Huarong Zhou

Network screening of Goto-Kakizaki rat liver microarray data during diabetic progression

BackgroundType 2 diabetes mellitus (T2DM) is a complex systemic disease, with significant disorders of metabolism. The liver, a central energy metabolic organ, plays a critical role in the development of diabetes. Although gene expression levels are able to be measured via microarray since 1996, it is difficult to evaluate the contributions of one altered gene expression to a specific disease. One of the reasons is that a whole network picture responsible for a specific phase of diabetes is missing, while a single gene has to be put into a network picture to evaluate its importance. In the aim of identifying significant transcriptional regulatory networks in the liver contributing to diabetes, we have performed comprehensive active regulatory network survey by network screening in 4 weeks (w), 8-12 w, and 18-20 w Goto-Kakizaki (GK) rat liver microarray data.ResultsWe identify active regulatory networks in GK rat by network screening in the following procedure. First, the regulatory networks are compiled by using the known binary relationships between the transcriptional factors and their regulated genes and the biological classification scheme, and second, the consistency of each regulatory network with the microarray data measured in GK rat is estimated to detect the active networks under the corresponding conditions. The comprehensive survey of the consistency between the networks and the measured data by the network screening approach in the case of non-insulin dependent diabetes in the GK rat reveals: 1. More pathways are active during inter-middle stage diabetes; 2. Inflammation, hypoxia, increased apoptosis, decreased proliferation, and altered metabolism are characteristics and display as early as 4weeks in GK strain; 3. Diabetes progression accompanies insults and compensations; 4. Nuclear receptors work in concert to maintain normal glycemic robustness system.ConclusionNotably this is the first comprehensive network screening study of non-insulin dependent diabetes in the GK rat based on high throughput data of the liver. Several important pathways have been revealed playing critical roles in the diabetes progression. Our findings also implicate that network screening is able to help us understand complex disease such as diabetes, and demonstrate the power of network systems biology approach to elucidate the essential mechanisms which would escape conventional single gene-based analysis.

APG: an Active Protein-Gene Network Model to Quantify Regulatory Signals in Complex Biological Systems

Synergistic interactions among transcription factors (TFs) and their cofactors collectively determine gene expression in complex biological systems. In this work, we develop a novel graphical model, called Active Protein-Gene (APG) network model, to quantify regulatory signals of transcription in complex biomolecular networks through integrating both TF upstream-regulation and downstream-regulation high-throughput data. Firstly, we theoretically and computationally demonstrate the effectiveness of APG by comparing with the traditional strategy based only on TF downstream-regulation information. We then apply this model to study spontaneous type 2 diabetic Goto-Kakizaki (GK) and Wistar control rats. Our biological experiments validate the theoretical results. In particular, SP1 is found to be a hidden TF with changed regulatory activity, and the loss of SP1 activity contributes to the increased glucose production during diabetes development. APG model provides theoretical basis to quantitatively elucidate transcriptional regulation by modelling TF combinatorial interactions and exploiting multilevel high-throughput information.

A computational procedure for identifying master regulator candidates: a case study on diabetes progression in Goto-Kakizaki rats

BackgroundWe have recently identified a number of active regulatory networks involved in diabetes progression in Goto-Kakizaki (GK) rats by network screening. The networks were quite consistent with the previous knowledge of the regulatory relationships between transcription factors (TFs) and their regulated genes. To study the underlying molecular mechanisms directly related to phenotype changes, such as diseases, we also previously developed a computational procedure for identifying transcriptional master regulators (MRs) in conjunction with network screening and network inference, by effectively perturbing the phenotype states.ResultsIn this work, we further improved our previous method for identifying MR candidates, by listing them in a more reliable manner, and applied the method to reveal the MR candidates for diabetes progression in GK rats from the active networks. Specifically, the active TF-gene pairs for different time periods in GK rats were first extracted from the networks by network screening. Another set of active TF-gene pairs was selected by network inference, by considering the gene expression signatures for those periods between GK and Wistar-Kyoto (WKY) rats. The TF-gene pairs extracted by the two methods were then further selected, from the viewpoints of the emergence specificity of TF in GK rats and the regulated-gene coverage of TF in the expression signature. Finally, we narrowed all of the genes down to only 5 TFs (Etv4, Fus, Nr2f1, Sp2, and Tcfap2b) as the candidates of MRs, with 54 regulated genes, by merging the selected TF-gene pairs.ConclusionsThe present method has successfully identified biologically plausible MR candidates, including the TFs related to diabetes in previous reports. Although the experimental verifications of the candidates and the present procedure are beyond the scope of this study, we narrowed down the candidates to 5 TFs, which can be used to perform the verification experiments relatively easily. The numerical results showed that our computational method is an efficient way to detect the key molecules responsible for biological phenomena.

Pancreatic hyperplasia after gastric bypass surgery in a GK rat model of non-obese type 2 diabetes

Gastric bypass surgery produces clear antidiabetic effects in a substantial proportion of morbidly obese patients. In view of the recent trend away from bariatric surgery and toward metabolic surgery, it is important to elucidate the enhancing effect of bypass surgery on pancreatic β-cell mass, which is related to diabetes remission in non-obese patients. We investigated the effects of gastric bypass surgery on glycemic control and other pancreatic changes in a spontaneous non-obese type 2 diabetes Goto-Kakizaki rat model. Significant improvements in postprandial hyperglycemia and plasma c-peptide level were observed when glucose was administered orally post-surgery. Other important events observed after surgery were enhanced first phase insulin secretion in a in site pancreatic perfusion experiment, pancreatic hyperplasia, improved islet structure (revealed by immunohistochemical analysis), striking increase in β-cell mass, slight increase in ratio of β-cell area to total pancreas area, and increased number of small islets closely related to exocrine ducts. No notable changes were observed in ratio of β-cell to non-β endocrine cell area, β-cell apoptosis, or β-cell proliferation. These findings demonstrate that gastric bypass surgery in this rat model increases endocrine cells and pancreatic hyperplasia, and reflect the important role of the gastrointestinal system in regulation of metabolism.

Reduction of pancreatic β-cell dedifferentiation after gastric bypass surgery in diabetic rats

Dear Editor, Type 2 diabetes mellitus (T2DM) develops only in insulin-resistant subjects when pancreatic b-cell compensation fails (Matveyenko and Butler, 2006). Decreased insulin secretory function and reduced cell mass are traditionally viewed as major contributing factors in b-cell insufficiency. A recent study using a diabetic rodent model suggests that progressive b-cell dedifferentiation is an important underlying mechanism in b-cell failure (Talchai et al., 2012). b-cell dedifferentiation in diabetes refers to the loss by healthy b-cells of key components characteristic of the differentiated state (Dor and Glaser, 2013), including insulin (for its secretory product), Glut2 (for glucose intake), and PDX-1 (for critical insulin transcription factor). b-cell dedifferentiation may be largely responsible for not only b-cell secretory dysfunction but also impaired b-cell identity. In view of findings that bariatric surgery in a rodent T2DM model led to increased b-cell mass and improved islet morphology (Strader et al., 2009), we investigated the effects of gastric bypass surgery on dedifferentiated b-cells. Roux-en-Y gastric bypass (RYGB), a type of bariatric surgery, is an effective surgical treatment for patients with morbid obesity. RYGB surgery also improved secretion of b-cells in response to intravenous glucose (Salinari et al., 2013) and completely resolved T2DM in a significant number of patients (Schauer et al., 2003). In animal studies, novel surgical approaches relieved diabetes in a rapid and sustained manner, independent of weight loss effects (Strader et al., 2009). In view of the trend away from bariatric surgery and toward metabolic surgery, we performed RYGB surgery on spontaneous T2DM Goto-Kakizaki (GK) rats (GK-S), a non-obese model with inherited b-cell deficits, to study weight lossindependent effects of RYGB on islets. Dedifferentiated b-cells were examined in this experimental group (GK-S) and two control groups: (i) sham-operated rats pair-fed with the GK-S group (GK-PF-Sham), and (ii) normal Wistar rats (Wistar). Pancreatic b-cell function after 3 months of RYGB surgery was evaluated by intravenous glucose tolerance tests (IVGTT). Both blood glucose levels at individual time points and total glucose level determined by area under the curve (AUC) reflected improvement of glycemic controls following RYGB surgery (Figure 1A-i,ii). Basal insulin levels were similar in the two GK groups. Plasma insulin levels after 2 min and 5 min of intravenous glucose load were significantly higher in GK-S group than in GK-PF-Sham group. AUC-insulin level within 5 min was alsosignificantlyhigher inGK-Sgroup, indicating improvement of first-phase insulin secretion following RYGB surgery (Figure 1A-iii,iv). These findings indicate that pancreatic b-cell secretory function is improved after RYGB surgery, in agreement with previous reports (Salinari et al., 2013). We then performed immunohistochemical analyses to assess islet structure in the pancreas. GK-PF-Sham pancreases contained predominantly irregularly shaped ‘broken’ islets, in which weaker insulin immunoreactive staining revealed uneven hypoglycemic factor expression in b-cells, whereas GK-S pancreases contained many normal islets indistinguishable from those in Wistar group (Figure 1B-i). Insulin content data showed a partial but significant recovery of insulin storage after surgery in total pancreaticb-cells within equivalent amounts of pancreas (Figure 1B-ii). Glucagon content data were comparable between the two GK groups (Figure 1B-iii). Thus, the insulin/ glucagon ratio was significantly higher after bypass surgery (GK-PF-Sham: 23.7 + 0.6 vs. GK-S: 32.4 + 1.1; P , 0.05, n 1⁄4 6). These morphological and pancreatic hormone content data also reflect functional improvement of GK-S b-cells. In a study by Talchai et al. (2012), the endocrine cell marker secretagogin (SCGN) was still present after the insulin staining disappeared in dedifferentiated b-cells. To distinguish immunoreactive areas between SCGN-positive and pancreatic hormone-positive cells, we measured expressions of four pancreatic hormones (glucagon, insulin, somatostatin, pancreatic polypeptide) and SCGN in islets by confocal microscopy (Supplementary Figure S1). Cells with strongly positive hormone staining were regarded as healthy endocrine cells, cells with reduced hormone staining were regarded as degranulated endocrine cells, and those SCGN-positive but hormone-negative were regarded as dedifferentiated endocrine cells (Figure 1C-i; Supplementary Figure S1). Endocrine cells consist of 15% 2 20% a-cells and 65% 2 80% b-cells. Since a-cells determined by glucagon expression display hyperfunction and increased mass in T2DM (Elayat, 1995; Figure 1B-iii and C-iii), we presume that the hormone ‘empty’ and ‘pale-staining’ endocrine cells were primarily dedifferentiated and degranulated b-cells. Islets from GK-S group, in comparison with GK-PF-Sham group, showed a significant reduction in doi:10.1093/jmcb/mju042 Journal of Molecular Cell Biology (2014), 6(6), 531–534 | 531 Published online October 31, 2014

Association of Rev-erbα in adipose tissues with Type 2 diabetes mellitus amelioration after gastric bypass surgery in Goto-Kakizaki rats

We estimated the key molecules related to Type 2 diabetes mellitus (T2DM) in adipose, liver, and muscle tissues, from nonobese diabetic Goto-Kakizaki (GK) rats and their Wistar controls, by computationally analyzing the expression profiles in open source data. With the aid of information from previous reports, Rev-erbα in adipose tissue emerged as one of the most plausible candidates. Here, in animal models, including GK rats surgically treated to ameliorate T2DM, we examined the association of Rev-erbα in adipose tissue with T2DM progression. After analyses of the Rev-erbα mRNA expression in the adipose tissue of our animal models, we compared the Rev-erbα protein expression levels in the adipose, liver, and muscle tissues of GK and Wistar controls at the ages of 1 mo (M), 3M, and 6M. The Rev-erbα protein levels in adipose tissue showed a distinctive pattern, with the negative correlation of an increasing trend in GK rats, and a decreasing trend in Wistar rats during aging, from those in liver and muscle tissues. Moreover, dysregulation of the circadian Rev-erbα expression in the adipose tissue of 6-mo-old GK rats was also observed. In particular, we ameliorated T2DM in GK rats by gastric bypass surgery, and revealed that T2DM amelioration in diabetic GK rats was associated with improved circadian Rev-erbα expression, in a comparison between the surgically treated and untreated GK rats. The roles of Rev-erbα in adipose tissue were further investigated by observations of Rev-erbα-related molecules, with reference to previous reports.

Phenotype-difference oriented identification of molecular functions for diabetes progression in Goto-Kakizaki rat

In general, molecular signatures of diseases are estimated by comparing the two sets of molecular data measured for the samples with distinctive phenotypes, and then molecular functions of the diseases are characterized by the following analyses of the signatures. Unfortunately, ambiguous relationships between molecular signatures and functions are observed in some cases, due to a posteriori justification from molecular level to phenotype level. Here, we propose a method for detecting molecular functions of the disease by a deductive justification from phenotype level to molecular level, and illustrate its performance by applying our method to the gene expression and phenotype data sets for diabetes progression in Goto-Kakizaki rat. By our method, the functions identified by the previous studies were well covered, and furthermore, some implications for molecular mechanisms were obtained. Our phenotype-difference oriented method provides some clues to bridge directly a gap between molecular signatures and phenotype data in diabetes.

Network clustering along diabetes progression in three tissues of Goto-Kakizaki rats

We investigated the macroscopic changes in the regulatory coordination of diabetes progression during three periods in three tissues, adipose, liver and muscle, of Goto-Kakizaki (GK) rats. For this purpose, we performed network clustering by the Newman algorithm for the regulatory networks inferred by a modified path consistency algorithm, and investigated the biological functions of each cluster by an enrichment analysis of the constituent genes. We then compared the network clusters characterized by biological functions with the diabetes progression of GK rats in each of the three tissues. The network structure, the number of clusters, and the number of clusters characterized by biological functions during the three periods showed similar patterns in the three tissues. In contrast, further scrutiny of the biological functions at coordinated clusters revealed characteristic differences between the three tissues along the diabetes progression. In particular, the hypothetical roles of each tissue emerged: adipose and liver function at the cellular and molecular levels at the early stage, respectively, and all three tissues are responsible for diabetes progression, under the control of various transcriptional regulators.

Exploration of Cellular Relationships from Characteristically Expressed Genes by Partial Canonical Correlation Analysis

In this report, we designed a procedure for exploring the cellular relationships from the sets of characteristic genes in distinctive cell types, by using partial canonical correlation analysis. We applied the present procedure to the characteristic gene sets of seven subtypes of testicular germ cell tumors in a previous report. The cellular relationships were well reconstructed, consistent with the general histologic lineage model, and new implications for the subtype differentiation were found. In particular, the correspondence between the inferred relationships and the functional characterizations of constituent genes suggests a hypothesis for the classification between seminoma and embryonal carcinoma. The partial canonical correlation analysis is also appropriate for revealing new features of cellular relationships, based on the transcriptional programs. Thus, the present procedure helps to create a macroscopic view of the cellular relationships, by following the detection of the characteristically expressed genes.

Identification of master regulator candidates for diabetes progression in Goto-Kakizaki Rat by a computational procedure

Recently, we have identified 39 candidates of active regulatory networks for the diabetes progression in Goto-Kakizaki (GK) rat by using the network screening, which were well consistent with the previous knowledge of regulatory relationship between transcription factors (TFs) and their regulated genes. In addition, we have developed a computational procedure for identifying transcriptional master regulators (MRs) related to special biological phenomena, such as diseases, in conjunction of the network screening and inference. Here, we apply our procedure to identify the MR candidates for diabetes progression in GK rat. First, active TF-gene relationships for three periods in GK rat were detected by the network screening and the network inference, in consideration of TFs with specificity and coverage, and finally only 5 TFs were identified as the candidates of MRs. The limited number of the candidates of MRs promises to perform experiments to verify them.