A collaborative study of an alternative in vitro potency assay for the Japanese encephalitis vaccine

Virus Research 223 (2016) 190–196

Contents lists available at ScienceDirect

Virus Research journal homepage: www.elsevier.com/locate/virusres

A collaborative study of an alternative in vitro potency assay for the Japanese encephalitis vaccine Byung-Chul Kim a,b,1 , Do-Keun Kim a,c,1 , Hyung-Jin Kim a , Seung-Hwa Hong a , Yeonhee Kim a , Jong-Mi Lim a , JiYoung Hong a , Cheol-Hee Kim b , Yong-Keun Park c , Jaeok Kim a,∗ a

National Center for Lot Release, Ministry of Food and Drug Safety, Chungcheongbuk-do 28159, Republic of Korea Department of Biology, Chungnam National University, Daejeon 34134, Republic of Korea c School of Life Sciences and Biotechnology, Korea University, Seoul 02841, Republic of Korea b

a r t i c l e

i n f o

Article history: Received 25 January 2016 Received in revised form 24 July 2016 Accepted 29 July 2016 Available online 3 August 2016 Keywords: Japanese encephalitis vaccines in vivo in vitro ELISA Plaque reduction neutralization test

a b s t r a c t The use of inactivated Japanese encephalitis (JE) vaccines has been ongoing in East Asia for 40 years. A mouse immunogenicity assay followed by a Plaque Reduction Neutralization (PRN) Test (PRNTest) is currently recommended for each lot release of the vaccine by many national authorities. We developed an alternative in vitro ELISA to determine the E antigen content of the Japanese encephalitis virus to observe the 3Rs strategy. A collaborative study for replacing the in vivo potency assay for the Japanese encephalitis vaccine with the in vitro ELISA assay was confirmed comparability between these two methods. The study demonstrated that an in vitro assay could perform faster and was more convenient than the established in vivo PRNTest. Moreover, this assay had better precision and reproducibility compared with the conventional in vivo assay. Additionally, the content of antigen determined using the in vitro ELISA correlated well with the potency of the in vivo assay. Furthermore, this method allowed discrimination between individual lots. Thus, we propose a progressive switch from the in vivo assay to the in vitro ELISA for JE vaccine quality control. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Japanese encephalitis (JE) is the cause of viral encephalitis in most Asian countries and parts of the western Pacific (Fischer et al., 2008; Halstead and Jacobson, 2008; Fischer et al., 2009; Wilder-Smith and Freedman, 2009; Misra and Kalita, 2010). JE has a case-fatality ratio of approximately 20–30%, and 30–50% of survivors have neurologic or psychiatric sequelae (Solomon et al., 2000; WHO, 2006). The JE vaccine is the only method available to prevent this disease, as no specific antiviral agent or other medication currently exists (Gould et al., 2008). Thus, the mass vaccination of children has led to a decrease in the number of JE cases in some countries (Arai et al., 2004; Gupta et al., 2008). Therefore, the importance of the vaccine is paramount. The following three types of JE

Abbreviations: JE, Japanese encephalitis; PRN, plaque reduction neutralization; ELISA, enzyme-linked immunosorbent assay; CEC, chick embryo cell; EMEM, Eagle’s minimum essential medium; CV, coefficient of variation. ∗ Corresponding author. E-mail address: [email protected] (J. Kim). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.virusres.2016.07.012 0168-1702/© 2016 Elsevier B.V. All rights reserved.

vaccines are licensed in the world: inactivated mouse brain-derived vaccines, inactivated Vero cell culture-derived vaccines and liveattenuated primary hamster kidney cell culture-derived vaccines (Halstead and Thomas, 2011). The World Health Organization recommends that a quality control test for each individual lot of licensed vaccine be performed by the manufacturer and by a national control laboratory prior to approving its release onto the market (WHO, 1992; WHO, 2010). It is essential to ensure the consistent quality of each lot, as vaccines are biological products used in a healthy population, and the safety and efficacy of the vaccine have a drastic impact on many individuals (Griffiths, 1996). Furthermore, vaccines have an inherent potential for variability. Therefore, independent testing that is based on a validated methodology is essential to ensure the consistent quality of each lot (Hendriksen et al., 1998). Currently, the potency test used as a lot release test for the inactivated JE vaccine is an in vivo assay that compares the amount of vaccine necessary to induce neutralizing antibodies in mice equated with the amount of reference preparation necessary to produce the same effect (WHO, 1998; Ashok and Rangarajan, 2000; MHW, 2006; WHO, 2007; MFDS, 2011). However, there are inher-

B.-C. Kim et al. / Virus Research 223 (2016) 190–196

ent problems with this assay. First, this method is time consuming, as there is a required period of time for the murine immunization and the preparation of the plaque reduction neutralization assay (Hendriksen, 2007). In addition, there are animal ethical issues and deviations of test results from the condition of the animal or the cultured cells or variability due to the proficiency of the technician. Moreover, the global trend is to discourage the use of experimental animals; thus, many developed countries are reducing the number of animals utilized in quality control assays of vaccines through the development of in vitro assays (De Mattia et al., 2011; Isbrucker et al., 2011; Shin et al., 2011). In Korea, it is estimated that the number of mice used in the national lot release tests for JE vaccines are approximately 2000 per year, with more than 10,000 mice used annually when including the manufacturers’ quality control tests. The Ministry of Food and Drug Safety (MFDS) developed an in vitro enzyme-linked immunosorbent assay (ELISA) potency test for the JE vaccine to solve the above problems and to join the global trend in a 3Rs strategy (Refinement, Reduction and Replacement of animal testing) (Hendriksen, 2002). Our previous study suggested that an alternative in vitro method for the measurement of the amount of the JE E-antigen be used, which consisted of an ELISA kit that included neutralizing monoclonal capture and detection of antibodies mapped to specific epitopes (Kim et al., 2012). Herein, we describe a study performed by the MFDS to assess the feasibility of replacing the in vivo assay with the in vitro ELISA for the lot release test, including a comparison between these two tests and collaborative tests with the manufacturers. 2. Materials and methods 2.1. Vaccine samples Two different inactivated liquid JE vaccines (labeled Samples A and B) commercially available in Korea containing the Nakayama strain derived from mouse brain were used for all tests with four participants. To obtain the reduced-potency vaccine samples, a vaccine lot was treated by incubating aliquots at 4 ◦ C, 25 ◦ C, 37 ◦ C and 56 ◦ C for 3 weeks. 2.2. Standard The standard for this study was a Korean national reference standard (code: 07/022) with an assigned titer of 2.661 log PFU/vial of antigen. This standard contained the inactivated and lyophilized Nakayama strain was established in 2007 by MFDS similarly to the Korean commercial vaccine. This was used to assess both the in vitro ELISA and the immunogenicity, via the in vivo animal PRN test, of the vaccine samples. 2.3. In vivo immunogenicity potency assay The in vivo potency assay was performed for two different commercial vaccines in accordance with the Korean Biologics Specification and Test Methods. Briefly, each group of 10, 4-week-old ICR (Institute of Cancer Research) strain mice was immunized subcutaneously on days 0 and 7 with 0.5 mL of standard or vaccine sample diluted in phosphate buffered saline. One week later, the mice were bled from the heart, and serum pooled from each mouse was prepared. The neutralizing antibody titer was measured with a 50% plaque-reduction end-point assay test in 60 mm culture dishes containing a monolayer of primary chick embryo cells (CECs), as reported previously (Oya et al., 1967; Gowal et al., 2010). These sera were diluted 1: 320 in 1 x Eagle’s minimum essential medium (EMEM) with 5% fetal bovine serum (FBS) and were mixed with 100 PFU of the challenge virus followed by incubation for 90 min at

191

36 ± 1 ◦ C. The incubated samples were added onto 4 culture dishes containing the CEC monolayer, and the cells were covered with 0.9% agarose overlay medium containing 1 x EMEM with 7% FBS. Following incubation for 48 h at 37 ◦ C, the covered cells were overlaid again with agarose containing 1 x EMEM and 0.02% neutral red. After incubation for 1–2 days at 37 ◦ C, the plaque numbers were counted, and the 50% reduction in potency was calculated as: Z = (Y-50)/47.7622 + log X, where Z: NT-Ab titer (log10), Y: plaque reduction rate (%), Y = 10–90% and X: reciprocal of serum dilution used (Abe et al., 2003; Borges et al., 2008). The plaque numbers of 10 culture dishes inoculated with 100 PFU of the challenge virus were counted concurrently to confirm the titer of the virus. The result has the unit of logPFU/mL 2.4. In vitro ELISA Anti-JEV ELISA kits were produced by JENO Biotech Inc., a Korean diagnostic medicine manufacturer. The kits consisted of an ELISA 96-well plate with 2 ␮g/mL of capture mAb for the JEV E-protein immobilized on the plate, wash solution (phosphate buffered saline containing 0.1% Tween 20), dilution buffer (wash solution containing 0.1% bovine serum albumin), Horseradish peroxidase (HRP)-labeled detector mAb, TMB (3,3 ,5,5 -tetramethylbenzidine) substrate, stop solution (2N sulfuric acid) and positive (1: 80 dilutions of national reference standard with tris-buffered saline and Tween 20 containing 0.1% bovine serum albumin) and negative controls (1 × EMEM) (Kim et al., 2012). The ELISA was performed as directed by the manufacturer’s manual. Briefly, serial ten-fold dilutions of the national reference standard and the commercial JEV vaccines with dilution buffer were added to the microplate (final volume 100 ␮L). The microplate was incubated with these samples, including the positive and negative controls, at 37 ± 1 ◦ C for 1 h ± 5 min. After incubation, the microplate was washed three times with wash solution and was incubated with HRP-labeled detector mAb for 1 h ± 5 min at 37 ± 1 ◦ C (final volume 100 ␮L). After washing three times, the microplate was incubated in the dark with 100 ␮L/well of TMB substrate. After 15 min, the reaction was stopped with 50 ␮L/well of stop solution. The absorbance was measured at 450 nm (reference wavelength 650 nm) using an ELISA plate reader (Molecular Devices, USA). The result has the unit of ELISA Unit (E.U.)/mL. 2.5. Method validation of in vitro ELISA 2.5.1. Specificity Dilution buffer, negative control, positive control, national reference standard and two commercial JE vaccines were assayed to confirm specificity for the E antigen of JEV. The national reference standard and the two JE vaccines were diluted to 1: 80 with dilution buffer. 2.5.2. Linearity Six concentrations of the value ranging from 1: 10–1: 320 were tested with three independent determinations to obtain the slope and correction coefficient (R2 ). 2.5.3. Accuracy The estimation of the accuracy was performed using the linearity result. The national reference standard was used to artificially determine the ELISA Unit (E.U.) to 1331 mE.U./mL. 2.5.4. Precision The precision was established to assess assay repeatability and intermediate precision. To determine the repeatability, three concentrations ranging from 1: 40–1: 160 were tested with three wells

192

B.-C. Kim et al. / Virus Research 223 (2016) 190–196

Fig. 1. Linearity of the in vitro ELISA method.

A: Intercept; B: Slope; R2: Correction coefficient. (a) 1st, (b) 2nd, (c) 3rd.‘

per plate by two technicians. The intermediate, inter-day and interanalyst precisions were assessed following a similar method.

Table 1 The specificity of the ELISA. Samples

Dilution

Dilution buffer Negative control Positive control Standard Vaccine (company A) Vaccine (company B)

– – – 80 80 80

2.6. Correlation study

O.D.

Mean

1st

2nd

3rd

0.044 0.047 0.274 0.276 0.568 0.681

0.043 0.046 0.268 0.269 0.567 0.659

0.042 0.047 0.27 0.252 0.557 0.671

0.0430 0.0467 0.2707 0.2657 0.5640 0.6703

The relationships between the incubation temperature and the in vitro or in vivo titers were assessed using regression analysis. Moreover, the corresponding coefficient of determination (R2 ) was calculated. To compare the methods, 4 participants were requested to perform 5 independent assays with 3 lots with both the in vivo and in vitro assays. For another study, five lots were analyzed with the in vivo and in vitro assays by the MFDS and participant A, and the remaining 5 lots were performed by the MFDS and participant B. Five lots from company A were produced from three different final bulk lots. Lot numbers 1 and 2 were filled from the same final bulk lot. Lot numbers 3 and 4 were from a different final bulk lots, and lot number 5 was from a third final bulk lot. Lot numbers 1–4 of company B were produced from the same final bulk lot, while lot number 5 was from a different final bulk lot. The relationship between the in vitro and in vivo titers was estimated via the analysis of 60 data points (3 lots x 5 times in 4 participants) and 20 data points (10 lots x MFDS and participant A or B). The participants referred to in this article are alphabetically listed in the Appendix A but are arbitrarily referred to as participants A or B.

the 1: 80 dilutions of the national reference standard and the 2 vaccines exhibited optical densities of no less than 0.2 of absorbance (Table 1).

2.7. Statistical analysis

3.2. Linearity and accuracy of the in vitro ELISA method

All ELISA results were analyzed using the Parallel Line Assay 1.2.0 software, which relates the log of the optical density to the log of the dose. The correlation study between the in vivo and in vitro assays was analyzed via the implementation of a two-sample t-test or a Wilcoxon rank sum test using SAS 9.2. The potency comparisons of inter-laboratory or inter-lot assays were analyzed using an analysis of variance.

The linearity of the ELISA was assessed using a dilution range from 1: 10–1: 320 of the national reference standard. The linear model had a correction coefficient R2 that was no less than 0.95, and the standard curves had similar slopes ranging from 0.994 to 0.999 (Fig. 1). The accuracy of the ELISA, assessed as the recovery rate (%), ranged from 92.8 to 109.3% (Table 2). This range was acceptable, as the specifications had been determined to range from 80 to 120%.

Table 2 The accuracy of the ELISA. Samples

Dilution

mE.U./mL

Standard [Code: 07/022]

10 20 40 80 160 320

133.1 66.6 33.3 16.6 8.3 4.2

Recovery(%) Mean

SD

%CV

98.1 109.3 97.9 92.8 96.8 106.8

2.8 5.2 4.5 6.1 5.6 5.8

2.8 4.7 4.6 6.6 5.7 5.4

3. Results 3.3. Precision of the in vitro ELISA method 3.1. Specificity of the in vitro ELISA method The specificity of the ELISA for the dilution buffer was confirmed using the negative control. The optical density of dilution buffer and negative control were no more than 0.1. The positive control and

The repeatability (intra-plate) and intermediate precision (inter-day and inter-analyst) were assessed with 2 commercial vaccines. The coefficients of variation (CV) of the repeatability was found to be below 15%. The intermediate precision was acceptable

B.-C. Kim et al. / Virus Research 223 (2016) 190–196

193

Fig. 2. Correlation between the in vivo and in vitro assays with (a) the reduced-potency vaccine (liquid form) and (b) the standard (lyophilized form).

Table 3 The precision of the ELISA. Samples (E.U./mL)

Conc.

Repeatability (%CV)

Intermediate Precision (%CV)

Analyst #1

Analyst #2

Intra-plate

vaccine (company A)

Lot #1

Lot #2

Lot #3

vaccine (company B)

Lot #1

Lot #2

Lot #3

200% 100% 50% 200% 100% 50% 200% 100% 50% 200% 100% 50% 200% 100% 50% 200% 100% 50%

plate #1

plate #2

1.3 3.8 3.4 3.2 2.7 3.2 4.0 8.8 2.7 1.2 1.6 3.4 4.2 2.2 1.6 1.0 2.2 4.3

9.0 2.5 2.7 4.0 3.7 0.6 2.4 5.7 5.6 9.4 4.2 2.9 3.3 2.1 1.8 0.8 4.6 2.0

(≤20%) with an inter-day CV of 2.8–12.6% and inter-analyst CV of 2.7–7.9% (Table 3). 3.4. Correlation between the in vivo and in vitro assays The results of the correlation between the in vivo and in vitro assays were analyzed with the reduced-potency vaccines and the standard, as shown in Fig. 2. The relative potency of the in vivo assay was calculated using the potency of the vaccine divided by the potency of the standard. Formulations of JEV vaccine samples were liquid and the reference standards were lyophilized. Lyophilization generally results in improved stability profiles. Regardless of different formulation, the correlation value of the two methods was 0.832 (P = 0.010) with a Spearman correlation analysis, there was a statistically significant relationship between the in vivo and in vitro assays at the 95% confidence level (Fig. 2, Table 4). 3.5. Collaborative study The feasibility studies and training were performed by all participants to define standard operating procedures and to identify critical parameters for the collaborative study (results not shown). Four participants performed 5 in vivo assays and 5 in vitro assays,

Intra-plate

Inter-day

Inter-analyst

4.0 3.9 4.1 1.5 2.9 3.2 2.6 2.1 2.5 4.9 4.9 2.8 3.2 6.9 5.8 9.4 8.3 6.2

11.6 3.8 5.1 6.9 4.1 7.6 8.6 8.1 8.6 8.8 11.8 2.8 3.5 7.8 7.3 12.6 8.0 10.0

2.7 6.0 3.6 3.0 3.4 3.0 3.2 5.9 4.1 3.9 3.4 5.3 6.8 4.7 3.9 7.9 6.8 5.9

respectively, with 3 different lot samples as requested per protocol. The raw data of the in vitro assays were submitted for central calculations at the MFDS using the parallel line assay model. The PLA 1.2.0 software was used to produce the most satisfactory parallelism and linearity. The means and CV of the potency estimates per laboratory were displayed in Table 5. The overall mean potencies for the three different lots were 2.94, 2.93 and 2.93 for the in vivo assay and 3.19, 3.01 and 2.82 for the in vitro assay. The overall CV ranged from 7.20 to 8.19% for the in vivo assay and 4.52–6.27% for the in vitro assay.

3.6. Determination of potency in 10 commercial vaccine lots The differences of potency for ten commercial vaccine lots from two Korean pharmaceutical companies were evaluated by the MFDS and manufacturer. The mean potencies of lots 1–5 from company A were 3.02, 2.95, 2.04, 2.10 and 2.83, respectively, for the in vitro assays and 3.04, 3.08, 2.95, 3.00 and 3.02, respectively, for the in vivo assays. The mean potencies of the company B vaccines were 3.22, 3.09, 2.99, 2.91 and 4.60, respectively, for the in vitro assays and 3.09, 3.14, 3.05, 3.04 and 3.09, respectively, for the in vivo assays (Fig. 3).

194

B.-C. Kim et al. / Virus Research 223 (2016) 190–196

Table 4 Correlation between the in vivo and in vitro assays. Method

N

Mean

SD

CV

Correlation with in vitro

p-value†

p-value*

p-value**

in vivo in vitro

8 8

2.36 0.89

0.81 0.81

34.34 91.27

0.83216 –

0.0104 –

0.0028 –

0.0172 –

† * **

p-value for correlation analysis with in vitro. p-value for two sample t-test between in vivo and in vitro. p-value for Wilcoxon rank sum test between in vivo and in vitro.

4. Discussion

Fig. 3. Comparison of the in vivo and in vitro assays with different lots manufactured by (a) company A and (b) company B. (a) Lots #1 and #2 made from final bulk lot Number F1. Lots #3 and #4 made from final bulk lot Number F2. Lot #5 made from final bulk lot Number F3. (b) Lots #1, #2, #3 and #4 made from final bulk lot Number F1. Lot #5 made from final bulk lot Number F2. Table 5 Collaborative study of the in vivo and in vitro assays. Lab.

Potency (in vivo)

Antigen Content (in vitro)

Sample 1 Sample 2 sample 3 Sample 1 Sample 2 sample 3 A B C D All

Mean CV% Mean CV% Mean CV% Mean CV% Mean CV%

3.16 (4.35) 2.79 (2.21) 3.14 (0.88) 2.45 (2.48) 2.94 (7.95)

3.16 (4.18) 2.82 (2.11) 3.07 (1.08) 2.67 (2.57) 2.93 (7.20)

3.16 (3.81) 2.78 (3.55) 3.13 (1.63) 2.65 (2.56) 2.93 (8.19)

3.26 (1.91) 3.18 (4.89) 3.21 (5.45) 3.09 (4.75) 3.19 (4.52)

3.01 (6.73) 3.00 (4.72) 3.03 (5.56) 2.99 (2.05) 3.01 (4.67)

2.80 (6.14) 2.74 (4.23) 2.89 (9.41) 2.85 (4.52) 2.82 (6.27)

The Japanese encephalitis (JE) vaccine is the most effective tool to control national disease; it has been used to prevent JE virus infection in healthy children in Korea since it was licensed in 1969. This vaccine, similar to other biologicals, should be quality controlled via a national lot release test by a national control laboratory (NCL) prior to use in the market. The in vivo PRNTest for the JE vaccine has been used as a quality control for decades and has been used in many countries. However, there are many challenges to replacing the gold standard in vivo test, and there are many benefits in favor of the development of an in vitro method. Prior to replacing the gold standard in vivo test method with an in vitro test, however, it will be necessary to accumulate a range of correlative data for the in vivo and in vitro assays. The MFDS developed an in vitro ELISA estimating the JE E-antigen as a test for antigen content and undertook this correlation study to analyze the in vivo vs. the in vitro methods (Kim et al., 2012). The reason for selecting the E-antigen as a measurement is that anti-E protein antibodies have neutralization activity, which prevents attachment of the JE virus to the host cell receptor (Srivastava et al., 1987; Butrapet et al., 1998; Konishi et al., 1999). Prior to undertaking the study of the correlation of the novel in vitro method and the established in vivo method, we determined that this ELISA was able to measure a range of 4.2 − 133.1 mE.U./mL during the validation study (Table 2). Furthermore, a study with a reduced-potency sample demonstrated that the two methods had a strong correlation (r = 0.832) (Fig. 2, Table 4). Because, the different temperatures can generally impact stability and potency of JE vaccine. The incubation at high temperature for a certain period of time is known to alter the viral particles in many viral vaccines (Poirier et al., 2010). These results confirmed that our monoclonal antibodies used for detection and capture had neutralization activity for the Japanese encephalitis virus. These findings are in accordance with a previous study in which the monoclonal antibodies used for this ELISA exhibited neutralization activity via a PRNTest. The collaborative study with the commercial lots revealed that mean potency is not consistent among laboratories, although the overall CV% of the in vivo test was not higher than that of the inherent variable animal test (Fig. 4). In contrast, results from the in vitro test had a low CV% and similar antigen content in both intra- and inter-laboratory analysis (Fig. 5). The experimental variation of the in vivo assay appeared low because the test results were expressed on a logarithmic scale; however, the mismatched potencies of the inter-laboratory results meant that the variation between the lots tested with the in vivo assay was greater. Nevertheless, the difference between the laboratories did not affect the pass or fail rate for the vaccine because the acceptance specifications of the products were used with the relative potency compared with the national standard material. As the shown in Fig. 5, each relative potency was acceptable for each specification, and the relative potencies were not different among the laboratories. In the study with the commercial vaccines, the potency of the in vivo test was nearly constant regardless of the original lot or company product; thus, the results of the in vivo assay showed a similar potency. While these results may appear to produce consis-

B.-C. Kim et al. / Virus Research 223 (2016) 190–196

195

Fig. 4. Collaborative study of the in vivo assay with 4 participants. (a) Lot 1, (b) Lot 2, (c) Lot 3, (d) standard.

Fig. 5. Collaborative study of the in vitro assay with 4 participants. (a) Lot 1, (b) Lot 2, (c) Lot 3.

tent results from every lot, it may actually demonstrate consistence because the test variation is greater than the manufacturing variation. In the in vitro assay results, the test results of lots produced from the same final bulk lot showed nearly the same antigen content, while antigen contents of lots produced from different final bulk lots were different (Fig. 3). These results suggested that the in vitro method is more suitable for determining the consistency of the JE vaccine lots. In addition, the in vitro method was more favorable in relation to quality control because the antigen content was affected if a modification of the antigen occurred during produc-

tion: the coating monoclonal antibody recognizes a conformational epitope of the E-protein, and the conjugate monoclonal antibody reacts to a linear epitope of the C-terminal of the E-antigen (Kim et al., 2012). The capacity of the ELISA kit to measure the antigen content is important for the quality control of the JE vaccine. Thus, we determined the specifications for quality control for producing the ELISA kit in addition to the specifications for the antibodies, conjugate and positive and negative controls included in the ELISA kit (data now shown). In addition, it was established that the ELISA kit should be

196

B.-C. Kim et al. / Virus Research 223 (2016) 190–196

used within 12 months based on real-time and accelerated-stability test results. In addition, we performed method validation studies of this in vitro ELISA for the inactivated Vero cell culture-derived Beijing-1 strain vaccine and the live-attenuated primary hamster kidney cell culture-derived SA-14-14-2 strain vaccine. The in vitro methods were met all the method validation acceptance criteria (data now shown). These results will be provided in further study. Although the direct replacement of the gold standard in vivo potency assay for all viral vaccines cannot be determined herein, an alternative in vitro method has been used successfully for some viral vaccines, including the hepatitis B and hepatitis A vaccines (Poirier et al., 2010). We are convinced that JE vaccine potency can be determined by an in vitro ELISA of antigen content validated against the in vivo PRN assay of antibody response. Thus, upon further collaborative study of pass/fail data in available commercial lots, and the establishment of new national reference standard for both in vitro ELISA and in vivo PRN assay, this in vitro ELISA may be interchangeable with the in vivo PRN assay. In conclusion, we have demonstrated that this ELISA is a reliable and suitable method for measuring the in vitro antigen content of the mouse brain derived JE vaccine as a consistent quality control test method for the final product. Conflict of interest The authors declare no conflicts of interest. Acknowledgements This study was supported by a grant (11171MFDS352) from Ministry of Food and Drug Safety (MFDS) in 2011. We are grateful to the following vaccine manufacturers for their kind collaboration. Appendix A. List of Study participants (listed in alphabetical order by company) • Boryung Biopharma Co., Republic of Korea • Green Cross Corp., Republic of Korea • Korea Vaccine Co., Republic of Korea References Abe, M., Kuzuhara, S., Kino, Y., 2003. Establishment of an analyzing method for a Japanese encephalitis virus neutralization test in Vero cells. Vaccine 21, 1989–1994. Arai, S., Matsunaga, Y., Takasaki, T., Tanaka-Taya, K., Taniguchi, K., Okabe, N., Kurane, I., 2004. Japanese encephalitis: surveillance and elimination effort in Japan from 1982 to 2004. Jpn. J. Infect. Dis. 61 (5), 333–338. Ashok, M.S., Rangarajan, P.N., 2000. Evaluation of the potency of BIKEN inactivated Japanese Encephalitis vaccine and DNA vaccines in an intracerebral Japanese Encephalitis virus challenge. Vaccine 15–19 (2–3), 155–157. Borges, M.B., Kato, S.E., Damaso, C.R., Moussatche, N., Da Silva Freire, M., Lambert Passos, S.R., et al., 2008. Accuracy and repeatability of a micro plaque reduction neutralization test for vaccinia antibodies. Biologicals 36, 105–110. Butrapet, S., Kimura-Kuroda, J., Zhou, D.S., Yasui, K., 1998. Neutralizing mechanism of a monoclonal antibody against Japanese encephalitis virus glycoprotein E. Am. J. Trop. Med. Hyg. 58, 389–398. De Mattia, F., Chapsal, J.M., Descamps, J., Halder, M., Jarrett, N., Kross, I., et al., 2011. The consistency approach for quality control of vaccines − a strategy to improve quality control and implement 3Rs. Biologicals 39 (1), 59–65. Fischer, M., Hills, S., Staples, E., Johnson, B., Yaich, M., Solomon, T., 2008. Japanese encephalitis prevention and control: advances, challenges, and new initiatives, 93. In: Scheld, W.M., Hammer, S.M., Hughes, J.M. (Eds.), Emerging Infections 8. ASM Press, Washington, DC, pp. 93–124.

Fischer, M., Griggs, A., Staples, J., 2009. Japanese encephalitis, 74. In: Brunette, G. (Ed.), Health Information for International Travel 2010. US Department of Health and Human Services, Public Health Service, Atlanta, pp. 74–81. Gould, E.A., Solomon, T., Mackenzie, J.S., 2008. Does antiviral therapy have a role in the control of Japanese encephalitis? Antiviral Res. 78, 140–149. Gowal, D., Kumar, S., Sharma, A., 2010. Stability of freeze-dried inactivated Japanese encephalitis vaccine under different experimental conditions. Am. J. Biomed. Sci. 2, 184–189. Griffiths, E., 1996. Assuring the quality of vaccines: regulatory requirements for licensing and batch release. Methods Mol. Med. 4, 269–288. Gupta, N., Chatterjee, K., Karmakar, S., Jain, S.K., Venkatesh, S., Lal. Bellary, S., 2008. India achieves negligible case fatality due to Japanese encephalitis despite no vaccination: an outbreak investigation in 2004. Indian J. Pediatr. 75 (1), 31–37. Halstead, S.B., Jacobson, J., 2008. Japanese encephalitis vaccines, 311. In: Plotkin, S.A., Orenstein, W.A., Offit, P. (Eds.), Vaccines. , 5th edn. Elservier, Philadelphia, pp. 311–352. Halstead, S.B., Thomas, S.J., 2011. New Japanese encephalitis vaccines: alternatives to production in mouse brain. Expert Rev. Vaccines 10 (3), 355–364. Hendriksen, C., Spieser, J.M., Akkermans, A., Balls, M., Bruckner, L., Cussler, K., et al., 1998. Validation of alternative methods for the potency testing of vaccines. ATLA 26, 747–761. Hendriksen, C., 2002. Refinement, reduction, and replacement of animal use for regulatory testing: current best scientific practices for the evaluation of safety and potency of biologicals. ILAR J. 43 (Suppl), S43–S48. Hendriksen, C., 2007. Three rs achievements in vaccinology. AATEX 14, special issue. In: Proc. 6th World Congress on Alternatives & Animal Use in the Life Sciences, Tokyo, Japan, pp. 575–579. Isbrucker, R., Levis, R., Casey, W., McFarland, R., Schmitt, M., Arciniega, J., et al., 2011. Alternative methods and strategies to reduce, refine, and replace animal use for human vaccine post-licensing safety testing: state of the science and future directions. Proced. Vaccinol., 47–59. Ministry of Food and Drug Safety, 2011. Korean Biological Products Specification and Test Methods, 120–121. Kim, D.K., Kim, H.Y., Kim, J.Y., Ye, M.B., Park, K.B., Han, E., et al., 2012. Development of an in vitro antigen-detection test as an alternative method to the in vivo plaque reduction neutralization test for the quality control of Japanese encephalitis virus vaccine. Microbiol. Immunol. 56, 463–471. Konishi, E., Yamaoka, M., Khin Sane, W., Kurane, I., Takada, K., Mason, P.W., 1999. The anamnestic neutralizing antibody response is critical for protection of mice from challenge following vaccination with a plasmid encoding the Japanese encephalitis virus premembrane and envelope genes. J. Virol. 73, 5527–5534. Ministry of Health and Welfare, 2006. Japanese Government. Minimum Requirements for Biological Products, 93–94. Misra, U.K., Kalita, J., 2010. Overview. japanese encephalitis. Prog. Neurobiol. 91, 108–120. Oya, A., Kan, K., Kurokawan, M., Oshida, S.A., 1967. A new formula for direct calculation of neutralizing antibody titer against Japanese encephalitis virus from plaque reduction rate on chick embryo tissue culture, 5. In 2nd Report of Japanese Vaccine Research Commission, 5–10. Poirier, B., Variot, P., Delourme, P., Maurin, J., Morgeaux, S., 2010. Would an in vitro ELISA test be a suitable alternative potency method to the in vivo immunogenicity assay commonly used in the context of international Hepatitis A vaccines batch release? Vaccine 28, 1796–1802. Shin, J., LeiD. Conrad, C., Knezevic, I., Wood, D., 2011. International regulatory requirements for vaccine safety and potency testing: a WHO perspective. Proced. Vaccinol., 164–170. Solomon, T., Dung, N.M., Kneen, R., Gainsborough, M., Vaughn, D.W., Khanh, V.T., 2000. Japanese encephalitis. J. Neurol. Neurosurg. Psychiatry 68, 405–415. Srivastava, A.K., Aira, Y., Mori, C., Kobayashi, Y., Igarashi, A., 1987. Antigenicity of Japanese encephalitis virus envelope glycoprotein V3 (E) and its cyanogen bromide cleaved fragments examined by monoclonal antibodies and Western blotting. Arch. Virol. 96, 97–107. World Health Organization, 1992. Guidelines for national authorities on quality assurance for biological products. annex 2 (WHO technical report series, No. 822). In: In WHO Expert Committee on Biological Standardization, Geneva. World Health Organization, 1998. Requirements for japanese encephalitis vaccine (inactivated) for human use. thirty-eighth report, annex 6 (WHO technical report series No. 771). In: In WHO Expert Committee on Biological Standardization, Geneva. World Health Organization, 2006. Japanese encephalitis vaccines. Wkly Epidemiol. Rec. 81, 331–340. World Health Organization, 2007. Proposed revision: recommendations for japanese encephalitis vaccine (inactivated) for human use. In: 58st Meeting of the WHO Expert Committee on Biological Standardization, Geneva. World Health Organization, 2010. Guidelines for independent lot release of vaccines by regulatory authorities. In: 61st Meeting of the WHO Expert Committee on Biological Standardization, Geneva. Wilder-Smith, A., Freedman, D.O., 2009. Japanese encephalitis: is there a need for a novel vaccine? Expert Rev. Vaccines 8, 969–972.