TNF-α-induced Inflammation Stimulates Apolipoprotein-A4 via Activation of TNFR2 and NF-κB Signaling in Kidney Tubular Cells

Apo-A4 expression was increased in tissues from chronic kidney disease (CKD) patients compared to that in normal kidney tissue. We determined the association of apo-A4 and its regulatory signals following acute kidney injury and elucidated the effects of apo-A4 on cell signaling pathways related to kidney injury in vitro and in vivo. Tumor necrosis factor (TNF)-α, which causes inflammatory cell injury, induced significantly increased expression of apo-A4 protein levels, and these levels were related to pro-inflammatory acute kidney injury in human kidney cells. Apo-A4 expression was also increased in experimented rat kidney tissues after ischemic reperfusion injury. The expression of tumor necrosis factor receptor (TNFR) 2 was increased in both kidney cell lines and experimented rat kidney tissues following acute kidney injury. The expression of apo-A4 and TNFR2 was increased upon treatment with TNF-α. Immunohistochemistry revealed positive apo-A4 and TNFR2 staining in ischemic reperfusion injury rat kidneys compared with levels in the sham operation kidneys. After neutralization of TNF-α, NF-κB expression was only observed in the cytoplasm by immunofluorescence. Therefore, the apo-A4 expression is increased by stimulation of injured kidney cells with TNF-α and that these effects occur via a TNFR2-NFκB complex.

to demonstrate the association of apo-A4 and regulatory signals of apo-A4 following acute kidney injury (AKI) and elucidate the effects on cell signaling pathways related to kidney injury in vitro and in vivo.

Results
Localization and expression of apo-A4 in kidney cell lines and human kidney tissues. To investigate the expression of apo-A4 in the kidney, we used two kidney cancer cell lines (caki-1 and caki-2) and two normal kidney cell lines (HK-2 and HEK-293). We measured apo-A4 expression levels in these cells using western blotting (Fig. 1A). We observed apo-A4 protein expression in all four kidney cell lines. Also, we examined apo-A4 expression in normal kidney and CKD tissues by immunohistochemistry (IHC) using a mouse antiapo-A4 antibody. IHC revealed apo-A4 expression in normal and CKD tissues (Fig. 1B). We detected apo-A4 expression in the tubules as well as in the glomeruli, and interestingly, the intensity of apo-A4 expression in CKD tissue was stronger than in healthy kidney tissue.
Apo-A4 expression in kidney cell lines following renal injury model. To determine which type of renal injury model affects apo-A4 expression in kidney cells, we exposed human kidney cells to conditions known to induce an AKI. We incubated Caki-1, HK-2, and HEK293 cells in a hypoxic chamber for 72 h ( Fig. 2A); yet, we detected no significant expression of apo-A4 in any of these cell-lines as a result. We induced chemical hypoxia using CoCl 2 (0, 100, and 200 µM) for 8, 16, and 24 h in HK-2 cells, but these conditions also had no effect on apo-A4 expression in these cell types (Fig. 2B). Also, we treated HK-2 cells with the nephrotoxic cancer drug cisplatin; the calcium ionophore A23187, which induces calcium influx that causes apoptosis; and the pro-inflammatory cytokines TNF-α and IL-1β, which are known to cause inflammatory injury 9,10 . Following treatment, we measured apo-A4 protein expression levels by western blotting (Fig. 2C). While cisplatin and IL-1β failed to modify the expression of apo-A4 in these cells, TNF-α induced significantly increased expression of apo-A4. Although we observed an increase in apo-A4 following treatment with A23187, this change was not significant. Immunofluorescence assays revealed an increase in apo-A4 expression 24 h after TNF-α treatment compared with levels in the untreated control group (Fig. 2D). Furthermore, the expression of apo-A4 in the nucleus of HK-2 cells was increased in accord with TNF-α treatment. Thus, of the renal injury agents tested, only TNF-α significantly affects the expression of apo-A4.

Inflammatory AKI induced by TNF-α increased apo-A4 Expression.
To examine whether AKI induced by TNF-α increased apo-A4 expression in vivo, we induced ischemic reperfusion injury in rat kidneys (Fig. 3A) 7 . We removed kidney tissue 48 h after reperfusion and isolated protein lysates to measure apo-A4 expression levels by western blotting. Apo-A4 protein expression was elevated in ischemic reperfusion injury tissue compared with levels in control tissue (Fig. 3B). Also, we isolated serum samples from these rats and confirmed the increased apo-A4 levels by enzyme-linked immunosorbent assay (ELISA). Data presents 10 samples,  and were shown as mean ± SEM and analyzed by students t-test (*P ≤ 0.05; Fig. 3C). Together, these findings indicate that apo-A4 expression is increased by inflammatory AKI in vitro and in vivo.

Inflammatory AKI induced by TNF-α increases apo-A4 via TNF Receptor 2 (TNFR2). Expression
of TNFR2 increases in tubular epithelial cells during ischemic reperfusion injury 11 . Therefore, we measured the expression of apo-A4 and TNFR2 upon induction of AKI. The expression of apo-A4 and TNFR2 were increased upon treatment of TNF-α as assessed by western blotting (Fig. 4A). Also, we detected the expression of apo-A4 and TNFR2 by IHC in rat kidney tissue. Furthermore, IHC revealed positive apo-A4 and TNFR2 staining in ischemic reperfusion injury rat kidneys compared with the sham operation kidneys (Fig. 4B). These experiments revealed a positive correlation between the expression of apo-A4 and TNFR2 following treatment with the pro-inflammatory cytokine TNF-α.

Increased apo-A4 expression caused by inflammatory AKI is associated with the NF-κB
Complex. The NF-κB complex plays a major role in inflammation 12 . We treated HK-2 cells with TNF-α, TNF-α antibody and then measured IκB and apo-A4 expression by western blotting. We observed an inverse correlation between apo-A4 and IκB expression levels ( Fig. 5A). Western blotting revealed expression in apo-A4 increased by TNF-α concentration. When 100, 200 ng/ml of TNF-α were administered, apo-A4 increased by 1.45 and 1.2. Moreover, apo-A4 levels decreased significantly to 0.41 for TNF-α antibody treated group. The expression of IκB was similar to control when treated TNF-α antibody (Fig. 5A). Also, we neutralized TNF-α using a specific TNF-α antibody and then examined NF-κB localization using an immunofluorescence assay. NF-κB expression was only observed in the cytoplasm after neutralization of TNF-α in comparison with localization of this cytokine in untreated HK-2 cells (Fig. 5B). This finding suggests that NF-κB translocation is promoted by TNF-α, which elevates apo-A4 expression levels.

Discussion
AKI occurs in the initial stages of kidney disease and is followed by CKD 13 . Previous work has demonstrated that apo-A4 expression in kidney was significantly elevated during AKI and that serum apo-A4 increases during the earliest phases of renal insufficiency [3][4][5] . We had previously shown that apo-A4 expression levels are strongly reduced in pig kidneys when ischemic preconditioning was performed prior to renal ischemic reperfusion injury 14 . Also, we found that the down-regulation of apo-A4 expression affected the recovery of kidney cells from ischemic reperfusion injury 14 . Thus, we hypothesized that kidney cell injury is influenced by increased apo-A4 expression level just after AKI. To test this hypothesis, we examined the role of apo-A4 and the apo-A4 associated pro-inflammatory pathway under conditions of AKI.
First, we observed endogenous apo-A4 expression levels in kidney cell lines and tissues. Apo-A4 has been reported to be expressed in proximal and distal tubular cells, capillaries, and blood vessels but not in the glomeruli of normal kidneys; however, we observed apo-A4 expression in kidney tubular cells as well as in the glomeruli of normal and injured kidney tissues 1,4 . We anticipated apo-A4 accumulation in the kidney during ischemic reperfusion injury, and indeed, apo-A4 expression levels were elevated in CKD tissue compared with those in normal kidney tissue 4 . This result indicates that apo-A4 is up-regulated in response to kidney injury. The accumulation of apo-A4 has been reported in the blood plasma and urine of CKD patients during initial phases of the disease, suggesting that apo-A4 expression can predict the progression of kidney disease 2, 3, 5, 8 .
Also, we induced injury in human kidney cells using different AKI conditions. AKI is caused by decreased renal blood flow (ischemia of kidney tissue), exposure to harmful substances, inflammatory processes, or urinary tract obstructions, which impede the flow of urine 15,16 . To mimic renal ischemia, we exposed human kidney cells to physical and chemical hypoxic conditions. Additionally, we induced nephrotoxicity by treating kidney cells with cisplatin and induced apoptosis by treatment of kidney cells with the calcium ionophore (A23187), which increases intracellular Ca 2+ levels 17,18 . We did not observe any correlation between apo-A4 expression and renal injury model induced by cisplatin and A23187.
Finally, we investigated the effect of inflammatory injury on apo-A4 expression. Inflammatory processes comprise a critical response to pathogenic infection; however, the excessive production of inflammatory cytokines, such as TNF-α and IL-1β, may induce cell injury. TNF-α is a group of pro-inflammatory cytokines implicated in a wide variety of biological processes. This cytokine is produced by multiple cell types, including macrophages, lymphocytes, fibroblasts, and keratinocytes, as a response to inflammation, infection, and other environmental stresses 17,19 . We found that the increase in apo-A4 expression level caused by TNF-α was related to pro-inflammatory AKI in human kidney cells. In vivo, we induced ischemic reperfusion injury in rat kidneys to investigate the effect of apo-A4 expression on specific conditions of inflammatory kidney injury. We invoked ischemic reperfusion injury using the clamping of renal hilar vessels (renal artery and renal vein) to elicit an inflammatory reaction in rat kidneys 7 . We provoked reperfusion injury and isolated protein lysates to measure apo-A4 expression levels by western blotting. Apo-A4 protein expression was elevated in ischemic reperfusion injury tissue compared with levels in control tissue (Fig. 3B). Also, we isolated serum samples from these rats and confirmed the increased apo-A4 levels by enzyme-linked immunosorbent assay (ELISA). Data presents 10samples were shown statistically significant (*P ≤ 0.05; Fig. 3C). In vitro and in vivo, these findings indicate that inflammatory AKI increases apo-A4 expression.
TNFR1 acts like cell death, altered target gene transcription, and cytokine production. TNFR2 interacts either in cooperation with or independently with TNFR1, acts as cell proliferation and cell survival. TNFR2 may have an antiapoptotic effect acting through an NF-kB pathway that results in exacerbated expression of several inflammatory mediators 20 . TNFR2 plays an important role during the inflammatory process 12,19 . This receptor was up-regulated during intestinal inflammation, and concordantly, less intestinal inflammation was observed in TNFR2-deficient mice 11,19,21 . We demonstrated that apo-A4 expression was increased during inflammatory injury stimulated by TNF-α and that the expression of TNFR2 was increased in both kidney cell lines and experimented rat kidneys following AKI. Therefore, inflammatory injury stimulated by TNF-α induced TNFR2 activation. TNF-α-mediated activation of the NF-κB complex initiates phosphorylation and proteasomal degradation of IκB within the cytoplasm, promoting nuclear translocation of the p65 subunit of NF-κB and enabling heterodimeric NF-κB to enter the nucleus. In the nucleus, NF-κB binds to enhancers of its target genes 12,22,23 . In our research, TNF-α inhibited IκB and activated NF-κB. Also, the expression of Apo-A4 decreased when IκB activity was high (Fig. 5A). TNF-α was found to be a major factor in the expression of apo-A4, because of the apo-A4 lower than the normal control when TNF-α antibody was administered.
We conduct an additional experiment that NF-κB regulates apo-A4 expression. It has reported that NF-κB inhibitors attenuate the inflammatory response in injured kidney animal model 24,25 . Pyrrolidine dithio-carbamate ammonium (PDTC) is known to inhibit the NF-κB system and attenuated renal damage and inflammation 24,25 . Apo-A4 was decreased when HK-2 cell treated by PDTC, compared to control group ( Supplementary Fig. S4A,B). Apo-A4 was increased in the AKI model treated with TNF-α. The administration of TNF-α and PDTC in HK-2 cell attenuated apo-A4 expression ( Supplementary Fig. S4A,B). Based on these results, we proved that NF-κB regulates apo-A4.
We showed that inflammatory damage induced TNFR2 activation via NF-κB in human kidney cell lines and tissues, and this finding is supported by the fact that the promoter region of TNFR2 contains binding sites Figure 5. Inflammatory AKI increases apo-A4 expression through the NF-κB complex. (A) Western blotting revealed an inverse relationship between apo-A4 and IκB protein expression in HK-2 cells following TNF-α treatment. The loading control was GAPDH and the normalized ratio of each extract pair (A set to 1) determined using Image Jv1.42q software (http://rsb.info.nih.gov/ij). (B) Immunofluorescence assay for apo-A4 and NF-κB expression in HK-2 cells after 24-h TNF-α treatment. TNF-α antibody was used to neutralize the TNF-α effect. NF-κB (green) immune-localization is shown. And NF-κB regulates apo-A4 (see Supplementary Fig. S4).
Furthermore, we suggest that continuous kidney damage induced by inflammatory AKI disrupts apo-A4 metabolism, which promotes apo-A4 deposition in kidney cells, possibly leading to irreversible kidney injury or amyloidosis 4 . The mechanisms of apo-A4 function and metabolism in injured kidney tubular cells, however, remain unclear (Fig. 6B). Also, the factor that causes apo-A4 deposition during CKD requires identification 3,4,13 . Further investigation is needed to determine whether the protective role of apo-A4 is specific to NF-κB, and our results provide encouragement for further examinations of the therapeutic effects of apo-A4 in injured kidney cells.
In summary, we demonstrated that apo-A4 expression is increased by stimulation of injured kidney cells and tissues with TNF-α and that these effects occur via a TNFR2-NF-κB complex.

Materials and Methods
Human Tissues. Normal or ischemic human renal specimens were collected with the informed consent of patients according to the Institutional Review Board at Yonsei University (College of Medicine, Yonsei University), and the study was approved by the ethics committee. All human kidney biopsies were stored in liquid nitrogen.
Animal Models. The study was approved by the Institutional Animal Care and Use Committee of Yonsei University Health System in accordance with the 'Guide for the Care and Use of Laboratory Animals' published by the National Institutes of Health and was conducted according to the principles of the Declaration of Helsinki(2015-0149). Eight-week-old male Sprague-Dawley rats were used for the renal ischemic reperfusion injury model. For the single kidney model, we performed right nephrectomy, and we initiated left kidney ischemia using renal hilar clamping for 30 min followed by reperfusion. Rats were sacrificed 48 h after ischemia-reperfusion of the kidney, and kidney tissues and blood samples were collected and stored in liquid nitrogen 7 .
Cell Culture. HK-2 cells were cultured in keratinocyte serum-free medium (K-SFM, Gibco, Grand Island, NY, USA) including 0.05 mg/ml bovine pituitary extract, 5 ng/ml human recombinant epidermal growth factor, 10% fetal bovine serum, and 1% penicillin/streptomycin under an atmosphere of 5% CO2 and 95% O2 at 37 °C. Cells were subcultured every 7 days using 0.02% ethylenediaminetetraacetic acid and 0.05% trypsin. The medium was replaced with fresh medium every 2 days. Western blot. Total cellular protein extracts were prepared on ice using a PRO-PREP protein extract solution (Intron, Seoul, Korea). Cell lysates were loaded onto a sodium dodecyl sulfate-polyacrylamide gel and transferred to polyvinylidene fluoride membrane for 1 hour. Membranes were incubated at 4 °C overnight with HIF-1α, apo-A4, TNFR2, IκB, and α-tubulin primary antibodies (Abcam, Cambridge, MA, USA) diluted 1:1000 with 5% BSA Tris-buffered saline-Tween 20 (TBS-T). After incubation, the membranes were washed with TBS-T and then secondary antibodies (1:10000, horseradish peroxidase-conjugated anti-mouse IgG, anti-rabbit IgG) were added for incubation at room temperature for 1 hour. Labeled bands were detected by West Pico chemiluminescent kit (Thermo Scientific, Rockford, IL, USA). Immunofluorescence Assay. Cells were washed in phosphate-buffered saline (PBS) and fixed in 4% paraformaldehyde. Cells were subjected to permeation in 0.5% Triton X-100/ PBS for 5 min and then blocked in 5% BSA/PBS for 20 min. Cells were then incubated with mouse anti-apo-A4 antibody (Abcam) at 37 °C for 1 h and then incubated with Alexa Fluor 488 anti-mouse secondary antibody (1:200, Invitrogen) at 4 °C for 1 h. Images were acquired by confocal microscope (LSM Meta 700, Carl Zeiss, Oberkochen, Germany) and analyzed with LSM Image Browser software.

IHC.
Kidney tissue sections were incubated in formalin and treated overnight with 0.1 M citrate buffer antigen retrieval solution (pH 6.0; Dako, Carpinteria, CA, US) at 85 °C. Endogenous peroxidase was blocked with 3% H 2 O 2 for 5 min. After blocking in 1% BSA in Tris-HCl buffer (BSA-T, pH 8.6) for 1 h, sections were incubated with the anti-human apo-A4 primary antibody (Abcam) diluted 1:400 in BSA-T for 1 h at room temperature. Then, specific immunoreaction was detected using a secondary horseradish peroxidase-conjugated goat anti-rabbit antibody (Dako). After washing in Tris-HCl buffer, 3-3-diaminbenzidine tetrahydrochloride (DAB + , Dako) was used for visualization. Sections were counterstained with hematoxylin, dehydrated, mounted in DePeX (Serva, Heidelberg, Germany), and examined with a microscope.
ELISA. Rat plasma apo-A4 concentrations were determined using a double-antibody ELISA. The amounts of apo-A4 in the rat plasma were estimated using the apo-A4 rat ELISA kit (Mybiosource, Vancouver, Canada) according to the manufacturer's instructions.

Statistical analysis.
Quantitative values were expressed as mean ± S.E.M. Statistical differences were compared using the Kruskal-Wallis test and Mann-Whitney U-test with SPSS software version 23.0 (IBM SPSS Statistics, IBM Corp., Armonk, N.Y., United States). P < 0.05 was considered statistically significant.