A partial pathway- and network-based transformation reveals the synergistic mechanism of JA and UA against cerebral ischemia-reperfusion injury

Dear Editor, Ischemic stroke falls under the category of complex diseases, which do not obey the single-gene dominant or single-gene recessive Mendelian pattern of inheritance but rather arises from the combination of numerous genetic variants and environmental factors (Cho et al., 2012). Given this degree of complexity, a single compound could never be expected to manage or reverse the effects of a stroke. Recently there has been accumulating evidence suggesting that combination therapies could exert a significant effect on areas affected by a cerebral ischemic stroke. Combination therapies that simultaneously impact multiple targets have promise in the control of complex diseases by way of synergy (Li et al., 2011). Jasminoidin (JA) and ursodeoxycholic acid (UA) are compounds extracted from Qingkailing injection, which was extensively used for treating acute ischemia stroke for decades in China (Cheng et al., 2012). There was synergistic effect observed with the combination of JA and UA with regard to the reduction of infarction volumes, a decrease in the neurological deficit score and the MRI results in ischemic rats (Liu et al., 2012). However, till now, little was known regarding the synergistic mechanism of JA and UA in the treatment of ischemia stroke until the implementation of an integrated network analysis of global gene expression profiling. In the present study, we aim to investigate the synergistic mechanism of JA and UA against cerebral ischemia on a systemic level. The IPA system was applied to identify the differential networks and canonical pathways as well as to determine the overlap of these diverse biological functions using an integrated network analysis of the global gene expression profiling in the JA, UA, JU (JA + UA combination) treatment groups. Firstly, the IPA results revealed 11 statistically significant networks between the JU and vehicle groups. The top 5 networks included 35, 35, 35, 34 and 35 nodes (genes; proteins) (Table.S1). Network 1 includes gene expression, cell cycle and cellular compromise (Fig. 1A); network 2 functions in DNA replication, recombination and repair, drug metabolism, and endocrine system development and function (Fig. 1B); network 3 involves endocrine system development and function, small molecule biochemistry, and organ morphology (Fig. 1C); network 4 includes cell death, neurological disease, and connective tissue disorders (Fig. 1D); and network 5 functions in lipid metabolism, molecular transport, and small molecule biochemistry (Fig. S1). Secondly, the IPA analysis also identified 33 statistically significant canonical pathways out of 149 pathways involving the differentially expressed genes between the JU and vehicle groups (Table.S2). Of these 33 canonical pathways, the combination JU treatment regulated as many as 13 pathways independently, including renin-angiotensin signaling and cAMP-mediated signaling, whereas the JU, JA and UA groups showed 6 overlapping pathways. Meanwhile 6 overlapping pathways were identified between the JA and JU groups, and 8 overlapping pathways were discovered between the UA and JU groups (Fig. 2A). Furthermore, 10 out of 13 non-overlapping pathways regulated by JU separately were related to cerebral ischemia. These pathways were largely related to inflammation, apoptosis/ necrosis, and neurogenesis/angiogenesis with regard to cerebral ischemia. However, 6 non-overlapping pathways in the JU group were stimulated by microglial activation, which plays a key role in the inflammatory reaction during cerebral ischemia (Fig. 2B). More importantly, 3 non-overlapping pathways were identified to be involved in cerebral ischemia for the first time, including molecular mechanisms of cancer, PXR/RXR activation, and caveolar-mediated endocytosis. Thirdly, compared with the vehicle group, a total of 66 overlapping biological functions were identified among the JA, UA and JU groups, which accounted for 79.52 % of all the functions. Interestingly, 5 overlapping functions were discovered between the JA and UA groups, one overlapping


MCAO model and drug treatment
Mice were subjected to the middle cerebral artery occlusion (MCAO) model, in which they were ligated with an intraluminal filament for 1.5 h and then reperfused for 24 h. The mice in the sham-operated group were surgically prepared for the insertion of the filament, but the filament was not inserted.
The animals were randomly divided into the following 6 groups: JA-treated group (50 mg/kg), UA-treated group (14 mg/kg), JU-treated group (combination of JA and UA at 1:1), NI-treated group (80 mg/kg), vehicle group (0.9% NaCl), and sham-operated group (0.9% NaCl). JA and UA were chemically standardized products from either the China Natural Institute for the Control of Pharmaceutical and Biological Products or Beijing University of Traditional Chinese Medicine, and their tested purities exceeded 98% by fingerprint chromatography. All of the drugs were dissolved in 0.9% NaCl just prior to use and injected into the tail vein of mice 1.5 h after the induction of focal cerebral ischemia. Each mouse in the vehicle and the sham-operated groups were injected with saline alone (2 ml/kg body weight) into the tail vein at the same time point.

TTC Staining and Histological analysis
The infarct volume from 9 mice from each group was measured by TTC (2, 3, 5triphenyl tetrazolium chloride) staining. The brain was sliced into coronal sections and immersed into a 1% TTC solution for 30min at 37°C. After washing with normal saline, the slices were transferred into a 4% formaldehyde solution for fixation.
Images of brain sections were captured using a digital camera (Color CCD camera TP-6001A,Topica Inc., Japan). The area of ischemic damage was measured using pathology image analysis system software (Topica Inc., Japan), and the ratio of infarct volume to the total slice was calculated.
An additional 6 mice from each group were subjected to hippocampus 3 histological analysis. After washing with normal saline, mouse brains were perfused with cold 4% formaldehyde for 30 min to induce polymerization. The brains were fixed in 4% formaldehyde for at least 24 h, sectioned coronally into 5-μm slices, and then stained with 0.2% thionine. The hippocampal CA1 region was observed under a microscope.

RNA isolation
The hippocampi of 9 mice from each group were homogenized in Trizol Reagent (Invitrogen, USA), and the total RNA was isolated according to the manufacturer's instructions. The purification and concentration of RNA were performed using an RNeasy Micro Kit (Qiagen, Valencia, CA) according to the manufacturer's protocol.
The RNA quality was evaluated by determining the 26S/18S ratio with a bioanalyzer microchip (Agilent, Palo Alto, CA).

Microarray
Gene expression profiling was conducted using a microarray chip containing the whole genome array (Boao Capital, Beijing, China) for mice (16,463 oligoclones, Incyte Genomics, Santa Clara, CA, USA). On one given chip, each clone generated 4 technical replicates via duplicate spots printing. After smoothing spline normalization, a single intensity value for each clone was determined by averaging the quadruplet measurements. All of the clones had been verified by DNA sequencing. The total RNA from the vehicle group was labeled with Cy3 as the pool control, whereas RNA from the other groups was labeled with Cy5. The microarrays were then hybridized, washed, and scanned according to standard protocols. The experiment was repeated in triplicate, with technical quadruplets for each group within each array.

IPA analysis of Microarray Data
All microarray data were uploaded to the ArrayTrack system (Food and Drug Administration, USA).Experimental analysis was performed using the Minimum Information About a Microarray Experiment (MIAME) guidelines and the MicroArray Quality Control (MAQC) Project standards. Analysis results were submitted to the Array Express database, and the microarray data were normalized by carrying out a locally weighted linear regression (Lowess) to reduce systematic bias as best as possible (smoothing factor: 0.2; robustness iterations: 3). One-way analysis of variance and significance analysis of the microarrays were applied to compare the means of the altered genes between the sham and vehicle, JA and vehicle, UA and vehicle, and JU and vehicle groups. Genes that showed significant changes in expression with a P value less than 0.05 and a fold change greater than 1.5 were selected for further analysis. Additionally, genes that had >1.5-fold increase in their expression levels were indicative of upregulation, and genes that had <0.5-fold decrease of expression levels suggested downregulation.
After the initial analysis, genes showing significant changes in expression were uploaded to the IPA system (http://www.ingenuity.com/).A cutoff was set to identify network eligible molecules whose differential expression was significantly regulated.
Based on the connectivity of these network-eligible molecules, networks were algorithmically generated. The right-tailed Fisher's exact test was applied to calculate a P value determining if each biological function assigned to that network is due to chance alone. The significance of correlation between these genes and the canonical pathway was determined in two ways: (1) a ratio of the number of genes from the data set that map to the pathway divided by the total number of genes that map to the canonical pathway; and (2) Fisher's exact test to calculate a P value determining if the correlation between the genes and the canonical pathway can be explained by chance alone. A P value <0.05 was considered to be statistically significant. Finally, we screened out and analyzed all canonical pathways that had with a P value < 0.05 and a fold change >1.5.

Discussions
In our current study, based on the analysis of the top 5 differential networks, the targets genes were primarily related to ERK, NF-κB, p38 MAPK, Akt, Fos, Jnk, Ras, JUNB, EGR1, PI3K, and STAT3 in both the single and combination treatment groups.
It has been reported that these genes are involved in the process of cerebral ischemia.
In recent years, the altered expression of proteins regulating the cell cycle were confirmed (resulting in cell death), and infarction was reduced significantly by inhibiting the cell cycle after cerebral ischemia (Zhang et al., 2009).Furthermore, compounds such as tricyclodecan-9-yl-xanthogenate (D609) and berberine can prevent mature neurons from entering the cell cycle at the early reperfusion, which is possibly mediated by interference due to the delayed proliferation of microglia and macrophages (Adibhatla and Hatcher; Chai et al.). Previous studies have proven that cerebral injuries can induce distinct abnormalities of the endocrine system (Schwarz et al., 2003). Dysregulation of the hypothalamic-pituitary axis can induce hypercortisolism, abolish the hormonal circadian pattern of β-endorphins and cortisol, and elevate nocturnal prolactin release (Schwab et al., 1997;Wallaschofski et al., 2006).Additionally, other abnormalities have been shown to manifest after stroke such as lowered levels of gonadotropins, growth hormone, and thyroid-stimulating hormone as well as elevated nocturnal prolactin release and impaired thyrotropin-releasing hormone-stimulated secretion of thyroid-stimulating hormone (Akhoundi et al.; Vespa). There have been few studies reporting on how the functions of drug metabolism, organ morphology and molecular transport are involved in acute cerebral ischemia. Therefore, this study may reveal novel mechanisms of the synergistic effect of JA and UA in treating cerebral ischemia.
With regard to the networks identified in the NI group(positive control), Ca 2+ was identified as one of node molecules, which was consistent with its pharmacological effect of Ca 2+ -antagonist against cerebral ischemia (Greenberg et al., 1990;Zhu et al., 1999). Furthermore, other node molecules, such as ERK, NF-κB, p38 MAPK, Jnk, Fos, Junb, Caspase, Ap1were also found in the NI group. It was indicated that nimodipine may act on multiple targets in treating with cerebral ischemia.