Genetics and Disease

"Let me begin with an assertion," says Paul Berg, professor of biochemistry at Stanford University and director of Stanford's Center for Molecular and Genetic Medicine, "that all human disease is genetic in origin, or, more accurately, that most diseases are the result of interactions between our genes and our environment." The cases for which this is most obvious are the diseases involving gross disturbances in the number or arrangement of a person's chromosomes. The classic example is Down syndrome, in which an extra copy of one of the smallest human chromosomes causes mental retardation, congenital heart defects, and increased susceptibility to infection. Other diseases of this type arise from additional copies of other chromosomes, missing chromosomes, or chromosomes in which the genetic message has become garbled through insertions, deletions, or other obvious rearrangements.

Another category of hereditary diseases is made up of those diseases associated with a defect in a single gene. These diseases are generally divided into two groups, dominant and recessive, depending on their mode of inheritance. The DNA in human cells is divided into 46 chromosomes organized into 23 pairs (Figure 1-1). Each pair consists of a copy of one of the 23 chromosomes in the father's sperm cell and a copy of one of the 23 chromosomes in the mother's egg cell. With the exception of the X and Y chromosomes in males, the two members of each pair are very similar. They contain virtually the same genes in virtually the same order. But the pairs of genes are not identical. Over 30 percent of the corresponding genes in a chromosome pair differ in some way, reflecting the overall genetic variability of the human population.

FIGURE 1-1

The DNA in human cells consists of 23 pairs of chromosomes. Twenty two of these pairs, which are numbered roughly in order of descending length, have very similar members, reflecting the equal genetic contributions of mother and father. In addition, each (more...)

A dominant genetic disease occurs when an individual receives a defective gene from either parent. In other words, the presence of a functioning gene on one member of a chromosome pair is not enough to overcome the effects of the defective gene on the other member. An example is familial hypercholesterolemia, in which an inability to clear cholesterol from the blood can cause children to suffer fatal heart attacks as early as 18 months. A person with a dominant genetic disease has a 50-50 chance of passing it on to a child.

Recessive genetic defects require that a person inherit defective copies of a gene from both parents. A person who has one defective and one functioning version of such a gene, known as a carrier, suffers little or no ill effects. But each child of two carriers has a one in four chance of inheriting two defective genes and suffering from the disease. Examples of recessive diseases are sickle cell anemia, cystic fibrosis, Duchenne muscular dystrophy, and Tay-Sachs disease.

Over 3,000 single-gene, or monogenic, diseases have been identified. Although individually rare, together they account for a great deal of human suffering. They affect more than 1 percent of liveborn infants and cause almost 10 percent of deaths among children.

The role of genetics is not as clear-cut in multigenic diseases, those involving the interactions of a number of genes with the environment. Examples of multigenic diseases include "hypertension, schizophrenia, manic depression, juvenile diabetes, heart diseases, rheumatoid arthritis, and a host of others," according to Berg. Biologists know that these diseases can be inherited, because they can cluster in families. But they do not yet know the number or identity of most of the genes involved in these diseases.

Cancer should also be classed as a genetic disease, Berg argues, even though it is not a hereditary disease. Cancer is caused by defects in the genetic signals that regulate cell growth and reproduction, and human families clearly have genetic predispositions to certain kinds of tumors. Also, genetic defects in the human immune system, in DNA repair systems, and in the ability to metabolize carcinogens are associated with a higher frequency of certain tumors.

Even infectious diseases, according to Berg, can be seen as genetic diseases. But in this case, the genes of the infecting organism and their relationship with the host and with the environment determine the course of the disease.

Tracking Disease to Its Source

Molecular biology has already demonstrated the amazing precision it can offer in analyzing genetic diseases. Research has shown, for instance, that sickle cell disease, which is characterized by anemia, impaired growth, and increased susceptibility to infection, is caused by a change in a single genetic base. This change leads to the substitution of one amino acid for another in hemoglobin, the protein that carries oxygen in the blood, causing the molecule to misfunction when oxygen is scarce. Another example is phenylketonuria (PKU), which can cause severe mental retardation if undiagnosed; it is caused by a defect in the gene coding for an enzyme that converts one amino acid into another. "Every few months another gene is isolated and the structural defect is identified," says Berg. "It's not too optimistic, I think, to predict that the genes and their defects for many of the monogenic diseases will be known within the next five years."

The impressive accomplishments of the past emphasize how much remains to be learned, however. Of the over 3,000 monogenic diseases now recognized, the responsible gene has been identified in only about 100 cases. The genes involved in multigenic diseases and disease susceptibilities remain largely unknown.

The ideal situation would be to know the gene or genes involved in every human disease, their location on the chromosomes, the nature of the defect associated with the disease, and the way in which the defect contributes to the disease. This information is known for very few diseases, and attaining it for the majority of diseases will take many decades. But biologists are starting to systematically pursue some of the early stages of such a program. They are constructing maps of the human chromosomes that give the locations of known genes. By mapping these locations, researchers can develop genetic probes to determine whether a person has a normal or a defective gene, leading to great advances in the diagnosis and treatment of disease. Genetic maps can also reveal patterns in the way genes are organized and regulated, leading to a greater understanding of how genes function in health and disease.

The scientific community has also been considering a much more ambitious proposal: a plan to determine the exact sequence of the billions of genetic bases making up the human genome. This information would be an invaluable resource for biologists. It would allow them to identify all of the genes within a given region of a chromosome, including many currently unknown genes. Genetic differences among individuals could be compared, revealing a great deal about the functioning of normal and defective genes. The mechanisms that control the expression of genes and the processes involved in development could be deciphered. The structure and organization of genes in different species could be contrasted, leading to insights into the evolutionary processes that resulted in those species.

The complete sequence would not answer every question in biology. It would not establish exactly how genes are controlled or how gene products function within a cell or organism. It would not completely explain how people are different or how humans have evolved. But it is unlikely that these questions can be answered without a deep understanding of the human genome.

Putting Genes on a Map

Geneticists were mapping genes to chromosomes well before they knew how chromosomes are constructed. The first genetic map was made in 1913, for five traits carried on the X chromosome of the fruit fly Drosophila melanogaster. Today, maps of human chromosomes are being made using the same principles.

The method used to make such maps involves analyzing how traits are passed down from generation to generation. In a parent's reproductive system, each chromosome pair separates during the formation of egg and sperm cells, reducing the original 46 chromosomes to 23. If this were all that occurred during reproduction, inheritance would be a straightforward matter; chromosomes would remain intact and be passed unchanged between generations. But chromosomes do not remain intact. Rather, before the formation of egg and sperm cells, each chromosome pair can exchange parts through a process known as crossing-over.

Imagine three different genes located on a pair of chromosomes (Figure 1-2). Each chromosome can have different versions of these genes, which interact according to the usual rules for dominant and recessive genes. For instance, gene B in Figure 1-2 may be for brown eyes-a dominant trait-while gene b is for blue eyes.

FIGURE 1-2

During the formation of egg and sperm cells, a process known as crossing-over can unlink versions of genes located on the same chromosome. In the above diagram, A and a are versions of a specific gene located on a chromosome pair, as are B and b and (more...)

If crossing-over and other forms of genetic recombination never occurred, the versions of genes located on a given chromosome would always be inherited together. In genetic terms, they would be permanently linked (top half of Figure 1-2). But crossing-over can reshuffle the genes on a chromosome pair, resulting in new genetic combinations.

The trick in genetic mapping is to observe how often this process separates the versions of genes on a chromosome. As shown in the bottom half of Figure 1-2, genes that are close together tend to stay together, whereas distant genes are easier to separate. By noting the frequency of separation, researchers can calculate the distance between genes and assign them relative locations on a chromosome.

Another way to map genes draws on the appearance of chromosomes under a microscope. If the chromosomes are stained during cell division with certain chemicals, alternating bands of light and dark regions appear, with up to several dozen bands on a single chromosome (Figure 1-3). These cytogenetic bands distinguish the chromosomes and provide broad landmarks along their length. But they are still quite large, with each band containing an average of 100 genes.

FIGURE 1-3

Several dozen disease-causing genes have been mapped to human chromosome 1. The gene responsible for Gaucher disease, type 1, a recessive disease characterized by an enlarged spleen, skin pigmentation, and bone lesions, has been mapped to a section of (more...)

Some people with genetic disorders have recognizable abnormalities in their cytogenetic banding patterns. For instance, the gene for Duchenne muscular dystrophy was mapped to a specific region of the X chromosome by noting that some sufferers of the disease were missing that portion of the chromosome. Translocations, in which part of one chromosome has broken away and become attached to another chromosome, and fragile sites that are susceptible to breakage have also pointed to the locations of specific genes.

If the base sequence of the gene being mapped is known, it can be located on a chromosome by making radioactive copies of part or all of the gene (by synthesizing or cloning the gene using radioactive constituents). These DNA probes can then be mixed with human chromosomes under chemical conditions that cause the DNA strands to temporarily separate. When a single-stranded probe encounters a matching single strand of DNA on a chromosome, the two combine or hybridize to produce double-stranded DNA. The radioactivity given off by the probe then functions as a marker to find the chromosome and the approximate location of the gene.

Another method of locating genes on chromosomes involves an intriguing technique known as somatic cell hybridization. When human cells and mouse tumor cells are grown together under the proper conditions, they tend to fuse and form hybrid cells. As these hybrid cells grow and divide, they lose most of their human chromosomes. But often one or a few human chromosomes will become stably established in a particular cell. The result is a mouse tumor cell line containing specific human chromosomes. By looking for human proteins produced by a given gene, that gene can be assigned to a chromosome carried in a cell line. Genes with known sequences can also be located using DNA hybridization.

All of these techniques can establish the relative or approximate location of a gene on a chromosome. But to map a gene to an exact chromosomal location requires the techniques of recombinant DNA.

A Genetic Scalpel

The ability of biologists to recombine DNA at will originated in the discovery of enzymes in bacteria that can cut DNA at specific sequences of four to ten genetic bases. These so-called restriction enzymes, which bacteria evolved to fight invasions of foreign DNA, were a marvelous gift to biologists. They allow researchers to slice DNA at specific locations, so that large DNA molecules can be cut into smaller, more manageable pieces. Then, using a class of enzymes known as ligases, biologists could splice together any two pieces of DNA formed with the same restriction enzymes. In this way, they could create mosaics of DNA, known as chimeras, that have never before existed in nature. (In Greek mythology, the chimera is an animal with the head of a lion, the body of a goat, and a serpent for a tail.)

Recombinant DNA has transformed the mapping of the human genome. It has enabled researchers to cut complex genomes into pieces, each of which can then be cloned many times over. These pieces can be organized into a library, creating a complete collection of all the DNA in an organism. Individual volumes in the library can be further analyzed to map specific genes or pieces of DNA. The logical conclusion of this process is the next step beyond mapping: the sequencing of genetic bases in part or all of an organism's genome.

One of the products of recombinant DNA technology has been a family of genetic markers much more precise and powerful than the banding patterns on chromosomes. Human beings are much more alike genetically than they are unalike. Only about one in every hundred base pairs of DNA are different between any two people. Many of these differences have no effect on the functioning of the genes, but others contribute to the crucial differences that make us unique: differences in physical appearance, aspects of personality, susceptibility to disease. Without these genetic differences, people would all be as alike as identical twins.

Using recombinant DNA, genetic differences among individuals can also serve as markers along human chromosomes. First DNA from a person's cells is cut into pieces using a specific restriction enzyme, and the resulting pieces are placed along the edge of a gel. Under normal conditions, DNA has a slightly negative electric charge, so when an electrical potential is applied across the gel the DNA is attracted to the positive charge on the other side of the gel. The smaller pieces of DNA move faster through the gel than do the larger ones. When the DNA pieces are spread across the gel, this process, known as electrophoresis, is stopped. The result is a virtually continuous series of bands, with each band corresponding to a DNA fragment of a different length. Researchers can then highlight specific bands using radioactive DNA probes that combine with given DNA sequences.

If two individuals have a difference, known as a polymorphism, in the region of DNA being tested, that difference can cause the restriction enzyme to produce DNA fragments of different lengths. For instance, say a person has a difference in a sequence recognized by a restriction enzyme, causing the sequence to remain uncut (Figure 1-4). If so, the DNA fragments from that person will have different lengths than fragments from a person without that polymorphism. The difference in the DNA responsible for the different fragment sizes is called a restriction fragment length polymorphism or RFLP (pronounced ''riflip").

FIGURE 1-4

A point mutation at a single genetic base can keep a restriction enzyme from cleaving a genetic location that is cleaved in another individual. This results in DNA fragments of different lengths, which can be separated and distinguished using electrophoresis. (more...)

The presence or absence of a particular RFLP in a person's DNA can act as a genetic marker on a chromosome. As with entire genes, it can be tracked from generation to generation and used in linkage studies. If a RFLP is located close to a defective gene, it will tend to be inherited with that gene, with the degree of linkage depending on the actual distance between the RFLP and the gene. If the identity of the gene is unknown-as is still the case for the great majority of genetic diseases-the RFLP can serve as a surrogate marker for the gene.

RFLPs and Disease Diagnosis

The implications for genetics of RFLPs are "extraordinary," according to Berg. Geneticists have already located hundreds of RFLPs scattered throughout the human genome. At first, a RFLP may be used as a flag to indicate the presence of a disease-causing gene closely linked to the RFLP. Later, RFLPs may be used to track down the actual location of that gene, so that the nature of its defect can be determined.

The use of RFLPs and other genetic probes to diagnose genetic diseases "will change medicine in a very profound way," according to Hood. For instance, it will eventually be possible to detect the genes responsible for virtually every monogenic disorder. These genes could be detected after birth or prenatally, giving parents an option to terminate a pregnancy. Carriers of defective genes could also be identified, allowing them to decide whether to have children and risk passing on the disease.

The identification of carriers and prenatal testing could greatly reduce the toll that many monogenic diseases take on the human population, Berg points out. An example is Tay-Sachs disease, a recessive disease occurring most frequently among Jews of European descent that causes retardation, paralysis, and early death. A test for Tay-Sachs disease that can identify carriers and affected fetuses has been available for a number of years. This test "has virtually eliminated Tay-Sachs disease from the Jewish population," Berg says.

As scientists learn more about the role of genes in multigenic diseases, it will increasingly be possible to determine an individual's susceptibility to such diseases. In time, diagnostic tests should be available to determine a person's susceptibility to such common diseases as cancer, heart disease, and diabetes.

Such tests could provide "a new opportunity for preventive medicine," says Hood. Now, genetic testing is confined largely to fetuses and expecting parents. But in the future it will be possible to diagnose the disease susceptibilities of anyone and enter that information into a medical record. That information could help a person to avoid certain habits, diets, or environmental conditions that might lead to the disease. New therapeutics might also be developed that could lower the chance of contracting a disease when a susceptibility exists. "We can start to think about changing some fundamental aspects of human health care," says Hood.

As an interesting sidelight, RFLPs and other DNA probes have also assumed a growing role in forensic medicine. The patterns in a person's DNA are just as distinctive as a fingerprint, and in some crimes DNA samples are easier to find than fingerprints. Already, genetic tests have been used to identify rape suspects, murder suspects, and the parents of children applying for immigration. "Very soon we will be in a position to translate genetic patterns into electronic databases and compare particular DNA fingerprints with any others that come up," says Hood. "This raises challenging social and ethical questions."

The Price of Knowledge

The social and ethical questions posed by genetic analysis extend far beyond DNA fingerprinting. As testing for diseases and disease susceptibilities increases, it will be possible for people to learn a great deal about themselves and their futures. If it were possible for a person to know what he or she was most likely to die of, would that person want to know? Today, genetic testing is often done as the prelude to some therapeutic intervention, from the abortion of a severely affected fetus to dietary or pharmacological interventions. But treatment may not be possible for some of the diseases that will be detectable in the future, at least until researchers apply new understandings of genetic information to the development of new therapies.

A stark example of the dilemma posed by the gap between diagnosis and treatment is Huntington's disease, which threatens about 125,000 people in the United States. The disease is caused by a dominant gene inherited from either the mother or the father; therefore, a person with an affected parent has a 50-50 chance of also having the disease. The symptoms of the disease, beginning with loss of control over movement and proceeding to dementia and eventual death, generally do not appear until after the childbearing years. As with many monogenic diseases, there is no treatment for Huntington's disease.

Through extensive analyses of RFLPs from families afflicted by Huntington's disease, geneticists have narrowed the location of the Huntington's gene to a million-base-pair region at the end of chromosome 4. By studying the RFLPs of individual families, researchers can usually find markers that signal the presence of the defective gene with a high degree of accuracy. The question then becomes, Will a person at risk of Huntington's disease want to know whether he or she carries the gene for the disease? If a person is trying to decide whether to have a child and risk passing the gene on to the next generation, the information is clearly useful. But if a person is past the childbearing years or does not want to have children, why undergo a genetic test that has the potential to predict future illness and early death?

As genetic testing becomes more powerful, similar dilemmas will emerge. The ability to test for the presence of a gene is an inevitable step in the process of understanding a disease and developing ways to treat it. But some diseases will always remain untreatable, and for others the ability to diagnose a disease is imminent whereas the treatment for it is much less certain.

Information from genetic testing will also raise a number of contentious social questions. Should employers or insurance companies have access to such information, even though in some cases it could lead to loss of a job or loss of insurance coverage? What kinds of traits or conditions should prospective parents be able to test for in deciding whether to have children or abort a pregnancy? Guidelines will have to be established for the use of such information to preserve individual confidentiality and autonomy.

Increased genetic testing will also require that people understand the information they are being given. An increased susceptibility to a disease does not mean that a person will inevitably get that disease. In fact, the chance of getting the disease may be quite small. Yet such a diagnosis could greatly increase a person's anxiety and affect important decisions.

Even in cases in which a diagnosis is more certain, the potential for misinterpretation will be large. People who have a disease caused by a single gene can show enormous variations in the manifestations and severity of the disease. This variation is so great, says Berg, "that some physicians who see these patients cannot accept that they're all caused by the same genetic defect." In part, says Berg, these differences are caused by a person's broad genetic background, which probably will not enter into the diagnosis. They are also caused by the environmental influences a person experiences, which can be endlessly variable. People will therefore have to understand the full range of outcomes that could accompany a specific diagnosis.

These are not new issues in human genetics. Similar situations have already been addressed by researchers, clinicians, patients, and policymakers. But the issues will become more numerous and widespread as biologists learn more about the connection between a person's genetic heritage and disease.

From Mapping to Sequencing

From a purely scientific standpoint, biologists generally agree that a program to map the human genome through the construction of a large library of RFLPs and other genetic markers is a worthwhile goal. But the new techniques of molecular biology allow more to be done. Within a few years it should be possible to begin systematically sequencing the 3 billion base pairs in the human genome. It would be the largest project ever undertaken in biology, with characteristics much different than those of normal biological research. There is much less consensus among biologists about how to undertake such a program.

One thing is clear, however: it will be a formidable task. Apart from the technical difficulties involved in sequencing, the project will produce a tremendous amount of information to store, analyze, and disseminate. There are about 3 billion bases that would need to be sequenced in the human genome (technically, only one member of each chromosome pair needs to be sequenced, since the members of a pair are so similar) If every letter in this book represented one of these bases, it would take roughly 10,000 of these books to represent the entire sequence. Furthermore, sequencing programs will have to acquire sequences from different individuals and from other species to make the best use of the human sequence. So far, only about 2 million base pairs of the human genome have been sequenced and stored in a central data base, less than 0.1 percent of the total. Without a special effort to sequence the human genome, it will not be done for many years, if ever.

There are several broad objections to a sequencing program that must be answered for the project to proceed, according to Hood. The first asks whether obtaining the full sequence would in fact be scientifically uninteresting and therefore a misallocation of resources. Hood agrees that much of the mapping and sequencing would be routine, repetitive work. Ideally, he says, much of it can be automated (see box, pages 20-21). It is the sequence itself that will generate "the staggering and exciting kinds of scientific endeavors." Also, Hood questions whether the United States can afford not to undertake a sequencing program. "I would argue that the impetus provided by this program is going to have a major impact on whether the United States stays competitive in the industrial biotechnology area."

Box

Building a Better Sequencer. The standard laboratory method used to sequence DNA involves labeling fragments of the unknown DNA with a radioactive marker and passing the fragments through four different electrophoretic gels, one for each kind of base. (more...)

A second objection is that such a project will draw funds from other areas of biology, and Hood finds this to be "a much more serious objection." But the technologies and information developed through a sequencing program will find applications throughout biology, Hood points out. He also believes that the project will be "sufficiently compelling" to generate new sources of funding for itself. Several committees have studied sequencing proposals and have concluded that funding in the neighborhood of $200 million per year would be appropriate for such a program. While this is a substantial amount of money, it is only about 3 percent of the total amount of funds spent by the federal government on biological research each year.

A third objection is that the project will be "big science" and therefore at odds with the traditional approach to biological research, in which most work is done by small, independent groups of scientists. But Hood contends that "it is not big science in the same sense that projects such as the superconducting supercollider and the space shuttle are big science." The costs of the instruments required are modest, and the instruments can be widely disseminated. Research proposals for aspects of mapping and sequencing will be peer reviewed, and this research will also be widely distributed. Furthermore, once the information is available, it will greatly increase the power and range of what small groups can do, and they will no longer need to spend great amounts of time on routine mapping and sequencing.

The final objection is that the technology is not yet adequate to do the job, and this is the one objection that Hood finds convincing. As currently performed, sequencing is tedious, time-consuming, labor-intensive, and expensive, according to Hood. "Technology development is what we really should concentrate on," he says. "This is an area of major deficiency at this time." Sequencing the human genome will involve building up a library of overlapping pieces of the genome, which can be stored in a central location, easily reproduced, and sent to investigators to be sequenced and studied. This is now relatively easy to do with small pieces of DNA. They can be cut with restriction enzymes, inserted into pieces of bacterial or viral DNA known as vectors, and maintained in culture as clones. The problem is one of scale. Conventional vectors can hold a maximum of about 40,000 bases, meaning that with an average overlap of 10,000 bases, it would take over 100,000 different clones to store the entire human genome. Maintaining a culture library of this size would require an administrative structure much larger than any in existence today. Furthermore, some parts of the human genome are difficult to clone, and there are concerns about the stability of DNA in a culture library.

Progress is being made on many of these problems. Researchers are developing new vectors in yeast that may be able to carry 500,000 to 1 million DNA bases, over ten times the size of the fragments that can be maintained with current systems. Instead of needing over 100,000 clones to encompass the entire human genome, it could be done with several thousand. Restriction enzymes have recently been discovered that cut DNA into very large pieces, and a new method of electrophoresis using pulsed electric fields can separate DNA fragments 200 times larger than the maximum possible with conventional electrophoresis. These are the kinds of technological developments that will be necessary to make sequencing the human genome economically and scientifically feasible. (The box on pages 24-25 discusses several other technologies that will be essential to sequencing efforts.)

Box

Protein Sequencers and Synthesizers. Recombinant DNA technology is not the only driving force behind the rapid advances now occurring in genetics. Another powerful influence has been the development of microchemical instruments that can analyze and synthesize (more...)

Hood advocates that an effort to sequence the human genome proceed in stages. During the first stage, new technologies would be developed to increase the efficiency of DNA sequencing five-to ten-fold. At the same time, detailed mapping of the human genome could be under way, which would provide a framework for the sequencing effort. A phased approach would also allow systems to be developed to collect and disseminate DNA clones and to store and analyze the huge amounts of data that will be generated. Once this technological infrastructure is in place, Hood says, the complete sequencing of the genome could begin.

Prospects for Human Gene Therapy

The knowledge provided by mapping and sequencing the human genome may make it possible to achieve one of the most provocative of the new biotechnologies: human gene therapy. Interest in human gene therapy arises from a depressing fact: for the majority of monogenic diseases, no effective therapies are available. For some single-gene defects, dietary restrictions, the use of drugs or biologic agents, or transplantation of tissues or organs may alleviate part or all of the symptoms. But for many such genetic defects there is no alternative to "debilitating and progressive disease leading to suffering and early death," according to Berg.

Any disease to be treated with human gene therapy must meet a number of "very formidable" criteria, Berg says. The gene responsible for the disease and the molecular nature of its defect must be known. The disease must involve cell types that are accessible and well-characterized. The disease cannot begin to exert its harmful effects until after birth, since a variety of techoical and ethical constraints prohibit gene therapy before birth. Also, the defect must be limited to a single gene. "We certainly enter the realm of wishful thinking when the therapy aims to modify more than one gene or to rectify the defects resulting from chromosomal abnormalities," notes Berg.

The gene therapy being considered for humans involves placing normal genes into somatic, or body, cells, not into germ line, or sex, cells. In this way, "the therapy stays with the treated patient and is not transmitted to offspring," Berg says, "so any undesirable effects of the integration are not going to be propagated to future generations." Given the criteria that candidate diseases must meet, only a handful have received serious consideration. The most attention has focused on a rare disease caused by defects in the enzymes necessary for normal immune system development. Children with the disease, known as severe combined immunodeficiency disease, must live in totally sterile "bubbles," since their immune systems cannot protect them from common viruses, bacteria, and fungi.

Researchers have developed an experimental procedure in mice that could be applied to humans if it proves effective and safe (Figure 1-5). First, bone marrow cells are removed from the animals and mixed with an infectious agent known as a retrovirus. When retroviruses infect a cell, they insert a copy of their genetic information into the genome of the host. By genetically engineering the retroviruses to contain a normal version of the defective gene, scientists can insert the normal gene into the bone marrow cells. The cells are then reimplanted into the mice where, presumably, they will produce the missing gene product.

FIGURE 1-5

Much of the research on human gene therapy has focused on retroviruses, infectious agents that can insert their own genetic material into the DNA of cells they infect. The genome of a retrovirus consists of RNA, which is enzymatically copied into DNA (more...)

Unfortunately, says Berg, the results to date "have been rather discouraging and may prompt others to begin looking at other potential vectors.'' The genes can be inserted into mouse bone marrow cells and made to produce the gene product, but only for a limited time. "For reasons that are totally mystifying at the moment, the gene that has been introduced by the virus is turned off over some varying period of time. Although the animal has acquired the foreign DNA, it no longer expresses the gene of interest." Much effort has gone into trying to maintain the expression of the introduced gene, but without success. "My feeling is that this probably reflects our lack of knowledge about the development of the cell," says Berg.

Human gene therapy has other potential problems that are only beginning to be addressed, notes Berg. One is that the retrovirus inserts its genetic message into the cell's DNA at random and uncontrollable locations. It is possible that the foreign DNA would integrate into the middle of a gene essential for the survival of the cell, which would destroy it. Or the gene could insert itself in such a way that it increases the activity of a gene regulating cellular growth, leading to tumors.

"The inability to control the site at which the vector DNA integrates is one of the major handicaps at the moment in this approach," Berg points out. As a result, he and his coworkers are examining ways to target the introduced DNA to specific locations. Perhaps vectors can be developed that would recognize sequences within the cell's genome and integrate its DNA in a predictable way around that sequence. Even more desirable would be some sort of agent that homes in on the defective gene and somehow repairs it in place. But knowing whether any of these approaches are feasible will require a much more sophisticated understanding of the workings of the human genome.

ESSAY-Science and Scientists from the Public's Perspective

"The end of man is knowledge, but there is one thing he can't know. He can't know whether knowledge will save him or kill him. He will be killed, all right, but he can't know whether he is killed because of the knowledge which he has got or because of the knowledge which he hasn't got and which if he had it, would save him."

—Robert Penn Warren

All the King's Men

When scientists confront the dilemma posed by Robert Penn Warren, they tend to respond optimistically, says Maxine Singer, president of the Carnegie Institution in Washington, D.C. "We believe that the acquisition of new knowledge is both wondrous and good. We also believe that the quest for better comprehension is a fundamental human trait, one which sets Homo sapiens apart from other living species. We reinforce this belief by a conviction that original research is a creative endeavor, linked to artistic and literary creativity in method and talent. And we believe that these last considerations say that science is a thoroughly human enterprise, that through science we are expressing human as well as humane values."

But the public's perception of science can be ambivalent and often pessimistic, Singer acknowledges. "For the most part, nonscientists find the continuing quest for knowledge somewhat frightening," she says. To some extent, this fear extends even to scientists. "Often at Washington cocktail parties, where everyone asks strangers, 'What do you do?' I find that the answer 'molecular biologist' is likely to drive the questioner to the far end of the room." The same attitude surfaces in the popular media. In television shows and best-selling books, scientists are often depicted as "frightening, usually unsympathetic, almost inhuman."

Despite these widespread impressions, Singer does not believe that the public's view of science is totally negative. Public attitudes are too numerous, diverse, and at times contradictory to characterize one-sidedly. People are eager to make use of the fruits of scientific research, and many avidly follow reports of scientific developments. Biologists and other scientists have also devoted considerable time in recent years to explaining their work to the public. Partly as a result, Singer notes, over 50 percent of all Americans say that they know what DNA is.

But the tension between the acquisition of new knowledge and the fear of that knowledge remains widespread in society. It is a troubling tension to scientists, Singer believes, because scientists in the United States rely on the public for support. On a purely financial level, the majority of scientific research is paid for with public funds. And more broadly, in a democracy, scientific work on controversial subjects can be slowed or halted by public opposition, even if engendered by unwarranted fears.