ProxTox

HTS Predictive Nephrotoxicity Assay

A new tool for accurate in vitro high-throughput human nephrotoxicity prediction.

Description

ProxTox-HTS is a unique in vitro approach for nephrotoxicity prediction that combines high-throughput imaging of human renal proximal tubular cells (PTCs), or HK-2 cells, quantitative phenotypic profiling, and machine learning methods. The technology was developed by Dr. Daniele Zink and Dr. Lit-Hsin Loo at the Institute of Bioengineering and Nanotechnology and Bioinformatics Institute, A*STAR, Singapore. SOLVO Biotechnology is pleased to offer these tailored services in partnership with CellBae.

Cellbae Solvo Biotechnology

Features

  • Primary Proximal Tubule Cells or HK-2 cells
  • 384-well format
  • No predefined endpoints
  • Analysis of 129 cellular features through high-content imaging
  • Validated with 44 chemically diverse compounds

Advantages

  • True HTS due to 384 well format, low cost per datapoint
  • Prediction of toxicity based on cellular phenotypes, cell injury mechanisms do not need to be defined a priori. This is especially useful for building models for a diverse set of xenobiotic compounds that may injure similar tissues and cell types, but through different biochemical mechanisms
  • Thoroughly validated using machine learning
iPSC Derived PTC Like Cells

 

Scientific Background

Renal PTCs are particularly susceptible to xenobiotic-induced damage due to their role in compound transport, metabolism and clearance. The human kidney can be damaged by a wide range of xenobiotics, such as industrial chemicals, environmental toxicants and drugs. Xenobiotic-induced injuries impair cellular functions and lead to changes in cellular phenotypes, such as reorganizations and changes of cellular and subcellular structures.

The ProxTox-HTS platform was developed without using pre-defined endpoints, and instead predictive cellular changes were identified by phenotypic profiling after compound treatment.

Analysis of 129 cellular features was conducted by powerful high-content imaging and predictive features were identified by machine learning by comparing in vitro results to known clinical effects of the tested compounds on human kidneys. ProxTox-HTS has been pre-validated with 44 chemically diverse compounds, with non-PTC-toxic nephrotoxicants and non-nephrotoxic compounds serving as negative reference compounds (Table 1).

These reference compounds included commonly used industrial chemicals, antibiotics, antivirals, chemotherapy drugs, mycotoxins, agricultural chemicals and other compounds. The predictive performance of ProxTox-HTS was determined with separate training and test sets, and the test balanced accuracies were 82% and 89% using primary PTCs or HK-2 cells, respectively (Table 2).

Furthermore, the model was able to positively identify both acute nephrotoxicants as well as those resulting in chronic nephrotoxicity in vivo.

Table 1 - LibrarySelected of 44 compounds used for pre-validation of ProxTox-HTS

Compound Direct PTC Toxicity in Humans
5-Fluorouracil Toxic
Aristolochic acid Toxic
Arsenic (III) oxide Toxic
Bismuth (III) oxide Toxic
Cadmium (II) chloride Toxic
Cephaloridine Toxic
Cephalosporin C Toxic
Cephalothin Toxic
Cisplatin Toxic
Citrinin Toxic
Copper (II) chloride Toxic
Cyclosporin A Toxic
Gentamicin Toxic
Germanium (IV) oxide Toxic
Gold (I) chloride Toxic
Lead (IV) acetate Toxic
Ochratoxin A Toxic
Paraquat Toxic
Potassium dichromate Toxic
Puromycin Toxic
Rifampicin Toxic
Tacrolimus Toxic
Tenofovir Toxic
Tetracycline Toxic
Acarbose Non-toxic
Acetaminophen Non-toxic
Ciprofloxacin Non-toxic
Dexamethasone Non-toxic
Ethylene glycol Non-toxic
Furosemide Non-toxic
Glycine Non-toxic
Hydrocortisone Non-toxic
Ibuprofen Non-toxic
Levodopa Non-toxic
Lincomycin Non-toxic
Lindane Non-toxic
Lithium chloride Non-toxic
Melatonin Non-toxic
Metformin hydrochloride Non-toxic
Phenacetin Non-toxic
Ribavirin Non-toxic
Triiodothyronine Non-toxic
Valacyclovir Non-toxic
Vancomycin Non-toxic

Table 2 - Predictive performance of the ProxTox-HTS renal in vitro model with primary PTCs or HK-2 cells.

PTC Model Test Sensitivity Test Specificity Test Balanced Accuracy
Human Primary PTC (Mean of 3 donors) 84% 80% 82%
HK-2 cells 99% 79% 89%

file_download See our ProxTox-HTS application note to learn more (PDF 614Kb)

For further reference:

Su R, Xiong S, Zink D, Loo LH., High‑throughput imaging‑based nephrotoxicity prediction for xenobiotics with diverse chemical structures. Arch Toxicol. 2016 Nov;90(11):2793-2808. Epub 2015 Nov 27

Webinar

Rapid and accurate high-throughput in vitro prediction of nephrotoxicity in humans with ProxTox-HTS

Date: March 01 2018
Presenter: Daniele Zink, PhD, Team Leader and Principal Research Scientist, IBN Singapore

Time:

7:00 am (PDT)

10:00 am (EST)

3:00 pm (GMT)

4:00 pm (CET)

5:00 pm (IST)

12:00 pm (Tokyo, Seoul)

Summary:

The human kidney can be damaged by a wide range of xenobiotics, such as industrial chemicals, environmental toxicants and drugs. Major targets for compound-induced damage are renal proximal tubule cells (PTCs) due to their role in compound transport, metabolism and clearance. ProxTox-HTS is the first high- throughput in vitro screening method developed to predict nephrotoxicity in humans with high accuracy. The HTS platform was developed without using pre-defined endpoints, instead, predictive cellular changes were identified by phenotypic profiling after compound treatment. Analysis of 129 cellular features was conducted by powerful high content imaging and their changes were compared by machine learning to the known clinical effects of the compounds on human kidneys. It has been pre-validated with 44 chemically diverse compounds, with well-characterized effects. The method utilizes primary human renal proximal tubule or HK-2 cells in 384-well format, the test balanced accuracies are 82% (primary PTCs) and 89% (HK-2 cells), respectively. These data were obtained using relatively short compound exposure, and  also included compounds that typically lead to chronic kidney injury gave positive results. ProxTox-HTS can therefore be used for the identification of acutely as well as chronically nephrotoxic compounds.

ProxTox-HTS a new tool for in vitro high-throughput nephrotoxicity prediction.

Speaker's Bio:

Daniele Zink studied Biology at the University of Bielefeld (Germany) and earned her PhD from the University of Heidelberg in 1995 where she worked in the fields of developmental biology and epigenetics. After becoming a postdoctoral fellow at the Institute of Human Genetics of the University of Heidelberg, she was appointed in 1996 as Research Assistant Professor at the Institute of Anthropology and Human Genetics of the Ludwig-Maximilians University (LMU) of Munich, Germany, where she completed her Habilitation in Cell Biology in 2001. In 2001, she was awarded a grant from the Volkswagen-Foundation for the funding of a Junior Group. She was head of the Junior Group located at the LMU (Department of Biology II) until she moved to the Institute of Bioengineering and Nanotechnology (IBN) of the Agency for Science, Technology and Research, Singapore in 2007, as Team Leader and Principal Research Scientist. Her research at IBN focuses on the development of predictive models for in vitro toxicology/nanotoxicology. In addition, she performs research on stem cells, which is directed towards applications in predictive toxicology and the development of cost-effective fully synthetic cultivation systems for industrial applications. The work on predictive in vitro models has been awarded by the US Society of Toxicology and has won the prestigious LUSH Prize (2016, Science Category). The predictive in vitro models are used in collaboration with the US Environmental Protection Agency for the prediction of the human nephrotoxicity of ToxCast compounds. Zink holds 12 patents/patent applications, is Editorial Board Member of Scientific Reports and has > 70 publications, which include publications in Nature, Nature Reviews Cancer, Archives of Toxicology and Nanotoxicology.

Contact us!