, Minneapolis, MN [LOD = 0.6 mIU/ml]). A Gentaur ELISA kit (Molecular Products, Kampenhout, Belgium) and a Phospho-Akt (Ser473) ELISA kit (RayBiotech) were used, respectively, for assessment of PTEN and pAkt levels in serum and PBL lysates. The LOD of the PTEN kit was 0.12 pg PTEN/ml (serum) and 0.115 ng/ml (lysate). The LOD of the pAKT kit was 0.49 pg pAkt/ml (serum) and 0.6 ng/ml (lysate). All studies were done according to manufacturer instructions. Each sample was measured at least in duplicate.

Continuous variables are expressed as means ± SD and as medians and interquartile ranges if non-normally distributed. Normality was assessed by the Shapiro–Wilk test. Differences between groups were assessed by one-way analysis of variance (ANOVA), while Friedman's test, and if it was breached we used Welch's correction. The association between the parameters was determined using the Pearson's correlation coefficient. All reported probability values were two-tailed, and a p value <0.05 was considered statistically significant. Statistical analysis was performed using the SPSS version 20.0 statistical software package (SPSS Inc., Chicago, IL).

Patient characteristics and laboratory test results before and after antibiotic therapy are illustrated in Table 1 and Figure 1. After clinical assessment of sepsis, the laboratory results of the septic AKI patients' sera revealed culture-positive sepsis (N = 40, Gram-negative bacteria), higher levels of SCr (2.3-fold), BUN (1.8-fold), uric acid (2.3-times) and WBC (2.3-fold), in addition to lower levels of albumin (15.7%), creatinine (44.7%) and BUN/Creatinine ratios (23.8%), compared to normal subjects. The results also revealed the patients had elevated serum levels of inflammatory TNFα (4.4-fold) and TGFβ1 (3.2-fold) compared with control values (p < 0.05). Further, serum HEPC and HIF-1α levels were also elevated in the AKI patients (8.8- and 3.6-fold) relative to control subject levels, while EPO levels were decreased to 77.8% of control (Table 2).

Table 2. Serum levels of circulating endothelial dysfunction indicators inflammatory markers and PBL intracellular signaling markers in control and septic AKI patients before/after antibiotic therapy.

• 3 Data for HEPC, HIF-1α and pAKT here are presented as medians (25th–75th percentile range). Parametric data were analyzed via ANOVA; non-parametric data were analyzed via Friedman's test. All other data are presented as means ± SD.
• 4 aSignificantly different from the normal group at p < 0.05; bsignificantly different from AKI patients before therapy at p < 0.05.

After 12 weeks of antibiotic therapy, all septic AKI patients showed significant improvement of serum markers relative to corresponding pre-treatment values. Specifically, changes in terms of decreased septic levels of SCr (7.9% decrease), BUN (8.2%), uric acid (20.6%) and WBC (38.4%), as well as increased levels of albumin (2.6%), BUN/Creatinine ratios (12.2%) and creatinine (20%) were indicated. Moreover, a significant reduction in serum TNFα and TGFβ1 (25.5% and 26.2%, respectively) levels was also evident in these subjects compared to their pretreatment values (p < 0.05). Lastly, serum HEPC and HIF-1α levels were significantly decreased in post-treatment AKI patients (31.6% and 19.3% reductions, respectively) versus pretreatment levels, as illustrated in Figure 1. Serum EPO levels post-treatment underwent a 1.9-fold increase.

Both serum and PBL levels of pAKT and PTEN showed similar change patterns in septic AKI before treatment and afterward. Serum and PBL PTEN levels were lower in septic AKI patients (≈75 and ≈85%, respectively) compared to corresponding healthy host sample values. Each of these levels was significantly elevated after treatment (≈53 and ≈58%, respectively) from pre-treatment values. Similarly, serum and PBL levels of pAKT in septic AKI patients were significantly lower than in their normal subject counterpart samples (serum: 58–60%, PBL: 26–36%). These levels were significantly increased after treatment (serum: 63–78%, PBL: 43–60% from pretreatment values) (Figure 1).

Associations between HEPC, HIF-1α and EPO levels and the markers of AKI and inflammatory signals in septic AKI are presented in Table 3. There was a significant strong positive correlation between HEPC levels and the kidney function markers SCr and BUN in septic AKI patients before and after treatment, and strong negative correlation with serum albumin levels. The association between HEPC, EPO and HIF-1α levels revealed a positive strong association of HEPC with HIF-1α and a negative association with EPO; such findings which are in accordance with a negative regulatory effect of HEPC upon EPO expression. While HEPC levels before antibiotic treatment were positively associated with pAKT, the association with PTEN failed to reach significance; however, the association with PTEN attained a significant negative association after treatment. HIF-1α and EPO showed significant strong negative associations with HEPC levels before and after treatment. Weak associations between EPO/albumin and HIF/BUN were revealed after treatment.

Table 3. Correlation (r) values for AKI patient serum HEPC, HIF-1α and EPO levels with serum kidney function markers and signaling mediators PTEN and pAKT, before/after therapy.

5 Values statistically significant at p < 0.05 are indicated and values that were not significant are highlighted in bold italic font.

Concerning the serum levels of intracellular signaling indicators sPTEN and spAKT presented in Table 3, PTEN failed to show any significant correlation with any assessed markers before antibiotic treatment; however, after treatment, there was a strong negative correlation with HEPC, EPO and HIF-1α levels. On the other hand, pAKT showed a positive correlation with HEPC, HIF-1α and negative association with EPO levels, both before and after treatment.

With regard to TNFα and TGFβ1 presented in Table 4, both levels were found to positively correlate with the kidney function markers Scr, albumin and BUN but with a weaker association with TNFα, in both before and after treatment. Moreover, TNFα and TGFβ1 were strongly associated with HIF-1α levels, while being negatively associated with EPO levels, both before and after sepsis treatment (Table 4). Furthermore, a weaker association was revealed between TGFβ1 and HEPC. As for sPTEN correlation, there was a significant, strong negative correlation with TNFα and TGFβ1 levels only after treatment. spAKT showed significant, strong positive correlation with TNFα both before and after treatment and a weak one with TGFβ1 level before treatment (Table 4). Concerning the PBL-PTEN level, there were neither significant correlations with kidney markers nor with EPO and HIF-1α before treatment; with HEPC, there was no association before or after treatment. On the other hand, PBL-pAKT levels showed significant strong positive correlations with SCr, BUN, HEPC and HIF-1α levels both before and after treatment, as well as negative associations with albumin and EPO at both time points.

Table 4. Correlation (r) values for AKI patient serum TNFα and TGFβ1 levels as well as PBL-PTEN and pAKT levels with serum kidney function marker, HEPC, HIF-1α and EPO levels, before and after therapy.

6 Values statistically significant at p < 0.05 are indicated and values that were not significant are highlighted in bold.

In the cohort of patients here afflicted with sepsis-induced AKI, serum HEPC levels were significantly elevated compared to control patient values, an outcome in alignment with inflammatory injury and increases in the presence of inflammatory mediators TNFα and TGFβ1. Moreover, in parallel to the elevation in HEPC levels, HIF-1α levels increased significantly early after the onset of AKI, reflecting the expected AKI-induced hypoxia postulated to cause stimulation to EPO formation/release (Liu et al. [18]). Interestingly, here, expression of EPO responded in the opposite manner, i.e. there were significantly lower levels in septic AKI patients compared to in the control subjects. The induced response of this triad reflects the complex interplay between positive and negative regulatory processes controlling serum HEPC levels in sepsis-induced AKI. Here, two key determinants are postulated to be involved in regulating HEPC levels; the hypoxia induced by sepsis and consequent AKI (symbolized by changes in HIF-1α presence) and an increasingly inflammatory milieu (reflected by changes in the presence of TNFα and TGFβ1).

Nemeth and Ganz ([23]) previously investigated the regulatory pattern of HEPC and reported that the regulation of HEPC expression in the liver occurred at the transcriptional level and was modulated by iron status, inflammation and hypoxia. As HEPC is known to be positively regulated by inflammation and negatively so by hypoxia (Haase et al. [8]) – and both phenomena are considered key in the pathophysiology of AKI – it was interesting to observe the dominating impact of sepsis-induced inflammation over the hypoxic response in this particular AKI setting. Nangaku et al. ([22]) focused on the effect of hypoxia and HIF-1α in chronic renal diseases and stated that in patients with chronic kidney diseases, renal chronic hypoxia is the end result of multiple pathologic processes and mechanisms. The cellular response to hypoxia is centered on HIF-1α, designated a "master gene" switch that results in broad coordinated downstream reactions. HIF-1α is constitutively transcribed and translated, and its levels are primarily regulated by rates of degradation (i.e. via HIF-1α hydroxylation). Thus, under normoxic conditions, HIF-1α is targeted for ubiquitinylation and is rapidly degraded by proteasomes. Conversely, under hypoxic conditions, viz. AKI-induced hypoperfusion, there is less breakdown of the protein and so a surge in HIF-1α levels. This type of HIF "activation" is stimulated in other pathological conditions, including diabetes, acute renal ischemia (Hill et al. [10]) and experimental models of rhabdomyolysis (Katavetin et al. [14]).

Morrell et al. ([21]) provided crucial clarification to understanding the pathogenesis of septic AKI in humans such that it should not be exclusively viewed in the context of distributive shock-associated ischemia but rather also within the context of a dysregulated ill-defined inflammatory response to septic stimuli. This clarification was in accord with previous experimental data revealing conflicting results about HIF-1α levels during AKI. Despite a measured decrease in kidney perfusion in their septic mouse model, Tran and associates found no elevation in HIF1α transcription factor that is normally upregulated in hypoxic and ischemic rat kidneys (Tran et al. [28]). These authors hypothesized this was due to significantly reduced oxygen consumption by kidney proximal tubule cells during sepsis. While increased expression of hypoxia-inducible factor-1α was not seen in mice exposed to lipopolysaccharide (Huet et al. [12]), sepsis has been associated with a significant intrarenal oxidative and nitrosative stress mediated by inflammatory cytokines rather than by local hypoxia or ischemia (Heyman et al. [9]). Rosenberger et al. ([26]) utilized an ex vivo model of the isolated perfused rat kidney with controlled oxygen consumption to provide evidence for a "window of opportunity" for HIF-1α activation under a moderate sublethal hypoxia, whereas a more severe hypoxia resulted in HIF-1α suppression and induction of apoptotic cell death. The idea of "cellular stunning", where a cell within a septic environment is trying to defend itself from death by decreasing its metabolic activity, has been postulated as a mechanism not only in sepsis (Hotchkiss & Karl [11]) but also in models of ischemic heart disease (Sawyer & Loscalzo [27]).

Hypoxia is postulated to cause HEPC suppression and to stimulate EPO-induced erythropoiesis (Haase et al. [8]); however, contrary results were revealed by the current findings. Liu et al. ([18]) previously clarified this relationship and reported that in hypoxic conditions, HIF-1α did not have a direct inhibitory effect on HEPC but acted through HIF-induced erythropoiesis. This was based on the fact that HIF-1α-induced EPO production causes increased erythropoietic activity and the latter has a direct negative regulating effect on HEPC. In the current study, antibiotic treatment reduced the previously elevated levels of HEPC and HIF-1α, while levels of EPO were found to be further increased. Concerning the negative association of EPO with HEPC observed here, this was previously also noted in a population-based study of octogenarians (den Elzen et al. [3]). That study found EPO levels were highest in participants registering within the lowest HEPC quartile, and vice versa; those authors expounded that EPO itself may inhibit HEPC production.

We postulate here that one likely positive regulator of HEPC – responsible for striking increase in its serum levels – is sepsis-induced renal inflammatory processes in AKI. Both TNFα and TGFβ1 are inflammatory mediators whose increased expression here paralleled that of the HEPC levels. Numerous signaling pathways are activated to direct tubule regeneration after AKI, but several persist after repair. Among these, TGFβ1 is particularly important because it controls epithelial differentiation and profibrotic cytokine production. TNFα has been shown to upregulate renal apoptosis in animal models by binding with TNF receptor (TNFR)-1 on glomerular cells and TNFR2 on renal tubular cells (Wan et al. [30]). Increased renal TGFβ1 expression, potentially induced by the injury factors including TNFα, suggests that TGFβ1 may be a critical factor for renal repair. However, the function of TGFβ1 in the cellular process of injury and repair is often contradictive. It has been previously shown that stimulation with TNFα significantly enhanced TGFβ1 biosynthesis. In TGFβ1-deficient mice, kidneys were protected from ischemia-reperfusion injury. Upregulation of TGFβ1 resulted in an increase in antiapoptotic bcl2 expression, a protein implicated in a survival mechanism for resistance to injury in the kidney and protection of renal cells from apoptosis (Lan et al. [16]).

A plethora of studies suggested a strong association between cytokine levels (IL-6, IL-10 and macrophage migration inhibitory factor) and the development of sepsis-induced AKI (van Eijk et al. [29]), suggesting an important role of systemic inflammatory mediators in this process. During sepsis, infection triggers a host response, in which inflammatory mechanisms contribute to clearance of infection and tissue recovery on the one hand and organ injury on the other. The released proinflammatory mediators can act in an autocrine and paracrine fashion and may contribute to further tubular cell damage (Gomez et al. [7]). In humans, increased HEPC levels have also been found in diseases characterized by overt inflammation, like rheumatoid arthritis and sepsis. In addition, HEPC levels have been found to be elevated in anemic patients with chronic inflammation, chronic kidney disease or cancer. HEPC expression is, therefore, considered a potentially good marker of anemia or inflammation (van Eijk et al. [29]; Wu et al. [31]). In patients with lupus nephritis, changes in urine HEPC-20 and -25 measures predicted renal flares (Zhang et al. [33]). Those authors also reported the intrarenal expression of HEPC by infiltrating leukocytes in the patients, raising the possibility that during a renal disease flare, HEPC was produced within the kidney rather than simply being filtered.

HEPC, other than being a highly conserved antimicrobial β-defensin-like peptide, has a dual role in innate immunity: it enhances intracellular sequestration of iron by degrading the membrane iron transporter ferroportin (thereby depriving pathogens of this essential mineral) and it has direct antimicrobial activity against bacteria and viruses (Prowle et al. [25]; Wu et al. [31]; Zeng et al. [32]). This could help explain the significant elevations in serum HEPC in septic AKI patients, i.e. when its antimicrobial actions are stimulated/essential, and its decline after antibiotic treatment.

The current study also reports that the ischemic effect of AKI caused a decreasing effect on the PTEN and an increasing one on pAKT. PTEN is among the proteins modified by TGFβ1 and is a biological antagonist of PI3K signaling. Matsuda et al. ([19]) clarified the antagonistic relationship between PTEN and AKT and stated that PTEN has dual lipid and tyrosine-specific phosphatase activities that, if not physiologically regulated, could fundamentally alter cell behavior. Lan et al. ([16]) found that tubule PTEN was suppressed when TGFβ1 signaling after AKI was persistently elevated and investigated its role in tubule repair. They suggested that PTEN deficiency interfered with physiological recovery of tubules regeneration after injury and resulted in tissues that were dysfunctionally profibrotic. In contrast, PTEN recovery was associated with epithelial differentiation, normal tubule repair and less fibrosis.

Loss of PTEN function also leads to over-activation of AKT via activation of NF-κB and TNFα that, in turn, represses PTEN expression (Engelman et al. [4]; Carracedo & Pandolfi [2]). Moreover, the elevated levels of pAKT that secondarily evolved as a result of bacteria-induced sepsis, also impact on EPO levels. The relationships between AKT and EPO-reported to have a protective effect against kidney ischemia-reperfusion injury – are based, in part, on the inhibitory impact of EPO on pAKT and NF-κB (Li et al. [17]). When taking together, these inter-connected expression patterns suggesting that pAKT, PTEN and TNFα and their associated cell-signaling pathways are involved in the pathology of sepsis-AKI. These also reiterate the reno-protective role of EPO, on one hand, could be due in great part to the inhibition of AKT phosphorylation, and decreases in NF-κB – thereby affecting the expression of TNFα and mitigating inflammation.

This study concludes that there is a strong involvement of HEPC, EPO and HIF-1α in the pathogenesis of sepsis-induced AKI. Furthermore, the data provided here suggest the possible utility of their inclusion as predictive markers for the diagnosis of sepsis-induced AKI in larger cohort studies. We conclude this while at the same time acknowledging the small sample size included in the study (due to the difficulty in recruitment of patients in the ICU setting) needs to be expanded to strengthen our claims here. Lastly, based on the results here, we conclude that while HEPC, EPO and HIF-1α are involved in the pathogenesis of sepsis-induced AKI, the effects of inflammatory determinants dominate over hypoxia-related complications. This conclusion is based upon the inability of the AKI-induced increases in HIF-1α to inhibit HEPC levels versus the ability of the induced inflammation to stimulate these levels.

The authors would like to acknowledge the guidance and help provided by Dr. Mohamed Ahmed, Nephrology Department, Kasr Eleini Teaching Hospitals at Cairo University, Cairo, Egypt.

The authors report no conflicts of interest. The authors alone are responsible for the content of this manuscript.