College of American Pathologists
Printable Version

  Breast cancer—on the path to better treatment


CAP Today




March 2009
Feature Story

Karen Lusky

Whether you view the medicine cabinet for breast cancer as partly full or partly empty, there’s hope that it could relatively soon contain new non-chemotherapy options and personalized therapies. Medical scientists are also finding ways to target the genetic underpinnings of metastasis so as to prevent it—efforts that could revolutionize breast cancer treatment overall.

The biggest news of a breakthrough involves an old drug, tamoxifen, which is known to have variable success in preventing recurrence. Mayo Clinic researchers have honed in on a common reason the drug doesn’t work in some cases—one that can be identified by readily available genotype testing.

In a presentation at the December 2008 San Antonio Breast Cancer Symposium, John Hawse, PhD, reported that he and colleagues at Mayo Clinic in Rochester, Minn., have found that endoxifen, a metabolite of tamoxifen and produced by the liver enzyme CYP2D6, has a mechanism of action unique to the parent drug, tamoxifen, and is the most important tamoxifen metabolite mediating tamoxifen drug effect.

Previous work has demonstrated that endoxifen is an active metabolite and produced by the CYP2D6 enzyme, and people who are genetically poor metabolizers of CYP2D6 produce little endoxifen. (Genotype testing for CYP2D6 reveals wheth­er a person is a poor, intermediate, extensive, or ultra-rapid metabolizer.)

As Dr. Hawse reported in his symposium pre­sentation and in a paper accepted by Cancer Research, endoxifen appears to block regulation of gene transcription so that some genes get turned on or off. “If you expose ER-positive breast cancer cells to estrogen, you can show through proliferation assays that the cancer grows faster,” Dr. Hawse says. And low concentrations of endoxifen, which occur in poor metabolizers, do not block that estrogeninduced growth, but high concentrations mimicking an extensive (normal) metabolizer of CYP2D6 do, he says.

Mayo oncologist Matthew Goetz, MD, says the research also shows that endoxifen appears to ­degrade the estrogen receptor, whereas 4HT (4-­hydroxtamoxifen), the other active tamoxifen metabolite, and tamoxifen itself simply stabilize it. 4HT is present in very small concentrations in women taking tamoxifen.

Dr. Goetz presented a separate Mayo Clinic study at the symposium showing that poor metabolizers of CYP2D6 who take tamoxifen had an approximately fourfold higher risk of early recurrence of breast cancer compared with extensive metabolizers while all were still taking the drug.

To conduct the study, Dr. Goetz and colleagues performed a retrospective analysis of patient outcomes in an Austrian clinical trial (ABCSG 8), which he says is unique in that it included one arm of patients who switched to anastrozole, an aromatase inhibitor, after two years on tamoxifen. And the CYP2D6 poor metabolizers who switched to anastrozole, which is not metabolized by CYP2D6, did not show an increased risk of recurrence. Dr. Goetz says the study did not show an increased risk for tamoxifen takers who are intermediate metabolizers. The researchers did not look at outcomes for ultra-rapid metabolizers.

Mayo’s original study data published in 2005 and again in 2007 on CYP2D6 status and outcomes found that poor metabolizers had a 32 percent risk of relapse or death within the first two years of starting tamoxifen, compared with a two percent risk of recurrence or death for patients with extensive metabolism, Dr. Goetz says. Intermediate metabolizers tended to also have an increased risk of recurrence compared with extensive metabolizers, but the difference was not statistically significant, he says.

Also important, says Dr. Goetz, is that taking certain selective serotonin reuptake inhibitor antidepressants and other medications can turn someone into a poor metabolizer for tamoxifen. “So variability in CYP2D6 enzyme activity, and thus endoxifen concentrations, is related to both genetic as well as environmental factors”—that is, “when a caregiver unknowingly prescribes a drug concomitantly with tamoxifen that inhibits CYP2D6.”

Some of the CYP2D6 inhibitors, says David Flockhart, MD, PhD, include fluoxetine (Prozac), paroxetine (Paxil), bupropion (Wellbu­trin), and, to a lesser extent, sertraline (Zoloft). Over-the-counter diphenhydramine (Benadryl), which patients may take on their own without telling a physician, is also a problem. While people seldom take Benadryl for years, “it’s such a potent inhibitor of the enzyme that even taking it for a short period could be like going off tamoxifen for a time,” says Dr. Flockhart, chief of the Division of Clinical Pharmacology at Indiana University School of Medicine, Indianapolis.

John Horn, PharmD, professor of pharmacy at the University of Washington in Seattle, says that “probably all of the sedating anti-histamines have the ability to inhibit CYP2D6, whereas the nonsedating ones don’t appear to” do so. (For a list of CYP2D6 inhibitors, log on to

As for the clinical implications of the Mayo study findings, Dr. Goetz says, “Our recommendation is to first identify those patients [CYP2D6 poor metabolizers] who will never make enough endoxifen and avoid giving them tamoxifen.” Perhaps the rest of the patients could be monitored by checking serial endoxifen levels, he adds. Studies that are underway are looking at that strategy.

As for treatment options, a postmenopausal woman who is a poor or even intermediate metabolizer might do better on an aromatase inhibitor, Dr. Goetz says. These medications are not effective, however, for wo­men with functioning ovaries. Thus, he tells premenopausal patients who test as poor metabolizers that although they can take an aromatase inhibitor if they also take Lupron, a gona­dotropin releasing hormone agonist that suppresses ovarian function, the long-term risks of that strategy are not known, at least for now. A worldwide randomized clinical trial (SOFT trial) is expected to answer the question of how best to treat premenopausal patients with hormonally responsive breast cancer, he notes.

Despite the aromatase inhibitors being what Dr. Goetz calls “wonderful drugs,” his belief is that they are no different than tamoxifen in that they work very well in some patients and not in others. He points to published data showing that aromatase inhibitors lower estrogen levels very effectively in some patients and not at all in others. “I suspect that genetic differences in the enzymes responsible for the metabolism uptake and distribution of these drugs may affect this variability. Once we understand this, we will take another step forward toward fully individualizing therapy” based on the patient’s genetic makeup.

Indiana University’s Dr. Flockhart says he and colleagues have been doing a lot of work to define what affects aromatase inhibitors and have two comprehensive papers coming out on the subject. But “we don’t think meta­bolism of these drugs is going to be a big headline like it has been for tamoxifen because the parent drugs are active,” he says.

In the meantime, there’s interest in developing endoxifen as a drug, which would sidestep the CYP2D6 metabolism issue and could potentially deliver a more powerful therapeutic punch.

Mayo’s Dr. Hawse says he and his colleagues have conducted a number of different studies that are aimed at mimicking a patient receiving a pharmacologic dose of endoxifen (1,000 nM), and have found that this concentration is even more effective than a 100 nM concentration, the level typically produced by an extensive metabolizer of CYP2D6.

“We see no induction of growth by estrogen in cell lines treated with 1,000 nanomolar concentrations of endoxifen, and we may actually be seeing some cell death,” he says.

The other active tamoxifen metabolite, 4HT, which appears at very low concentrations in the body, could also be a candidate for a drug, Dr. Hawse says. He points to “thousands of papers” that have examined the effects of high concentrations of 4HT (100–1,000 nM) on breast cancer cells and tumors. “But a patient never produces that high of a level of 4HT,” though that’s not to say that if you made a drug out of it that it might not work, he adds.

Whether endoxifen would produce negative side effects remains to be seen. “We have initiated studies to look at its effect on bone, which is a highly endocrine-responsive tissue, and on endometrium and ovaries,” says Dr. Hawse.

A growing understanding of how certain genes and cellular signaling pathways drive metastasis is paving the way for therapies that could prevent it. While researchers are studying different genes and systems in that regard, the underlying functionality of the genes is probably the same, says Larry Norton, MD, deputy physician-in-chief and director of breast cancer programs at Memorial Sloan-Kettering Cancer Center in New York. And that functionality entails three things, he says: the ability of the cells to get into the blood and, second, to make a niche in a distant site where they “make the soil attractive” to grow. Third, the cells have to be able to attract white blood cells from the bone marrow to form blood vessels and support the tumor’s structure and growth.

Cancer researcher Lawrence Lum, PhD, says “the different pathways related to cancer have the same functionality in that you can consider them stem cell pathways,” essential for maintaining the stem cell compartments in an adult.

“These pathways were extremely important when we were embryos and making cellular decisions ... whether cells will be hair or teeth, etc.,” says Dr. Lum, assistant professor of cell biology at the University of Texas Southwestern Medical Center in Dallas. “But when we become adults, those types of behaviors are primarily restricted to stem cell compartments and activated when there is tissue damage.” Thus, if a person scrapes his or her knee, the body activates a pathway to recover cell linings.

“Stem cell pathways are very important for cancerous stem cells to set up shop in a tissue to which they are normally not privy—one that they would normally never seed,” he says.

The increased understanding of how metastasis works is making it more of a bull’s-eye for therapeutics in its own right, says Patricia S. Steeg, PhD, of the National Cancer Institute, who gave a talk on the topic at the December breast cancer symposium. In the past, metastasis was “written off,” she says, because of the view that once it occurs, “the horse is out of the barn.”

Yet growth in the primary tumor versus growth in a distant site are different, said Dr. Steeg, chief of the NCI’s Women’s Cancers Section, Laboratory of Molecular Pharmacology. For example, metastasis suppressor genes (the first of which, nm23, Dr. Steeg discovered) have no effect on primary tumor size when you express them in a tumor cell line. “But they do suppress metastasis—that is, growth in a distant site,” Dr. Steeg told CAP TODAY.

Also, drugs tend to have very different effects on metastases versus the primary tumor, Dr. Steeg says. “Yet we rarely do that kind of study. We usually inject some cancer cells under the skin of mice, give the drugs, and look at the growth two weeks later.”

When Dr. Steeg and her research team found the nm23 metastasis suppressor gene, they tried in vain to turn the gene back on in metastatic tumor cells by using genetic modifying agents, such as DNA methylation inhibitors and HDAC (histone deacetylase) inhibitors. Then the researchers tried high-dose progestin, and “very unusually,” she says, found that it worked in triple-negative breast cancer.

Whose idea was it to use progestin? “Not mine,” says Dr. Steeg with a quick laugh. When looking at the promoter of nm23 to see what would turn it on, she and her colleagues found glucocorticoid response elements. So they initially tried dexamethasone, a classic steroid, to see if that worked. And it would, “if we stripped the culture medium of all the endogenous steroids,” she says. Yet in the presence of certain levels of steroids normally coursing through the blood, dexamethasone would not induce nm23.

Undaunted, Dr. Steeg showed the work to someone who studies steroids; this person suggested trying a progestin, explaining that it would interact with the glucocorticoid receptor. “And darn if it didn’t work! It wasn’t what I was after because it has side effects such as weight gain.”

“It turned out the progestin can act through the progesterone receptor, glucocorticoid receptor, and androgen receptor,” says Dr. Steeg. But “in our model, progestin [medroxyprogesterone acetate, or MPA] elevated nm23 expression through the glucocorticoid receptor.”

In an experiment, Dr. Steeg and her team administered high-dose MPA to mice implanted with triple-negative breast cancer cells. “We injected the cells in their tails, which will lodge the tumor cells in the lungs, and let the [cells] grow for a month, at which time there were micro mets in the lungs, so we were looking at metastatic colonization. Then we randomized mice to progestin or a vehicle, and the progestin caused a 33 to 60 percent, depending on the dose, reduction in the number of mets that grew out.”

Now a multicenter phase two clinical trial is testing MPA for patients with ER- and PR-negative breast cancer to examine its effect on the nm23 metastasis suppressor gene. Participants in the clinical trial will have their tumors tested for nm23, Dr. Steeg says. The trial will explore whether the progestin induces clinical responses when used alone or with metronomic chemotherapy. “We have unpublished data that it has an antiangiogenic role as well,” she adds.

Candidates for the trial include patients with triple-negative or HER2-positive stage four breast cancer, says Kathy D. Miller, MD, an oncologist at Indiana University Simon Cancer Center in Indianapolis, who is heading up the trial. She says a tumor’s HER2 status will not affect progestin’s mechanism of action in turning on nm23, though hormonally responsive tumors would.

In the phase two trial, which Dr. Miller expects to be completed in about a year, researchers will look at whether they can achieve the level of MPA in the trial participants that produced changes in female mice. They will also conduct serial blood testing and tumor biopsies (or skin biopsies if tumor isn’t availa­ble) on participants to see if MPA at the desired level alters various signaling pathways—primarily a change in nm23 levels, and changes in the glucocorticoid receptor, says Dr. Miller.

“I don’t think any of us expect to see bulky advanced disease shrink with this therapy,” Dr. Miller says. But if the trial shows changes in the signaling pathways, it would be “very reasonable,” she says, to try MPA in early stages of disease to try to prevent metastasis. If a phase three trial pans out, women with triple-negative breast cancer might have a helpful nonchemotherapy option. “They have a lot of options for chemotherapy,” says Dr. Miller, “but no one is ever all that thrilled about getting chemotherapy.”

In another popular symposium presentation, Emory University researcher Benjamin Barwick, MS, a bioinformatic analyst, presented preliminary data showing that the Wnt pathway appears to be upregulated in triple-negative breast cancer, providing a potential target for therapy using medications already on the market for type 2 diabetes.

Wnt was discovered in 1982 by Harold Varmus, MD, Nobel laureate and former director of the National Institutes of Health and an advisor to president Barack Obama.

“Wnt is a pathway not only integral to development but evolutionarily conserved,” Barwick says—in everything “from flies to fish to worms to mice to humans.”

A 1982 landmark paper by Dr. Varmus and Roel Nusse, PhD, described Wnt as pivotal to mouse mammary tumor virus’ integration and subsequent oncogenesis, Barwick says. Later in that decade, researchers “discovered that when Wnt was knocked out of Drosophila, it resulted in a wingless phenotype.” The combination of the words wingless and integration gave Wnt its name.

Wnt is also a pathway known to be exploited by cancer. For example, says Barwick, it’s known that “deleterious germline mutations of APC—an infamous familial oncogene for colorectal cancer—can lead to Wnt activation in colon carcinomas,” and potentially in basal cell carcinomas.

“In triple-negative breast cancer, the Wnt pathway tends to become more activated than in other subtypes of breast cancer.” Barwick has his own hypothesis, which he admits is far from being proved, for why that might occur: Obesity and high sugar diet may have something to do with it, he says.

Triple-negative breast cancer is known for being prone to metastasize. Emory researchers hypothesized that Wnt signaling might be playing a role in promoting its “migratory and invasive” nature, Barwick says.

To test that hypothesis, the researchers treated a triple-negative cell line with troglitazone, a drug at one time approved for type 2 diabetes. Troglitazone is a PPARgamma agonist, a class of drugs known to have an ancillary inhibitory action on Wnt, as well as the ability to attenuate DNA and mRNA overall production, Barwick says. The treatment appeared to work, reducing the cells’ migratory action in a migration assay.

While troglitazone was removed from the market because of liver toxicity, two more PPARgamma agonists (rosiglitazone and pioglitazone) are still on the market, he says. And Emory is considering using those medications in clinical trials for triple-negative breast cancer, though, Barwick says, “we have to crawl before we can walk.”

“Showing efficacy of a Wnt inhibiting modality would require at least animal xenograft models prior to clinical trials,” he notes. And if that panned out, the “walking component” would be to develop a diagnostic for the modality. That will require identifying the best biomarker of Wnt activation in terms of specificity and sensitivity. “That’s a ways off in our lab—at least, I’d say, a couple of years,” Barwick says.

In a study published in a recent issue of Nature Chemical Biology, researchers described two potential targets in the Wnt pathway that are sensitive to small molecules that could be developed into drugs. One target is “the porcupine gene, which encodes an extracellular enzyme essential for Wnt function,” says UT Southwestern Medical Center’s Dr. Lum, one of the researchers (Chen B, et al. Nat Chem Biol. 2009;5[2]:100–107. Epub Jan. 4, 2009). “The other target is a gene called axin that suppresses the Wnt pathway.”

The research didn’t focus on breast cancer, though Dr. Lum points out that “maintaining the stem cell population in many different types of tissues appears to be a function of Wnt, including possibly the breast and certainly in the GI epithelium.” And “if cancer cells are related to normal stem cells, then it’s conceivable that they’d be supported by the same types of cellular signaling systems that support stem cells—and that would make them potentially visible targets.”

“We describe two different approaches to chemically disrupting the Wnt pathway,” as Dr. Lum sums up the work, “and that’s the beginning of a chess game.”

In a study at Princeton University published in January, researchers identified a gene that appears to be associated with breast cancer metastasis and chemotherapy resistance: Metadherin, or MTDH, which resides on the eighth chromosome. When overexpressed, which MTDH appears to be in 30 to 40 percent of breast tumors, it portends distant metastasis and resistance to chemotherapy, according to the study (Hu G, et al. Cancer Cell. 2009;15[1]:9–20).

As for the gene’s role in metastasis, the only function that researchers can establish at this point is that “it promotes adhesions of the breast cancer cells to the blood vessels and allows them to seed the tumor to distant organs,” says Princeton molecular biologist Yibin Kang, PhD, who led the study.

It also appears that MTDH allows tumor cells to survive chemotherapy by activating a number of different chemo­resistance genes, including “c-Met, Hsp90, and ALDH3A1, and suppressing pro-apoptosis genes, such as TRAIL and BNIP3,” Dr. Kang says. “Patients with breast cancer tumors that overexpress MTDH might benefit from a higher dose of chemotherapy or more frequent treatments and followup,” says Dr. Kang. However, this has not been tested in breast cancer patients, he adds.

Thus far, the researchers have not tied MTDH overexpression to any particular type of breast cancer. As an explanation for that preliminary finding, Dr. Kang says: “Even among ER-positive, and HER2- and triple-negative breast cancer, you can classify tumors as being high versus low in terms of risk of recurrence and metastasis based on the level of MTDH expression.” This conclusion is consistent, he notes, with another independent study in which MTDH was called AEG1 (Li J, et al. Clin Cancer Res. 2008;14[11]:3319–3326).

In the Princeton study, Dr. Kang and colleagues looked at the gene expression at the protein level through immunostaining. This “tells you that the MTDH protein is overexpressed,” he notes. The researchers found that two-thirds of the cases they saw of protein overexpression are probably due to amplification of the genomic DNA—“that is, too many copies” of MTDH—and the rest are due to transcription activation occurring for unknown reasons, he says.

“We can’t really say whether the gene promotes metastasis to one organ or another based on our data,” but reports indicate it seems to favor the lung, says Dr. Kang. “But this will have to be examined more carefully in the clinical setting by looking at the metastasis outcome of patients.”

“We are trying to develop inhibitors, but right now we are mostly in the stage of figuring out how this protein works. ... We have identified interacting proteins that bind to MTDH protein.”

As the next step for the gene, Dr. Kang says, “LabCorp wants to discuss with us developing a marker for prognosis.”

MTDH is included in Mamma­Print, a 70-gene molecular profiling test that is FDA cleared for both hormone-driven and hormone-independent breast cancer and that’s used to predict prognosis and identify high-risk patients who will benefit from cytotoxic chemotherapy. The gene is not in any of the other FDA-cleared or CLIA lab-developed molecular profile tests for breast cancer, says oncologist Richard A. Bender, MD, chief medical officer for Agendia, which sells MammaPrint.

Dr. Bender, who maintains an active oncology clinical practice, predicts that eventually molecular science will be able to give physicians an individual genomic profile for each patient and their tumor. The profile, he says, will predict prognosis and identify which drugs will turn pathways on or off to either kill the tumor or deprive it of key nutrients, enzymes, or co-factors required for it to grow.

“We will be able to use drugs that aren’t 30 percent effective but rather 70 percent or more effective,” reducing toxicity to a minimal level and taking into account an individual patient’s ability to metabolize a drug, he says.

Dr. Lum foresees physicians eventually having a medicine cabinet with, say, 50 inhibitors of various pathways from which to choose in customizing a patient’s treatment. “When you compare that to what we are doing now, it is a dream,” he says, “but that’s where we hope to be in the not-too-distant future.”

Karen Lusky is a writer in Brentwood, Tenn.