59390 
Federal Register / Vol. 46. No. 233 / Friday. December 4. 1981 / Notices 
Cloning of animal viruses is 
considered here with respect to the 
question of whether recombinant DNA 
experiments per se pose any unique 
hazards not associated with handling 
animal viruses in general. The following 
I comments are offered in light of the 
consideration that all work with animal 
viruses poses some potential risk, 
unrelated to recombinant DNA. 
especially to investigators. It is essential 
to employ good laboratory techniques 
appropriate to the particular virus being 
studied. 
• Cloning of viral genomes in 
prokaryotes 
Our increased understanding of the 
differences in gene expression between 
prokaryotes and eukaryotes renders it 
difficult to imagine ways in which 
1 complete animal viral particles could be 
I synthesized in recombinant DNA 
containing bacteria. Animal viruses 
. have complex post-transcriptional 
modification requirements, and use host 
I functions in processing and assembly 
' which are not provided by a bacterial 
I host. 
I • Cloning of Viral Genomes in 
Eukaryotic Cells 
Because animal cells in culture are so 
fragile, they in themselves constitute an 
excellent containment system. The 
i major potential risk would be that a 
I recombinant DNA molecule could 
I replicate, be encapsidated as a normal 
j virus, and be able to propagate as a 
I virus. There are several potential 
j constraints on such a system, primarily 
I the fact that there is an upper limit on 
the size of the genome which may be 
encapsidated. Under such conditions, 
some of the original viral genomes 
would most probably be deleted, 
(rendering the recombinant virus 
I defective. However, such a defective 
virus might be propogated in the 
presence of a helper virus. Such systems 
are known to exist both naturally (avian 
'retroviruses and adeno-associated 
Iviruses) and in the laboratory (adeno- 
,SV40 hybrid viruses, most of which are 
I defective, but may be propogated) 
jRecombinants might include genes from 
jdifferent viruses or cellular genes 
jinserted into a viral genome. Examples 
jof both natural types of recombinants 
are generated by recombination in cell 
^culture (e g. the adeno-SV40 hybrid 
viruses and the highly oncogenic types 
of both avian and murine retroviruses — 
those which have incorporated a Sre 
|g(-ne of cellular origin). Interestingly, 
none of the known examples have 
proven particularly hazardous for 
^ ((experimenters to handle. Presumably a 
"*^hiiman retrovirus might be especially 
hazardous but this situation could 
obtain independently of any 
recombinant DNA experiments. 
Recombinants between viruses which 
are not closely related (different 
Families) are likely to have such basic 
differences in mechanisms of replication 
and infection as to be unlikely to work 
together well enough to produce viable 
entities; therefore the recombinant in 
this case will be equivalent to. or more 
defective, than the parent viruses. 
Recombinants between viruses which 
are more closely related may be more 
likely to create new useable functions, 
but the close relationship may suggest 
that such a recombination will have 
already occurred naturally. 
In summary recombinant DNA 
experiments involving animal viruses in 
eukaryotic cells might well lead to the 
construction of some viruses with 
properties which are novel in detail. 
However, there is no reason to think 
that these agents would be novel in 
terms of host range or potential 
virulence compared to viruses generated 
in other ways. 
Much of the summary of animal virus 
recombinant DNA risks is based upon 
the conclusions of virologists at the 
EMBO meeting in Ascot. England, 
January 27-28 1978 (Federal Register, 
July 28, 1978), and at the ASM meeting in 
Miami Beach, Florida. April 13, 1981. 
Reports on these meetings are available. 
B. Human Genetic Engineering 
The implications of the recombinant 
DNA technique for inserting 
recombinant DNA into humans has 
served as a backdrop for much of the 
recombinant DNA debate. The 
guidelines as now written deal with this 
issue only obliquely, by requiring 
special RAC consideration for most such 
experiments. Currently, human 
experimentation falls under the control 
of human experimentation review 
groups (Institutional Review Boards. 
IRBs). If a decision to remove the 
recombinant DNA oversight of this 
technique is made, it will still be 
oversighted by IRBs. It is not clear that 
recombinant D.NA perse poses a special 
problem in this area, distinct from those 
posed by non-recombinant 
transformation and genetic 
manipulation procedures. 
III. Costs 
Expenditures for risk assessment 
testing, administration of RAC. and the 
functioning of the Office of Recombinant 
DNA Activities were approximately 
S700.000 in the peak year of 1979. Risk 
assessment costs accounted for 
approximately one-half of that total. In 
all, risk assessment and vector 
development contracts supported by 
NIAID have cost over $2 million since 
1976. There are also expenses 
associated with the functioning of 
Institutional Biosafety Committees, and 
with the oversight of physical 
containment facilities. 
Costs in time and energy for 
individual scientists are difficult to 
estimate but continue to be significant. 
IV. Conclusions 
A. Summary Analysis of Risks 
Given the above analysis of the risks 
associated with recombinant DNA. we 
have come to the following major 
conclusions: 
(1) That accidental combinations of 
genes, rising out of "shotgun" cloning 
experiments or experiments where 
expression is not specifically 
engineered, are extremely unlikely to 
lead to serious problems. Both the 
barriers to expression of foreign genes 
in most organisms, the necessity for new 
activities to function as an integrated 
part of an existing pathway, and the 
selective disadvantage given to an 
organism by recombinant DNA inserts 
will interfere with such organisms 
establishing themselves in the 
environment and thus, ultimately with 
their potential to cause harm. Therefore, 
for these experiments, the minimal 
controls associated with good 
laboratory practice should be sufficient. 
(2) A particular subset of experiments 
may still pose some possibility of risk. 
While there is no evidence that this risk 
is qualitatively different from the risks 
associated with other kinds of genetic 
research, the possibility for improving 
the virulence, host range, or 
survivability of some pathogens does 
seem to exist. In most cases, in these 
experiments the problems of expression 
of foreign functions will have been 
bypassed, or normal functions will have 
been engineered to operate more 
efficiently. 
In many cases, even the best 
engineered strain will be at a major 
disadvantage in the environment, or will 
require artificial selections to maintain 
recombinant DNA information. The 
issue to be faced here, however, is (a) 
how serious is this risk? (b) What is the 
most effective, non-obtrusive 
mechanism for guarding against any 
untoward consequences of such work? 
B. Possible Responses 
Re-evaluation of the guidelines might 
in theory lead to changes in the 
guidelines' containment requirements, or 
the oversight procedures, or both. We 
will consider here the two extreme 
cases of (1) maintaining the current 
[279] 
