|Year : 2019 | Volume
| Issue : 2 | Page : 105-110
Latex agglutination test for rapid on-site serodiagnosis of Japanese encephalitis in pigs using recombinant NS1 antigen
MR Grace1, Himani Dhanze1, Pranita Pantwane1, M Sivakumar1, Baldev Raj Gulati2, Ashok Kumar3
1 Division of Veterinary Public Health, ICAR-Indian Veterinary Research Institute, Izatnagar, Uttar Pradesh, India
2 ICAR-National Research Centre on Equines, Hisar, India
3 Indian Council of Agricultural Research, New Delhi, India
|Date of Submission||20-Jan-2018|
|Date of Acceptance||02-Apr-2018|
|Date of Web Publication||31-Jul-2019|
Division of Veterinary Public Health, ICAR-Indian Veterinary Research Institute, Izatnagar, Uttar Pradesh
Source of Support: None, Conflict of Interest: None
Background & objectives: Japanese encephalitis (JE) is a mosquitoe-borne viral zoonotic disease and globally around three billion people are at the risk of disease. The occurrence of JE cases has shown a rising trend during last decade in India. Pig is the amplifying host for JE virus and serves as a suitable sentinel model for the prediction of disease outbreak in humans. The development of a diagnostic test that is suitable for surveillance of JE in pigs is the need of the hour. The existing tests require elaborate laboratory facilities which make their application in rural settings difficult. Therefore, realizing the need for a rapid test, efforts were made to standardize a latex agglutination test (LAT) for serodiagnosis of JE in pigs.
Methods: Standardization of LAT by physical adsorption of recombinant NS1 (non-structural) protein of JE virus onto latex beads was done by altering six different variables, namely the antigen concentration, sensitization condition, surface blocking agent, blocking condition, particle concentration and reaction time. The standardized latex-protein complex was used for screening 246 pig serum samples under optimal conditions.
Results: The test was standardized with a diagnostic sensitivity and specificity of 82.24 and 87.83%, respectively. Screening of 246 field pig serum samples using standardized LAT showed a seropositivity of 50.4%. The results were available within 5 min after addition of test serum sample to the sensitized beads.
Interpretation & conclusion: The findings of the study highlight the potential of LAT as a rapid on-site assay for JE diagnosis in pigs which would aid in predicting JE outbreaks in humans.
Keywords: Japanese encephalitis; latex agglutination test; NS1 antigen; serodiagnosis
|How to cite this article:|
Grace M R, Dhanze H, Pantwane P, Sivakumar M, Gulati BR, Kumar A. Latex agglutination test for rapid on-site serodiagnosis of Japanese encephalitis in pigs using recombinant NS1 antigen. J Vector Borne Dis 2019;56:105-10
|How to cite this URL:|
Grace M R, Dhanze H, Pantwane P, Sivakumar M, Gulati BR, Kumar A. Latex agglutination test for rapid on-site serodiagnosis of Japanese encephalitis in pigs using recombinant NS1 antigen. J Vector Borne Dis [serial online] 2019 [cited 2022 May 20];56:105-10. Available from: https://www.jvbd.org/text.asp?2019/56/2/105/263717
| Introduction|| |
Japanese encephalitis (JE) is a re-emerging mosquito- borne zoonotic disease caused by Japanese encephalitis virus (JEV), which belongs to the genus Flavivirus of family Flaviviridae. The JEV is a small, enveloped virus around 50 nm in length, with a single-stranded, positivesense RNA genome of approximately 11,000 nucleotides. It is capable of infecting most vertebrates, although clinical disease is mainly limited to humans, pigs and horses,. More than 60,000 human cases are estimated to occur every year in the affected areas with case fatality rate ranging from 20–30% with neurologic or psychiatric sequelae in 30–50% of survivors.
India along with China influence JE figures on a global scale because the number of people living in JE endemic areas is concentrated in these two countries. In India, since the confirmation of first clinical case of JE from Vellore in 1955, the geographic area affected by JEV has expanded in the last 60 years with higher epidemic activity in the north and central India.
Pigs are the amplifier host of JEV since they allow manifold virus multiplication without suffering from disease and maintain prolonged viraemia for 3–5 days. Also, pigs seroconvert 2–3 wk before infection occurs in human and thus serve as a relevant sentinel model, the surveillance of which could predict a potential JE outbreak in a human population nearby. Antibody-based surveillance is preferred in case of JEV in pigs since there is no specific disease symptom and various tests like, virus neutralization test (VNT), haemagglutination inhibition (HI), immunofluorescence antibody assay (IFA) and enzyme-linked immunosorbent assay (ELISA) are currently being employed.
The VNT is the gold standard for serological diagnosis of flaviviral infections. Though specific, there are many disadvantages of this test like, labour intensive, requirement of skilled personnel, a minimum of five days to get results, the handling of live virus; and the accuracy of interpretation requires comparison of end-point titres of other flaviviruses endemic to the given area. Haemag-glutination inhibition for JE has been described for many years but it cannot be used for early diagnosis as it requires paired serum samples. Sensitivity to pH change, fragility of HA factor, need for continuous supply of avian eryth-rocytes and cross-reactive nature necessitating a simultaneous assessment of other flaviviruses endemic to the area are some of the other practical limitations of this test. Therefore, now it is not a preferred method for serodiag-nosis of JE,.
Immunofluorescence assay is used to differentiate the IgM and IgG responses to flaviviral infection. Cross-reactivity with closely related flaviviruses is a major disadvantage and there is requirement for a fluorescent microscope to evaluate the results. IgM and IgG capture ELISAs have been developed for JE which have become the accepted standard for diagnosis of Japanese encepha-litis. But since ELISA requires elaborate laboratory facilities and trained manpower, the test is more suited to diagnostic centres than to the rural settings. Therefore, recognizing the importance of developing a rapid diagnostic test for serosurveillance of JEV in pigs, which can be easily applied in field settings, efforts were made to standardize the LAT for JE.
The JEV contains three structural and seven non-structural polypeptides, which are encoded by a single long open reading frame (ORF). The 5’-one-third of the genome codes for the three structural proteins, namely the capsid protein [C], membrane protein [M], and the envelope protein [E], while rest 3’-two-third of the genome codes for the seven non-structural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5). The NS1 protein is secreted from the infected mammalian cells and is hence considered as the most suitable antigen for use in diagnostic techniques. The NS1 protein has been reported to solve the problem of cross-reactivity and can differentiate between JE and related viruses,. Therefore, the NS1 protein is considered a better alternative and was chosen as the recombinant protein for the present study.
| Material & Methods|| |
A total of 246 pig serum samples were collected from the Indian states of Uttar Pradesh, Maharashtra, Goa, Karnataka, Tamil Nadu, Kerala and from the Union Territory of Chandigarh during the period from August 2015 to April 2016. The samples were collected from apparently healthy animals having the age of >3 months. Samples were transported to the laboratory under chilled condition and the serum samples were stored at –20 °C until further use.
Preparation of antigen
The JEV non-structural recombinant protein (rNS1) expressed in pET32a vector previously at the Division of Veterinary Public Health, Indian Council of Agricultural Research–Indian Veterinary Research Institute (ICAR-IVRI) was selected as the antigen for coating the latex beads. Recombinant NS1 protein was produced in bulk and purified by nickel chelating affinity chromatography using imidazole gradient method. The SDS-PAGE analysis revealed that the recombinant NS1 protein of 59 kDa size started appearing with 40 mM imidazole concentration and increased steadily thereafter with maximum concentration appearing in II and III fraction of 100 mM imidazole, followed by gradual decline in subsequent fractions, i.e. 200 and 500 mM im-idazole concentrations [Figure 1]. The fractions of 40, 80 and 100 mM imidazole concentrations were pooled and dialyzed by sequential reduction in the concentration of urea and finally against phosphate buffered saline (PBS) for proper refolding of the proteins. The pooled protein was concentrated using Merck Millipore 30 K centrifugal filters and the concentration was estimated to be 750 μg/ml by Bradford reagent. The protein was stored at –80 °C for further use as antigen for standardizing LAT.
|Figure 1: SDS-PAGE revealing different elutes of purified rNS1 protein. Lane M: Protein marker; Lane 1: Protein flow through; Lane 2: Protein wash through; Lane 3: 40 mM imidazole elute 2; Lane 4: 40 mM imidazole elute 3; Lane 5: 80 mM imidazole elute 2; Lane 6: 80 mM imidazole elute 3; Lane 7 : 100 mM imidazole elute 2; Lane 8: 100 mM imidazole elute 3; Lane 9: 200 mM imidazole elute 2; Lane 10: 200 mM imidazole elute 3; and Lane 11: 500 mM imidazole elute 3.|
Click here to view
The hyperimmune serum which has been previously raised in rabbits against rNS1protein by the Division of Veterinary Public Health, ICAR-IVRI, was used in the present study, after taking due permission (No. 1-53/2012- 13-JD.Res dated 10/9/2013) from the Institute Animal Ethics Committee.
Standardization of LAT
Standardization and optimization of reaction conditions of LAT with rNS1 fusion protein of JEV was attempted using physical adsorption. Briefly, a 2.5% suspension of the latex beads (violet-dyed polystyrene microspheres with 0.80 μ diam; Polyscience, USA) were washed thrice using 0.06 M sodium carbonate-bicarbonate buffer by centrifugation at 4000 g for 5 min at 4 °C followed by re-suspension in the same buffer. Further the beads were coated with rNS1 antigen in different concentrations (250, 500 and 750 μg/ml) using 0.06 M sodium carbonate-bicarbonate buffer. The protein was incubated with the beads for 6 h at two different temperatures (room temperature and at 4 °C) and two different incubation conditions (with continuous mixing by keeping over a rocker, and under static condition without mixing). The protein-bead mixture after 6 h incubation was centrifuged and the supernatant was collected to determine the protein concentration using Bradford reagent. The pelleted beads were washed once using sterile PBS solution, pH 7.2 by re-suspension and mixing, followed by centrifugation.
Further, for finding the ideal blocking agent for the unbound sites on the latex beads, the pelleted beads were re-suspended in different blocking buffers as a 1% suspension in PBS containing 10 mg/ml of bovine serum albumin (BSA); PBS containing 20 mg/ml of BSA; PBS containing 20 mg/ml of BSA and 0.5% tween 20, and in PBS containing 10 mg/ml glycine. The latex beads were left overnight with constant shaking at three different temperature conditions (37, 4 °C and at room temperature). Latex beads were centrifuged as before and re-suspended as a 2.5% suspension in storage buffer and stored at 4 °C until used. The storage solutions used were either 0.05% BSA in PBS or 0.05% BSA in 0.01% PBST or 0.05% glycine in PBS in accordance with the blocking agent used. A panel of confirmed JE positive and JE negative serum was prepared by testing the collected pig serum samples using VNT.
The LAT was performed on glass slides using serial dilutions of hyperimmune serum raised in rabbits against the rNS1 antigen, the panel of confirmed JE positive pig serum, JE negative pig serum and PBS. Three different reaction conditions, using 5 μl beads and 5 μl serum; 5 μl beads and 10 μl serum and 10 μl beads and 10 μl serum were tested. The slide was rocked briefly for 2 min to mix the coated beads and serum samples. The reaction was read/visualized at three different time periods (5, 10 and 15 min) and the reaction conditions yielding best visual perception of agglutination were selected to further screen the samples using LAT .
The result of the test was graded according to the degree of agglutination reaction. The reactions where the mixture becomes completely transparent with strongly agglutinated particles, which tend to settle at the edge of the beads-serum mixtures was graded as strong positive (+ + +); reactions wherein most of the latex particles agglutinated but fluid was slightly opaque was graded as moderately positive (+ +); reactions wherein about 50% latex agglutinated but the liquid was opaque was graded as weak positive (+), and reactions where the mixture remained homogenous was graded as negative.
The storage life of beads was evaluated by periodic testing of the prepared beads. For this the beads were prepared and stored at 4 °C, and a specific set of serum samples out of the collected samples was aliquoted into small units and stored at –80 °C. These samples were periodically subjected to the developed LAT assay using the stored beads, and sensitivity and specificity was calculated as described by Thrusfield.
Diagnostic sensitivity and specificity
Out of 246 pig serum samples collected, 181 serum samples were screened using VNT; of which 107 were VNT positive and 74 were VNT negative. The same set of samples was screened using rNS1 LAT to access diagnostic efficacy of the test. The diagnostic efficacy of rNS1 protein-based LAT in terms of sensitivity, specificity and predictive value in comparison with VNT were calculated as described by Thrusfield.
| Results|| |
Optimal conditions for LAT
A series of experiments was performed to determine the optimal conditions for the LAT assay and to adjust the test sensitivity to an appropriate level. Among three different concentrations of the protein used (250, 500 and 750 μg/ml); 750 μg/ml showed highest sensitivity up to 1 : 16 dilution of hyperimmune serum [Figure 2]. Incubation of protein with the beads at 4 °C for 6 h with continuous mixing was found to yield the best results. Among the various blocking agents used, BSA at 1% concentration in PBS gave the least number of false positive reactions. Blocking at 4 °C was found to be better in terms of sensitivity of reaction when compared with blocking at 37 and 25 °C (room temperature).
|Figure 2: LAT by physical adsorption with: (a) Negative sample; and (b) Positive sample (Hyperimmune sera in 1 : 16 dilution).|
Click here to view
The best visual perception of agglutination was observed when 5 μl beads and 10 μl serum was reacted. The reaction time up to 5 min was found to be the optimum, with longer time leading to settling and drying of the mixture.
The storage life of the beads was evaluated by keeping the beads at 4 °C and checking at periodic intervals. There was no appreciable change in sensitivity when the beads were reacted after 15 days’ interval; but when tested after 40 days, reduction was observed in sensitivity and specificity.
Application of LAT for screeningfield samples
All the 246 pig serum samples were screened following the optimum reaction conditions for LAT [Table 1]. The reactions were graded from strong positive to negative as per the criteria described earlier [Figure 3]. Out of 246 pig serum samples, 124 were found to be positive for JE antibodies revealing an overall seroposi- tivity of 50.4%.
|Figure 3: LAT using field swine serum samples: (a) Negative field sample—The mixture remains opaque and no agglutination was observed; (b) Strong positive field sample—The particles agglutinated and migrated to the periphery leaving a clear solution; and (c) Positive field sample—Most of the particles have agglutinated but fluid is slightly opaque.|
Click here to view
Diagnostic efficacy of LAT
Diagnostic sensitivity, specificity, efficiency, positive and negative predictive values of LAT in comparison with VNT were 82.24, 87.83, 84.53, 90.72 and 77.38%, respectively [Table 2] and [Table 3].
|Table 2: Contingency table to calculate diagnostic efficacy of LAT in comparison with VNT|
Click here to view
|Table 3: Consolidated table showing diagnostic efficacy of LAT in comparison with VNT based on Table 2.|
Click here to view
| Discussion|| |
Development of a rapid diagnostic test for diagnosis and serosurveillance of JEV in pigs, which can be easily applied in field settings, is necessary to fill the gap between existing diagnostic methods and surveillance programmes. In the present study, Latex agglutination test was standardized for serodiagnosis of JE in pigs using recombinant NS1 protein. Physical adsorption of protein on to latex surface and its efficiency in agglutination reactions is determined by many factors. In the present study, six different variables, namely the antigen concentration, sensitization condition, surface blocking agent, blocking condition, particle concentration, and reaction time, that are important in controlling the physicochemical processes of immunological recognition and agglutination in latex particle assays were analyzed. Finally, the latexprotein complex was used for screening 246 pig serum samples under optimal conditions.
Violet-dyed polystyrene beads (0.80 μ diam), were used for standardization of LAT in the present study. The particle diameter of 0.2 to 0.9 μm is required for visual slide agglutination. The commercial latex beads are usually supplied in medium containing various inhibitors for the attachment of ligands. Thus, cleaning of beads was done by washing using carbonate bicarbonate buffer. This cleaning may not be sufficient to eliminate all the surfactant groups. However, maximum adsorption which requires totally clean particles is not required in slide agglutination test, so cleaning using carbonate bicarbonate buffer is sufficient,.
The study showed that the highest sensitivity was obtained at the highest protein concentration used (750 μg/ ml) and this is because the amount of total linked protein increases with the concentration of added protein thereby increasing the sensitivity. The best sensitization condition was observed when the protein was allowed to bind at 4 °C for 6 h with continuous mixing. However, many researchers have reported effective binding of protein at elevated temperatures such as room temperature, and 37 °C,,. In the present study, BSA at 1% concentration in PBS as the blocking agent gave the least number of false positive reactions however, when BSA was used in combination with tween 20, the number of false negative reactions increased. This might be due to the surfactant activity oftween 20 causing desorption of bound proteins and hence, decreasing the sensitivity. Glycine, though reported as the best blocking agent; in the present study caused non-specific agglutination even with PBS. The reason behind this remains obscure. The most ideal blocking condition was found to be 4 °C and this is in accordance with the findings of other researchers. The ideal reaction condition was the one in which 5 μl beads at 2.5% concentration and 10 μl serum reacted. At low particle concentration, the naturally occurring proteins found in serum samples can interfere with immunoassays and at high particle concentration there is decreased formation of immunocomplexes due to imbalance between antigen and antibody concentrations or due to increased steric hindrance or non-specific agglutination.
The immunoagglutination assay does not reach an end point and thus reaction time analysis is an important factor to consider when optimizing an assay. In the present study, it was found that, t=5 min was the most ideal. When the reaction time was increased, there was difficulty in interpreting the result due to settling of the particles and drying of the mixture leading to false positive interpretations.
The grading of this assay whether it is strong or intermediate or weak positive requires trained personnels in field conditions. However, while doing serosurveillance in pigs, differentiation between positive and negative samples will serve the purpose, which can easily be done.
The evaluation of storage life of the sensitized beads indicated that there was reduction in sensitivity when the beads were reacted after 40 days interval; though there was no appreciable change when the beads were reacted after 15 days interval. Researchers have reported reduction in activity of latex beads sensitized by physical adsorption due to the partial desorption of adsorbed protein that normally occurs during its storage.
The diagnostic efficacy of rNS1 protein-based LAT was calculated in comparison with VNT. This test with 82.24% diagnostic sensitivity and 87.83% diagnostic specificity has great potential to be developed as a robust rapid on-site diagnostic test for JE with some further refinement to enhance its stability. We succeeded in increasing the specificity of LAT which is the most challenging parameter of agglutination assays by using appropriate blocking buffer and reaction conditions.
In a study, LAT using purified whole virus antigen was described. The slow growing nature of JEV makes it difficult to be produced in bulk for use as antigen. Moreover, the preparation of whole virus antigen is costly and also associated with biohazard risk. The use of recombinant protein in the present study, overcomes the biohazard risk. Further, in that study, a total of 35 samples were screened using both HI and LAT, and the coincidence rate for the two methods was 91.4%. However, HI as a diagnostic test suffers from many shortcomings in terms of sensitivity, specificity and cross-reactivity. Therefore, in the present study we have used VNT for the comparison.
| Conclusion|| |
The present study was an attempt to develop a rapid on-site test for serodiagnosis of JE in pigs. The results indicate that LAT and NS1 antigen of JEV are potential candidates for the development of rapid on-site assay for serodiagnosis of JE in pigs. Further, this test can also be applied to screen the samples from humans and other species of animals for JE antibodies in the field conditions.
Conflict of interest
The authors declare that they have no conflict of interest.
| Acknowledgements|| |
The authors are thankful to the Director, ICAR-IVRI for providing necessary facilities. The work was carried out under the ICAR funded project ‘Outreach Programme on Zoonotic Diseases’.
| References|| |
Singh A, Saxena SK, Srivastava AK, Mathur A. Japanese encephalitis: A persistent threat. Proc Natl Acad Sci Sect B Biol Sci
Solomon T. Control of Japanese encephalitis-Within our grasp? N Engl J Med
2006; 355(9): 869-71.
Wang H, Liang G. Epidemiology of Japanese encephalitis: Past, present, and future prospects. Ther Clin Risk Manag
Erlanger TE, Weiss S, Keiser J, Utzinger J, Wiedenmayer K. Past, present, and future of Japanese encephalitis. Emerg Infect Dis
2009; 15(1): 1-7.
Williams DT, Daniels PW, Lunt RA, Wang LF, Newberry KM, Mackenzie JS. Experimental infections of pigs with Japanese encephalitis virus and closely related Australian flaviviruses. Am J Trop Med Hyg
Flohic GL, Porphyre V, Barbazan P, Gonzalez JP. Review of climate, landscape, and viral genetics as drivers of the Japanese encephalitis virus ecology. PloS Negl Trop Dis
2013; 7(9): e2208.
Hobson-Peters J. Approaches for the development of rapid se- rological assays for surveillance and diagnosis of infections caused by zoonotic flaviviruses of the Japanese encephalitis virus serocomplex. JBiomedBiotechnol
Solomon T, Dung NM, Kneen R, Gainsborough M, Vaughn DW, Khanh VT. Japanese encephalitis. J Neurol Neurosurg Psychiatry
Laboratory diagnosis of JE virus infection. In: Manual for the laboratory diagnosis of Japanese encephalitis virus infection
. Geneva: World Health Organization 2007; p. 29-30.
Kabilan L, Rajendran R, Arunachalam N, Ramesh S, Sriniva- san, Philip P, et al
. Japanese encephalitis in India: An overview. Indian JPediatr
2004; 71(7): 609-15.
Konishi E, Shoda M, Ajiro N, Kondo T. Development and evaluation of an enzyme-linked immunosorbent assay for quantifying antibodies to Japanese encephalitis virus nonstructural 1 protein to detect subclinical infections in vaccinated horses. J Clin Microbiol
Kitai Y, Shoda M, Kondo T, Konishi E. Epitope-blocking enzyme-linked immunosorbent assay to differentiate West Nile virus from Japanese encephalitis virus infections in equine sera. Clin Vaccine Immunol
2007; 14(8): 1024-31.
Shu PY, Chen LK, Chang SF, Yueh YY, Chow L, Chien LJ, et al
. Antibody to the nonstructural protein NS1 of Japanese encephalitis virus: Potential application of mAb-based indirect ELISA to differentiate infection from vaccination. Vaccine
2001; 19(13-14): 1753-63.
Dhanze H, Bhilegaonkar KN, Rawat S, Kumar A, Chethan HB. Recombinant NS1 protein based indirect IgG ELISA for sero- surveillance of Japanese encephalitis, and kit thereof. Patent application No. 201611024016. New Delhi, India: Indian Patent Office 2016.
Grace MR, Dhanze H, Pantawane PB, Sivakumar M, Kumar A. Seropositivity of Japanese encephalitis virus in swine using Virus Neutralization Test. J Vet Public Health
2016; 141(1): 9-12.
Thrusfield M. Veterinary epidemiology
. III edn. London, UK: Blackwell Science Ltd. 2005; p. 313-6.
Molina-Bolivar JA, Galisteo-Gonzalez F. Latex immunoagglutination assays. J Macromolecular Sci (Part C: Polym Rev)
Sumithra TG, Chaturvedi VK, Gupta PK, Sunita SC, Rai AK, Kutty MVH, et al
. Development of a simple and rapid method for the specific identification of organism causing anthrax by slide latex agglutination. Lett Appl Microbiol
Yap KL. Development of a slide latex agglutination test for rotavirus antigen detection. Malaysian JPathol
1994; 16(1): 49-56.
Kasempimolporn S, Saengseesom W, Lumlertdacha B, Sit- prija V. Detection of rabies virus antigen in dog saliva using a latex agglutination test. J Clin Microbiol
Garcia VS, Gonzalez VDG, Caudana PC, Vega JR, Marcipar IS, Gugliotta LM. Synthesis of latex-antigen complexes from single and multiepitope recombinant proteins: Application in immunoagglutination assays for the diagnosis of Trypanosoma cruzi
infection. Colloid Surf Biointerfaces
Dey S, Mohan CM, Ramadass P, Nachimuthu K. Recombinant antigen-based latex agglutination test for rapid serodiagnosis of leptospirosis. Vet Res Commun
2007; 31(1): 9-15.
Xinglin J, Huanchun C, Qiagi H, Xiang W, Bin W, Dexin Q. The development and application of the latex agglutination test to detect serum antibodies against Japanese encephalitis virus. Vet Res Commun
2002; 26(6): 495-503.
Garcia VS, Gonzalez VDG, Marcipar IS, Vega JR, Gugliotta LM. Optimisation and standardization of an immunoagglutination assay for the diagnosis of Trypanosoma cruzi
infection based on latex- (recombinant antigen) complexes. Trop Med Int Health
2014; 19(1): 37-46.
Tripathi NK, Kumar JS, Biswal KC, Rao PVL. Production of recombinant nonstructural 1 protein in Escherichia coli
for early detection of Japanese encephalitis virus infection. Microb Bio- technol
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3]