Download part 2b risk assessment of genetically modified human and animal

Part 2B
PART 2B
RISK ASSESSMENT OF GENETICALLY MODIFIED HUMAN
AND ANIMAL VIRUSES AND VIRAL VECTORS
The Contained Use Regulations 2000
Introduction
1. This part is intended to provide guidance on
the risk assessment of work that either involves
the genetic modification of human or animal
viruses (including site-directed mutagenesis and
approaches involving reverse genetics) or the
cloning of foreign genes into eukaryotic viral
vectors. This supplements the general guidance
in Parts 1 and 2 of the Compendium. For the
purposes of this guidance, the term "animal" is
taken in its broadest sense, and includes both
vertebrates and invertebrates.
3. The guidance has been updated to take
account of the new requirements of the
Genetically Modified Organisms (Contained Use)
Regulations 20004. These new Regulations take
into account amendments to EU directive 90/219
that came into force in June 2000.
4. Schedule 3 of the Contained Use Regulations
2000 sets out factors which must be considered in
risk assessments and the steps that have to be
included.
5. The main difference from the previous
legislation is the requirement for the work to be
given an activity classification (which replaces the
Gene therapy
old classification of the GMM into Group I or
Group II). Otherwise the steps and considerations
2. The use of viral vectors for gene therapy
that are required in the risk assessment process
requires rigorous control of production and safety are the same as those in previous editions of
testing methods. Approval from the Medicines
ACGM’s Compendium. The order of the steps set
Control Agency1 and the Gene Therapy Advisory
out in the following guidance and in the risk
Committee via the Secretariat at the Department
assessment examples at Annex IV is slightly
of Health2 should be obtained before any gene
different to that in Schedule 3. This is because in
therapy research is attempted on human subjects. Schedule 3 the human health and environmental
In addition some gene therapy projects are
considerations have been combined. However, it is
covered by the Contained Use Regulations. Where permissible to change the order of consideration
this is so, projects should be risk assessed to
and we continue to recommend separate
determine the risk class. Particular attention
consideration of human health and environmental
should be paid to the possibility that patients may
safety.
shed the virus following treatment (either at the
point of inoculation or in bodily excretions). It is
Work identified by ACGM as being likely to
possible that, if it is foreseeable that there will be
raise important safety issues
significant shedding, the gene therapy may
constitute a deliberate release3 and be regulated
6. When replication competent viruses are being
under the Environmental Protection Act and the
constructed, or are likely to be generated in the
Genetically Modified Organisms (Deliberate
course of work, there are some types of
Release) Regulations 1992 (as amended).
experiment where particular caution should be
However, if the gene therapy is a contained use
taken. These are cases where the pathogenicity of
then it is likely to fall into risk class 1 and will,
the viral vector may be enhanced (e.g. the
therefore, not require notification. This is because insertion of a gene involved in evading the host
such projects are likely to use replication defective immune response) or where the host range or
viral vectors which are unlikely to cause harm.
tissue tropism may be extended (e.g. the insertion
of a capsid or envelope gene from a related virus).
In such cases serious consideration should be
1
2
3
4
Medicines Control Agency, Market Tower, Nine Elms Lane, London SW8 5NQ.
Department of Health, Room 417, Wellington House, 133-135 Waterloo Road, London SE1 8UG Tel: 020 7972 4911
For deliberate releases contact the Biotechnology Unit at the Department of the Environment, Transport and the Regions.
Refered to from now on as the Contained Use Regulations 2000
Issued: March 2000
Part 2B
given to increasing the physical containment of the (i) Consideration of the predicted properties of
GMM by at least one level as compared to the
the genetically modified virus to determine if
parental virus.
there are any potential mechanisms by which
it could represent a hazard to human health
Access, Expression and Damage
7. The scheme outlined in the Annexes to Part
2A consisting of numerical factors under the
headings of Access, Expression and Damage is
not applicable for viral vectors.
and how severe the consequences might be.
(a) Hazards associated with the recipient virus
or viral vector
Structure of the guidance
9. Particular care must be given to the
assessment of viruses or viral vectors with an
8. The following procedures for risk assessment actual or potential ability to infect humans or
and the assignment of containment and control
human cells. The COSHH Regulations 1999
measures are recommended.
require that all biological agents (in this context,
any virus or viral vector that may cause any
- (i) Consideration of the predicted properties of
infection, allergy, toxicity or any other hazards to
the genetically modified virus to determine if
human health ) are classified into one of four
there are any potential mechanisms by which
hazard groups by reference to the Approved List of
it could represent a hazard to human health
biological agents, or the classification criteria set
and how severe the consequences might be.
out in COSHH, if the virus in question does not
- (ii) Consideration of the likelihood that, in the appear in the Approved List. Further guidance can
event of exposure, the GMM could actually
be found in Part 1 of the Compendium and in the
cause harm to human health - including
latest edition of the Advisory Committee on
consideration of uncertainty.
Dangerous Pathogens (ACDP) publication
"Categorisation of biological agents according to
- (iii) Assignment to a provisional containment
hazard and categories of containment."5 Specific
level. This represents an initial judgement as
guidance on certain commonly used viral vectors
to the containment measures that are
is given in Annex III of this part of the guidance.
appropriate to safeguard human health by
comparison of the GMM with the biological
agents hazard groups or other suitable
classification schemes.
Viral vectors with reduced pathogenicity
- (iv) Consideration of the nature of the work to
be undertaken and a detailed review of the
control measures necessary to safeguard
human health.
10. COSHH requires the prevention of exposure to
a biological agent by substituting a biological
agent which is less hazardous, where it is
reasonably practicable. For genetic modification
- (v) The identification of any hazards to the
work involving viruses with a human host range
environment and then, on the assumption that
this can be equated to a statutory requirement to,
the controls necessary to safeguard human
whenever possible, use disabled or attenuated viral
health have already been applied, the
vectors with a reduced pathogenicity. Furthermore,
assignment of any additional containment
where appropriate, use of a vector without a
measures to protect the environment.
human host range should be considered.
- (vi) Assignment of the final activity Class (1,
2, 3 or 4).
11. The origin and mechanism of attenuation
should be well understood and will form an
important part of the risk assessment. In
assessing whether a viral vector is adequately
disabled the possibility of reversion or
RISK ASSESSMENT FOR HUMAN HEALTH
5
Categorisation of biological agents according to hazard and categories of containment (4th edition, 1995 ISBN
0-7176-1038-1)
Issued: March 2000
Part 2B
complementation should be considered and it
should be confirmed that the virus is disabled after
modification. The likelihood of reversion will
depend on the mechanism of attenuation; deletion
mutants are less likely to revert to wild-type than
point mutations or conditional lethal mutants.
(b) Hazards arising directly from the inserted
14. Further specific information on disabled
vectors, including adenovirus, alphavirus,
baculovirus, poxvirus, herpes simplex and
retrovirus vectors, is set out in Annex III. If there is
any doubt about the correct biological agent
hazard group for an attenuated or disabled virus,
you are advised to contact HSE in Bootle.
Oncogenes
gene product
16. The insertion of additional nucleic acid
sequences into a viral vector can give rise to
potential adverse effects. These may result either
12. Insertion of a gene into the site of any
from the direct effects of an expressed gene
disabling mutation is expected to reduce the
product or as a consequence of an alteration in
likelihood that recombination events could result in the overall properties of the GMM (see section c
the generation of replication competent virus
below). In considering the direct effects particular
expressing the inserted gene. Insertion at the site attention should be paid to the level of expression
of disablement thus increases the effective
and site of insertion of the gene(s) and whether
biological containment. This approach should be
there is a known or suspected pharmacological or
followed whenever practicable, especially when
physiological effect, including the possibility of
working with harmful genes. Where it is proposed effects other than those being sought in the
to insert a harmful gene into a site other than the
construction.
site of a disabling mutation, full justification should
- For example, a non-harmful human protein
be given in the risk assessment.
(such as myelin pre-protein) expressed in
vaccinia virus may provoke auto-immune
13. Where the viral vector is an attenuated or
disease if an operator were to be accidentally
disabled derivative of a human pathogen it may be
infected.
acceptable for it to be reclassified into a hazard
group different from that of the parental virus in the
17. Particular attention should be paid to the
Approved List of Biological Agents. As an
insertion of genes which may alter the growth,
example, wild-type adenovirus is a Hazard Group
replication or differentiation of cells, for example;
2 biological agent but an E1a deletion derivative
oncogenes, potentially oncogenic sequences, or
cannot replicate unless the deletion is
genes encoding biologically active proteins (e.g.
complemented in trans. Such an E1a-deleted
cytokines, growth factors or toxins) into viruses
vector can therefore be considered a biological
capable of infecting human cells. Work with such
agent "unlikely to cause human disease" and
modified viruses may pose serious consequences
assigned to Hazard Group 1, with containment
for people who are occupationally infected or
level 1 as the minimum level of containment.
exposed. Additional containment and control
(Note that the reclassification applies only to the
measures over and above those required for the
disabled parental virus; any harmful properties
viral vector will generally be necessary and must
associated with the insert or the final genetically
be applied at the correct level following the risk
modified virus may present an increased risk and
assessment.
warrant additional control measures.)
15. Experiments using viral vectors that do not
normally infect human cells in culture, and for
which there is no evidence of human infection, are
considered to represent a minimal risk to the
operator and ACGM containment level 1 is
sufficient to protect human health. A higher
standard of containment may however be required
to control risks to other species.
Issued: March 2000
18. ACGM recognises that there is no precise
definition of an oncogene. Genes known to be
involved in the generation of tumours in humans
and other animals could form the basis of a
definition but many other genes generate
phenotypes in cultured cells that suggest that
they could also be involved in tumorigenesis. Such
phenotypes include transformation, density
independent growth, anchorage independent
growth and immortalisation, terms which
themselves are difficult to define.
19. Users, in consultation with their local GMSC,
should pay special attention to the potential risks
of work with sequences that may be regarded as
oncogenic and should feel free to request further
Part 2B
advice from HSE. The following points may assist
users with their risk assessments.
23. There is separate guidance (Part 1 of the
Compendium and in Annex III to Part 2A) on
handling naked oncogenic DNA and there are
duties under the COSHH Regulations to assess
the potential risks of such substances. There are
few examples of naked DNA producing tumours in
animals but in one case, activated ras (admittedly
in a large dose and after scarification of the skin)
induced transformation of mouse skin endothelial
cells. Results from experiments on DNA
immunisation show that gene expression can
occur from injected naked DNA, so all DNA should
be handled with caution, particularly where it
includes harmful sequences and in situations
where sharps are being used.
20. The formation of a cancer requires the
activation (by mutation or over-expression) of
oncogenes and the inactivation (by mutation or
deletion) of tumour suppresser genes. It is
normally a multi-step process requiring the
activation or inactivation of several genes, with the
cells becoming progressively more tumorigenic as
genetic changes accumulate. The introduction of
one change into a small number of cells is unlikely
to cause cancer (although it has been observed in
some experimental systems). Nevertheless,
particularly if the gene is stably introduced into a
stem cell, that cell and its progeny may be one
step nearer to forming a cancer. Such a potentially
(c) Hazards arising from the alteration of
serious outcome should not be dismissed lightly.
existing pathogenic traits
21. Oncogenes can often induce tumorigenicity in
cells in culture that already carry genetic
24. Whilst many modifications of viruses do not
alterations (e.g. conferring immortalisation) but it
involve the insertion of foreign genes whose
is rare for single oncogenes to induce
products are inherently harmful, adverse effects
tumorigenicity in cultured primary cells.
may nevertheless arise as the result of
Combinations of oncogenes can be more effective.
exacerbation or alteration of existing pathogenic
The introduction of an adenovirus type 12
traits. This may arise as the result of the product
sequence that includes both the E1a and E1b
of an inserted gene acting alongside existing
genes into human retinoblasts can cause them to
pathogenic determinants. Alternatively it is
form tumours when inoculated into the brain (but
possible that either the modification of a normal
not the skin) of immunodeficient mice. Activated
viral gene or its substitution with a gene from a
ras does not induce tumorigenicity in primary cells
related virus may alter pathogenicity. Therefore,
from a number of different species but, in
the following points should be considered (the list
combination with a second oncogene (e.g. myc)
is not exhaustive):
can induce tumorigenicity in primary rat embryo
fibroblasts. Sequences that inactivate tumour
- Alteration of tissue tropism or host range:
suppresser genes (e.g. dominant negative
Is there a possibility that the structure of the
mutants, antisense constructs) may also
receptor binding site will be altered or will the
co-operate with oncogenes.
product of the inserted gene be incorporated
on the virus surface with the possibility of
22. Many known oncogenes and tumour
forming a novel receptor binding capacity?
suppresser genes function in cellular signalling
Cell or tissue tropism may also be affected
pathways and almost any gene that encodes a
by alterations in the transcriptional control of
protein involved in cell-to-cell or intracellular
viral genes.
signalling, interaction with the environment, cell
- Increase in infectivity or pathogenicity:
cycle control, differentiation or programmed cell
Could the modified virus show a decreased
death (apoptosis), could be regarded as potentially
susceptibility to host defence mechanisms,
oncogenic in some circumstances (e.g. perhaps if
for example, as a consequence of the
expressed constitutively at high levels). For
insertion of a gene involved in evading the
example, expression of some genes (e.g.those
host immune response? Is the recombinant
encoding growth factors) can allow proliferation of
likely to have enhanced effects upon an
cells which otherwise would not grow in culture
immuno-compromised host, beyond those
(e.g. interleukin-2 expression in T lymphocytes)
normally expected with the parent virus?
and expression of other genes (e.g. the E6 gene of
human papilloma virus type 16) can confer an
- Availability of prophylaxis or therapy: Will
extended life span on cells in culture which,
viral susceptibility to anti-viral drugs (where
nevertheless, still undergo senescence.
these are available) be affected by the genetic
modification? Can vaccination or normal
immune status be expected to protect
Issued: March 2000
Part 2B
against the modified virus?
Deliberate alteration of tissue tropism or
specificity
25. There is increasing interest in the
modification of virus tropism (usually by
modification of the receptor binding protein) for
scientific or therapeutic exploitation. Given our
current understanding of viral pathogenesis, the
consequences of changes in tropism are difficult
to predict. The techniques available for modifying
tropism are in their infancy, but are likely to
develop rapidly. In assessing the risk of
manipulations designed to modify tropism it must
be assumed that the experiments will be
successful. In general, experiments designed to
generate replication-competent viruses with novel
tropism or other novel pathogenic characteristics
will attract high levels of containment, until the
biological characteristics of the recombinant have
been determined. As a general rule, it should be
assumed that a virus capable of infecting cell
types that are not susceptible to the parental virus
will require a higher level of containment than the
parental virus. During the risk assessment of
such work a number of questions need to be
considered. For example,
- could the route of transmission of the
modified virus be altered?
virus that might be regarded as seriously harmful.
If sequences are inserted at the site of the
disabling mutation, it is reasonable to assume that
the repair of the disabling mutation would result in
the loss of the insert.
(ii) Consideration of the likelihood that, in the
event of exposure, the genetically modified
virus could actually cause harm to human
health
27. The first stage in the risk assessment
process that has been outlined above involved
identifying those features of the GMM which have
the potential to cause harm to humans. It is,
however, recognised that in some cases, while it
may be possible to draw up theoretical scenarios
to suggest that a modified virus may be hazardous
to human health, there can sometimes be
justification to say that the likelihood of these
scenarios being realised is small.
28. Factors which come into play when
considering likelihood include a judgement as to
the fitness of the modified virus and the analysis of
the probability that rare events may occur (e.g.
mutations which overcome disabling mutations).
29. As part of such considerations it should be
recognised that during the course of evolution
viruses have proved particularly adept at
responding to selective pressures by infecting new
cell types or host organisms. This is a
(d) Transfer of harmful sequences to related
consequence of the high level of genetic variability
viruses
exhibited by viruses. In particular, RNA viruses
replicate by an error-prone mechanism which
means that mutant genomes are continually being
26. Whilst the phenotype of the recombinant virus
generated. Normally these variants will be
that is under construction is the primary
maintained at low frequencies by negative
consideration, some thought must also be given to
selection. However, in a situation where such a
the possibility that harmful sequences may be
virus replicates in an environment that is different
transferred as the result of recombination.
from that in which it is normally found, the
Important scenarios which need to be considered
probability of one of the genetic variants becoming
at this stage are the possibilities that a disabled
dominant will be much higher. When undertaking
vector might recombine with the parental virus or
risk assessments of genetically modified viruses it
with viral sequences present in a packaging cell
is important to have some awareness of this
line. One way in which the former scenario might
genetic variability. Even if the virus that is initially
arise is as the result of a worker, who is already
constructed is not well adapted to growth in a
carrying an infection with the wild-type virus,
particular cell line or host organism there is a
becoming exposed to the disabled virus. Another
possibility that it will adapt as new variants arise.
is as the result of accidental cross contamination
Of course it would be possible to take such
in a laboratory handling both disabled and
scenarios to extreme levels but this is certainly
wild-type virus. Particular care should be taken if
not the intention of this guidance. It is merely
the repair of a single disabling mutation is
intended to indicate that it is necessary to
reasonably foreseeable and would give rise to a
- what are the predicted effects of the modified
virus in tissues it would not normally infect?
Issued: March 2000
Part 2B
proceed with caution and where possible use
recipient viruses that are replication defective,
thereby virtually eliminating problems relating to
genetic variability.
of a foreign gene results in a construct that is
close to the packaging constraints of the virus
there is a tendency for the foreign gene to be
rapidly deleted. Similarly, the loss of a gene which
conferred environmental protection, for example
30. Issues relating to the likelihood of harm arising resistance to desiccation or UV light, would also
will, by their very nature, be very difficult to handle reduce the potential for spread.
in situations where there is no firm data on which
to make a judgement. Therefore, a great deal of
33. Consideration of the fitness of a virus is a
caution must be applied when seeking to discount legitimate part of a risk assessment, but should
on the basis of likelihood those predicted
not be based merely on supposition, but on
properties of the modified virus which have been
established scientific knowledge. Until it can be
identified in Section (i) as being potentially
demonstrated that a particular type of modification
harmful. In general, the weighting given to
will render a virus less fit than the parental virus
information used in the consideration of likelihood (for example by experimental data, or through the
should reflect the quality of the supporting data.
literature etc.) the precautionary approach should
Where the information is either anecdotal or based be followed. This is particularly so where
on a series of roughly-drawn assumptions it may
counter-arguments can be made for the foreign
be necessary to assume the worst and act
insert giving the virus a competitive advantage.
accordingly.
(a) The ability of a GM virus to establish an in
vivo infection and the efficiency of subsequent
(b) Probabilistic considerations that relate to
the likelihood of occurrence of rare events
in vivo propagation. Assessment of the
34. In some instances it may be possible to
assign a frequency - precise or approximate - to
an event. This is particularly true in the case of
recombination and reversion frequencies, leading
31. It is common for discussion of the potential
to the production of replication competent viruses,
hazards of viral work to focus on pathogenicity.
where there may be published data. In other
However, pathogenicity is not the only factor which
cases, it may be possible to adopt only a
determines the potential of a virus to cause harm
semi-quantitative frequency or descriptive
and it is equally valid to consider factors which
assessment of the probability, based on
relate to the ability of a virus to spread in vivo i.e.
experience with this or other comparable viruses
viral 'fitness'.
or with the particular working methods.
'fitness' of genetically modified viruses.
32. This concept of fitness is difficult to define but
it will clearly be important in assessing the
potential for the spread of a virus, if there were to
be a breach of containment. For example, a
clearly established example of the fitness of a
virus being reduced by genetic modification is the
case of insertion of a foreign gene into the E3
locus of adenovirus. Such an insertion makes the
virus more susceptible to immune surveillance and
so the modified virus can be considered as less
likely to establish an infection and spread in the
community. However, it is interesting to note that
in this case it can be argued that the
pathogenicity of the virus is actually increased.
Thus in the event of viral infection being
established, for example in an
immuno-compromised individual, there can be a
more severe inflammatory response than would be
the case with wild-type virus. Another example
relating to viral fitness has been demonstrated with
a number of vector systems. When the insertion
Issued: March 2000
35. For example, the fact that co-infection of a
single cell with related forms of the same virus has
not been observed in vivo should not be taken as
firm evidence of a specific partition mechanism to
prevent this happening. The lack of any observed
co-infection would be relevant when making a
judgement on the likelihood of recombination or
complementation, but such an event could not be
totally discounted in the absence of firm scientific
data.
36. The judgements surrounding the assessment
of likelihood can be illustrated by considering the
situation of a harmful gene cloned into a vector
containing a single disabling mutation. If the
insert was located at a separate site from the
disabling mutation homologous recombination
could produce a replication competent virus
carrying the harmful gene. This could occur as a
result of cross contamination with the wild-type
virus in the laboratory, recombination with viral
Part 2B
sequences in the packaging cell line, or following
accidental inoculation of a person already infected
with wild-type virus (particularly in the case of a
virus that is present in the general population in a
latent or persistent form). Such events would be
deemed sufficiently likely to require additional
control measures. However, if the harmful gene
was inserted at the site of the disabling mutation,
although it would be theoretically possible that two
illegitimate recombination events could restore the
replication function of the disabled virus, such an
event would be considered as highly unlikely, and
can be discounted in most cases.
(iii) Assignment of a provisional containment
level
37. The next step is to assign a provisional
containment level. This is based on consideration
of the following factors:
- the biological agents hazard group of the
parental virus (or the rating according to
another suitable classification scheme);
- any identifiable hazards arising as a
consequence of the genetic modification;
- the severity of any harmful consequences,
and the likelihood that, in the event of
exposure, they might occur.
hazardous or about the same as the recipient.
Thus by comparing the GMM with the relative
hazards presented by other organisms in the
biological agents classification a provisional
containment level can be derived.
40. At this stage only a broad brush estimate of
the containment necessary to safeguard human
health is made. In many cases it is likely to
correspond to the containment level that is
appropriate for the recipient virus (if it is a human
pathogen). However, in some cases there may be
reason to believe that the GMM will be
considerably more hazardous than the recipient
virus (e.g. where a harmful gene has been inserted
into a replication-competent virus). In such cases
it may be appropriate to assign the GMM to a
higher provisional containment level than that
appropriate for the recipient virus.
41. Because the provisional containment level is
being assigned on the basis of risks to human
health some animal pathogens may be given a
provisional classification of 1, as level 1 may be
sufficient to protect human health. The risk to
animals and the wider environment will then be
considered separately as set out below in the
section on environmental risk assessment.
(iv) Consideration of the nature of the work to
38. When assigning the provisional containment
be undertaken and a detailed review of the
level it should be noted that the potential to cause
control measures necessary to safeguard
harm may involve a combination of the factors
identified within different subsections of the hazard human health
identification process.
- For example, the fact that the disabling
mutation in a viral vector shows a high
reversion frequency may not appear to be
particularly significant, if the wild-type is only
weakly pathogenic. However, this high
reversion frequency could have very serious
consequences if the vector were to be used
for the cloning of a toxin or oncogene and
there would need to be some additional
containment and control measures.
39. Where it is possible to identify the biological
agents hazard group of the recipient virus, this will
provide a basis for assigning a provisional
containment level. (Hazard group 1 = level 1,
hazard group 2 = level 2 etc). A judgement can
then be made as to whether the modification will
result in a GMM which is more hazardous, less
Issued: March 2000
42. Stage (iii) involved the assignment of a
provisional containment level. However, this does
not take into account the nature of the actual work
or detailed consideration of the required control
measures. It provides only a broad brush
assignment. Stage (iv) therefore involves
refinement of the control measures to safeguard
human health. Two aspects are considered:
- (a) whether the minimum requirements of the
provisional containment level are sufficient to
control all of the potential harmful properties
of the GMM, or whether some additional
measures might be required. (It is also
possible that this more detailed consideration
will show that some specific measures from
the provisional containment level are not
necessary in which case a derogation may
be sought);
Part 2B
- (b) the nature of the activity to be undertaken,
especially consideration of any non-standard
operations.
(a) Assignment of any additional control
measures
43. Further consideration should be given to
whether the minimum requirements for the
provisional containment level are adequate or
whether some additional measures over and above
the minimum need to be applied. It may be
possible to identify some particular aspect of the
experimental design or work procedures which
should be improved in order to minimise the risk to
human health and safety. For example, some
projects may be assigned to containment level 2
with one or two additional measures taken from
the requirements of containment level 3.
Management systems may also need to be
implemented or improved e.g. increased
monitoring by internal inspections and systems to
ensure that workers are adequately trained and
fully aware of the potential hazards.
44. Further guidance on containment measures
is provided in Part 3A or can be obtained from
HSE, Bootle.
45. The Contained Use Regulations 2000 also set
out the underlying principles of containment and
control measures for all GMMs. These include the
principles of good microbiological practice and
good occupational safety and hygiene. These
measures are also required for work with biological
agents under COSHH. In the case of GM viruses
capable of infecting human cells, ACGM feels that
it is prudent to also recommend the following:
- Measures should be taken to prevent cross
contamination during laboratory work to
minimise the possibility of adverse
consequences resulting from recombination
or complementation. It is therefore not good
practice to use aliquots taken from the same
bottle of medium for culturing different virus
infected cell lines and laboratory workers
should be discouraged from sharing bottles of
medium. Moreover, the handling of wild-type
viruses in the same laboratory as
recombinant viruses, should be avoided
wherever possible. Where it is not possible
(due to space constraints), there should be
both spatial and temporal separation of the
work. This can be achieved by designating
cabinets for particular types of work, or
fumigating before working with wild-type after
Issued: March 2000
work with recombinant virus. Care should
also be taken when storing or incubating
cultures with, if possible, separate incubators
or freezers being used. If that is not practical,
it is important to clearly label cultures, and
store wild-type and recombinant viruses on
separate shelves.
- Consideration should be given to the need for
testing to detect the presence of adventitious
agents and replication competent virus
(RCV). This issue may be of particular
importance in situations where the risk
assessment relies heavily on the premise
that RCV are not present.
- The person responsible for the work should
be satisfied that the laboratory local rules give
effective guidance on the maintenance of
laboratory discipline and on avoiding
accidental inoculation. Moreover there should
be a programme of internal inspections and/or
active monitoring by the BSO (or other
competent person) to ensure that the local
rules are satisfactorily implemented.
- All workers should be trained in good
laboratory techniques before commencing
work and should be fully aware of the
potential hazards of the work. In particular,
they should have a working knowledge of the
nature and importance of any disabling
mutations which provide biological
containment.
- In order to minimise the risk of accidental
colonisation with infected cell lines, users
should not infect cultures of their own cells,
nor, as a general rule, those of their
immediate family or other members of the
laboratory.
(b) Consideration of the nature of the work
46. This involves a consideration of whether the
work that will be undertaken involves any
non-standard operations that may involve risks
that are not accounted for in the general
requirements for a containment level. Examples of
such activities might include the following:
- inoculation of animals with modified virus;
- the use of equipment likely to generate
aerosols e.g. sonication or mixing;
- the use of high titres of virus.
Part 2B
47. If it is decided that any such non-standard
operations are likely to generate risks that are not
accounted for in the provisional containment
assigned in Section (iii), additional control
measures should be applied.
justification. For such cases it can be assumed
that the risks to the environment will be negligible.
50. If the virus is covered by any of the Animal
Health Orders (see Annex II) or if it may infect any
other animals (vertebrates or invertebrates), then
- For example, when working with a highly
the assessment should consider the risks posed
concentrated stock of a recombinant virus
to the environment. Attention should also be paid
and undertaking a particular work procedure,
to any viruses which are known to be pathogenic
that is likely to generate aerosols, particular
to wildlife (vertebrates and invertebrates) and, in
care may be necessary as compared to
particular, any endangered species which could be
undertaking the same procedure with low titre
affected; advice on endangered species may be
virus. Therefore, if such work is to be
obtained from the DETR Directorate of Rural
undertaken in a Class II cabinet particular
Affairs6.
care would have to be taken to ensure that it
provided an adequate level of operator
51. Any additional risks to the environment
protection. It may, therefore, be appropriate to
caused by the modification or the inserted
instigate a more rigorous testing regime than
sequences should be assessed by consideration
normal. In this situation it would be best
of the following points (the list is not exhaustive):
practice to test the cabinet using the KI
discus method on a six-monthly basis. In
- Survivability: is there reason to suspect that
addition, the laboratory may need to be
the modification carried out to the virus may
equipped with a general ventilation system
result in altered survivability in the
that is designed to provide an inward airflow
environment? Special attention should be
(negative pressure).
given to effects on UV tolerance, temperature
and resistance to desiccation. If the virus is
capable of long term survival in the
environment and there are indigenous species
with which it can recombine/ re-assort, then
further considerations will be the likelihood of
harmful sequences being transferred to
closely related viruses and the possibility that
the selective pressures could lead to the
RISK ASSESSMENT FOR ENVIRONMENTAL
emergence of mutant derivatives that are
PROTECTION
more harmful than the recombinant virus.
(v) The identification of any hazards to the
environment and the assignment of any
additional containment measures
48. There is a requirement under the Contained
Use Regulations 2000 to consider the risks to the
environment. The primary consideration here is
whether the virus is capable of infecting animals
(vertebrates and invertebrates). Note that this
guidance does not cover work involving GM plant
viruses; appropriate guidance can be found in Part
2C.
49. If the virus cannot infect any species other
than humans the risk assessment should include
a statement to this effect together with some
6
- Alteration of tissue tropism or host range:
is the modification likely to alter the tissue
tropism or host range of the recombinant
virus?
- Increase in infectivity or pathogenicity: is
the modification likely to increase the
infectivity or pathogenicity of the virus vector?
Is the modified virus likely to show altered
susceptibility to host defence mechanisms?
- Effects on other organisms: does the insert
code for a protein(s) with known or suspected
inhibitory, detrimental or other physiologically
active effects on other organisms?
Consideration should be given to possible
effects other than those being sought in the
construction.
- Environmental release: are all potential
routes of transmission or escapes to the
environment known? If so, will such routes
allow the modified virus and/or its products
Directorate of Rural Affairs 2X, DETR, Rm 902, Tollgate House, Houlton St., Bristol BS2 9DJ
Issued: March 2000
Part 2B
access to the organisms in which effects
may be manifested?
- Availability of control agents: will virus
susceptibility to control agents (where these
are available) be affected by genetic
modification? Can vaccination (in domestic
animals) or normal immune status (in any
animals) be expected to protect against the
modified virus?
52. Any hazards identified from these
considerations should be assessed using the
approach in Part 2A, including Annex IV of the
Compendium (including Table 6), that is, by
estimating the likelihood that identified hazards
will be realised and the consequences. The
assessment should be done on the basis that
those measures necessary to protect human
health that have been identified in sections (i) to
(iv) are already in place. Given the presence of
these containment measures a judgement should
be made as to whether there is any possibility of
accidental release (escape). If accidental release
is a possibility then the consequences must be
assessed. If the virus is to be used at high levels
of containment because of the risks to human
health and safety it is likely that the control
measures will also be sufficient to protect the
environment.
53. There may be cases where a virus is known
to have limited survivability in the environment or is
known not to infect UK hosts. In such cases, the
likelihood that a hazard will be realised in the
environment could be considered as "low" or
"effectively zero". In considering survival it is
important to determine the likely route of the virus
into the environment. In an aerosol, the probability
of survival may be poor, but the virus may survive
well in infected animal material. The ability of the
virus to infect hosts and replicate within them are
also important characteristics to take into
account. The assessment of risks should be
made with reference to Table 6, Annex IV in Part
2A, which should allow a final estimate of risk to
the environment to be made.
54. If the modified virus has been assigned to a
low level of containment on the basis of risks to
human health, and the final risk in terms of
environmental safety is not considered to be "low
or effectively zero", additional controls may need
to be adopted and the environmental risk
assessment repeated to ensure that all risks are
reduced to an acceptable level. The additional
measures should seek to reduce the likelihood of
7
environmental exposure. Particular attention will
need to be given to the possible routes of escape
including the disposal of infected material, in order
to minimise risks of accidental spread of virus
beyond the laboratory. For certain viruses the
possibility of airborne spread will need to be
considered, e.g. through ventilation systems or via
insect vectors.
55. The containment level for viruses pathogenic
to animals should be, as a minimum, that
specified by the appropriate Agriculture
Department for viral pathogens controlled by
Animal Health Orders. As well as notification or
consent required by the Contained Use
Regulations, work with such viruses may require a
licence from the appropriate Agriculture
Department (Annex II).
(vi) Assignment of the final activity Class (1, 2,
3 or 4)
56. To decide the final activity Class, users
should compare the control measures selected
during steps (iii) (iv) and (v) above with the
appropriate table of containment measures in
Schedule 8 to the Contained Use Regulations
2000. (These tables are also reproduced in Part
3A, 3C and 3D as appropriate.)
57. In many cases the required containment
measures will correspond to a single level of
containment. In such cases the containment level
will be a direct indication of the Class (eg an
activity requiring level 2 containment will be Class
2). Where the required containment measures are
a mixture from two different levels the activity
Class corresponds to the higher level. For
instance, the activity may require mainly level 2
measures, with the addition of one or two items
from level 3. The Class in this example would be
Class 3. Further explanation of the classification
system can be found in the Guide to the
Genetically Modified Organisms (Contained use)
Regulations 20007.
58. There is one further, and very important,
aspect about classification that applies to the
genetic modification of non-disabled animal
pathogens. Class 1 activities are described in the
Contained Use Regulations as being of “no or
negligible risk”. Therefore, because pathogens
cannot be said to be of no or negligible risk, such
work will always be Class 2 or higher. Since work
ISBN:0-7176-1768-0 , available as a priced publication from HSE Books, Tel: 01787 881165, Fax: 01787 313995
Issued: March 2000
Part 2B
with animal pathogens will almost invariably
require at least some of the measures required at
containment level 2 (eg the presence of an
autoclave in the building and the restriction of
access) it would not normally be possible to
assign a classification of Class 1. However, there
may be very rare cases where no level 2
containment measures are needed. Even when
that is the case assignment to Class 1 is
inappropriate.
59. Remember that classification into Class 2
does not necessarily mean that you will always
have to apply the full level 2 containment. If it can
be justified by the risk assessment you can ask
the competent authorities to the Contained Use
Regulations for agreement not to apply some of
the measures.
62. In some instances, especially when users are
working with animal pathogens, there may also be
additional requirements imposed by other
legislation such as the Specified Animal
Pathogens Order 1998, or the Importation of
Animal Pathogens Order 1980. Where there is any
discrepancy in requirements you must apply the
most stringent measures, irrespective of which
legislative regime they are derived from. Of course,
given that requirements in relation to working with
animal pathogens will often have formed part of the
risk assessment under the Contained Use
Regulations such discrepancies are unlikely. (See
Part 3, paragraphs 20 - 22 for further discussion.)
Risk Assessment Examples
63. Annex IV contains two example risk
assessments set out using the steps
60. The importance of the activity Class is twofold: recommended in this guidance. They are not
intended to exemplify “perfect” risk assessments.
Rather they are intended to be practical
- it determines the appropriate notification
illustrations of the procedure. For this reason
requirements. (See Part 1 for further
technical detail and justification have been kept to
guidance, as well as the Guide to the
a minimum. “Real” risk assessments will have to
Regulations cited above);
provide a suitable and sufficient level of detail,
including justification of statements made. Where
- it determines the minimum containment and
control measures which must be applied. For appropriate, references to the scientific literature
should be included.
Class 1 activities, level 1 containment must
be applied as a minimum, for Class 2, level 2
etc. The exception to this is when the user
has the agreement of the Competent
Authority to not apply the full containment
level corresponding to the Class. (See Part 3
for further guidance on such derogations.)
Two final aspects of classification
61. It is possible that when assigning control
measures the risk assessment will indicate some
measures which are not actually in the tables in
Schedule 8. In such cases the Class is
determined only by those items listed in the
tables. As with any activity, the Class identified
would indicate the minimum containment level you
would have to apply (unless you have agreement
from the authorities for a derogation). However, the
general requirement to reduce to the lowest level
that is reasonably practicable exposure of humans
and the environment to the GMMs and the
requirement to apply the principles of good
microbiological practice and good occupational
safety and hygiene would mean that the items
identified as necessary, but not in the tables,
would also have to be applied.
Issued: March 2000