Archives of Toxicology

, Volume 89, Issue 1, pp 101–106 | Cite as

Interleukin-19 as a translational indicator of renal injury

  • Paul Jennings
  • Daniel Crean
  • Lydia Aschauer
  • Alice Limonciel
  • Konrad Moenks
  • Georg Kern
  • Philip Hewitt
  • Karl Lhotta
  • Arno Lukas
  • Anja Wilmes
  • Martin O. Leonard
Organ Toxicity and Mechanisms

Abstract

Accurate detection and prediction of renal injury are central not only to improving renal disease management but also for the development of new strategies to assess drug safety in pre-clinical and clinical testing. In this study, we utilised the well-characterised and differentiated human renal proximal tubule cell line, RPTEC/TERT1 in an attempt to identify markers of renal injury, independent of the mechanism of toxicity. We chose zoledronate as a representative nephrotoxic agent to examine global transcriptomic alterations using a daily repeat bolus protocol over 14 days, reflective of sub-acute or chronic injury. We identified alterations in targets of the cholesterol and mevalonate biosynthetic pathways reflective of zoledronate specific effects. We also identified interleukin-19 (IL-19) among other inflammatory signals such as SERPINA3 and DEFB4 utilising microarray analysis. Release of IL-19 protein was highly induced by an additional four nephrotoxic agents, at magnitudes greater than the characterised marker of renal injury, lipocalin-2. We also demonstrate a large increase in levels of IL-19 in urine of patients with chronic kidney disease, which significantly correlated with estimated glomerular filtration rate levels. We suggest IL-19 as a potential new translational marker of renal injury.

Keywords

Nephrotoxicity Chronic kidney disease Proximal tubule Biomarker 

Introduction

The search for new and better predictive markers of nephrotoxicity, particularly those which can predict an adverse effect before any long lasting damage occurs, is a fundamental strategy behind many studies involved not only in disease monitoring but also those seeking to build better systems for pre-clinical and clinical nephrotoxicity compound screening. Animal studies are used on a routine basis to assess organ specific injury as part of drug development and disease investigative programs. Despite the potential for species specific differences, these models have played a significant role in the identification and development of relevant biomarkers such as lipocalin-2 (LCN2) a protein marker of renal injury (Kieran et al. 2003; Mishra et al. 2003). It is important to understand that these models have limitations in their ability to reflect and predict tissue specific injury in humans (Olson et al. 2000), which has led investigators to explore other model systems to model tissue injury and explore biomarker development. Particular attention has been given to developing better in vitro culture of highly differentiated human cells.

Within the kidney, the proximal tubule plays a major role in the reabsorption and secretion of key substances such as glucose, amino acids and certain metabolites. It is also particularly sensitive to injury and has a prominent role in the pathology of many conditions where the kidney is a target of injury (Thevenod 2003; Vallon 2011; Witzgall 1999). Primary cell culture models represent in vivo function to a high degree but are limited by their loss of differentiation over time and their limited proliferative capacity (Brown et al. 2008; Courjault-Gautier et al. 1995; Detrisac et al. 1984; Sharpe and Dockrell 2012). This has led to the development of many immortalised human proximal tubular epithelial cells such as the RPTEC/TERT1 cell line (Wieser et al. 2008), which represents one of the most similar to primary culture and highly differentiated cell line available (Aschauer et al. 2013; Wieser et al. 2008). These cells also have the added benefit of forming stable differentiated monolayers capable of being maintained over many weeks allowing previously difficult long-term exposure protocols to be performed with relative ease (Limonciel et al. 2011; Wilmes et al. 2013). The use of cells with a highly differentiated phenotype to study human biomarker development represents a complementary approach to those currently applied to clinical sample sets. Distinct advantages include high throughput and the ability to use global molecular screening or “Omics” strategies in well-controlled experimental designs that not only identify important targets for further validation but also give important mechanistic insight, which is often difficult to ascertain from in vivo animal and human studies.

The aim of this study was to investigate sub-acute or chronic toxic exposure in RPTEC/TERT1 cells up to 14 days in order to identify markers that could represent or predict toxic exposure in vivo. Using the bisphosphonate zoledronate as a model stimulus of renal injury, transcriptomic analysis identified interleukin-19 (IL-19) as the most highly differentially regulated gene. Moreover, IL-19 could be detected in the supernatant medium and displayed time and concentration dependent profiles. IL-19 could also be measured in urine samples from patients with chronic kidney disease (CKD) and correlated with renal function and other urinary markers of renal injury including NAG and LCN2. This work is an important example of how identification of a target molecule in highly differentiated in vitro culture can be used to explore new and better markers of human renal injury.

Results

For a detailed description of the methods used in this article, please refer to the online supplementary information. Initially, an IC10 for viability of RPTEC/TERT1 cells at 14 days treatment on filter supports for zoledronate was calculated using a combination of lactate and resazurin measurements as 15 μM (data not shown) and was used to assess changes in gene transcription at 1, 3 and 14 days of the treatment protocol. A no observable effects limit (NOEL) of 5 μM was also used. Separate arrays were used for each biological replicate (N = 3), for each time point and treatment condition including matched day controls (27 arrays in total). A profile of those genes which were differentially expressed at low and high doses of zoledronate is displayed (Fig. 1a). Examination of high-dose differential expression at 3 days revealed a large number of genes involved in inflammatory processes including IL-19, SERPINA3, β-defensins, lymphotoxin-β and complement component 4BPα (Fig. 1b). For comprehensive details of differentially expressed genes across all treatments, see online supplementary information. Ingenuity pathway analysis of day 3 differentially expressed genes again revealed a high significance of inflammatory processes including granulocyte migration and IL-17-induced genes (Fig. 1c—left panel). Cholesterol biosynthetic pathways were identified in both analyses at day 3 and day 14 (Fig. 1c—right panel). Zoledronate has its pharmacological action through inhibition of the mevalonate arm of the cholesterol biosynthesis pathway through farnesyl diphosphate synthase (FDPS) inhibition (Russell 2011) and IPA-predicted activation of this pathway may be considered as part of a feedback mechanisms to restore mevalonate levels (Fig. 1c).
Fig. 1

Zoledronate induced alterations in global gene expression. Differentiated cells cultured on aluminium oxide porous trans-well filters were exposed to two concentrations of zoledronate (5 and 15 μM) in a repeat dosing protocol every 24 h for either 1, 3 or 14 days. Messenger RNA was isolated from three independent experiments and assessed for gene expression alterations using microarray analysis. Genes with expression significantly different (p value) from matched day control values are displayed as fold change versus significance. a The top ten up-regulated genes at day 3 of treatment by zoledronate (15 μM) are displayed ranked based on log2 fold change compared with control (LFC). b IPA analysis of those differentially regulated genes at day 3 and 14 by zoledronate (15 μM) was performed at a cut-off of 1.5-fold change. Predicted canonical pathways are displayed. c Statistics for presentation are based on the Fisher’s exact test (p value) and ratio

We explored IL-19 as a predictive indicator of chronic nephrotoxicity by analysing protein release into the cell culture media using ELISA not only for zoledronate but also for 4 other nephrotoxic compounds. Interestingly, we observed an increase in IL-19 production up to 30-fold at 3 days exposure (Fig. 2a) paralleling observations from our microarray analysis. Concentrations of adefovir dipivoxil, cisplatin, cidofovir and ifosfamide which did not exceed IC10 significantly increased IL-19 protein release. LCN2 has been used extensively as an indication of nephrotoxic potential both in vivo and in vitro. LCN2 mRNA was induced by zoledronate at 3 days exposure from the microarray analysis but not to the same extent as observed for IL-19 (data not shown). Zoledronate and adefovir dipivoxil exposure resulted in an increased supernatant LCN2 content but again this was not to the same extent as observed for IL-19 protein levels (Fig. 2). We also examined the release of SERPINA3 in response to these compounds as it was highly regulated by zoledronate on microarray analysis. We observed increases in release of this protein only in response to zoledronate. We did not pursue this molecule as a general marker of renal injury.
Fig. 2

Nephrotoxin induced release of interleukin-19 protein. Differentiated cells cultured on trans-well filters were exposed to zoledronate (15 μM), adefovir dipivoxil (8 μM), cisplatin (2 μM), cidofovir (46 μM) or ifosfamide (5,000 μM) in a repeat dosing protocol every 24 h for either 1, 3, 5, 7 or 14 days (X axis labels). At these time points, equal volumes of cell culture media from apical and basolateral sides of the culture monolayer were combined and analysed for interleukin-19 (IL-19). a Lipocalin-2 (LCN2) b or alpha-1-antichymotrypsin (SERPINA3) c protein expression using ELISA. Levels were expressed as fold over average untreated control (F.O.C.) mean ± SEM for three independent experiments, and statistical significance was calculated using one-way ANOVA with Dunnett’s post-test, where asterisk indicates a p value <0.05, **0.01, ***0.001 and ****0.0001 for comparisons to control

Having established the potential for IL-19 to act as a marker of chronic toxicity in vitro, we next wanted to explore whether these observations could be translated to clinical observations of chronic kidney disease (CKD). We therefore investigated a population of patients with CKD in comparison to healthy volunteers. Estimated glomerular filtration rate in CKD patients was significantly reduced when compared to control patients (Fig. 3a). Typical of individuals with compromised renal function, we observed an increase in the levels of urinary protein and NAG (Fig. 3a). Also as previously described, we observed a significant increase in urinary LCN2 in the CKD population (Fig. 3a). When we examined urinary levels of IL-19, we also observed a significant increase in CKD patients (Fig. 3a). Correlations were then examined between levels of LCN2 and IL-19 and all other parameters (Fig. 3b). IL-19 displayed an inverse correlation to estimated glomerular filtration rate to a greater statistically significant level than LCN2. Both LCN2 and IL-19 correlated with other markers of renal function including urinary protein and NAG, another marker indicative of proximal tubular injury. There was also a strong correlation between IL-19 and LCN2 urinary levels.
Fig. 3

Urinary interleukin-19 as an indicator of renal function. Healthy volunteers’ (CTRL) and chronic kidney disease (CKD) patients’ urinary levels of IL-19, protein, NAG and LCN2 together with serum creatinine were determined as described in the methods (see Supplementary information) and expressed as mean ± SEM of values expressed per g of urinary creatinine. Estimated glomerular filtration rate (eGFR) was also determined. a Comparisons were made between CTRL and CKD patients for each of these parameters, and statistical differences were calculated using the nonparametric Mann–Whitney test where asterisk indicates a p value <0.05, **0.01, ***0.001 and ****0.0001 for comparisons to CTRL. b Levels of LCN2 and IL-19 were correlated with all other parameters using the Spearman’s statistical test. R and the p values are given

Discussion

New strategies are being continuously developed to identify more specific and predictive markers of renal injury. Targets are ideally released into the extracellular environment allowing detection in readily accessible bio-fluids such as urine. The characteristics of such markers are typical of those peptide signals found as part of inflammatory responses. Bearing this in mind, using our highly differentiated in vitro model, we examined the nephrotoxic effects of zoledronate across a dosing regimen of IC10 and NOEL concentrations at 14 days to reflect or recapitulate sub-acute or chronic injury responses. Using transcriptomic analysis, we identified a number of pathways and genes with biomarker potential and focused on interleukin-19, an inflammatory peptide as the highest regulated gene on transcriptomic analysis.

Interleukin-19 is a member of the IL-10 family of cytokines which have indispensable functions in many inflammatory processes (Ouyang et al. 2011). It is also a member of the IL-20 subfamily and binds to the IL-20 receptor to regulate processes such as anti-microbial, wound healing and tissue remodelling (Ouyang et al. 2011). Cellular sources have been characterised as of myeloid and epithelial origin fitting with the significant levels found in the RPTEC/TERT1 renal epithelial cell system. The precise function of IL-19 in our system is difficult to ascertain given the in vitro nature of the model, but we have demonstrated that recombinant IL-19 protein induces the release of IL-6 and TNF-α from these cells (data not shown). The signalling mechanisms responsible for the induction of IL-19 were suggested by IPA analysis to include IL-17. Interestingly, IL-17 when used in combination with TNF-α has been observed to induce many of the genes altered in our model system, including IL-19, DEFB4 and LCN2 (Johnston et al. 2013). While insights into how IL-19 is regulated are important, in the context of biomarker discovery, it is equally important to characterise whether such signals can be used as predictive markers of tissue specific injury. Evidence to support our hypothesis that IL-19 expression in renal cells can reflect toxicity, independent of the type of injury or dysfunction, has recently been reported (Hsu et al. 2013). This study demonstrated increased expression of IL-19 using in vivo mouse models of ischaemia reperfusion and HgCl2-induced acute kidney injury. This study, however, did not look at changes in IL-19 in any clinical samples.

When we examined urine from patients with chronic kidney disease, we demonstrated a large increase in IL-19 protein levels as compared to control volunteers, with levels of change similar to other urinary markers of renal damage such as protein, NAG and LCN2. We also observed a more significant negative correlation of IL-19 with estimated GFR than when compared to LCN2, the implication of which is interesting to speculate upon. From our urinary analysis, it is unclear what the cellular source for IL-19 is but as NAG levels are increased in these samples indicating proximal tubular injury, it is possible these cells may be acting as a source. Some of our unpublished observations have demonstrated lack of IL-19 expression from liver- and CNS-derived cells in response to toxic stimuli, hinting at some degree of renal specificity. This, however, would have to be explored in greater detail. Levels of IL-19 expression from RPTEC/TERT1 cells seemed to precede general toxicity (data not shown) but whether this is the case in human urinary samples is unclear. We can say, however, that based on our results that it is very likely IL-19 is a novel clinical marker of ongoing proximal tubular injury. Lastly, it should also be highlighted that translation of targets identified in vitro to an in vivo human patient cohort gives great confidence to future studies which aim to use such cellular models to predict renal injury in humans.

Notes

Acknowledgments

This project was funded by the European Union’s 7th Framework Programme (FP7/2007–2013) under grant agreement No. 202222, Predict-IV. The funding agency had no input into study design and in the decision to publish.

Conflict of interest

The authors declare no conflicts of interest.

Supplementary material

204_2014_1237_MOESM1_ESM.xlsx (6.9 mb)
Supplementary material 1 (XLSX 7077 kb)
204_2014_1237_MOESM2_ESM.docx (36 kb)
Supplementary material 2 (DOCX 35 kb)
204_2014_1237_MOESM3_ESM.xlsx (40 kb)
Supplementary material 3 (XLSX 40 kb)

References

  1. Aschauer L, Gruber LN, Pfaller W et al (2013) Delineation of the key aspects in the regulation of epithelial monolayer formation. Mol Cell Biol 33(13):2535–2550. doi:10.1128/MCB.01435-12 PubMedCentralPubMedCrossRefGoogle Scholar
  2. Brown CD, Sayer R, Windass AS et al (2008) Characterisation of human tubular cell monolayers as a model of proximal tubular xenobiotic handling. Toxicol Appl Pharmacol 233(3):428–438. doi:10.1016/j.taap.2008.09.018 PubMedCrossRefGoogle Scholar
  3. Courjault-Gautier F, Chevalier J, Abbou CC, Chopin DK, Toutain HJ (1995) Consecutive use of hormonally defined serum-free media to establish highly differentiated human renal proximal tubule cells in primary culture. J Am Soc Nephrol: JASN 5(11):1949–1963PubMedGoogle Scholar
  4. Detrisac CJ, Sens MA, Garvin AJ, Spicer SS, Sens DA (1984) Tissue culture of human kidney epithelial cells of proximal tubule origin. Kidney Int 25(2):383–390PubMedCrossRefGoogle Scholar
  5. Hsu YH, Li HH, Sung JM, Chen WT, Hou YC, Chang MS (2013) Interleukin-19 mediates tissue damage in murine ischemic acute kidney injury. PLoS One 8(2):e56028. doi:10.1371/journal.pone.0056028 PubMedCentralPubMedCrossRefGoogle Scholar
  6. Johnston A, Fritz Y, Dawes SM et al (2013) Keratinocyte overexpression of IL-17C promotes psoriasiform skin inflammation. J Immunol 190(5):2252–2262. doi:10.4049/jimmunol.1201505 PubMedCentralPubMedCrossRefGoogle Scholar
  7. Kieran NE, Doran PP, Connolly SB et al (2003) Modification of the transcriptomic response to renal ischemia/reperfusion injury by lipoxin analog. Kidney Int 64(2):480–492. doi:10.1046/j.1523-1755.2003.00106.x PubMedCrossRefGoogle Scholar
  8. Limonciel A, Aschauer L, Wilmes A et al (2011) Lactate is an ideal non-invasive marker for evaluating temporal alterations in cell stress and toxicity in repeat dose testing regimes. Toxicol In Vitro 25(8):1855–1862. doi:10.1016/j.tiv.2011.05.018 PubMedCrossRefGoogle Scholar
  9. Mishra J, Ma Q, Prada A et al (2003) Identification of neutrophil gelatinase-associated lipocalin as a novel early urinary biomarker for ischemic renal injury. J Am Soc Nephrol: JASN 14(10):2534–2543PubMedCrossRefGoogle Scholar
  10. Olson H, Betton G, Robinson D et al (2000) Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul Toxicol Pharmacol 32(1):56–67. doi:10.1006/rtph.2000.1399 PubMedCrossRefGoogle Scholar
  11. Ouyang W, Rutz S, Crellin NK, Valdez PA, Hymowitz SG (2011) Regulation and functions of the IL-10 family of cytokines in inflammation and disease. Annu Rev Immunol 29:71–109. doi:10.1146/annurev-immunol-031210-101312 PubMedCrossRefGoogle Scholar
  12. Russell RG (2011) Bisphosphonates: the first 40 years. Bone 49(1):2–19. doi:10.1016/j.bone.2011.04.022 PubMedCrossRefGoogle Scholar
  13. Sharpe CC, Dockrell ME (2012) Primary culture of human renal proximal tubule epithelial cells and interstitial fibroblasts. Method Mol Biol 806:175–185. doi:10.1007/978-1-61779-367-7_12 CrossRefGoogle Scholar
  14. Thevenod F (2003) Nephrotoxicity and the proximal tubule. Insights from cadmium. Nephron Physiol 93(4):p87–p93. doi:10.1159/000070241 PubMedCrossRefGoogle Scholar
  15. Vallon V (2011) The proximal tubule in the pathophysiology of the diabetic kidney. Am J Physiol Regul Integr Comp Physiol 300(5):R1009–R1022. doi:10.1152/ajpregu.00809.2010 PubMedCentralPubMedCrossRefGoogle Scholar
  16. Wieser M, Stadler G, Jennings P et al (2008) hTERT alone immortalizes epithelial cells of renal proximal tubules without changing their functional characteristics. Am J Physiol Renal Physiol 295(5):F1365–F1375. doi:10.1152/ajprenal.90405.2008 PubMedCrossRefGoogle Scholar
  17. Wilmes A, Limonciel A, Aschauer L et al (2013) Application of integrated transcriptomic, proteomic and metabolomic profiling for the delineation of mechanisms of drug induced cell stress. J Proteomics 79:180–194. doi:10.1016/j.jprot.2012.11.022 PubMedCrossRefGoogle Scholar
  18. Witzgall R (1999) The proximal tubule phenotype and its disruption in acute renal failure and polycystic kidney disease. Exp Nephrol 7(1):15–19. doi:10.1159/000020579 PubMedCrossRefGoogle Scholar

Copyright information

© Crown Copyright 2014

Authors and Affiliations

  • Paul Jennings
    • 2
  • Daniel Crean
    • 3
  • Lydia Aschauer
    • 2
  • Alice Limonciel
    • 2
  • Konrad Moenks
    • 4
  • Georg Kern
    • 2
  • Philip Hewitt
    • 5
  • Karl Lhotta
    • 6
  • Arno Lukas
    • 4
  • Anja Wilmes
    • 2
  • Martin O. Leonard
    • 1
  1. 1.Centre for Radiation, Chemical and Environmental HazardsPublic Health EnglandChilton, Didcot, OxonUK
  2. 2.Division of Physiology, Department of Physiology and Medical PhysicsInnsbruck Medical UniversityInnsbruckAustria
  3. 3.School of Medicine and Medical ScienceUniversity College DublinDublinIreland
  4. 4.Emergentec Biodevelopment GmbHViennaAustria
  5. 5.Merck KGaA, Merck Serono, Nonclinical SafetyDarmstadtGermany
  6. 6.Department of Nephrology and DialysisAcademic Teaching Hospital FeldkirchFeldkirchAustria

Personalised recommendations