those used for work with common pathogens corresponding to the conditions of our level I. The four levels are described in detail in the code of practice (Appendix II) but may be briefly summarised as follows:

I. conditions conforming to the standard necessary for the maintenance of good microbiological practice, as required for work with common pathogens in a medical microbiology laboratory;

II. conditions as at I above in a laboratory which is additionally sited away from areas used by the general public and has controlled air flow and an exhaust-protective cabinet for aerosol-producing operations;

III. conditions as at II above but with access to the laboratory only through an airlock containing washing facilities and with an autoclave and provision for decontamination of all effluent from the laboratory: IV. conditions provided by a laboratory equivalent to the category A toxic laboratory referred to in Appendix III to the Report of the Working Party on the Laboratory Use of Dangerous Pathogens, ie those at III above with, additionally, provision for a full change of clothing and showering on leaving the laboratory and a double-ended autoclave.

2.9 Categorisation

There is a continuous spectrum of hazard but any categorisation must be done in steps. We assume that there are conditions of biological containment and nucleic acid purity that will allow an experiment to be moved from one category to another but these cannot be absolutely defined without reference to the individual experiment. We therefore give some examples to illustrate the principles we have adopted in the following table. It will be seen that, experiments involving recombination between the nucleic acids of non-pathogenic bacteria are considered to require only category I containment. Depending on the degree of biological containment achieved, and on the nature of the nucleic acid sequence inserted, certain other classes of experiment can be acceptably undertaken with Category I or II containment and some examples of such experiments are given in the following table. We stress that the table gives illustrative examples of some typical experiments and is intended to serve only as a guide and not as a definition of the whole area of possible work in each category. It will of course be for the GMAG to advise on specific cases and to establish a more complete categorisation as part of the "case law" as the work proceeds.

• For example, streptococci, staphylococci, salmonellae and other pathogens commonly present in the community and not especially liable to cause laboratory-acquired infection.

IV-10

Suggested categorisations for some typical experiments in Categories I, II, III

The term "purified" means fractions with little chance of including any unrecognised extraneous sequences (see paragraph 2.6.1). It is of course possible to have sequences selected because of their pathogenicity and these would raise the level of containment required.

Random Phage or plasmid/bacteria,

II

disabled

IV-11

3.1 We were asked to draft a central code of practice for laboratories undertaking experiments involving the techniques of genetic manipulation. We have done this on the basis of the four levels of physical containment (I-IV) summarised in paragraph 2.8 above: the code of practice is set out in full in Appendix II, together with a table illustrating the major differences between the four containment levels. The following headings are used in the code, as far as these are applicable to the particular level of containment being described: Laboratory (premises and facilities)

Biological Safety Officer

Staff-selection

training

supervision

protective clothing

health

discipline

Packaging and transport of samples

Security

Special requirements of experiments involving laboratory animals or plants. 3.2 In drawing up the code of practice, we have taken account of the code set out in the Report of the Working Party on the Laboratory Use of Dangerous Pathogens and in view of the similar requirements of our category IV containment level and the category A level for dangerous pathogens, we have tried as far as possible to follow the provisions and wording of that Working Party's code of practice. We also noted that a code of practice covering three containment levels is being prepared by Sir James Howie's Working Party on the Prevention of Infection in Clinical Laboratories.

3.3 As indicated in the introduction to the code of practice, we have not attempted to deal with all the technical queries that may arise on various aspects of the recommended containment levels: we have sought rather to provide guidelines for the GMAG to build on in drawing up the detailed specifications that may be necessary for assessment of individual laboratories.

IV-12

3.4 We recommend that every laboratory conducting genetic manipulation experiments must have both a properly constituted and representative local safety committee and a Biological Safety Officer answerable to the administrative head of the establishment or department, who must ensure that the Biological Safety Officer, on whose role we place particular emphasis, has the necessary training, experience and authority to enable him to carry out his duties. As we explain in Section 5 below, these are factors that we believe should be taken into account by the GMAG when it advises on whether or not a particular experiment should be conducted at a particular laboratory. We discuss training further in Section 4.

Experimental Animals

3.5 From time to time there will be a need to introduce live bacteria, viruses or phages bearing introduced genetic material into laboratory animals, initially to study whether genetic information is indeed transferred to animals in particular circumstances and to study the distribution, survival and replication of the vector in the animal. If this is done, the animals used must be kept in appropriate isolation facilities offering at least the level of containment appropriate to the experiment in question and desirably a higher level in view of the increased risk of dissemination when animals are involved.

3.6 Once the results of such experiments are available it may be possible to reduce the strictness of isolation for further experiments of the same type, if for example it has been shown that genetic information is unlikely to be transferred from the vectors. But it will need to be remembered that host-parasite relationships are rather specific and although there may be no undesirable sequels when a vector is put into one species, the outcome may be different if a different vector of species of animal is used, or if germ-free animals are used instead of conventional animals.

Plants

3.7 When suitable vectors become available, there will undoubtedly be proposals to introduce foreign nucleic acid into whole plants, especially if some of the exciting possibilities for genetic manipulation referred to in the Ashby Report (paragraph 3.4) are to be explored. Suitable measures of containment for the plants will be needed and we include some comments in Section 34 of Appendix II. These will vary with the nature of the vector and of the foreign nucleic acid etc, as in paragraphs 2.4-2.7 above.

3.8 In view of the availability of experience and expertise in plant pathology in the Agriculture Departments, and of the rather general nature of the measures required under the present licensing system for known plant pathogens, as summarised in Section 34 of the code of practice, which in any case apply only to imported organisms, we feel it would be illogical to attempt to make detailed recommendations for the containment of plants inoculated with vectors of recombinant nucleic acid that might involve a hazard to plant populations. We therefore recommend that, for the present, all experiments involving the introduction of recombinant nucleic acid into plants should require the prior approval of the GMAG and that the GMAG should, at an early stage, agree with the Agriculture Departments a procedure for specifying suitable containment measures for such experiments proposed to it. However, it should be borne in mind that by no means all such experiments will necessarily require measures as strict as those needed for the pathogens at present covered by licences and it may be possible to reduce the strictness of isolation procedures as more knowledge of the properties of the new genetic entities becomes available.

IV-13

4.1 The Ashby Report (paragraph 7.3) recommended that all those who work with these techniques should have training in the handling of pathogens. We were asked to make recommendations for the provision of necessary training facilities.

4.2 The first generation of workers in genetic manipulation will be largely self-taught, but we consider that appropriate training should be made available to all research workers, technicians and Biological Safety Officers intending to work in the field and that the head of each establishment or department, advised by the Biological Safety Officer, should ensure that anyone selected for the work has satisfactorily completed an appropriate course of training.

4.3 Biological Safety Officers will need:

i. to understand both the nature and implications of the genetic experiments being undertaken, and also the methods advised for physical containment;

ii. to judge the training requirements of the research workers and technicians and, as necessary, to recommend appropriate courses, or arrange to supplement a new entrant's experience with in-house training;

iii. to be familar with health monitoring procedures and to be able to collaborate with the supervisory medical officer in organising them; and

iv. to understand the legal and medical provisions relating to work in laboratories (for example, the Health and Safety at Work Act, relevant codes of practice and the organisation of local health services) and the procedures for consulting the Genetic Manipulation Advisory Group (GMAG).

Training for Biological Safety Officers will best be provided by special courses, preferably at, or in association with, a laboratory that has the necessary containment facilities and experience of containment techniques.

4.4 For research workers and technicians, we see a need for three types of training:

i. general education in the science of genetic manipulation, in the nature of the potential hazards and of the special precautions required to counter them. This should be required for all research workers without relevant experience and would include molecular biology and genetics, microbiology, epidemiology and relevant aspects of other disciplines. Such training might be provided at a university summer school;

ii. as a minimum, practical training in the procedures and manipulations appropriate to work in a genetic manipulation laboratory at the physical containment level of categories I and II above. This might be provided by industrial employers or at special university, polytechnic or other suitable courses; and

iii. special training for anyone intending to work in a genetic manipulation laboratory at the physical containment level of categories III or IV above. Such training should include practice in the use of such facilities and there is therefore a need for specimen physical containment laboratories which can be made available for training as well as for research.

IV-14