Intranasal immunization with a helper-dependent adenoviral vector expressing the codon-optimized fusion glycoprotein of human respiratory syncytial virus elicits protective immunity in BALB/c mice
© Fu et al.; licensee BioMed Central Ltd. 2013
Received: 9 March 2013
Accepted: 3 June 2013
Published: 7 June 2013
Human respiratory syncytial virus (RSV) is a serious pediatric pathogen of the lower respiratory tract. Currently, there is no clinically approved vaccine against RSV infection. Recent studies have shown that helper-dependent adenoviral (HDAd) vectors may represent effective and safe vaccine vectors. However, viral challenge has not been investigated following mucosal vaccination with HDAd vector vaccines.
To explore the role played by HDAd as an intranasally administered RSV vaccine vector, we constructed a HDAd vector encoding the codon optimized fusion glycoprotein (Fsyn) of RSV, designated HDAd-Fsyn, and delivered intranasally HDAd-Fsyn to mice.
RSV-specific humoral and cellular immune responses were generated in BALB/c mice, and serum IgG with neutralizing activity was significantly elevated after a homologous boost with intranasal (i.n.) application of HDAd-Fsyn. Humoral immune responses could be measured even 14 weeks after a single immunization. Immunization with i.n. HDAd-Fsyn led to effective protection against RSV infection on challenge.
The results indicate that HDAd-Fsyn can induce powerful systemic immunity against subsequent i.n. RSV challenge in a mouse model and is a promising candidate vaccine against RSV infection.
Human respiratory syncytial virus (RSV) is a serious pediatric pathogen of the lower respiratory tract [1, 2] and causes significant illness in the elderly and adults with underlying risk factors such as immunodeficiency [3, 4]. Several approaches have been used to develop vaccines against RSV infection including live/attenuated vaccines, viral and bacterial vector vaccines, DNA vaccines, and adjuvant subunit vaccines. Although three vaccine candidates, two live attenuated and a viral vector vaccine, have been evaluated in seronegative infants and seropositive children, respectively [1, 5], no RSV vaccines have been approved for use in humans [3, 4, 6].
The replication-deficient first-generation adenovirus serotype 5 (FGAd5) vector can be readily grown and purified in large quantities and is able to express high levels of the transgene in dividing and non-dividing cells. Therefore, it is considered to be an attractive vaccine vector. Studies have shown that FGAd-based vaccines can elicit robust protective immune responses against RSV infection and represent promising candidates for an RSV vaccine [7–9]. However, the majority of the human population, particularly in developing countries and in adult populations, has been exposed to Ad5 and has pre-existing neutralizing antibodies that can attenuate Ad5-vector vaccine delivery and efficacy [10, 11]. For Ad5-vector vaccines against RSV infection, the situation is somewhat different. Ad5-vector vaccines encoding RSV antigens can be administered in Ad5-seronegative children aged between 6 and 24 months, when maternal antibodies have declined substantially and a rapid increase in Ad5 seroprevalence from natural infections is absent . It would also have a major impact on morbidity by preventing a first or second RSV infection in these infants .
In contrast to the FGAd vector, the helper-dependent adenoviral (HDAd) vector has all the adenovirus (Ad) coding regions deleted and exhibits lower Ad-specific cellular immunity and stronger longer-term gene expression in vivo[12–14]. Therefore, HDAd vectors have been constructed as a safer vector platform with increased transgene capacity compared to FGAd vectors [12–14]. For efficient packaging, the HDAd must often include “stuffer” DNA. The choice of stuffer DNA is important with regard to vector stability and replication efficiency. In general, noncoding eukaryotic DNA is preferable while repetitive elements and unnecessary homology with the helper virus should be avoided. Recent data from our group and others have demonstrated that HDAd vectors are capable of generating stronger immune responses against transgenes such as reporter proteins and HIV-env than FGAd vectors in mice [12–15]. However, there has been no investigation of viral challenge following mucosal vaccination with HDAd vector vaccines.
In this study, mice were immunized intranasally with a HDAd vector containing the codon-optimized full-length fusion glycoprotein (Fsyn) of RSV. The animals were then monitored for induction of RSV-specific humoral and cellular immune responses and protection against i.n. RSV challenge. We report that the HDAd-Fsyn vaccine candidate induces both a serum antibody response and an RSV-specific CD8+ T-cell response. To the best of our knowledge, this is the first study showing that an i.n. HDAd vector vaccine is highly effective at stimulating RSV F-specific humoral and cellular immunities. Our findings will be useful in the development of vaccines against RSV infection.
Materials and methods
Preparation and titration of RSV stock
Subgroup A RSV Long strain (kindly provided by Prof. Y. Qian, Capital Institute of Pediatrics, Beijing, China) was propagated in HEp-2 cells (ATCC, Rockefeller, MD, USA) in Dulbecco’s modified Eagle medium (DMEM; Invitrogen, Carlsbad, CA, USA) supplemented with 2% fetal calf serum (Invitrogen), L-glutamine (2 mol/L), penicillin G (40 U/mL), streptomycin (100 μg/mL) and 0.2% sodium bicarbonate. The infectivity of the resulting RSV was titrated using the immunoenzyme assay described by Wang et al.  with slight modifications.
Construction and purification of HDAd-Fsyn
For construction of HDAd-Fsyn and improving the expression level of F gene, the F sequence was codon optimized and synthesized by Geneart (Regensburg, Germany) (GenBank database entry EF566942 ). The Fsyn gene expression cassette with the cytomegalovirus (CMV) enhancer/promoter and the bovine growth hormone (BGH) polyA was cloned into the HDAd shuttle plasmid pSC11. The expression cassette was digested with restriction enzyme I-Sce I and I-Ceu I and cloned into the HDAd backbone plasmid pSC15B  to produce pSC15B-Fsyn. The resulting plasmid was linearized using the restriction enzyme Pme I and transfected into 293Cre4 cells (293 cells expressing Cre recombinase; Microbix, Toronto, Canada) using the calcium phosphate transfection method. The cells were infected with E1-E3-deleted helper virus H14 (Microbix) 16 h after transfection. HDAd-Fsyn was amplified by serial coinfection of 293Cre4 cells with helper virus and crude lysates from the previous passage. Large-scale production of HDAd-Fsyn was achieved by infecting 293Cre4 in 150-mm dishes with crude HDAd-Fsyn and helper virus, purified by CsCl banding. The concentration of purified HDAd-Fsyn was determined using a previously described method . The amount of contaminated helper virus was also ascertained by determining the number of plaque-forming units (pfu/mL) after titration of the purified HDAd-Fsyn in 293 cells [18, 19]. HDAd vector encoding enhanced green fluorescent protein (HDAd-EGFP) was constructed as previously described and used as a control vector .
Specific pathogen-free female BALB/c mice, aged between 6 and 8 weeks, were purchased from Vital River Laboratories (Beijing, China) and kept under specific pathogen-free conditions. All animal studies were performed according to the guidelines of our Institutional Animal Care and Use Committee. The protocol was approved by the Committee on the Ethics of Animal Experiments of Beijing Jiaotong University (Permit number: 2010–0013). All surgeries were performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering.
Immunization and challenge
BALB/c mice were divided into four groups: HDAd-Fsyn (once), HDAd-EGFP (once), HDAd-Fsyn (prime + boost), and HDAd-EGFP (prime + boost). After BALB/c mice being lightly anesthetized with pentobarbital sodium (4 μg/kg weight), a dose of 5×108 virus particles (in 50 μL) of either HDAd-Fyn or HDAd-EGFP was delivered intranasally either once (in week 0) or twice (in weeks 0 and 4). Three weeks after the final immunization, mice that received two doses of HDAd-Fsyn or HDAd-EGFP were subjected to i.n. challenge with 100 μL of subgroup A RSV Long strain (106 pfu/mL).
Collection of splenocytes
Spleens from vaccinated mice were harvested and placed in mouse lymphocyte separation medium. After BALB/c mice being lightly anesthetized with pentobarbital sodium (4 μg/kg weight), the mice were were sacrificed by cervical dislocation. The spleens were triturated and passed gently through cell strainers (Becton-Dickinson, San Jose, CA, USA) to obtain single-cell suspensions, which were centrifuged at 800g for 30 min. The splenocytes were collected and washed with complete RPMI 1640 medium (Invitrogen).
Preparation of lung homogenates
After sacrifice, the left lung was harvested from vaccinated mice in each group. Each lung was washed three times in phosphate-buffered saline (PBS), weighed, placed in PBS (1 ml per 0.1 g of tissue) containing 0.1% bovine serum albumin (BSA; Huashengyili, Beijing, China), and homogenized with a glass tissue grinder. The homogenates were centrifuged (10,000g for 5 min) and the supernatant was analyzed for IgA by ELISA as previously described .
Analysis of RSV-F-specific antibody production
Blood was obtained from the retro-orbital plexus using a capillary tube and collected in an Eppendorf tube. After centrifugation (5000g for 15 min), sera were stored at −20°C. RSV-F-specific antibody responses in immunized mice were measured by ELISA as previously described [7, 20]. In brief, RSV was adsorbed onto ELISA plates overnight in carbonate buffer (pH 9.8) at 4°C. The plates were blocked with 5% fetal bovine serum (FBS) in PBS for 2 h at 37°C. After thorough washing with 1% BSA in PBS, serum or homogenate was added to the plate and incubated for 1 h at 37°C. After further washing, the plates were incubated for 1 h with horseradish peroxidase (HRP)-conjugated anti-mouse IgA (1:500 dilution), IgG1 (1:5000 dilution), IgG2a (1:5000 dilution), or IgG (1:5000 dilution) antibodies (Santa Cruz). The plates were developed with 100 μl of tetramethyl benzidine (TMB; Sigma, St. Louis, MO, USA) substrate solution, stopped with 50 μl of 2 mol/L H2SO4, and analyzed at 450 nm using an ELISA plate reader (Tecan, Grödig, Austria).
RSV-specific neutralizing antibody assay
To analyze the RSV-specific neutralizing antibody titer, serum samples were heat-inactivated at 56°C for 30 min. Serial two-fold dilutions of the mouse sera were prepared in virus diluent (Eagle’s minimal essential medium with L-glutamine containing 2% FBS, 2.5% HEPES (1 mol/L) and 1% antibiotic/antimycotic). To each serial diluted sample, 50 pfu of RSV virus suspension was added and incubated at 37°C for 1 h. RSV-specific neutralizing antibody titers were analyzed using the immunoenzyme assay described above. Neutralization titers are expressed as the reciprocal of the serum dilution giving a 50% reduction in pfu number in control wells.
CD8+ T cell responses to RSV F
To determine the number of cytokine-producing cells, IFN-γ was measured using an ELISPOT kit (BD Biosciences, San Diego, CA, USA) according to the manufacturer’s instructions. In brief, 2×105 freshly isolated splenocytes were added to each of three replicate wells coated with purified anti-mouse IFN-γ monoclonal antibody and stimulated with peptides F85–93 and F249–258 (10 μg/mL), corresponding to two known H-2 Kd-restricted RSV F protein epitopes (KYKNAVTEL  and TYMLTNSEL , respectively; purity ≥95%), for 24 h at 37°C in a 5% CO2 incubator. Unstimulated splenocytes were used to measure background cytokine production. The cells were then lysed with deionized water and the plates were incubated at room temperature with biotinylated IFN-γ antibody for 2 h and peroxidase-labeled streptavidin for 1 h. After washing with PBS, 100 μL of the final substrate solution (BD Biosciences) was added to each well and spot development was monitored. The plates were washed with distilled water to terminate the reaction. IFN-γ spot-forming cells (SFCs) were counted automatically using a CTL ELISPOT reader (BD Biosciences) and analyzed using ImmunoSpot image analyzer software v4.0 (BD Biosciences).
RSV titer in lungs
Mice were sacrificed on day 5 after the challenge. The left lung was harvested from mice in each group, weighed, placed in sterile stabilizing buffer (1 ml/g lung), and homogenized with a glass tissue grinder. The homogenates were centrifuged (10,000g for 1 min) and the RSV titers in the supernatants were measured by immunoenzyme assay as follows. HEp-2 cells (2×104 cells/well) were added to 96-well tissue culture plates (Corning Incorporated, Corning, USA) and incubated overnight at 37°C After removal of the medium, serial ten-fold dilutions of the clarified homogenate supernatants were adsorbed onto cell monolayers that had been washed with DMEM without serum in duplicate. Plates were incubated at 37°C for 45 min, the supernatant was removed, and 100 μl of DMEM containing 1.0% carboxymethyl cellulose (Sigma) was added to each well. After 3 days of incubation at 37°C, the overlay was removed, the cells were washed with PBS, and the monolayers were fixed with 1 ml of cold 95% alcohol for 10 min. After washing, plates were blocked in 5% nonfat dry milk in PBS for 1 h at RT. Anti-F monoclonal antibody (1:200 dilution) was added to the wells and incubated for 2 h, followed by 1 h incubation with HRP-conjugated anti-mouse IgG antibodies (1:5000 dilution in 5% nonfat milk; Santa Cruz) at 37°C. After washing twice with PBS containing 0.5% Tween 20 (Sigma), the plates were developed with 100 μl of TMB substrate solution. Individual plaques were counted under an inverted microscope and expressed as pfu.
Statistical analyses were performed using SPSS 11.5 software (SPSS, Chicago, IL, USA). Differences were compared using the Tukey test. P<0.05 was considered statistically significant.
Characterization of antibody responses induced by HDAd-Fsyn
Induction of cellular immune responses
Protective efficacy of HDAd-Fsyn against RSV challenge
Durability of antibody responses after immunization
Several studies have shown that HDAd vectors are superior to FGAd vectors for vaccine development in terms of their ability to produce immune responses against transgenes and are a robust platform for vaccination [12–14]. However, as pioneering explorations, most of the studies only investigated humoral and cellular immune responses induced by HDAd vector vaccines. There is only one report involving rectal HIV challenge following intramuscular administration of a HDAd vector vaccine encoding the envelope antigen . This situation prompted us to explore the protective effect following mucosal vaccination with HDAd vector vaccine. Our study provides the first demonstration of in vivo immunogenicity and efficacy of i.n. HDAd-Fsyn vaccination. The HDAd-Fsyn mucosal vaccine induced long-lasting RSV-F-specific cellular and humoral immune responses and protected against RSV infection in BALB/c mice on challenge. The reduction in viral load observed was approximately 1100-fold after prime–boost immunization in mice after i.n. HDAd-Fsyn vaccination when compared with the control group. This indicates that the HDAd vector is a very efficient and convenient mucosal vaccine platform in a murine model.
A few studies have addressed stimulation of high-titer antibodies against transgenes by HDAd vector vaccines . However, none of the studies investigated the neutralizing activity of these antibodies [12–14]. The presence of serum neutralizing antibodies is vital in protecting against viral infection. Our results clearly show that HDAd-Fsyn is capable of inducing serum neutralizing antibodies against RSV, indicating that HDAd-Fsyn has promise as a candidate RSV vaccine.
Long-lasting protective immunity is indispensable for a potential vaccine. McGinnes et al. explored the potential of virus-like particles (VLPs) as an RSV vaccine candidate and found that mice immunized with VLPs generated antibodies that had neutralizing activity up to 120 days. We evaluated RSV-F-specific immunity and the duration of the immune response induced by a single dose of i.n. HDAd-Fsyn vaccine. We found that HDAd-Fsyn evoked RSV-F-specific serum neutralizing antibody responses lasting for at least 14 weeks. Two mechanisms may be responsible for the transgene-specific immunity: (i) HDAd vaccines can express the transgene durably in vivo and achieve a high level of transgene protein in infected dendritic cells (DCs) [13, 24]; and (ii) HDAd vaccines can produce lower anti-Ad T-cell responses . After i.n. RSV challenge, the lower virus titers in lung tissues from vaccinated mice demonstrated induction of efficient protective immunity compared with the control group.
In summary, HDAd-Fsyn mucosal vaccine induced RSV-F-specific cellular and humoral immune responses and protected against subsequent i.n. RSV infection in a murine model. Our results suggest that HDAd has promise as a candidate i.n. RSV vaccine vector. Further development, such as co-expression of protective antigens or a combination with a DC-targeting strategy, might significantly enhance the efficacy of these vectors.
Animal experiments that is reported in the manuscript has been performed with the approval of the Committee on the Ethics of Animal Experiments of Beijing Jiaotong University (Permit number: 2010–0013). All surgeries were performed under sodium pentobarbital anesthesia, and all efforts were mase to minimize suffering.
We thank Dr. Chang-Xin Shi (Mayo Clinic, Scottsdale AZ) for technical assistance and for help in manuscript preparation. This work was supported by the Natural Science Foundation of China (grants 30671965 and 30972604), Fundamental Research Funds for the Central Universities (grants 2009JBM104 and S12JB00120), the Scientific and Technological Major Special Project of China (grant 2009ZX10004-216), the Beijing Municipal Natural Science Foundation (grant 7092053), and the China Postdoctoral Science Foundation (grant 20100470196).
- Graham BS: Biological challenges and technological opportunities for respiratory syncytial virus vaccine development. Immunol Rev 2011, 239: 149-166. 10.1111/j.1600-065X.2010.00972.xPubMedPubMed CentralView ArticleGoogle Scholar
- Hall CB: Respiratory syncytial virus in young children. Lancet 2010, 375: 1500-1502. 10.1016/S0140-6736(10)60401-1PubMedView ArticleGoogle Scholar
- Collins PL, Crowe JE: Respiratory syncytial virus and metapneumovirus. In Fields virology 5th edition. Edited by: Knipe DM, Howley PM, Griffin DE, Lamb RA, Martin MA, Roizman B, Straus SE. 2007, 1601-1646.Google Scholar
- Karron RA: Respiratory syncytial virus and parainfluenza virus vaccines. In Vaccines. 5th edition. Edited by: Plotkin SA, Orenstein WA, Offit P. Philadelphia: Saunders, Elsevier; 2008:1283-1293.Google Scholar
- Karron RA, Wright PF, Belshe RB, Thumar B, Casey R, Newman F, Polack FP, Randolph VB, Deatly A, Hackell J, et al.: Identification of a recombinant live attenuated respiratory syncytial virus vaccine candidate that is highly attenuated in infants. J Infect Dis 2005, 191: 1093-1104. 10.1086/427813PubMedView ArticleGoogle Scholar
- Power UF: Respiratory syncytial virus (RSV) vaccines–two steps back for one leap forward. J Clin Virol 2008, 41: 38-44. 10.1016/j.jcv.2007.10.024PubMedView ArticleGoogle Scholar
- Fu Y, He J, Zheng X, Wu Q, Zhang M, Wang X, Wang Y, Xie C, Tang Q, Wei W, et al.: Intranasal immunization with a replication-deficient adenoviral vector expressing the fusion glycoprotein of respiratory syncytial virus elicits protective immunity in BALB/c mice. Biochem Biophys Res Commun 2009, 381: 528-532. 10.1016/j.bbrc.2009.02.075PubMedView ArticleGoogle Scholar
- Kohlmann R, Schwannecke S, Tippler B, Ternette N, Temchura VV, Tenbusch M, Uberla K, Grunwald T: Protective efficacy and immunogenicity of an adenoviral vector vaccine encoding the codon-optimized F protein of respiratory syncytial virus. J Virol 2009, 83: 12601-12610. 10.1128/JVI.01036-09PubMedPubMed CentralView ArticleGoogle Scholar
- Kim S, Jang JE, Yu JR, Chang J: Single mucosal immunization of recombinant adenovirus-based vaccine expressing F1 protein fragment induces protective mucosal immunity against respiratory syncytial virus infection. Vaccine 2010, 28: 3801-3808. 10.1016/j.vaccine.2010.03.032PubMedView ArticleGoogle Scholar
- Abbink P, Lemckert AA, Ewald BA, Lynch DM, Denholtz M, Smits S, Holterman L, Damen I, Vogels R, Thorner AR, et al.: Comparative seroprevalence and immunogenicity of six rare serotype recombinant adenovirus vaccine vectors from subgroups B and D. J Virol 2007, 81: 4654-4663. 10.1128/JVI.02696-06PubMedPubMed CentralView ArticleGoogle Scholar
- Thorner AR, Vogels R, Kaspers J, Weverling GJ, Holterman L, Lemckert AA, Dilraj A, McNally LM, Jeena PM, Jepsen S, et al.: Age dependence of adenovirus-specific neutralizing antibody titers in individuals from sub-Saharan Africa. J Clin Microbiol 2006, 44: 3781-3783. 10.1128/JCM.01249-06PubMedPubMed CentralView ArticleGoogle Scholar
- Weaver EA, Nehete PN, Buchl SS, Senac JS, Palmer D, Ng P, Sastry KJ, Barry MA: Comparison of replication-competent, first generation, and helper-dependent adenoviral vaccines. PLoS One 2009, 4: e5059. 10.1371/journal.pone.0005059PubMedPubMed CentralView ArticleGoogle Scholar
- Harui A, Roth MD, Kiertscher SM, Mitani K, Basak SK: Vaccination with helper-dependent adenovirus enhances the generation of transgene-specific CTL. Gene Ther 2004, 11: 1617-1626. 10.1038/sj.gt.3302332PubMedView ArticleGoogle Scholar
- Weaver EA, Nehete PN, Nehete BP, Buchl SJ, Palmer D, Montefiori DC, Ng P, Sastry KJ, Barry MA: Protection against mucosal SHIV challenge by peptide and helper-dependent adenovirus vaccines. Viruses 2009, 1: 920. 10.3390/v1030920PubMedPubMed CentralView ArticleGoogle Scholar
- Fu YH, He JS, Zheng XX, Wang XB, Xie C, Shi CX, Zhang M, Tang Q, Wei W, Qu JG, Hong T: Intranasal vaccination with a helper-dependent adenoviral vector enhances transgene-specific immune responses in BALB/c mice. Biochem Biophys Res Commun 2010, 391: 857-861. 10.1016/j.bbrc.2009.11.152PubMedView ArticleGoogle Scholar
- Wang XB, He JS, Fu YH, Zheng XX, Fang X: Research on the methods for titrating respiratory syncytial virus. Zhonghua Shi Yan He Lin Chuang Bing Du Xue Za Zhi 2010, 24: 147-149.PubMedGoogle Scholar
- Shi CX, Graham FL, Hitt MM: A convenient plasmid system for construction of helper-dependent adenoviral vectors and its application for analysis of the breast-cancer-specific mammaglobin promoter. J Gene Med 2006, 8: 442-451. 10.1002/jgm.867PubMedView ArticleGoogle Scholar
- Ng P, Parks RJ, Graham FL: Preparation of helper-dependent adenoviral vectors. Methods Mol Med 2002, 69: 371-388.PubMedGoogle Scholar
- Hillgenberg M, Schnieders F, Loser P, Strauss M: System for efficient helper-dependent minimal adenovirus construction and rescue. Hum Gene Ther 2001, 12: 643-657. 10.1089/104303401300057342PubMedView ArticleGoogle Scholar
- Yuan-hui F, Jin-sheng H, Xiao-bo W, Xian-Xian Z, Qiang W, Can X, Mei Z, Wei W, Qian T, Jing-dong S, Jian-guo Q, Tao H: A prime–boost vaccination strategy using attenuated Salmonella typhimurium and a replication-deficient recombinant adenovirus vector elicits protective immunity against human respiratory syncytial virus. Biochem Biophys Res Commun 2010,395(1):87-92. 10.1016/j.bbrc.2010.03.144View ArticleGoogle Scholar
- Chang J, Srikiatkhachorn A, Braciale TJ: Visualization and characterization of respiratory syncytial virus F-specific CD8(+) T cells during experimental virus infection. J Immunol 2001, 167: 4254-4260.PubMedView ArticleGoogle Scholar
- Johnstone C, de Leon P, Medina F, Melero JA, Garcia-Barreno B, Del Val M: Shifting immunodominance pattern of two cytotoxic T-lymphocyte epitopes in the F glycoprotein of the Long strain of respiratory syncytial virus. J Gen Virol 2004, 85: 3229-3238. 10.1099/vir.0.80219-0PubMedView ArticleGoogle Scholar
- McGinnes LW, Gravel KA, Finberg RW, Kurt-Jones EA, Massare MJ, Smith G, Schmidt MR, Morrison TG: Assembly and immunological properties of Newcastle disease virus-like particles containing the respiratory syncytial virus F and G proteins. J Virol 2011, 85: 366-377. 10.1128/JVI.01861-10PubMedPubMed CentralView ArticleGoogle Scholar
- Roth MD, Cheng Q, Harui A, Basak SK, Mitani K, Low TA, Kiertscher SM: Helper-dependent adenoviral vectors efficiently express transgenes in human dendritic cells but still stimulate antiviral immune responses. J Immunol 2002, 169: 4651-4656.PubMedView 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.