WHO biosafety risk assessment and guidelines for the ... WHO/INFLUENZA/DRAFT/NOVEMBER 2017 3 ENGLISH ONLY 4 5 WHO biosafety risk assessment and guidelines for the production 6 and

1

WHO/INFLUENZA/DRAFT/NOVEMBER 2017 2

ENGLISH ONLY 3

4

WHO biosafety risk assessment and guidelines for the production 5

and quality control of novel human influenza candidate vaccine 6

viruses and pandemic vaccines 7

(Proposed revision of WHO TRS No. 941, Annex 5; H1N1 specific update 2009; and H7N9 8

update 2013) 9

10

NOTE: 11

This document has been prepared for the purpose of inviting comments and suggestions on the 12

proposals contained therein, which will then be considered by the Expert Committee on 13

Biological Standardization (ECBS). Publication of this early draft is to provide information 14

about the proposed WHO biosafety risk assessment and guidelines for the production and 15

quality control of novel human influenza candidate vaccine viruses and pandemic vaccines to a 16

broad audience and to improve transparency of the consultation process. 17

The text in its present form does not necessarily represent an agreed formulation of the 18

Expert Committee. Written comments proposing modifications to this text MUST be 19

received by 12 February 2018 in the Comment Form available separately and should be 20

addressed to the World Health Organization, 1211 Geneva 27, Switzerland, attention: 21

Department of Essential Medicines and Health Products (EMP). Comments may be submitted 22

electronically to the Responsible Officer: Dr TieQun Zhou at email: [email protected]. 23

The outcome of the deliberations of the Expert Committee will be published in the WHO 24

Technical Report Series. The final agreed formulation of the document will be edited to be in 25

conformity with the "WHO style guide" (WHO/IMD/PUB/04.1). 26

© World Health Organization 2017 27

All rights reserved. Publications of the World Health Organization can be obtained from WHO Press, World Health 28

Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: 29

[email protected]). Requests for permission to reproduce or translate WHO publications – whether for sale or for 30

non-commercial distribution – should be addressed to WHO Press, at the above address (fax: +41 22 791 4806; e-31

mail: [email protected]). 32

The designations employed and the presentation of the material in this publication do not imply the expression of 33

any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, 34

territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on 35

maps represent approximate border lines for which there may not yet be full agreement. 36

WHO/INFLUENZA/DRAFT/NOVEMBER 2017

Page 2

The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or 1

recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. 2

Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters. 3

All reasonable precautions have been taken by the World Health Organization to verify the information contained in 4

this publication. However, the published material is being distributed without warranty of any kind, either expressed 5

or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall the 6

World Health Organization be liable for damages arising from its use. 7

The named authors [or editors as appropriate] alone are responsible for the views expressed in this publication. 8

9

10

11


Page 3

Contents 1

2

Summary 3

1. Introduction 4

2. Scope 5

3. Terminology 6

4. Hazard identification 7

5. Safety Testing of candidate vaccine viruses 8

6. Risk assessment and management 9

Authors and acknowledgements 10

References 11

Appendix 1. Testing for attenuation of influenza vaccine viruses in ferrets 12

13

14

15


Page 4

Abbreviations 1

A/Len/17 Influenza A/Leningrad/134/17/57 virus

BSL Biosafety Level

CVV Candidate vaccine virus

EID Egg infectious dose

GAP WHO Global Action Plan for Influenza Vaccines

GISRS WHO's Global Influenza Surveillance and Response System

GMP Good manufacturing practice

HA Haemagglutinin

H&E haemotoxylin and eosin (staining)

HEPA High-efficiency particulate air filtration

HP High pathogenicity

HPAI Highly pathogenic avian influenza virus

IVPI (Chicken) Intravenous pathogenicity index. Any virus with an index greater

than 1.2 is considered an HPAI

LP Low Pathogenicity

LPAI Low-pathogenic avian influenza virus

MDCK Madin-Darbey Canine Kidney (cells)

MVA Modified Vaccine Ankara

NA Neuraminidase

OIE World Organization for Animal Health

PAPR Powered air-purifying respirators

PFU Plaque forming unit

PPE Personal protection equipment

PR8 Influenza A/Puerto Rico/8/34 virus (A/PR8/34)

PTC Pass through cabinets

RG Reverse Genetics

TCID Tissue culture infective dose

VCM WHO Vaccine Composition meeting for Influenza Vaccines

VSV Vesicular stomatitis virus

WHO CC WHO Collaborating Centre for Influenza

WHO ERL WHO Essential Regulatory Laboratory for Influenza

wt Wild-type


Page 5

Summary 1

International biosafety expectations for both the pilot-scale and large-scale production of 2

human vaccines in response to a pandemic influenza virus, and the quality control of these 3

vaccines, are described in detail in these WHO Guidelines. Tests required to evaluate the safety 4

of candidate influenza vaccine viruses (CVVs) prior to release to vaccine manufacturers are 5

also specified in this document, which is thus relevant to both development and production 6

activities, and also to vaccine and biosafety regulators. A detailed risk assessment is presented 7

that concludes that the likelihood of direct harm to human health would be high if reassorted 8

H5 or H7 viruses or other subtypes that have high in vivo pathogenicity are used for vaccine 9

production; moreover, low pathogenic avian influenza viruses that are highly virulent for 10

humans may also present a high risk and all such viruses could also pose a significant risk to 11

animal health. Stringent vaccine biosafety control measures, defined as Biosafety Level (BSL)3 12

enhanced, are defined to manage the risk from vaccine production and quality control using 13

such viruses in the pre-pandemic period. For all other vaccine viruses, for example reassortants 14

containing a highly attenuated backbone (e.g. from PR8) or the handling of low pathogenic 15

avian influenza viruses that are not highly virulent for humans, the direct risk to human health 16

is considered to be low. Nevertheless, there is an indirect risk to human health due to a 17

theoretical risk of secondary reassortment with circulating human influenza viruses, resulting in 18

a novel virus capable of replicating in humans. Although extremely unlikely, such a secondary 19

reassortant could become adapted to human infection and transmission and result in serious 20

public health consequences. The biosafety control measures that are proposed, defined as BSL2 21

enhanced (pandemic influenza vaccine), take this and also potential risks to animal health, into 22

account. Specifications for personal protection are provided for both BSL2 enhanced and BSL3 23

enhanced biosafety levels, and guidance is provided on biosafety management and 24

implementation within a vaccine production facility. Tests to be performed on CVVs prior to 25

release to vaccine manufacturers depend on the type of virus but include, at a minimum, in vivo 26

tests in ferrets and, where appropriate, chickens and embryonated eggs embryos, plaque assays 27

and sequencing. 28

29

30


Page 6

1. Introduction 1

Due to the inherent risks in handling pandemic influenza virus(es) for the production of 2

vaccines where an uncontrolled release could have a significant public health impact, risk 3

assessment, biosafety and biosecurity precautions are needed in laboratory and manufacturing 4

environments. WHO developed a biosafety risk assessment and guidelines for the production 5

and quality control of human influenza pandemic vaccines in 2005, published as Annex 5 to 6

TRS 941, in response to the threat of a pandemic posed by the highly pathogenic avian 7

influenza (HPAI) viruses A(H5N1) and the need to begin development of experimental 8

vaccines (2). This threat still persists and several countries have produced and stockpiled H5N1 9

vaccine. Moreover, other threats have arisen since the appearance of H5N1, such as the 10

H1N1pdm09 subtype virus causing the pandemic in 2009 and the emergence of low pathogenic 11

avian influenza (LPAI) viruses A(H7N9) that were able to replicate in humans causing severe 12

disease with a high fatality rate (mostly in adults and the elderly). The document was updated 13

to include pandemic influenza A(H1N1) virus in 2009 and A(H7N9) in 2013 (3-5). However, it 14

has been over a decade since the original publication of the WHO guideline and a revision was 15

requested by industry, regulators and GISRS laboratories during several WHO informal 16

consultations, such as the WHO Vaccine Composition Meetings (VCM), Global Action Plan 17

for Influenza Vaccines (GAP) meetings and ‘Switch’ meetings (Meetings on Influenza vaccine 18

response at the start of a pandemic), in order to review and reduce testing timelines for CVVs, 19

which have been identified as one of the bottlenecks to rapid vaccine responses (6-10). 20

Moreover, since the initial publication of TRS 941 Annex 5 guidance experience with viruses 21

of pandemic potential and with pandemic viruses has increased globally and is reflected in this 22

update. Further experience gained to date from the development and testing of CVVs derived 23

by reverse genetics (RG) from HPAI viruses has also been incorporated. WHO convened a 24

Working Group meeting during 9-10 May 2017 attended by experts and representatives of 25

WHO Collaborating Centres (CC), Essential Regulatory Laboratories (ERL) and other national 26

regulatory authorities for vaccine and biosafety regulation, manufacturers, and the World 27

Organisation for Animal Health (OIE) to review the up-to-date experience, discuss the revision 28

of TRS 941, Annex 5 and reach consensus on the outline and key issues for the revision (11). 29

This document follows the risk-assessment scheme used in the WHO biosafety guidance for 30

pilot-lot vaccine production, but is extended to include considerations relating to the greater 31

production-scale needed to supply large quantities of vaccines (12, 13). The risks associated 32

with large-scale production are likely to be different from pilot lots, e.g. the “open” aspect of 33


Page 7

some production processes and quantity of virus-containing waste. It also takes into account the 1

considerable experience gained from handling HPAI viruses and those classed as low 2

pathogenic in avian species but highly virulent in humans, as well as the hazards associated 3

with such viruses. 4

Furthermore, the range of options for vaccine development is broader than originally 5

considered in the WHO risk assessment for pilot lot production and the present document has 6

been expanded to encompass current vaccine development pathways. 7

2. Scope 8

Efforts have continued into the development and manufacture of H5N1 vaccines and the 9

guidance presented in this document reflects experience gained with this virus and our greater 10

knowledge of H5 subtype viruses in general. Moreover, a great deal has also been learned from 11

our experience with the A/H1N1pdm09 viruses, and from the production of vaccines to this 12

pandemic virus. It is, nevertheless, intended that the guidance will also be applicable to threats 13

from other potential pandemic viruses (e.g., H2, H9, H7 subtype viruses, etc.), as well as low 14

pathogenic avian and mammalian influenza viruses of various subtypes, which are potentially 15

virulent in humans. Manufacturers and laboratories handling HPAI should consult their 16

national regulatory authorities to determine whether additional biosecurity measures are 17

required to handle these viruses. 18

Transmission and pathogenicity of influenza viruses are multi-factorial traits that are currently 19

not completely understood (14). There is therefore a significant range of diversity in the 20

pathogenicity of viruses used to make CVVs which are then used in vaccine production, not 21

only for humans but also for other mammals and avian species. The haemagglutinin (HA) 22

protein has been identified as a major determinant for virulence of avian influenza viruses (15). 23

For example, H5N1 HPAI viruses that can also cause fatal disease in humans have been used to 24

produce reassortant viruses containing an HA that has been genetically modified to generate 25

viruses of low pathogenicity for chickens. In case of viruses that are inherently less pathogenic 26

for humans, wild-type (wt) virus might be used directly for vaccine production (16). Thus, 27

reassortants derived by classical means, RG, or synthetic means, which may or may not be 28

genetically modified, as well as wt viruses are within the scope of these guidelines. 29

Eggs have traditionally been used for the production of influenza vaccines from defined 30

influenza virus reassortants, but cell culture techniques and recombinant protein technologies 31

have also been established and licensed for seasonal influenza vaccine production with 32


Page 8

international expectations for production and quality control specifications defined (17). The 1

principles outlined in this guidance should be applicable to these and other technology 2

platforms for pandemic vaccine development. 3

Most effort to date with candidate pandemic vaccine development has been targeted towards 4

inactivated vaccines; however, in both North America and Europe live attenuated pandemic 5

virus vaccines have been produced and a ‘mock dossier’ approach has been accepted by the 6

European Medicines Agency (EMA). This raises important issues beyond the risks to humans, 7

namely the potential for excreted viruses or their derivatives to infect and replicate in non-8

human species particularly in those raised for commercial purposes, which could have a 9

significant economic impact as well as ramifications for international trade. Developers and 10

regulators will need to assess both the human and the agricultural risk of live pandemic virus 11

vaccines for their ability to shed and to replicate in different hosts. Both inactivated and live 12

vaccines are therefore covered in the scope of these guidelines. 13

Technologies not covered by this guideline, although the general principles could be applied, 14

include new generation technology platforms that do not use live influenza vaccine viruses for 15

production such as expressed recombinant proteins, virus-like particles, DNA- and RNA-based 16

vaccines and vectored vaccines. 17

Furthermore, it is intended that the guidelines on containment measures should apply to all 18

facilities and laboratories that handle live vaccine virus. This includes not only the vaccine 19

manufacturing facility but also applies to the quality control laboratories of the manufacturer as 20

well as National Control Laboratories and specialist laboratories as appropriate. The transport 21

of live virus materials within and between sites should comply with international and national 22

specifications (18). 23

Finally, it should be noted that the risk assessments for vaccine manufacture will vary 24

according by whether production is occurring in an interpandemic period, in a pandemic alert 25

period (as for example early in 2004 when H5N1 was threatening to circulate extensively in 26

South East Asia) or in a pandemic period. These guidelines are intended to describe steps to 27

minimize risks involving vaccine manufacture with emphasis on the interpandemic period, 28

while indicating modifications that may be found appropriate during other periods. 29

3. Terminology 30

The definitions given below apply to the terms used in these guidelines. They may have 31

different meanings in other contexts. 32


Page 9

Aerosol: A dispersion of solid or liquid particles of microscopic size in a gaseous medium. 1

Airlock: Areas found at entrances or exits of rooms that prevent air in one space from entering 2

another space. These generally have two doors and a separate exhaust ventilation system. In 3

some cases a multiple-chamber airlock consisting of two or more airlocks joined together is 4

used for additional control. 5

Biosafety Committee: An institutional committee of individuals versed in the subject of 6

containment and handling of infectious materials. 7

Biosafety manual: A comprehensive document describing the physical and operational 8

practices of the laboratory facility with particular reference to infectious materials. 9

Biosafety officer: A staff member of an institution who has expertise in microbiology and 10

infectious materials, and has the responsibility for ensuring the physical and operational 11

practices of various biosafety levels are carried out in accordance with the standard procedures 12

of the institution. 13

Biosafety Level 2, enhanced or BSL3: A specification for the containment of pandemic 14

influenza during vaccine manufacture and quality control testing with specialized air handling 15

systems, waste effluent treatment, immunization of staff, specialized training, and validation 16

and documentation of physical and operational requirements. 17

Biosecurity: Laboratory biosecurity describes the protection, control and accountability for 18

valuable biological materials within laboratories, in order to prevent their unauthorized access, 19

loss, theft, misuse, diversion or intentional release. 20

Decontamination: A process by which an object or material is freed of contaminating agents. 21

EID 50: Egg Infetious Dose. A potency unit for measuring infectious activity of a biologic 22

product or infectious agent equal to a base-10 logarithm of amount of product or agent 23

preparation that causes infection in the 50% of embryos. 24

FFP3: A Mask that protect health professional against airborne particles and microorganisms 25

in pharmaceutical work. 26

Fumigation: The process whereby gaseous chemical is applied to an enclosed space for the 27

purpose of sterilizing the area. 28


Page 10

Good manufacturing practice (GMP): That part of quality assurance which ensures that 1

products are consistently produced as controlled to the quality standards appropriate to their 2

intended use and as required by the marketing authorization. 3

HEPA filter: A filter capable of removing at least 99.97% of all airborne particles with a mean 4

aerodynamic diameter of 0.3 micrometres. 5

Highly pathogenic avian influenza (HPAI) virus: Avian influenza viruses of subtypes H5 or 6

H7 containing a cleavage site in HA with multiple inserted amino acids (also referred to in the 7

literature as ‘multibasic’ cleavage site) and causing a systemic infection in poultry associated 8

with mortality of up to 100% (fowl plague). The designation “highly pathogenic” does not refer 9

to the virulence of these viruses in mammalian and human hosts. 10

Inactivation: To render an organism inert by application of heat, chemicals (e.g. formalin, 11

beta-propiolactone), UV irradiation or other means. 12

Low-pathogenic avian influenza (LPAI) virus: Avian influenza viruses containing an HA 13

with a single basic amino acid preceding the site of proteolytic cleavage (also referred to as 14

‘monobasic’ cleavage site) causing localized infections in poultry leading to mild or moderate 15

disease. The designation “low-pathogenic” does not refer to the virulence of these viruses in 16

mammalian and human hosts. 17

N95: A surgical mask, also known as a procedure mask, designed to catch the microorganisms 18

shed in droplets and aerosols from the wearer's mouth and nose with 95% efficiency at 19

removing particles 0.3 micron and larger when correctly fitted to the user. 20

Positive pressure laminar flow hood: An enclosure with unidirectional outflowing air, 21

generally used for product protection. 22

Primary containment: A system of containment, usually a biological safety cabinet or closed 23

container, which prevents the escape of a biological agent into the immediate working 24

environment. 25

Respirator hood: A respiratory protective device with an integral perimeter seal, valves and 26

specialized filtration, used to protect the wearer from toxic fumes or particulates. 27

Risk assessment: A formalized documented process for analysing risks involving a systematic 28

process of evaluating the potential risks that may be involved in a projected activity or 29

undertaking. 30


Page 11

TCID 50: median tissue culture infective dose 50%. The quantity of a virus suspension that 1

will infect 50% of tissue culture inoculated with the suspension, expressed as TCID50/ml. 2

Validation: The documented act of proving that any procedure, process, equipment, material, 3

activity, or system actually leads to the expected results. 4

4. Hazard identification 5

Hazards associated with pandemic vaccine manufacturing and laboratory testing are dependent 6

on the type of pandemic vaccine virus (reassortant or wild-type), method of production (egg-7

based or cell-based or other), whether it is an inactivated, live (attenuated) virus or a 8

recombinant virus-vectored vaccine, and whether or not there are any deliberate modifications 9

of the virus for attenuation or for enhanced immunogenicity and/or increased yield. For 10

recombinant virus-vectored vaccines, (e.g. MVA, adenovirus, or VSV) which use replicating 11

recombinant constructs based on viruses other than influenza, virus homogeneity, the nature of 12

the transgene, shedding and the potential for recombination are important factors to evaluate 13

(19, 20). 14

4.1 Candidate vaccine viruses (CVV) 15

CVVs for live and inactivated influenza vaccines are generally produced via reassortment with 16

well-defined backbones from established parent viruses, such as human virus A/Puerto 17

Rico/8/34 (PR8), A/Ann Arbor/6/60, or A/Leningrad/134/17/57 (A/Len/17); wt viruses may 18

also be used for vaccine production. Further, new donor strains for reassortment are being 19

developed and evaluated to enhance vaccine yields. It is likely that a high growth reassortant 20

will provide the basis for pandemic vaccine development, although it is conceivable that a wt 21

virus could be used. 22

4.1.1. Reassortants 23

The genome of influenza A viruses is composed of eight individual single-stranded RNA 24

segments of negative polarity. Segments 4 and 6 encode the two surface glycoproteins, HA and 25

neuraminidase (NA), respectively. The HA is the major surface antigen of the virus and 26

antibodies directed against HA can protect from infection. Antibodies to NA are not 27

neutralizing but can inhibit virus release from infected cells thereby reducing severity and 28

duration of disease and limiting viral shedding. The remaining six RNA segments ("internal 29

genes") encode internal and non-structural viral proteins. The segmented structure of the 30


Page 12

genome allows for the exchange of the individual RNA segments between influenza viruses, a 1

process defined as reassortment, upon coinfection of a single cell with two or more viruses. 2

The classical or conventional method for reassortment involves preparing CVVs by co-3

inoculation of a WHO recommended wt virus and a backbone donor virus with a high growth 4

phenotype in embryonated chicken eggs. Concomitant infection allows for reassortment of 5

genetic segments between the two viruses. Antiserum is used for negative selection against the 6

donor virus surface glycoproteins; amplification in eggs results in positive selection for growth. 7

The resulting high growth reassortant CVV must contain the HA and NA genes of the wt target 8

virus. This system takes time and is likely to take several weeks for the production of a high 9

growth reassortant. 10

An alternative approach is through the use of RG methodology to produce a reassortant vaccine 11

virus. (21). This process is based on incorporating the six RNA segments encoding the internal 12

genes of the (high growth) donor virus and the two segments encoding the HA and NA from 13

the circulating wt virus into plasmids. The plasmids are subsequently transfected into cells to 14

rescue the reassortant to be used for vaccine manufacturing. This system allows for the direct 15

genetic manipulation of the influenza gene segments and is generally faster and more accurate 16

than the use of classical reassortment. Moreover, if an HPAI virus is used in the RG process, 17

the HA gene can be easily modified to remove a specific motif of amino acids at the HA 18

cleavage site that is known to convey high pathogenicity in poultry (22). The reassortant can 19

thus be specifically designed to serve as a low pathogenic CVV. The RG system has been 20

reported to be able to produce a CVV within 9 to 12 days (23). 21

The distribution and receipt of the circulating wt virus as a source of RNA for construction of 22

RG HA and NA plasmids adds extra time to the reassortant process and can be bypassed if the 23

reassorting laboratories can produce plasmids by site‐directed mutagenesis of the template 24

DNA of an exisiting related virus, or by the use of synthetic DNA (24, 25). 25

4.1.1.1 Backbone donor viruses 26

PR8 is a common donor virus for generating reassortant vaccine viruses as it reaches a high 27

titre in embryonated chicken eggs. It was originally used in the late 1960s to produce “high 28

growth reassortants” and the use of such reassortants as vaccine viruses increases vaccine yield 29

many-fold. Moreover, PR8 has had over 100 passages in each of mice, ferrets and embryonated 30

chicken eggs. The result of such a passage history is complete attenuation of the virus rendering 31

it incapable of replication in humans (26). 32


Page 13

Improved donor viruses are under development in an effort to enhance vaccine yields from 1

CVVs used to manufacture inactivated influenza virus vaccines. Such donor viruses may be 2

new derivations of PR8 but also may have genes from viruses other than PR8, be synthetically 3

generated, and be optimized for specific HA and NA subtypes (23). Demonstration of adequate 4

attenuation for new/improved donor viruses will be needed. 5

Live attenuated seasonal influenza vaccines are licensed in some countries using reassortants 6

with a 6:2 gene constellation based on live attenuated donor viruses such as the A/Ann 7

Arbor/6/60 and A/Len/17 backbones. The attenuated A/Ann Arbor/6/60 virus has been used as 8

a backbone in 6:2 reassortant live attenuated vaccines in clinical studies for over 40 years using 9

approximately 30 different vaccine viruses, and the data demonstrate that the Ann Arbor/6/60 10

based reassortant vaccine viruses are attenuated for humans (27). To date, live attenuated 11

seasonal influenza vaccines derived from the A/Ann Arbor virus have been licensed in North 12

America and Europe while the A/Len/17 backbone has been licenced in Russia, China, 13

Thailand and India (28). These donors may also be used for the development of pandemic 14

influenza vaccine and an adequate level of attenuation has been shown for modified reassortant 15

viruses of various subtypes (29). For each candidate pandemic vaccine virus, this should be 16

verified by testing as described in section 5.1. H5N1-specific LAIV versions (LPAI) made 17

from A/Ann Arbor/6/60 reassortants have been licensed as pandemic preparedness vaccines in 18

several countries. 19

4.1.1.2 Gene segments from wild type viruses 20

Reassortants with a 6:2 gene constellation are considered to be ideal but are not always possible 21

to isolate by traditional co-cultivation methods and for seasonal vaccines some CVVs with 22

different gene constellations, such as 5:3, have been used. As the resultant gene constellation is 23

less predictable using classical methods there is a theoretical possibility of developing 24

reassortants with more than two wt parental genes or even of selection of a mutant (non-25

reassortant) wt virus with improved growth characteristics. The gene constellation of 26

reassortants derived by traditional co-cultivation methods should be determined. 27

4.1.1.3 Virulence factors associated with HA of wild type viruses 28

The CVV gene products derived from the wt virus will be, at a minimum, the HA and NA. For 29

reassortants derived from highly pathogenic H5 and H7 avian viruses by RG, the HA should be 30

modified so that the inserted amino acids at the HA cleavage site are reduced to a single basic 31

amino acid; for some H5 viruses additional nucleotide substitutions can be introduced in the 32


Page 14

vicinity of the cleavage site in order to increase the genetic stability of the created monobasic 1

motif during virus amplification for large-scale manufacture. Also, the modification of the 2

cleavage site alone is not a guarantee of low pathogenicity as there are LPAI that are virulent in 3

humans due to the presence of other virulence factors (30). It should also be noted that this 4

procedure will increase the safety of the reassortants for avian species as cleavage site 5

modifications have consistently resulted in a reduction of their pathogenicity in avian embryos 6

and in poultry (31). Reassortants for vaccine production are expected to be of low 7

pathogenicity in poultry compared to the highly pathogenic wt parental viruses (32). 8

9

The hazards associated with reassortants depend in part on HA receptor specificity. If a 10

reassortant has a preference for avian cell receptors (i.e. α2,3-linked terminal sialic acid) the 11

hazard is considered to be minimal to humans; however, if a reassortant has a preference for 12

mammalian cell receptors (α2,6-linkages, e.g. human H2N2 pandemic virus from 1957), or 13

possesses both avian and mammalian receptor specificities (e.g. H9N2), there is a greater risk 14

of human infection. For H5 reassortants, the HA retains a preference for α2,3-linked terminal 15

sialic acid residues, so the ability of the H5N1 reassortants to bind to and replicate in human 16

cells should be reduced. It is envisaged that an H5N1 reassortant derived by RG according to 17

WHO guidance (33) would be attenuated for humans compared to the wt H5 virus. 18

Nevertheless, it should also be noted that the human lower respiratory tract contains α2,3-19

linked sialic acid receptors and exposure to high doses of H5N1 viruses represents a risk of 20

infection. Moreover, humans are immunologically naive to H5 and many other avian subtypes, 21

which is also an important risk factor. 22

4.1.1.4 Other factors associated with influenza virus virulence 23

Although it is clear from experience in south-east Asia from 1997 to the present that H5N1 24

influenza viruses that display preference for α2,3-linked sialic acids can still replicate in 25

humans, it must be noted that influenza virus pathogenicity does not depend solely on HA, but 26

is a polygenic trait. The 1997 H5N1 virus had unusual PB2 and NS1 genes that influenced 27

pathogenicity whereas the 2004 H5N1 viruses possessed complex combinations of changes in 28

different gene segments that affected pathogenicity in ferrets (34, 35). In these cases even a 29

virus with a poor affinity for the mammalian receptors was able to replicate in humans 30

(although it was not transmissible). Further, prior to the 2003 outbreak in The Netherlands, 31

only two cases of transmission of H7 viruses from birds to humans were documented (36). Also 32

during the many years of laboratory handling of high-titre avian viruses (of which 33


Page 15

A/FPV/Dobson/27 is known to contain a gene which adapts it for replication in mammalian 1

cells (37, 38)), there has only been one report in the literature of a laboratory worker being 2

affected by these viruses. This was a laboratory worker in Australia who developed 3

conjunctivitis after accidentally being exposed to an H7N7 virus directly in the eye (39). The 4

PR8/H5N1 6:2 reassortants and the A/Ann Arbor/6/60 live attenuated 6:2 reassortants created 5

by RG for the production of H5N1 vaccine do not contain the gene constellation considered 6

necessary for pathogenicity in chickens, mice and ferrets and in contrast have internal genes 7

that are likely to confer sensitivity to the innate immune response. 8

Compared to HA, the NA protein of influenza viruses has a less prominent record as a 9

virulence factor. It is known that a balance of the counteracting activities of HA (receptor 10

binding) and NA (receptor destruction and virus release) is required for efficient viral 11

replication (40, 41). Further, specific adaptations in NA have been identified upon transmission 12

from wild aquatic birds to poultry. However, specific determinants for the adaptation to and 13

virulence in humans have so far not been detected in the NA protein, although there is some 14

evidence that the NA can mediate HA cleavage in rare H1N1 viruses (42, 43). Of note, 15

resistance to the viral inhibitors oseltamivir and zanamivir is caused by specific mutations in 16

either NA or HA, the latter reducing the affinity of HA to the receptor determinant. 17

4.2. Manufacture 18

Vaccine manufacture requires the establishment of both WHO Good manufacturing practices 19

(GMP) and appropriate biosafety requirements for biological products, as well as related 20

national regulations, technical standards, and guidelines (44, 45). GMP requires the protection 21

of the product from the operator and protection of the operator and the environment from the 22

infectious agent, thus minimizing the risk of any hazards associated with production. It should 23

be noted that reassortants derived from PR8 have been used routinely for the production of 24

inactivated influenza vaccines for the past 40 years. This work involves the production of many 25

thousands of litres of infectious egg allantoic fluids, which creates substantial aerosols of 26

reassortant virus within manufacturing plants. Most of the reassortants were made from wt 27

human influenza viruses. Although staff in the manufacturing facility may have some 28

susceptibility to infection with the wt virus, there have been no anecdotal or documented cases 29

of work-related human illness resulting from occupational exposure to the reassortants. 30

Similarly, reassortants derived from the A/Ann Arbor/6/60 virus have been used for the 31

production of LAIV for many years and no anecdotal or documented cases of work-related 32

human illness have been reported. While no conclusive study yet has been conducted to detect 33


Page 16

silent infections for either the PR8 or live-attenuated viruses, the attenuation status of these 1

CVVs continues to be supported by their excellent safety record to date. 2

Furthermore, for pandemic human CVVs that express avian influenza genes there may be 3

potential consequences for agricultural systems. If influenza A viruses of H5 or H7 subtype or 4

any influenza A virus with an IVPI greater than 1.2 are introduced into poultry (46), the 5

presence of infection would be notifiable to OIE and should lead to implementation of 6

biosecurity measures that aim at preventing the spread of disease (46, 47). Infection with 7

influenza A viruses of high pathogenicity in birds other than poultry, including wild birds, 8

should be reported to OIE; however, a Member Country should not impose bans on the trade in 9

poultry and poultry commodities in response to such a notification, or other information on the 10

presence of any influenza A virus in birds other than poultry, including wild birds (46, 47). 11

4.2.1 Production in eggs 12

Influenza vaccine has been produced in embryonated hens’ eggs since the early 1940s. Much 13

experience has been gained and some facilities are capable of handling large numbers of eggs 14

on a daily basis with the aid of mechanized egg handling, inoculation and harvesting machines. 15

Hazards may occur during the production stages and quality control laboratory activities prior 16

to virus inactivation. During egg inoculation the virus used is dilute and of a relatively small 17

volume. The most hazardous production stage is egg harvesting when the eggs have to be 18

opened to harvest the allantoic fluid, as the open nature of the operations may lead to a greater 19

exposure to aerosols and spills. The allantoic fluid that is harvested from the eggs is invariably 20

manipulated thereafter in closed vessels and hazards arising from live virus during downstream 21

processing and during the virus inactivation process, if used, are therefore less than during virus 22

harvest. Collection and disposal of egg waste is potentially a major environmental hazard. Safe 23

disposal of the waste from egg-grown vaccines, both within the plant and outside, is therefore 24

critical. 25

4.2.2 Production in cell cultures 26

For pandemic influenza vaccines produced on cell cultures, the biosafety risks associated with 27

manufacturing will depend primarily on the nature of the cell culture system employed. Closed 28

systems, such as bioreactors, normally present little to no opportunity for exposure to live virus 29

during normal operation, but additional safety measures must be taken during procedures where 30

samples are introduced into or removed from the bioreactor. Roller bottles and cell culture 31

flasks used for virus production may allow exposure to live virus through aerosols, spills, and 32


Page 17

other operations during virus production and, thereafter, additional risks are associated with the 1

inactivation and disposal of the large quantities of contaminated liquid and solid waste, 2

including cellular debris, generated by this method. 3

The possibility exists that genetic mutations may be selected in pandemic vaccine viruses 4

during passage in mammalian cells that render them more adapted to humans. Sequence 5

analysis of HA may be useful to detect such changes most likely occuring within or close to the 6

receptor-binding domain of the protein; however, it should be noted that little is known about 7

the relation between cell substrate and virus reversion or adaption. Beare et al. (48) tried to de-8

attenuate PR8 by multiple passage in organ cultures of human tissue, but failed, whereas 9

studies with MDCK cells demonstrated that human viruses that retained their α2,6 receptor 10

specificity (human-like) were likely to mutate to an α2,3 specificity (avian-like) as this 11

provided a replicative advantage on MDCK cells, rather than the reverse (49). Overall, hazards 12

arising from the inherent properties of a reassortant or wild type (wt) virus are likely to be far 13

greater than the probability of adaptation of the virus to a more human-like phenotype. 14

4.3 Hazards from the vaccine 15

Inactivated pandemic influenza vaccines present no biosafety risks provided that the results of 16

the inactivation steps show complete virus inactivation, as the viral vaccine is rendered 17

incapable of replication. 18

In an interpandemic or pandemic alert period, pilot-scale live attenuated pandemic influenza 19

vaccines may be developed for clinical evaluation. As there is some uncertainty concerning the 20

biosafety risks associated with shedding or other unintentional release into the environment 21

following vaccination, subjects participating in clinical trials in the interpandemic or pandemic 22

alert phase should be kept under appropriate clinical isolation conditions. If this is not done, 23

indirect hazards for humans could arise. 24

While it is unlikely that a reassortant will be harmful to humans, an indirect hazard may exist 25

through secondary reassortment with a human or animal influenza virus (50); however, recent 26

studies have shown that there is a clear time dependence of coinfections and the generation of 27

reassortants limiting this possibility (51). For secondary reassortants to be generated, several 28

events need to occur: 29

• Infection of production staff (or recipients of live attenuated vaccine in clinical trials) 30

with the reassortant virus 31


Page 18

• A worker to simultaneously contract infection with a wt influenza virus and the 1

reassortant virus 2

• A reassortment event would have to take place. 3

Evidence to date indicates that the probability of generating secondary reassortants is low. For 4

example, manufacturers have more than 30 years of experience with large-scale production of 5

vaccines based on PR8 reassortants and there have been no reported cases of human illness. 6

Moreover, containment procedures have significantly improved over the last 30 years and 7

production staff can be vaccinated to reduce the chances of an infection with a circulating wt 8

virus thus minimising the risk for secondary reassortment and appropriate personal protective 9

equipment (PPE) can also be provided. 10

5. Safety testing of candidate vaccine viruses (CVVs) 11

CVVs can be developed by either a WHO laboratory, or by a laboratory approved by a national 12

regulatory authority. The following tests and specifications have been developed based on 13

experience gained with the evaluation of CVVs derived from viruses of various subtypes. The 14

required safety testing of different CVVs and proposed containment levels are summarised in 15

Table 1. The information summarised in this table should be considered as guidance and 16

changes to these requirements may be determined on a case-by-case basis by WHO and/or 17

national authorities. It is recommended that for CVVs developed from newly emerging viruses 18

of pandemic potential, an appropriate WHO expert group review the associated data from 19

safety testing and advise WHO; WHO will then provide further guidance as to the appropriate 20

biocontainment requirements through its expert networks such as GISRS. 21

The requirement to conduct or complete some or all of these tests prior to the distribution of a 22

CVV may be relaxed based on additional risk assessments; these assessments should take into 23

account WHO pandemic phase, evolving virological, epidemiological and clinical data as well 24

as national regulatory requirements applying to shipment and receipt of infectious substances. 25

26

5.1 Tests to evaluate pathogenicity of CVVs 27

The recommended tests for CVVs are dependent on the parental viruses from which they are 28

derived (Table 1). The parental viruses determine the appropriate biocontaiment level for 29

conducting tests. The tests are described in the following sections. 30


Page 19

CVVs that are genetically similar (i.e. have HA and NA derived from the same or a genetically 1

very closely related wt virus and on the nominally same backbone) to a CVV that has already 2

undergone complete safety testing may not require full testing; it may be sufficient to confirm 3

sequence and genetic stability in this case. 4

5.1.1 Attenuation in ferrets 5

Ferrets were chosen because they have been used extensively as a good indicator of influenza 6

virus virulence for humans (52). Typically, seasonal influenza viruses cause mild to no clinical 7

signs in ferrets and virus replication is usually limited to the respiratory tract. PR8 virus has 8

been assessed in ferrets and found to cause few or no clinical signs, and virus replication is 9

limited to the upper respiratory tract. However, some wt HPAI viruses can cause severe and 10

sometimes lethal infections (53, 54). Thus, in the absence of human data, the ferret is the best 11

model to predict whether a virus will be pathogenic or attenuated in humans. 12

CVVs should be shown to be attenuated in ferrets in accordance with Table 1, except when 13

virus-specific risk assessments suggest a different approach (ie, performance of ferret test 14

where Table 1 does not require it or no ferret test where Table 1 requires it). These tests should 15

be conducted in well-characterized ferret models standardised by the use of common reference 16

viruses available from WHO CCs/ERLs for influenza. Detailed test procedures are described in 17

Appendix 1. Attenuated viruses would be those that meet the attenuation criteria as defined in 18

Appendix 1, with reference to the reference/standard viruses1. One or more laboratories may 19

have ferret pathogenicity data on parental wt viruses that could be used by all testing 20

laboratories as a further benchmark for comparison. Assessing the transmissibility of CVVs 21

between ferrets is not required. 22

5.1.2 Pathogenicity in chickens 23

For CVVs derived from HPAI H5 or H7 parental viruses determining the chicken intravenous 24

pathogenicity index (IVPI) is recommended and may be required by national authorities. The 25

procedure should follow that described in the OIE guidance. Any virus with an index greater 26

than 1.2 is considered an HPAI (55). 27

5.1.3 The ability to plaque in the presence or absence of added trypsin 28

1 Until the establishment of reference viruses for pathogenicity testing, CVVs are expected to be

compared to their respective wt parent virus; attenuation is then defined as relative to the pathogenicity of the wt virus.


Page 20

HPAI viruses can replicate in mammalian cell culture in the absence of added trypsin, whereas 1

LPAI viruses generally do not. This test is recommended for all CVVs derived from HPAI H5 2

or H7 parental viruses. It is recommended that these tests be established and characterized by 3

use of known positive and negative control viruses. 4

5.1.4 The ability to cause chicken embryo death 5

HPAI viruses cause rapid chicken embryo death upon inoculation into eggs and this test has 6

been used to complement in vivo IVPI assays (55). This test is recommended for all CVVs 7

derived from H5 or H7 parental viruses. 8

5.1.5 Genetic sequence analysis and stability 9

Genetic sequencing is important to confirm identity and/or to verify the presence of attenuating 10

and other phenotypic markers, (e.g., cold adaptation and temperature sensitivity in the case of 11

LAIV CVVs). Genetic sequencing is recommended to verify retention (stability) of the markers 12

of relevant phenotypic traits, where such markers are known, after 10 passages beyond 13

production level in relevant substrates i.e., embryonated chickens’ eggs or cultured cells. These 14

tests should be conducted on all CVVs, including wt CVVs. 15

16


Page 21

Table 1 1

Required safety testing of different CVVs and proposed containment levels for vaccine 2

production 3

4

aTest performed by WHO reference laboratory. 5

bWith deletion of additional residues at HA connecting peptide 6

c Highly pathogenic avian influenza virus 7

Candidate vaccine

virus

Tests needed on CVVsa Proposed containment for

vaccine production

Reassortantsb and

modified viruses derived

from H5 and H7 HPAIc

virusesd

Ferret, chickene, sequence,

plaquing, egg embryo,

genetic stability

BSL2 Enhanced (pandemic

influenza vaccine)

Reassortantsb and


from synthetic DNA

representing H5 and H7

HPAIc virusesf

Ferret, sequence, plaquing,

egg embryo, genetic

stability


influenza vaccine)

Reassortants and


from H5 and H7 LPAIg

viruses

Ferret, sequence, egg

embryo, genetic stability


influenza vaccine)

Reassortants and


from non-H5, H7 viruses

Ferret, sequence, genetic

stability


influenza vaccine)

Wild type H5, H7 HPAI

viruses

Sequence, genetic stabilityh BSL3 Enhanced (pandemic

influenza vaccine)

Wild type H5, H7 LPAI

viruses

Ferret, sequence, genetic

stabilitye, i

BSL2 Enhanced i (pandemic

influenza vaccine)

Wild type non-H5, H7

viruses

Ferret. sequence, genetic

stabilitye


influenza vaccine)


Page 22

dThis category refers to viruses derived by reverse genetics technology using as starting material wt 1

HPAI virus (or RNA extracted from wt HPAI virus) 2

ethe requirement for performance of the chicken pathogenicity test (IVPI) is dependent on national 3

regulatory requirements which are currently under review in some countries and may change 4

fthis category refers to viruses derived by reverse genetics technology using as starting material 5

synthetic DNA; the nucleotide sequence of the synthetic DNA is equivalent or similar to wt HA and NA 6

genes from HPAI viruses, with the exception that additional residues at the at HA connecting peptide 7

have been removed 8

gLow pathogenic avian influenza viruses. 9

hif changes are identified pathogenicity tests may be required 10

iif a virus specific hazard assessment identifies that additional control measures are appropriate, 11

containment level may be increased, for example Asian-lineage H7N9, and additional pathogenicity 12

testing required 13

14

6. Risk assessment and management 15

6.1 Nature of the work 16

Production of influenza vaccines in embryonated chicken eggs or cell culture require 17

propagation of live virus. Modifications of some CVVs will result in viruses that are expected 18

to be attenuated in humans (12, 21). Several production steps have the potential to generate 19

aerosols containing live virus. The virus concentration in aerosols will depend on the 20

production step and is highest during the harvest of infected allantoic fluid. Small amounts of 21

virus containing liquid or very dilute virus suspensions are used during seed virus preparation 22

and egg inoculation. Appropriate biosecurity and biosafety measures, such as use of laminar air 23

flows, positive pressure laminar flow hoods, cleaning and decontamination of equipment, waste 24

management and spill kits, will be in place to prevent accidental exposure in the work 25

environment and the release of virus into the environment. 26

6.2 Health protection 27

6.2.1 Likelihood of harm to human health 28

Wild-type influenza viruses are able to infect humans and cause serious illness. Certain 29

phenotypic features that are associated with virulence can be modified and the resulting CVV 30

will have lower probability of causing harm to human health. 31


Page 23

6.2.1.1 Reassortant CVVs 1

Reassortant CVVs containing a backbone of genes except segments 4 and 6 (HA and NA) from 2

PR8, A/Ann Arbor/6/60 or А/Len/17 have been widely used for production of seasonal 3

influenza vaccines and vaccines against the pandemic H1N1pdm09 influenza viruses. A 4

significant body of data is available, suggesting that a reassortant virus composed of RNA 5

segments coding for HA and NA derived from a pandemic virus and the genes coding for the 6

internal and non-structural proteins ("internal genes") from PR8, A/Ann Arbor/6/60 or 7

A/Len/17 will have only a low probability of causing harm to human health (12, 21). 8

Viruses that were rescued from plasmids will have the desired 6 "internal" genes of the helper 9

virus, however this is not always the case for reassortants prepared by mixed infection and 10

antibody selection ("classical reassortment"). Risks associated with RNA segments, other than 11

segments 4 and 6, that were derived from the pandemic virus must be carefully evaluated. 12

Efforts have been made to modify backbones that were derived from PR8 or other viruses to 13

achieve attenuation or to increase the yield during vaccine manufacturing. Until more field 14

experience has become available, the potential of vaccine viruses that contain modified 15

backbones to cause harm to human health needs careful assessment. 16

HA cleavage site: Most HPAI viruses contain a sequence of basic amino acids at the cleavage 17

site separating the HA1 and HA2 subunit. HPAI viruses in which the basic amino acids have 18

been removed from the HA by genetic engineering are likely to be attenuated. Modification of 19

the HA cleavage site is associated with reduced virulence in animal models and is expected to 20

result in attenuation in humans. Modification of the HA cleavage site only applies to HPAI H7 21

and H5 viruses. It should be noted that LPAI, e.g. H7N9 wt viruses also have caused serious 22

human illness (56). 23

Receptor specificity: Preferential binding of the HA to α2,6 receptors is associated with 24

transmissibility of pandemic influenza and circulating human viruses in humans (57,58). 25

However, it must be noted that viruses with a preference to bind to α2,3 receptors, such as 26

H7N9 viruses, have also been causing serious human illness (59). While receptor specificity 27

should be considered during risk assessment as a factor in reducing the risk for a virus to cause 28

harm to human health, it is not in itself sufficient for virus attenuation. 29

Secondary reassortment: It is conceivable that reassortment between a CVV, containing HA 30

and NA from a pre-/pandemic virus, and a seasonal human wt influenza virus could occur 31


Page 24

during simultaneous infection of humans with both viruses. Such a reassortant may be 1

replication-competent in humans, while having pre-/pandemic surface proteins. The likelihood 2

that these events occur and lead to secondary reassortment is considered to be low. Laboratory 3

and production facilities have biosafety control measures in place to prevent exposure of staff 4

to live virus. In case of accidental exposure, it is is unlikely that a CVV would replicate 5

efficiently or transmit to human contacts. Virus shedding would be expected to be well below 6

the titres considered to be needed for human infection; however, although there is no known 7

precedent of secondary reassortment of using PR8 reassortants for vaccine manufacturing in 8

more than 40 years, the public health consequences of such an event could be serious. 9

6.2.1.2 Wild-type CVVs of low pathogenicity (LP) 10

Pathogenicity of wt viruses with low in vivo pathogenicity, without multiple basic amino acids 11

at the HA cleavage site is likely to be low in humans (60). However, it should be noted that 12

some H7N9 viruses that are LP in poultry have caused severe illness in humans. 13

Transmissibility of wt viruses with avian receptor specifity is likely to be low in humans, 14

whereas transmissibility of wt viruses with mammalian receptor specificity (e.g. human H2N2 15

and H9N2) is unknown. 16

In case of accidental exposure there is a risk of secondary reassortment with human viruses and 17

the resulting reassortant viruses may be able to replicate in humans. Appropriate biosafety 18

control measures will be in place during manipulation of these viruses to prevent exposure of 19

staff. 20

6.2.1.3 Wild-type CVVs of high pathogenicity (HP) 21

The use of highly pathogenic wt CVVs is restricted to cell-culture-based production, which 22

allows the use of closed systems. For diagnostic tests and vaccines for terrestrial animals, HPAI 23

viruses should not be used. Instead, CVVs produced by reverse genetics and containing the 24

haemagglutinin gene of an HPAI virus that had the cleavage site sequence altered to that of a 25

LPAI H5/H7 virus are preferred (47). Appropriate biosecurity and biosafety measures during 26

production, analytical testing and waste disposal are required to protect staff and prevent 27

release of infectious virus into the environment. 28


Page 25

6.2.1.4 Susceptibility of CVVs to neuraminidase inhibitors 1

Influenza viruses/CVVs that are sensitive to NA inhibitors (or other licensed drugs, once 2

available) should be used for production if at all possible. Sequence verification may be enough 3

to confirm drug susceptibility. 4

6.3 Environmental protection 5

6.3.1 Environmental considerations 6

Influenza A viruses are endemic throughout the world in some farm animals (pigs and horses) 7

and some populations of wild birds, specifically birds of the families Anseriformes (ducks, 8

geese and swans) and Charadriiformes (shorebirds) (61). 9

A number of influenza A viruses, such as H5, H7 and H9 can cause disease in domestic 10

poultry. H5 and H7 can be highly pathogenic in poultry, whereas H9 is less so. In addition, 11

sporadic infections by influenza A viruses have been reported in farmed mink, wild whales and 12

seals, dogs and captive populations of big cats (tigers and leopards) (62). In big cats, the 13

infections followed consumption of dead chickens infected with H5N1 viruses. Influenza A 14

virus infection in dogs were caused by H3N8 or H3N2 viruses closely related to endemic 15

equine and avian viruses, respectively. The H3N8 and H3N2 viruses continue to circulate in 16

dogs in North America (63). Recently, LPAI avian A(H7N2) virus was identified as the cause 17

of an outbreak of respiratory disease in domestic shelter cats in northeastern United States, 18

indicating yet another mammalian host for avian influenza viruses (64) 19

It is expected that many prepandemic viruses will have avian receptor specificity and thus birds 20

would theoretically be the species most susceptible. Several studies indicate that viruses with 21

PR8 backbones are attenuated in chickens (65). A reassortant containing HA (with a single 22

basic amino acid at the cleavage site) and NA from the 1997 Hong Kong H5N1 virus and the 23

genes coding for the internal and non-structural proteins of PR8 was barely able to replicate in 24

chickens and was not lethal (65). Similar studies have been performed with the 2003 25

Hong Kong H5N1 virus at the WHO Collaborating Centre, Memphis, USA (66), where the 6:2 26

PR8 reassortant did not replicate or cause signs of disease in chickens. The removal of the 27

multiple basic amino acids from the H5 x PR8 reassortants in both studies probably played a 28

major role in reducing the risk for chickens. 29

Hatta et al. (34) have demonstrated that acquisition of only one PR8 gene by an avian influenza 30

virus can abolish virus replication in ducks. 31


Page 26

It is likely that the temperature sensitive phenotype of the cold-adapted vaccine virus would 1

also be attenuated in avian species due to the elevated body temperature of birds. 2

Pigs have both α2,3 and α-2,6 –linked sialic acid receptors in abundance (67) and must be 3

considered susceptible to most influenza viruses, including LAIV and CVVs with a backbone 4

of PR8 genes. 5

6.4 Assignment of containment level 6

The production of influenza vaccine reassortant CVVs from highly pathogenic wt viruses 7

should take place in BSL3 enhanced or BSL4 facilities, as advised by WHO (68) or national 8

authorities. WHO Collaborating Centres, other specialized laboratories and some vaccine 9

manufacturers provide characterized reassortant CVVs to all interested vaccine manufacturers 10

who may develop vaccine seeds and vaccines from these materials. 11

In consideration of the hazards associated with egg and cell culture vaccine production and 12

quality control with "classical" or "RG" derived reassortant or wt viruses of pandemic potential 13

which have been demonstrated to be attenuated in ferrets and/or have demonstrated low 14

pathogenicity in chickens where applicable, as specified in section 5.1, the assigned 15

containment level is BSL2 enhanced (pandemic influenza vaccine), as defined below. This 16

applies to both pilot-scale and large-scale production during the interpandemic phase and 17

pandemic alert period (68). Any subsequent relaxation of the levels of containment to the 18

standard used for seasonal production during the developing pandemic, should be authorized on 19

a case-by-case basis by the national competent authorities after careful evaluation of the risks. 20

Special consideration should be given to hazards associated with cell culture vaccine 21

production and quality control of HPAI or "classical" or "RG" reassortant pandemic viruses of 22

unknown pathogenicity. The assigned containment level is BSL3 Large Scale manufacturing as 23

defined below. This applies to both pilot-scale and large-scale production during the 24

interpandemic phase and pandemic alert period (68). The parts of the facility where such work 25

is done (both production and quality control) should additionally meet the OIE requirements for 26

containment, which include not only biosafety and biosecurity but also requirements to limit 27

the introduction and spread within animal populations (55). Any subsequent relaxation of the 28

levels of containment during the developing pandemic, should be authorized on a case-by-case 29

basis by the national competent authorities after careful evaluation of the risks. 30


Page 27

In view of the open nature of large-scale egg-based vaccine production, it is not feasible to 1

operate at BSL3 enhanced (pandemic influenza vaccine). Therefore egg-based vaccine 2

production from HPAI H5 or H7 wt viruses is not recommended. 3

Containment conditions must be defined, based on an activity based risk assessment, taking 4

into account the scale of manipulations, the titres of live virus, and whether an activity involves 5

virus amplification. Biosafety control measures must be reconciled with rules and regulations 6

governing the manufacture and testing of medicinal products known as good manufacturing 7

paractices (GMP) (44). It should be noted that biosafety control measures apply to 8

manipulations involving live virus; they no longer apply once virus has been inactivated by a 9

validated process. 10

6.5 Environmental control measures 11

Appropriate containment measures to prevent release of live virus into the environment must be 12

in place. 13

Local biosafety/biosecurity regulations provide guidance on the disposal of potentially 14

infectious waste. Notably, contaminated waste from current production facilities may reach 15

high virus titres. All decontamination methods should be validated. If possible, 16

decontamination of waste should take place on site. Where this is not possible, it is the 17

responsibility of each manufacturer to ensure that procedures are in place to safely contain 18

material during transport prior to decontamination off site. Guidance on regulations for the 19

transport of infectious substances is available from WHO (18) and national competent 20

authorities. In all cases the procedures must be validated to ensure that they function at the 21

scale of manufacturing. 22

Medical surveillance program of staff should be established prior to vaccination. 23

Procedures should be in place to provide antiviral medicines, in case the situation warrants it, 24

e.g. in case of accidental exposure. Where antiviral medicines are only available as prescription 25

medicines, care should be taken to have stocks available in pharmacies. 26

Suitable PPE must be available to prevent exposure of staff to live virus. In order to further 27

reduce the risk of secondary reassortment in case of accidental exposure, the use of seasonal 28

influenza vaccines should be recommended to staff. For the same reason, the use of a pre-29

/pandemic vaccine can be considered if available and if marketing authorization has been 30

received. 31


Page 28

Each manufacturer should also assess the risk of contamination of birds, horses, pigs or other 1

susceptible animals based on the likelihood of their being in the vicinity of the manufacturing 2

plant. Following occupational exposure, staff or other personnel entering the area potentially 3

exposed to live virus should avoid visiting pig, horse or bird facilities (e.g. farms, equestrian 4

events, bird sanctuaries) for at least 14 days following occupational exposure. If conjunctivitis 5

or respiratory signs indicating the potential development of influenza infection or disease 6

develop during this 14 day period this period, quarantine time should be extended to 14 days 7

after the symptoms have resolved. 8

Stringent measures to control rodents, other mammals and birds should be in place. 9

6.5.1 Specifications for "BSL2 enhanced (pandemic influenza vaccine) facilities" 10

Specifications for BSL2 enhanced (pandemic influenza vaccine) facilities include the following 11

in addition to the principles for BLS2 facilities as specified in the WHO Laboratory biosafety 12

manual (1). 13

6.5.1.1 Facility 14

The facility should be designed and operated according to the stage of the manufacturing 15

process to meet the demands of protection of the recipient of the vaccine, the staff producing 16

and testing the vaccine, the environment and the population at large. It is noted that different 17

solutions may be needed depending on the risks inherent in the operation(s) conducted in an 18

area. Specialized engineering solutions will be required that may include: 19

— use of relative negative pressure biosafety cabinets 20

— use of high-efficiency particulate air (HEPA) filtration of air prior to exhaust into public 21

areas or the environment 22

— use of positive pressure barrier and/or negative pressure in-line sinks prior to exhausting 23

to areas where no live virus is handled 24

In addition, the following decontamination procedures should take place: 25

— decontamination of all waste from BSL2 enhanced (pandemic influenza vaccine) areas 26

— decontamination of all potentially contaminated areas at the end of a production 27

campaign through cleaning and validated decontamination, for example gaseous 28

fumigation. 29


Page 29

6.5.1.2 Personal protection 1

• Full-body protective laboratory clothing (for example, Tyvek® disposable overalls) 2

• If activities cannot be contained by primary containment and open activities are being 3

conducted, the use of respiratory protective equipment, such as N95, FFP3 (69) or 4

equivalent respirators is strongly recommended. Minimal specifications for the 5

filtering/absorbing capacity of such equipment should be met, and masks, if used, must be 6

fitted properly and the correctness of fit tested. 7

• All personnel, including support staff and others who may enter the production or QC areas 8

where pandemic viruses may be handled, should be instructed, in a written document to 9

which they sign their agreement, not to have any contact with susceptible animals such as 10

ferrets or farm animals, in particular birds, horses or pigs, for 14 days after departure from 11

the facility where vaccine has been produced. This period should be extended to 14 days 12

after the symptoms resolve if conjunctivitis or respiratory signs indicating the potential 13

development of influenza infection or disease develop during this 14 day period. Currently 14

the risks involved in contact with household dogs and cats are not considered to be 15

significant, but the available scientific evidence is sparse. 16

• Wherever possible, it is strongly recommended that staff should be prophylactically 17

vaccinated with seasonal inactivated influenza vaccines. 18

• It is anticipated that pilot lots of pandemic virus vaccine will have already been produced 19

before large scale vaccine production is attempted. Experimental vaccines inducing 20

protective antibody levels are recommended for use by staff before large scale vaccine 21

production commences, provided they are available and marketing authorization has been 22

received. 23

• Procedures should be in place to provide antiviral treatment in case the situation warrants it 24

(e.g. accidental exposure). 25

6.5.1.3 Monitoring of decontamination 26

Cleaning and decontamination methods need to be validated and reviewed periodically as part 27

of a master validation plan to demonstrate that the protocols, reagents and equipment used are 28

effective in the inactivation of pandemic influenza virus on facility and equipment surfaces, 29

garments of personnel and waste materials, and within cell growth and storage containers. Once 30

decontamination procedures for influenza virus have been fully described and validated, there 31


Page 30

is no need to repeat them for each new virus. Validation studies using influenza viruses may be 1

supplemented by studies with biological (for example bacterial) markers selected to be more 2

difficult to inactivate than influenza. 3

6.5.2 Specifications for "BSL3 Large Scale Manufacturing" 4

It is assumed that ferret pathogenicity testing will be conducted on all strains of unknown 5

pathogenicity even given the assumptions above (Section 5.1) related to the low probability of 6

a PR8 reassorted virus being pathogenic in humans. This assumption is based on the experience 7

currently held relating to reassortants of HA subtypes other than H1 & H3; i.e., H5, H7, and 8

H9. 9

Under these circumstances, a facility will need to meet the requirements for protection of 10

personnel handling potentially dangerous micro-organisms found in the WHO Laboratory 11

Biosafety Manual, Third edition (1), which lays out a Risk Survey to be undertaken prior to 12

rating a laboratory space as either BSL- 1, 2 or 3. Similar requirements are found in the 13

European Directive 2000/54/EC (2007) (70) on the protection of workers from risks related to 14

exposure to biological agents at work and the US CDC’s Biosafety in Microbiological and 15

Biomedical Laboratories Guide (Edition 5) (71). 16

Following review of the requirements for Biosafety at Level BSL3 from the sources noted 17

above, it is proposed that a facility meeting the criteria detailed below, and with the noted 18

operator protection in place, would be suitable to manufacture vaccine, at large scale using a 19

virus seed prepared by RG or classical reassortant methodology, whilst pathogenicity remains 20

unknown. 21

6.5.2.1 Facility 22

The facility should be designed and operated to meet the demands of protection of the recipient 23

of the vaccine, the staff producing and testing the vaccine and of the environment. This will 24

require specialized engineering solutions that may include: 25

• Appropriate signage and labeling related to the activities being undertaken when a virus of 26

unknown pathogenicity is in use. 27

• The facility must be designed and constructed as a contained GMP space. The surfaces and 28

finishes must comply with GMP requirements which will ensure they are appropriately 29

sealed and easily cleaned and decontaminated. 30


Page 31

• The air cascades within the facility should be such that any live virus is contained within 1

the work zones in which it is used. All work with infectious virus must be conducted within 2

these contained zones. 3

• Access to the contained areas must be gained via double door entry airlocks. The airlocks 4

should operate at a pressure either lower or higher than that on either side. In this way the 5

airlocks become either a sink or a pressure barrier, maintaining an airflow containment of 6

the facility. Note that in the case where the airlocks provide a low pressure ‘sink’ it will be 7

required that the entry and exit doors be interlocked or alarmed with a suitable delay or 8

alarm system to prevent both being opened at the same time. The air pressure cascade 9

within the negatively pressure ‘contained’ zone should allow for compliance with GMP (i.e. 10

higher pressures in cleanest zones) requirements for clean rooms. 11

• All supply and exhaust air must be passed through HEPA filtration, again maintaining the 12

required containment and GMP conditions. The air conditioning systems within the facility 13

must be the subject of rigorous risk assessment related to potential failure modes and ‘fail 14

safe’ systems implemented where deemed necessary. The facility should be under constant 15

monitoring related to environmental factors including the maintenance of appropriate room 16

pressure differentials. Should there be any undesirable fluctuation the situation should be 17

alarmed and appropriate action taken. 18

• All equipment to be reused is either cleaned in place, decontaminated by means of 19

autoclaving or otherwise cleaned and decontaminated by validated, dedicated systems prior 20

to washing for reuse. 21

• Areas of potential liquid spill should be assessed and bunded (dammed) to ensure 22

containment of any spill. This should include waste treatment plants and processes. 23

Procedures must be in place such that spills are contained, cleaned and the contaminated 24

materials properly disposed so as not to compromise the integrity of the facility. 25

• Materials entry to the ‘contained’ zone should be via separately HEPA filtered, interlocked, 26

double ended Pass Through Cabinets (PTC) or double-ended autoclaves. 27

• All facility waste, including egg waste, should be discarded via validated on-site waste 28

effluent systems or by autoclaving. The validation of waste systems should specifically 29

include viral inactivation testing. Any items which follow the process from the external 30

environment, through the manufacturing process and are returned to the external 31

environment must be the subject of particular attention (e.g. the plastic egg maché). 32


Page 32

Dedicated washing and decontamination of equipment and/or procedures must be provided 1

which, again, must be fully validated for viral inactivation. 2

6.5.2.2 Personal protection 3

• All clothing worn outside the facility should be replaced by manufacturing facility garments 4

upon entry into the facility. 5

• It should be noted that gowning requirements in viral zones should always include full suit, 6

overshoes, eye protection and double gloves. 7

• The provision of full hood powered air-purifying respirators (PAPR) for all personnel 8

working within the containment areas of the manufacturing facility. These powered hoods 9

are to be worn at all times when the facility is in operation under these enhanced biosafety 10

requirements. Figure 1 shows the type of operator protection to be provided. 11

12

Figure 1: Operator Protection 13

• All facility clothing is to be removed on exit, with soiled clothing transferred out of the 14

facility via a decontamination autoclave or similar method. The respirator hoods may be 15

decontaminated with 70% alcohol or similar and stored in the ‘grey’ zone of the entry/exit 16

airlock. 17

• Specific operational procedures should be developed and implemented for operation under 18

enhanced biosecurity conditions. 19

• Wherever possible it is strongly recommended that staff should be prophylactically 20

vaccinated with a seasonal influenza vaccine. In the case of pandemic strains, it is 21


Page 33

anticipated that before large scale vaccine production is attempted, pilot lots of vaccine will 1

have already been produced. If available and if marketing authorization has been received, 2

experimental vaccines inducing protective antibody levels are recommended for use by staff 3

before large scale vaccine production commences 4

• Procedures should be in place to provide antiviral treatment in case the situation warrants it. 5

(e.g. accidental exposure). 6

• Daily temperature monitoring before commencing work (with an exclusion policy for ill 7

workers) where this is possible 8

• On-site Occupational Health and Safety and medical support should be maximized with 9

such additional measures as: 10

o Medical consultation / training in recognition of ‘flu-like’ symptoms 11

o Out of hours referral plan to medical facilities with quarantine facilities 12

• Taking a full body shower upon exit from the BSL3 Large Scale Manufacturing 13

containment facility is recommended. It is mandatory following situations when staff may 14

have been exposed to vaccine virus. 15

• All personnel, including support staff and others who may enter the production or QC areas 16

where pandemic viruses may be handled, should be instructed, in a written document to 17

which they sign their agreement, not to have contact with animals, in particular farm 18

animals 14 days following departure from the facility where vaccine has been produced. 19

This period should be extended to 14 days after the symptoms resolve if conjunctivitis or 20

respiratory signs indicating the potential development of influenza infection or disease 21

develop during this 14 day period. Currently the risks involved in contact with household 22

dogs and cats are not considered to be significant, but the available scientific evidence is 23

sparse. 24

6.6 Biosafety management and implementation within a vaccine production facility 25

6.6.1 Management structure 26

The implementation of the biosafety levels described in these guidelines requires that the 27

institution employ a biosafety officer who is knowledgeable in large-scale virus production and 28

containment, but is independent of production in his or her reporting structure. The biosafety 29

officer is responsible for the independent oversight of the implementation of the biosafety 30

practices, policies and emergency procedures in place within the company or organization and 31


Page 34

should report directly to the highest management levels within the company. A biosafety 1

officer is needed in addition to a qualified person who, in some countries, has overall 2

responsibility for a medicinal product. 3

There should also be a Biosafety Committee comprising representatives of virus production and 4

quality control that is responsible for reviewing the biosafety status within the company and for 5

coordinating preventive and corrective measures. The institutional biosafety officer must be a 6

member of the Committee. The chairperson should be independent of both the production and 7

quality control functions. The Committee should report to the executive management of the 8

manufacturing company to ensure that adequate priority and resources are made available to 9

the Committee to implement the required measures. 10

6.6.2 Medical surveillance 11

Occupational health departments at vaccine manufacturers of pandemic virus influenza 12

vaccines should provide training in recognizing the clinical signs of influenza infection to 13

company physicians, nurses and vaccine manufacturing staff including supervisors, who must 14

make decisions on the health of personnel associated with the manufacture and testing of 15

pandemic virus influenza vaccine. Local medical practitioners caring for personnel from the 16

manufacturing site should receive special training in the diagnosis and management of 17

pandemic influenza infection. Any manufacturer embarking on large-scale production should 18

have documented procedures for dealing with influenza-like illness in the staff involved, or 19

their family members, including diagnostic procedures and prescribed treatment protocols. 20

Manufacturers should ensure staff understand that they have an obligation to seek medical 21

attention and to report any influenza-like illness to the occupational health department or 22

equivalent. Manufacturers should ensure that procedures are in place to provide antiviral 23

treatment if the situation warrants it (e.g. accidental exposure) and have defined means or 24

arrangements of advising staff with any influenza–like illness as necessary. 25

6.6.3 Implementation 26

A detailed and comprehensive risk analysis should be conducted to define possible sources of 27

contamination of personnel or the environment that may arise from the production or testing of 28

live influenza virus within the establishment. For each procedure or system, this analysis 29

should take into account the concentration, volume and stability of the virus at the site, the 30

potential for inhalation or injection that could result from accidents, and the potential 31

consequences of a major or minor system failure. The procedural and technical measures to be 32


Page 35

taken to reduce the risk to workers and the environment should be considered as part of this 1

analysis. The results of this risk analysis should be documented. 2

A comprehensive Biosafety Manual, or equivalent must be created and implemented to fully 3

describe the biosafety aspects of the production process and of the quality control activities. It 4

should define such items as emergency procedures, waste disposal, and the requirements for 5

safety practices and procedures as identified in the risk analysis. The manual should be made 6

available to all staff of the production and quality control units, with at least one copy present 7

in the containment area(s). The manual should be reviewed and updated when changes occur 8

and at least biannually. 9

Comprehensive guidelines outlining the response to biosafety emergencies, spills and accidents 10

should be prepared and made available to key personnel for information and for coordination 11

with emergency response units. Rehearsals of emergency response procedures are helpful. 12

These guidelines should be reviewed and updated annually. 13

The implementation of the appropriate biosafety level status in the production and testing 14

facilities should be verified through an independent assessment. National requirements 15

concerning verification mechanisms should be in place and complied with. 16

17


Page 36

Authors and acknowledgements 1

The first draft was prepared by Gary Grohmann (consultant) and TieQun Zhou (WHO), based 2

on input from the participants of the Working Group Meeting on Revision of WHO TRS 941, 3

Annex 5: WHO Biosafety Risk Assessment and Guidelines for the Production and Quality 4

Control of Human Influenza Pandemic Vaccines, held on 9-10 May 2017: 5

Othmar G Engelhardt, National Institute for Biological Standards and Control, Medicines and 6

Healthcare products Regulatory Agency, Potters Bar, United Kingdom; Glen Gifford; World 7

Organisation for Animal Health, Paris, France; Shigeyuki Itamura, Centre for Influenza Virus 8

Research, National Institute of Infectious Diseases, Tokyo, Japan; Jacqueline M. Katz, WHO 9

Collaborating Centre for the Surveillance, Epidemiology and Control of Influenza, Centers for 10

Disease Control and Prevention, Atlanta, United States of America; Changgui Li, National 11

Institutes for Food and Drug Control, Beijing, People's Republic of China; John McCauley, 12

WHO Collaborating Centre for Reference and Research on Influenza, Crick Worldwide 13

Influenza Centre, The Francis Crick Institute, London, United Kingdom; Jennifer Mihowich, 14

Centre for Biosecurity, Health Security Infrastructure Branch, Public Health Agency of 15

Canada, Ottawa, Canada; Ralf Wagner, Paul-Ehrlich-Institut, Langen, Germany; Richard 16

Webby, WHO Collaborating Centre for Studies on the Ecology of Influenza in Animals, St. 17

Jude Children's Research Hospital, University of Tennessee, Memphis, United States of 18

America; Jerry Weir, Centre for Biologics Evaluation and Research, Food and Drug 19

Administration, Maryland, United States of America; Gert Zimmer, Institute of Virology and 20

Immunology, Mittelhäusern, Switzerland. Representatives of international federation of 21

pharmaceutical manufacturers & associations (IFPMA): Matthew Downham, 22

AstraZeneca/Medimmune; Lionel Gerentes, Sanofi Pasteur, France; Elisabeth Neumeier, GSK 23

Vaccines, Germany; Beverly Taylor, Seqirus, UK. Representatives of developing countries 24

vaccine manufacturers network (DCVMN): Pradip Patel, Cadila Healthcare Limited, 25

Ahmedabad, India; Kittisak Poopipatpol and Ponthip Wirachwong, Government 26

Pharmaceutical Organization, Bangkok, Thailand. WHO headquarters: Mustapha Chafai, 27

Prequalification Team, Essential Medicines and Health Products (EMP) Department, Health 28

Systems and Innovation (HIS) Cluster, World Health Organization (WHO), Geneva, 29

Switzerland; Gary Grohmann, HIS/WHO, Switzerland; Ivana Knezevic, Technologies 30

Standards and Norms (TSN) Team, EMP/HIS/WHO, Geneva, Switzerland; Wenqing Zhang, 31

High Threat Pathogens (PAT), Infectious Hazard Management (IHM), Health Emergencies 32


Page 37

Programme (WHE), WHO, Geneva, Switzerland; TieQun Zhou, TSN/EMP/HIS/WHO, 1

Geneva, Switzerland. 2

3

References 4

1. Laboratory biosafety manual, 3rd ed. Geneva, World Health Organization, 2004 5

(WHO/CDS/CSR/LYO/2004.11). 6

http://www.who.int/csr/resources/publications/biosafety/WHO_CDS_CSR_LYO_2004_11/en/, 7

accessed 22 Nov 2017. 8

2. WHO biosafety risk assessment and guidelines for the production and quality control of 9

human influenza pandemic vaccines, World Health Organization, WHO Technical Report 10

Series No 941, 2007, Annex 5. 11

http://www.who.int/biologicals/publications/trs/areas/vaccines/influenza/Annex%205%20huma12

n%20pandemic%20influenza.pdf, accessed 22 Nov 2017. 13

3. WHO H1N1 updates (2009): 14

http://www.who.int/entity/biologicals/publications/trs/areas/vaccines/influenza/H1N1_vaccine_15

production_biosafety_SHOC.27May2009.pdf?ua=1, accessed 22 Nov 2017. 16

4. Update of WHO biosafety risk assessment and guidelines for the production and quality 17

control of human influenza vaccines against avian influenza A(H7N9) virus (2013): 18

http://www.who.int/biologicals/BIOSAFETY_RISK_ASSESSMENT_21_MAY_2013.pdf?%219

520ua, accessed 22 Nov 2017. 20

5. Recommendations to assure the quality, safety and efficacy of influenza vaccines 21

(human, live attenuated) for intranasal administration. WHO TRS 977, Annex 4, 2009. 22

http://www.who.int/biologicals/areas/vaccines/influenza/TRS_977_Annex_4.pdf?ua=1, 23


6. Global Influenza Surveillance and Response System (GISRS): 25

http://www.who.int/influenza/gisrs_laboratory/en/. 26

7. WHO Global Influenza Programme: http://www.who.int/influenza/en/. 27

8. Influenza Vaccine Response during the Start of a Pandemic. Report of a WHO 28

Informal Consultation held in Geneva, Switzerland, 29 June - 1 July 2015. 29

http://apps.who.int/iris/bitstream/10665/207751/1/WHO_OHE_PED_GIP_2016.1_eng.pdf?ua=30

1, accessed 22 Nov 2017. 31

9. Report: 2nd WHO Informal Consultation on Influenza Vaccine Response during the 32

Start of a Pandemic, 21-22 July 2016. 33

http://www.who.int/influenza/resources/publications/influenzavaccineresponse_meeting0 2/en/. 34


Page 38

10. WHO Global Action Plan for Influenza Vaccines (GAP) 1

http://www.who.int/influenza_vaccines_plan/en/. 2

11. Report. WHO Working Group Meeting on Revision of WHO TRS941, Annex 5: WHO 3

Biosafety Risk Assessment and Guidelines for the Production and Quality Control of Human 4

Influenza Pandemic Vaccines, Domaine de Penthes, Geneva, Switzerland, 9 – 10 May 2017. 5

http://who.int/biologicals/areas/vaccines/INFLUENZA_WG_mtg_TRS941_Annex_5_12_Oct_6

2017_for_web.pdf?ua=1, accessed November 12, 2017. 7

12. Production of pilot lots of inactivated influenza vaccines from reassortants derived from 8

avian influenza viruses. Interim biosafety risk assessment. WHO/CDS/CSR/RMD/2003.5, 9

World Health Organization, 2003. 10

http://www.who.int/csr/resources/publications/influenza/influenzaRMD2003_5.pdf, accessed 11

22 Nov 2017. 12

13. Recommendations for production and quality control of inactivated influenza vaccines. 13

In: WHO Expert Committee on Biological Standardization. Fifty- fourth report. Geneva, World 14

Health Organization, 2005, Annex 3 (WHO Technical Report Series, No. 927). 15

http://www.who.int/biologicals/publications/trs/areas/vaccines/influenza/ANNEX%203%20Infl16

uenzaP99-134.pdf?ua=1, accessed 22 Nov 2017. 17

14. Neumann G, Kawaoka Y. Transmission of influenza A viruses. Virology. 2015 18

May;479-480:234-46. doi: 10.1016/j.virol.2015.03.009. Epub 2015 Mar 24. 19

15. Böttcher-Friebertshäuser et al. (2014). The hemagglutinin: a determinant of 20

pathogenicity. Current Topics Microbiol Immunol 385:3-34. 21

16. Robertson JS and Engelhardt OG, 2010, Developing vaccines to combat pandemic 22

influenza Feb;2(2):532-46. doi: 10.3390/v2020532. Epub 2010 Feb 2. 23

17. Recommendations for the production and control of influenza vaccine (inactivated), 24

WHO TRS 927, Annex 3. 2005. 25

http://www.who.int/entity/biologicals/publications/trs/areas/vaccines/influenza/ANNEX%2 26

03%20InfluenzaP99-134.pdf?ua=1. 27

18. Guidance on regulations for the transport of infectious substances. Geneva, World 28

Health Organization, 2005 (WHO/CDS/CSR/LYO/2005.22). 29

19. Tripp RA, Tompkins SM. Virus-Vectored Influenza Virus Vaccines. Viruses. 30

2014;6(8):3055-3079. doi:10.3390/v6083055. 31

20. Lambe T. Novel Viral Vectored Vaccines for the Prevention of Influenza. Molecular 32

Medicine. 2012;18(1):1153-1160. doi:10.2119/molmed.2012.00147. 33

21. Engelhardt, OG. Many ways to make an influenza virus--review of influenza virus 34

reverse genetics methods. Influenza Other Respir Viruses. 2013 May;7(3):249-56. doi: 35

10.1111/j.1750-2659.2012.00392.x. 36


Page 39

22. Garten W, Bosch FX, Linder D, Rott R, Klenk HD. Proteolytic activation of the 1

influenza virus hemagglutinin: The structure of the cleavage site and the enzymes involved in 2

cleavage. Virology. 1981 Dec;115(2):361-74. 3

23. Dormitzer et al, Synthetic generation of influenza vaccine viruses for rapid response to 4

pandemics. Sci Transl Med. 2013 May 15;5(185):185ra68. doi: 10.1126/scitranslmed.3006368. 5

24. Fulvini AA, Ramanunninair M, Le J, Pokorny BA, Arroyo JM, Silverman J, et al. 6

(2011) Gene Constellation of Influenza A Virus Reassortants with High Growth Phenotype 7

Prepared as Seed Candidates for Vaccine Production. PLoS ONE 6(6): e20823. 8

https://doi.org/10.1371/journal.pone.0020823. 9

25. Verity et al. (2011). Rapid generation of pandemic influenza virus vaccine candidate 10

strains using synthetic DNA. Influenza and Other Respiratory Viruses DOI:10.1111/j.1750‐11

2659.2011.00273.x. 12

26. Beare AS, Schild GC, Craig JW. Trials in man with live recombinants made from 13

A/PR/8/34 (H0N1) and wild H3N2 influenza viruses. Lancet, 1975, ii:729–732. 14

27. Coelingh KL, Luke CJ, Jin H, Talaat KR. Development of live attenuated influenza 15

vaccines against pandemic influenza strains. Expert Rev Vaccines. 2014 Jul;13(7):855-71. doi: 16

10.1586/14760584.2014.922417. 17

28. Rudenko L, Yeolekar L, Kiseleva I, Isakova-Sivak I. Development and approval of live 18

attenuated influenza vaccines based on Russian master donor viruses: Process challenges and 19

success stories. Vaccine. 2016;34(45):5436-5441. doi:10.1016/j.vaccine.2016.08.018. 20

29. Chen GL, Subbarao K. Live attenuated vaccines for pandemic influenza. Curr Top 21

Microbiol Immunol. 2009;333:109-32. doi: 10.1007/978-3-540-92165-3_5. 22

30. Sun X, Tse LV, Ferguson AD, Whittaker GR. Modifications to the Hemagglutinin 23

Cleavage Site Control the Virulence of a Neurotropic H1N1 Influenza Virus . Journal of 24

Virology. 2010;84(17):8683-8690. doi:10.1128/JVI.00797-10. 25

31. Nicolson C, Major D, Wood JM, Robertson JS. Generation of influenza vaccine viruses 26

on Vero cells by reverse genetics: an H5N1 candidate vaccine virus produced under a quality 27

system. Vaccine, 2005, 23:2943–2952. 28

32. Webby, RJ et al. 2004. Responsiveness to a pandemic alert: use of reverse genetics for 29

rapid development of influenza vaccines. The Lancet, Volume 363, Issue 9415, 1099 – 1103 30

33. WHO guidance on development of influenza vaccine reference viruses by reverse 31

genetics. Geneva, World Health Organization, 2005 (WHO/CDS/CSR/ GIP/2005.6). 32

34. Hatta M, Gao P, Halfmann P, Kawaoka Y. Molecular basis for high virulence of Hong 33

Kong H5N1 influenza A viruses. Science, 2001, 293:1840–1842. 34

35. Govorkova EA et al. Lethality to ferrets of H5N1 influenza viruses isolated from 35

humans and poultry in 2004. Journal of Virology, 2005, 79:2191–2005. 36


Page 40

36. Kurz J, Manvell RJ, Banks J. Avian influenza virus isolated from a woman with 1

conjunctivitis. Lancet, 1996, 348:901–902. 2

37. Yao Y, Mingay LJ, McCauley JW, Barclay WS. Sequences in Influenza A Virus PB2 3

Protein That Determine Productive Infection for an Avian Influenza Virus in Mouse and 4

Human Cell Lines. Journal of Virology. 2001;75(11):5410-5415. doi:10.1128/JVI.75.11.5410-5

5415.2001. 6

38. Almond JW. A single gene determines the host range of influenza virus. Nature, 1977, 7

270:617–618. 8

39. Taylor, H.R., Turner, A.J. A case report of fowl plague keraconjunctivitis. British 9

Journal of Ophthalmology, 1977; 61: 86–88. 10

40. Wagner R, Matrosovich M, Klenk HD. Functional balance between haemagglutinin and 11

neuraminidase in influenza virus infections. Rev Med Virol. 2002 May-Jun;12(3):159-66. 12

41. Gaymard A, Le Briand N, Frobert E, Lina B, Escuret V. Functional balance between 13

neuraminidase and haemagglutinin in influenza viruses. Clin Microbiol Infect. 2016 14

Dec;22(12):975-983. doi: 10.1016/j.cmi.2016.07.007. Epub 2016 Jul 15. 15

42 Goto, H, Kawaoka Y. A novel mechanism for the acquisition of virulence by a human 16

influenza A virus. Proc Natl Acad Sci USA 1998. 95, 10224-8. 17

18

43 Tumpey TT, Basler CF. Aguilar PV, Zeng H, Solorzano A, Swayne DE, Cox, NJ, Katz 19

JM, Taubenberger JF, Palese P, Garcia-Sastre A. Characterization of the reconstructed 1918 20

Spanish influenza oandmeic virus. Science 2005; 310:77-80. 21

44. WHO good manufacturing practices for biological products (Replacement of Annex 1 22

of WHO Technical Report Series, No. 822). In: WHO Expert Committee on Biological 23

Standardization: sixty-sixth report. Geneva: World Health Organization; 2015: Annex 2 (WHO 24

Technical Report Series, No. 999; 25

http://www.who.int/entity/biologicals/areas/vaccines/Annex_2_WHO_Good_manufacturing_pr26

actices_for_biological_products.pdf?ua=1, accessed 24 September 2017) 27

45. PHARMACEUTICAL INSPECTION CO-OPERATION SCHEME (PIC/S) (2014). 28

Guide to good manufacturing practice for medicinal products. Pharmaceutical Inspection 29

Convention, PIC/S, Geneva, Switzerland. Available at: 30

http://www.picscheme.org/publication.php?id=4. 31

46. Terrestrial Animal Health Code. Chapter 10.4. Infection with avian influenza viruses. 32

World Organisation for Animal Health (OIE). Available at: 33

http://www.oie.int/index.php?id=169&L=0&htmfile=chapitre_avian_influenza_viruses.htm#ch34

apitre_avian_influenza_viruses_0 35

47. Manual of diagnostic tests and vaccines for terrestrial animals. World Organisation for 36

Animal Health (OIE). Chapter 2.3.4. Avian Influenza (Infection with avian influenza viruses). 37

Available at: http://www.oie.int/fileadmin/Home/eng/Health_standards/tahm/2.03.04_AI.pdf. 38


Page 41

48. Beare AS, Hall TS. (1971) Recombinant influenza A viruses as live vaccines for man. 1

Lancet, ii:1271–1273. 2

49. Robertson JS, Cook P, Attwell A-M, Williams SP. Replicative advantage in tissue 3

culture of egg-adapted influenza virus over tissue-culture derived virus: implications for 4

vaccine manufacture. Vaccine, 1995, 13:1583–1588. 5

50. Nicolle Marshall, Lalita Priyamvada, Zachary Ende, John Steel, and Anice C. Lowen. 6

2013. Influenza Virus Reassortment Occurs with High Frequency in the Absence of Segment 7

Mismatch. PLoS Pathogens Volume 9, Number 6. Pages 1-11. 8

9

51. Dan Dou, Iván Hernández-Neuta, Hao Wang, Henrik Östbye, Xiaoyan Qian, et al., 2017, 10

Analysis of IAV Replication and Co-infection Dynamics by a Versatile RNA Viral Genome 11

Labeling Method. Cell Reports,V 20, Issue 1, p251–263. July 5, 2017. DOI: 12

http://dx.doi.org/10.1016/j.celrep.2017.06.021. 13

52. Belser, JA, Katz, JM, Tumpey, TM. The ferret as a model organism to study influenza 14

A virus infection Dis Model Mech 2011;4(5);575-579. 15

53. Govorkova EA et al. Lethality to ferrets of H5N1 influenza viruses isolated from 16

humans and poultry in 2004. Journal of Virology, 2005, 79:2191–2005. 17

54. Zitzow LA et al. Pathogenesis of avian influenza A (H5N1) viruses in ferrets. Journal of 18

Virology, 2002, 76:4420–4429. 19

55. Manual for standards for diagnostic tests and vaccines for terrestrial animals. 5th ed. 20

Paris, Office International des Epizooties (OIE), 2004. 21

56. Li Q, Zhou L, Zhou M, et al., Epidemiology of human infection with avian influenza 22

A(H7N9)virus in China. N Engl J Med 2014;370:520-32 23

57. Nelson J et al. Influenza virus strains selectively recognize sialyooligo- saccharides on 24

human respiratory epithelium: the role of the host cellin selection of haemagglutinin receptor 25

specificity. Virus Research, 1993, 29:155–165. 26

58. Matrosovich MN et al. Human and avian influenza viruses target different cell types in 27

cultures of human airway epithelium. Proceedings of the National Academy of Sciences, 2004, 28

101:4620–4624. 29

59. Chen Y et al. Human infections with the emerging avian influenza A H7N9 virus from 30

wet market poultry: clinical analysis and characterisation of viral genome. Lancet. Volume 381, 31

Issue 9881, 1–7 June 2013, Pages 1916-1925. https://doi.org/10.1016/S0140-6736(13)60903-4. 32

60. Beare AS, Webster RG. Replication of avian influenza viruses in humans. Archives of 33

Virology, 1991, 119:37–42. 34

61. Swayne DE, Halvorson, DA. Influenza. In: Saif YM et al. eds. Diseases of Poultry. 11th 35

ed. Ames, Iowa, Iowa State University Press, 2003:135–160. 36


Page 42

62. Keawcharoen J et al. Avian influenza H5N1 in tigers and leopards. Emerging Infectious 1

Diseases, 2004, 10:2189–2191. 2

63. Pecoraro HL et al., Epidemiology and eclogy of H3N8 canine influenza viurses in US 3

shelter dogs. J Vet Intern Med 2014:28:311-18. 4

5

64. Lee CT et al., Outbreak of influenza A(H7N2) among cats in an animal shelter with cat-6

to-human transmission-New York City, 2016. Clin Infect Dis 2017;65:1927-29. 7

65. Subbarao K et al. Evaluation of a genetically modified reassortant H5N1 influenza A 8

virus vaccine candidate generated by plasmid-based reverse genetics. Virology (2003) 305: 9

192–200. 10

66. Webby RJ, Perez DR, Coleman JS, Guan Y, Knight JH, Govorkova EA, McClain-Moss 11

LR, Peiris JS, Rehg JE, Tuomanen EI, Webster RG. Responsiveness to a pandemic alert: use of 12

reverse genetics for rapid development of influenza vaccines. Lancet. 2004 Apr 13

3;363(9415):1099-103. 14

67. Ito T, Couceiro JN, Kelm S, Baum LG, Krauss S, Castrucci MR, Donatelli I, Kida H, 15

Paulson JC, Webster RG, et al. Molecular basis for the generation in pigs of influenza A 16

viruses with pandemic potential. J Virol. 1998 Sep;72(9):7367–7373. 17

68. WHO global influenza preparedness plan. The role of WHO and recommendations for 18

national measures before and during a pandemic. WHO/CDS/CSR/GIP/2005.5. World Health 19

Organization, 2005. 20

http://www.who.int/csr/resources/publications/influenza/WHO_CDS_CSR_GIP_2005_5.pdf, 21


69. European Union. EN 149:2001. Respiratory protection devices — Filtering half-masks 23

to protect against particles — Requirements, testing, marketing. 24

70. European Directive 2000/54/EC (2007). 25

https://osha.europa.eu/en/legislation/directives/exposure-to-biological-agents/77, accessed 22 26

Nov 2017. 27

71. Biosafety [website]. Atlanta, GA: Centers for Disease Control and Prevention, 28

https://www.cdc.gov/biosafety/, accessed 22 Nov 2017. 29

30


Page 43

Appendix 1 1

Testing for attenuation of influenza vaccine viruses in ferrets 2

3

Laboratories conducting testing for attenuation of influenza viruses in ferrets should make use 4

of a panel of standard/or reference viruses (‘pathogenicity standards’ in the following section) 5

and defined experimental outcomes for pathogenicity testing. The pathogenicity standards (to 6

be established by WHO laboratories) serve as benchmarks for the pathogenicity test in ferrets 7

and delineate the expected outcomes. The use of these standards will ensure that attenuation of 8

CVVs is being measured against common parameters independently of subtype. The CVV to 9

be tested must show parameters of pathogenicity below the predefined values of a high 10

pathogenicity standard and in line with those of an attenuation standard in order to be 11

designated as attenuated. Comparative attenuation in comparison with the parental wild-type 12

virus is not necessary in this case. However, laboratories that have the capacity to evaluate 13

attenuation of a CVV compared with the parental wild-type virus can continue to do so. To 14

account for the expected experimental variability of results across different laboratories, the 15

pathogenicity standards should be tested in ferrets at each testing laboratory according to the 16

experimental protocol shown below when establishing the ferret model for pathogenicity 17

testing and at regular intervals thereafter. The outcomes of these tests should fall within the 18

limits described for the pathogenicity standards. In cases of discrepancy, a review of the ferret 19

model should be conducted and advice should be sought from experienced WHO laboratories. 20

Test virus 21

The 50% egg or tissue culture infectious dose (e.g. EID50, TCID50) or plaque-forming units 22

(PFU) of the reassortant CVV or pathogencitiy standard will be determined. The infectivity 23

titres of viruses should be sufficiently high to allow infection with 107 to 106 EID50, TCID50 or 24

PFU of virus and diluted not less than 1:10. Where possible, the pathogenic properties of the 25

donor PR8 should be characterized thoroughly in each laboratory. 26

Laboratory facility 27

Animal studies with the CVV and the pathogenicity standards should be conducted in animal 28

containment facilities in accordance with the proposed containment levels shown in Table 1. 29

For untested CVVs, the biocontainment level to be used for the ferret safety test is the one 30

shown for the respective wild-type virus, except for ‘Reassortants and modified viruses derived 31


Page 44

from synthetic DNA representing H5 and H7 HPAI viruses’, for which the containment level 1

proposed in Table 1 (BSL2 enhanced) can be used (1). An appropriate occupational health 2

policy should be in place (2). 3

Experimental procedure 4

Outbred ferrets 4-12 months of age that are serologically negative for currently circulating 5

influenza A and B viruses and the test virus strain are anesthetized by either intramuscular 6

administration of a mixture of sedatives (e.g. ketamine (25 mg/kg) and xylazine (2 mg/kg) and 7

atropine (0.05 mg/kg)) or by suitable inhalant anesthetics. A standard virus dose of 107 to106 8

EID50 (or TCID50 or PFU) in 0.5 to 1 ml of phosphate-buffered saline is used to inoculate 9

animals. The dose should be the same as that used for pathogenesis studies with the wild-type 10

parental virus, if used, or the pathogenicity standards previously characterised and regularly 11

assessed in the laboratory. The virus is slowly administered into the nares of the sedated 12

animals, reducing the risk of virus being swallowed or expelled. A group of four to six ferrets 13

should be inoculated. One group of two to three animals should be euthanized on day 3 or 4 14

after inoculation and samples should be collected for estimation of virus replication from the 15

following organs: spleen, intestine, lungs (samples from each lobe and pooled), brain (anterior 16

and posterior sections sampled and pooled), olfactory bulb of the brain, and nasal turbinates. If 17

gross pathology demonstrates lung lesions similar to those observed in wild-type viruses, 18

additional lung samples may be collected and processed with haematoxylin and eosin (H&E) 19

staining for histopathological evaluation. The remaining animals are observed for clinical signs, 20

which may include weight loss, lethargy [based on a previously published index (3)], 21

respiratory and neurologic signs, and increased body temperature. Collection of nasal washes 22

from animals anesthetized as indicated above should be performed to determine the level of 23

virus replication in the upper airways on alternate days after inoculation for up to seven days. 24

At the termination of the experiment on day 14 after inoculation, a necropsy should be 25

performed on at least two animals and organs collected. If signs of substantial gross pathology 26

are observed (e.g. lung lesions), the organ samples should be processed as described above for 27

histopathology. 28

Expected outcome 29

Clinical signs of disease such as lethargy and/or weight loss should be within the predefined 30

ranges of acceptable pathogenicity defined by the pathogenicity standards. Viral titres of the 31

vaccine strain in respiratory samples should be within the ranges of acceptable virus replication 32


Page 45

defined by the pathogenicity standards. Replication of the CVV should be restricted to the 1

respiratory tract. Virus isolation from the brainis not expected. However, detection of virus in 2

the brain has been reported for seasonal H3N2 viruses (4); this is due to detection of virus in 3

the olfactory bulb. Therefore, should virus be detected in the anterior or posterior regions of 4

the brain (excluding the olfactory bulb) the significance of such finding may be confirmed by 5

performing immunohistochemistry to detect viral antigen and/or histopathological analysis of 6

brain tissue collected on day 14 after inoculation. Detection of viral antigen and/or neurological 7

lesions in brain tissue would confirm virus replication in the brain. Presence of neurological 8

signs and confirmatory viral antigen and/or histopathology in brain tissue would indicate a lack 9

of suitable attenuation of the CVV. 10

11

References for Appendix 1 12

13

1. World Organization for Anima Health (OIE). CHAPTER 2.3.4. AVIAN INFLUENZA, 14

Appendix 2.3.4.1- “Biosafety guidelines for handling highly pathogenic avian influenza viruses 15

in veterinary diagnostic laboratories” (adopted in May 2015), in Manual of Diagnostic Tests 16

and Vaccines for Terrestrial Animals 2017. http://www.oie.int/en/international-standard-17

setting/terrestrial-manual/access-online/, accessed 22 November 2017. 18

2. WHO laboratory biosafety guidelines for handling specimens suspected of containing avian 19

influenza A virus. 12 January 2005. 20

http://www.who.int/influenza/resources/documents/guidelines_handling_specimens/en/, 21

accessed 22 November 2017. 22

3. Reuman PD, Keely S, Schiff GM. Assessment of signs of influenza illness in the ferret 23

model. Laboratory Animal Science, 1989, 42:222–232. 24

4. Zitzow LA et al. Pathogenesis of avian influenza A (H5N1) viruses in ferrets. Journal of 25

Virology, 2002, 76:4420–4429. 26

27

28

Documents

WHO biosafety risk assessment and guidelines for the ... WHO/INFLUENZA/DRAFT/NOVEMBER 2017 3 ENGLISH ONLY 4 5 WHO biosafety risk assessment and guidelines for the production 6 and