Report on the WHO collaborative study to establish the 1st International Standard for antiserum to Respiratory Syncytial Virus

Transcription

1 ENGLISH ONLY EXPERT COMMITTEE ON BIOLOGICAL STANDARDIZATION Geneva, 17 to 20 October 2017 Report on the WHO collaborative study to establish the 1st International Standard for antiserum to Respiratory Syncytial Virus Jacqueline U McDonald 1, Peter Rigsby 2, Thomas Dougall 2, Othmar G Engelhardt 1 and Study Participants 1 Division of Virology and 2 Biostatistics National Institute for Biological Standards and Control (NIBSC), South Mimms, Potters Bar, Herts, EN6 3QG, UK 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 T. Zhou at zhout@who.int. World Health Organization 2017 All rights reserved. Publications of the World Health Organization are available on the WHO web site ( or can be purchased from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: ; fax: ; 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: ( 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.

2 Page 2 Summary A collaborative study was conducted with the aim to establish the 1 st International Standard for antiserum to RSV. Two candidate standards were produced from serum samples donated by healthy adult individuals in the US. The candidate standards are intended to standardize RSV neutralization assays across multiple assay formats. These assays are particularly useful in the evaluation of immunogenicity of RSV vaccine candidates. The candidates were processed, filled and freeze-dried at NIBSC. The study consisted of 21 laboratories from 9 countries and included university laboratories, manufacturers/developers of RSV vaccines and public health laboratories. All participants used their own in-house virus neutralization assay and their own virus stocks. The study samples comprised the two candidate standards,16/284 and 16/322, naturally infected adult sera, age stratified naturally infected paediatric sera, sera from RSV vaccine clinical trials in maternal and elderly subjects, a monoclonal antibody to RSV (palivizumab), two cotton rat serum samples and samples from the BEI Resources panel of human antiserum and immune globulin to RSV. The collaborative study showed that between-laboratory variability in neutralization titres was significantly reduced when values were expressed relative to those of either of the two candidate international standards. Stability of 16/284 maintained for 6 months at different temperatures showed no significant loss of activity (relative to that at -20 o C storage temperature) at temperatures of up to +20 o C. Stability data are not yet available for 16/322. From these results, 16/284 is recommended as the 1 st international standard for antiserum to RSV, with 16/322 as a potential replacement for 16/284 in the future, with an assigned unitage of 1,000 and 960 International Units (IU) of anti-rsv neutralising antibodies per vial, respectively. Introduction Development of an RSV vaccine is recognised as a global priority by national governments, the World Health Organization, the pharmaceutical industry and not for profit health organisations. Activity in this area has increased significantly in recent years, with at least 51 RSV vaccine candidates in development, 14 of which are now in human clinical trials (PATH. 2017). RSV neutralising activity in serum has been reported to correlate with protection against RSV acute lower respiratory infection in both rodent models and human infants (Graham. 2016). Quantifying this neutralising activity is vital in the development of future RSV vaccines. RSV neutralization assays come in multiple formats and one of the challenges in RSV vaccine research is accurately comparing the neutralization titres in sera from multiple clinical trials, each using a different neutralization assay format (Hosken et al. 2017). A reference antiserum is needed to standardize clinical trials and outcomes. PATH conducted a multi-laboratory RSV neutralization assay survey study of 12 diverse assay formats and found that it was feasible to harmonize neutralization results using a standard (Hosken et al. 2017). Pooled human serum confirmed as seropositive for RSV is being proposed as the candidate material to be assessed for the international standard.

3 Page 3 Aim of the Study The aim of the study was to characterize two candidate anti-rsv sera in diverse RSV neutralization assays to assess their suitability to be used as the 1 st international standard for antiserum to RSV. Age stratified paediatric serum pools and vaccinee serum pools from 3 separate clinical trials were also evaluated to establish commutability of the standard. The BEI Resources panel of human antiserum and immune globulin to RSV (NR-32832) was also included to allow comparison with and potential calibration against the proposed international standard, as these materials are currently being used by some laboratories as working standards. Materials and Methods Bulk materials and processing Product summary A donation of sixty serum samples from healthy adults was provided by PATH as source material for the proposed international standard. The samples were confirmed to be negative for HBsAg, HIV, HBV, HCV and syphilis antibodies. All samples were positive for antibodies against RSV. From the 60 samples, 12 samples with high and medium RSV antibody titre, as determined in the PATH harmonization study (Hosken et al. 2017), were selected for two candidate pools. The two candidate pools were filled and freeze-dried at NIBSC. The first fill and freeze-dry was completed in October 2016 and the second pool was filled and freeze-dried in January For both materials, 0.5ml was filled in each vial and freeze-dried. The two candidates gave yellowish, robust cakes. Pooling, filling, freeze-drying and sealing Serum samples were received from PATH, as frozen samples in dry ice, and then stored at - 80 o C. On the day before filling, the appropriate bulk sera were thawed at room temperature, pooled in a sterile vessel and delivered to the filling plant; six samples were pooled for each candidate standard. The pooled serum was filtered prior to filling and kept at +4 o C. Filling was completed for both candidate pools from homogenous stirred bulks, which were maintained at between 4 o C and 8 o C during fill. 3ml ampoules were filled with 0.5ml of material. Freezedrying was carried out immediately after filling using a 4 day cycle, after which the completed product was kept at -20 o C for long term storage. Ampoules were sealed under boil-off gas from high purity liquid nitrogen (99.99%) and measurement of the mean oxygen head space after sealing served as a measure of ampoule integrity. The mean oxygen head space was measured non-invasively by frequency modulated spectroscopy (FMS 760, Lighthouse Instruments, Charlottesville, USA). Residual moisture content was measured using the colorimetric Karl Fischer method in a dry box environment (Mitsubishi CA100, A1 Envirosciences, Cramlington, UK) with total moisture expressed as a percentage of the mean dry weight of the ampoule contents.

4 Page 4 Product summary details for each of the filled samples are shown in Table 1. Fill dates for the candidate materials are detailed below: Anti-RSV serum Pool 1 (NIBSC 16/284) 20 th October 2016 Anti-RSV serum Pool 2 (NIBSC 16/322) 12 th January 2017 Post filling testing of freeze-dried samples One ampoule for each of the candidate standards 16/284 and 16/322 from the beginning, middle and end of the freeze drying process were tested in two repeat assays to check the effect of the freeze drying process on RSV neutralization antibody activity. This testing was performed against RSV/A2 virus strain, using a NIBSC in-house RSV neutralization assay. The GMTs of the beginning, middle and end samples of 16/284 were 1657, 1714 and 1493 respectively (119%, 123% and 107% compared to pre-lyophilised 16/284); and for 16/322 were 2071, 1731 and 1459 respectively (160%, 134% and 113% compared to pre-lyophilised 16/322). Overall there were no losses in activity for each of the candidates for any of the time points during the freeze drying process. The filled material was therefore fit to be used in the collaborative study as candidate standards. Study Samples A total of thirty-eight samples were available to participants. The samples were shipped in dryice and storage at -20 o C was recommended. All twenty-five participating labs were provided with the core panel and the panel for commutability listed in Table 2, except for 2 labs that were unable to receive the cotton rat and paediatric samples due to internal processes. Of the 25 labs, 19 were able to receive the BEI Resources panel. Six labs were not able to receive this panel due to administrative reasons. BEI samples were reconstituted according to instructions from BEI materials. Sample 014 (NR ) was diluted 1/10 before being sent to participants. All titres shown for this sample are for the 1/10 diluted product. The international unit assigned to this sample has been converted to take this dilution into account. A sample information sheet explaining how to store and handle the samples was sent with each package. There were no issues with the shipment and the receipt of samples for the majority of participating labs. Receipt of the samples was confirmed by 24 of the 25 labs. Design of Collaborative Study Participants Twenty six laboratories were invited to participate in the study. Twenty five laboratories from twelve countries agreed to participate. Twenty one labs returned data; two of these returned two sets of data, one lab using two distinct assay methods and the other using two virus strains, giving a total of twenty three datasets. They are referred to by a code number, allocated at random, and not reflecting the order of listing in Appendix 1.

5 Page 5 Study time-frame The study samples were sent to participants at the end of January 2017 and the coordinator requested data to be returned by the 20 th March Due to issues with import of study samples for some participants, some panels were sent late and the deadline for the affected participants was extended as much as possible. The majority of participants were able to return data by mid- April Study Plan Participants were requested to: - Follow the recommended protocols for storage and reconstitution of the study samples. - Determine the neutralization titre against RSV of each of the samples in the panel by performing four independent assays using their in-house method and using in-house reagents, including their own virus stocks. - Avoid multiple freeze-thaw cycles of the study samples. Laboratory methods Participants used their own in-house assays for determining RSV neutralization antibody titres. Each participant used their own virus stocks. Participants provided calculated titres based on their own in-house method of titre calculation. In most cases, an ED50 (or equivalent) was provided. In cases where this was not possible, participants provided the raw data and statisticians at NIBSC calculated an ED50 using Combistats. Details of participants in-house methods can be found in Table 3. Documentation of Study results Participants were requested to report their results electronically using standard forms provided by the study coordinator. The forms requested both raw data and calculated endpoint titres. Statistical Analysis Analysis was performed using ED50s reported by the participants or calculated at NIBSC and also using relative potencies, i.e. ED50s expressed relative to candidate standard samples. All ED50s and relative potency estimates were combined as Geometric Means (GM) and variability within laboratories (between assays) and between laboratories was expressed using Geometric Coefficients of Variation (%GCV), i.e. (10 s -1)x100%, where s is the standard deviation of the log 10 ED50s or potency estimates. In some cases, the endpoint ED50 was not covered by the range of dilutions used by the participant and results were reported as less than or greater than and all estimates for that sample in that laboratory were excluded from further analysis. Any exclusions due to high intra-assay or inter-assay variability within a laboratory are described in the results section of this report.

6 Page 6 An exploratory visual assessment of sample and laboratory differences was carried out by performing a simple correspondence analysis of potencies relative to candidate standard sample 11 (16/284). Further assessment of agreement in geometric mean potencies for each pair of laboratories was performed by calculating Lin s concordance correlation coefficient with log transformed potencies relative to candidate standard samples 11 or 24 (16/284 and 16/322 respectively). Laboratory 7, 11a & 13 These labs reported their data in a format other than ED50 or equivalent. To allow for like for like comparisons ED50 values for these datasets were calculated using Combistats. The recalculation could affect the within assay variations for these labs, as the assays are not optimized to calculate an ED50. Stability Studies Samples of the candidate standards were stored at -70 C, -20 C, +4 C, +20 C, + 37 C, +45 C and +56 C and 16/284 was tested after 6 months storage for RSV neutralization activity. Candidate 16/322, which was filled later than 16/284, will be tested at the 6 month mark. Samples will be tested at regular intervals to assess long term stability. The stability of candidate standards reconstituted in liquid form was also assessed. The candidate standards were reconstituted and stored at +4 C, +22 C and + 37 C for up to four weeks then tested for RSV neutralization activity. Samples for stability were tested using a NIBSC in-house RSV neutralization assay. Results Study data returned A total of 23 datasets were received from 21 out of 25 participants. Laboratory 11 returned 2 datasets using 2 different neutralization assay formats, and laboratory 14 returned 2 datasets using 2 different virus strains. The data from laboratory 6 were not included in the analysis, as the results returned did not give estimated values for either candidate standard but instead returned values of >1024. Intra-assay and inter-assay variability in ED50s Intra-assay variability was assessed using the coded duplicate samples included in the study. Three pairs of coded duplicates were included in the panel of samples, these were; 1 & 3, 9 & 29, and 16 & 26. Where any of the coded duplicate ED50s within an assay differed by more than a factor of 2.5, no result for that assay was used, see Figure 1. The excluded assays accounted for ~8% of returned assays.

7 Page 7 Inter-assay variability was assessed for each sample from the ratio of the maximum and minimum ED50 results across all assays within a laboratory. Where the ratio for a sample exceeded 3.5, all results for that sample for that laboratory were excluded. Further to the excluded assays, ~5% of samples were excluded, see Figure 2. For each laboratory and sample, GM ED50s and relative potencies, together with inter-assay GCVs are shown in Appendix 2, following the exclusions detailed above. Correspondence analysis outcomes A simple correspondence analysis of log transformed potencies relative to 16/284 was carried out after exclusion of sample 32 (negative) and laboratory 15 (only limited data available). This multivariate statistical technique allows an exploratory visual assessment of the results profiles observed for samples (across different laboratories) or laboratories (across different samples). In the small number of cases (less than 5%) where no result was reported, the median result for that sample from all other laboratories was included in place of the missing value for the purposes of this analysis. A scatterplot of the first three sample components is shown in Figure 3. Samples exhibiting similar results profiles across different laboratories would be expected to be in close proximity on this plot. In this case, there is evidence that results profiles across laboratories are different for the animal and monoclonal antibody (mab) panel samples. No particular groupings of laboratories were observed and plots of laboratory components are not shown. Agreement between laboratories Concordance correlation coefficients are summarised in Tables 4 and 5 for potencies relative to candidate standards 16/284 and 16/322 respectively, based on the human panel samples only. Scatterplots of log potencies relative to 16/284 and 16/322 for all samples and all laboratory pairs are shown in Figures 4a, 4b, 5a and 5b. Concordance was noted to be poor for laboratories 02, 12, 13, 14b, 15, 17, 19, 20 and 21, all having concordance correlation coefficients <0.80 in more than 50% of cases shown in Tables 4 and 5. Levels of inter-laboratory variability have been calculated with (set A) and without (set B) these laboratories (Table 7) to illustrate the reduced level of variability that was achieved between the remaining laboratories (n=13). Inter-laboratory variability in ED50s and relative potencies Variability between laboratories for ED50s and potencies relative to different candidate standards was assessed using the inter-laboratory GCV values and ratios of maximum and minimum estimates shown in Tables 6a and 6b and Figures 6a to 6f. A summary of the calculated GCVs can be found in Table 7 and median GCVs are summarised in Table 8. Animal samples 13 & 27, mab samples 10 & 22, and sample 32, an IgG deficient serum sample included as a negative control, will be excluded from all further discussions of data unless specifically related to these samples.

8 Page 8 ED50 Results Using the data without the removal of any laboratories (set A), with the exception of sample 7 (82%), all between-laboratory GCVs were above 100%. The GCVs ranged from 82% to 186%. When using data excluding 9 laboratories according to the concordance analysis described above (set B), the GCVs range from 54% to 173%. Potencies relative to 16/284 For all samples tested, the GCVs show better agreement between laboratories when the titres are expressed relative to 16/284 than when end point titres are used. When using data set A, the GCVs for relative potencies were between 26% and 86%. There was a reduction in variability for the median GCVs, from 143% to 54%. When using data set B, the GCVs for relative potencies ranged from 15% to 72%. There was a reduction in the median GCVs, from 114% to 33%. Potencies relative to 16/322 These also show better agreement between laboratories than end point titres for all samples tested. When using data set A, the GCVs for relative potencies were between 24% and 89%. There was a reduction in the median GCVs, from 143% to 56%. When using data set B, the GCVs for relative potencies ranged from 16% to 70%. There was a reduction in median GCVs, from 114% to 32%. Adult Human Sera (Natural Infection) The median GCV for this group of samples was 143% when using data set A, and 114% with data set B. When expressed relative to 16/284, the GCV was reduced to 48% with data set A, and 22% when using data set B. When expressed relative to 16/322, the GCV was reduced to 43% with data set A, and 31% when using data set B. BEI Resources Samples The median GCV for the BEI resources samples was 130% with data set A, and 65% with data set B. When expressed relative to 16/284, the GCV was reduced to 45% with data set A, and 25% with data set B. When expressed relative to 16/322, the GCV was reduced to 50% with data set A, and 28% with data set B. Potencies relative to samples 007, 014, and 019 were also calculated to assess the ability of these samples to perform as working standards. These samples were able to reduce the GCVs of all samples as successfully as the candidate standards (Tables 6b, 7 and 8). Paediatric Samples The median GCV for the paediatric samples was 158% with data set A, and 118% with data set B. When expressed relative to 16/284, the GCV was reduced to 60% with data set A, and 37% with data set B. When expressed relative to 16/322, the GCV was reduced to 55% with dataset A, and 30% with data set B. Vaccinee Samples The median GCV for the vaccinee samples was 139% with data set A and 116% with data set B. When expressed relative to 16/284, the median GCV was reduced to 65% with data set A, and

9 Page 9 35% with data set B. When expressed relative to 16/322, the median GCV was reduced to 60% with dataset A, and 35% with data set B. Animal Sera Cotton rat sera included in the study showed higher GCVs than all other sample groupings, with a median GCV of 222%, when using data set A, and a median GCV of 159% when using data set B. When expressed relative to 16/284, the GCV was reduced to 119% with data set A, and 67% with data set B. When expressed relative to 16/322, the GCV was reduced to 94% with data set A, and 53% with data set B. Monoclonal Antibodies The median GCV for the monoclonal antibody sample was 154% when using data set A, and 132% with data set B. When expressed relative to 16/284, the GCV was reduced to 83% with data set A, and 71% with data set B. When expressed relative to 16/322, the GCV was reduced to 88% with data set A, and 87% with data set B. Stability study results At the time of this report only data up to the 6 month time point for 16/284 were available for evaluation. No accelerated stability data are yet available for 16/322. After 6 months storage at elevated temperatures, two independent assays were performed for each temperature, and potencies relative to the -20 C baseline for 16/284 were obtained. The ED50s and GMTs are shown in Table 9 and the estimated RSV neutralization activity loss per month and year are shown in Table 10. The data show that there is no loss of activity at +4 C and minimal loss at +20 C (temperatures used during laboratory manipulation for assays), relative to the -20 C baseline. The low predicted loss in activity per year (<0.01%) when stored at -20 o C suggests 16/284 is sufficiently stable to serve as a WHO international standard. The stability of 16/284 and 16/322 will be monitored regularly throughout the life time of the standards. The stability of candidate standards reconstituted in liquid form was also assessed (Table 11a and 11b). Ampoules of the three candidate materials that had been reconstituted in 0.5ml of sterile glass distilled water and maintained at different temperatures were tested. Two independent assays were performed at each temperature and time point, and potencies relative to that of an ampoule stored at -20 C and freshly reconstituted were obtained for both candidates. The data suggest that there is no loss of activity for any of the reconstituted materials that had been stored at +4 C or +22 C for both standards for up to a week. After 4 weeks at +4 C, 16/284 had lost 24% of its neutralization activity. At +37 C, 16/284 showed approximately 20% loss after 2 weeks, and 16/322 had lost >40% activity after 1 week. Discussion The data from this study show that both 16/284 and 16/322 are suitable as reference standards to measure RSV neutralization activity in a range of sample types, particularly human serum, due to their ability to reduce GCVs across the various assay methods included in this study. We propose 16/284 as the first International Standard for antiserum to RSV, with an assigned

10 Page 10 potency of 1000 IU/ampoule, as stability data for this candidate are already available. Based on this proposal and the geometric mean potency of 16/322 relative to 16/284 (0.96; sample 24 in Table 6a; no laboratory results detected as outliers), 16/322 can be assigned a potency of 960 IU/ampoule and may be used as a replacement standard, once 16/284 has been depleted. The data generated showed high GCVs across the laboratories for all sample types. This can be attributed to the varying formats and methods used to assess RSV neutralizing antibody titres. A deliberate decision was made that no aspect of the methods used would be controlled, to mimic the real world variation of the assays currently being used. However, once the data were expressed as potencies relative to the candidates, the GCVs were significantly decreased across all laboratories. Using data with the exclusion of 9 laboratories further reduced the GCVs. However, there was no common aspect of the methods between the excluded laboratories, or an obvious reason why concordance correlation coefficients should be so low. There were two exceptions to this: laboratory 14b, which looked at neutralization titres against RSV B, and laboratory 17 which included guinea pig complement in their assay. Further collaborative studies will be needed to determine the usefulness of this standard against RSV B viruses. Stability data for 16/284, based on storage at elevated temperatures up to six months, gave a low predicted loss in activity per year (<0.01%) when stored at -20 o C, suggesting suitable stablilty to serve as a WHO IS. A long term programme of monitoring stability will be needed to show that 16/284 remains stable over its life time. Stability data for 16/322 is not currently available but it will also be monitored for stability over its life time. Furthermore, stability analysis showed that the candidate standards were also stable after reconstitution. Both candidates showed loss of activity at 37ºC after 2 weeks, with 16/322 showing a greater loss than 16/284. This study has achieved its stated aims and an International Standard is proposed, with an International Unitage (IU). This study has shown that the standard is useful for multiple sample types across a wide variety of assay formats; however, the analysis suggests that the cotton rat serum samples and monoclonal antibody samples behave differently from the human serum samples, and that a more suitable standard should be considered for those sample types. This is not an issue for this International Standard, as its main role will be to look at neutralising antibody activity in human serum, mostly produced in RSV vaccine clinical trials. The BEI Resources panel consisted of a BEI materials panel of human antisera and immunoglobulin. These were included to assess their ability to act as working standards and data from the collaborative study support their suitability, as they are able to reduce GCV when used as standards. The proposed international unitage of these samples can be found in Table 12. Commutability for vaccine serum samples and paediatric samples was assessed in this collaborative study. Sera from both maternal and elderly clinical trials were included and both candidate standards were able to reduce GCVs for these samples. Naturally infected paediatric sera were also assessed and both candidates were able to reduce the GCVs across all laboratories for these samples. Vaccinee sera from paediatric clinical trials were not available for testing. As well as the inter-laboratory GCVs, correspondence analysis did not suggest differences in the behaviours of the vaccinee sera and paediatric sera in comparison to the adult human naturally

11 Page 11 infected sera, which further indicates that the candidate standards are commutable with these sample types. Recommendations It is proposed that the candidate 16/284 should be established as the 1 st International Standard for antiserum to RSV to be used for the standardization of RSV neutralization assays. The assigned potency for this should be: 1,000 IU/ampoule It is proposed that the candidate 16/322 should be retained as a potential secondary/replacement International Standard for antiserum to RSV to be used for the standardization of RSV neutralization assays. The assigned potency for this should be: 960 IU/ampoule Comments from Participants All participants were requested to comment on the draft report. Three out of the 21 did not respond to request for comments. There were no disagreements with the suitability of the candidate IS (NIBSC 16/284) to serve as the 1 st international standard for antiserum to RSV. Some responders had minor queries or suggestions for editorial changes and these have been addressed. Acknowledgements We gratefully acknowledge the important contributions of the collaborative study participants. We would also like to thank NIBSC Standards Production and Dispatch for the filling, freeze-drying and distribution of the candidate material. We also greatfully acknowledge Glaxo Smith Kline (Belgium), MedImmune (US), Novavax (US), Professor Andrew Pollard (Oxford University, UK) and Professor Pedro A. Piedra (Baylor College of Medicine, US) for donating samples included in this study. We also acknowledge PATH for arranging blood collection from those who gave consent to the use of their sera in preparing the candidate material, donating said material to NIBSC and funding this study. References - Graham B. Vaccines against respiratory syncytial virus: The time has finally come. Vaccine Jun 24;34(30): DOI: /j.vaccine PATH. RSV Vaccine and mab Snapshot 3rd March Hosken N, Plikaytis B, Trujillo C, Mahmood K, Higgins D, Participating Laboratories Working Group. A multi-laboratory study of diverse RSV neutralization assays indicates

12 Page 12 feasibility for harmonization with an international standard. Vaccine May 25;35(23): DOI: /j.vaccine

13 Page 13 Table 1. Product Summary Production Summary For Cadidate International Standards NIBSC code 16/284 16/322 Presentation 3ml DIN ampoules 3ml DIN ampoules Number of containers Validation of ampoule integrity Visual inspection plus Oxygen headspace on 12 random samples Visual inspection plus Oxygen headspace on 12 random samples Sealing gas Nitrogen from liquid nitrogen 99.99% pure Nitrogen from liquid nitrogen 99.99% pure Oxygen headspace measured by Near Infra-Red spectroscopy Near Infra-Red spectroscopy Residual moisture measured by Karl Fischer reagent Karl Fischer reagent Mean fill mass g g CV fill mass 0.95% 0.62% Number of fill weights measured Mean dry weight g g CV of dry weight 0.27% 0.39% Number of dry weights measure 6 6 Mean residual moisture 1.04% 0.32% CV of residual moisture 25.50% 11.86% Number of residual moisture measurements Mean oxygen headspace 0.44% 0.44% CV of oxygen headspace 26.02% 35.69% Number of oxygen tests carried out Date of fill 20/10/ /01/2017 Storage temperature -20 C -20 C Microbial contamination None detected None detected

14 Page 14 Table 2. Samples Included in the Collaborative Study Sample Panels Name Sample Code International Standard Candidate 1- NIBSC 16/ International Standard Candidate 2 NIBSC 16/ International Standard Candidate 1 Pre-lyophilised 009 International Standard Candidate 1 Pre-lyophilised 029 International Standard Candidate 2 Pre-lyophilised 018 Core Human Panel P-0015 Individual Adult Human Serum 001 P-0015 Individual Adult Human Serum 003 P-0029 Individual Adult Human Serum 031 P-0035 Individual Adult Human Serum 038 P-0041 Individual Adult Human Serum 037 P-0056 Individual Adult Human Serum 016 P-0056 Individual Adult Human Serum 026 Core Monoclonal Palivizumab 0.1mg/ml 010 Antibody Panel Palivizumab 1mg/ml 022 Core Animal Panel Cotton Rat Serum Pool - RSV A 027 Cotton Rat Serum Pool - RSV B 013 Maternal sera pool from RSV F Trial Maternal sera pool from RSV F Trial Maternal sera pool from RSV F Trial Maternal sera pool from RSV F Trial Maternal sera pool from RSV F Trial Panel for Elderly sera pool from RSV F Trial Commutability- Vaccinee Elderly sera pool from RSV F Trial Elderly sera pool from RSV F Trial Elderly sera pool from RSV F Trial Adult sera pool from RSV F Trial - 1 (Low) 028 Adult sera pool from RSV F Trial - 2 (Med) 035 Adult sera pool from RSV F Trial - 3 (High) 030 Panel for Commutability - Paediatric BEI Resources Panel Pediatric Serum Pool Age < 1 year 005 Pediatric Serum Pool Age 1-2 years 023 Pediatric Serum Pool Age 2-3 years 033 Pediatric Serum Pool Age 3-4 years 008 NR-4020: Human Reference Antiserum to RSV 007 NR-4021: Human Antiserum to RSV, High Control 019 NR-4022: Human Antiserum to RSV, Medium Control 002 NR-4023: Human Antiserum to RSV, Low Control 036 NR-21973: Human Reference Immunoglobulin to RSV 014 NR-49447: Human IgG-Depleted Serum (Negative control) 032

15 Table 3. Details of methods used by participants Lab Code Plate Format Duration of Assay Cell Line Complement/ % and type RSV Strain [Virus] added per well Dilution Series (#dilutions / fold dilution) Time/temp (RT=room temp) Titre Determination Read-out well A2 Recombinant 5,000 GFPfluorescence Vero No 9 / 3 fold 30 mins/ RT IC50 hours GFP pfu/cell well 2 days HEp-2 No A2 200 pfu/well 12 / 2-fold 1hr/4 C ED50 Plaque counts well 2 days A549 No Recombinant 2x Firefly 10 / 8 fold EC50 FFL-A2 PFU/ml hr/37 C luciferase Direct to well 3 days Vero No A2 1,000 pfu 10 / 3-fold Veros at EC50 ELISA A C well 3 days HEp-2 No A2 75 pfu 9 / 2-fold 30 min/rt ED50 by Excel F protein EIA well 24 hours HEp-2 No A2 100 pfu/well 8 / 2-fold 1hr/37 C ED50 by Excel Plaque counts well 6 days vero No RSV A (long) well 6-7 days HEp-2 No RSV/A/Tracy well 4 days Vero No rrsv/v35px EGFP 100 pfu/well 3^3 TCID100/.05 or 3^5.5 TCID50 3,750 pfu/ml 7 / 2-fold 1hr/33 C ED60 by Excel 12 / 2-fold 90 minutes/ 36 C 7 / 2-fold 48 hr/37 C The last dilution with 50% or greater intact Hep-2 monolayer. 2-log titer at 50% by Graphpad showing LogEC50 as results Fluorescent plaques (stained with primary RSV AB and secondary FITC) and counted with fluorescence microscope reader CPE Plaque counts by spots with fluorescence

16 Page 16 Lab Code Plate Format Duration of Assay Cell Line Complement/ % and type well 24 hours Vero No RSV Strain A2 Recombinat GFP [Virus] added per well 652 ffu/well Dilution Series (#dilutions / fold dilution) Time/temp (RT=room temp) 8 / 3-fold 1hr/RT 11a 24 well 5 days HEp-2 No A pfu 6 / 4-fold 1hr/37 C 11b 96 well 4 days Vero well 7 days Vero No A well No A2 24 hours Vero No A2 14a 96 well 6 days HEp-2 No RSV/A (Long) 14b 96 well 6 days HEp-2 No RSV/B (B1 WT) well well 7 days 3 days A549 No RSV A (M37) Vero E well 2 days A549 No 5% guinea pig complement A2 A2 expressing Renilla Luciferase 500-1,000 PFU/well 200 pfu/well 500-5,000 TCID50/ well 100 TCID TCID ffu/well pfu/well 250 pfu/100 ul/well 10/3-fold 1hr/37 C Titre Determination IC50 using 2-pt interpolation of dilutions that surround 50% virus reduction 60% plaque reduction IC50 by SoftMax Pro Read-out Foci/well Plaque counts OD at 450nm 11 / 2-fold 1hr/37 C ED50 by Excel Plaque counts 2-fold 1hr/37 C ND90, EXCEL 8 / 2-fold 1hr/37 C 8 / 2-fold 1hr/37 C 10/2.5-fold 6 or 12 /2- fold 30 min/37 C 30 +/-5 mins /37 C ED50 (Reed & Muench by excel) ED50 (Reed & Muench by excel) 2-pt interpolation of dilutions that surround 50% virus reduction ED50 by loglinear regression in Excel 12 / 2 -fold 60mins/37 C ED50 by Excel RSV N gene amplicon CPE CPE Fluorescent foci Plaque counts Luciferase activity

17 Page 17 Lab Code Plate Format Duration of Assay Cell Line Complement/ % and type well 24 hours HEp-2 NO well 24 hours well 7 days HEp- 2 HEp- 2 No No well 5 days Hep-2 No RSV Strain RSV/A2 carrying the monomeric Katushka fluorescent protein (RSV A mkate) Recombinant mkate-rsv (A2) RSV A (long) RSV-A Memphis 37b [Virus] added per well MOI=4, cells/well 2x10 4 pfu/well 3500 pfu/ml Approx 2.75 log10 TCID50/ ml Dilution Series (#dilutions / fold dilution) Time/temp (RT=room temp) Titre Determination 10 / 4-fold 1 hr/37 C IC50 3 or 4 / 9 37 C 12 / 2-fold 90 mins/37 C 7 / 3-fold 30 mins/37c IC50 by curve fitting with Prism software Last dilution with > 50% intact Hep-2 monolayer ED50 (Reed & Muench by Excel) Karber calculation Read-out Mean fluorescence intensity per well Fluorescence Intensity CPE CPE

18 Page 18 Table 4. Concordance correlation coefficients for log potencies relative to 16/284 (human panel samples only); values 0.8 shaded Lab a 11b a 14b a b a b

19 Page 19 Table 5. Concordance correlation coefficients for log potencies relative to 16/322 (human panel samples only); values 0.8 shaded Lab a 11b a 14b a b a b

20 Page 20 Table 6a. Sample geometric mean ED50 and potency estimates relative to 16/284 or 16/322 Type Animal Human mab BEI Resources Paediatric Sample ED50 Potencies v 16/284 Potencies v 16/322 GM Max:Min GCV N GM Max:Min GCV N GM Max:Min GCV N

21 Page 21 Vaccinee GM: Geometric Mean Max:Min: Ratio of maximum and minimum laboratory geometric means GCV: Geometric Coefficient of Variation (%) N: Number of laboratories used in calculation of GM and GCV

22 Page 22 Table 6b. Sample geometric mean potency estimates relative to BEI Resources samples 7(NR-4020), 14(NR-21973) and 19(NR-4021) Type Animal Human mab BEI Resources Paediatric Sample Potencies v Sample 007 Potencies v Sample 014 Potencies v Sample 019 GM Max:Min GCV N GM Max:Min GCV N GM Max:Min GCV N

23 Page 23 Vaccinee GM: Geometric Mean Max:Min: Ratio of maximum and minimum laboratory geometric means GCV: Geometric Coefficient of Variation (%) N: Number of laboratories used in calculation of GM and GCV

24 Page 24 Table 7. Summary of inter-laboratory GCV values Type Human BEI Resources Paediatric Vaccinee ED50s Potencies relative to different reference standards Sample (All (With Set A (All Data) Set B (With Exclusions) Data) Exclusions) 16/284 16/322 S7 S14 S19 16/284 16/322 S7 S14 S

25 Page 25 Animal mab Shading indicates level of inter-laboratory variability: darker red = increased variability Blue shading: animal and mab samples are shaded in blue as they behave differently to human serum samples Grey shading: empty cell

26 Page 26 Table 8. Median inter-laboratory GCV values ED50 (All Data) ED50 (With Exclusions) Potencies relative to different reference standards Set A (All Data) Set B (With Exclusions) 16/284 16/ /284 16/ All Human BEI Resources Paediatric Vaccinee Animal mab Shading indicates level of inter-laboratory variability: darker red = increased variability Blue shading: animal and mab samples are shaded in blue as they behave differently to human sera samples Grey shading: empty cell

27 Page 27 Table 9. Thermal degradation assessment of 16/284; anti-rsv neutralization titres after 6 months storage Storage Temperature ( C) 1 2 GM % of -20 C Table 10. Thermal degradation assessment of 16/284; estimated percentage loss per month and year Storage Temperature ( C) % loss per month % loss per year Table 11a. Thermal degradation assessment of reconstituted 16/284; percentage relative to 16/284 stored at -20 C and reconstituted on day of assay Storage Temperature ( C) 1 week 2 weeks 4 weeks Table 11b. Thermal degradation assessment of reconstituted 16/322; percentage relative to 16/322 stored at -20 C and reconstituted on day of assay Storage Temperature ( C) 1 week 2 weeks 4 weeks

28 Page 28 Table 12. Proposed IU/mL values for BEI Resources Materials Sample Name IU/mL NR-4020: Human Reference Antiserum to RSV 1236 NR-4021: Human Antiserum to RSV, High Control 3654 NR-4022: Human Antiserum to RSV, Medium Control 843 NR-4023: Human Antiserum to RSV, Low Control 756 NR-21973: Human Reference Immune Globulin to RSV 7346

29 Page 29 Figure 1. Ratios of ED50s for coded duplicate samples in each assay and the cut-off for acceptability

30 Page 30 Figure 2. Ratios of the maximum and minimum ED50s for each sample in each laboratory and the cut-off for acceptability

31 Page 31 Figure 3. Scatterplot of row (sample) principal coordinates from correspondence analysis

32 Page 32 Figure 4a. Scatterplots of log potencies relative to 16/284 for individual laboratory pairs excluding laboratories 02, 12, 13, 14b, 17, 19, 20 and 21; solid line indicates agreement

33 Page 33 Figure 4b. Scatterplots of log potencies relative to 16/284 for individual laboratory pairs for laboratories 02, 12, 13, 14b, 17, 19, 20 and 21 only; solid line indicates agreement

34 Page 34 Figure 5a. Scatterplots of log potencies relative to 16/322 for individual laboratory pairs excluding laboratories 02, 12, 13, 14b, 17, 19, 20 and 21; solid line indicates agreement