Igor Matushansky & Robert G Maki. American Scientist. Volume 93, Issue 5. Sep/Oct 2005.
A 50-year-old Finnish woman was having mild stomach pains when she went to see her doctor in 1996. The physician in Helsinki found a large abdominal mass, and things got worse from there. Further exams discovered tumors, 7 and 10 centimeters in diameter, on her stomach, plus many small nodules of spreading cancer. Surgeons removed as much of it as they could find, but the diagnosis was grim. It was a gastrointestinal stromal tumor, or GIST, a cancer of the connective tissue in the gut that was inevitably fatal if surgery failed.
Two years later the cancer was back, and doctors had to operate again to remove growths on the liver and abdominal wall. Another surgery that year excised more tumors on the liver and ovary. The woman’s doctors tried to slow the proliferating cells with an intense barrage of combined chemotherapeutics-seven cycles using four different drugs over a five-month period-without success. As the cancer spread, it blocked the patient’s intestine, requiring yet another operation. When the surgeon went in to cut away the blockage, he found and removed 45 additional tumors. The patient began taking large daily doses of two cutting-edge, immune-system-enhancing drugs, to little effect.
Having exhausted other options, the woman’s oncologist, Heikki Joensuu at the Helsinki University Central Hospital, suggested an experimental drug, STI571, which had just begun phase I testing for chronic myelogenous leukemia-a completely different kind of cancer from the soft-tissue tumors his patient carried. It was a desperate attempt to save the patient’s life, so despite the lack of any clinical supporting data, the hospital agreed to let him try.
Two weeks later, an MRI exam showed the woman’s tumors were 40 percent smaller. Two months later, they had shrunk half as much again. At eight months, they were further reduced in size; about a quarter were no longer detectable. What’s more, the tumor cells that remained had stopped dividing and no longer showed the molecular signature of cancer. It was an incredible improvement.
Why did Joensuu think to try this particular drug? Because he knew, based on the work of others, the molecular basis of GIST: The abnormal protein that caused his patient’s tumor was similar to the one that caused the leukemia for which the drug was approved. Furthermore, some reports indicated that STI571 could work on both types of proteins-at least in a dish of cultured cells. In the end, this success owed much to many. The paper describing this striking case included as coauthors doctors and scientists from Helsinki, Turku University (also in Finland), Massachusetts Institute of Technology, Harvard, the Oregon Health Sciences University and the pharmaceutical company Novartis, which made the compound (now called imatinib and sold under the name Gleevec).
Is this drug the long-sought silver bullet, the cure for all types of cancer? No. But it does illustrate how discoveries in research labs can quickly pay off in clinics. It is an early fruit from what promises to be a great harvest of medical advances made possible by two decades of accelerating progress in understanding how cells work. And it can’t come soon enough.
Divide and Conquer
In the United States in 2005, almost two and a half million people will be diagnosed with some form of cancer, and about 570,000 people are expected to die of this disease. Indeed, cancer passed heart disease this year as the top killer of people under age 85. Although these survival statistics would be even grimmer without modern medicine, the best efforts remain inadequate.
The three pillars of treatment for cancer are surgery, radiation and chemotherapy. Of these three, chemotherapy is the least discriminating. Whereas surgical excision and radiation therapy are site-specific, standard chemotherapy kills dividing cells everywhere in the body. The rationale for administering such poison is that the cells that split frequently-usually cancer cells-should suffer the most. Although the strategy works well in some instances, say for treating leukemias, other types of cancer do not grow rapidly and are resistant to chemotherapy. Furthermore, normal cells that divide often-those in hair follicles and the lining of the gut, for example-are destroyed too, causing hair loss and diarrhea. Additional medications can alleviate some of these side effects but do not solve the fundamental problem, which is a lack of specificity.
Although new therapies based on advances in molecular biology have begun to enter clinical practice, the “holy grail” of oncology-a treatment that is both effective and completely specific to cancer cells-remains unknown. In fact, a single cure now seems more elusive than ever as physicians continue to learn about the many physiological changes that distinguish different forms of cancer from one another and from normal cells.
This heterogeneity among cancers is reflected in a relatively rare, highly diverse class of tumors called sarcomas. Thus, sarcomas are good (and popular) subjects for the study of therapies that physicians can use on other types of malignancies: More than a dozen medical centers and hospitals around the country specialize in sarcoma research and treatment.
The word … appears in the writings of the physician-philosopher Galen, who lived during the latter part of the second century A.D. In its original Greek, the term sarcoma describes a fleshly growth. Doctors use the word today to describe cancers derived from connective tissues such as bone, muscle, fat or cartilage. Each year in the united States, clinicians diagnose approximately 10,000 new sarcoma cases, encompassing 50 different types or cancer, each with its own distinctive biology.
Because relatively few people are afflicted with each sarcoma subtype, the disease doesn’t lend itself to studies that require many patients, such as large-scale searches for susceptibility genes or big, randomized trials. Instead, most investigators look at pathological mechanisms using smallscale studies that are carried out on the cellular level. Consequently, scientists now know more about how sarcomas work than their scarcity might suggest, and clinical trials for sarcoma treatments are likely to reflect specific molecular data from experiments in a laboratory.
Our own research at Memorial Sloan-Kettering Cancer Center in New York strives to understand how sarcomas arise and to exploit that knowledge in the development of new treatments. Our work and the work of our fellow physician-scientists has led to several therapies for individual sarcomas, some of which are also proving useful in the fight against common types of tumors, such as those found in some lung cancers. This review highlights some of these advances.
Going After the Genes
Like other cancers, sarcomas are products of genetic mutations, which can take many forms. One particular category of genetic errors, called chromosomal translocations, is responsible for several sarcomas.
A chromosome is a single long strand of DNA-thousands of times longer than a cell is wide. When a human cell is preparing to divide, it copies each of its 23 pairs of chromosomes so that each daughter cell can receive a complete set. Occasionally, a strand of DNA will break during this process. The cell usually mends these fractures correctly, or, if it cannot, trips the self-destruct switch (leading to programmed cell death, or apoptosis). But sometimes a cell will incorrectly join two or more different chromosomes, yielding a translocation. If the cell subsequently escapes its own destruction, daughter cells can inherit too many or too few copies of that piece of chromosome. Furthermore, if the DNA is snapped and incorrectly repaired in a region that specifies a protein, that valuable piece of the genetic code-that gene-may be destroyed, leaving the cell with only one remaining copy on the unbroken partner in the chromosome pair. Another possibility is that the improperly repaired DNA will encode a “fusion protein,” made from the sequences of two different genes spliced together. Many such fusion proteins are merely ineffective, like a bicycle with oars instead of pedals, but some can be dangerous. If the original performed some critical function in the cell, such as regulating cell division, the fusion protein can cause big problems.
These breaking-and-joining events do not happen randomly, and certain translocations cause specific kinds of cancer. One example is the abnormal inheritance of an extra copy of the long arm of chromosome 12, which causes a version of the most common soft-tissue cancer, liposarcoma-the class of sarcoma that develops from fat cells. Thanks to recent advances in “gene profiling” (a technique that measures gene activity), scientists have identified a cause of one subset of this class, the so-called dedifferentiated liposarcomas.
With two normal copies of chromosome 12 plus the extra fragment, cells manufacture too much of the protein encoded by one of the resident genes: the cyclin-dependent kinase 4 gene, a mouthful that usually goes by the shorthand CDK4. As its name indicates, the CDK4 protein is a kinase-an enzyme that adds phosphate groups onto other proteins as a means of controlling how active or inactive they are. It so happens that this particular kinase acts on one of the master switches of cell division, the stoutly named retinoblastoma tumor suppressor, or RB, which acts through a DNA-binding protein called E2E A glut of CDK4 causes RB to have an excessive number of phosphate groups attached, thereby jamming the cell-division switch in the on position-a hallmark of cancer.
Once scientists understood this chain of events, they hypothesized that blocking CDK4 might slow the spread of liposarcoma. One candidate drug is flavopiridol, which inhibits several kinases, including CDK4. Our colleagues Gary K. Schwartz and Samuel Singer at Memorial Sloan-Kettering have shown that this drug destroys liposarcomas in the culture dish and in mice that carry human liposarcomas. Several clinical trials are now testing flavopiridol for various cancers.
Sarcoma Antigens 101
Some tumors produce characteristic proteins. For example, melanomas churn out the pigment melanin and related molecules. Thus, training the patient’s immune system to attack such proteins-by using a vaccine, for example-can help the body identify the malignant cells and get rid of the cancer. This process of teaching the immune system to create antibodies against cancer-specific (or in the case of melanin, cancer-enriched) proteins is called immunotherapy.
Sarcomas are excellent candidates for irnmunotherapy because many have chromosomal translocations that generate fusion proteins not seen in any other cell of the body. The immune system can thus attack these specific molecules (and the cancer cells that contain them) without harming healthy tissue. Of course, the body does not automatically recognize the mutant proteins as foreign. If it did, the tumor never would have appeared in the first place. Part of the problem is access: The abnormal proteins may not show up on the surface of the cell at all; they may be present only within, where they are hidden from immunesystem surveillance. Fortunately, in some instances, fusion proteins do appear on the cell membrane, albeit in quantities too small to stimulate the immune system. The challenge then is only to generate antibodies to these cancer-specific proteins. A few years ago, investigators led by Akinobu Matsuzaki at the Graduate School of Medical Sciences in Fukuoka, Japan, succeeded in doing so in the treatment of an 11-year-old girl with synovial sarcoma that had spread to other parts of her body.
The Japanese team started by collecting and isolating the patient’s own dendritic cells, a part of the immune system whose job it is to engulf invading microbes, chew them up and wear the pieces on their outer membranes. (This dismemberment allows another type of immune cell, the B cell, to make antibodies that recognize each chunk, or antigen.) Matsuzaki and his colleagues already knew that a translocation of one of this patient’s X chromosomes and a chromosome 18 had fused two genes together so that they encoded a unique protein. In the lab, the investigators combined the dendritic cells with that protein and then returned the “stimulated” cells to the patient. The girl’s immune system began to recognize the sarcoma, and the treatment temporarily suppressed the growth of the cancer that had spread to her lung. Although such outcomes have not yet been tested systematically, the strong scientific rationale and individual successes like that of Matsuzaki’s patient have sparked dozens of clinical trials that use a person’s own dendritic cells as a form of therapy for advanced cancers of the kidney, prostate, breast, colon and lung.
Sarcomas and Sick Kids
In a report published in 1997, Edmond S. Massuda and his colleagues at the University of Wisconsin Children’s Hospital in Madison described their experiments with a type of cancer called alveolar rhabdomyosarcoma-the most common type of soft-tissue sarcoma among children. This sarcoma is typically characterized by a translocation of chromosomes 2 and 13 that fuses the genes PAX3 and FKHR, both of which encode transcription factors (proteins that bind to DNA and switch other genes on and off). Normally, the PAX3 protein organizes embryonic muscle development, and FKHR is widespread. The fusion protein created from the melding of these genes causes muscle cells to remain immature and hence susceptible to other cancer-promoting events.
The investigators showed that this fusion protein is about a hundred times more effective at activating PAX3-regulated genes than PAX3 itself. Massuda and his coworkers took advantage of this property by taking the gene for diphtheria toxin A-a potent cellular poison-and modifying it so that it would be switched on only in the presence of the PAX3 protein. They then added that carefully crafted DNA to different strains of cultured cells, some of which carried the PAX3-FKHR mutant. Sure enough, the additional DNA selectively killed those cells that manufactured the fusion protein. Furthermore, when they added the fusion gene to otherwise normal cells, those cells also died in the presence of the PAX3-regulated toxin. Although the thought of injecting toxin genes into patients and trusting a bit of DNA alongside to keep them from harming the rest of the body may seem dangerous, the results from these experiments give hope that the effects will be specific to the tumor.
Like the cancer Massuda is trying to cure, Ewing sarcoma is one of the most common connective-tissue tumors in children and young adults, although it more commonly affects bone rather than soft tissues. It also stems from a chromosomal rearrangement, in this case the fusion of a gene from chromosome 22 with one from chromosome 11. The resultant protein transforms normal bone cells into cancer cells. In laboratory experiments, cultured bone cells proliferate rapidly when one adds this fusion protein and, conversely, stop dividing when it is removed-making the aberrant protein an ideal therapeutic target.
One new technology for getting rid of a particular protein in the cell is called small interference RNA (siRNA). In 2004, a team of clinical scientists headed by Howard A. Chansky at the University of Washington School of Medicine in Seattle reported on their successful application of this tool, which is both powerful and specific, to the task of suppressing the fusion protein responsible for Ewing sarcoma. The approach uses many short pieces of RNA that carry the complementary sequence to the fusion gene’s RNA. The interfering RNA binds to its target to form a double-stranded molecule that the cell perceives as a virus and snips apart. Without its RNA template, new copies of the fusion protein cannot be made, and existing copies are soon degraded. Using this strategy, Chansky and his coworkers silenced the mutant gene in Ewing sarcoma cells in culture, thereby preventing further cell division. These results represent the first use of siRNA to target the RNA that cancer cells make, and this approach will almost certainly lead to new therapies.
Although siRNA and similar experimental avenues for fighting cancer are still far from routine clinical application, the use of imatinib for treating GIST (sparked by the recovery of the Finnish woman) has now become accepted for certain advanced cases. How exactly does the drug work? GIST is usually caused by mutations in a gene called-no kidding-v-kit Hardy-Zuckerman 4 feline sarcoma viral oncogene homolog. This tongue-twister, called KIT for short, codes for a protein also named KIT, a kinase like CDK4, but one that adds phosphate groups only to tyrosine-one of the 20 different types of amino acids that make up human proteins.
Tyrosine kinases are often embedded in the outer membrane of the cell. There, they can receive signals from the immediate environment and transmit them to the nucleus (via a chain of other messengers), thereby helping to determine which genes are turned on or off. Overactive tyrosine kinases cause many kinds of tumors.
Although the KIT protein shows up in other cancers, such as small-cell lung cancer and seminoma of the testes, only GIST contains mutations in the gene that cause unregulated activity. In a 2001 paper, David A. Tuveson, then a postdoctoral fellow in the laboratory of Tyler Jacks at MIT, showed that imatinib interferes with normal and mutant KIT and inhibits the growth of cultured GIST cells that contain the latter. These observations helped to hasten the Food and Drug Administration’s approval of imatinib for the treatment of GIST tumors that have spread so widely that surgery is impossible-granted a scant 10 months after the publication of the original case study. (Customarily, experimental drugs are first approved for use in people who have exhausted their other options.)
To grasp the striking success of imatinib, one needs to understand that prior to its use there were no good treatments for GISTs. Most of these tumors are highly resistant to chemo- and radio-therapy, and multiple surgeries were the only palliative option. Now, the combination of surgery and imatinib benefits more than 80 percent of patients. Unfortunately, this upswing is only a respite in some cases. The tumors vary in their genetic make-up, which presumably explains the slow remedy seen by some patients and the unresponsiveness of others. The latter group often carries tumors that have little or no KIT protein, a variety that can also be found among patients who initially respond well to the drug, but worsen as the susceptible cells die off, leaving others to spawn new tumors that resist imatinib.
The success of imatinib has led to investigation of a growing number of compounds that interfere with the development of more-common (nonsarcoma) tumors in the lung, colon, breast and prostate. Among the approved drugs are gefitinib (Iressa) and erlotinib (Tarceva), which treat non-smallcell lung cancer (NSCLC) by inhibiting a tyrosine kinase called the epidermalgrowth-factor receptor (EGFR). This protein is overactive in many solid tumors and is typically associated with a poor prognosis.
Early clinical trials of gefitinib in the United States and Japan showed that almost half of the patients improved while taking the drug-remarkable considering their tumors had resisted standard chemotherapy. However, a different set of studies showed that adding gefitinib to conventional therapy did not provide an additional benefit. Nevertheless, the FDA quickly approved gefitinib for advanced NSCLC in patients whose condition had worsened under standard therapy.
Erlotinib also works by inhibiting the EGFR. In clinical trials, oncologists saw modest success treating NSCLC and bronchoalveolar carcinoma, another form of lung cancer. Evidently, the people who responded the best to erlotinib were those who carried mutations in their EGFR. In November of 2004, the FDA approved, after priority review, erlotinib for the treatment of patients with advanced non-small-cell lung cancer who did not improve after traditional chemotherapy. Because many of the most common cancers contain EGFR proteins, it makes sense to test all the available inhibitors of this protein to find out which are best for each combination of tumor type and genetic constitution. Such studies are now going on, and they should help improve the treatment of these types of cancer within the next few years.
Over the past decade, scientists have made remarkable discoveries of the molecular mechanisms that cause sarcomas and other cancers and are just now seeing the payoff in the form of treatments that specifically target genes or proteins of those cancer cells. However, investigators have much more to learn about the translocationspecific sarcomas, not to mention the large majority of cancers that do not carry a known genetic abnormality.
It is our hope that the knowledge obtained from the study of the better-understood sarcomas will apply to their uncharacterized relatives and to solid tumors as a whole, as demonstrated by recent advances in therapies for lung cancer. We believe strongly that the systematic analysis of these remarkable tumors will result in an enormous benefit to patients within the next few years.