18 F-Fluoride positron emission tomography/computed tomography for noninvasive in vivo quantification of pathophysiological bone metabolism in experimental murine arthritis
© Irmler et al.; licensee BioMed Central Ltd. 2014
Received: 4 March 2014
Accepted: 9 July 2014
Published: 22 July 2014
Evaluation of disease severity in experimental models of rheumatoid arthritis is inevitably associated with assessment of structural bone damage. A noninvasive imaging technology allowing objective quantification of pathophysiological alterations of bone structure in rodents could substantially extend the methods used to date in preclinical arthritis research for staging of autoimmune disease severity or efficacy of therapeutical intervention. Sodium 18 F-fluoride (18 F-NaF) is a bone-seeking tracer well-suited for molecular imaging. Therefore, we systematically examined the use of 18 F-NaF positron emission tomography/computed tomography (PET/CT) in mice with glucose-6-phosphate isomerase (G6PI)–induced arthritis for quantification of pathological bone metabolism.
F-fluoride was injected into mice before disease onset and at various time points of progressing experimental arthritis. Radioisotope accumulation in joints in the fore- and hindpaws was analyzed by PET measurements. For validation of bone metabolism quantified by 18 F-fluoride PET, bone surface parameters of high-resolution μCT measurements were used.
Before clinical arthritis onset, no distinct accumulation of 18 F-fluoride was detectable in the fore- and hindlimbs of mice immunized with G6PI. In the course of experimental autoimmune disease, 18 F-fluoride bone uptake was increased at sites of enhanced bone metabolism caused by pathophysiological processes of autoimmune disease. Moreover, 18 F-fluoride signaling at different stages of G6PI-induced arthritis was significantly correlated with the degree of bone destruction. CT enabled identification of exact localization of 18 F-fluoride signaling in bone and soft tissue.
The results of this study suggest that small-animal PET/CT using 18 F-fluoride as a tracer is a feasible method for quantitative assessment of pathophysiological bone metabolism in experimental arthritis. Furthermore, the possibility to perform repeated noninvasive measurements in vivo allows longitudinal study of therapeutical intervention monitoring.
Rheumatoid arthritis (RA) is one of the most common autoimmune diseases, affecting approximately 1% of the population in Europe and North America. Bone erosion can be detected as early as several weeks after the onset of the first clinical signs and symptoms of RA. Pathogenesis in preclinical models is similar to that in clinical RA; it is characterized by inflammation and bone destruction. In the past few decades, histopathological evaluation of joint sections has mainly been used for assessment of inflammation and bone destruction in small-rodent models of arthritis. Recently, we showed the benefit of state-of-the-art imaging modalities for visualization and quantitative assessment of inflammation in glucose-6-phosphate isomerase (G6PI)–induced arthritis using 2-deoxy-2-18 F-fluoroglucose (18 F-FDG) positron emission tomography/computed tomography (PET/CT). Also, it has been shown that cell proliferation can be detected in experimental arthritis with the PET proliferation tracer 3′-deoxy-3′-18 F-fluorothymidine.
18 F-fluoride not only can be used to label glucose and other molecules of physiologic relevance but also has favorable properties in the form of sodium 18 F-fluoride (18 F-NaF) as a radiotracer per se in noninvasive in vivo imaging to investigate musculoskeletal diseases. The use of 18 F-NaF as a bone imaging probe was established by Blau et al. in the early 1960s, but it was subsequently replaced by 99mTc-labeled tracers due to their availability, lower costs and the lower energy of 140-keV photons, allowing the use of γ-cameras. 18 F-fluoride PET is an increasingly used molecular imaging modality, not only in human skeletal disorders but also in small-animal preclinical research[6–8]. This is facilitated by the distribution and use of three-dimensional PET scanners with high spatial resolution in clinical medicine and advantages over 99mTc-labeled bone agents used for skeletal scintigraphy, such as higher diagnostic accuracy and reduced scanning time. Compared to γ-cameras, molecular imaging using PET provides the advantages of higher spatial resolution, higher sensitivity and three-dimensional tomographic image reconstruction. Furthermore, the combination of PET with μCT enables attenuation correction of radiotracer signaling, allowing quantitative measurements using 18 F-fluoride PET/CT.
Applied 18 F-NaF, dissociated into Na+ and 18 F−, is rapidly cleared from the blood and accumulates in the bone, where, on the hydroxyapatite surface, an OH− ion is replaced by an 18 F− ion to form fluorapatite. The incorporation of 18 F-fluoride in the bone is determined by vascular perfusion and bone surface accessible for ion exchange, indirectly reflecting bone formation and bone resorption. This means that 18 F-NaF can be used not only for the common measurement of bone mineral deposition but also for visualization of osteolytic increases of exposed bone surface, such as in the context of musculoskeletal autoimmune disease.
In clinical oncology, primary bone tumors and skeletal metastasis can reliably be detected by 18 F-fluoride PET[4, 11]. In mice, pathological osteoblastic activity can be detected even earlier by 18 F-fluoride PET/CT imaging than by radiography and corresponds to histological evaluation of increased bone formation. As with bone tumor pathogenesis, a pathologically increased bone metabolism is a central feature of chronic arthritis, resulting in functional disorders of the joints. Therefore, in our present study, we examined the use of 18 F-NaF small-animal PET/CT for the quantitative and noninvasive in vivo assessment of pathophysiological bone metabolism in acute and chronic stages of experimental G6PI-induced murine arthritis. 18 F-NaF PET quantification of bone destruction, visible as lesions in cortical bone surface and dysregulated bone formation, was validated using high-resolution CT measurements of the paws.
Glucose-6-phosphate isomerase–induced arthritis
DBA/1 mice were bred at the animal facility of the Jena University Hospital (Jena, Germany). All animal studies were approved by the local commission for animal protection (Thüringer Landesamt für Verbraucherschutz, Bad Langensalza, Germany; registered number 02-045/08). Arthritis was induced as described elsewhere. In brief, DBA/1 mice were immunized subcutaneously with 400 μg of recombinant human G6PI in emulsified complete Freund’s adjuvant (Sigma-Aldrich, Taufkirchen, Germany). Macroscopic evaluation of arthritis was performed according to severity of clinical manifestations in wrist and ankle joints, in metacarpophalangeal and metatarsophalangeal joints, and in digits and toes. Swelling and redness in wrist and ankle joints and in metacarpophalangeal and metatarsophalangeal joints, respectively, were graded from 0 to 3. A score of 0 indicates no macroscopically recognizable signs of arthritis, 1 indicates swelling and redness, 2 means strong swelling and redness and 3 indicates massive swelling and redness. Additionally, the number of digits and toes with inflamed joints were divided in half to avoid assessment imbalance because inflammation in the G6PI-induced arthritis model is located mainly in proximal joints of the paws. To calculate the total clinical score per animal, results from all paws were summed. For PET/CT measurements, mice were anesthetized with 1.5% to 2% isoflurane (Deltaselect, Dreieich, Germany) vaporized in oxygen (1.5 L/min) to prevent animal movement and reduce imaging artifacts. Respiration of mice under anesthesia was monitored. 18 F-fluoride (half-life = 109 minutes; Eckert & Ziegler, Bad Berka, Germany) with an activity of 10.4 ± 0.8 MBq was injected intravenously into the lateral tail vein. Longitudinal arthritis imaging was performed at various time points of acute and chronic clinical arthritis (n = 3 to 6 mice per time point). We obtained dynamic PET scans for kinetic analysis of tracer uptake in nonimmunized mice (n = 3).
Positron emission tomography/computed tomography in vivo imaging
In vivo18 F-NaF imaging was performed using a Siemens Inveon small-animal multimodality PET/CT system (Preclinical Solutions, Siemens Healthcare Molecular Imaging, Knoxville, TN, USA), characterized by the combination of two independently operating PET and CT scanners. Radial, tangential and axial resolutions at the center of the field of view of the PET module are 1.5 mm for this imaging modality[14, 15]. PET image acquisition was carried out with a coincidence timing window of 3.4 ns and an energy window of 350 to 650 keV. PET data acquisition was performed for 1,800 seconds, starting 50 minutes after tracer application. In kinetic analysis, PET data acquisition lasted 5,400 seconds and started concomitantly with radiotracer injection. Images were reconstructed into three-dimensional images using Fourier rebinning and three-dimensional ordered-subset expectation maximization algorithm. Attenuation of PET data was corrected based on the CT measurements. The CT module consists of a cone-beam X-ray μCT source (50-μm focal spot size) and a 3,072 × 2,048–pixel X-ray detector. In our μCT imaging protocol, we used an axial–transaxial resolution of 2,048 × 2,048–pixel, 80 kVp at 500 μA, 360° rotation and 360 projections per bed position for paw measurements and 3,072 × 2,048–pixel, 220° rotation and 120 projections per bed position for whole-animal attenuation scans, respectively. CT images were reconstructed using a Shepp-Logan filter and cone-beam–filtered back projection. To reduce stress due to overly prolonged measurement times, only high-resolution μCT data from hindpaws were acquired for correlation analysis of bone surface assessment with 18 F-fluoride PET and CT.
Assessment of pathophysiological bone metabolism with 18 F-fluoride or μCT
Quantitative analysis of 18 F-fluoride accumulation in the period from 50 to 80 minutes after 18 F-fluoride injection or 0 to 90 minutes, respectively, was enabled by image fusion technology in Siemens Inveon Research Workplace 4.0 software. For dynamic PET analysis, scans were started 5 seconds before radiotracer injection and continued for 90 minutes. The 90-minute data set was divided into 45 time frames (6 × 10 seconds, 6 × 20 seconds, 7 × 60 seconds, 10 × 120 seconds, 10 × 180 seconds and 6 × 300 seconds) during histogramming to construct time–activity curves. For static analysis, data from 75 to 80 minutes after 18 F-fluoride injection were analyzed in a single time frame. Radioisotope activity in the venous blood pool or in fore- or hindpaws, reflecting bone incorporation of 18 F-fluoride, was measured as standardized uptake value (SUV; g/ml) using PMOD 3.15 software (PMOD Technologies Ltd, Zurich, Switzerland) or Siemens Inveon Research Workplace 4.0 software. Guided by maximum intensity projection images, volumes of interest were drawn as spheres (ellipsoids) over anatomic structures of bones and joints. A threshold of 40% (max–min) was used to separate the site of tracer accumulation from background tissue signaling. Visualization of skeletal elements of the hindpaws was also done using PMOD 3.15 software.
Analyses of metatarsal CT data were performed using Definiens Developer XD™ 2.0.3 build 2015 software (Definiens, Munich, Germany). For segmentation of metatarsal bones, the automatic threshold routine was used. Then, individual bones were discriminated by watershed segmentation. Oversegmented fragments were merged manually to achieve three-dimensional objects of each bone. After identification of metatarsals 1 to 5, a local threshold was applied on each metatarsal bone to refine bone margins and remove remaining nonbone pixels. The threshold was calculated as the difference between mean pixel gray value and standard deviation of pixel gray values. Quantitative measurements of bone surface and bone volume were calculated for each metatarsal bone.
Statistical differences between groups were evaluated using the nonparametric Mann-Whitney U test. Correlation analyses (Spearman test) were performed for PET data and clinical scoring or PET and CT data, respectively, whereas CT surface data of all metatarsals, per paw, were summed. Statistical significance was accepted at P < 0.05 (*P < 0.05, **P < 0.01, ***P < 0.001). In bar charts and text, data are given as arithmetic means and standard errors of the mean (SEM). All calculations were performed using the software package IBM SPSS Statistics version 20.0 (IBM, Ehningen, Germany).
Inflammation and bone destruction in glucose-6-phosphate isomerase–induced arthritis
Kinetics of 18 F-fluoride bone uptake in mice
Concomitantly, the quantitative time–activity curve analysis of 18 F-fluoride paw uptake revealed rapid radiotracer accumulation beginning immediately after injection and tracer signaling remaining at almost the same level from 40 to 90 minutes postinjection (Figure 2B). At 45 minutes postinjection, 90% of 18 F-fluoride uptake was achieved (SUVmean = 1.8), compared to bone radiotracer signaling after 90 minutes (SUVmean = 1.9). In contrast, blood pool tracer activity measured in the venae cavae region showed a rapid decline after an activity maximum (SUVmax = 15.2), within minutes following injection, reflecting rapid clearance of 18 F-fluoride from the blood.
Localization of exact sites of 18 F-fluoride paw bone uptake in glucose-6-phosphate isomerase–induced arthritis
Quantification of pathophysiological bone metabolism using 18 F-fluoride PET in progressing murine arthritis
To examine the general effects of arthritis on 18 F-fluoride uptake in elements of the skeleton, we quantified PET signaling in medial and distal femoral bones. SUVs revealed high tracer uptake (SUV > 10) in trabecular distal areas of the femur (Figure 4C). Except for late remitting arthritis (d39), SUVs remained at a similar level at different stages of G6PI-induced arthritis, indicating only a slight influence of pathological bone metabolism associated with arthritis on distal femoral tracer uptake. In contrast to trabecular distal areas, mean SUVs in median cortical areas of the femur ranged from 2.1 to 2.2 in all stages of arthritis (Figure 4D), additionally negating general effects of arthritis on bone accumulation of 18 F-fluoride at sites without clinical manifestations of disease. 18 F-fluoride PET/CT measurements in healthy mice without any pathological conditions revealed a nonequal pattern of bone tracer uptake (Additional file1: Figures S1A and S1B). There were hotspots of PET signaling in proximal joints of the pectoral girdle; in the pelvis, including the knees; and in spinal and cranial bones. Volume-rendering of the knee joints revealed a coincidence of increased 18 F-fluoride accumulation and trabecular bone (Additional file1: Figure S1C). Therefore, we assume that the increased surface in trabecular bones, meeting the demands of distinct mechanical stress or saving weight, is the cause for the observed heterogeneous pattern of 18 F-fluoride uptake in skeletal bone.
Quantification of pathophysiological bone metabolism using μCT
Regression analysis of clinical parameters and quantitative PET and CT assessment of bone parameters
The usefulness of the small-animal PET/CT imaging modality using 18 F-NaF as a radiotracer is not restricted to the detection of primary bone tumors and skeletal metastasis in cancer research. 18 F-fluoride PET/CT imaging is also a valid technique for use in the assessment of disease severity according to pathological bone turnover in the field of preclinical arthritis research. In this study, in which we employed the G6PI-induced arthritis model, we focused on the feasibility of using 18 F-fluoride PET to visualize and quantify pathological bone metabolism in distal murine arthritic joints noninvasively and in vivo. In addition to joint inflammation, which can easily be quantified by 18 F-FDG PET/CT according to activation and proliferation of resident cells and migrated cells of the immune system, bone damage is the second major parameter used for assessment of arthritis severity.
In contrast to 18 F-FDG, which is trapped in cells at sites of inflammation due to pathologically increased glucose metabolism, 18 F-fluoride represents for specific radiotracer accumulation in the bone. Erosive processes degrading bone and cartilage and dysfunctional bone repair mechanisms in RA and its animal models are associated with an increased bone surface. Therefore, the increased mineral-binding capacity results in site-specific 18 F-fluoride uptake in arthritic joints, which can easily be used for visualization and, more importantly, provides a measurement method for the quantification of pathological bone metabolism in preclinical arthritis models. The results of our studies show increased uptake of 18 F-fluoride in the paws of arthritic mice and reveal that the 18 F-fluoride PET/CT quantification of pathologic bone metabolism in arthritis pathogenesis significantly correlated with our assessment of pathophysiologic bone surface alterations based on high-resolution μCT measurements.
Radioisotope imaging is, to date, one of the most applicable imaging modalities for functional and whole-body in vivo quantification of bone metabolism in mice. Compared to other radiopharmaceuticals used for bone imaging, such as 99mTc, 18 F-fluoride has some beneficial attributes. First, 18 F-fluoride has a high affinity to bone, resulting in favorable skeletal kinetics. Within 60 minutes after intravenous injection, only 10% of the injected dose is still located in the bloodstream. Thus the concurrence of rapid bone uptake and fast blood-pool clearance yields a favorable bone-to-background ratio. Additionally, 18 F-fluoride does not accumulate in inflamed soft tissue and only minimally binds to serum proteins. Furthermore, small-animal 18 F-NaF PET measurements have an excellent reproducibility in animal models of bone disease[6, 17].
One limiting factor in 18 F-fluoride PET imaging might be vascularization of the tissue restricting tracer delivery. In contrast to epithelial tissue, the circulation in well-vascularized bone tissue is less affected by exogenous factors, allowing reproducible data acquisition. In experimental arthritis, the increased vascularization and blood flow in inflamed tissue may influence tracer delivery and, therefore, PET signaling at stages of acute inflammation. Nevertheless, at time points of maximal tracer uptake in our experiments, clinical arthritis was already remitting and macroscopically visible signs of inflammation were diminished. Another important aspect of 18 F-fluoride PET imaging is bone structure. Skeletal elements consisting of cortical and trabecular bone result in strong baseline PET signaling. This is presumably caused by the manifold increase in bone surface and less by increased metabolism at these sites, as binding capacities of cortical and trabecular bone are only slightly different. However, both decrease significantly if bone matrix is demineralized. Therefore, if bones with cancellous and noncancellous structures are located near regions of interest, data analyses require a high degree of anatomical accuracy, which can be achieved with μCT. In quantification of experimental arthritis severity, we found that the effect of high baseline signaling in tibial and radial bones was of negligible relevance for 18 F-fluoride PET imaging, because signaling hotspots of arthritis pathogenesis were located in tarsal, metatarsal and carpal and metacarpal bones. Another source of a high and pathological 18 F-fluoride baseline PET signaling independent of autoimmune disease pathogenesis might be the preoccurrence of osteoarthritis resulting in mechanical bone erosion. Whereas this aspect is not relevant for imaging in experimental arthritis models, it should be considered in RA bone imaging, where age is a risk factor.
In our present study, we demonstrate the capability of 18 F-fluoride PET to monitor and quantify pathological bone conditions in the model of G6PI-induced arthritis. Furthermore, we validated this bone imaging technique successfully by using bone surface μCT data. Because 18 F-fluoride PET is a noninvasive and nondestructive way to measure bone metabolism in vivo, this molecular imaging modality is not only useful for numerous applications in basic animal science, where pathophysiological bone metabolism is of interest, but also is a valuable tool for use in preclinical arthritis research, where pathological bone destruction and inflammation are the major parameter used for assessment of disease severity.
- 18 F-FDG:
18 F-labelled fluorodeoxyglucose
Glucose-6-phosphate isomerase–induced arthritis
Positron emission tomography
Standardized uptake value.
The authors express their gratitude to Alexander Heidrich, Cornelia Huettich and Annett Krause for excellent technical and organizational assistance and to the Bundesministerium für Bildung und Forschung for financial support (BMBF; FKZ 0316040B).
- Backhaus M, Kamradt T, Sandrock D, Loreck D, Fritz J, Wolf KJ, Raber H, Hamm B, Burmester GR, Bollow M: Arthritis of the finger joints: a comprehensive approach comparing conventional radiography, scintigraphy, ultrasound, and contrast-enhanced magnetic resonance imaging. Arthritis Rheum. 1999, 42: 1232-1245. 10.1002/1529-0131(199906)42:6<1232::AID-ANR21>3.0.CO;2-3.View ArticlePubMedGoogle Scholar
- Irmler IM, Opfermann T, Gebhardt P, Gajda M, Bräuer R, Saluz HP, Kamradt T: In vivo molecular imaging of experimental joint inflammation by combined 18 F-FDG positron emission tomography and computed tomography. Arthritis Res Ther. 2010, 12: R203-10.1186/ar3176.PubMed CentralView ArticlePubMedGoogle Scholar
- Fuchs K, Kohlhofer U, Quintanilla-Martinez L, Lamparter D, Kötter I, Reischl G, Röcken M, Pichler BJ, Kneilling M: In vivo imaging of cell proliferation enables the detection of the extent of experimental rheumatoid arthritis by 3′-deoxy-3′-18 F-fluorothymidine and small-animal PET. J Nucl Med. 2013, 54: 151-158. 10.2967/jnumed.112.106740.View ArticlePubMedGoogle Scholar
- Even-Sapir E, Metser U, Mishani E, Lievshitz G, Lerman H, Leibovitch I: The detection of bone metastases in patients with high-risk prostate cancer: 99mTc-MDP planar bone scintigraphy, single- and multi-field-of-view SPECT, 18 F-fluoride PET, and 18 F-fluoride PET/CT. J Nucl Med. 2006, 47: 287-297.PubMedGoogle Scholar
- Blau M, Nagler W, Bender MA: Fluorine-18: a new isotope for bone scanning. J Nucl Med. 1962, 3: 332-334.PubMedGoogle Scholar
- Berger F, Lee YP, Loening AM, Chatziioannou A, Freedland SJ, Leahy R, Lieberman JR, Belldegrun AS, Sawyers CL, Gambhir SS: Whole-body skeletal imaging in mice utilizing microPET: optimization of reproducibility and applications in animal models of bone disease. Eur J Nucl Med Mol Imaging. 2002, 29: 1225-1236.View ArticlePubMedGoogle Scholar
- Hsu WK, Virk MS, Feeley BT, Stout DB, Chatziioannou AF, Lieberman JR: Characterization of osteolytic, osteoblastic, and mixed lesions in a prostate cancer mouse model using 18 F-FDG and 18 F-fluoride PET/CT. J Nucl Med. 2008, 49: 414-421. 10.2967/jnumed.107.045666.PubMed CentralView ArticlePubMedGoogle Scholar
- Grant FD, Fahey FH, Packard AB, Davis RT, Alavi A, Treves ST: Skeletal PET with 18 F-fluoride: applying new technology to an old tracer. J Nucl Med. 2008, 49: 68-78.View ArticlePubMedGoogle Scholar
- Tomlinson RE, Silva MJ, Shoghi KI: Quantification of skeletal blood flow and fluoride metabolism in rats using PET in a pre-clinical stress fracture model. Mol Imaging Biol. 2012, 14: 348-354. 10.1007/s11307-011-0505-3.PubMed CentralView ArticlePubMedGoogle Scholar
- Blau M, Ganatra R, Bender MA: 18 F-fluoride for bone imaging. Semin Nucl Med. 1972, 2: 31-37. 10.1016/S0001-2998(72)80005-9.View ArticlePubMedGoogle Scholar
- Langsteger W, Heinisch M, Fogelman I: The role of fluorodeoxyglucose, 18 F-dihydroxyphenylalanine, 18 F-choline, and 18 F-fluoride in bone imaging with emphasis on prostate and breast. Semin Nucl Med. 2006, 36: 73-92. 10.1053/j.semnuclmed.2005.09.002.View ArticlePubMedGoogle Scholar
- Schett G, Gravallese E: Bone erosion in rheumatoid arthritis: mechanisms, diagnosis and treatment. Nat Rev Rheumatol. 2012, 8: 656-664. 10.1038/nrrheum.2012.153.PubMed CentralView ArticlePubMedGoogle Scholar
- Schubert D, Maier B, Morawietz L, Krenn V, Kamradt T: Immunization with glucose-6-phosphate isomerase induces T cell-dependent peripheral polyarthritis in genetically unaltered mice. J Immunol. 2004, 172: 4503-4509. 10.4049/jimmunol.172.7.4503.View ArticlePubMedGoogle Scholar
- Constantinescu CC, Mukherjee J: Performance evaluation of an Inveon PET preclinical scanner. Phys Med Biol. 2009, 54: 2885-2899. 10.1088/0031-9155/54/9/020.PubMed CentralView ArticlePubMedGoogle Scholar
- Visser EP, Disselhorst JA, Brom M, Laverman P, Gotthardt M, Oyen WJ, Boerman OC: Spatial resolution and sensitivity of the Inveon small-animal PET scanner. J Nucl Med. 2009, 50: 139-147.View ArticlePubMedGoogle Scholar
- Blake GM, Park-Holohan SJ, Cook GJ, Fogelman I: Quantitative studies of bone with the use of 18 F-fluoride and 99mTc-methylene diphosphonate. Semin Nucl Med. 2001, 31: 28-49. 10.1053/snuc.2001.18742.View ArticlePubMedGoogle Scholar
- Frost ML, Blake GM, Park-Holohan SJ, Cook GJ, Curran KM, Marsden PK, Fogelman I: Long-term precision of 18 F-fluoride PET skeletal kinetic studies in the assessment of bone metabolism. J Nucl Med. 2008, 49: 700-707. 10.2967/jnumed.107.046987.View ArticlePubMedGoogle Scholar
- Mitterhauser M, Toegel S, Wadsak W, Mien LK, Eidherr H, Kletter K, Viernstein H, Kluger R, Engel A, Dudczak R: Binding studies of [18 F]-fluoride and polyphosphonates radiolabelled with [99mTc], [111In], [153Sm] and [188Re] on bone compartments: verification of the pre vivo model?. Bone. 2005, 37: 404-412. 10.1016/j.bone.2005.04.027.View ArticlePubMedGoogle 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/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.