Identifying pre-disease signals before metabolic syndrome in mice by dynamical network biomarkers

Keiichi Koizumi,Makito Oku,Yoshiko Igarashi,Makoto Kadowaki,Akiko Inujima,Shusaku Hayashi,Kazuyuki Tobe,Shigeru Saito,Kazuyuki Aihara,Luonan Chen,Naotoshi Shibahara

doi:10.1038/s41598-019-45119-w

Abstract

The establishment of new therapeutic strategies for metabolic syndrome is urgently needed because metabolic syndrome, which is characterized by several disorders, such as hypertension, increases the risk of lifestyle-related diseases. One approach is to focus on the pre-disease state, a state with high susceptibility before the disease onset, which is considered as the best period for preventive treatment. In order to detect the pre-disease state, we recently proposed mathematical theory called the dynamical network biomarker (DNB) theory based on the critical transition paradigm. Here, we investigated time-course gene expression profiles of a mouse model of metabolic syndrome using 64 whole-genome microarrays based on the DNB theory, and showed the detection of a pre-disease state before metabolic syndrome defined by characteristic behavior of 147 DNB genes. The results of our study demonstrating the existence of a notable pre-disease state before metabolic syndrome may help to design novel and effective therapeutic strategies for preventing metabolic syndrome, enabling just-in-time preemptive interventions.

Highlights

Metabolic syndrome is a state wherein a blood glucose level, blood pressure, and/or a triglyceride level are elevated higher than those in normal ranges, mainly due to abdominal obesity[1]
Our results were consistent with those of previous studies reporting that the metabolic syndrome onset in TSOD mice was seen at approximately 8–12 weeks of age[20,24,27]
The two strains of TSOD and TSNO mice were almost completely separated in the principal component analysis (PCA) plot using PC1 and PC2 (Fig. 1C)

Summary

Introduction

Metabolic syndrome is a state wherein a blood glucose level, blood pressure, and/or a triglyceride level are elevated higher than those in normal ranges, mainly due to abdominal obesity[1]. Increases in the strength of intervariable correlations[11] have been reported This is not necessarily because the coupling between variables becomes stronger by itself near a critical transition. The DNB theory provides statistical methods to select relevant variables for detection of the pre-disease state. The basic assumption is that a small number of closely related variables, called DNBs, convey early warning signals for the impending critical transition. The average standard deviation of DNB variables and the average absolute value of correlation coefficients between DNB variables are easy to calculate and are widely applicable measures Simultaneous changes in these measures are regarded as a reliable early warning signal, suggesting that a critical transition to a disease state will occur. Researches on the improvement of the statistical methods[14,16] and refinement of the theory[19] are in progress

Methods

Results

Conclusion