Scientific American, 6/97, R. Michael Blaese, page 91
Inserted genes could in theory arrest tumor growth or even AIDS
In 1997 an estimated 1.38 million Americans will be newly diagnosed with cancer Sadly, the main treatments currently available - surgery, radiation therapy and chemotherapy - cannot cure about half of them. This sobering fact has spurred serious efforts to develop additional strategies for treating the disease - ones based on the biology behind it. To that end, scientists are turning toward gene therapies, which involve introducing into the body genes that can potentially combat tumors.
Researchers initially explored gene therapies for remedying conditions caused by defective genetic instructions, or mutations, passed on from one generation to the next. Most cancers are not inherited in this way but instead result from acquired mutations, produced by external factors such as tobacco smoke or high doses of radiation-or just pure bad luck. These mutations ac cumulate in cells over time, ultimately rendering the cells unable to control their own growth-an inability that leads to cancer [see "What You Need to Know about Cancer," special issue of SCIENTIFIC AMERICAN; September 1996].
Gene therapies in general deliver instructions-in the form of DNA sequences - to diseased cells so that they will produce a therapeutic protein of some kind. This type of therapy is possible because viruses, bacteria, plants and people all share the same genetic code. Researchers have learned a great deal in little time about how certain genes govern the fundamental processes of life and how they contribute to disease. And because genes from one species can be read and understood by an other, experimenters can transfer genes between cells and species in their efforts to devise treatments.
For treating cancer, experimental gene therapies take varied forms: some involve imparting cancer cells with genes that give rise to toxic molecules. When these genes are expressed (that is, used by cells to make proteins), the resulting proteins then kill the cancer cells. Other designs aim to correct or compensate for acquired genetic mutations. Still others attempt to activate the processes by which such defects are normally re paired. And a host of ideas are coming from insights into how tumors evade recognition and destruction by the immune system, how they spread away from their sites of origin, how they gain a new blood supply and how they accomplish other feats that allow them to endure and spread.
Most of these approaches have yet to pass even the most preliminary clinical tests demonstrating their overall safety and efficacy, but these ideas may lead to better cancer treatments in the future.
Aside from promising actual treatments, gene therapy techniques have thus far helped physicians evaluate existing interventions. For example, in re cent years doctors have increasingly re lied on bone marrow transplants for treating cancers that fail to respond to traditional therapies. Frequently, the procedure is used for battling advanced stages of leukemia, a cancer that affects white blood cells, which are made by bone marrow. Before performing transplantation, oncologists take bone mar row from leukemia patients in remission. This apparently healthy bone marrow is stored away, and the patients are given superhigh doses of chemotherapy or radiation to kill off any residual cancer. Because these high doses destroy normal bone marrow, such an aggressive treatment would ordinarily kill the patient as well as the cancer In a transplantation, though, the patient is "rescued" by receiving a transfusion of his or her own saved bone marrow.
This method should, in theory, cure leukemia. But sometimes the disease re curs anyway. Clinicians have often wondered what goes wrong. Do the high dose radiation treatments sometimes fail to kill off all residual cancer cells, or are there sometimes undetected leukemia cells in the presumed disease-free marrow stored away during remission? To distinguish between these possibilities, scientists needed to devise a nontoxic and permanent tag so that they could mark the extracted bone marrow cells and find them later in the body. Malcolm K. Brenner of St. Jude Children's Research Hospital in Memphis, Tenn., did just that starting in late 1991, by inserting a unique sequence of bacterial DNA not found in humans into the patient's saved bone marrow Brenner knew that if he detected this bacterial DNA in the recovering patient's blood and bone marrow after transplantation, it would prove that the "rescue marrow" was restoring the blood system. In addition, detection of this tag in recurrent cancer cells would prove that the rescue mar row was a source of leukemic cells.
This is in fact what Brenner and others have found in some cases, forcing a critical reappraisal of the use of bone marrow transplantation. For instance, it is now recognized that for certain types of cancer, it may be necessary to give additional treatments to the mar row itself to rid it of any contaminating cancer cells before transplantation. To that end, marrow "gene-marking" studies, identical to those described above, are helping to find the best solution. These studies allow researchers to com pare sundry methods for purging the marrow of residual cancer. Different marker genes are used to tag marrow samples purged in different ways. As a result, physicians can determine how well marrow that has been purged in some way helps the patient following therapy. And, if the cancer recurs, they can also determine whether it has come from bone marrow that was purged in a particular way.
Gene Vaccinations
In terms of treatment, scientists have for more than three decades tried to find ways to sic the immune system on cancer - a tactic termed immunotherapy or vaccine therapy. And with good reason. Because immunity is a systemic reaction, it could potentially eliminate all cancer cells in a patient's body - even when they migrate away from the original tumor site or reappear after years of clinical remission. The problem with this strategy has been that the immune system does not always recognize cancer cells and single them out for attack. Indeed, many tumors manage to hide themselves from immune detection.
Recently, however, research in basic immunology has revealed means for unmasking such cancers. In particular, it now seems possible to tag cancer cells with certain genes that make them more visible to the immune system. And once awakened, the immune system can frequently detect even those cancer cells that have not been tagged.
The immune response involves many different cells and chemicals that work together to destroy in several ways invading microbes or damaged cells. In general, abnormal cells sport surface proteins, called antigens, that differ from those found on healthy cells. When the immune system is activated, cells called B lymphocytes produce molecules known as antibodies. These compounds patrol the body and bind to foreign antigens, thereby marking the antigen bearers for destruction by other components of the immune system. Other cells, called T lymphocytes, recognize foreign antigens as well; they destroy cells displaying specific antigens or rouse other killer T cells to do so. B and T cells communicate with one another by way of proteins they secrete, called cytokines. Other important accessory cells-antigen-presenting cells and dendritic cells-further help T and B lymphocytes detect and respond to antigens on cancerous or infected cells.
One gene therapy strategy being widely tested at the moment involves modifying a patient's cancer cells with genes encoding cytokines. First the patient's tumor cells are removed. Into these tumor cells, scientists insert genes for making cytokines, such as the T cell growth factor interleukin-2 (IL-2) or the dendritic cell activator called granulocyte-macrophage colony-stimulating factor (GM CSF). Next, these altered tumor cells are returned to the patient's skin or mus cle, where they secrete cytokines and thereby catch the immune system's attention. In theory, the altered cells should solicit vigorous immune cell activity at the site of the reinjected tumor. More over, the activated cells, now alerted to the cancer, could circulate through the body and attack other tumors.
In certain instances, these gene-modified tumor vaccines do seem to awaken the immune system to the presence of the cancer, and some striking clinical responses have been observed. All these clinical studies, however, are preliminary. In most cases, patient responses to these treatments have not been carefully compared with responses to conventional treatments alone. Also, the response patterns are not predictable, and they are not consistent from one tumor type to another or among patients who have the same type of cancer
Another problem with these studies is that nearly every person tested so far has had widely disseminated terminal cancer. Usually these patients have previously received intensive anticancer therapy, which has weakened their immune systems. Thus, even if gene vaccines did activate immunity in these individuals, the responses might not be easily notice able. Gene-modified tumor vaccines are most likely to prove beneficial in patients with minimal tumor burdens and robust immunity. Testing patients in this category, though, must wait until researchers are finished testing more seriously ill patient groups and have established the risks associated with the treatment. As this research so well illustrates, the development of new cancer therapies is a very complex and lengthy process.
A related gene therapy involves antigens that are found predominantly on cancer cells. During the past three to four years, scientists have made remarkable progress in identifying antigens produced by tumor cells. In addition, they have uncovered the genes that encode these tumor-associated antigens, particularly those on the most serious form of skin cancer, malignant melanoma. Now that at least some of these antigens have been described, it might be possible to develop a vaccine to prevent cancer, much like the vaccines for preventing tetanus or polio. The approach might also help treat existing tumors.
Preventive Immunizations
As with the cytokine vaccines, these antigen-based cancer vaccines re quire gene transfer. They work best when administered to cells that are readily accessed by the immune system. For example, Philip L. Felgner of Vical in San Diego and Jon A. Wolff of the University of Wisconsin and their colleagues observed that injecting a DNA fragment coding for a foreign antigen directly into muscle triggered a potent immune response to the antigen in mice [see "Nonviral Strategies for Gene Therapy," by Philip L. Felgner, on page 86]. The explanation for this reaction is simple: a bit of the foreign DNA enters the cells of the muscle or other nearby cells and directs them to produce a small amount of its protein product. Cells containing this newly synthesized foreign protein then present it to roving antibody-producing B cells and T cells. As a result, these sensitized immune components travel the body, prepared to attack tumor cells bearing the activating antigen.
The same basic strategy is revolutionizing the development of vaccines for preventing many infectious diseases. When these DNA immunizations are tested against cancer, the genes for newly identified tumor antigens are delivered directly into the body by way of vaccinia or adenovirus particles that have been rendered harmless or by such nonviral gene delivery systems as naked DNA. At present, the tests involve patients with widely spread cancer. It is clearly too late in these cases for DNA vaccines to prevent disease, but the studies should demonstrate whether the antigens can meet the essential requirement of eliciting a defensive response in the human body. Further, the studies offer a sense of whether DNA vaccines might have any merit for treating existing cancers. Given how sick many of these patients are, though, the results so far are difficult to interpret.
Yet another gene immunotherapy for cancer currently being tested in patients and in the laboratory involves antibodies. Thanks to highly variable regions on individual antibodies, these molecules are exquisitely specific. They can distinguish the slightest differences between foreign or mutated and very similar self antigens. As it turns out, specific anti body molecules exist naturally in the outer membranes of some cancer cells such as lymphomas that develop from B cells, which are committed to producing antibody molecules. Because a single lineage or clone of cells produces one specific antibody, all cancers of these cells will contain the same specific membrane molecule. This antibody then pro vides a unique molecular marker by which the cancer cells might be differentiated from similar but noncancerous antibody-producing cells.
Occasionally scientists have managed to produce antibodies to the antibodies found on cancer cell membranes. And some patients treated with these so called anti-idiotype antibodies have responded exceedingly well. Unfortunately, producing anti-idiotype antibodies is laborious. Thus, even though the approach can sometimes provide an effective treatment, it has seen only limited use. More recently, gene transfer techniques have offered other options. Be cause antibodies are gene products, scientists have been able to prepare anti-idiotype DNA vaccines that include the DNA encoding the critical cancer marker (the idiotype). This DNA sequence has then been linked with a gene encoding the cytokine GM-CSF. So far this double-whammy cancer vaccine has been tested only in laboratory animals, but it shows exciting promise.
Another double-whammy therapy in the works couples antibodies and T lymphocytes. Some rare patients have cancers that their T cells do recognize. But the T cells from these patients usually at tack only their own tumor cells or those from a small fraction of cases with the same type of cancer and tissue type. Also, people rarely produce antibodies to tumors. In contrast, mice immunized with human cancers do make antibodies that react strongly to those same human cancer cells. In some cases, the mouse antibodies bind to nearly all the tumor cells of one cancer in a test tube even if they have been taken from many different individuals with the same kind of cancer The mouse antibodies, though, are usually not effective in killing the cancer cells in humans. Even if the murine antibodies do have cancer-killing activity in a patient, the response is usually very short-lived because the patient soon produces inactivating antibodies against the mouse antibodies.
Therefore, oncologists have long hoped to find some way to combine the targeting ability of the murine antitumor antibodies with the killing ability of human T cells. Recombinant DNA technology offers the necessary tools. Researchers have successfully isolated the antitumor antibody genes from mouse cells and recombined parts of them with gene segments encoding the receptor that killer T cells use to recognize their targets. The modified receptor gene redirects killer T cells, which often do not recognize cancers, to see what the less discriminating mouse antibodies see. Indeed, killer T cells rearmed with chimeric T cell receptors kill cancer cells in a test tube quite efficiently. Early clinical experiments using this strategy are now under way in cancer patients, as well as in those infected with HIV, the AIDS-causing agent [see box on these two pages], and other pathogens.
Other Gene Therapies
Immunotherapy aside, cancers can be battled on other genetic fronts. There has been intense interest in identifying the precise DNA defects that cause cancer. Some mutations, scientists have learned, are associated with specific types of cancer. Other mutations occur in many varieties. Furthermore, there are different kinds of mutations. Some activate genes, called oncogenes, that drive uncontrolled growth in cells. Other mutations-those in so-called tumor suppressor genes-result in the loss of a nor mal brake on uncontrolled cell growth.
One of the most commonly mutated tumor suppressor genes in human cancer is p53, a gene whose protein product normally monitors the DNA in a cell as it divides. If the DNA is flawed, the p53 protein may halt cell division until the damage can be fixed or may induce cell suicide (apoptosis). When a normal copy of p53 is reintroduced to cancer cells in tissue culture, those cells return to a more regular growth pattern or self-destruct. Either outcome would be useful in cancer treatment, and so a great deal of effort has gone into developing methods for inserting normal p53 genes into cancers growing in the body.
There are still major roadblocks: as Theodore Friedmann notes in the first article of this section on page 80, cur rent technologies for delivering genes to specific organs or cell populations are inefficient. In addition, there are no perfected means for extending the effects of such locally delivered genes to other areas in the body. Until physicians can do so, these gene therapies will help tackle only tumors at isolated sites.
Even so, animals have shown significant improvements when the p53 gene is delivered either through the blood stream (in complexes with lipids that allow cells to take up the gene) or to tumors directly (using modified viruses to shuttle the gene into cells). An early clinical trial has reported some tumor regressions at local sites. In theory, though, there is one major limitation to using gene transfers to activate tumor suppressor genes or to neutralize oncogenes the corrective gene must be delivered to every tumor cell. Otherwise, the unaccessed cells will continue growing uncontrollably. It is impossible to correct the genes in every tumor cell-even those in a single site-using current technology. And although additional treatments might help correct more tumor cells, repeated gene transfers using modified viruses often are not feasible; the immune system frequently recognizes the virus the second time and destroys it before it can deliver genes to tumors.
Fortunately, though, the beneficial effects of an initial injection sometimes appear to reach cells that have not been gene corrected. Indeed, several different experimental gene therapies for cancer report the appearance of a "bystander effect." This phenomenon is invoked to explain why a treatment sometimes kills a higher proportion of tumor cells than can be accounted for by the number of cells actually expressing some new gene. Researchers have reported this kind of discrepancy in some p53 gene therapy trials but cannot yet explain it: presumably, if normal p53 genes did generate a bystander effect, cancer would not develop in the first place. But the bystander effect has been seriously studied in conjunction with other treatments, such as "suicide" gene therapy, in which a gene inserted into a cancer cell renders it supersensitive to some drug that ordinarily has no anticancer effect.
In the original application of suicide gene therapy, my colleagues Edward H. Oldfield, Zvi Ram and Ken Culver and I inserted the gene for an enzyme called thymidine kinase (tk) from a herpesvirus into cancerous brain cells of patients. In cells infected with the herpes simplex virus, this enzyme can convert the otherwise nontoxic drug ganciclovir into a metabolite, or by-product, that acts as a potent viral killer We found that this same toxic metabolite could kill dividing cancer cells; in some tumors, it killed neighboring cancer cells as well. To create this bystander effect, the toxic metabolite spread from the cell in which it was produced to its neighbors via gap junctions-channels that allow small compounds to move between cells. In the original clinical trial testing this treatment for brain tumors, about one quarter of the patients responded. And clinicians are testing other suicide gene therapies involving different anticancer compounds, some of which are also expected to produce a bystander effect.
In various early explorations of gene therapy technology, researchers are just beginning to learn about its potential and its limitations. As with so many other new and unexplored areas of science, some ideas will probably prove useful; many more will fall by the wayside. Ideas that are unworkable now may eventually become highly successful, when our technological capabilities in crease. Even though in the future cur rent methods for using genes for treatment will be looked back on as crude and inefficient, these methods have al ready offered important lessons. And they have indicated many new paths in the quest for cancer control.