WHO/BS/2017.2307 ENGLISH ONLY
EXPERT COMMITTEE ON BIOLOGICAL STANDARDIZATION Geneva, 17 to 20 October 2017
A WHO collaborative study to evaluate a candidate International Standard for anti-Typhoid capsular Vi polysaccharide IgG (Human)
Sjoerd Rijpkema1*, Jason Hockley2, Alastair Logan1, Peter Rigsby2, Eleanor Atkinson2, Celina Jin3 and the anti-Vi IgG working group
1Division of Bacteriology and 2Biostatistics Group, National Institute for Biological Standards and Control, Potters Bar, Hertfordshire, EN6 3QG; 3Oxford Vaccine Group, University of Oxford, Centre for Clinical Vaccinology and Tropical Medicine, Churchill Hospital, Oxford OX3 7LE,
United Kingdom
*(Corresponding author, Email address: Sjoerd.Rijpkema@nibsc.org NOTE:
This document has been prepared for the purpose of inviting comments and suggestions on the proposals contained therein, which will then be considered by the Expert Committee on Biological Standardization (ECBS). Comments MUST be received by 18 September 2017 and should be addressed to the World Health Organization, 1211 Geneva 27, Switzerland, attention: Technologies, Standards and Norms (TSN). Comments may also be submitted electronically to the Responsible Officer: Dr Insoo SHIN at email: shini@who.int.
© World Health Organization 2017
All rights reserved. Publications of the World Health Organization are available on the WHO web site (www.who.int) or can be purchased from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264;
fax: +41 22 791 4857; e-mail: bookorders@who.int).
Requests for permission to reproduce or translate WHO publications – whether for sale or for noncommercial distribution – should be addressed to WHO Press through the WHO web site: (http://www.who.int/about/licensing/copyright_form/en/index.html).
The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement.
The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters.
All reasonable precautions have been taken by the World Health Organization to verify the information contained in this publication.
However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall the World Health Organization be liable for damages arising from its use. The named authors alone are responsible for the views expressed in this publication.
Summary
Candidate International Standard (IS) 16/138 is a serum pool of 9 volunteers vaccinated with Vi polysaccharide Tetanus Toxoid (Vi-TT) conjugate and 7 volunteers vaccinated with plain Vi, which contains IgG directed against the Vi capsular polysaccharide of Salmonella enterica subspecies enterica serovar Typhi. At the National Institute for Biological Standards and Control (NIBSC), 2551 ampoules were filled with 1 mL of candidate IS 16/138, freeze dried and stored at -20oC. As part of a collaborative study, 7 laboratories from 6 countries assessed the suitability of candidate IS 16/138 as a reference for anti-Vi IgG serum (human), alongside U.S. reference reagent Vi-IgGR1, 2011 (Center for Biologics Evaluation and Research, Food & Drug Administration, USA), NIBSC 10/126 (the previous candidate IS), 3 post-vaccination sera and one pre-vaccination serum in several Vi ELISA formats. Datasets of the VaccZyme ELISA (Binding Site, n=7), a biotinylated Vi ELISA developed at NIBSC (NIBSC ELISA, n=7) and in-house ELISAs based on locally prepared or modified Vi (n=7) were used to calculate the potency of test samples relative to candidate IS 16/138 and U.S. reference reagent Vi-IgGR1, 2011. In the majority of cases, valid estimates were obtained for the potency of coded samples relative to candidate IS 16/138 or U.S. reference reagent Vi-IgGR1, 2011, and for the relative potency of the two standards to each other. Six datasets (NIBSC ELISA, n=2; Vacczyme ELISA, n=1; and in-house ELISAs, n=3) were not included in the final analysis, because data did not meet criteria for parallelism and repeatability, or did not allow estimates of relative potency to be calculated. The outcomes of the study were the same for both candidate IS NIBSC 16/138 and U.S. reference Vi-IgGR1, 2011. Candidate IS 16/138 is suitable as a reference standard for anti-Vi IgG serum (human) in the VaccZyme ELISA and in in-house ELISAs where the commutability of 16/138 with the coded samples and U.S. reference reagent Vi-IgGR1, 2011 was evident. Assuming an arbitrary unitage of 100 IU is assigned per ampoule or 100 IU mL-1 to NIBSC 16/138, based on the VaccZyme ELISA, then the potency of U.S. reference reagent Vi-IgGR1, 2011 is 163 IU per vial and the potency of NIBSC 10/126 is 54 IU per ampoule.
Introduction
Typhoid is caused by an infection with Salmonella enterica subspecies enterica serovar Typhi (S.
Typhi). S. Typhi expresses a capsular polysaccharide labelled Vi, which is a virulence factor and considered the main protective antigen (7,19). In developing countries, Typhoid affects in particular infants, school-age children and young adults and is a considerable cause of morbidity and mortality. Although the true incidence of typhoid fever remains unknown, a recent study estimated an incidence of 12 million cases and 129,000 deaths each year (15).
Vaccination is the most cost effective preventative strategy to control Typhoid, especially in areas where multidrug resistant strains are endemic. However, the oral live attenuated Ty21a vaccine and the injectable plain Vi vaccine are not suitable for use in children under five and two years of age respectively. Furthermore plain Vi vaccines are unable to provide a booster response. In 2001, a prototype Vi conjugate vaccine, Vi recombinant Pseudomonas aeruginosa exoprotein A conjugate (Vi-rEPA), proved successful in clinical trials: vaccination induced a booster response in 2-5 year olds and the vaccine was highly efficacious, immunogenic and safe (11,12,20,22). In addition, Vi- rEPA induced long term protection against Typhoid in infants under 2 year of age and in pre-school age children (23). Recently, two Vi Tetanus Toxoid conjugate (Vi-TT) vaccines were licensed in India (14,16). Currently, 9 Vi conjugate vaccines using a variety of carrier proteins are in development, 5 of these have entered clinical trials and one is close to being licensed. In 2013, Vi- IgGR1, 2011 was established as a U.S. reference reagent for anti-Vi IgG, it consists of purified IgG derived from plasma of volunteers vaccinated with Vi-rEPA, contains 33 µg anti-Vi IgG mL-1 and has a potency relative to in-house reference Vi-IgGNIH of 26.6 EU mL-1 by indirect Vi ELISA (21).
U.S. reference reagent Vi-IgGR1, 2011 is available through the Center for Biologics Evaluation and Research (CBER) of the Food & Drug Administration (FDA) of the Department of Health and Human Services (U.S.A.).
The 1st WHO international reference preparation of anti-Typhoid serum, equine (TYS) was developed by Felix and Pitt in 1935, and has anti-Vi reactivity (7,17). In 2009, the Expert Committee on Biological Standardization of WHO (ECBS) endorsed the replacement of TYS by an anti-Vi IgG serum (human). In 2010, sera from nine volunteers of a Phase II clinical trial with Typbar TCVTM a Vi-TT vaccine of Bharat Biotech International Ltd (BBIL) were donated to the National Institute for Biological Standards and Control (NIBSC) for the production of the then candidate IS NIBSC 10/126 (18).
In 2012, a working group under the auspices of the Korean FDA and WHO reviewed the scientific basis for regulatory evaluation of Typhoid conjugate vaccines (TCVs) with emphasis on manufacturing, quality control, non-clinical and clinical evaluation and standardisation (9). The meeting confirmed that WHO guidelines and WHO reference materials for Vi and human anti-Vi antibodies are required to allow comparison of clinical trial studies of various TCV formulations and their analysis by calibrated physicochemical assays and immunoassays (2). At the time, U.S.
reference reagents Vi-IgGR1, 2011 and Vi PS lot 05 became available. It is likely that use of either the U.S. reference reagent or the WHO international standard (IS) for anti-Vi IgG will lead to confusion and hamper comparisons between clinical trial studies. The meeting concluded that ‘a serum standard containing defined antibody contents, either in a mass unit or in an arbitrary international unit (IU) that is inter-convertible with mass unit is required to support cross calibration’ (9).
Collaborative study (CS) 507 was designed with the aim to compare U.S. reference reagent Vi- IgGR1, 2011 and NIBSC 10/126 head to head (18). The outcome of CS 507 showed that the potency of NIBSC 10/126 relative to U.S. reference reagent Vi-IgGR1, 2011 was highly variable among in-house ELISAs, including the ELISA based on Vi PS lot 05. However, both reagents performed consistently in the commercially available VaccZyme ELISA (Binding Site).
ECBS concluded that: 1) the reliance on a single commercial ELISA combined with the low number of datasets available provided insufficient evidence to establish NIBSC 10/126 as an IS and 2) an investigation into the cause of the high level of inter-laboratory and inter-assay variation was required. In practical terms, ECBS proposed that a set of anti-Vi positive sera from a clinical trial should be included in the study to establish commutability of the reference reagents and that a reliable non-commercial alternative for the Vi ELISA should be developed (3,18).
Analysis of Vi lot 05 revealed a relatively high level of co-purified LPS and for NIBSC 10/126 the level of (cross-) reactive anti-LPS antibodies was high compared to U.S. reference reagent Vi-IgGR1, 2011 (18). Whether these two factors contributed to the variance in potency of NIBSC 10/126 relative to U.S. reference reagent Vi-IgGR1, 2011 observed for ELISAs based on Vi 05 or for that matter other in-house ELISAs remains to be determined. Previous studies have reported poor binding of bacterial polysaccharides to the micro-plate surface as a cause for inconsistent ELISA results and pre-coating or co-coating with a protein or a chemical modification enhanced binding of polysaccharides (4,5,13). Ferry et al. developed an ELISA based on biotinylated Vi that provided a specific and sensitive screen for anti-Vi IgG (8). An ELISA based on biotinylated Vi as coating antigen was also developed in-house at NIBSC and by participant 6 of this CS.
Sera from previous clinical trials of Vi conjugate vaccines proved difficult to access in sufficient quantity. Following a meeting with representatives of the Oxford Vaccine Group (OVG, University of Oxford) and the Bill and Melinda Gates Foundation, an immunisation study was designed to immunize volunteers with a licensed plain Vi vaccine or a Vi-TT conjugate vaccine, licensed in India, to produce post-vaccination sera of sufficient volume and anti-Vi titre. A selection of post- vaccination sera were pooled to use as a source for candidate IS 16/138 and 12 individual pre- and post-vaccination sera were selected for a quality control (QC) panel.
CS 566 was designed to analyse the potency of candidate IS 16/138 and includes U.S. reference reagent Vi-IgGR1, 2011 and NIBSC 10/126 as comparators. In addition, the commutability of candidate IS 16/138 is assessed by comparing the reactivity of individual anti-Vi PS IgG sera from the QC panel. An ELISA, based on biotinylated Vi developed at NIBSC (NIBSC ELISA) was chosen as the non-commercial alternative to the VaccZyme ELISA. The NIBSC ELISA performed well in–house and the specificity and sensitivity of this ELISA format for the detection of anti-Vi IgG was found to be excellent in a study of naïve and vaccinated individuals (8). In addition essential assay reagents were stocked at NIBSC and were available for distribution among study participants. Both assays were made available free of charge, and participants of CS566 were asked to test study samples in the VaccZymeELISA and the NIBSC ELISA, and were encouraged to use their in-house assays as well. The intended use of candidate IS 16/138 is as a benchmark for immunoassays used in clinical trial studies to allow a comparison of the immunogenicity of various TCVs.
The aims of CS 566 were:
1) assess the suitability of candidate IS 16/138 to serve as a WHO IS for anti-Vi PS IgG (human) in various ELISAs.
2) assess the commutability of candidate IS 16/138 with individual anti-Vi PS IgG positive sera in various ELISAs.
3) assign a unitage to candidate IS 16/138 and a relative unitage to U.S. reference reagent Vi-IgGR1, 2011 and 10/126.
4) assess the performance and reproducibility of the NIBSC ELISA, as an alternative assay for the VaccZyme ELISA, across multiple laboratories.
Materials and Methods
Participating laboratories and assay codification
Seven participating laboratories, including vaccine manufacturers, national control laboratories and research laboratories, from six countries tested study samples and supplied test data for CS 566 (see Appendix Table A3 for details). Data were collected and analysed at NIBSC. Throughout the study, participating laboratories and assays were identified by a randomly assigned code number to maintain confidentiality (see Table 2). Each participant received a sample set comprising coded ampoules (A-F), one ampoule of U.S. reference reagent Vi-IgGR1, 2011, one ampoule of candidate IS 16/138, VaccZyme ELISA kits and essential reagents for the NIBSC ELISA . Participants also received Instructions for Use (IFU) and Material Safety Data Sheets for reagents and study samples and a Standard Operating Procedure and manufacturer’s instruction for ELISAs.
Description of test samples and ELISA reagents
A brief characterisation of the study samples, study codes, NIBSC codes and their reactivity in the VaccZyme ELISA is given in Table 1. All samples tested negative for antibodies to HIV 1/2 and Hepatitis C and Hepatitis B surface antigen. Freeze dried samples 16/138 and A-F were reconstituted with water as described in the IFU. CBER donated 9 vials of U.S. reference reagent Vi-IgGR1, 2011 free of charge for use among participants. Details of the production and characterisation of Vi-IgGR1, 2011 are given in Szu et al. (2013). The freeze dried material was reconstituted with water as described in the product circular. Also included were 6 VaccZyme ELISA kits and reagents for the NIBSC ELISA: one ampoule of 5 mg streptavidin from Streptomyces avidinii (S4762, Sigma-Aldrich); one ampoule of 1 mg freeze dried Biotinylated Vi derived from Citrobacter freundii (NIBSC12/244 biotinylated by Innova Biosciences) and one ampoule of goat anti-Human IgG horse radish peroxidase (HRP) conjugate (A8667, Sigma-Aldrich).
Six plate layouts were provided in the study protocol (see Appendix). All samples were sent on dry ice, with the exception the VaccZyme ELISA kits sent at 2-8oC.
The samples used to analyse the stability of candidate IS 16/138 were distributed on dry ice.
Characterization of the candidate International Standard 16/138
Sera from volunteers vaccinated with licensed Typhoid vaccines were supplied by OVG. Ethical approval was obtained from South Central Oxford A Ethics Committee (14/SC/1427) and the protocol was approved by the Medicines and Healthcare Products Regulatory Agency (Eudract
2014-002978-36). The Human Materials Advisory Committee of NIBSC(16/004SR) approved the use of these sera as a source for candidate IS 16/138.
In brief, consenting healthy volunteers (age range 18 to 57 years) received one dose of either plain Vi (Typhim ViTM, Sanofi Pasteur; n=15) or Vi-TT (Typbar TCVTM, BBIL; n=14). Sera were collected at week 4-6, frozen, transferred to NIBSC and stored at -80oC until further use. Pre- and post-immunisation sera of each individual were tested for anti-Vi IgG in the VaccZyme ELISA and the NIBSC ELISA. Post-vaccination sera which showed a >3 fold increase in ELISA Units in both assays were selected as source material for candidate IS 16/138 (results not shown). Seven sera from volunteers vaccinated with Typhim ViTM and 9 sera from volunteers vaccinated with Typbar TCVTM were pooled (see Table 1). The pool of appr 3 L was dispensed in 1.0 mL aliquots into glass ampoules coded NIBSC 16/138. The mean fill weight for 98 ampoules was 1.007 g (CV of 0.21 %).
On the same day, freeze-drying under vacuum started and was completed after four days. Ampoules were back filled with pure N2 (moisture content <10 ppm). Residual moisture, measured by the Karl-Fischer method, for 12 ampoules was 0.5246 % (CV of 6.92 %). An integrity test was not performed; however the oxygen level in the headspace of 12 ampoules was checked and found to be 0.39% (CV 29.97 %). This implies these ampoules passed a test for integrity because the presence of cracks would be associated with an oxygen level of around 20%, similar to that found in the atmosphere. Twelve ampoules were rejected during the production process, 30 ampoules were held for accelerated degradation studies and 2551 ampoules were stored at -20 oC. These are available for distribution by NIBSC.
The effect of freeze drying on levels of specific anti-Vi IgG titre was determined by NIBSC ELISA in liquid and freeze dried samples of candidate IS 16/138. As shown in Figure 1, the anti-Vi IgG titre was not affected.
Sera of the Quality Control panel
Twelve individual pre- and post-vaccination sera of volunteers, immunised with plain Vi or Vi-TT and representing a range of anti-Vi IgG levels, were used to assemble the QC panel. Sera were dispensed in 1.0 mL aliquots, freeze dried and labelled 16/140, 16/142, 16/144, 16/146, 16/148, 16/150, 16/168, 16/172, 16/174, 16/176, 16/178 and 16/180. Approximately 80 ampoules of each sample were produced and stored at -20 oC. These sera are available for distribution by NIBSC.
Four QC samples (coded B, D, E and F) were selected for CS 566.
Characterization of NIBSC 10/126
Approximately 1 L of human serum was collected from a total of 9 consenting volunteers (24-30 years old) who had received two doses of Typbar TCVTM (BBIL) as part of a phase II clinical trial (see Table 1). The vaccinations were given 4 weeks apart and sera were collected at week 6 after the booster. The use of source material for candidate IS 10/126 was approved by the Human Materials Advisory Committee of NIBSC 10_002SR. At NIBSC, samples were thawed and pooled. The serum pool was dispensed as 0.5 mL aliquots into glass ampoules coded 10/126. The mean fill weight for 78 ampoules was 0.5181 g (CV of 0.19 %). On the same day, freeze-drying under vacuum started and was completed after four days. Ampoules were back filled with pure N2 (moisture content <10 ppm). Residual moisture, measured by the Karl-Fischer method, for 6 ampoules was 0.23 % (CV of 17.60 %). An integrity test was not performed; however the oxygen
level in the headspace of 12 ampoules was checked and found to be 0.59% (CV 29.55%). This implies these ampoules passed a test for integrity because the presence of cracks would be associated with an oxygen level of around 20%, similar to that found in the atmosphere. Twenty-six ampoules were rejected during the production process, 44 ampoules were held for accelerated degradation studies and 1654 ampoules are stored at -20 oC and available for distribution by NIBSC.
The effect of freeze drying on levels of specific anti-Vi IgG titres was determined in samples of 10/126 taken before and after freeze drying and results showed no effect on the titre of IgG directed against Vi (18).
Test methods
Study samples were tested in the commercially available VaccZyme ELISA and in-house ELISAs.
ELISA formats and the participant codes are given in Table 2. A summary of the operating procedures of all ELISAs is given below. Two ELISA formats were made available to all participants: the VaccZyme ELISA based on S Typhi Vi, and the NIBSC ELISA based on a complex of streptavidin-biotinylated C freundii Vi, developed at NIBSC. All participants performed the NIBSC ELISA, the Vacczyme ELISA as well as their in-house ELISA. Samples were tested repeatedly on 6 ELISA plates except by participant 2 (4 plates for the Vacczyme ELISA and 2 plates for the in-house ELISA) and participant 6 (5 plates for the Vacczyme ELISA). Six different plate layouts were provided in the study protocol (see Appendix). Starting dilutions of sera and reference reagents of 1:100 were made in appropriate assay diluents, with the exception of candidate IS16/138, which had a dilution of 1:25. A description of ELISA procedures accompanied the raw data returned by the participant for analysis.
VaccZyme ELISA. The VaccZyme ELISA was carried out according to manufacturer’s instructions (Binding Site, UK). The VaccZyme ELISA is based on Vi of S. Typhi. Five calibrators, covering a range from 7.4, 22.2, 66.7, 200 and 600 EU mL-1, and a high and a low positive control are included in each run. The assay has a sensitivity of 7.4 EU mL-1. Plates were read at OD450nm. NIBSC ELISA. The NIBSC ELISA uses a coating of streptavidin to capture biotinylated Vi from C freundii (NIBSC 12/244; Innova Biosciences). The procedure is adapted from the internal NIBSC document ‘Biotinylated Vi PS ELISA for serology’ (1), which is a modification of the ELISA described by Ferry et al. (8). A 1 mL suspension of 1 mg biotinylated Vi from C freundii and 10 mg Trehalose (a cryopreservant) was dispensed in each ampoule and freeze dried at NIBSC. Wells of a micro-titre plate (Nunc) were pre-coated with 100 µL of 5 µg Streptavidin mL-1 (Sigma Aldrich, S4762) in sterile water and incubated uncovered overnight at 37oC. The content of one ampoule with freeze dried biotinylated Vi was reconstituted with 1 mL water as per IFU. On the following day, 3 µg biotinylated Vi mL-1 in PBS was prepared from stock, and 100µL was added to wells of a streptavidin-coated plate and incubated for 3 h at 37oC. Plates were washed three times and 250 µL Blocking Buffer 1(BB1: 10% non-fat dry milk (w/v) in PBS [pH 7.3-7.5]) was added to each well.
Plates were incubated for 2 h at 4oC and washed 3x with PBS-Tween 0.05% (PBS-T). Sera and reference reagents were diluted in Assay Diluent 1 (AD1; 0.05% Tween-20 in BB1). 200 µL each of starting dilutions was added to appropriate wells and 100 µL AD1 was added to the remaining wells.
Doubling dilutions were performed on the plate. Plates were covered and incubated for 2 h at RT (range 20-25oC) and washed 3x with PBS-T. 100µL of HRP conjugated goat anti-human IgG (Sigma Aldrich, A8667) diluted 1:10,000 in AD1 was added to each well. Plates were covered, incubated for 1 h at RT and washed 3x with PBS-T. 100µL of TMBlue substrate (Leinco
Technologies) was added to each well. Plates were developed in the dark for 10 mins and the reaction was stopped with 50 µL H2SO4 per well. Plates were read at OD450nm.
In-house ELISA of participant 1. For this indirect ELISA, plate surfaces were pre-coated with Poly-L- Lysine hydrobromide (Sigma, P1399) followed by a coat with Vi from S. Typhi produced locally. Wells were pre-coated with 100 µL of 10 µg poly-L-Lysine mL-1 in PBS, incubated at RT for 2 h and washed 6x with 0.85% NaCl and 0.1% Brij35 in ultrapure water (NaCl-Brij 35). 100 µL of 2 µg Vi mL-1 in PBS was added to the wells. Plates were incubated overnight at 37℃ and washed 6x with NaCl-Brij 35. 250 µL BB2 (1% BSA in PBS) was added to each well. Plates were incubated for 1 h at 37℃ and washed six times. Dilutions of test samples were prepared in AD2 (BB2 with 0.1% Brij35). Then 100 µL of diluted tested sample was dispensed to appropriate wells of the plate.
Incubate at 37℃ for 1h and washed 6x with NaCl-Brij 35. Then 100 µL of alkaline-phosphatase (AP) conjugated mouse anti-human IgG Fc (Abcam, AB99764) diluted 1:2,000 in AD2 was added to each well, incubated at 37℃ for 1h and washed 6x with NaCl-Brij 35. 100 µL of 4-Nitrophenyl Phosphate liquid substrate (Sigma-Aldrich, N-7653) and incubate at RT in dark for 1h, followed by 50 µL of 3M NaOH stop solution to each well. Plates were read at OD405nm.
In-house ELISA of participant 2. For this indirect ELISA, Vi from C. freundii produced in–
house, was coated on the plate surface. Wells were coated with 100 µL of 4 μg Vi mL-1 diluted in PBS and incubated overnight at 2 – 8 ºC. Plates were washed 1x with PBS-T using a plate washer and plates were tapped dry on tissue paper. Wells were blocked with 200 μL BB3 (PBS + 5%
skimmed milk) and incubated at RT for 1 h. Plates were washed 3x with PBS-T. Sera were diluted in AD3 (PBS-T + 0.1 % BSA) and 100 μL of each dilution was added to wells in duplicate. Plates were incubated at RT for 2 h. Plates were washed 3x with PBS-T. Goat anti Human IgG-AP (Sigma- Aldrich, A9544) was diluted 1:15,000 in AD3 and 100 μL was added to each well. Plates were incubated at RT for 1 h. Plates were washed 5x with PBS-T. 4-Nitrophenyl Phospate liquid substrate (Sigma-Aldrich, N-7653) was warmed to RT and 100μL was added per well. Plates were incubated at RT for 1 hr. The reaction was stopped by adding 100μL of 2M NaOH to each well. The plate was read at 405 nm and 490 nm. Background OD490 was subtracted from OD405.
In-house ELISA of participant 3. This indirect ELISA used Vi lot 05 (CBER, FDA) as antigen and followed the procedure by Szu et al. with minor modifications (21). Wells (Nunc Maxisorp) were coated with 100 µL of 2 μg Vi mL-1 diluted in PBS. Plates were covered, incubated overnight at 25ºC, washed 2x with NaCl-Brij and shaken dry. Wells were blocked with BB2 for 2 h at 25ºC, washed 2x with NaCl-Brij and shaken dry. Starting dilutions of test samples were prepared in AD2.
Wells of the first row received 200 µL of the starting dilution and all other wells received 100 µL AD2. Top down dilutions were prepared by transfer of 100 µL of the diluted sample to subsequent wells. Plates were covered, incubated overnight at 25ºC, washed 2x with NaCl-Brij and shaken dry.
Hundred µL of AP conjugated goat anti-human IgG (Sigma) diluted 1:5,000 in AD2 was added to each well and incubated at 37℃ for 4h. Plates were washed 2x with NaCl-Brij, shaken dry and 100 µL of 4-Nitrophenyl Phosphate disodium salt hexahydrate substrate (Sigma) in Tris/3mM Mg-buffer (pH 9.8) was added to each well and incubated at RT for 10-30 mins. Plates were read at OD405nm. In-house ELISA of participant 4. The procedure for this indirect ELISA is identical to the in- house ELISA of participant 1; except for the use of an AP conjugated mouse anti-human IgG Fc (Abcam, AB99764) diluted 1:5,000.
In-house ELISA of participant 5. This indirect ELISA used methylated Human Serum Albumine (mHSA, NIBSC 12/176) as a co-coating agent for Vi from C freundii (NIBSC 12/244) and is a modification of the procedure of Arakere and Frasch (4). In brief, 100 µL of a mixture consisting 5 µg Vi mL-1 and 5 µg mHSA mL-1 in PBS was used to coat wells of a microtiter plate (Nunc). The plate was incubated overnight at 4oC and washed 3x with PBS-T. Wells were blocked with 250 µL BB1 for 2 hours at 4oC and washed 3x with PBS-T. Sera and reference reagents were diluted in AD1. 200 µL each of starting dilutions was added to appropriate wells and 100 µL AD1 was added to the remaining wells. Doubling dilutions were performed on the plate. Plates were covered and incubated for 2 h at RT and washed three times. 100 µL HRP conjugated goat anti-human IgG (Sigma) diluted 1:10,000 in AD1 was added to each well. Plates were covered, incubated for 1 h at RT and washed 3x with PBS-T. 100µL of TMBlue substrate (Leinco Technologies) was added to each well. Plates were developed in the dark for 10 mins and the reaction was stopped with 50 µL 2M H2SO4 per well. Plates were read at OD450nm.
In-house ELISA of participant 6. This ELISA, like the NIBSC ELISA, is a modification of the ELISA described by Ferry et al., which uses a streptavidin coated surface to capture biotinylated Vi from S Typhi. In brief, streptavidin from S avidinii (Sigma, S4762) was prepared in a solution of 3 μg mL-1 and 100 μL was added to each well of flat bottom microtiter plates (Nunc Maxisorp) and incubated uncovered at 37°C overnight, allowing the solution to evaporate to dryness. Pre-coated plates were stored at 4°C until used for coating. For antigen coating, 2 μg mL-1 of biotinylated S.
Typhi Vi PS produced locally was prepared in PBS (pH 7.4) and 100 μL was added to the wells of pre-treated plates. Plates were incubated for 3 h at 37oC and washed with PBS-T and wells were blocked overnight with 250 μL PBS containing 10% non-fat dry milk at 4oC. Plates were washed 6x with PBS-T with two mins soaking period in between washes. Sera and reference reagents were diluted in AD1 and added to the plates. All plates were incubated for 1 h at 37oC. Following incubation and washing as described above, plates were incubated with HRP-labeled goat anti- Human IgG (Jackson Immunoresearch, 109-035-008) diluted in AD1 for 1h at 37oC. The plates were washed 6x with PBS-T and 100 μl of TMB Microwell Peroxidase Substrate (KPL, SeraCare Life Sciences Inc) was added to each well, incubating for 15 mins in the dark (with agitation). The reaction was stopped by adding 100 μL of 1M Phosphoric acid per well. Plates were read immediately at OD450nm.
In-house ELISA of participant 7. This indirect ELISA used in-house Vi PS from S. typhi produced locally as antigen and follows the procedure by Szu et al. (2013) with some modifications.
Wells (Costar, 2592) were coated with 100 µL of 2 μg Vi mL-1 in PBS diluted from 1 mg mL-1 in Milli-Q. Plates were covered, incubated overnight at 4ºC, washed 5x with 1% Brij L23 in 0.8%
NaCl (NaCl-Brij 23). Wells were blocked with BB2 for 2 h at 37ºC, washed 5x with NaCl-Brij 23 and shaken dry. Starting dilutions of test samples were prepared in BB2 with 0.1% Brij L23 (AD4).
Wells of the first row received 200 µL of the starting dilution and all other wells received 100 µL AD4. Top down dilutions were prepared by transfer of 100 µL of the diluted sample to subsequent wells. Plates were covered, incubated overnight at 25ºC, washed 5x with NaCl-Brij 23 and shaken dry. Hundred µL of AP conjugated goat anti-human IgG (Southern Biotech, 2040-04) diluted 1:2,000 in AD4 was added to each well and incubated at 37℃ for 2h. Plates were washed 5x with NaCl-Brij 23, shaken dry and 100 µL of 4-Nitrophenyl Phosphate disodium salt hexahydrate substrate (Sigma, 71768) in Tris/3 mM Mg-buffer (pH 9.8) was added to each well and incubated at RT for 90-120 mins, then stopped by 50 μl 3 M NaOH. Plates were read at OD405nm.
Statistical analysis
Raw data from all participants were analysed at NIBSC using a four-parameter logistic model (sigmoid curves) or a parallel line model. Details of the models used for individual laboratories and assay methods are shown in Table 3. Analysis was performed using EDQM CombiStats Software Version 5.0 (6). Fitted models were used to estimate the potency of coded samples A-E relative to candidate IS 16/138 and U.S. reference reagent Vi-IgGR1, 2011 and the potency of the two reference standards relative to each other. Parallelism of dose-response relationships was concluded if the slope ratio calculated by CombiStats was within the range 0.80 - 1.25 and no relative potency was reported where this was not the case.
All mean results shown in this report are unweighted geometric means (GM). Variability has been expressed using geometric coefficients of variation (GCV = {10s-1}×100% where s is the standard deviation of the log10 transformed estimates). Individual assay estimates of relative potency were log transformed and a mixed effects model used to determine intra-laboratory and inter-laboratory variance components (also expressed as %GCV) for each sample in NIBSC and VaccZyme ELISAs.
Further assessment of agreement in geometric mean results for each pair of laboratories was performed by calculating Lin’s concordance correlation coefficient with log transformed data, although should be noted that these values are only based on a small number of samples (n=6) in each case. Calculations were performed using the R package ‘DescTools’.
Results
Data returned
All participants reported results for candidate IS 16/138 or U.S. reference reagent Vi-IgGR1, 2011 and samples A-E in all ELISAs. Six plates were tested in all cases except for VaccZyme assays by Laboratory 2 (4 plates) and Laboratory 6 (5 plates) and in-house assays by Laboratory 2 (only 2 plates so results are included in Appendix Tables A1 and A2 for information only).
Assay validity
In the majority of cases, valid estimates were obtained for the potency of coded samples A-E relative to candidate IS 16/138 or U.S. reference reagent Vi-IgGR1, 2011 and for the relative potency of the two standards to each other. Details of all individual potency estimates calculated are shown in Appendix Tables A1 and A2.
The following exceptions were not included in further analysis:
Laboratory 2: The NIBSC ELISA showed poor repeatability in relative potency estimates (average GCV >75% compared to <25% in all other labs and assay methods). The results are shown in Appendix Tables A1 and A2 for information.
Laboratory 3: The raw data of the NIBSC and in-house ELISAs did not allow potency estimates to be calculated by parallel line analysis as the dose-response relationships showed high OD values with no plateau in dose-response and no linear section covering more than two dilutions. The results shown for information in Appendix Tables A1 and A2 are based on estimates reported by Laboratory 3 in U mL-1.
Laboratory 7: For the in-house ELISA a high level of assay invalidity was observed (68% of cases) and VaccZyme assays showed poor repeatability in relative potency estimates (average GCV >75%
compared to <25% in all other labs and assay methods). Results from these assays are shown in Appendix Tables A1 and A2 for information only
The relative potency of candidate IS 16/138 and sample C (a coded duplicate of 16/138) was generally close to the expected value of 1.00 in most ELISAs (see Table 4, Figure 2 and Table A1), although a systematic bias was observed in some cases (e.g. VaccZyme ELISAs of Laboratories 1 and 6). The reasons for this are unclear as the laboratories followed the study protocol in all cases and no further exclusions have been made other than those noted above. The impact of this observation on the proposed outcomes is discussed further below.
Comparison of NIBSC and VaccZyme ELISAs
Laboratory GM potencies, overall GM potencies, intra-lab GCVs and inter-lab GCVs are shown in Tables 4 and 5. Figures 2 and 3 show boxplots of the individual assay potency estimates obtained in NIBSC and VaccZyme ELISAs. Inter-lab variability was generally higher in the NIBSC ELISA (GCV~26% using candidate IS 16/138, ~35% using U.S. reference reagent Vi-IgGR1, 2011) compared to VaccZyme (~19% using candidate IS 16/138, ~23% using U.S. reference reagent Vi-IgGR1, 2011).
No agreement in relative potencies was observed between the NIBSC and VaccZyme ELISAs. In addition, as shown in Figures 2 & 3 where the coded samples have been shown in decreasing order of potency in the VaccZyme ELISA, a different ranking of the coded samples is observed in the two different assay formats.
Concordance between NIBSC, VaccZyme and in-house ELISAs
Tables 6 and 7 show the concordance correlation coefficients for all individual assay pairs. Figures 4, 5, 6 and 7 show scatterplots for assay pairs with fitted Deming regression lines included on each plot. Agreement is evident between the VaccZyme ELISAs and in-house ELISAs performed by Laboratories 1, 4 and 5, particularly when using candidate IS NIBSC 16/138 as reference with all values are >0.80 as shown in Table 6.
Results obtained for pre-vaccination serum sample F
Potency estimates for the pre-vaccination serum sample F relative to candidate IS 16/138 were approximated using the same analysis models used for the other samples and are summarised in Table 8. Most laboratories found sample F to be negative (relative potency ≤0.1) for anti-Vi IgG in all 3 ELISA formats. Notable exceptions are Laboratory 3, which gives sample F a high relative potency value in the NIBSC ELISA and their in-house ELISA (based on reported results in U mL-1, not calculated at NIBSC), and the in-house ELISA of Laboratory 2.
Potencies of U.S. reference reagent Vi-IgG
R1, 2011and NIBSC 10/126 relative to candidate IS 16/138
Potency estimates for U.S. reference reagent Vi-IgGR1, 2011 and NIBSC 10/126 relative to candidate IS 16/138 are summarised in Table 9. The overall GM potency of U.S. reference reagent Vi-IgGR1,
2011 calculated relative to candidate IS 16/138 in the VaccZyme ELISA is 1.63 (95% confidence limits 1.44 to 1.85; n=6) or 163 IU mL-1 if candidate IS 16/138 is assigned an arbitrary unitage of 100 IU per ampoule, equivalent to 100 IU mL-1. Inclusion of the in-house ELISAs by Laboratories 1, 4 and 5 gives an estimate of 1.69 (95% confidence limits 1.55 to 1.85; n=9). As shown in Table 9, the overall values were not affected by the inclusion of sample C (coded duplicate of candidate IS 16/138) in these calculations. The impact on the level of agreement in potency estimates obtained for coded samples A-E when calculated relative to NIBSC 16/138 (assigned 100 IU per ampoule) or relative to Vi-IgGR1, 2011 (assigned 163 IU mL-1) is shown in Figure 8.
The overall GM potency of NIBSC 10/126 calculated relative to candidate IS 16/138 in the VaccZyme ELISA is 1.09 (95% confidence limits 0.94 to 1.27; n=6) or 54 IU per ampoule equivalent to 109 IU mL-1, if 16/138 is assigned a unitage of 100 IU per ampoule. Inclusion of the in-house ELISAs by Laboratories 1, 4 and 5 gives an estimate of 1.08 (95% confidence limits 0.98 to 1.19; n=9).
In the VaccZyme ELISA, the overall GM potency of NIBSC 10/126 calculated relative to U.S.
reference reagent Vi-IgGR1, 2011 is 0.67 (n=6) which is close to the value of 0.65 reported previously (18).
Stability studies
Freeze dried samples of candidate IS 16/138 were stored for 8.8 months at -20oC and at elevated temperatures: +4oC, +20oC, +37oC and +45oC. Two ampoules of candidate IS 16/138 exposed to each temperature were tested in duplicate by Laboratory 5 in the VaccZyme ELISA, the NIBSC ELISA and the mHSA Vi ELISA. The potencies relative to samples stored at -20oC are given in Table 10.
The results were analysed by DEGTEST-R (in-house NIBSC software) and used to fit Arrhenius equations relating the degradation rate to absolute temperature assuming first-order decay and hence predict the degradation rates when stored at -20°C (10). Only results of the VaccZyme ELISA were valid and gave a predicted loss of activity for candidate IS 16/138 stored at -20°C of 0.002% per year. The data from the NIBSC ELISA were analysed but the chi-squared test for goodness of fit showed these not to be appropriate for the model. The data from the Vi-MHSA ELISA could not be analysed due to the non-linearity of the dose responses curves. Relative potencies calculated from NIBSC ELISA data sets are shown for information only.
Discussion
The collaborative study reported here has demonstrated that candidate IS 16/138 is suitable to serve as a reference standard for anti-Vi IgG serum (human) in various ELISA formats used in Vi serology where the commutability of candidate IS 16/138 with coded test samples A-E and U.S.
reference reagent Vi-IgGR1, 2011 was evident (VaccZyme ELISA and in-house ELISAs by Laboratories 1, 4 and 5).
The outcomes of this study (assay validity, variability in potency estimates, agreement between methods and laboratories) are similar for both candidate IS NIBSC 16/138 and U.S. reference Vi- IgGR1, 2011. Assuming an arbitrary unitage of 100 IU per ampoule is assigned to NIBSC 16/138, based on the VaccZyme ELISA, the potency of U.S. reference reagent Vi-IgGR1, 2011 is 163 per vial equivalent to 163 IU mL-1.
Previous candidate standard NIBSC 10/126 was also included as a coded test sample in this study and gave an estimated potency relative to U.S. reference reagent Vi-IgGR1, 2011 in the VaccZyme ELISA that showed excellent agreement with the value reported previously (18). Assuming an arbitrary unitage of 100 IU per ampoule is assigned to NIBSC 16/138, the potency of NIBSC 10/126 based on the VaccZyme ELISA is 54 IU per ampoule equivalent to 109 IU mL-1.
Good concordance was seen between the in-house ELISAs of Laboratories 1, 4 and 5 and the VaccZyme ELISA. By contrast there was a lack of concordance between the NIBSC ELISA, the in- house ELISA of Laboratory 6 and the VaccZyme ELISA. We attribute the degree of concordance between ELISAs to the choice of coating procedures for Vi: the in-house ELISAs of Laboratories 1, 4 and 5 use a protein to enhance the binding of native Vi to the plate surface, whereas the NIBSC ELISA and the in-house ELISA by Laboratory 6 use streptavidin to bind a Vi modified by biotinylation to the plate surface. The coating procedure for Vi used in the VaccZyme ELISA was not disclosed and is unknown.
The study results indicate that the NIBSC ELISA, despite a good performance among participants of CS566, is not suitable as a non-commercial alternative for the VaccZyme ELISA as no agreement in relative potency estimates or ranking of coded test samples was observed between the two types of ELISA (Figures 2-5; Tables 6 and 7). However, either one of the in-house ELISAs 1, 4 or 5 could be a non-commercial alternative for the VaccZyme ELISA (Figures 6 and 7; Tables 6 and 7). The in- house ELISA used by participants 1 and 4, is easy to establish in other laboratories, as shown in this CS, and therefore this ELISA format would be a suitable alternative. The coating agents for in-house ELISA of participant 5 (mHSA Vi ELISA) are stored at NIBSC and available for distribution, making this assay a credible candidate as well. A follow-up collaborative study will be required to demonstrate if these ELISA formats can be an alternative for the VaccZyme ELISA.
Recommendation to ECBS
Candidate IS 16/138 contains a freeze dried residue of 1.0 mL pooled post-vaccination sera from volunteers immunised with plain Vi vaccine or conjugated Vi-TT vaccine. The candidate IS was filled to WHO specifications and is suitable to serve as a WHO IS for anti-Vi IgG (human) in various ELISA formats used in Vi serology. On the basis of data from this study we propose that:
1) A unitage of 100 IU is assigned per ampoule of candidate IS 16/138 2) A relative unitage of 54 IU is assigned per ampoule of NIBSC 10/126
3) A follow up collaborative study is required to confirm whether (one of) the in-house ELISAs from participants 1, 4 and 5 can be an alternative for the VaccZyme ELISA and are (is) fit for use in Vi serology
Replies from participants
Participants were asked to correct full names, addresses, method details, to check that their data was correctly represented and to add any other comments. In addition, participants were asked if they agreed with the overall recommendation of the draft report. Replies were received from 6 participants, who all agreed with the outcome and conclusions of CS566.
Participant 1: Agreed with the analysis of their results and the conclusion of the report.
Five comments were made:
1) Affiliation needs correcting.
2) Results: In-house assay of Lab 1, 4, 5 showed good correlation with Vacczyme. Further studies including stability tests need to be tested by in-house ELISA of lab1, 4, 5.
3) Discussion: in-house ELISA of lab 1, 4 is easy to establish in the laboratories as non- commercial alternative for Vacczyme kit. Therefore, it needs to be considered as non- commercial alternative.
4) Table 10 [?]. The results using NIBSC Vi ELISA showed higher variation in titer comparted to in-house ELISA of lab1, 4, 5. Therefore, it would be recommended that further studies include in-house ELISA of lab1, 4, 5.
5) Tables in appendix: Please confirm NL of A sample in plate 1 of VaccZyme by lab1 in the Tab A1 & A2, and confirm NP of B sample in plate 4 of VaccZyme by lab1 in the Tab A1.
Our reply to comments1-4:
1) Amended affiliation and names of staff.
2) The stability of 16/138 has been assessed in three ELISAs. It requires extra effort and cost to ship stability samples on dry ice to Laboratories 4 and 5. It is unlikely that the additional data would not extend our findings.
3) A sentence to this extend has been added to the discussion and as a recommendation ECBS.
The participant is interested to take part in the follow up study, thereby facilitating a comparison of their in-house ELISA.
4) This observation adds to our own analysis of the non-commutability between VaccZyme and NIBSC ELISA. This is stated in the Discussion section.
Participant 2: Agreed with the conclusion of the report and asked to update the affiliation.
Our reply: amended affiliation.
Participant 3: Agreed with the analysis of their results and the conclusion of the report.
Two comments were made:
1) Affiliation and two typos in the ‘Introduction section’ need correcting.
2) Results: The participant expressed surprise at the lack of repeatability and parallelism of their in-house ELISA as noted in the report. Their in-house assay had been validated and has performed well in previous studies. The participant is now undertaking a re-validation of their assay and has requested samples for a retest.
Our reply to comments1 and 2:
1) Amended affiliation and typos.
2) The raw data of their in-house ELISA results were not analysed, as noted in the report.
There is also no evidence of poor repeatability in their in-house assays. However, their
reported in-house assay results (as U/mL; see Table 3), are included in Appendix table A1 and A2 for information. These results do not correlate with the VaccZyme ELISA results from all labs (including theirs).We agree with the action the participant proposes with regards to their in-house ELISA.
Participant 4: Head of the group agreed with the report agreed with the representation of their results and the conclusion of the report. The participant accepted our reply.
Two comments were made:
1) Affiliation needs correcting.
2) In Table 3 it was mentioned ‘was assumed that the dilution 1:100 on plate.’ Why?
Our reply to comments1 and 2:
1) Amended affiliation and names of staff.
2) The data from Lab 4 VaccZyme plate 2 have ODs for the top dilution of 16/138 which only makes sense if the starting dilution is 1/100 rather than 1/25. This would mean that the responses and dilutions match up with the data for Sample C, the coded duplicate, on that plate. If it were 1/25 then all of the coded sample estimates from plate 2 would all be roughly 4-fold different to the other 5 plates.
Participant 5: Agreed with representation of their results and the conclusion of the report.
Participant 7: Agreed with representation of their results and the conclusion of the report.
Comment: Update affiliation.
Our reply: Amended names of staff.
Acknowledgements
We gratefully acknowledge the staff of the participating laboratories for their important contributions in time, expertise and effort, which were indispensable for the completion of this study. We are grateful to Professor AJ Pollard of the Oxford Vaccine Group for his guidance and the volunteers of the VAST B study for their dedication and for donating the sera. We thank J Cipollo of the Division of Biological Standards and Quality Control, the Office of Compliance and Biologics Quality (Department of Health & Human Services, USA) for making reference reagent Vi-IgGR1, 2011 available. We would also like to thank Drs D Goldblatt, L-F Ma and D Steele of the Bill and Melinda Gates Foundation, Dr I Feavers of NIBSC, Dr B Plikaytis of BioStat Consulting, LLC and Dr K Gao of WHO for their guidance and support. Finally we thank Mr M Harris and Mrs S-J Holmes and their colleagues at the Centre for Biological Reference Materials for processing 16/138 and the QC sera, coding and packaging of samples and reagents and organising the distribution of sample packs for the study.
References
1. Anonymous. Biotinylated Vi PS ELISA for serology. NIBSC Workbench document 2017;
7503 v2
2. Anonymous. WHO Guidelines on the quality, safety and efficacy of typhoid conjugate vaccines. 64th Meeting of the WHO Expert Committee on Biological Standardization, 21-25 October 2013. WHO Tech Rep Series 2014; 987:18-19, Annex. Also see: WHO Typhoid BS2215 doc v1.14_Web_Version
3. Anonymous. Proposed First WHO International Standard for anti-Vi polysaccharide serum (human). 66th Meeting of the WHO Expert Committee on Biological Standardization, 12 to 16 October 2015. WHO Tech Rep Series 2016; 999:82.
4. Arakere G, Frasch CE. Specificity of antibodies to O-acetyl-positive and O-acetyl-negative group C Meningococcal polysaccharides in sera from vaccinees and carriers. Infect Immun 1991;59:4349-56.
5. Barra A, Schulz D, Aucouturier P, Preud'homme JL. Measurement of anti-Haemophilus influenzae type b capsular polysaccharide antibodies by ELISA. J Immunol Methods 1988;115:111–7.
6. EDQM – Council of Europe. CombiStats v5.0. www.combistats.eu.
7. Felix A, Pitt RM. A new antigen of B. typhosus. Its relation to virulence and to active and passive immunisation. Lancet 1934;227:186-91.
8. Ferry BL, Misbah SA, Stephens P, Sherrell Z, Lythgoe H, Bateman E, Banner C, Jones J, Groome N, Chapel HM. Development of an anti-Salmonella typhi Vi ELISA: assessment of immunocompetence in healthy donors. Clin Exp Immunol 2004;136:297-303.
9. Jones C, Lee CK, Ahn C, Shin J, Knezevic I. Working Group on quality, safety and efficacy of typhoid Vi capsular polysaccharide conjugate vaccines, Jeju, Republic of Korea, 5-7 September 2012. Vaccine 2013;31:4466–9.
10. Kirkwood TBL. Predicting the stability of biological standards and products. Biometrics, 1977;33: 736-42.
11. Lanh MN, Phan VB, Vo AH, Tran CT, Lin FY, Bryla DA, Chu C, Schiloach J, Robbins JB, Schneerson R, Szu SC. Persistent efficacy of Vi conjugate vaccine against typhoid fever in young children. N Engl J Med 2003;349:1390-1.
12. Lin FY, Ho VA, Khiem HB, Trach DD, Bay PV, Thanh TC, Kossaczka Z, Bryla DA, Shiloach J, Robbins JB, Schneerson R, Szu SC, Lanh MN, Hunt S, Trinh L, Kaufman JB.
The efficacy of a Salmonella typhi Vi conjugate vaccine in two-to-five- year-old Children. N Engl J Med 2001;344:1263-9.
13. Losonsky GA, Ferreccio C, Kotloff KL, Kaintuck S, Robbins JB, Levine MM.
Development and evaluation of an enzyme-linked immunosorbent assay for serum Vi antibodies for detection of chronic Salmonella typhi carriers. J Clin Microbiol 1987;25: 2266-9.
14. Mitra M, Shah N, Ghosh A, Chatterjee S, Kaur I, Bhattacharya N, Basu S. Efficacy and safety of vi-tetanus toxoid conjugated typhoid vaccine (PedaTyph™) in Indian children: School based cluster randomized study. Hum Vaccin Immunother 2016;12:939-45.
15. Mogasale V, Maskery B, Ochiai RL, Lee JS, Mogasale VV, Ramani E, Kim YE, Park JK, Wierzba TF. Burden of typhoid fever in low-income and middle-income countries: a systematic, literature-based update with risk-factor adjustment. Lancet Glob Health 2014;2:e570-80.
16. Mohan VK, Varanasi V, Singh A, Pasetti MF, Levine MM, Venkatesan R, Ella KM. Safety and immunogenicity of a Vi polysaccharide-tetanus toxoid conjugate vaccine (Typbar-TCV) in
healthy infants, children, and adults in typhoid endemic areas: a multicenter, 2-cohort, open- label, double-blind, randomized controlled phase 3 study. Clin Infect Dis. 2015;61:393-402.
17. Rijpkema S, Durrani Z, Lemercinier X, Jones C. Detection of O-acetylated Vi polysaccharide of Salmonella enterica subspecies typhi by enzyme immunoassay. Biologicals 2004;32:11-6.
18. Rijpkema S, Hockley J, Last V, Marwaha A, Rigsby P. A WHO collaborative study to evaluate a candidate International Standard for anti-Typhoid capsular Vi polysaccharide IgG (Human). WHO/BS/2014.2234.
19. Robbins JD, Robbins JB. Re-examination of the protective role of the capsular polysaccharide (Vi antigen) of Salmonella typhi. J Infect Dis 1984;150:436-49.
20. Szu SC, Taylor DN, Trofa AC, Clements JD, Shiloach J, Sadoff JC, Bryla DA, Robbins JB. Laboratory and preliminary clinical characterization of Vi capsular polysaccharide-protein conjugate vaccines. Infect Immun 1994, 62(10):4440-44.
21. Szu SC, Hunt S, Xie G, Robbins JB, Schneerson R, Gupta RK, Zhao Z, Tan X. A human IgG anti-Vi reference for Salmonella typhi with weight-based antibody units assigned. Vaccine 2013;31:1970–4.
22. Szu SC, Klugman K, Hunt S. Re-examination of immune response and estimation of anti-Vi IgG protective threshold against typhoid fever-based on the efficacy trial of Vi conjugate in young children. Vaccine 2014;32:2359–63.
23. Thiem VD, Lin FY, Canh DG, Son NH, Anh DD, Mao ND, Chu C, Hunt SW, Robbins JB, Schneerson R, Szu SC. The Vi conjugate typhoid vaccine is safe, elicits protective levels of IgG anti-Vi, and is compatible with routine infant vaccines. Clin Vaccine Immunol 2011;18:730-5.
Table 1: Characteristics of study samples
Study code NIBSC code Description Reactivity in
VaccZyme ELISA Candidate
IS 16/138
NIBSC 16/138 Pooled Anti-Vi IgG sera from 16 volunteers 631 U mL-1 a - Vi-IgGR1, 2011 Purified Anti-Vi IgG from 5 volunteers 1280 U mL-1 b
32.8 ± 1.2 µg mL-1 c A
B C D E F
NIBSC 10/126 NIBSC 16/150 NIBSC 16/138 NIBSC 16/144 NIBSC 16/146 NIBSC 16/140
Pooled Anti-Vi IgG sera from 9 volunteers Post-Typbar TCV vaccination serum Pooled Anti-Vi IgG sera from 16 volunteers Post-Typhim vaccination serum
Post-Typhim vaccination serum Pre-Vaccination serum
26.6 ± 1.2 EU mL-1 c 842 U mL-1 d
350 UmL-1 d 631 UmL-1 a 345 UmL-1 d 524 UmL-1 d 22 UmL-1 d
a: Based on an estimate from 3 VaccZyme ELISAs
b: Based on an estimate from 4 VaccZyme ELISAs
c: Based on 12 repeats of the in-house ELISA reportedby Szu et al. (21)
d: Estimate from one VaccZyme ELISA
Table 2: Vi ELISA formats used by participants
Participant code ELISA method Vi characteristics Antigen coating procedure Name Format Status Biological origin
1, 2, 3, 4, 5, 6,7 VaccZyme Indirect - S Typhi Unpublished procedure by Binding Site UK 1, 2, 3, 4, 5, 6,7 NIBSC Capture Biotinylated C freundii Streptavidin coat to bind biotinylated Vi
1 In-house Indirect Naive S Typhi Coat with poly–L-Lysine to enhance binding of Vi
2 In-house Indirect Naive C freundii Vi coating only
3 In-house Indirect Naive C freundii Vi coating only
4 In-house Indirect Naive S Typhi Coat with poly–L-Lysine to enhance binding of Vi 5 In-house Indirect Naive C freundii Mixture with mHSA to enhance binding of Vi 6 In-house Capture Biotinylated S Typhi Streptavidin coat to bind biotinylated Vi
7 In-house Indirect Naive S Typhi Vi coating only
Figure 1: Titration of anti-Vi IgG in liquid ( ) and freeze dried ( ) formulations of candidate IS 16/138 by the NIBSC ELISA.
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
0 200 400 600 800 1000
FD 16/138 Liquid 16/138
Dilution
OD 450
Table 3: Summary of analysis models used
ELISA method Lab Analysis Model
Response Transformation
Notes
NIBSC 1 Sigmoid curves Log
NIBSC 2 Sigmoid curves Log
NIBSC 3 - - No analysis of raw data performed;
results reported by lab (U/ml) shown for information in Tables A1 & A2
NIBSC 4 Sigmoid curves Log
NIBSC 5 Sigmoid curves Log
NIBSC 6 Parallel line Log Insufficient plateau in dose-response curves to fit sigmoid curves model
NIBSC 7 Sigmoid curves Log
VaccZyme 1 Sigmoid curves Log
VaccZyme 2 Sigmoid curves Log
VaccZyme 3 Sigmoid curves Log
VaccZyme 4 Sigmoid curves Log Starting dilution for 16/138 assumed to be 1/100 on plate 2
VaccZyme 5 Sigmoid curves Log
VaccZyme 6 Sigmoid curves Log
VaccZyme 7 Sigmoid curves Log
In-house 1 Sigmoid curves Log In-house 2 Sigmoid curves Log
In-house 3 - - No analysis of raw data performed;
results reported by lab (U/ml) shown for information in Tables A1 & A2 In-house 4 Sigmoid curves Log
In-house 5 Sigmoid curves None In-house 6 Parallel line Log In-house 7 Parallel line Log
Table 4: Results summary for potencies relative to candidate IS 16/138
ELISA method
Sample Code
GM Overall
GM
Inter-lab GCV
Intra-lab Lab 1 Lab 2 Lab 3 Lab 4 Lab 5 Lab 6 Lab 7 GCV
NIBSC
Vi-IgGR1, 2011 2.80 3.50 2.61 3.75 1.80 2.80 31% 22%
A 1.34 1.47 1.62 1.70 1.15 1.44 16% 10%
B 0.73 0.77 0.74 1.03 0.73 0.79 15% 11%
C 0.85 0.92 0.90 1.21 1.07 0.98 15% 13%
D 0.46 0.50 0.66 0.25 0.25 0.39 54% 20%
E 0.33 0.39 0.38 0.43 0.60 0.42 25% 10%
VaccZyme
Vi-IgGR1, 2011 1.60 2.02 1.60 1.59 1.39 1.63 1.63 11% 13%
A 0.84 1.04 1.18 1.09 1.23 1.21 1.09 15% 10%
B 0.27 0.48 0.47 0.41 0.31 0.39 0.38 24% 9%
C 0.70 0.97 1.01 1.08 0.94 1.22 0.97 20% 11%
D 0.21 0.18 0.25 0.31 0.31 0.29 0.25 24% 14%
E 0.44 0.60 0.65 0.71 0.66 0.69 0.62 19% 13%
In-house
Vi-IgGR1, 2011 1.88 1.89 1.70 5.18
A 0.93 1.17 1.09 1.51
B 0.46 0.41 0.53 1.25
C 1.04 1.01 0.90 1.15
D 0.22 0.33 0.23 0.14
E 0.68 0.66 0.47 0.42
GM: Geometric Mean
GCV: Geometric Coefficient of Variation (values ≥25% shaded)
Table 5: Results summary for potencies relative toU.S. reference reagent Vi-IgGR1, 2011
ELISA method Sample Code
GM Overall
GM
Inter-lab GCV
Intra-lab Lab 1 Lab 2 Lab 3 Lab 4 Lab 5 Lab 6 Lab 7 GCV
NIBSC
16/138 0.36 0.29 0.38 0.27 0.56 0.36 31% 22%
A 0.46 0.47 0.65 0.45 0.56 0.51 14% 16%
B 0.25 0.22 0.28 0.27 0.38 0.28 18% 22%
C 0.29 0.27 0.36 0.32 0.59 0.35 34% 22%
D 0.16 0.15 0.24 0.07 0.17 0.15 58% 23%
E 0.12 0.11 0.14 0.12 0.34 0.15 57% 28%
VaccZyme
16/138 0.62 0.50 0.63 0.63 0.72 0.61 0.61 11% 13%
A 0.55 0.51 0.73 0.70 0.84 0.75 0.67 20% 13%
B 0.17 0.23 0.29 0.26 0.23 0.24 0.23 20% 13%
C 0.45 0.48 0.63 0.68 0.65 0.76 0.60 22% 12%
D 0.13 0.09 0.16 0.19 0.23 0.18 0.16 40% 16%
E 0.27 0.29 0.40 0.43 0.46 0.43 0.37 24% 17%
In-house
16/138 0.53 0.53 0.59 0.19
A 0.48 0.58 0.62 0.29
B 0.24 0.22 0.31 0.24
C 0.54 0.53 0.53 0.22
D 0.12 0.18 0.13 0.03
E 0.35 0.35 0.27 0.08
GM: Geometric Mean
GCV: Geometric Coefficient of Variation (values ≥25% shaded)
Figure 2: Boxplot of potencies relative to candidate IS 16/138 in NIBSC and VaccZyme ELISAs
Figure 3: Boxplot of potencies relative toU.S. reference reagent Vi-IgGR1, 2011 in NIBSC and VaccZyme ELISAs
Table 6: Concordance correlation coefficients (ρC) for log potencies relative to candidate IS 16/138
NIBSC ELISA VaccZyme ELISA In-House
1 4 5 6 7 1 2 3 4 5 6 1 4 5 6
NIBSC ELISA 1 0.98 0.96 0.89 0.80 0.67 0.75 0.73 0.70 0.65 0.70 0.72 0.78 0.82 0.78 4 0.98 0.97 0.92 0.79 0.64 0.74 0.72 0.69 0.63 0.69 0.71 0.77 0.78 0.80 5 0.96 0.97 0.82 0.70 0.56 0.63 0.64 0.63 0.59 0.63 0.60 0.71 0.70 0.66 6 0.89 0.92 0.82 0.87 0.66 0.83 0.78 0.71 0.63 0.73 0.78 0.76 0.83 0.96 7 0.80 0.79 0.70 0.87 0.77 0.95 0.95 0.90 0.81 0.90 0.94 0.90 0.96 0.81
VaccZyme ELISA 1 0.67 0.64 0.56 0.66 0.77 0.91 0.87 0.85 0.89 0.85 0.89 0.86 0.90 0.64
2 0.75 0.74 0.63 0.83 0.95 0.91 0.97 0.92 0.88 0.93 0.99 0.93 0.97 0.81 3 0.73 0.72 0.64 0.78 0.95 0.87 0.97 0.98 0.95 0.98 0.98 0.97 0.97 0.71 4 0.70 0.69 0.63 0.71 0.90 0.85 0.92 0.98 0.97 0.99 0.96 0.99 0.92 0.63 5 0.65 0.63 0.59 0.63 0.81 0.89 0.88 0.95 0.97 0.97 0.91 0.96 0.88 0.55 6 0.70 0.69 0.63 0.73 0.90 0.85 0.93 0.98 0.99 0.97 0.96 0.98 0.92 0.65
In-House
1 0.72 0.71 0.60 0.78 0.94 0.89 0.99 0.98 0.96 0.91 0.96 0.95 0.96 0.75 4 0.78 0.77 0.71 0.76 0.90 0.86 0.93 0.97 0.99 0.96 0.98 0.95 0.93 0.67 5 0.82 0.78 0.70 0.83 0.96 0.90 0.97 0.97 0.92 0.88 0.92 0.96 0.93 0.78 6 0.78 0.80 0.66 0.96 0.81 0.64 0.81 0.71 0.63 0.55 0.65 0.75 0.67 0.78
ρC ≥ 0.90 0.80 ≤ ρC < 0.90
Table 7: Concordance correlation coefficients (ρC) for log potencies relative to U.S. reference reagent Vi-IgGR1, 2011
NIBSC ELISA VaccZyme ELISA In-house
1 4 5 6 7 1 2 3 4 5 6 1 4 5 6
NIBSC ELISA 1 0.97 0.87 0.79 0.46 0.60 0.59 0.43 0.36 0.33 0.36 0.48 0.47 0.62 0.47 4 0.97 0.80 0.81 0.39 0.54 0.54 0.38 0.32 0.30 0.33 0.42 0.42 0.54 0.49 5 0.87 0.80 0.59 0.51 0.59 0.51 0.49 0.43 0.43 0.44 0.44 0.55 0.65 0.31 6 0.79 0.81 0.59 0.51 0.60 0.74 0.48 0.39 0.33 0.41 0.59 0.47 0.66 0.81 7 0.46 0.39 0.51 0.51 0.76 0.81 0.93 0.88 0.77 0.86 0.88 0.87 0.93 0.33
VaccZyme ELISA 1 0.60 0.54 0.59 0.60 0.76 0.93 0.84 0.81 0.78 0.82 0.93 0.92 0.90 0.39
2 0.59 0.54 0.51 0.74 0.81 0.93 0.84 0.76 0.68 0.77 0.96 0.85 0.93 0.53 3 0.43 0.38 0.49 0.48 0.93 0.84 0.84 0.98 0.93 0.97 0.90 0.93 0.93 0.32 4 0.36 0.32 0.43 0.39 0.88 0.81 0.76 0.98 0.97 0.99 0.85 0.93 0.85 0.25 5 0.33 0.30 0.43 0.33 0.77 0.78 0.68 0.93 0.97 0.97 0.76 0.87 0.77 0.21 6 0.36 0.33 0.44 0.41 0.86 0.82 0.77 0.97 0.99 0.97 0.85 0.91 0.85 0.27
In-House
1 0.48 0.42 0.44 0.59 0.88 0.93 0.96 0.90 0.85 0.76 0.85 0.93 0.94 0.40 4 0.47 0.42 0.55 0.47 0.87 0.92 0.85 0.93 0.93 0.87 0.91 0.93 0.90 0.28 5 0.62 0.54 0.65 0.66 0.93 0.90 0.93 0.93 0.85 0.77 0.85 0.94 0.90 0.44 6 0.47 0.49 0.31 0.81 0.33 0.39 0.53 0.32 0.25 0.21 0.27 0.40 0.28 0.44
ρC ≥ 0.90 0.80 ≤ ρC < 0.90
Figure 4: Scatterplots of log potencies relative to candidate IS 16/138 in NIBSC and in-house ELISAs for individual lab pairs; solid line indicates agreement, dashed line indicates fitted Deming regression
Figure 5: Scatterplots of log potencies relative to U.S. reference reagent Vi-IgGR1, 2011 in NIBSC and in-house ELISAs for individual lab pairs; solid line indicates agreement, dashed line indicates fitted Deming regression
Figure 6: Scatterplots of log potencies relative to candidate IS 16/138 in VaccZyme and in-house ELISAs for individual lab pairs;
solid line indicates agreement, dashed line indicates fitted Deming regression
Figure 7: Scatterplots of log potencies relative to U.S. reference reagent Vi-IgGR1, 2011 in VaccZyme and in-house ELISAs for individual lab pairs; solid line indicates agreement, dashed line indicates fitted Deming regression
Figure 8: Boxplot of log potencies (IU mL-1), calculated relative to candidate IS 16/138 (assigned 100 IU per ampoule) or Vi-IgGR1,
2011 (assigned 163 IU per mL-1) in VaccZyme ELISAs (V) and in-house ELISAs by Laboratories 1, 4 and 5 (IH)
