|Year : 2016 | Volume
| Issue : 1 | Page : 13-18
Bio-informatics analysis of renal carcinoma gene matrix metalloproteinase-7
L Li, LX Wang, GL Xu, F Yang, QL Gao, H Niu, B Shi, X Jiang
Department of Medical Oncology, The First Affiliated Hospital of Zhengzhou University, and Henan Cancer Hospital, Affiliated to Zhengzhou University, Zhengzhou, Henan, China
|Date of Web Publication||28-Apr-2016|
L X Wang
Department of Medical Oncology, The First Affiliated Hospital of Zhengzhou University, and Henan Cancer Hospital, Affiliated to Zhengzhou University, Zhengzhou, Henan
Source of Support: This project was supported by medical
science research project in Henan province (No. 201303229). In
addition, it also got financial support from ShengYuan Science
and Technology Cooperation Projects in Henan Province (No.
122106000042) and Science and Technology Open Cooperation
Projects in 2013 in Henan Province (No. 132106000064), Conflict of Interest: None
Background: Renal cancer is one of the common malignant tumors of the urinary system, seriously threatening human being's health. The current discoveries, however, are far enough for efficient and secure treatment of renal cancer. Aims: The aim was to explore the mechanism of matrix metalloproteinase-7 (MMP-7) protein in renal carcinoma cell metastasis by bioinformatics analysis. Materials and Methods: Bioinformatics methods were used to analyze the composition of amino acids, as well as transmembrane structure, coiled coils, subcellular localization, signal peptide, functions and structures at all levels. Results and Conclusions: It showed that the gene MMP-7 totally had 1131 bp. A peptide chain containing 267 amino acids was encoded in the coding region. Based on random coil, α helix, and further super-helix, it had formed a stable neutral hydrophilic protein. The subcellular location analysis indicated that the protein was located outside the cell. The mature peptide started from the 18th amino acid, and its front-end was the sequence of the signal peptide, belonging to the secreted protein. Analysis of the functional domain showed that this protein had two functional domains, the PG binding domain, and the zinc finger binding domain. Moreover, the protein, which was cross-linked with it, was also one related to cancer cell proliferation and metastasis. To sum up, MMP-7 is a stable neutral hydrophilic secreted protein, and it may play a vital role in the invasion and metastasis of cancer cells.
Keywords: Bioinformatics, matrix metalloproteinase-7, renal carcinoma
|How to cite this article:|
Li L, Wang L X, Xu G L, Yang F, Gao Q L, Niu H, Shi B, Jiang X. Bio-informatics analysis of renal carcinoma gene matrix metalloproteinase-7. Indian J Cancer 2016;53:13-8
|How to cite this URL:|
Li L, Wang L X, Xu G L, Yang F, Gao Q L, Niu H, Shi B, Jiang X. Bio-informatics analysis of renal carcinoma gene matrix metalloproteinase-7. Indian J Cancer [serial online] 2016 [cited 2020 Jul 4];53:13-8. Available from: http://www.indianjcancer.com/text.asp?2016/53/1/13/180835
| » Introduction|| |
Renal cancer, also called renal cell carcinoma (RCC), is one of the common malignant tumors of the urinary system, seriously threatening human being's health. Its incidence accounts for 3% of the total malignant tumors in adults. In recent years, as the population of obesity and hypertension are becoming large, the incidence of renal cancer is increasing at a rate of 2%/year. There are no obvious symptoms in early renal cancer. Imaging examination is mainly used for the diagnosis of renal cancer. Radical resection procedure of kidney and open operation are the only two effective means in treating renal carcinoma. However, the occurrence of cancer, regulated by multiple genes, is a process with multistages. Examination of the abnormities of renal carcinoma in time is of vital significance in preventing renal carcinoma. Currently, reports on kidney cancer genes are still rare. With the approach of postgenomic era, there will be more and more studies about the related genes of renal cancer, such as the antioncogenes VHL (von Hippel-Lindau), PTEN (phosphatase and tensin homolog), matrix metalloproteinase-2 (MMP-2), proto-oncogene pokemon  and the specific genes G250, GYLZ-RCC18 and CD117 and so forth. Those preliminary findings, however, are far enough for efficient and secure treatment of renal cancer. Therefore, it is especially necessary to make a deep analysis of the characteristics of the genes.
Matrix metalloproteinase is a kind of endopeptidase functioning in the extracellular matrix. It promotes the angiogenesis of cancer through transforming extracellular matrix to achieve the invasion and metastasis of cancer. Since the first discovery of MMP-1 in 1962, 23 members of the MMPs family have been discovered successively. It can be divided into five categories. Matrix metalloproteinase (MMP-7) belongs to the type of stromlysin, with extensive substance specificity and strong extracellular matrix degradation function. It is mainly produced by cancer cells and the surrounding mesenchymal cells. And almost all the components of extracellular matrix can be degraded. In the meantime, it can promote the angiogenesis of cancer through transforming extracellular matrix to achieve the invasion and metastasis of cancer. Many studies have proved the abnormal expression of MMP-7 in gastric cancer, colorectal cancer, mammary cancer, liver cancer  and prostatic cancer  and MMP-7 is closely related to the metastasis of cancer cells. Through bioinformatics analysis of the structure of MMP-7, the study aims to discover the potential functions of MMP-7 and provide theoretical evidence for corresponding treatment of targeted sites.
| » Materials and Methods|| |
NCBI accession number of MMP-7 is GenBank: BC003635.2 and protein accession number is AAH03635.1.
Software ProtParam, Compute pI/Mw and ProtScale of ExPASy (ExPASy Bioinformatics Resource Portal, Swiss Institute of Bioinformatics, Swiss) server were used to predict amino acid composition, molecular weight, isoelectric points, hydrophilic/hydrophobic ability, instable index and fat coefficient of MMP-7 protein. Kyte and Doolittle method was adopted to predict the hydrophilic/hydrophobic ability of MMP-7 amino acid sequence. Various methods were used for the analysis of MMP-7. TEPRED tool was selected to analyze the transmembrane region and direction of MMP-7 protein; software TMHMM (Center for Biological Sequence Analysis Prediction Servers, Denmark) was used to analyze the topological structure of transmembrane of MMP-7; COILS coiled coils prediction tool was used to analyze and predict MMP-7 protein; prediction of the secondary structure of MMP-7 protein was carried out through neural network model of PBIL LYON-GERLAND information bank; PSORT (WoLF PSORT, Computational Biology Research Center, Japan) was selected to predict subcellular location; SignalP3.0 (CBS Prediction Servers, Denmark) Server was used to predict signal peptide; SWISS-MODE was adopted to predict the tertiary structure; Scansite was used to analyze the functional and structural domains of MMP-7 protein sequence; NCBI conserved domain database was used to analyze the conserved domain of MMP-7 protein sequence; STRING 9.1 (NNF Center for Protein Research, Copenhagen) was used for mutual analysis of the interlink between MMP-7 protein and other proteins.
| » Results and Analysis|| |
Sequence analysis of the primary structure of matrix metalloproteinase-7
ExPASy server package was used to predict the primary structure parameters of MMP-7 protein. According to ProtParam method, when the value of instable index was <40, the predicted protein was stable in the experiment, otherwise it would show instability. In [Table 1], the instable index of MMP-7 protein was 31.67, showing the protein was stable. MMP-7 totally had 1131 bp, including the coding region and the noncoding region. A peptide chain containing 267 amino acids was encoded in the coding region. The average hydrophobicity of the peptide chain encoded by this gene was −0.364, showing the protein may be a hydrophilic protein.
In [Table 2], it shows that 20 kinds of amino acids composed MMP-7 encoding protein. The content of leucine (Leu) was the highest, 10.1%; the content of glycine (Gly) was a secondary, 9.7%; the contents of cysteine (Cys) and tryptophane (Trp) were the lowest, respectively 1.1% and 1.9%.
Hydrophilic/hydrophobic ability is the inherent characteristics of amino acids. The stability of protein structure largely depends on the hydrophobic ability within molecules. Prediction and analysis of hydrophilic/hydrophobic ability are of great value in the prediction and functional analysis of protein secondary structure. Hence, the paper adopted Kyte and Doolittle scale to calculate, and the window was set 21. Then evident hydrophilic signal was gained. In [Table 2], a positive value represented hydrophobic ability and the negative represented the hydrophilic ability. The results in [Figure 1] showed that the N-terminal of the protein appeared to be obviously hydrophobic while C-terminal appeared to be strongly hydrophilic.
|Figure 1: Analysis results of hydrophilia/hydrophobicity of matrix metalloproteinase-7 amino acid sequence|
Click here to view
Transmembrane structure prediction of matrix metalloproteinase-7 protein
The transmembrane structural domain, made up of 20 or so hydrophobic amino acids, is where the combination of the membrane intrinsic protein and membrane lipid happens. It mainly forms the helix α, attached to a cell membrane with the role of anchoring. Therefore, the prediction and analysis of protein transmembrane structural domain has some implications for the knowledge of protein structure, function and classification. Take the transmembrane protein database Tmbase for example. Generally speaking, if the score was >500, the transmembrane structure was thought to be of the high possibility. It showed in [Figure 2]a that there were three transmembrane helical regions from the inside to the outside and the positions of amino acids were respectively 3–20 (score 1338), 147–165 (score 469) and 204–222 (score 68). There was one transmembrane helical region from the outside to the inside, and the position of amino acid was 1–18 (score 1378). Software TMHMM was used to analyze the transmembrane topological structure of MMP-7 [Figure 2]b, finding that the protein was an extracellular protein, and transmembrane structural domain may exist at N-terminal. The above results proved that MMP-7 secreted within cells and then moved to the extracellular region through the transmembrane domain of N-terminal. Hydrolysis was carried out in the extracellular region.
|Figure 2: Prediction on transmembrane region of matrix metalloproteinase-7 protein based on Tmbase and TMHMM software|
Click here to view
Coiled helix prediction of matrix metalloproteinase-7 protein
Coiled helix is also called super-helix. Generally, it refers to the helix formed by the mutual entanglement of two or more helical polypeptide chains. Studies showed that proteins containing coiled helical structures were mainly membrane protein, the structural protein or transcription factor. Their functions were mainly molecular chaperones, metabolic regulation and membrane channel. COILS coiled helix prediction tool of EMBnet was used to get the prediction results of MMP-7 protein [Figure 3]. The protein contained coiled helical structure, which indicating it may belong to membrane protein, metabolic regulation protein or transcription factor.
|Figure 3: Analysis results of coiled coil of matrix metalloproteinase-7 protein|
Click here to view
Analysis of matrix metalloproteinase-7 secondary structure
Hopfield neural net (HNN) of PBIL LYON-GERLAND database (http://www.npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_hnn.html) was used to predict secondary structure of protein sequence. And it showed that random coil of the secondary structures of MMP-7 protein occupied the largest proportion, reaching 54.68%, while alpha helix and the extended chain occupied the proportions of 29.21% and 16.10% respectively [Table 3] and [Figure 4].
|Figure 4: Prediction on secondary structure of matrix metalloproteinase-7 protein|
Click here to view
Functional classification of subcellular location of matrix metalloproteinase-7 protein
The location of proteins in cells are closely related to the functions of proteins. Through subcellular location, the position of the protein in cells can be primarily achieved. PSORT (WoLF PSORT, Computational Biology Research Center, Japan) was used to predict the subcellular location of MMP-7 protein. [Table 4] shows that the location of MMP-7 protein was possible among intercellular space outside cells. Software TargetP (CBS Prediction Servers, Denmark) was used to make another analysis of subcellular location and [Table 5] was obtained. RC represented credibility, which sloped down from 1 to 5. The results showed that MMP-7 was secreted protein, and the value of credibility was 1. Hence, the protein was a secreted protein located in the intercellular space. In the eukaryotic and prokaryotic cells, there was a short hydrophobic peptide chain called signal peptide at N-terminal, whose function was to lead the secreted proteins to enter or implant the endoplasmic reticulum in the process of translation. Because N-terminal was hydrophobic, it can be predicted the signal peptide was at N-terminal.
|Table 4: Prediction on subcellular location of MMP-7 protein based on PSORT software (WoLF PSORT, computational biology research center, Japan)|
Click here to view
|Table 5: Prediction on subcellular location of MMP-7 protein based on targetP software (CBS prediction servers, Denmark)|
Click here to view
Prediction and analysis of the signal peptide of matrix metalloproteinase-7 protein
Signal peptide consists of 20–30 amino acid residues at N-terminal of newly formed peptide chain of secreted protein. It modifies some amino acid residues and is often used to direct the transmembrane metastasis of proteins. SignalP-4.1 Server was adopted to predict the signal peptide of MMP-7 protein and the results are shown in [Figure 5]. HNN analysis showed that the 18th amino acid residue had the highest score of primary splice site 0.830 and the signal peptide score of the secondary amino acid residue was 0.925. The signal peptide score of 1–17th amino acid residue was 0.904 and the comprehensive splice site score of 18th amino acid was the highest 0.868. The splice site between the signal peptide and mature peptide chain was probably between the 17th amino acid and the 18th amino acid. Therefore, it was inferred that the N-terminal of MMP-7 contained a section of the signal peptide, and the mature peptide chain may start from the 18th amino acid. It is in accordance with the former results.
|Figure 5: Prediction on signal peptide of matrix metalloproteinase-7 protein|
Click here to view
Prediction and analysis of matrix metalloproteinase-7 tertiary structure
The method of SWISS-MODEL online homology modeling was used to analyze the MMP-7 protein [Figure 6]. It was inferred that there were two proteins in tertiary structure that were homogenous to MMP-7. The first protein matched 169 amino acid residues, from 92nd to 260th, and the degree of similarity was 100%. The second protein matched 229 amino acid residues, from 34th to 262nd, and the degree of similarity was 52.17%.
Threading method of PHYRE2 Protein Fold Recognition Server was used to establish the model of MMP-7 protein [Figure 7]. The credibility of the following model was 100% and the coverage rate was 92%, which meant 245 amino acid residues were accurately described.
Prediction of matrix metalloproteinase-7 functional domain
By means of the analysis of conserved structural domain database of NCBI, it was found that there were two conserved domains in MMP-7 protein, respectively PG combined domain and zinc finger domain [Figure 8]. The sequence numbers of the two domains were respectively PS00546 and PS00142 through Scansite analysis [[Table 6].
|Figure 8: Analysis of functional sites of matrix metalloproteinase-7 protein|
Click here to view
|Table 6: Structural and functional sites obtained through ScanProsite search|
Click here to view
Interlinked effect of proteins
STRING9.0 interactive database was used to do cross-linked analysis of proteins. It was found that functional proteins were mainly ZBTB33, NGF, COL18A1, MMP1, TIMP1, CTNNB1, JUP, CD44, SDC1 and heparin binding epidermal growth factor (HBEGF). ZBTB33 also had zinc finger structure and was related to the abnormal proliferation of the cancer. TIMP1 was the inhibitor one of matrix metalloproteinase, participating in the concentration of multiple cancers.,, The HBEGF belongs to carcinogens. It can express adequately in cancer cells and promote the metastasis of cancer cells., Hence, MMP-7 probably played its part in the proliferation and metastasis of cancer cells when it was interacting with these proteins [Figure 9].
| » Discussion and Conclusion|| |
Renal cancer is the common malignant tumor of the urinary system, originated in the renal tubular epithelial cells. In China, it ranked number two in urological cancers, second to bladder cancer. Moreover, the morbidity in different regions also varies, but keeping an increasing trend. As studies on tumors going deep in recent years, it is widely believed by most experts that local infiltration and metastasis are the basic biological behaviors in malignant tumors, and it is unavoidable. Generally speaking, the relation between cancer cells and the environment is complex. The metabolic disorder of extracellular matrix can cause the invasion and metastasis of cancers. The process can be divided into three stages. First, cancer cells adhere to extracellular matrix, releasing or inducing proteomic enzymes to degrade extracellular matrix and matrix membrane. Then, due to the induction of chemotactic factors, translation of cancer cells took place. At the last, through blood and lymph circulation, the tumor cells are transferred to the distant tissues where the hyperplasia and metastasis have formed. MMP-7 is the main enzyme to degrade extracellular matrix, meaning that it has paved that way for the external expansion of tumors at the first step of tumor infiltration and metastasis.
In this study, the predictive molecular weight of MMP-7 protein was 29.66 kD, which was accordant with the studies of Sarkissian et al. through the method of two-dimensional electrophoresis and Western blot hybridization. As to the isoelectric point, it was predicted as 7.1. Sarkissian et al. found it was 7.7 after analysis, showing that the predictive results were basically in accordance with the experimental results. Catania and his team  conducted pathophysiological analysis of MMPs and it showed that the sequence of MMP-7 contained external secreted section, protein body fragments and the catalytic reaction zone. This is also in consistent with the prediction analysis of this paper that MMP-7 is an extracellular protein containing signal of secretion. This study indicates that the HBEGF protein, interacted with MMP-7, shows high expression in cancer cells, and it is in line with the study made by Lu et al. Through the analysis of MMP-7 gene expression of 98 patients with renal cancer, Lu et al. found the expression ratio of MMP-7 in cancer cells was obviously higher than that in normal tissues. The expression ratio of MMP-7 became larger as the pathological grading increased. MMP-7 has been treated as the marker of renal cancer in the patent application in 2011. In this patient, 177 patients with renal cancer and 116 cases (for control) were used as the study subjects. Form the ROC curve, the area value of MMP-7 as the marker of renal cancer was 0.926. By detecting the content of serum MMP-7 protein of 30 fresh cases, 30 patients and 40 contrast patients, Sarkissian et al. found that the content of MMP-7 protein in patients group was 3.5 times of that in the healthy group, which proved that MMP-7 can be regarded as the serum marker of renal cancer. Therefore, this study has offered certain theoretical basis in exploring the mechanism of MMP-7, and it will provide new ideas for anticancer treatment in the future.
To sum up, the occurrence and development of renal cancer is a complex process with polygenes and multi-factor interacted in each other. And the exploration of the mechanism of renal cancer from the gene level is of the positive effect on developing specific gene diagnosis and targeted therapeutic drugs.
| » References|| |
Zhang YZ, Yang GQ, Zhang SW, Zheng RS, Cao L, Chen WQ. An analysis of incidence and mortality of kidney and unspecified urinary organs cancer in China, 2009. China Cancer 2013;22:333-7.
Rini BI, Campbell SC, Escudier B. Renal cell carcinoma. Lancet 2009;373:1119-32.
Pan SH, Wang S, Xu G, Wu GF, Yan JJ. Clinical comparison between the treatment of renal cancer through the retroperitoneal approach and heminephrectomy under transabdominal laparoscope. J Clin Urol 2010;25:651-3.
Sun FD, Wang DW, Mi ZG. Present situation and research progress of Molecular targeted therapy for renal cell carcinoma. J Contemp Urol Reprod Oncol 2012;4:1-3.
Zhang T, Luo ZG. Research progress of renal cancer gene. Contemp Med 2011;17:25-6.
Hu XH, Chen HB, Zhu SL, Liao ZL, Zhao CN. Expression and clinical significance of PTEN and matrix metalloproteinase-2 in kidney cancer. China Med Her ISTIC 2012;9:63-4.
Hou LJ, Zheng JH, Geng J, Gu WY, Wang B. Expression of pokemon mRNA and Its clinical significance in renal cell carcinoma. China Mod Doct 2011;49:17-9.
Yan GB. Matrix metalloproteinases. Chin J Joint Surg (Electron Version) 2010;4:033.
McDonnell S, Navre M, Coffey RJ Jr, Matrisian LM. Expression and localization of the matrix metalloproteinase pump-1 (MMP-7) in human gastric and colon carcinomas. Mol Carcinog 1991;4:527-33.
Sentani K, Matsuda M, Oue N, Uraoka N, Naito Y, Sakamoto N, et al.
Clinicopathological significance of MMP-7, laminin γ2 and EGFR expression at the invasive front of gastric carcinoma. Gastric Cancer 2014;17:412-22.
Koskensalo S, Louhimo J, Nordling S, Hagström J, Haglund C. MMP-7 as a prognostic marker in colorectal cancer. Tumour Biol 2011;32:259-64.
Bucan V, Mandel K, Bertram C, Lazaridis A, Reimers K, Park-Simon TW, et al.
LEF-1 regulates proliferation and MMP-7 transcription in breast cancer cells. Genes Cells 2012;17:559-67.
Wang X, Xia Y, Liu Y, Qi W, Sun Q, Zhao Q, et al.
Dual-luminophore-labeled gold nanoparticles with completely resolved emission for the simultaneous imaging of MMP-2 and MMP-7 in living cells under single wavelength excitation. Chemistry 2012;18:7189-95.
Shin S, Oh S, An S, Janknecht R. ETS variant 1 regulates matrix metalloproteinase-7 transcription in LNCaP prostate cancer cells. Oncol Rep 2013;29:306-14.
Wei X, Zeng XG, Zhou HM. Progress on the study of coiled coils in protein structures. Chin J Biochem Mol Biol 2004;20:565-71.
Chaudhary R, Pierre CC, Nanan K, Wojtal D, Morone S, Pinelli C, et al.
The POZ-ZF transcription factor Kaiso (ZBTB33) induces inflammation and progenitor cell differentiation in the murine intestine. PLoS One 2013;8:E74160.
Fowell AJ, Collins JE, Duncombe DR, Pickering JA, Rosenberg WM, Benyon RC. Silencing tissue inhibitors of metalloproteinases (TIMPs) with short interfering RNA reveals a role for TIMP-1 in hepatic stellate cell proliferation. Biochem Biophys Res Commun 2011;407:277-82.
Luo J, Qiao F, Yin X. Hypoxia induces FGF2 production by vascular endothelial cells and alters MMP9 and TIMP1 expression in extravillous trophoblasts and their invasiveness in a cocultured model. J Reprod Dev 2011;57:84-91.
Chen Y, Wei X, Guo C, Jin H, Han Z, Han Y, et al.
Runx3 suppresses gastric cancer metastasis through inactivation of MMP9 by upregulation of TIMP-1. Int J Cancer 2011;129:1586-98.
Zhou ZN, Sharma VP, Beaty BT, Roh-Johnson M, Peterson EA, Van Rooijen N, et al.
Autocrine HBEGF expression promotes breast cancer intravasation, metastasis and macrophage-independent invasion in vivo
. Oncogene 2014 17;33:3784-93.
Schwenke M, Knöfler M, Velicky P, Weimar CH, Kruse M, Samalecos A, et al.
Control of human endometrial stromal cell motility by PDGF-BB, HB-EGF and trophoblast-secreted factors. PLoS One 2013;8:E54336.
Chen WQ, He YT, Zhang SW, Zou XN. Analysis of renal cell carcinoma mortality in China, 2004-2005 – An analysis of the third national death sampling retrospective survey. Chin J Cancer Prev Treat 2011;18:252-5.
Sarkissian G, Fergelot P, Lamy PJ, Patard JJ, Culine S, Jouin P, et al.
Identification of pro-MMP-7 as a serum marker for renal cell carcinoma by use of proteomic analysis. Clin Chem 2008;54:574-81.
Catania JM, Chen G, Parrish AR. Role of matrix metalloproteinases in renal pathophysiologies. Am J Physiol Renal Physiol 2007;292:F905-11.
Lu H, Yang Z, Zhang H, Gan M, Zhou T, Wang S. The expression and clinical significance of matrix metalloproteinase 7 and tissue inhibitor of matrix metalloproteinases 2 in clear cell renal cell carcinoma. Exp Ther Med 2013;5:890-6.
Darbouret B, Sarkissian G.In vitro
method for diagnosing and monitoring renal cellcarcinoma (RCC) using MMP-7 as humoral biomarker for RCC: U.S. Patent 7,981,621 [P]; 2011.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8], [Figure 9]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]
|This article has been cited by|
||Circular RNA hsa_circRNA_0007334 is Predicted to Promote MMP7 and COL1A1 Expression by Functioning as a miRNA Sponge in Pancreatic Ductal Adenocarcinoma
| ||Jinghui Yang,Xianling Cong,Ming Ren,Hongyan Sun,Tao Liu,Gaoyang Chen,Qingyu Wang,Zhaoyan Li,Shan Yu,Qiwei Yang |
| ||Journal of Oncology. 2019; 2019: 1 |
|[Pubmed] | [DOI]|
||The role of kinesin KIF18A in the invasion and metastasis of hepatocellular carcinoma
| ||Weiwei Luo,Minjun Liao,Yan Liao,Xinhuang Chen,Chunyan Huang,Jiyuan Fan,Weijia Liao |
| ||World Journal of Surgical Oncology. 2018; 16(1) |
|[Pubmed] | [DOI]|