William Check, PhD
Once upon a time, television was accurately called "broad casting." Advertisers
bought time on the three major networks to reach large heterogeneous audiences.
Then technology gave rise to hundreds of channels—and thus "narrowcasting."
Now an advertiser can target a more defined demographic—say, bass
fishermen under the age of 40. A similar revolution is taking place in
chemotherapy of hematopoietic malignancies.
Conventional cytotoxic chemotherapy attacks all dividing cells, with the premise that cancer cells will prove more susceptible. Like broadcasting ads, this approach is often successful. But cytotoxic therapy takes a heavy toll on healthy tissues as well. A more directed type of chemotherapy is emerging for leukemias, lymphomas, and myelomas, based on the identification of mutated growth factors that are central to malignant transformation and the development of agents that selectively inhibit these oncoproteins. Just as with the change in television, targeted chemotherapy has been made possible by technology—the many methods of genetic analysis that have become commonplace in the last decade.
"It is clear that we are now into the era of targeted therapy," says Jorge Cortes, MD, professor of medicine and deputy chair in the Department of Leukemia at the M.D. Anderson Cancer Center, Houston. "There are an increasing number of both targets that can be approached by intervention and agents that are coming along to block those targets or, in some instances, to stimulate them." Already targeted therapies are "clearly improving clinical outcomes," Dr. Cortes says.
Targeted therapy of chronic myeloid leukemia, or CML, with imatinib mesylate is "pioneering" the field of molecular medicine in blood cancers, says Jerald P. Radich, MD, member in the clinical research division of the Fred Hutchinson Cancer Research Center, Seattle, and professor of medicine at the University of Washington. CML is also pioneering molecular testing for biomarkers such as the BCR-ABL fusion transcript. "Ongoing trials with CML will influence how we will be using these tests in clinical trials," Dr. Radich says. "There are many markers in other leukemias. If molecular monitoring of response works in CML, we will gradually see a shift to [the use of biomarkers in] clinical trials, as well as in the therapeutic situation." This shift will signal a major change in laboratory practice as well. Dr. Radich believes that molecular monitoring is becoming standard of care in routine CML therapy.
"It is phenomenal what is happening," says Suzanne Kamel-Reid, PhD, ABMG, director of molecular diagnostics in the Department of Pathology at the University Health Network, head of the genetics program at Toronto Medical Laboratories/UHN, and professor of laboratory medicine and pathobiology at the University of Toronto. "We are better able to identify mutations present in different leukemias and lymphomas, and we can begin to think of a more directed approach toward therapy." As an example she cites diffuse large B cell lymphoma. "This lymphoma has been a real grab bag," she notes. "Now we can subclassify it according to mutations and begin to see more specificity in therapy, which is a result of better genetic tools."
Dr. Kamel-Reid agrees that these changes also herald a new era in laboratory testing for blood cancers. "The tests that clinicians ask for now are very specific," she says. "Rather than just looking at a bone marrow smear and seeing blasts, we can begin to use our genetic knowledge to see if certain mutations are present." In her view, hematopathologists should be integrating such information into their practice and into comprehensive reports. "Complex genetic information can’t stand on its own. It needs to be integrated with the other data that are available, such as clinical, biochemical, hematological, and morphological data, enabling us to see the whole picture."
As molecular analysis of hematopoietic malignancies other than CML progresses, additional tests will arise. For instance, recent findings in multiple myeloma mean that "molecular workup at diagnosis is critical" in this malignancy, says A. Keith Stewart, MD, professor of hematology at Mayo Clinic, Scottsdale. "My take-home message is that it is time to start testing these people," Dr. Stewart says.
Molecular pathologists are getting that message. "Assays for molecular markers have become standard of care quickly," says Janina Longtine, MD, chief of molecular diagnostics in the Department of Pathology at Brigham and Women’s Hospital, Boston, and associate professor of pathology at Harvard Medical School. "Many of them have been adopted and validated in individual molecular laboratories, so general pathologists may have to play catch-up a bit. It will be a challenge to do that in a rational manner using the resources that are available to us."
CML is the model for targeted therapy. Its treatment has been revolutionized
by the small-molecule inhibitor imatinib mesylate, which suppresses the
constitutive activity of the ABL tyrosine kinase (tk) coded for by t(9;22)—the
defining component of the Philadelphia (Ph) chromosome. At an Association
for Molecular Pathology session last November organized by Dr. Kamel-Reid,
Dr. Radich related the newest twists in this story. In the IRIS trial,
comparing imatinib mesylate to interferon-alpha (Ifn) plus cytarabine
(ara-C) in patients in the chronic phase of CML, a far greater proportion
of patients treated with imatinib mesylate had a complete cytogenetic
response, 68 percent versus seven percent (O’Brien SG, et al.
N Engl J Med. 2003;348:994-1004). Among the 83 percent
of patients with a major cytogenetic response on imatinib mesylate, time
to progression was impressive: 91 percent were progression-free at 42
months by cytogenetics.
But IRIS went beyond cytogenetic monitoring—it also included molecular
monitoring of response and disease progression. "Most work with BCR-ABL
testing by PCR had demonstrated that it predicted probability of relapse,"
says Dr. Radich, whose laboratory was one of the reference laboratories
for the polymerase chain reaction work in IRIS. Use of PCR to detect BCR-ABL
transcripts was "developed and honed in the transplant setting," Dr. Radich
says. At six to 12 months after bone marrow transplant, PCR assay of the
BCR-ABL fusion transcript was an independent predictor of relapse (Radich
JP, et al. Blood.
1995;85:2632-2638). "Once imatinib was found to produce cytogenetic remissions,"
Dr. Radich explains, "people realized that we needed a more sensitive
way to look at BCR-ABL burden. It was a logical step to use PCR testing."
Results from the IRIS study showed that molecular response at 12 months was a good indicator of overall outcome, Dr. Radich says. An estimated 39 percent of all patients treated with imatinib mesylate (but only two percent of those given Ifn/ara-C) had at least a three-log reduction in BCR-ABL transcript levels. For patients treated with imatinib mesylate who had a major cytogenetic response and a reduction in transcripts of more than three-log, progression-free survival at 12 months was 98 percent; if the reduction was less than three-log, progression-free survival dropped to 90 percent.
Molecular monitoring in IRIS also showed that patients don’t need to
be completely free of Ph-positive cells to be free of progression. Indeed,
a published model of imatinib action suggested that "imatinib is a potent
inhibitor of the production of differentiated leukaemic cells, but does
not deplete leukaemicstem cells" (Michor F, et al. Nature.
2005;435:1267-1270). "As you do more of these molecular markers," Dr.
Radich says, "you find it is more the rule than the exception" that "remission"
or "cure" does not mean extinguishing all cells that are positive by molecular
tests. This may explain why patients with excellent responses who have
gone off therapy have relapsed. Leukemia may become a chronic condition
like HIV, with lifelong therapy needed to control it. This would be a
radical departure from current concepts about cancer treatment.
Because of the predictive power of molecular monitoring, an upcoming trial of imatinib mesylate combination therapy will use quantitative PCR as a surrogate endpoint. A four-log reduction in transcripts at 12 months will be the goal. "This is the first time we’ve ever used molecular response as an endpoint in cancer trials," Dr. Radich says. Molecular monitoring could allow earlier intervention, such as transplantation, for patients who do not achieve the target response and drastically reduce the current three to four years needed to conduct trials.
Dr. Radich now monitors CML patients on imatinib mesylate by cytogenetics initially; for those who achieve a major cytogenetic response, he follows every three months by quantitative PCR. He considers a fivefold change in transcript level to be a "real" difference. "There are not a ton of labs that do [PCR monitoring for BCR-ABL transcripts] yet," Dr. Radich notes. To promote broader use of molecular monitoring for CML, it is being incorporated into National Cancer Care Network guidelines. Dr. Radich predicts that real-time quantitative PCR will eventually become the preferred method.
Even with imatinib mesylate treatment, some CML patients relapse. Many are
found to have mutations in the BCR-ABL binding domain (Branford S, et
2003;102:276-283). Quantitative PCR predicts which patients will relapse:
Those who have more than a twofold rise in BCR-ABL transcripts have a
greater chance of getting a point mutation and eventually relapsing. "You
can sometimes overcome the mutation with a higher imatinib dose," Dr.
Radich says, "but it is not clear if the effect lasts."
Dr. Cortes says mutations during imatinib therapy are "something that we understand better but still need to work on." Though mutations are found in some patients who develop imatinib mesylate resistance, many other patients become resistant to imatinib mesylate in whom mutations are not found. "Still others have mutations and the cancer doesn’t behave as we expect in the presence of mutations," Dr. Cortes says. "So it is not clear what role mutations play along with other mechanisms." Newer therapeutic agents are being developed for patients who become resistant to IM. Two that are in trials are dasatinib and AMN107. "They are still investigational," Dr. Cortes says, "but early results suggest that they are active."
While CML is a model for targeted therapy in hematopoietic malignancies, it is an imperfect model. "CML is not a good model for all hematopoietic malignancies or even all leukemias," says Donald Small, MD, PhD, professor of oncology, pediatrics, and cellular and molecular medicine in the Sidney Kimmel Comprehensive Cancer Center at the Johns Hopkins Medical Institutions. Dr. Small, who works on the genetics of acute myeloid leukemia, or AML, says that CML is really a myeloproliferative disorder—at least in its chronic phase. The distinction is crucial, since it is only in chronic-phase CML that imatinib mesylate is so effective. In phase two trials imatinib mesylate monotherapy produced complete hematologic responses in 88 percent of patients in chronic-phase CML, 63 percent of those in the accelerated phase, and only 26 percent of patients in blast crisis. Phase three trials included only chronic-phase patients.
"It’s possible that cells [in chronic phase] are affected only by the
BCR-ABL translocation," Dr. Small says. "So imatinib may be treating the
responsible mutation." A comparable condition is polycythemia vera, another
myeloproliferative disorder, which has recently been found to be associated
with a mutation in a single growth factor—the receptor tk JAK2 (James
C, et al. Nature.
2005;434:1144-1148). "Hopefully polycythemia vera will be amenable to
single-agent JAK2 inhibitors," Dr. Small says. "Since JAK2 may be the
only mutation in those cells, they will probably respond in the way CML
responds to imatinib."
In contrast, AML and other acute leukemias have multiple mutations affecting multiple pathways. Treating with an agent against a single mutated growth factor would block only one pathway by which cells are transformed. "That would be more analogous to treating CML in blast crisis or Ph-pos acute lymphocytic leukemia, where imatinib does not do what it does in chronic-phase CML," Dr. Small says.
Dr. Cortes reinforces this point. "The main difference between myeloproliferative diseases and leukemias, and between AML and CML, is that you have a more constant molecular abnormality in CML and myeloproliferative diseases," he says. "In the other malignancies we don’t know one common constant molecular pathway that you can target. That makes therapy much more complex." Relevant molecular targets are being identified in other malignancies, revealing that some conditions like AML may be several entities on the molecular level. In Dr. Cortes’ view, targeted therapy of more complex blood cancers will emerge, but it won’t be with single-agent chemotherapy.
A good example of a blood cancer that is being dissected into a number of different entities on the molecular level is multiple myeloma. Dr. Kamel-Reid says she picked multiple myeloma as the other topic to be discussed at the Association for Molecular Pathology session because it has not had a lot of attention. "It’s not a disease that we knew a lot about, but in the last few years our understanding of its underlying biology has really progressed," she says. "It’s analogous to where the CML story was a few years ago."
Dr. Stewart noted at the outset of his AMP presentation that our understanding of multiple myeloma on the molecular level is several years behind that of CML. "Multiple myeloma is behind most blood cancers," he told CAP TODAY, "because we were not able to find any translocations till the mid-1990s." Chromosomal abnormalities in multiple myeloma are small and at the telomeric end of the chromosome, so they are hard to see by conventional cytogenetics. Methods that are more sensitive, such as FISH, are required. As a result, the ability to translate translocations into clinical outcomes and therapeutics has been slower than for other hematologic malignancies. "But we are catching up very fast," Dr. Stewart says.
Though all myelomas look the same under the microscope, he says, "variable
outcomes in multiple myeloma suggested multiple myelomas plural." In the
last few years a suite of characteristic genetic abnormalities has been
uncovered in multiple myeloma. Initially a number of recurring translocations
were found that affected IgH structure—t(4;14), t(11;14), and t(14;16)—as
well as a deletion on chromosome 13 (Fonseca R, et al. Blood.
2002;100:1417-1424; Fonseca R, et al. Blood.
2003;102:2562-2567). Some translocations were associated with a hyperdiploid
state, others with nonhyperdiploidy.
Abnormalities in cyclin D1 and D2 production were also discovered. Some
were direct, while others were secondary to mutations in genes for other
growth regulators, such as c-maf, MMSET, and fibroblast growth factor
receptor 3. And Dr. Stewart’s laboratory found that a deletion in the
gene for p53 portended poor outcome (Chang H, et al. Blood.
A group of investigators at Mayo Clinic used three confirmed genetic
factors—IgH translocations, deletion of a large segment of chromosome
13, and deletion of the p53 gene on chromosome 17—to produce three prognostic
groups with post-transplantation survival ranging from 25 to 51 months
(Fonseca R, et al. Blood.
2003;101:4569-4575). The classification shows differing prognoses, they
wrote, but also and "more importantly it provides further compelling evidence
that MM is composed of subgroups of patients categorized according to
underlying genomic aberrations."
More recently, a classification of multiple myeloma was devised using
two genetic factors-translocations and expression of cyclin D genes-that
produced eight groups, which the authors speculated were "defined by early,
and perhaps initiating, oncogenic events" (Bergsagel PL, et al. Blood.
2005;106:296-303). They noted that "[D]espite subsequent progression events,
these groups have differing gene expression profiles and also significant
differences in the prevalence of bone disease, frequency at relapse, and
progression to extramedullary tumor." In this analysis, all groups had
upregulation of one cyclin D family member. "Cyclin D upregulation is
the unifying event in myeloma," Dr. Stewart says.
He notes that these classifications verify that there are "many and multiple myelomas" with variable prognoses that may require individualized treatment. For instance, poor prognosis of some groups after transplant suggests that those patients should not get that intervention.
In principle, small-molecule inhibitors could be discovered that would block each of the pathways leading to or from one of the cyclins. Phase one trials are now ongoing with several substances that act against fibroblast growth factor receptor 3. The hypothesis, as Dr. Stewart enunciates it, is that "Pharmacologic abrogation of tyrosine kinase signaling by FGFR3 in myeloma cells will result in a tumor-specific cytotoxicity." One difference from the CML/imatinib story will be that "targeted therapies will still be added to broad-based therapy for myeloma, just as in acute leukemias," Dr. Stewart says. He considers it "unlikely" for an imatinib-like effect to happen in myeloma, for the reason cited already-myeloma is the equivalent of CML in blast crisis. "No one drug will treat all myelomas," he says.
For diagnosis, Dr. Stewart says, "All patients should be worked up" for
the known genetic abnormalities in multiple myeloma. "Such thorough workups
are not widely adopted yet, but that needs to change," he insists. In
addition to conventional cytogenetics, workup should include detection
of translocations giving rise to IgH mutations-t(4;14), t(11;14), and
t(14;16)-as well as deletion 13 and 17 and trisomies. "FISH is probably
still the best way to pick up IgH rearrangements," Dr. Stewart says. He
also notes that, to do FISH properly, you have to be looking at malignant
cells. One method of identifying malignant cells is to use tricolor FISH
and to look for kappa or lambda chains—whichever the myeloma is—and only
score clonal cells staining positive for that subtype. Like multiple myeloma,
acute myeloid leukemia is being dissected into several entities by molecular
analysis. Dr. Small has been working on the most important known genetic
change, mutations in the FLT3 (FMS-like tyrosine kinase 3) gene. He was
the first to clone the gene, which was then found by a group in Japan
to be mutated in about one-third of cases of AML. Mutated FLT3 genes are
found in stem cells of AML patients (Levis M, et al. Blood.
2005;106:673-680) and are associated with poor prognosis. "FLT3 is a receptor
tyrosine kinase very analogous to BCR-ABL in CML," Dr. Small says. Just
as the BCR-ABL fusion constitutively activates ABL kinase, mutations in
FLT3 permanently turn on its kinase activity, which, in cooperation with
several other mutations, gives rise to transformation.
Dr. Small used an assay he developed to screen thousands of small molecules
in a commercial biopharmaceutical company’s library and became the first
scientist to identify a small molecule, CEP-701, that inhibits FLT3 activity.
He showed in vitro that CEP-701 kills leukemic cell lines and primary
leukemic samples from patients whose AML cells have FLT3 mutations (Brown
P, et al. Blood.
2004;104:1841-1849). He is working now with a group doing a phase two
clinical trial of CEP-701 plus broad-based chemotherapy in relapsed adults
with AML and FLT3 mutations. Several other FLT3 inhibitors are also in
As single agents, FLT3 inhibitors have shown the ability to clear peripheral
blood of leukemic cells but not to clear bone marrow (Smith BD, et al.
2004;103:3669-3676). Responses are usually transient, lasting only a few
months. "We really expected to see the degree of response that we saw,"
Dr. Small says. "AML patients have other mutations in addition to FLT3,
so when we inhibit FLT3 we are only taking care of one pathway by which
those cells are transformed.
"Some people were disappointed about monotherapy against FLT3," he continues, "but I think people will get excited again when results come out with FLT3 inhibitors plus chemotherapy." Studies combining chemotherapy with FLT3 inhibitors appear to show promise in improving remission rates, both at initial diagnosis and in relapsed patients. Dr. Small finds these results encouraging: "I hope this will be the start of data showing that small-molecule inhibitors combined with chemo ther apy will improve the cure rate for AML patients." In pediatric AML, FLT3-bearing cancers have a cure rate of 10 to 20 percent, much lower than the 50 percent cure rate for cases without an FLT3 mutation. Dr. Small’s hope is to bring the cure rate for FLT3-positive AML cases up to that of non-FLT3 patients, and maybe even exceed it.
Dr. Small envisions a time when AML can be treated with a combination of an FLT3 inhibitor and two or three other targeted small-molecule inhibitors against different mutated growth factors. "When we know what the other mutations are, we can get small-molecule inhibitors against them as well," he says. "That is absolutely the way to go." But finding these other mutated transcription factors will pose a challenge. It’s easier to find mutations when a translocation breakpoint is known. Then you can examine the genes at the breakpoint, as was done in CML. Point mutations, however, are much more difficult to discover. "That may take sequencing the entire genome of many leukemia samples," Dr. Small says.
A second obstacle is financial. What we call AML may in one case have mutated FLT3 plus three other mutations and in another case have mutated FLT3 plus an entirely different set of mutations, requiring truly patient-specific treatment. "The market for small molecules that inhibit individual mutations will be much smaller than for imatinib," Dr. Small notes. Some may be important in only one percent of AML cases. "Is any company going to be interested in developing a drug against a molecule that is mutated in only one percent of leukemias?" he asks.
For now, the focus is on FLT3 inhibitors. Clinical success of these agents could have implications for the laboratory. While some patients may respond nonspecifically to an FLT3 inhibitor, a much higher percentage of FLT3-positive patients is likely to respond. In the trial Dr. Small is working on, treatment with CEP-701 is restricted to patients who qualify with the FLT3 assay. So far other trials are not using a qualifying assay. But Dr. Small thinks other companies are starting to see that they are much better off selecting patients for these trials. "In phase one you may take all comers," he says. "But when you want to look at response rates, more companies are coming around to agree with us that you want to restrict treatment to patients with a mutation." So a qualifying assay could become standard for FLT3 inhibitor treatment in AML, as is done with Herceptin in breast cancer.
The major molecular assay for targeted therapy in blood cancers that is standard of care now remains quantitative PCR monitoring for BCR-ABL transcripts in CML. Dr. Kamel-Reid uses this assay much as Dr. Radich does—to monitor for minimal residual disease in patients with a major cytogenetic response. "When a patient is losing response, you can look for mutations in the ABL kinase domain by sequencing," she says. Tests for ABL kinase mutations are now available. Therapeutic choices in patients with mutations are to offer dasatinib in a clinical trial for certain mutations (some mutations don’t respond to either imatinib or dasatinib), to go to bone marrow transplantation for eligible patients, or to use interferon.
A second, less common molecular test in routine use is for the PML-RAR (retinoic acid receptor) alpha fusion in acute promyelocytic leukemia. The fusion is detected by RT-PCR, FISH, or cytogenetics. "Detection of the PML-RAR alpha transcript is the most sensitive predictor of relapse," Dr. Kamel-Reid says, adding that this was one of the first examples of targeted therapy in hematopoietic disease. PML-RAR alpha fusion disrupts the retinoic acid receptor resulting in deregulated retinoid signaling; treatment is with high-dose retinoic acid in combination with chemotherapy.
To Dr. Longtine, the AMP presentations of Drs. Radich and Stewart "represent the ideal of the way that many of us in pathology think our discipline should be practiced.
"Not only do we make a diagnosis, we also provide a guide to therapeutic options," she says.
To do that, Dr. Longtine says, "we must be willing to incorporate other technologies or other disciplines that we may not have that much expertise in. We have to become facile in our understanding of how to incorporate these technologies into our workup of patients." The pathologist is often the first person with whom a clinician may have a dialogue. "We need to be comfortable in advising clinicians how to use resources," she says.
Regarding genetic factors that are prognostic for different classes of multiple myeloma, Dr. Longtine has been working with the cytogeneticist in her department to devise an algorithm to define probes that would be informative but not overwhelm their financial capacity. "All these ancillary tests are quite expensive," she says. "We need to work closely with our colleagues in other laboratories to come to consensus so that we can use our limited resources to get clinically important biological information."
At the same time, she says, pathologists need to tend to technical issues, such as the perennial one of standardizing molecular assays. She sees Dr. Radich’s talk "as sort of a heralding call for the AMP committee to move forward our efforts toward standardization of these techniques." One of the challenges: There have not been commercial resources and reagents to promote standardization. "Dr. Radich highlighted how important that is from a clinical perspective," Dr. Longtine says. "We in AMP are in a position to put forward some consensus statements. We have done a wonderful job of bringing new technology into the workplace." The next step is to standardize it.
"It’s an exciting time," Dr. Longtine says. "This is why many of us became pathologists-so that by understanding the biology of tumors we could help in their treatment."
With the aid of sophisticated genetic analyses, that goal is now within reach.
William Check is a medical writer in Wilmette, Ill.