MSH2 and CXCR4 involvement in malignant VIPoma
© Muller et al.; licensee BioMed Central Ltd. 2012
Received: 6 August 2012
Accepted: 24 November 2012
Published: 11 December 2012
Vasoactive intestinal polypeptide secreting tumors(VIPomas) are rare endocrine tumors of the pancreas with an estimated incidence of 0.1 per million per year. The molecular mechanisms that mediate development of VIPomas are poorly investigated and require definition.
A genome- and gene expression analysis of specimens of a primary pancreatic VIPoma with hepatic metastases was performed. The primary tumor, the metastases, the corresponding healthy tissue of the liver, and the pancreas were compared with each other using oligonucleotide microarrays and loss of heterozygosity (LOH).
The results revealed multiple LOH events and several differentially expressed genes. Our finding of LOH and downregulation was conspicuous in the microarray analysis for the mismatch repair gene MSH2 in the primary pancreatic VIPoma tumor, the hepatic metastasis but not in the corresponding healthy tissue. Further a strong overexpression of the chemokine CXCR4 was detected in the hepatic metastases compared to its pancreatic primary. With a review of the literature we describe the molecular insights of metastatic development in VIPoma.
In VIPoma, defects in the mismatch repair system especially in MSH2 may contribute to carcinogenesis, and increased CXCR4 may be associated with liver metastasis.
KeywordsVIPoma MSH2 CXCR4 Microarrays LOH
Vasoactive intestinal polypeptide secreting tumor (VIPoma) is a rare tumor associated with watery diarrhea, hypokalemia, and achlorhydria (WDHA) . The estimated incidence is approximately 0.1 per million per year . VIPomas are mostly located in the pancreas, although they have been already detected in the bronchial system, colon, liver, adrenal gland, and sympathetic ganglia [3, 4]. The major causes of death concerning VIPomas are dehydration and renal failure . Clinical symptoms lead to the diagnosis. When symptoms present several patients already have metastasis. The only curative approach is surgical resection . Until now the molecular mechanisms that mediate development of metastases of VIPomas require definition and were not investigated yet. Thus, we present this rare entity by investigating a genome- and gene expression analysis of specimens of a primary pancreatic VIPoma with hepatic metastases.
Patients and tissue specimens
Histological proven tissue samples (primary pancreatic VIPoma with hepatic metastases and corresponding healthy tissue of the liver and the pancreas) of this patient were investigated. All tumor specimens were collected by the Department of General, Visceral and Transplant Surgery, University of Tübingen, Germany. Specimens were snap frozen in liquid nitrogen and stored at −80°C until use. Each tumor sample was cryosected, stained with hematoxylin and eosin, classified by two experience pathologists, and re-evaluated by an experienced surgical pathologist. To ensure a tumor content >80%, microdissection was performed regarding cancer specimens. The patients provided signed informed consent. The study was approved by the local Ethics Committee.
RNA was extracted using the NucleoSpin RNA II Kit (Macherey-Nagel, Dueren, Germany). The RNA quality and quantity were verified using the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) and the NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA).
Microarray data generation and statistical analysis
Microarray analysis was performed using oligonucleotide microarrays (65mer) produced at the Max Planck Institute, Tübingen, Germany. The arrays contained oligonucleotides for about 900 transcripts, and each oligonucleotide was printed twice. Using 5 oligonucleotide microarrays (including dye-swap) we compared the primary tumor, the metastases, the corresponding healthy tissue of the liver, and the pancreas. Amplification of the sample RNA was performed using Ambion’s Amino AllylMessageAmp™ II aRNA Amplification Kit (Ambion Inc., Austin, TX, USA). Dye-coupling reaction was performed using Amersham CyDye Post-labelling Reactive Dye Pack (GE Healthcare, Buckinghamshire, UK). After an aRNA fragmentation using Ambion’s Fragmentation Reagents (Ambion Inc., Austin, TX, USA), hybridization was carried out at 48°C for 14 h. The slides were scanned in a microarray scanner (Genetix Limited, Hampshire, UK). The photomultiplier tube voltage was set to 100% for both red and green channels. The two resulting green and red images were overlaid using ImaGene 5 (BioDiscovery, Inc., El Segundo, CA, USA). Raw data collection was performed using ImaGene v.5.0. Further statistical and bioinformatic analyses were performed using the R language (http://www.r-project.org) and the ‘limma’ (Linear Models for Microarray) package from the Bioconductor project (http://www.bioconductor.org/). As a first step in signal extraction, for each channel we used the mean of the pixel distribution for the foreground signal and the median for the background of each spot as estimators of the raw signal values. All spots were used, regardless of their flag status. The data were normalized using loess normalization on the normexp-background corrected expression values, followed by a dye-swap normalization and in-between-array quantile normalization. Both the loess and quantile normalization methods were used as provided in the limma package. On the basis of the M values computed, differentially expressed genes were detected by applying the Welch one-sample t-test as implemented in R.
M32: primary tumor
M32I: liver metastasis 1
M32II: liver metastasis 2
M32N: normal liver tissue
Data derived from these samples were compared to results derived from blood DNA.
Microsatellite analysis was performed applying an extended marker panel for endocrine tumors which was previously established in our department . In brief, markers BAT25, BAT26, D2S123, and D17S250 were amplified in a multiplex PCR (multiplex 1) using the QIAGEN Multiplex PCR Master Mix (Qiagen) under following conditions: final concentration of each forward primer (dye labeled) was 0.2 μM, of each reverse primer 1 μM, DNA concentration was 100 ng. Asymmetric PCR was used to reduce the ratio between primer signal and amplicon signal thereby improving the fluorescence detection threshold. Thermal cycling conditions were 94°C for 5 min (94°C for 30 s, 54°C for 90 s, 72°C for 60 s), 30 cycles, 60°C for 5 min. PCR products were purified using the QIAquick PCR purification kit (QIAGEN).
Microsatellite markers D2S443, D16S752, D21S1436 (multiplex 2) and D1S104, D3S1284 (multiplex 3) were amplified in two multiplex PCR reactions under following conditions: 200 ng DNA, 1 × PCR buffer, 0.08 mM dNTPs each, 2 mM MgCl2, 1 U Taq, 0.8 μM primer each, ad 50 μL H2O; 94°C for 2 min (94°C for 40 s, 53°C (multiplex 2), and 61°C (multiplex 3) for 40 s, respectively, 72°C for 1.30 min), 40 cycles, 72°C for 2 min. PCR products derived from the same patients sample were combined and purified using the PCR-purification kit (Qiagen). LOH analysis was performed on an automated capillary sequencer CEQ 8000 (Beckman Coulter) and evaluated with the CEQ 8000 Fragment analysis software (Beckman Coulter).
The tumor was considered to be positive for LOH if the allele peak ratio was ≤0.7, in analogy with allelic signal reduction of at least 30% . Homozygous peaks were classified as not informative and were not evaluated in LOH statistics.
A Medline search was conducted for the term ‘VIPoma’to December 2011. A total of 489 publications were retrieved. All English written publications reporting on oncogenesis were included. Articles dealing with clinical findings, diagnostics, therapy options, and clinical outcomes were excluded.
Gene expression analysis was performed using a custom-designed oligo microarray. We found a number of genes that were up- or downregulated comparing the primary tumor, the metastases, the corresponding healthy tissue of the liver, and the pancreas.
Comparison of gene up- and downregulation of the primary pancreatic tumor against its corresponding healthy pancreatic tissue (>log2)
Homo sapiens SMAD family member 5 (SMAD5)
Homo sapiens catenin (CTNNAL1)
Homo sapiens neural cell adhesion molecule 2 (NCAM2)
Homo sapiens adhesion regulating molecule 1 (ADRM1)
Homo sapiens cadherin-like 26 (CDH26)
Homo sapiens histonedeacetylase 2 (HDAC2)
Homo sapiens S100 calcium binding protein A9 (S100A9)
Homo sapiens cyclin L2 (CCNL2)
Homo sapiens mediator of DNA damage checkpoint 1 (MDC1)
Homo sapiens MSH2 (MSH2)
Homo sapiens adhesion molecule, (AMICA1)
Homo sapiens SMAD family member 6 (SMAD6)
Homo sapiens guaninenucleotide binding protein (G protein)
Homo sapiens cyclin-dependent kinase 5, (CDK5R2)
Comparison of gene up- and downregulation of the primary pancreatic tumor against the hepatic metastases (>log2)
Homo sapiens cyclin-dependent kinase 7 (CDK7)
Homo sapiens acyl-CoA synthetase, (ACSM1)
Homo sapiens cadherin 4 (CDH4)
Homo sapiens cadherin 20, type 2 (CDH20)
Homo sapiens cyclin D2 (CCND2)
Homo sapiens HUS1 checkpoint (HUS1B)
Homo sapiens gapjunctionprotein, beta 5 (connexin 31.1) (GJB5)
Homo sapiens chemokine (C-X-C motif) receptor 4 (CXCR4)
Homo sapiens lipase, hormone-sensitive
Homo sapiens caudal type homeobox ranscription factor 2 (CDX2)
Comparison of gene up- and downregulation of the hepatic metastases and its corresponding healthy liver tissue (>log2)
Homo sapiens neuronal cell adhesion molecule (NRCAM)
Homo sapiens guanine nucleotide binding protein-like 2 (GNL2)
Homo sapiens GNAS complex locus (GNAS)
Homo sapiens alanine-glyoxylate aminotransferase homolog (TLH6)
Homo sapiens cell division cycle and apoptosis regulator 1 (CCAR1)
Homo sapiens MSH2 (MSH2)
Homo sapiens cyclin-dependent kinase inhibitor (CDKN1C)
Homo sapiens gap junction protein, beta 1 (GJB1)
Homo sapiens cytochrome c oxidase subunit VIIa polypeptide 2 (liver)
Homo sapiens caspase 1 (CASP1)
Homo sapiens mucin and cadherin-like (MUCDHL)
Homo sapiens cyclin D1 (CCND1)
Regarding the gene expression of MSH2 and CXCR4 we were able to detect a downregulation (−2.80-fold) of MSH2 in hepatic metastases and (−2.63-fold) of MSH2 in the primary pancreatic tumor compared to both corresponding healthy tissue. Comparing the hepatic metastases against the primary pancreatic tumor an upregulation (10.41-fold) was detected.
Microsatellite markers used for MSI and LOH detection in pancreatic tumor tissue, liver metastasis tissue, and normal liver tissue
Sixty-three percent of markers analyzed showed LOH. Nine percent of markers were not informative for LOH calculation and 27% of markers showed microsatellite stability. LOH was detected in primary tumor DNA as well as in liver metastasis. Interestingly, two markers (D5S326 and D21S1436) revealed LOH in normal liver tissue. All tumor DNAs analyzed showed LOH for marker D2S123 which is located close to the MSH2.
VIP-producing tumors arise from pancreatic islet cells and are mostly located in the pancreatic body or tail . Malignancy is solely defined by existing distant metastasis . The clinical symptoms accompanying the tumor are watery diarrhea, hypokalemia, and metabolic acidosis. At the time of presentation, over 70% of patients had hepatic metastases . Hence, understanding the molecular mechanisms of tumor progression is of paramount importance. Our results revealed defects in the mismatch repair system especially in MSH2. MSH2 showed significantly lower expression in the primary pancreatic tumor and its liver metastases than in healthy pancreatic tissue. Further an overexpression of CXCR4 in the hepatic metastasis compared with the primary was detected.
Reviewing the literature,eight cohorts including 161 patients and 133 case reports exist on VIPoma. Only six publications deal with the molecular mechanisms of oncogenesis and metastasis development in VIPoma, reporting no more than two cases in each publication [13–18]. The remaining publication reports on the clinical presentation, diagnostics, and therapeutic options. Amongst others, mutations in the tumor suppression genes DPC4 , AIM1 (‘absent in melanoma1’) and PTPRK (receptor type protein-tyrosine phophastase kappa)  are shown in metastatic VIPomas while mutations of the proto oncogene BRAF known to play an important role in thyroid cancer and melanoma were not found [15, 18]. A series of 35 neuroendocrine tumors of the pancreas showed that defects in DNA mismatch repair are rather rare with the limitation, that no VIPoma case was included .
The DNA mismatch repair system (MMR) corrects base mismatches after DNA replication, avoids recombination between non-identical DNA sequences, and induces apoptotic and checkpoint responses after DNA damage . Besidespost-replication repair, MMR proteins are assumed to a great extent to be involved in carcinogenesis . So far hundreds of different predisposing mutations are known, mainly affecting the MMR genes, MLH1, MSH2, and MSH6. Predisposed individuals exhibit a deficient copy of a MMR gene in each cell. Deficient MMR, characteristic in the colon can lead to proceeding accumulation of mutations and cancer, such as hereditary non-polyposis colon cancer (HNPCC), is known to be a syndrome of insufficient DNA mismatch repair genes . Instability at short tandem repeat sequences, microsatellites, is a representative manifestation of MMR deficiency and beside HNPCC also occurs in other tumors . In particular deficient MSH2 is associated with an increased risk of cancer .
Chemokines and their respective receptors have been identified as contributing metastatic factors in numerous cancers [23, 24]. The common human chemokine system includes about 50 ligands and 20 G protein-coupled receptors which control migration and activation of leukocytes and influences angiogenesis and tumor growth . Recent discoveries assume that tumor cells themselves are able to secrete chemokines [25–27]. Therefore disseminated tumor cells that express members of the CXCR family and invade the circulation may be attracted and arrested by their corresponding ligand. Hence these cells gain the ability to infiltrate in distinct organs. CXCR4 is assumed to be involved in metastasis of non-small-cell lung, breast, pancreatic, prostate, gastric cancer, and peritoneal carcinomatosis [27–32]. Deschamps et al. found CXCR4 to be associated with the malignant metastatic progression of neuroendocrine tumors of the ileum correlated with a lower survival . So far, mutation of MSH2 and CXCR4 and its involvement in tumor growth and metastasis development have not been reported in VIPoma.
Altogether, the altered expression of MSH2 and CXCR4 in a metastatic VIPoma might reflect higher tumor aggressivity with the potential development of distant metastases. Owing to the rarity of this disease, we believe these results will provide a valuable resource for future work on this serious condition with the potential of diagnostic applications.
The authors gratefully acknowledge the effort and contributions to the present study, made by the technologists and staff, such as Melanie Hauth, Juergen Winter, Silvia Wagner, Christine Kienzle, and Caroline Zug for their help in sampling the tissues as well as Georg Otto and Robert Geisler, members of the Max Planck Institute in Tübingen, Germany, for their encouragement concerning microarray manufacture.
- Verner JV, Morrison AB: Islet cell tumor and a syndrome of refractory watery diarrhea and hypokalemia. Am J Med. 1958, 25: 374-380. 10.1016/0002-9343(58)90075-5.View ArticlePubMed
- Friesen SR: Update on the diagnosis and treatment of rare neuroendocrine tumors. Surg Clin North Am. 1987, 67: 379-393.PubMed
- Ectors N: Pancreatic endocrine tumors: diagnostic pitfalls. Hepatogastroenterology. 1999, 46: 679-690.PubMed
- Ghaferi AA, Chojnacki KA, Long WD, Cameron JL, Yeo CJ: Pancreatic VIPomas: subject review and one institutional experience. J Gastrointest Surg. 2008, 12: 382-393. 10.1007/s11605-007-0177-0.View ArticlePubMed
- Peng SY, Li JT, Liu YB, Fang HQ, Wu YL, Peng CH, Wang XB, Qian HR: Diagnosis and treatment of VIPoma in China: (case report and 31 cases review) diagnosis and treatment of VIPoma. Pancreas. 2004, 28: 93-97. 10.1097/00006676-200401000-00015.View ArticlePubMed
- Grant CS: Surgical management of malignant islet cell tumors. World J Surg. 1993, 17: 498-503. 10.1007/BF01655109.View ArticlePubMed
- Zieker D, Zieker J, Dietzsch J, Burnet M, Northoff H, Fehrenbach E: CDNA-microarray analysis as a research tool for expression profiling in human peripheral blood following exercise. Exerc Immunol Rev. 2005, 11: 1186-1196.
- Zieker D, Fehrenbach E, Dietzsch J, Fliegner J, Waidmann M, Nieselt K, Gebicke-Haerter P, Spanagel R, Simon P, Niess AM, Northoff H: cDNA microarray analysis reveals novel candidate genes expressed in human peripheral blood following exhaustive exercise. Physiol Genomics. 2005, 23: 287-294. 10.1152/physiolgenomics.00096.2005.View ArticlePubMed
- Kupka S, Haack B, Zdichavsky M, Mlinar T, Kienzle C, Bock T, Kandolf R, Kroeber SM, Konigsrainer A: Large proportion of low frequency microsatellite-instability and loss of heterozygosity in pheochromocytoma and endocrine tumors detected with an extended marker panel. J Cancer Res Clin Oncol. 2008, 134: 463-471. 10.1007/s00432-007-0307-9.View ArticlePubMed
- Karnik P, Paris M, Williams BR, Casey G, Crowe J, Chen P: Two distinct tumor suppressor loci within chromosome 11p15 implicated in breast cancer progression and metastasis. Hum Mol Genet. 1998, 7: 895-903. 10.1093/hmg/7.5.895.View ArticlePubMed
- Capella C, Polak JM, Buffa R, Tapia FJ, Heitz P, Usellini L, Bloom SR, Solcia E: Morphologic patterns and diagnostic criteria of VIP-producing endocrine tumors. A histologic, histochemical, ultrastructural, and biochemical study of 32 cases. Cancer. 1983, 52: 1860-1874. 10.1002/1097-0142(19831115)52:10<1860::AID-CNCR2820521017>3.0.CO;2-F.View ArticlePubMed
- Jensen RT: Pancreatic endocrine tumors: recent advances. Ann Oncol. 1999, Suppl: 4170-4176.
- Barghorn A, Komminoth P, Bachmann D, Rutimann K, Saremaslani P, Muletta-Feurer S, Perren A, Roth J, Heitz PU, Speel EJ: Deletion at 3p25.3–p23 is frequently encountered in endocrine pancreatic tumours and is associated with metastatic progression. J Pathol. 2001, 194: 451-458. 10.1002/path.886.View ArticlePubMed
- Barghorn A, Speel EJ, Farspour B, Saremaslani P, Schmid S, Perren A, Roth J, Heitz PU, Komminoth P: Putative tumor suppressor loci at 6q22 and 6q23-q24 are involved in the malignant progression of sporadic endocrine pancreatic tumors. Am J Pathol. 2001, 158: 1903-1911. 10.1016/S0002-9440(10)64658-5.PubMed CentralView ArticlePubMed
- Bartsch D, Hahn SA, Danichevski KD, Ramaswamy A, Bastian D, Galehdari H, Barth P, Schmiegel W, Simon B, Rothmund M: Mutations of the DPC4/Smad4 gene in neuroendocrine pancreatic tumors. Oncogene. 1999, 18: 2367-2371. 10.1038/sj.onc.1202585.View ArticlePubMed
- Missiaglia E, Moore PS, Williamson J, Lemoine NR, Falconi M, Zamboni G, Scarpa A: Sex chromosome anomalies in pancreatic endocrine tumors. Int J Cancer. 2002, 98: 532-538. 10.1002/ijc.10223.View ArticlePubMed
- Missiaglia E, Dalai I, Barbi S, Beghelli S, Falconi M, della Peruta M, Piemonti L, Capurso G, Di Florio A, delle Fave G, Pederzoli P, Croce CM, Scarpa A: Pancreatic endocrine tumors: expression profiling evidences a role for AKT-mTOR pathway. J Clin Oncol. 2010, 28: 245-255. 10.1200/JCO.2008.21.5988.PubMed CentralView ArticlePubMed
- Tannapfel A, Vomschloss S, Karhoff D, Markwarth A, Hengge UR, Wittekind C, Arnold R, Horsch D: BRAF gene mutations are rare events in gastroenteropancreatic neuroendocrine tumors. Am J Clin Pathol. 2005, 123: 256-260. 10.1309/YQBR9C05RU4DD3RN.View ArticlePubMed
- Arnason T, Sapp HL, Rayson D, Barnes PJ, Drewniak M, Nassar BA, Huang WY: Loss of expression of DNA mismatch repair proteins is rare in pancreatic and small intestinal neuroendocrine tumors. Arch Pathol Lab Med. 2011, 135: 1539-1544. 10.5858/arpa.2010-0560-OA.View ArticlePubMed
- O’Brien V, Brown R: Signalling cell cycle arrest and cell death through the MMR System. Carcinogenesis. 2006, 27: 682-692.View ArticlePubMed
- Peltomaki P: Deficient DNA, mismatch repair: a common etiologic factor for colon cancer. Hum Mol Genet. 2001, 10: 735-740. 10.1093/hmg/10.7.735.View ArticlePubMed
- Fishel R: Mismatch repair, molecular switches, and signal transduction. Genes Dev. 1998, 12: 2096-2101. 10.1101/gad.12.14.2096.View ArticlePubMed
- Kim J, Takeuchi H, Lam ST, Turner RR, Wang HJ, Kuo C, Foshaq L, Bilchik AJ, Hoon DS: Chemokine receptor CXCR4 expression in colorectal cancer patients increases the risk for recurrence and for poor survival. J Clin Oncol. 2005, 23: 2744-2753.View ArticlePubMed
- Kim J, Mori T, Chen SL, Amersi FF, Martinez SR, Kuo C, Turner RR, Ye X, Bilchik AJ, Morton DL, Hoon DS: Chemokine receptor CXCR4 expression in patients with melanoma and colorectal cancer liver metastases and the association with disease outcome. Ann Surg. 2006, 244: 113-120. 10.1097/01.sla.0000217690.65909.9c.PubMed CentralView ArticlePubMed
- Gerard C, Rollins BJ: Chemokines and disease. Nat Immunol. 2001, 2: 108-115. 10.1038/84209.View ArticlePubMed
- Bottazzi B, Polentarutti N, Acero R, Balsari A, Boraschi D, Ghezzi P, Salmona M, Mantovani A: Regulation of the macrophage content of neoplasms by chemoattractants. Science. 1983, 220: 210-212. 10.1126/science.6828888.View ArticlePubMed
- Muller A, Homey B, Soto H, Ge N, Catron D, Buchanan ME, McClanahan T, Murphy E, Yuan W, Wagner SN, Barrera JL, Mohar A, Verastequi E, Zlotnik A: Involvement of chemokine receptors in breast cancer metastasis. Nature. 2001, 410: 50-56. 10.1038/35065016.View ArticlePubMed
- Oonakahara K, Matsuyama W, Higashimoto I, Kawabata M, Arimura K, Osame M: Stromal-derived factor-1alpha/CXCL12-CXCR 4 axis is involved in the dissemination of NSCLC cells into pleural space. Am J Respir Cell Mol Biol. 2004, 30: 671-677. 10.1165/rcmb.2003-0340OC.View ArticlePubMed
- Saur D, Seidler B, Schneider G, Algul H, Beck R, Senekowitsch-Schmidtke R, Schwaiger M, Schmid RM: CXCR4 expression increases liver and lung metastasis in a mouse model of pancreatic cancer. Gastroenterology. 2005, 129: 1237-1250. 10.1053/j.gastro.2005.06.056.View ArticlePubMed
- Taichman RS, Cooper C, Keller ET, Pienta KJ, Taichman NS, McCauley LK: Use of the stromal cell-derived factor-1/CXCR4 pathway in prostate cancer metastasis to bone. Cancer Res. 2002, 62: 1832-1837.PubMed
- Wang J, Wang J, Dai J, Jung Y, Wei CL, Wang Y, Havens AM, Hogg PJ, Keller ET, Pienta KJ, Nor JE, Wang CY, Taichman RS: A glycolytic mechanism regulating an angiogenic switch in prostate cancer. Cancer Res. 2007, 67: 149-159. 10.1158/0008-5472.CAN-06-2971.View ArticlePubMed
- Yasumoto K, Koizumi K, Kawashima A, Saitoh Y, Arita Y, Shinohara K, Minami T, Nakayama T, Sakurai H, Takahashi Y, Yoshie O, Saiki I: Role of the CXCL12/CXCR4 axis in peritoneal carcinomatosis of gastric cancer. Cancer Res. 2006, 66: 2181-2187. 10.1158/0008-5472.CAN-05-3393.View ArticlePubMed
- Deschamps L, Bacha D, Rebours V, Mebarki M, Bretagnol F, Panis Y, Bedossa P, Ruszniewski P, Coulevard A: The expression of the hypoxia markers CA9 and CXCR4 is correlated with survival in patients with neuroendocrine tumours of the ileum. Neuroendocrinology. 2012, 95: 214-222. 10.1159/000329873.View ArticlePubMed
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.