gene in this case). If animal studies were successful, 
liver cells could he obtained by biopsy from patients 
with phenylketonuria, the normal phenylalanine 
hydroxylase gene could be inserted into the liver cells, 
and then the cells could be reimplanted under condi- 
tions (such as partial liver resection) that encourage 
multiplication of the implanted cells. For other dis- 
eases, skin fibroblasts could be used. 
Will the new gene get out of the target cells and 
spread to other cells and tissues? Recent evidence in- 
dicates that genes inserted into cells in the form of 
gene-containing plasmids (plasmids are bacterial 
cxtrachromosomal DNA molecules used as molecu- 
lar vehicles to carry genetic material in experiments 
on gene transfer) may exist not only as integrated 
genes but also as free extrachromosomal elements 
within the cell." It is not known whether or to what 
extent these gene-containing plasmids may move be- 
tween cells. Furthermore, it is not known whether the 
new gene will persist unchanged inside the cell. A 
plasmid now being commonly used for gene transfer 
(known as pBR322) appears to manifest recombina- 
tion and deletion events with itself (including any in- 
serted genes) when placed in mammalian cells, thus 
leading to replication of mutant genes within the host 
cell." 
These unsettling questions concerning the stability 
of new genes transferred into target cells can proba- 
bly be answered in experiments with animals. The 
presence of gene-containing plasmids or the genes 
themselves can be monitored in various animal tissues 
after administration of the gene to a specific target 
organ. However, such experiments will take time to 
complete, and they are just now being initiated. If the 
new genes are found in non-target tissues, their level 
of function must be assessed together with the result- 
ing effects, if any, on the non-target cell. 
The New Gene Should Be Regulated Appropriately 
Appropriate regulation (the ability of the gene to 
correct or ameliorate the genetic defect) docs not 
necessarily mean normal regulation. For example, it 
may take many years to understand normal regula- 
tion of the hemoglobin genes (i.e., the proper switch- 
ing on and off of the genes for the embryonic, fetal, 
and adult globin chains)." At present, although glo- 
bin genes have been inserted into animal cells, it has 
not been possible to make them function except at 
ineffectively low levels. Correction of the defect 
in /d-thalasscmia and alleviation of the defect in sick- 
le-cell anemia may be accomplished by linking a 
“naked” bcta-globin gene (i.e., one without any nor- 
mal regulatory sequences) to a known regulatory re- 
gion that will cause the beta-globin gene to express at 
the right level (i.e., to produce enough normal beta 
globin to equal the amount of alpha globin produced 
in d-ihalasscmia or to dilute the sickle hemoglobin 
in sickle-cell anemia). This latter approach is now 
under active investigation. However, to find the most 
effective regulatory region and to line tune it to en- 
sure the beta-globin gene will make the right amount 
of product will undoubtedly require considerable ex- 
perimentation in animal cells. Fortunately, there arc- 
strains of mice that have a disease analogous to thal- 
assemia. 1 '' 1 ’ Experimental studies with these animals 
should be helpful. 
On the other hand, enzymes such as thymidine ki- 
nase or hypoxanthine phosphoribosyl transferase 
(HPRT) that are required in every cell may be regu- 
lated efficiently in whatever cell they are located. The 
evidence appears to indicate that a viral thymidine ki- 
nase gene functions effectively when inserted into any 
of a number of different mammalian cells." The 
human HPRT gene might, therefore, be regulated 
appropriately if inserted into the cells of a patient with 
Lesch-Nyhan disease. 
The Presence of the New Gene Should Not Harm the Cell 
This requirement will be the most difficult to study 
since it will require determining what problems may 
develop years or generations after gene therapy. 
Treated animals (e g., mice) and their offspring 
should be studied carefully throughout their lifetimes 
to determine whether any pathologic process is detect- 
able. It is possible that the introduction of a new gene 
into a cell would interfere with the normal regulatory 
system of that cell and cause, for example, slow 
growth, organ-system dysfunction, reduced fertility, 
increased incidence of cancer, or early death. Fortu- 
nately, the short life span of mice should allow mean- 
ingful data to be obtained fairly rapidly. The extent of 
studies in larger animals and primates that should be 
completed before studies in human beings begin is 
difficult to estimate. The decision about how- much is 
adequate should be based on the realization that re- 
sults in larger animals will take a long lime to obtain, 
that human beings may react differently from other 
animals and, therefore, may respond better or have 
fewer side effects, and that we are attempting to treat 
tragic and lethal human diseases. It would be just as 
inappropriate to delay treatment of patients while we 
are awaiting long-term results in primates as it would 
be to rush into experimentation with patients before 
studies of small animals have been completed. 
What Can and Cannot Be Expected From 
Research on Gene Therapy in Human Beincs 
Many fear where research on gene therapy may be 
leading us. Despite the enormous amount of informa- 
tion obtained on DNA in recent years, our basic un- 
derstanding of cells and how they function is still ex- 
ceedingly elementary. At this point we can purify only 
genes whose products are known and can be isolated. 
We can transfer such a gene into animal cells, in 
which, in a few cases, it will function. But we are deal- 
ing only with single genes. We are currently attempt- 
ing to influence in a known way only single-gene 
defects. Intelligence, personality, fertility, organ struc- 
[316] 
Recombinant DNA Research, Volume 12 
