Analysis of protein expression regulated by lumican in PANC‑1 cells using shotgun proteomics
- Authors:
- Published online on: July 11, 2013 https://doi.org/10.3892/or.2013.2612
- Pages: 1609-1621
Abstract
Introduction
Lumican is a member of the class II small leucine-rich proteoglycan (SLRP) family. Members of this family have relatively small molecular sizes, with core proteins of approximately 40 kDa, and possess 6–10 leucine-rich repeat units in the core protein (1,2). Amino acid sequencing indicates that lumican has 4 potential sites for N-linked keratan sulfate (KS) or oligosaccharides (3,4). Therefore, lumican includes a core protein, glycoprotein and proteoglycan forms due to glycosylation (5). Lumican is a secreted collagen-binding extracellular matrix protein of the cornea, dermis and tendon stroma, arterial wall, and intestinal submucosa (6–9). Corneal opacity, as well as skin and tendon fragility due to disorganized and loosely packed collagen fibers in lumican-null mice suggest that lumican plays an important role in collagen fibrillogenesis (10,11).
Lumican was first reported as one of the major KS proteoglycans in the chicken cornea (12). In addition to the cornea, lumican expression has been reported in various human tissues, including malignant tumor tissues (5,13–27). Among the clinicopathological characteristics of pancreatic ductal adenocarcinoma (PDAC), the localization of lumican in the stromal tissue adjacent to cancer cells correlates with advanced cancer stage, retroperitoneal and duodenal invasion, and residual tumor, and tends to correlate with shorter survival (21). These reports suggest that lumican localized in the stromal tissue is secreted from cancer cells and affects cancer cells through an autocrine and paracrine mechanism. We previously reported that PANC-1 cells, one of the PDAC cell lines, secrete only 70-kDa glycosylated lumican into the extracellular space. We also demonstrated that the secreted lumican stimulated cell growth through ERK activation and inhibited cell invasion and matrix metalloproteinase (MMP)-9 activation using lumican-overexpressing PANC-1 cells and lumican-downregulated PANC-1 cells (28). However, the mechanism of how lumican affects cell growth and invasion remains unclear.
In the present study, we performed shotgun liquid chromatography (LC)/mass spectrometry (MS)-based global proteomic analysis using protein from lumican-overexpressing PANC-1 cells and lumican-downregulated PANC-1 cells to examine how lumican regulates cell growth and invasion in PDAC cells. We identified 24 candidate proteins that may play an important role in cell growth and invasion and could be regulated by lumican.
Materials and methods
Materials
The following materials were purchased from Wako Pure Chemical Industries (Osaka, Japan): urea, 3-(3-cholamidopropyl) dimethylammonio-1-propanesulphonate (CHAPS), dithiothreitol (DTT), Tris (2-carboxyethyl) phosphine hydrochloride (TCEP) and iodoacetamide (IAA); Amicon Ultra 0.5-ml 3K was from Millipore (Tokyo, Japan), and thiourea from Nacalai Tesque, Inc. (Kyoto, Japan). All other chemicals and reagents were purchased from Sigma Chemical Corp. (St. Louis, MO, USA).
PDAC cell line
PANC-1 cells were obtained from the Cell Resource Center for Biomedical Research, Institute of Development, Aging and Cancer, Tohoku University (Sendai, Japan).
Protein extraction of lumican-regulated PANC-1 cells
The lumican-overexpressing PANC-1 cells, lumican-downregulated PANC-1 cells, and control cell lines (Mock and NC, respectively) were prepared as previously described (28). The lumican-regulated PANC-1 cells were cultured at a density of 5×105 cells in a 100-mm dish in RPMI-1640 medium with 10% fetal bovine serum (FBS) for 72 h. Then, cells were solubilized in urea lysis buffer (7 M urea, 2 M thiourea, 5% CHAPS, 1% Triton X-100). Protein concentration was measured using the Bradford method.
In-solution trypsin digestion
A gel-free digestion approach was performed in accordance with the protocol described by Bluemlein and Ralser (29). In brief, 10 μg of protein extract from each sample was reduced by addition of 45 mM DTT and 20 mM TCEP and was then alkylated using 100 mM IAA. Following alkylation, samples were digested with Proteomics Grade Trypsin (Agilent Technologies, Inc., Santa Clara, CA, USA) at 37°C for 24 h. Next, digests were evaporated in a vacuum concentrator centrifuge and the residue was resuspended in 0.1% trifluoroacetic acid/5% acetonitrile. The digests were filtered through Amicon Ultra 0.5-ml 3K to remove undigested proteins and the flow-through was used in the following analyses.
LC-MS/MS analysis of protein identification
Approximately 2-μg peptide samples were injected into a peptide L-trap column (Chemicals Evaluation and Research Institute, Tokyo, Japan) using an HTC PAL autosampler (CTC Analytics, Zwingen, Switzerland) and further separated through an Advance-nano UHPLC using a Reverse-Phase C18-column (Zaplous column α, 3-μm diameter gel particles and 100 Å pore size, 0.1×150 mm; both from AMR, Inc., Tokyo, Japan). The mobile phase consisted of solution A (0.1% formic acid in water) and solution B (acetonitrile). The column was developed at a flow rate of 500 nl/min with a concentration gradient of acetonitrile from 5 to 45% B over 120 min. Gradient-eluted peptides were analyzed using an amaZon ETD ion-trap mass spectrometer (Bruker Daltonics, Billerica, MA, USA). The results were acquired in a data-dependent manner in which MS/MS fragmentation was performed on the 10 most intense peaks of every full MS scan.
All MS/MS spectra data were searched against the SwissProt Homo sapiens database using Mascot (v2.3.01; Matrix Science, London, UK). The search criteria was set as follows: enzyme, trypsin; allowance of up to two missed cleavage peptides; mass tolerance ± 0.5 Da and MS/MS tolerance ± 0.5 Da; and modifications of cysteine carbamidomethylation and methionine oxidation.
Semi-quantitative analysis of identified proteins
The fold-changes in expressed proteins on the base 2 logarithmic scale were calculated using Rsc based upon spectral counting (30). Relative amounts of identified proteins were also calculated using the normalized spectral abundance factor (NSAF) (31). Differentially expressed proteins were chosen so that their Rsc satisfy >1 or <−1, which correspond to fold-changes of >2 or <0.5.
Bioinformatics
Functional annotations for the identified proteins whose expression level was regulated by lumican were processed using the Database for Annotation, Visualization and Integrated Discovery (DAVID), v6.7 (http://david.abcc.ncifcrf.gov/home.jsp) (32–34).
Results
Protein identification and profile in lumican-regulated PANC-1 cells
To examine the effect of lumican on cell growth and invasion of PDAC cells, we created two types of PANC-1 cells whose lumican expression level was regulated: lumican-overexpressing PANC-1 cells and lumican-downregulated PANC-1 cells (28). We then investigated the molecular profile of proteins whose expression level was regulated by lumican using shotgun proteomics. Fig. 1 shows the Venn map for the identified proteins in lumican-regulated PANC-1 cells. In lumican-overexpressing PANC-1 cells (Lum), 321 proteins were identified, and 347 were identified in control cells (Mock) under the search parameter settings used (Fig. 1A). On the other hand, 388 proteins were identified in lumican-downregulated PANC-1 cells (siLum) and 287 in control cells (NC) (Fig. 1B). Among the 448 proteins identified from lumican upregulated cells, 220 (49.1%) proteins were identified in both cell lines, while 101 (22.6%) and 127 (28.3%) proteins were unique to Lum and Mock, respectively (Fig. 1A). Of the 451 total proteins identified from lumican downregulated cells, 224 (49.7%) proteins were identified in both cell lines, whereas 164 (36.3%) and 63 (14.0%) proteins were unique to siLum and NC, respectively (Fig. 1B).
Semi-quantitative comparison of identified proteins in lumican-regulated PANC-1 cells
Next, we performed a label-free semi-quantitative method based on spectral counting, as described in the Materials and methods section, to find proteins whose expression levels were regulated by lumican. The Rsc value was plotted against the corresponding protein (X-axis) from left to right for proteins identified in the Lum and Mock groups (Fig. 2A). The positive and negative Rsc values indicate increased and decreased expression, respectively, in the Lum group. The NSAF value (bar) plotted against the corresponding protein (X-axis), NSAF of Lum (black bar) and Mock (gray bar) proteins are indicated above and below the X-axis, respectively (Fig. 2A). Proteins with either a high positive or negative Rsc value were considered candidate proteins whose expression level was regulated by lumican. Fig. 2B shows the Rsc and NSAF values of the siLum and NC groups as described above. In the lumican upregulated cells, the Rsc of metallothionein (MT)-1X was positive, and the Rsc of galectin-1 was negative (Fig. 2A). In the lumican downregulated cells, the Rsc of Annexin A5 was positive and the Rsc of vinculin was negative (Fig. 2B). Housekeeping proteins such as β-actin, histone H4 and GAPDH were located near the center of the X-axis (Fig. 2).
As a result of semi-quantification, 174 differentially expressed proteins were identified in lumican upregulated PANC-1 cells (Table I), and a total of 143 differentially expressed proteins were identified in lumican downregulated PANC-1 cells (Table II). However, the expression level of housekeeping proteins such as β-actin, GAPDH and histone H4 was not changed in lumican upregulated cells or lumican downregulated cells. In order to identify proteins whose expression levels were regulated by lumican, we selected proteins whose Rsc value was inversely associated between lumican upregulated cells and lumican downregulated cells. Twenty four proteins were identified as candidates (Table III).
Functional annotation of proteins whose expression level is regulated by lumican
Gene ontology (GO) analyses were performed using the identified candidate proteins for each molecular function (Fig. 3), biological process (Fig. 4) and cellular component (Fig. 5) using DAVID. We also analyzed pathway terms, but no significant category was found. Functional annotations were counted by normalizing to the total number of proteins identified. Since a multifunctional protein yields more than one annotation and some proteins are not defined by GO terms yet, the total number of classified proteins resulted in more or less than 100% (Fig. 3). Major GO molecular function categories of the identified proteins were 34.8% structural molecule activity, 17.4% structural constituents of ribosomes, and 21.7% RNA binding proteins (Fig. 3).
Discussion
In the present study, we used a gel-free LC-MS-based proteomics approach to examine the effect of lumican on cell growth and invasion. Using semi-quantitative methods based on spectral counting, we successfully identified several proteins whose expression levels were altered more than 2-fold in both lumican upregulated PANC-1 cells and lumican downregulated PANC-1 cells. A limitation of spectral counting is in its accurate quantitative capacity (35). Therefore, we selected candidate proteins whose expression level was regulated by lumican using the analysis results of two types of cells; lumican upregulated cells and lumican downregulated cells. Thus, we identified 24 proteins whose Rsc values were inversely correlated among differentially expressed proteins in lumican upregulated and downregulated cells as candidate proteins (Fig. 3).
To examine the role of these candidate proteins in the effect of lumican on cell growth and invasion, we performed a functional classification of the candidate proteins by GO analysis. Although the GO terms for molecular function, biological processes, and cellular components were examined, we focused on molecular function. The molecular functions of candidate proteins were mainly classified in the ‘structural molecule activity’ category. Since structural molecule activity proteins contribute to the formation of complexes within or outside of the cell, such candidate proteins may be related to cell invasion regulated by lumican.
Annexin A5 expression may be regulated by lumican since Annexin A5 expression was negatively correlated with the lumican expression level. Annexin A5, also known as Annexin V, is widely known as a marker of early stage apoptosis (36,37). We previously demonstrated that the PDAC cell line secretes 70-kDa glycosylated lumican and that this secreted lumican stimulates cell growth (28). Thus, the induction of cell growth by lumican may be related to an inhibition of apoptosis through Annexin A5 expression. However, previous reports suggest that lumican plays an important role in apoptosis induction (22,38–40). This discrepancy may be derived from the differences between glycosylated lumican secreted from the PDAC cell line and other cells.
Furthermore, MT-1X expression levels positively correlated with the lumican expression level. MT family proteins are encoded by 10 functional isoforms (MT-1A, MT-1B, MT-1E, MT-1F, MT-1G, MT-1H, MT-1X, MT-2A, MT-3 and MT-4), and seven non-functional isoforms (MT-1C, MT-1D, MT-1I, MT-1J, MT-1K, MT-L and MT-2B). MT family proteins are involved in essential metal homeostasis, cellular free radical scavenging, cell proliferation, apoptosis, and metal detoxification. MT-1X and MT-2A transcripts were significantly upregulated under hypoxia in human prostate cancer cell lines, and siRNA-MT-2A treatment inhibited cell growth and induction of apoptosis, but an effect of MT-1X on cell growth and apoptosis was not demonstrated (41). Zinc is an abundant metal in the human prostate, and zinc inhibits cell growth and induces apoptosis in human prostate cancer cell lines (42,43). These findings suggest that MT-2A may play an important role in cell growth and apoptosis in prostate cancer through intracellular zinc homeostasis. MT-1X function in cancer cells, particularly PDAC, is not well understood. MT-1X is a known zinc-binding protein, and MT-1X mRNA expression was induced, as well as MT-2A, under hypoxic conditions (41). Thus, MT-1X may inhibit apoptosis as well as MT-2A. Furthermore, Ryu et al(44) suggested that MT-1E could enhance the migration and invasion of human glioma cells by inducing MMP-9 activation. MT-1X may have functions resembling MT-1E, since MMP-9 is a zinc-requiring enzyme, and MT-1X is classified into isoforms such as MT-1E. As mentioned above, it may be postulated that MT-1X plays an important role in cell growth and invasion by lumican. Further study is required to validate the expression levels of these candidate proteins, and to clarify the effect of candidate proteins, including MT-1X, on cell growth and invasion that are affected by lumican.
In conclusion, we identified more than 400 proteins from both lumican upregulated and lumican downregulated cells using global shotgun proteomics. A label-free semi-quantitative method based on spectral counting led to 24 candidate proteins whose expression was regulated by lumican. These candidate proteins included apoptosis-related and invasion-related proteins. Therefore, lumican may be involved in cell growth and invasion by altering the expression of these proteins.
Acknowledgements
The authors thank K. Teduka and T. Fujii for their technical assistance (Department of Pathology, Integrative Oncological Pathology). The present study was supported by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science to T.Y. (C, no. 24591019) and Z.N. (C, no. 23590477).
References
|
Nikitovic D, Katonis P, Tsatsakis A, Karamanos NK and Tzanakakis GN: Lumican, a small leucine-rich proteoglycan. IUBMB Life. 60:818–823. 2008. View Article : Google Scholar | |
|
Iozzo RV: Matrix proteoglycans: from molecular design to cellular function. Annu Rev Biochem. 67:609–652. 1998. View Article : Google Scholar : PubMed/NCBI | |
|
Grover J, Chen XN, Korenberg JR and Roughley PJ: The human lumican gene. Organization, chromosomal location, and expression in articular cartilage. J Biol Chem. 270:21942–21949. 1995. View Article : Google Scholar : PubMed/NCBI | |
|
Chakravarti S, Stallings RL, SundarRaj N, Cornuet PK and Hassell JR: Primary structure of human lumican (keratan sulfate proteoglycan) and localization of the gene (LUM) to chromosome 12q21.3–q22. Genomics. 27:481–488. 1995.PubMed/NCBI | |
|
Naito Z: Role of the small leucine-rich proteoglycan (SLRP) family in pathological lesions and cancer cell growth. J Nippon Med Sch. 72:137–145. 2005. View Article : Google Scholar : PubMed/NCBI | |
|
Rada JA, Cornuet PK and Hassell JR: Regulation of corneal collagen fibrillogenesis in vitro by corneal proteoglycan (lumican and decorin) core proteins. Exp Eye Res. 56:635–648. 1993. View Article : Google Scholar : PubMed/NCBI | |
|
Svensson L, Närlid I and Oldberg A: Fibromodulin and lumican bind to the same region on collagen type I fibrils. FEBS Lett. 470:178–182. 2000. View Article : Google Scholar : PubMed/NCBI | |
|
Chakravarti S: Functions of lumican and fibromodulin: lessons from knockout mice. Glycoconj J. 19:287–293. 2002. View Article : Google Scholar : PubMed/NCBI | |
|
Vogel KG, Paulsson M and Heinegård D: Specific inhibition of type I and type II collagen fibrillogenesis by the small proteoglycan of tendon. Biochem J. 223:587–597. 1984.PubMed/NCBI | |
|
Chakravarti S, Magnuson T, Lass JH, Jepsen KJ, LaMantia C and Carroll H: Lumican regulates collagen fibril assembly: skin fragility and corneal opacity in the absence of lumican. J Cell Biol. 141:1277–1286. 1998. View Article : Google Scholar : PubMed/NCBI | |
|
Jepsen KJ, Wu F, Peragallo JH, et al: A syndrome of joint laxity and impaired tendon integrity in lumican- and fibromodulin-deficient mice. J Biol Chem. 277:35532–35540. 2002. View Article : Google Scholar : PubMed/NCBI | |
|
Blochberger TC, Cornuet PK and Hassell JR: Isolation and partial characterization of lumican and decorin from adult chicken corneas. A keratan sulfate-containing isoform of decorin is developmentally regulated. J Biol Chem. 267:20613–20619. 1992. | |
|
Dolhnikoff M, Morin J, Roughley PJ and Ludwig MS: Expression of lumican in human lungs. Am J Respir Cell Mol Biol. 19:582–587. 1998. View Article : Google Scholar | |
|
Qin H, Ishiwata T and Asano G: Effects of the extracellular matrix on lumican expression in rat aortic smooth muscle cells in vitro. J Pathol. 195:604–608. 2001. View Article : Google Scholar : PubMed/NCBI | |
|
Baba H, Ishiwata T, Takashi E, Xu G and Asano G: Expression and localization of lumican in the ischemic and reperfused rat heart. Jpn Circ J. 65:445–450. 2001. View Article : Google Scholar : PubMed/NCBI | |
|
Onda M, Ishiwata T, Kawahara K, Wang R, Naito Z and Sugisaki Y: Expression of lumican in thickened intima and smooth muscle cells in human coronary atherosclerosis. Exp Mol Pathol. 72:142–149. 2002. View Article : Google Scholar : PubMed/NCBI | |
|
Ping Lu Y, Ishiwata T and Asano G: Lumican expression in alpha cells of islets in pancreas and pancreatic cancer cells. J Pathol. 196:324–330. 2002.PubMed/NCBI | |
|
Lu YP, Ishiwata T, Kawahara K, et al: Expression of lumican in human colorectal cancer cells. Pathol Int. 52:519–526. 2002. View Article : Google Scholar : PubMed/NCBI | |
|
Kelemen LE, Couch FJ, Ahmed S, et al: Genetic variation in stromal proteins decorin and lumican with breast cancer: investigations in two case-control studies. Breast Cancer Res. 10:R982008. View Article : Google Scholar : PubMed/NCBI | |
|
Troup S, Njue C, Kliewer EV, et al: Reduced expression of the small leucine-rich proteoglycans, lumican, and decorin is associated with poor outcome in node-negative invasive breast cancer. Clin Cancer Res. 9:207–214. 2003.PubMed/NCBI | |
|
Ishiwata T, Cho K, Kawahara K, et al: Role of lumican in cancer cells and adjacent stromal tissues in human pancreatic cancer. Oncol Rep. 18:537–543. 2007.PubMed/NCBI | |
|
Vuillermoz B, Khoruzhenko A, D’Onofrio MF, et al: The small leucine-rich proteoglycan lumican inhibits melanoma progression. Exp Cell Res. 296:294–306. 2004. View Article : Google Scholar : PubMed/NCBI | |
|
Shinji S, Tajiri T, Ishiwata T, Seya T, Tanaka N and Naito Z: Different expression levels of lumican in human carcinoid tumor and neuroendocrine cell carcinoma. Int J Oncol. 26:873–880. 2005.PubMed/NCBI | |
|
Seya T, Tanaka N, Shinji S, et al: Lumican expression in advanced colorectal cancer with nodal metastasis correlates with poor prognosis. Oncol Rep. 16:1225–1230. 2006.PubMed/NCBI | |
|
Matsuda Y, Yamamoto T, Kudo M, et al: Expression and roles of lumican in lung adenocarcinoma and squamous cell carcinoma. Int J Oncol. 33:1177–1185. 2008.PubMed/NCBI | |
|
Leygue E, Snell L, Dotzlaw H, et al: Expression of lumican in human breast carcinoma. Cancer Res. 58:1348–1352. 1998. | |
|
Naito Z, Ishiwata T, Kurban G, et al: Expression and accumulation of lumican protein in uterine cervical cancer cells at the periphery of cancer nests. Int J Oncol. 20:943–948. 2002.PubMed/NCBI | |
|
Yamamoto T, Matsuda Y, Kawahara K, Ishiwata T and Naito Z: Secreted 70kDa lumican stimulates growth and inhibits invasion of human pancreatic cancer. Cancer Lett. 320:31–39. 2012. View Article : Google Scholar : PubMed/NCBI | |
|
Bluemlein K and Ralser M: Monitoring protein expression in whole-cell extracts by targeted label- and standard-free LC-MS/MS. Nat Protoc. 6:859–869. 2011. View Article : Google Scholar : PubMed/NCBI | |
|
Old WM, Meyer-Arendt K, Aveline-Wolf L, et al: Comparison of label-free methods for quantifying human proteins by shotgun proteomics. Mol Cell Proteomics. 4:1487–1502. 2005. View Article : Google Scholar : PubMed/NCBI | |
|
Zybailov B, Coleman MK, Florens L and Washburn MP: Correlation of relative abundance ratios derived from peptide ion chromatograms and spectrum counting for quantitative proteomic analysis using stable isotope labeling. Analytical chemistry. 77:6218–6224. 2005. View Article : Google Scholar | |
|
Dennis G Jr, Sherman BT, Hosack DA, et al: DAVID: Database for Annotation, Visualization, and Integrated Discovery. Genome Biol. 4:P32003. View Article : Google Scholar : PubMed/NCBI | |
|
Huang da W, Sherman BT and Lempicki RA: Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc. 4:44–57. 2009.PubMed/NCBI | |
|
Huang da W, Sherman BT and Lempicki RA: Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res. 37:1–13. 2009.PubMed/NCBI | |
|
Lundgren DH, Hwang SI, Wu L and Han DK: Role of spectral counting in quantitative proteomics. Expert Rev Proteomics. 7:39–53. 2010. View Article : Google Scholar : PubMed/NCBI | |
|
Vermes I, Haanen C, Steffens-Nakken H and Reutelingsperger C: A novel assay for apoptosis. Flow cytometric detection of phosphatidylserine expression on early apoptotic cells using fluorescein labelled Annexin V. J Immunol Methods. 184:39–51. 1995. View Article : Google Scholar | |
|
Aubry JP, Blaecke A, Lecoanet-Henchoz S, et al: Annexin V used for measuring apoptosis in the early events of cellular cytotoxicity. Cytometry. 37:197–204. 1999. View Article : Google Scholar : PubMed/NCBI | |
|
Vij N, Roberts L, Joyce S and Chakravarti S: Lumican suppresses cell proliferation and aids Fas-Fas ligand mediated apoptosis: implications in the cornea. Exp Eye Res. 78:957–971. 2004. View Article : Google Scholar : PubMed/NCBI | |
|
Brezillon S, Venteo L, Ramont L, et al: Expression of lumican, a small leucine-rich proteoglycan with antitumour activity, in human malignant melanoma. Clin Exp Dermatol. 32:405–416. 2007. View Article : Google Scholar : PubMed/NCBI | |
|
Williams KE, Fulford LA and Albig AR: Lumican reduces tumor growth via induction of fas-mediated endothelial cell apoptosis. Cancer Microenviron. 4:115–126. 2010. View Article : Google Scholar : PubMed/NCBI | |
|
Yamasaki M, Nomura T, Sato F and Mimata H: Metallothionein is up-regulated under hypoxia and promotes the survival of human prostate cancer cells. Oncol Rep. 18:1145–1153. 2007.PubMed/NCBI | |
|
Liang JY, Liu YY, Zou J, Franklin RB, Costello LC and Feng P: Inhibitory effect of zinc on human prostatic carcinoma cell growth. Prostate. 40:200–207. 1999. View Article : Google Scholar : PubMed/NCBI | |
|
Feng P, Li TL, Guan ZX, Franklin RB and Costello LC: Direct effect of zinc on mitochondrial apoptogenesis in prostate cells. Prostate. 52:311–318. 2002. View Article : Google Scholar : PubMed/NCBI | |
|
Ryu HH, Jung S, Jung TY, et al: Role of metallothionein 1E in the migration and invasion of human glioma cell lines. Int J Oncol. 41:1305–1313. 2012.PubMed/NCBI |
