Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Sep;36(9):1137-49.
doi: 10.1002/jat.3278. Epub 2016 Jan 4.

Systems Toxicology of Chemically Induced Liver and Kidney Injuries: Histopathology-Associated Gene Co-Expression Modules

Affiliations
Free PMC article

Systems Toxicology of Chemically Induced Liver and Kidney Injuries: Histopathology-Associated Gene Co-Expression Modules

Jerez A Te et al. J Appl Toxicol. .
Free PMC article

Abstract

Organ injuries caused by environmental chemical exposures or use of pharmaceutical drugs pose a serious health risk that may be difficult to assess because of a lack of non-invasive diagnostic tests. Mapping chemical injuries to organ-specific histopathology outcomes via biomarkers will provide a foundation for designing precise and robust diagnostic tests. We identified co-expressed genes (modules) specific to injury endpoints using the Open Toxicogenomics Project-Genomics Assisted Toxicity Evaluation System (TG-GATEs) - a toxicogenomics database containing organ-specific gene expression data matched to dose- and time-dependent chemical exposures and adverse histopathology assessments in Sprague-Dawley rats. We proposed a protocol for selecting gene modules associated with chemical-induced injuries that classify 11 liver and eight kidney histopathology endpoints based on dose-dependent activation of the identified modules. We showed that the activation of the modules for a particular chemical exposure condition, i.e., chemical-time-dose combination, correlated with the severity of histopathological damage in a dose-dependent manner. Furthermore, the modules could distinguish different types of injuries caused by chemical exposures as well as determine whether the injury module activation was specific to the tissue of origin (liver and kidney). The generated modules provide a link between toxic chemical exposures, different molecular initiating events among underlying molecular pathways and resultant organ damage. Published 2016. This article is a U.S. Government work and is in the public domain in the USA. Journal of Applied Toxicology published by John Wiley & Sons, Ltd.

Keywords: adverse outcome pathways; co-expression modules; hepatotoxicity; histopathology; nephrotoxicity; systems toxicology; toxicogenomics.

Figures

Figure 1
Figure 1
The module for the specific histopathology endpoints has the highest Matthews correlation coefficient (MCC). Shown are the MCCs for the liver histopathology assessments, ordered according to Table 1 from cytoplasmic alteration as P1 to single cell necrosis as P11, and 11 modules for the liver, LM1 to LM11. LM1 corresponds to the module for liver P1 and has the highest MCC for that histopathology. Also shown are the MCCs for the eight kidney injuries, ordered as listed in Table 2. As with the liver modules, KM1 is the module for kidney P1 and has the highest MCC for P1.
Figure 2
Figure 2
Positive instances of the liver fibrosis histopathology activate the liver fibrosis module. Conditions (chemical‐time‐dose combination) with activation scores above the activation threshold (corresponding to P‐value <0.025) are considered activated (in red). The activation score is defined as the average of the absolute value of the expression Z‐scores of the genes in the liver fibrosis module (LM7). Positive instances of liver fibrosis are marked with stars. Liver fibrosis is predicted if the histopathology endpoint is observed and LM7 is activated for the condition (red rectangles with stars). ND indicates no data are available.
Figure 3
Figure 3
Positive instances of the kidney fibrosis histopathology activate the kidney fibrosis module. Conditions with KM5 activation scores above the activation threshold were considered activated (in red). Positive instances of kidney fibrosis are marked with stars. Liver fibrosis is predicted if the histopathology endpoint is observed and KM5 is activated for the condition (red rectangles with stars). ND indicates no data are available.
Figure 4
Figure 4
Module activation can predict histopathology‐causing conditions. Principal component analysis of the activation of the modules by histopathology‐graded chemical exposure conditions (red dots) for modules (A and C) and for genes (B and D). A condition was considered injury causing if at least one of the 11 liver (A and B) or 8 kidney (C and D) histopathology endpoints was positive. All other conditions not associated with histopathology graded damage are marked with black dots.
Figure 5
Figure 5
Injury severity is associated with correspondingly higher activation scores. In general, as the severity of the histopathology increases, the activation score of the module increases. The histopathology assessments were converted to numbers (with minimal = 1, slight = 2, moderate = 3, severe = 4) and the histopathology score was determined as the average over all replicates. Only conditions with at least one replicate having an observed histopathology were considered for the linear regression. The goodness of the fit was measured as the coefficient of determination (R 2).
Figure 6
Figure 6
The module selection assigns unique modules to closely related histopathology endpoints. In TG‐GATEs, chemical exposure conditions may cause a number of injuries. Four closely related histopathology assessments (cellular infiltration, fibrosis, bile duct proliferation and single cell necrosis) share a number of common chemicals. (A) All four injuries activate 61 modules in the liver, with the 61 modules having overlapping genes, as determined by the Sorensen‐Dice coefficient (Dice, 1945). (B) Using our module selection protocol, the number of overlapping genes among the modules is limited.
Figure 7
Figure 7
The activation of the modules is organ‐specific. Allopurinol and puromycin, two nephrotoxicants that cause a number of kidney injuries, including kidney fibrosis, generally activate kidney modules. Conversely, thiocetamide, a hepatotoxicant that causes liver fibrosis among other liver injuries, selectively activates liver injury modules. Also shown is carbon tetrachloride, a known hepatotoxicant. The four other chemicals causing liver fibrosis are shown in Supplementary Material Fig. S3.

Similar articles

See all similar articles

Cited by 6 articles

See all "Cited by" articles

References

    1. AbdulHameed MDM, Tawa GJ, Kumar K, Ippolito DL, Lewis JA, Stallings JD, Wallqvist A. 2014. Systems level analysis and identification of pathways and networks associated with liver fibrosis. PLoS One 9 e112193. DOI:10.1371/journal.pone.0112193. - DOI - PMC - PubMed
    1. Adams LA. 2011. Biomarkers of liver fibrosis. J. Gastroenterol. Hepatol. 26: 802–809. DOI:10.1111/j.1440-1746.2010.06612.x. - DOI - PubMed
    1. Bai JP, Abernethy DR. 2013. Systems pharmacology to predict drug toxicity: integration across levels of biological organization. Annu. Rev. Pharmacol. Toxicol. 53: 451–473. DOI:10.1146/annurev-pharmtox-011112-140248. - DOI - PubMed
    1. Bergmann S, Ihmels J, Barkai N. 2003. Iterative signature algorithm for the analysis of large‐scale gene expression data. Phys. Rev. E 67: 031902 DOI:10.1103/PhysRevE.67.031902. - DOI - PubMed
    1. Blomme EA, Yang Y, Waring JF. 2009. Use of toxicogenomics to understand mechanisms of drug‐induced hepatotoxicity during drug discovery and development. Toxicol. Lett. 186: 22–31. DOI:10.1016/j.toxlet.2008.09.017. - DOI - PubMed

Publication types

LinkOut - more resources

Feedback