Scientists submitted more than 5,100 abstracts to the annual meeting of the American Association for Cancer Research (AACR) in New Orleans, March 24-28. Many abstracts focused on hot topics such as angiogenesis and apoptosis. Studies of breast and prostate cancer abounded, as did jazzy work using DNA microarrays. A large bloc of intriguing abstracts, however, explored the less traveled byways of cancer research. Selected almost at random, a handful of such abstracts, and the posters and a talk elaborating on them, are discussed here in a reporter's notebook format. More information and the poster abstracts are available on the AACR Web site (www.aacr.org).
Early Warning of a Killer
A diagnosis of pancreatic cancer (PCA) is still virtually a death sentence. The five-year survival rate stands at 4 percent. In the United States, PCA is fourth among cancers in the number of deaths it causes, with 29,000 expected in 2001, according to the American Cancer Society. In an early-morning lecture at the AACR meeting, Margaret A. Tempero, deputy director of the Comprehensive Cancer Center at the University of California, San Francisco, noted that one-third to two-thirds of patients with one-centimeter tumors already show lymph-node metastases. "You almost have to diagnose [this disease] before it becomes cancer histologically," she says, adding that molecular-based diagnosis might be the key to detection and prevention strategies.
The posters: Epithelial cells secrete glycoproteins known as mucins, which lubricate and protect the body's ducts and lumina. A rat homolog of mucin-4 (muc4) increases lung tumor metastasis.1 A study described by Poster #1727 found that 75 percent of human PCA tumors and tumor cell lines produced muc4 mRNA, compared to no samples from pancreatitis cases or normal pancreases. Thus, muc4 was more specific than the reigning PCA marker, antigen CA19-9, which also increases in pancreatitis, notes Mahefatiana Andrianifahanana. He is a graduate student with study leader Surinder K. Batra, an associate professor of biochemistry and molecular biology at the University of Nebraska Medical Center.
In a finding suggestive of muc4's potential value as a clinical cancer marker, Poster #3311 revealed that muc4 mRNA was detectable in the white blood cells of 18 of 27 PCA patients, but not in cells from healthy people, pancreatitis patients, or patients with cancers besides PCA. Thus, the muc4 assay's sensitivity rate was 67 percent, compared to rates of 69 to 93 percent for the CA19-9 assay found by other studies. Jörg Ringel, Batra's former postdoc, says, "We believe there is an interaction between cancer cells and white blood cells" that somehow prompts the lymphocytes to produce muc4. But the small study uncovered no differences in prognosis between patients whose cells were positive and negative for muc4 mRNA, he adds.
The adjacent poster, from the lab of Michael Bouvet, was devoted to another possible PCA marker: parathyroid hormone-related protein (PTHrP). Bouvet, an assistant professor of surgery at the University of California, San Diego, School of Medicine, recently reported that human pancreatic tumors commonly produce PTHrP (which is made by other tumors as well).2 Poster #3312 showed that when pancreatic tumor samples were transplanted into the pancreases of nude mice, PTHrP levels were detectable in serum after five weeks and then rose in tandem with primary tumor weight. No PTHrP was seen in mice without tumors.
After the AACR meeting: Despite having found muc4 mRNA in white blood cells, Batra says, "We are still looking for a very simple test based on just serum," which is more available than lymphocytes to researchers and is easier for small hospitals to analyze. Pancreatic tumors release muc4 into the bloodstream, and Batra hopes to develop an ELISA test or radioimmunoassay to detect the protein.
Batra is planning larger clinical studies of the presence and role of muc4 in PCA. In collaboration with Henry T. Lynch of Creighton University School of Medicine, he intends to track families at high risk of PCA to determine when pancreatic cells first express muc4. Upon exposing pancreatic cells to carcinogens, he observed that the muc4 gene switched on after about 20 weeks. Batra is also completing papers that report on muc4's role in tumorigenesis and metastasis. "It changes the phenotype of pancreatic cells," he says. As for PTHrP, Bouvet reports that he is applying for grants to study its specificity and sensitivity as a marker in "as many patients with pancreatic cancer as I can test."
A New Look at an Old Nutraceutical
For centuries, various Asian populations have included curcumin, the yellow pigment of turmeric, in their diets. They also have used it to treat inflammation. In recent decades, this compound has caught scientists' attention, racking up more than 560 entries in the PubMed database. Studies of its antimutagenic and anticancer properties have appeared since the early 1980s.
The talk: Last year's AACR meeting included five abstracts on curcumin, according to the association's Web site; this year, the number was 14. The posters depicted curcumin's molecular effects in cell culture, lab animals, and even human blood. In contrast, a talk at a minisymposium (Abstract #3748) outlined a study of cancer patients.
Ricky A. Sharma, a clinical research fellow in the oncology department at the University of Leicester, presented the study that he co-led with William P. Steward, an oncology professor at the university. The subjects were 15 patients with advanced colorectal cancer who already had undergone standard chemotherapy. Each day, they consumed varying doses of curcumin-containing extracts. Encouragingly, five patients exhibited no increases in tumor size and no new metastases for three to four months. How the compound might have affected metastases and portions of a tumor outside the colon is something of a mystery. Poster #115 explained the poor bioavailability of orally administered curcumin in rats.
After the AACR meeting: To account for his study's inability to detect curcumin in plasma or blood cells, Sharma says, "There may be compartmentalization of curcumin within circulating leukocytes, as we have found in colon cells in vitro." He also suggests that the compound might operate therapeutically at unobservably low concentrations. To nail down curcumin's possible therapeutic effects, his group began a study last December that involves more patients and higher doses.
Mack T. Ruffin, meanwhile, offers another view of curcumin's potential. An associate professor of family medicine at the University of Michigan Medical Center in Ann Arbor, he proposes that the compound might be more valuable in preventing cancer than in treating it. He notes that colorectal cancer rates in India and Pakistan, where curcumin is ubiquitous in the diet, are lower than in Western nations. But he acknowledges that many factors might contribute to the lower rates.
Several years ago, Ruffin and Michigan oncology professor Dean E. Brenner conducted a study of curcumin's toxicity on 18 people considered at risk for colorectal cancer because they were older than 50. (The paper has been submitted for publication.) Curcumin is "an ideal product for colorectal cancer prevention since it really doesn't go anywhere other than the gut," says Ruffin. "That minimizes possible toxicities and other adverse effects." After determining the safe maximum dose, he hopes to investigate such questions as how curcumin affects markers associated with early tumor growth.
A Surprise COX-2 Effect
The enzyme cyclooxygenase-2 (COX-2) catalyzes a crucial step in the formation of prostaglandins, which are hormones with many biological functions. The results of several studies suggest that too much COX-2 increases a cell's odds of turning malignant.3 Thus, COX-2 overexpression lowers apoptosis and boosts tumor invasiveness. Conversely, tumorigenesis is inhibited in COX-2 knockout mice. Given the possibility that COX-2 inhibitors might fight cancer as well as reduce inflammation, it's little wonder that they are being touted as the latest miracle drugs.4
The posters: Poster #2265 described a COX-2 experiment with an unexpected outcome. The principal investigators were David K. Bol, a senior research scientist at Bristol-Myers Squibb Pharmaceutical Research Institute in Princeton, N.J. and Susan M. Fischer, a professor of carcinogenesis at the University of Texas M.D. Anderson Cancer Center in Smithville. They generated transgenic mice that overproduced COX-2 under the control of the human keratin-14 promoter, which drives gene expression in epithelial tissues. When their research team exposed the mice to a carcinogen in a classic skin-cancer protocol, it found, to its surprise, that the animals developed far fewer tumors than controls. In another study, however, the transgenic mice spontaneously developed neoplasias, cancer-like growths, in their prostates (Poster #1424).
After the AACR meeting: Using a keratin-5 promoter to overexpress COX-2 in mice, a German group reportedly has observed a lower incidence of skin cancer. Bol offers three possible reasons for his and Fischer's similar findings. He recalls seeing a low-grade skin inflammation in the mice, leading him to suspect that heightened immune-system surveillance was eliminating small tumors and cells on the verge of malignancy. Noting the animals' hair loss and sebaceous-gland abnormalities, he proposes that transgene expression might have altered the stem cells that could be the carcinogen's main target. Finally, he surmises that, given higher levels of apoptosis that appeared to be occurring in the skin, cells might have committed suicide before they began to form tumors.
Because the goal of his company project was to generate mice that develop COX-2-driven tumors, Bol remarks that his follow-up will be limited to examining whether COX-2 inhibitors reverse the transgene's effects. Fischer, however, plans to continue this line of research. Her lab is trying to develop mice "where we can turn COX-2 on and off at different stages," she says.
COX-2-overexpressing mice weren't produced earlier because, under the promoters that were used, the transgene proved lethal during early development, according to Timothy Hla, a physiology professor at the University of Connecticut Health Center in Farmington. His lab recently reported that COX-2 overexpression, driven by the murine mammary tumor virus promoter, causes female mice to develop breast tumors around the time of their third pregnancy.5 To explain why his results differed from Bol's and Fischer's skin-cancer findings, Hla says, "It may be that in different organ systems, prostaglandins have different roles." That's an interesting idea: Perhaps it will be the subject of a poster at next year's AACR meeting in San Francisco.
1. M. Komatsu et al., "Potentiation of metastasis by cell surface sialomucin complex (rat MUC4), a multifunctional anti-adhesive glycoprotein," International Journal of Cancer
, 87:480-6. 2000.
2. M. Bouvet et al., "Human pancreatic adenocarcinomas express parathyroid hormone-related protein," Journal of Clinical Endocrinology & Metabolism
, 86:310-6, January 2001.
3. K. Subbaramaiah et al., "Inhibition of cylooxygenase: a novel approach to cancer prevention," Proceedings of the Society for Experimental Biology and Medicine
, 216:201-10, 1997.
4. P. Smaglik, "Beyond inflammation: blocking COX-2 may provide therapy for multiple diseases," The Scientist, 13:14, April 12, 1999
5. C.H. Liu et al., "Over-expression of cyclooxygenase (COX)-2 is sufficient to induce tumorigenesis in transgenic mice," Journal of Biological Chemistry
, in press.