上海翻譯公司完成醫(yī)學(xué)“芯片研究”英文翻譯

Abstract
Background:
Paraquat Poisoning (herein referred to as PQ) can cause illness and death to people and wildlife, and its main cause of death is respiratory failure and lung fibrosis resulting in acute lung damage. Early detection and treatment is vital for prognosis advancement. At the present time, the mechanism that controls the gene's expression changes in the early stage of PQ still remains unknown.
Methods:
Paraquat intragastric administration was applied on C57/BL6 rodents to induce PQ to confirm the model via pathological examination. After a gene chip-based screening, the change of gene expression in the lung was further dissected by gene expression profiles, GO-Analysis, co-expression network construction and real-time RT-PCR, western blot and immunofluorescence.
Results: 2287 genes with different expression variations were identified at the very early stage of PQ. Of these genes, 76 which are connected to mitochondrion were isolated for further examination. Among these 76 genes, the PSTK gene is a phosphorylase kinase which plays an essential role in synthesizing selenium and serves a protective role in oxidative stress lung damage.   Drawn from our discovery that PSTK was only one of the most important of the ten genes in the co-expression network of 76 genes according to the degree, we constructed another co-expression network of centralization of PSTK with a co-efficiency of above 0.99.. In the PSTK co-expression network, there are an additional 30 genes which function as complex I assembly, mitochondrial apoptosis, mitochondrial fatty acid beta-oxidation, formation of acetoacetyl-CoA, and they could   be directly tied to the pathogenesis of PQ lung damage. In the end, we confirmed the levels of the PSTK gene and encoded protein in PQ induced rodents, and discovered that PSTK is lowered in rodent lungs.
Conclusions: Utilizing microarray analysis, we isolated PSTK as a potential core gene in the pathogenesis of PQ. The gene forms a network that has shown a strong connection with PQ development. In the early stages of PQ, the expression of PSTK gene involved in selenium protein synthesis changed significantly in lung tissue. The results we collected imply that PSTK is a novel gene connected to early lung injuries in rodent PQ models.
 
Introduction
Paraquat (PQ, 1,1′-dimethyl-4,4′-bipyridinium dichloride) is a herbicide commonly seen worldwide. Exposure to more than 20 mg of PQ ion per kg body weight in humans usually leads to illness and ¹ death due to pulmonary inflammation, fibrosis and respiratory failure.² When certain amount is absorbed, oxidative damage is caused by PQ aggregating in the lungs and producing large amounts of oxygen-free radicals. ³. Early hemoperfusion (herein referred to as HP) is very important for healing patients with acute PQ intoxication⁴. Despite this the death rates remain high despite the large differences in treatments containing HP, antioxidants and immunosuppression therapy. At the present time, therapeutic methods have little founding for increasing excretion, halting absorption, nor does the use of antioxidant treatments and immunosuppressive therapies. Additionally, it was shown that PQ exposure caused an increase in activity of manganese-superoxide dismutase (Mn-SOD) and mitochondrial superoxide creation. The activity levels of Mg+2 adenosine triphosphatase (ATPase) and complex I-III, complex II-III were lowered. ⁵ However, the genetic mechanism of the pathogenesis and development of PQ has not yet been identified.
 
Effective treatment and early detection is vital for prognosis advancement. At the present time we don't know much about gene expression in the early stage of paraquat poisoning (herein referred to as PQ).  Clues for early diagnosis may come from the recognition of genes tied to the early stage of PQ development. Previously, the Illumina gene chip had been used to test the genes in neuron in brain and lung^9-11 (this method was based upon the PCR and cDNA technology and the number of genes which were examined was limited to around 1000,) but these days, high-throughput gene chips are used to explore the genetic dynamics on a genome scale ^6-8. In our study we used the Affymetrix Rat Gene1.0 gene chip; this method allows for more than 27,000 genes to be examined, and provided higher sensitivity and intensity as compared with the previous tools. ¹²
 
    In specific conditions such as the very early stage of PQ, our method could also be used to identify gene expressions. The changes in gene expression could possibly be recognized as diagnostic standards. To help identify genes whose transcript profiles respond to organ damage, it is essential to use statistical methods. Consequently, it might help to explore the molecular mechanisms of organ damage in PQ models because of whole gene expression analysis that provides an understanding of the crucial factors that control the organ damage and pathogenesis of PQ.
 
Therefore, in our research, we used the PQ-induced rodent model lung tissue and selected microarray analysis, with Affymetrix being our principal detection tool. We began by analyzing the gene expression with whole genome mouse GeneChip (Affymetrix). We identified the phosphoseryl-tRNA [Ser] Sec kinase (PSTK) gene as vital gene tied to PQ-stimulated lung damage via an integrative analysis that combined changes in gene expression with gene function within a genetic network. Secondly, we discovered that PSTK encoded protein is lowered in rodent lungs and confirmed the level of PSTK gene and encoded protein in injured lungs from PQ rodents.
 
Methods
Animals
For microarray analysis, 9 4-to-6-week-old male C57/BL6 rodents were purchased from the Shanghai Laboratory Animal Center (Shanghai, China). Additional 4-to-6-week-old male C57/BL6 rodents were purchased from the Shanghai Laboratory Animal Center (Shanghai, China). These 30 rodents were divided into 3 groups (n=10), Group A:  Normal control; Group B: PQ 6hrs and Group C: PQ 24hr. The average weight was 21 g. All rodents were raised and cared for in accordance with the guidelines established by the National Science Council of the People’s Republic of China[C1]  and were kept at least 1 wk at 22°C and 55% relative humidity in a 12-h day/night cycle environment with open access to nutritional supplements and liquids.
 
Microarray Analysis
For Affymetrix microarray profiling, total RNA was isolated with TRIzol reagent (Invitrogen, Canada) and purified by means of  RNeasy Mini Kit (Qiagen, German), including a DNase digestion treatment. RNA concentrations were determined by the absorbance at 260 nm and quality control standards were A260/A280 = 1.8–2.1, by means of NanoDrop 2000 (Thermo, America).
cDNA of lung of PQ rodents, and controls was hybridized to Mouse Gene 1.0 ST GeneChip® arrays (Affymetrix, America) according to the User Manuals. Affymetrix® Expression Console Software (version 1.2.1) was used for microarray analysis. Raw data (CEL files) were normalized at transcript level by means of the robust multiaverage method (RMA workflow). Median summarization of transcript expressions was calculated. Gene-level data was then filtered to include only those probe sets that are in the ‘core’ metaprobe list, which stimulate RefSeq genes.
 
Data Analysis
Because the RVM F-test can raise degrees of freedom effectively in the cases of small samples, we utilized the RVM F-test to filter the differentially expressed genes for the control and experiment group. We then selected the differentially expressed genes according to the p-value threshold after the significant analysis and FDR analysis,. The threshold of truly significant genes was taken to be p-value < 0.05 and FDR <0.05.
 
Cluster Analysis
 We used the Hierarchical Clustering tab, a powerful and useful method for analyzing all sorts of large genomic datasets, to perform hierarchical clustering on the data. A variety of published applications of this analysis are provided in this article’s references. The main principle behind the method is to collect a set of items (genes or arrays) into a tree, where items are joined by very short branches if they are very similar to each other, and by increasingly longer branches as their similarity decreases. To achieve this, a cluster analysis currently performs four types of binary, agglomerative, hierarchical clustering. Part one is calculating the distance matrix between the gene expression data. After the data is computed, the clustering begins.  The two best matching items (those with the smallest distance) are joined by a node/branch of a tree, with the length of the branch set to the distance between the joined items in an agglomerative hierarchical process. The two items are taken off the list of items being processed and are replaced by an item that stimulates the new branch. The differences between this new result and all other remaining results are computed, and the process is done again until there is only a final product. Each cluster contained a certain number of genes that have similar expression patterns after PQ. The p-value significance of the observed number of genes that fit a profile beyond the expected number of gene clusters was then ranked.
 
STC Analysis
We selected differential expression genes at a logical sequence according to RVM （Random Variance Model）corrective ANOVA.  The raw expression values were converted into log₂ratio and, in accordance with the different signal density change tendencies of genes under different situations we identified a set of unique model expression tendencies. We developed a technique for clustering short time-series gene expression data and defined some distinctive profiles. The expression model profiles are tied to the expected or the actual number of genes assigned to each model profile. Major profiles have higher likelihood of being distinct, as expected by Fisher’s exact test and multiple comparison tests. ^{13, 14}
 
GO Analysis
Any genes falling in significant model profiles (p<0.001), a total of 686 genes, underwent a vital functional classification of NCBI; Gene Ontology analysis according to the Gene Ontology. All the GO terms assigned to these genes were examined, and obtained by Fisher’s exact test and x² test for calculating the level of significance¹⁵. The FDR was designed to correct the p-value, the smaller the error in judging the p-value, the smaller the FDR. In addition, enrichment also provides a measurement of the significance the function provides: the corresponding function is more specific under our experimental condition as the enrichment increases ¹⁶. The criterion of p value <0.05 was used to eliminate significant GO terms.               
 
Dynamic Gene Network Analysis
To build a co-expression network we used the normalized intensity of the significant differential. On the basis of choosing the significant correlation gene pairs, the Pearson’s correlation of each pair of genes was first calculated. According to the correlation between genes the gene-gene interaction network was outlined. Within the network analysis, nodes stimulate genes, and the edges between them stimulate the interaction between them ¹⁷. All the nodes were marked with a link numbers one node has to the other; also known as a degree. K-core in graph theory was applied to describe the characteristics of the network including the centrality of genes within a network and the complexity of the sub-networks. In addition, genes with larger degrees occupied a more central position in the network and had a more pronounced capacity for modulating adjacent genes. According to the relationship between genes, they were marked with different colors and divided into several subnetworks¹⁸.
 
Paraquat Administration
We gave rodents paraquat via intragastric pathway (Sigma Chemical Co., St. Louis, MO) in 100 μl PBS at a concentration of 25 mg/kg, and the same dose of PBS was used as negative control.
 
Histology
One lung was fixed with 4% phosphate-buffered paraformaldehyde and embedded in paraffin whereas the other lung was removed and was frozen in liquid nitrogen. Tissue sections (4 μm) were then examined under light microscopy after being prepared and stained with hematoxylin/eosin. The slices (4 µm thick) were stained with hematoxylin and eosin (H&E) and deparaffinized then sent to a neutral pathologist who was not privy to the design details. The morphological changes of the lung were measured based on the criteria described by Hofbauer. ¹⁹
 
Quantitative PCR
We reverse-transcribed with a cDNA synthesis kit (Invitrogen, Carlsbad, CA), and real-time PCR by means of the SYBR (Invitrogen) after the quantitative reverse transcription–PCR RNA was isolated. This analysis method has been described above. ²⁰
 
Immunofluorescence
PSTK antibodies (Santa Cruz, Inc) were used as the principal antibodies for immunofluorescence staining of PSTK in the lung. Paraffin embedded lung sections were placed in an optimal cutting temperature compound (Sakura Fintek, Torrance, CA). Sections of 3 mm were dried overnight after being cut and set in cold acetone for a period of 8 min. For control sections we used normal rat immunoglobulin. Sections were incubated with biotinylated polyclonal rabbit anti-rat immunoglobulin (Dako Corporation, Glostrup, Denmark) or RTU Vectastain Elite ABC Peroxidase Kit (Vector Laboratories, Burlingame, CA), and 3,3–diaminobenzidine chromogenic substrate solution (Dako) was applied and then washed. Slides were counterstained with hematoxylin (Sigma-Aldrich, Steinheim, Germany). Goat anti-mouse IgG2a/2b FITC conjugated antibodies (1∶200; BD Biosciences Pharmingen), goat anti-rabbit Texas red conjugated antibodies (1∶100; Calbiochem). Pictures were taken by means of Spot Advanced version 3.4 (Diagnostic Instruments Inc.) with a fluorescence microscope (Olympus BX51) and Spot RT Slider camera. This analysis method has been described above. ²⁰
 
Western Blot
Before transferring the lung tissue from each group onto nitrocellulose membrane (Axygen, Union City, CA, USA) they were digested by protein extraction kit (QIAGEN, US) and then transferred onto nitrocellulose membrane (Axygen, Union City, CA, USA). The membranes were incubated with the antibodies against PSTK (Santa Cruz, Inc) before blocking with 5% non-fat milk in Tris-buffered saline and then a horseradish peroxidase (HRP)-linked secondary antibodies (Cell Signaling) for 1hr at room temperature. Chemiluminescence phototope-HRP kit performed the detection according to manufacturer’s instruction (Cell Signaling). As essential, blots were stripped and re-probed with Anti-beta-Actin (Merck, Darmstadt, Germany) antibody as internal controls. We repeated all experiments three times.
 
Statistics
Included in the statistical tests was two-tailed student’s test by means of one-way analysis of variance with Tukey’s multiple comparison tests. Results are expressed as the mean±s.d. A P-value of 0.05 was considered statistically significant. Statistical analysis was done by means of Prism (Version 4, Graphpad, La Jolla, CA).
 
Results
1. PQ Induced Lung Injury
Compared with the control group, the rodents developed the lung inflammation after the administration of Paraquat in both 6hr and 24hr groups. The rodents that received PQ displayed a collapse of pulmonary architecture, edematous interstitial space (Figure 1A) and severe infiltration of mononuclear cells into intra-alveolar and interstitial spaces. Meanwhile, lung histopathological analysis revealed that the rodents in the control group exhibited normal structure and slight inflammatory cell infiltration. The results reflected that lung damage occurred at 6hr and 24hr after PQ, which corresponded to the histopathological evaluation score (p<0.01 and p<0.05 respectively) (Figure 1B)
 
2. Genes Screened by Gene-chip Study of Lung Tissue Multi Class Dif
To investigate gene expression our normal control lung tissue was compared with profiles that are connected to PQ, and mRNA expression in lung tissue was profiled by means of gene chips at 6hr, 24hr. When compared to the control (p<0.05) there was a total of 2287 genes we identified as showing statistically significant differential expression in PQ rodents at least at 6h and 24h (Figure2).
 
Cluster Analysis of Significant Differential Genes
We used Cluster 3.0 software to examine the temporal expression pattern of significant differential gene expression. After PQ, each profile contains a cluster of multiple genes which have similar expression patterns. We identified 16 significant clusters containing a total of 2287 genes (Figure 3). After STC analysis, five of these clusters, example profiles 9, 4, 7 13 and 15, are comprised of genes that were repressed gradually and expressed different levels across the time points; while genes in profiles 4 and 1 had reverse effects. (Figure 3A). After PQ, the profiles of 7 and 4 shared comparable descending expression at 6hr and 24 hr. (Figure 3B)
 
 
GO Analysis
For each pattern Gene Ontology (GO) analysis was conducted. The high enrichment GO terms included: mitochondron, membrane and cytoplasm but in patterns such as pattern 4 and 7 the patterns with descending tendency of time course had the highest enrichments of GO which are all connected to mitochondrion. (Figure 4)
 
Dynamic Gene Network Analysis Based on Mitochondral Dysfunction
We further narrowed down the gene data set for PQ regulated genes in light of the compelling inner correlation between PQ mitochondrial dysfunction and stimulated pulmonary damage. We selected the candidate genes from the total differential gene as a function of mitochondron by gene co-expression network with k-core algorithm by means of the Dynamic Gene network analysis. Gene co-expression networks are built in accordance with the normalized signal intensity of differentially expressed genes. All genes were then identified and analyzed by a gene co-expression network with k-core algorithm to determine which gene or genes may play a pivotal role in the early stage of PQ. Gene networks were constructed from functional gene associations based on the results from the GO analysis. The diagram shown in Figure 5A shows the connection between genes. Each node describes the given gene and the relationship between a pair of genes is represented with a line segment. A degree describes the number of links one gene has to others within the gene network, with the most central genes in the network having the largest degree values.
 
To stimulate the connectivity of the adjacent genes with the node a clustering coefficient was applied to evaluate the property of a node in the network; the more complex the interactions of one gene among its neighboring genes, the higher the clustering coefficient. A k-core of gene co-expression network usually contains cohesive groups of genes. K-core in graph theory was applied to describe the characteristics of the network. There is one main k-core sub-network in our results, including 76 genes. Intriguingly, these 76 genes all belong to STC pattern 4 and 7.
 
PSTK Gene Co-expression Network Analysis
We discovered the PSTK gene in the network and that the PSTK gene plays a protective role in oxidative stress lung damage in the co-expression network of 76 genes of patterns 4 and 7. We selected the PSTK gene as a potential target to investigate. We also discovered that PSTK was only one of the most important of the ten genes in the co-expression network of 76 genes; the core gene PSTK appeared at the center of 11 k-core sub-network. According to their degrees, PSTK directly controls 30 neighboring genes that interact. (Figure 5B)  These are higher than other genes (above 0.99) so their interactions depend strongly on PSTK. After referring to the Gene Ontology database, the additional 30 genes in the network are involved in mitochondrial apoptosis, formation of acetoacetyl-CoA, complex I assembly, mitochondrial fatty acid beta-oxidation, which are consistent with the results of the GO analysis as they were directly tied to the pathogenesis of PQ lung We also confirmed that the expression of PSTK mRNA is gradually lowered damage. From the heat map we constructed these 30 differential genes and confirmed that the expression of PSTK mRNA is gradually lowered. (Figure 5C-D)
 
Real-time RT-PCR and Protein Verification of the PSTK Gene
Real-time RT-PCR was used to verify PSTK in vivo and to confirm the expression of the PSTK gene. The expression of PSTK showed a statistically significant decrease (p < 0.05) when compared to control lungs at 6hrs and 24hrs. Additionally, the level of PSTK from PQ rodents lowered four-folds more than the one from normal controls. However, there was no statistical significance (p > 0.05) between 24hrs and 6hrs groups. (Figure 6A) The PSTK protein level was measured in lungs from PQ rodents by means of Western Blot. There was no difference between 6 hr and 24 hr groups but it was discovered that PSTK was significantly less expressed in both the 6hr and 24hr groups as compared with that in the normal group. (Figure 6B-C)
 
Immunofluorescence Staining of PSTK in Lungs
We discovered when utilizing immunofluorescence staining that PSTK in lungs has a lower number in lungs from 6hr and 24hr groups as compared with the normal groups. (Figure 7A).The number of PSTK positive staining from a normal control group has virtually two times more than from both 6hr and 24hr PQ groups under HPF magnitude microscopy. (Figure 7B)
 
 
Discussion
 
We used a gene chip-basis in our study to screen the change of gene expression in PQ lungs, and used gene expression profiles, gene network construction and real-time RT-PCR, GO-Analysis, western blot and immunofluorescence. Totally 2287 genes with different expression differences were identified, and we discovered that PSTK was the core gene tied to the 30 key genes which function as mitochondrial apoptosis, formation of acetoacetyl-CoA, mitochondrial fatty acid beta-oxidation, and complex I assembly. This implied that PSTK played a crucial role in oxidative stress lung damage. We finally confirmed the levels of the PSTK gene and discovered that PSTK was lowered in rodent lungs encoded protein in PQ stimulated rodents.
 
The rodents developed lung injuries at 6hr and 24hr after PQ administration, which were characterized by cell infiltration in pulmonary architecture, inflammation intra-alveolar and interstitial spaces, indicating that at 6hr and 24hr PQ model has been successfully established.
.
This technology could also be used to distinguish gene expression at a specific state such as the early stage of PQ because a high-throughput gene chip had been used to explore the transcription dynamics on a genome scale. It was reported by Yang et al. that long noncoding RNAs -HEIH is an oncogenic IncRNA which stimulates tumor progression and might serve as critical regulatory focal points in hepatocellular carcinoma (HCC) progression by means of high-throughput microarray²¹. With the use of a PQ mouse model, we had focused on lung gene expression during early stages of PQ. We identified PSTK whose expression was significantly altered at the very early stage and have shown large correlations with PQ by a series of gene chip analysis. Gene chips have the benefit of generating vast amounts of data. The high expenditure of micro-array experiments and limited sample accessibility created a small sample size for each treatment group. In the framework of patterns that can be anticipated in the course of growth, we established a certain sets of profiles. Any one pattern will be shared by many genes and can be expected to appear by chance due to the noise and the small number of points for each gene. To start with, we targeted a set of latent expression profiles. These sets of profiles stimulate a single temporal expression pattern and cover all of the possible profiles that can be generated by gene expression in the course of the disease. To measure the profile significance level enrichment of genes in each profile was used. Common functions attributed to genes that are co-expressed were indicated by significant profiles. Such functions mainly control the biological makeup ²². We can look at snapshots of protein-protein interaction, gene expression regulatory networks, and metabolism networks among different groups because of the plasticity of the network model based on the algorithmic predictions from high throughput gene expression tests. The intrinsic gene networks of a phenotype can stimulate gene function propriety of samples. Key attributes in the network include degrees of genes and K-cores. A k-core of a network is a sub-graph in which all genes are linked to at least k other genes in the sub-graph ²³. The k-core value rank rose with the increasing complexity of gene relationship. We wanted to define vital gene functions at each complexity level of the network by finding the main Gene Ontology (GO) assigned by the maximum numbers of genes in separately k-core and then define the key gene functions at each complexity level of the network ²⁴. For this analytical result, we concluded that the core functions at the core status of the networks which have a top k-core level. The network responds to the events such as mitochondrial matrix, ribosome protein synthesis, and extra- and intra-cellular membranes.
 
In our study, the PSTK gene was selected by means of the network and validated by Real-time RT-PCR. The expression of the PSTK gene appeared to be a statistically significant change (p < 0.05) when compared with normal controls. A critical concern regarding the results from a GeneChip-based analysis is their association to the actual biological processes. RNAi of genes encoding mitochondrial respiratory complexes I leads to an increased life span in flies with paraquat-stimulated free-radical generator models. ²⁶ Also, as revealed by the activation of caspase-3/-9, cleavage of PARP, depletion of mitochondrial membrane potential regulated by Bcl-2 family members, and overproduction of reactive oxygen species ²⁵ as reported by Copeland et al., all 31 genes identified in this PSTK co-expression network are involved in the mitochondria function of cells. Mitochondria-mediated apoptosis pathway contributed importantly to PQ-stimulated apoptosis.
 
We demonstrated that mutation in this locus affects lipid homeostasis and the oxidative stress response and that Enigma encodes a mitochondrial protein with homology to enzymes of the beta-oxidation of fatty acids. ²⁷ A clear implication is that PSTK could act as a strong bond to the part of mitochondrial function which plays a pivotal role in the process of PQ. As lung fibrosis and respiratory failure are the main reasons of death by PQ, the alteration of the level of the PSTK gene at the very early stage could be followed by severe damage to the lungs, the pathological characteristic of PQ was revealed as large area of pulmonary fibrosis, thereby contributing to potential diagnostic information for clinical PQ. As mentioned earlier, a PQ-stimulated rodent model has broadly been used as an experimental model of oxidative stress organ damage, therefore we employed a PQ animal model instead of human tissue in this study as human tissue is extremely difficult to obtain.
 
PSTK is the O-phosphoseryl-tRNA (Sec) kinase (PSTK) that changes Ser-tRNA (Sec) to Sep-tRNA (Sec), which further leads to the creation of selenium. Selenium proteins, such as GPx, have been reported to take on the function of anti-oxidative stress ²⁸. PQ damage, on the other hand, is characterized by the inner oxidative stress before finally leading to lung fibrosis²⁹. In this study, we discovered that PSTK dramatically lowered in lungs, possibly due to the counterbalance of body stress, particularly when the PQ-stimulated oxidative stress damage was developed at an early stage. Nonetheless, there are no reports demonstrating the connection between the PSTK gene and lung damage.
 
Both PSTK mRNA and protein was depressed in rodent samples without fail. PSTK has been shown to express largely in lungs, implying that PSTK might be involved in the routine hemostasis of the lung. However, we discovered that PSTK was lowered in injured lungs, implying that it contributes to the PQ-triggered inflammation in the lung. It is assumed here that PSTK is a protector in preventing oxidative stress because up to the present time there have been no reports that PSTK is a protective molecule in the process of oxidative stress in the lung. Furthermore, PSTK was found in surrounding and pulmonary alveolars; this strengthens the likelihood of PSTK's protective role.
 
Selenium's function as an anti-oxidant has been discussed above. Oxidative stress is understood to have a critical role in the development of acinar damage in experimental acute pancreatitis as noted by Hardman et al. Compound multiple antioxidant therapy has previously been demonstrated to ameliorate end-organ damage in the intra-peritoneal L-arginine rat model. Selenium is a key constituent of several antioxidant preparations as the principal co-factor for glutathione. Intravenous selenium administered 24 hours after induction of experimental acute pancreatitis was connected to dramatic reduction in bronchoalveolar lavage protein content and a decrease in the histological stigmata of pancreatic damage 30. Whether the supplement of selenium can help restore PQ stimulated lung damage remains unclear, however oxidative stress has been discovered to worsen lung damage³¹
 
The model resembles human PQ-stimulated lung injuries, as confirmed by our pathological and biochemical data. For that reason, our research is of important referential value for a clinical examination. It still must be noted that the genes identified in our study require further dissection and confirmation in human trails by other clinical-related studies.
 
Conclusion
This study is the first to select PSTK as a potential core gene in organ damage from PQ and utilize high-throughout gene chips to analyze the differential genes. Expression of the PSTK gene involved in selenium protein synthesis changed significantly in lung tissue at the early stages of PQ. The gene has shown a strong correlation with PQ development and the network that it forms. Therefore this study has presented the chance to make use of selenium protein as therapeutic target in curing PQ patients in clinics.
 
Acknowledgements
Our research was funded by the NFSC (National Science Fund Committee) (No. 09JC1412100), China.
MRNA microarray experiments and technical assistance in bioinformatic analysis were handled by Genminix Informatics Ltd., Shanghai, China.