- Open Access
[18F]Fluorodeoxyglucose accumulation as a biological marker of hypoxic status but not glucose transport ability in gastric cancer
- Ryusuke Takebayashi†1,
- Kunihiko Izuishi†1, 4Email author,
- Yuka Yamamoto2,
- Reiko Kameyama2,
- Hirohito Mori3,
- Tsutomu Masaki3 and
- Yasuyuki Suzuki1
© Takebayashi et al.; licensee BioMed Central Ltd. 2013
- Received: 14 February 2013
- Accepted: 22 May 2013
- Published: 29 May 2013
The use of [18F] 2-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET) for detection of gastric cancer is often debated because FDG uptake varies for each patient. The purpose of this study was to clarify the molecular mechanisms involved in FDG uptake.
Material and methods
Fifty patients with gastric cancer who underwent FDG-PET and gastrectomy were studied. Snap-frozen tumor specimens were collected and examined by real-time PCR for relationships between maximum standardized uptake value (SUV) and mRNA expression of the following genes: glucose transporter 1 (GLUT1), hexokinase 2 (HK2), hypoxia-inducible factor 1α (HIF1α), and proliferating cell nuclear antigen (PCNA).
Tumor size was the only clinicopathological parameter that significantly correlated with SUV. Transcripts for the genes evaluated were about three-fold higher in malignant specimens than in normal mucosa, although only HIF1α was significantly correlated with SUV. When divided into intestinal and non-intestinal tumors, there was a significant correlation between SUV and tumor size in intestinal tumors. Interestingly, the weak association between SUV and HIF1α expression in intestinal tumors was substantially stronger in non-intestinal tumors. No correlation was found between SUV and mRNA expression of other genes in intestinal or non-intestinal tumors.
SUV was correlated with HIF1α, but not PCNA, HK2, or GLUT1 expression. FDG accumulation could therefore represent tissue hypoxia rather than glucose transport activity for aggressive cancer growth.
- Positron emission tomography
- Gastric cancer
- Glucose transporter-1
- Hypoxia-inducible factor 1α
Radiology examinations provide important information for cancer treatment, and [18F] 2-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET) differs from conventional imaging through its use of cellular metabolic characteristics to detect a variety of tumors and metastases [1, 2]. FDG-PET detection rates tended to vary widely for gastric cancer, however, with 0–44% detection in early stages and 34–94% detection in advanced stages [1, 3–5]. Pseudolesions from physiological FDG uptake prevent a more precise diagnosis . Moreover, signet ring cell carcinoma was reported to significantly lower the standardized uptake value (SUV) of FDG compared to papillary or tubular adenocarcinomas [1, 7, 8]. The usefulness of FDG-PET detection for gastric cancer is thus a matter of debate.
Besides detecting tumors based on absolute value, FDG-PET can also assess the response to chemotherapy based on relative values before and after cancer treatment . Previous studies have suggested a significant association between the metabolic changes observed by FDG-PET and clinical or histopathological response [9–11]. One report in particular predicted patient prognoses by detecting early changes in glucose uptake after chemotherapy, which could help prevent the continuation of ineffective treatments. Ott et al. found that a reduction in FDG uptake of more than 35% for metabolic responders predicted a favorable response in gastric cancer patients two weeks after initiation of chemotherapy , while metabolic non-responders or FDG non-avid tumors received an unfavorable prognosis.
Cancer cells theoretically require a greater amount of glucose consumption than healthy tissue because of increased cell division [12, 13] or anaerobic respiration in tumors . Many cancers increase glucose transport through glucose transporter 1 (GLUT1) and glucose phosphorylation by hexokinase (HK) [15–17]. A correlation between FDG uptake and GLUT1 expression has been found in gastric cancer patients [1, 3, 7, 8], but these studies were conducted by non-quantitative immunohistochemistry analysis, such as negative or positive staining that can vary by evaluator. We therefore evaluated the expression of glucose metabolism-related proteins through quantitative reverse-transcription polymerase chain reaction (qRT-PCR) and compared the results to maximum SUV of FDG-PET. In addition, we also analyzed the expression of proliferating cell nuclear antigen (PCNA) as a valid marker of proliferation  and hypoxia-inducible factor 1 alpha (HIF1α) as a marker of hypoxia  to elucidate either of these mechanisms, i.e., tumor proliferation or tumor hypoxia, contribute to FDG uptake. We then discuss the significance and difficulties involved with the clinical application of FDG-PET in gastric cancer due to FDG uptake mechanisms.
This retrospective study involved 50 patients (29 male and 21 female; mean age ± standard error of measurement [SEM], 65.8 ± 1.4 years) with gastric cancer who underwent same FDG-PET system before gastrectomy in Kagawa University from July 2005 to March 2010. Tumor specimens were snap-frozen at the time of surgery, and stored at −80°C. Participants were divided into 25 cases of intestinal tumors and 25 cases of non-intestinal tumors based on histopathological diagnoses. When focal FDG uptake was not found in the stomach, SUV was calculated from a lesion determined by histology results after gastrectomy. The International Union Against Cancer staging system was used to determine clinicopathological parameters associated with FDG uptake. The protocol was approved by the institutional review board of our institution, and all patients provided written informed consent.
FDG-PET images were acquired with a PET scanner (ECAT EXACT HR+, Siemens/CTI, Knoxville, TN, USA). Patients fasted at least five hours before FDG injection. Images were reviewed on a Sun Microsystems workstation (Siemens/CTI) along transverse, coronal, and sagittal planes with maximum intensity projection images. The images were then interpreted independently by two experienced nuclear medicine physicians blinded to the clinical data. Tumor lesions were identified as areas of focally increased FDG uptake exceeding that of surrounding normal tissue. A region of interest was placed over each lesion to include the highest levels of radioactivity. Maximum SUV was calculated with the following formula: SUV = cdc/(di/w), wherein cdc is the decay-corrected tracer tissue concentration (Bq/g), di is the injected dose (Bq), and w is the patient’s body weight (g).
Immunohistochemical staining was performed to determine GLUT1 and HK2 levels in gastric cancer tumors. Briefly, resected specimens were fixed in 10% buffered formalin solution, embedded in paraffin, and sectioned at a thickness of 4 μm. Slides were then incubated overnight at room temperature with primary rabbit polyclonal antibody against GLUT1 (1:200) or HK2 (1:100). Avidin-biotin-peroxidase complex staining was performed according to the manufacturer’s instructions (Santa Cruz Biotechnology, CA, USA). Finally, nuclei were counterstained with hematoxylin .
Total RNA was isolated from specimens by guanidinium isothiocyanate-acid phenol extraction and quantified by absorbance at 260 nm. Total RNA (1 μg) was used for reverse transcription, and the resulting cDNA was analyzed by real-time PCR with Power SYBR Green PCR Master Mix and ABI Prism 7000 (Applied Biosystems, Foster, CA, USA). Target-specific oligonucleotide primers and probes were previously described [20, 21]. 18S rRNA was used as an endogenous control. Primers and probes for 18S rRNA were obtained in a Pre-Developed TaqMan Assay Reagent kit (Applied Biosystems, Stockholm, Sweden).
Data are expressed as mean ± SEM. Paired SUV results were compared by student’s t-test. Multiple one-way analysis of variance was used to assess differences in mRNA levels. Correlation analyses were performed with Spearman’s correlation analysis test. P<0.05 was considered statistically significant.
Relationship between mean SUV and clinicopathological data in gastric cancer
These results indicate that SUV was not dependent on the number of lymph node metastases or cancer stage. Maximum tumor diameter was the only parameter with a significant difference. To more precisely determine its correlation with SUV, we carried out quantitative analysis (Figure 1c). Spearman’s correlation analysis indicated a possible relationship between the factors (rs = 0.33, P < 0.05).
Expression of glucose transporter and glucose metabolizing enzymes in gastric cancer
Relationship between mean SUV and HIF1α or PCNA expression in gastric cancer
Expression of HK1, HK2, GLUT1, and G6Pase mRNA levels in intestinal and non-intestinal gastric cancers
Correlation between mean SUV and tumor size, HIF1α mRNA levels, or PCNA mRNA levels in intestinal and non-intestinal gastric cancers
FDG-PET has been used to not only detect cancerous lesions, but also predict therapeutic response after chemotherapy [1, 11, 23]. There are several possible mechanisms behind its ability to reveal malignant potential or cancer cell activity. Our results found that SUV in stage 4 gastric cancer patients was no higher than in stage 2 and stage 3 patients, and the main tumor SUV did not reflect the number of lymph node metastases. Only tumor size was associated with SUV, a correlation also reported in breast, pancreatic, and colorectal cancers [20, 24, 25]. These finding narrow the FDG-PET mechanism possibilities by suggesting that SUV reflects tumor size rather than tumor cell activity for each cancer stage.
Over expression of glucose metabolism-related protein in tumors
A molecular explanation for high FDG uptake in cancerous tissues is the overexpression of GLUT1, the molecule reported to be responsible for FDG uptake in various cancers [20, 26]. Glucose uptake ability as assessed by FDG-PET was significantly correlated with the doubling time of tumors  because increased uptake can provide additional energy to support tumor growth. Yamada et al.  determined from immunohistochemistry that GLUT1 expression was an important factor for FDG uptake and also a prognostic tool for gastric cancer. Alakus et al.  similar reported that FDG uptake in gastric cancer is dependent on the degree of GLUT1 staining. Our immunohistochemical staining also showed strong GLUT1 expression in cell membranes, as well as GLUT1 mRNA expression 3.3-fold greater in tumors than the surrounding mucosa; however, Spearman’s correlation analysis did not find a relationship between GLUT1 expression and SUV. HK2 also plays an important role in FDG catabolism, with its overexpression significantly associated with SUV in malignant tumors [15, 28]. We also found HK2 overexpression in gastric cancer tumors, but there was again no correlation between HK2 expression and SUV. Other complicated mechanisms, such as blood flow, accumulation of inflammatory cells, and cellularity might be also contribute to the intensity of FDG uptake based on malignant energy demand .
Hypotheses of the increased glucose uptake in tumor
Two major hypotheses have been presented to explain the increased glucose uptake in cancerous tissue, either that enhanced glucose consumption is associated with tumor proliferative activity [12, 13] or that tissue hypoxia induces anaerobic glycolysis to increase glucose metabolism . Our results indicate that FDG uptake associated significantly with hypoxia, reflected by HIF1α expression, but not with proliferative activity, reflected by PCNA expression; these gastric cancer findings correspond to our previous report on colorectal cancer . Rapid cancer growth induces a hypoxic environment in tumors. HIF1α acts as a sensor for hypoxic stress and upregulates angiogenic factors and promotes transcription of several genes, including glucose transporters and glycolytic enzymes such as GLUT1 and HK, for tumor survival . HIF1α may also be involved with oncogenic alterations to glucose metabolism because it activates cancer-related gene transcription and affects pathways such as angiogenesis, cell survival, glucose metabolism, and cell invasion . HIF1α overexpression has been associated with increased patient mortality rates in several cancers, while inhibited expression reduced tumor growth in an in vitro study . HIF1α could thus play a central role in cancer progression that FDG uptake represents.
Histological differences in the expression of glucose metabolism-related proteins
The non-intestinal gastric cancers, signet ring cell carcinoma and mucinous carcinoma, presented a very low FDG uptake compared to their intestinal counterparts due to low GLUT1 expression [1, 3, 7, 8]. Berger et al. reported that FDG-PET revealed an unusually high percentage (41%) of false-negative results in carcinoma with mucin. There was a positive correlation of FDG uptake with tumor cellularity but a negative correlation with the amount of mucin . Therefore, non-intestinal gastric cancers, which have characters of low cellularity and/or high mucin content, do not show high FDG uptake. Alakus et al. has reported that over expression of GLUT1 in papillary/tubular adenocarcinoma and signet ring cell carcinoma was 94% and 24%, respectively . Our results also indicate that GLUT1 expression in non-intestinal cancers was lower than in intestinal cancers. However, the reason why such aggressive cancers showed low GLUT1 expression is unknown. A previous study found that glutamine metabolism is upregulated in gastric cancer . Gastric cancer cells use glutamine as an energy source in a hypoxic tumor microenvironment, which may eliminate the necessity for glucose transport. This metabolic alteration accompanied with malignant transformation has been reported in other cancers . Interestingly, a glutamine-based PET is being developed; if successful, this contradiction could be disproved in the future.
On the other hand, HIF1α expression correlated with SUV in both types, although a more significant correlation was seen in non-intestinal specimens. The non-intestinal tumors may have been influenced more by hypoxia derived from tumor fibrosis due to a scattering tumor growth pattern than hypoxia due to increased tumor size. Further research will be needed to determine the exact reason.
Limitations of this study
There are several limitations in our study. First, we examined 50 cases of gastric cancer patients. The fewness of cases affects the statistical analysis and makes it difficult to get firm results in association of FDG uptake and the expression of the proteins. Second, we could not exclude the possibility of contribution of physiological FDG uptake in normal stomach on cancerous lesion. Finally, our results did not show the direct physiological relationship between HIF1α as a marker of hypoxic condition and FDG accumulation.
The usefulness of FDG-PET in the detection of malignant tumors or prediction of prognoses has been widely reported. However, our results indicate that the degree of FDG accumulation does not always suggest a prognosis in gastric cancer. This study is the first to show the correlation by evaluating FDG uptake in a quantitative manner. Upregulation of glucose transport due to increased GLUT1 expression was not an explanation for the different FDG uptakes observed, although tumor hypoxia and HIF1α expression may provide a reasonable mechanism. Further investigation is needed to confirm these results, but metabolic alternation through HIF1α induction in tumor hypoxia could increase FDG uptake in gastric cancer.
We are extremely grateful to all the clinical staff who cared for these patients. We also are thankful to Dr. Shoji Kimura for his reliable experimental suggestion.
- Shimada H, Okazumi S, Koyama M, Murakami K: Japanese gastric cancer association task force for research promotion: clinical utility of 18F-fluoro-2-deoxyglucose positron emission tomography in gastric cancer. A systematic review of the literature. Gastric Cancer. 2011, 14: 13-21. 10.1007/s10120-011-0017-5.View ArticleGoogle Scholar
- Murakami K: FDG-PET for Hepatobiliary and pancreatic cancer: advances and current limitations. World J Clin Oncol. 2011, 2: 229-236.Google Scholar
- Alakus H, Batur M, Schmidt M, Drebber U, Baldus SE, Vallböhmer D, Prenzel KL, Metzger R, Bollschweiler E, Hölscher AH, Mönig SP: Variable 18F-fluorodeoxyglucose uptake in gastric cancer is associated with different levels of GLUT-1 expression. Nucl Med Commun. 2010, 31: 532-538.Google Scholar
- Chen J, Cheong JH, Yun MJ, Kim J, Lim JS, Hyung WJ, Noh SH: Improvement in preoperative staging of gastric adenocarcinoma with positron emission tomography. Cancer. 2005, 103: 2383-2390. 10.1002/cncr.21074.View ArticleGoogle Scholar
- Mochiki E, Kuwano H, Katoh H, Asao T, Oriuchi N, Endo K: Evaluation of 18F-2-deoxy-2-fluoro-D-glucose positron emission tomography for gastric cancer. World J Surg. 2004, 28: 247-253. 10.1007/s00268-003-7191-5.View ArticleGoogle Scholar
- Kamimura K, Nagamachi S, Wakamatsu H, Fujita S, Nishii R, Umemura Y, Ogita M, Komada N, Sakurai T, Inoue T, Fujimoto T, Nakajo M: Role of gastric distention with additional water in differentiating locally advanced gastric carcinomas from physiological uptake in the stomach on 18F-fluoro-2-deoxy-D-glucose PET. Nucl Med Commun. 2009, 30: 431-439. 10.1097/MNM.0b013e3283299a2f.View ArticleGoogle Scholar
- Yamada A, Oguchi K, Fukushima M, Imai Y, Kadoya M: Evaluation of 2-deoxy-2-[18F]fluoro-D-glucose positron emission tomography in gastric carcinoma: relation to histological subtypes, depth of tumor invasion, and glucose transporter-1 expression. Ann Nucl Med. 2006, 20: 597-604. 10.1007/BF02984657.View ArticleGoogle Scholar
- Kawamura T, Kusakabe T, Sugino T, Watanabe K, Fukuda T, Nashimoto A, Honma K, Suzuki T: Expression of glucose transporter-1 in human gastric carcinoma: association with tumor aggressiveness, metastasis, and patient survival. Cancer. 2001, 92: 634-641. 10.1002/1097-0142(20010801)92:3<634::AID-CNCR1364>3.0.CO;2-X.View ArticleGoogle Scholar
- Ott K, Herrmann K, Krause BJ, Lordick F: The value of PET imaging in patients with localized gastroesophageal cancer. Gastrointest Cancer Res. 2008, 2: 287-294.Google Scholar
- Ott K, Herrmann K, Lordick F, Wieder H, Weber WA, Becker K, Buck AK, Dobritz M, Fink U, Ulm K, Schuster T, Schwaiger M, Siewert JR, Krause BJ: Early metabolic response evaluation by fluorine-18 fluorodeoxyglucose positron emission tomography allows in vivo testing of chemosensitivity in gastric cancer: long-term results of a prospective study. Clin Cancer Res. 2008, 14: 2012-2018. 10.1158/1078-0432.CCR-07-0934.View ArticleGoogle Scholar
- Ott K, Fink U, Becker K, Stahl A, Dittler HJ, Busch R, Stein H, Lordick F, Link T, Schwaiger M, Siewert JR, Weber WA: Prediction of response to preoperative chemotherapy in gastric carcinoma by metabolic imaging: results of a prospective trial. J Clin Oncol. 2003, 21: 4604-4610. 10.1200/JCO.2003.06.574.View ArticleGoogle Scholar
- Adler LP, Blair HF, Makley JT, Williams RP, Joyce MJ, Leisure G, al-Kaisi N, Miraldi F: Noninvasive grading of musculoskeletal tumors using PET. J Nucl Med. 1991, 32: 1508-1512.Google Scholar
- Avril N: GLUT1 expression in tissue and (18)F-FDG uptake. J Nucl Med. 2004, 45: 930-932.Google Scholar
- Vander Heiden MG, Cantley LC, Thompson CB: Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science. 2009, 324: 1029-1033. 10.1126/science.1160809.View ArticleGoogle Scholar
- Paudyal B, Oriuchi N, Paudyal P, Tsushima Y, Higuchi T, Miyakubo M, Ishikita T, Nakajima T, Endo K: Clinicopathological presentation of varying 18F-FDG uptake and expression of glucose transporter 1 and hexokinase II in cases of hepatocellular carcinoma and cholangiocellular carcinoma. Ann Nucl Med. 2008, 22: 83-86. 10.1007/s12149-007-0076-1.View ArticleGoogle Scholar
- Khan MA, Combs CS, Brunt EM, Lowe VJ, Wolverson MK, Solomon H, Collins BT, Di Bisceglie AM: Positron emission tomography scanning in the evaluation of hepatocellular carcinoma. J Hepatol. 2000, 32: 792-797.View ArticleGoogle Scholar
- Miyakubo M, Oriuchi N, Tsushima Y, Higuchi T, Koyama K, Arai K, Paudyal B, Iida Y, Hanaoka H, Ishikita T, Nakasone Y, Negishi A, Mogi K, Endo K: Diagnosis of maxillofacial tumor with L-3-[18F]-fluoro-alpha-methyltyrosine (FMT) PET: a comparative study with FDG-PET. Ann Nucl Med. 2007, 21: 129-135. 10.1007/BF03033991.View ArticleGoogle Scholar
- Baserga R: Growth regulation of the PCNA gene. J Cell Sci. 1991, 98: 433-436.Google Scholar
- Hong SS, Lee H, Kim KW: HIF-1alpha: a valid therapeutic target for tumor therapy. Cancer Res Treat. 2004, 36: 343-353. 10.4143/crt.2004.36.6.343.View ArticleGoogle Scholar
- Izuishi K, Yamamoto Y, Sano T, Takebayashi R, Nishiyama Y, Mori H, Masaki T, Morishita A, Suzuki Y: Molecular mechanism underlying the detection of colorectal cancer by 18F-2-fluoro-2-deoxy-D: -glucose positron emission tomography. J Gastrointest Surg. 2012, 16: 394-400. 10.1007/s11605-011-1727-z.View ArticleGoogle Scholar
- Kameyama R, Yamamoto Y, Izuishi K, Sano T, Nishiyama Y: Correlation of 18F-FLT uptake with equilibrative nucleoside transporter-1 and thymidine kinase-1 expressions in gastrointestinal cancer. Nucl Med Commun. 2011, 32: 460-465. 10.1097/MNM.0b013e32834209b5.View ArticleGoogle Scholar
- Kuang Y, Schomisch SJ, Chandramouli V, Lee Z: Hexokinase and glucose-6-phosphatase activity in woodchuck model of hepatitis virus-induced hepatocellular carcinoma. Comp Biochem Physiol C Toxicol Pharmacol. 2006, 143: 225-231. 10.1016/j.cbpc.2006.02.005.View ArticleGoogle Scholar
- Di Fabio F, Pinto C, Rojas Llimpe FL, Fanti S, Castellucci P, Longobardi C, Mutri V, Funaioli C, Sperandi F, Giaquinta S, Martoni AA: The predictive value of 18F-FDG-PET early evaluation in patients with metastatic gastric adenocarcinoma treated with chemotherapy plus cetuximab. Gastric Cancer. 2007, 10: 221-227. 10.1007/s10120-007-0438-3.View ArticleGoogle Scholar
- Heudel P, Cimarelli S, Montella A, Bouteille C, Mognetti T: Value of PET-FDG in primary breast cancer based on histopathological and immunohistochemical prognostic factors. Int J Clin Oncol. 2010, 15: 588-593. 10.1007/s10147-010-0120-3.View ArticleGoogle Scholar
- Izuishi K, Yamamoto Y, Sano T, Takebayashi R, Masaki T, Suzuki Y: Impact of 18-fluorodeoxyglucose positron emission tomography on the management of pancreatic cancer. J Gastrointest Surg. 2010, 14: 1151-1158. 10.1007/s11605-010-1207-x.View ArticleGoogle Scholar
- Usuda K, Sagawa M, Aikawa H, Ueno M, Tanaka M, Machida Y, Zhao XT, Ueda Y, Higashi K, Sakuma T: Correlation between glucose transporter-1 expression and 18F-fluoro-2-deoxyglucose uptake on positron emission tomography in lung cancer. Gen Thorac Cardiovasc Surg. 2010, 58: 405-410. 10.1007/s11748-010-0603-1.View ArticleGoogle Scholar
- Duhaylongsod FG, Lowe VJ, Patz EF, Vaughn AL, Coleman RE, Wolfe WG: Lung tumor growth correlates with glucose metabolism measured by fluoride-18 fluorodeoxyglucose positron emission tomography. Ann Thorac Surg. 1995, 60: 1348-1352. 10.1016/0003-4975(95)00754-9.View ArticleGoogle Scholar
- Ong LC, Jin Y, Song IC, Yu S, Zhang K, Chow PK: 2-[18F]-2-deoxy-D-glucose (FDG) uptake in human tumor cells is related to the expression of GLUT-1 and hexokinase II. Acta Radiol. 2008, 49: 1145-1153. 10.1080/02841850802482486.View ArticleGoogle Scholar
- Dang CV, Semenza GL: Oncogenic alterations of metabolism. Trends Biochem Sci. 1999, 24: 68-72. 10.1016/S0968-0004(98)01344-9.View ArticleGoogle Scholar
- Semenza GL: Targeting HIF-1 for cancer therapy. Nat Rev Cancer. 2003, 3: 721-732.View ArticleGoogle Scholar
- Berger KL, Nicholson SA, Dehdashti F, Siegel BA: FDG PET evaluation of mucinous neoplasms: correlation of FDG uptake with histopathologic features. AJR Am J Roentgenol. 2000, 174: 1005-1008. 10.2214/ajr.174.4.1741005.View ArticleGoogle Scholar
- Hirayama A, Kami K, Sugimoto M, Sugawara M, Toki N, Onozuka H, Kinoshita T, Saito N, Ochiai A, Tomita M, Esumi H, Soga T: Quantitative metabolome profiling of colon and stomach cancer microenvironment by capillary electrophoresis time-of-flight mass spectrometry. Cancer Res. 2009, 69: 4918-4925. 10.1158/0008-5472.CAN-08-4806.View ArticleGoogle Scholar
- Rajagopalan KN, DeBerardinis RJ: Role of glutamine in cancer: therapeutic and imaging implications. J Nucl Med. 2011, 52: 1005-1008. 10.2967/jnumed.110.084244.View ArticleGoogle Scholar
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.