William Check, PhD
The symposium in which George Netto, MD, participated at the 2009 American Society for Clinical Pathology meeting—“Immunohistochemistry-Molecular Interface: Diagnostic, Theranostic, and Genomic Applications”—was in synch with a new direction in molecular pathology. “We are seeing an influx of people interested in acquiring further knowledge in molecular diagnostics of solid tumors,” says Dr. Netto, associate professor of pathology, oncology, and urology in the Department of Pathology at Johns Hopkins Medical Institutions, judging from last year’s meeting of the Association for Molecular Pathology. Genetic markers have been used in hematopoietic neoplasms for some time, but the trend toward their widespread use in solid tumors is relatively recent. The strong attendance at the solid tumor plenary session on the Cancer Genome Atlas and workshop sessions on solid tumor markers at the AMP meeting was one manifestation of this sea change.
At the ASCP session, speakers discussed the use of molecular biomarkers in breast, lung, colon, genitourinary, and head and neck cancers, showing that both immunohistochemistry and molecular assays play a role in this complex diagnostic field. Breast, lung, and colon cancers were the subject of part one of this article (“Whole new world for IHC, molecular”).
Speaking of biomarkers in genitourinary cancers, Dr. Netto told attendees: “This field is not as mature as those of us with an interest in urologic malignancies would like it to be.” He explained to CAP TODAY later that he was referring mostly to theranostic applications in GU tumors. “I meant that in comparison to the tumors already discussed—breast, colon, and lung—in urologic malignancies we don’t yet have well-established tests to determine who can be put on a certain targeted agent or who is in need of a more aggressive treatment approach based on well-validated molecular prognosticators,” Dr. Netto says.
However, several promising biomarkers for detection, prognosis, and targeted therapeutics are now under evaluation. For prostate cancer, Dr. Netto focused on TMPRSS2-ERG, a fusion gene resulting from a genetic rearrangement between two genes on chromosome 21: TMPRSS2 (Transmembrane Protease, Serine 2) and ERG (ETS Related Gene). Transcription factors encoded by the ETS family of genes are central elements in tumorigenesis. In 2005, ERG, a member of the ETS transcription factor family, was found to be the most consistently overexpressed oncogene in prostate cancer cells (Petrovics G, et al. Oncogene. 2005;24:3847–3852). A few months later, capitalizing on the previous discovery, Arul Chinnaiyan’s group at the University of Michigan was able to show that ERG overexpression was driven by the TMPRSS2-ERG rearrangement, thus providing an androgen-driven mechanism for ERG overexpression in prostate cancer (Tomlins SA, et al. Science. 2005;310:644–648). Since that major discovery, several studies have linked the presence of TMPRSS2-ERG fusion to disease outcome. However, Dr. Netto cautions, “There are conflicting studies whether TMPRSS2-ERG is prognostically useful or not, given that recent large cohort studies have failed to show prognostic significance for the presence of the fusion. So this remains to be seen.” Dr. Netto and his colleagues at Hopkins conducted one of these studies not replicating the prognostic value of the gene fusion. Their finding: “Only chromosome 21 polyploidy regardless of TMPRSS2-ERG rearrangement was associated with increased risk” of recurrence.
On the other hand, Dr. Netto adds, “the fusion has the potential to be exploited diagnostically. The lack of specificity of serum PSA as a marker of clinically significant prostate cancer has fueled the search for better detection markers. Many believe that our current serum PSA-based screening biopsy strategy has led to detecting too many clinically insignificant prostate cancers.” If the detection of TMPRSS2-ERG fusion in urine samples is shown to be a more specific marker for aggressive prostate cancer, he says, a urine TMPRSS2-ERG-based molecular test could become a valuable addition to current detection strategies.
Other cancer-associated molecular alterations that could be detected in urine specimens are epigenetic changes of gene promoter hypermethylation. A group at Hopkins was able to show that GSTP1, APC, RASSF1a, PTGS2, and MDR1 genes were hypermethylated in up to 85 percent of prostate cancers examined (Yegnasubramanian S, et al. Cancer Res. 2004;64:1975–1986). A panel of these markers could distinguish benign from malignant prostate tissue with a very high degree of accuracy. Therefore, “analysis of the methylation status of a panel of several genes in a urine sample can be potentially used as a diagnostic test to help us detect prostate cancers of clinical significance,” Dr. Netto says.
Finally, Dr. Netto described work on the PTEN/PI3K/mTOR (mammalian target of rapamycin) pathway. “There are many studies showing the important role of mTOR pathway in cell growth, proliferation, and oncogenesis in prostate cancer,” he told CAP TODAY. PTEN is a negative regulator of this pathway. Loss of PTEN tumor suppressor gene activity and the ensuing mTOR pathway gene expression signature appear to be associated with poor prognosis in prostate cancer (Corcoran NM, et al. BJU Int. 2006;97:1149–1153). The mTOR pathway is also a potential target for prostate cancer treatment.
Dr. Netto and coworkers at Hopkins recently completed a collaborative pilot study in association with Duke University and the University of Michigan, in which patients with high Gleason grade prostatic adenocarcinoma were given neoadjuvant rapamycin treatment before radical prostatectomy. The aim of the safety and pharmacodynamic trial was to demonstrate that by inhibiting mTOR activity, rapamycin will “readjust” the pathway activation due to loss of PTEN. Indeed, at prostatectomy, rapamycintreated subjects had a significant change in a marker of drug activity (lower phos-S6 levels, measured by immunohistochemistry) relative to biopsy specimens, while untreated control subjects showed no change. “This type of study demonstrates that we can use IHC as a surrogate marker to see whether a drug is doing what it’s supposed to do at the molecular level—to rectify a defective pathway,” Dr. Netto says. “Analogs of rapamycin are being developed that might be clinically effective.”
Bladder cancer is the other urologic malignancy Dr. Netto discussed at the ASCP symposium. He pointed to the parallelism between the two distinct phenotypes of bladder cancer (superficial and muscle invasive disease) and their two divergent underlying genetic pathways, saying, “The two clinicopathologic phenotypes appear to be a reflection of two somewhat distinct oncogenic pathways.” Fortunately, most (up to 70 percent) bladder neoplasms fall into the superficial group. Noninvasive bladder cancer is more frequently associated with genetic alterations in the FGFR3 and HRAS gene pathways, Dr. Netto says. In contrast, muscle invasive urothelial cancers are associated with mutations in the P53 and RB (retinoblastoma) genes and tumor environment and angiogenesis molecular alterations. The two distinct genetic pathways can be “exploited,” he says, in identifying prognostic and therapeutic markers for bladder cancer. The current prognostication approach based on clinicopathologic parameters such as stage, grade, and tumor size can be refined with the addition of IHC panels of p53, Rb, and cell cycle markers expression. Furthermore, currently available therapeutics directed at tyrosine kinase receptor pathways or angiogenesis pathways, or both, may be of utility in treating bladder cancer and are being evaluated in clinical trials.
At the early detection and surveillance front, Dr. Netto pointed to the utility of molecular assays such as the UroVysion FISH assay and gene promoter methylation-based assays under evaluation. Several studies have pointed to the improved sensitivity and specificity of the FISH-based UroVysion assay over urine cytology alone, he says.
Dr. Netto believes it is crucial for the molecular diagnostic pathology community to be a visible partner in the design and implementation of ongoing genomic-based diagnostic and theranostic studies validating new genomic biomarkers and targeted therapeutic agents. He says, “We are best positioned to help determine what can trickle down from the complex wealth of genomic data to help identify which of these alterations in solid tumors can ultimately be used as a well-validated companion test, and hopefully influence which technologies are best used to detect the genetic alteration at the tissue/specimen level.”
Jennifer L. Hunt, MD, associate professor of pathology at Harvard Medical School and associate chief of pathology and director of quality and safety at Massachusetts General Hospital, spoke at the ASCP symposium about applying biomarkers to head and neck cancers. She described several types of molecular oncogenesis found in various head and neck tumors: oncogenes in mucoepidermoid carcinoma and papillary thyroid carcinoma; tumor suppressor genes in parathyroid carcinoma; and a virus (human papillomavirus, or HPV) in oral squamous cell carcinoma. (Nasopharyngeal carcinoma is also caused by a virus, Epstein-Barr, but that is well known, so Dr. Hunt didn’t include it.)
In mucoepidermoid carcinoma, the most common malignant salivary gland tumor, variant morphologies seen in fine-needle aspirates can make diagnosis and grading difficult. A translocation—t(11;19) (q21; p13)— produces a mutant fusion protein that disrupts the Notch signaling pathway. This translocation is found in 60 percent to 70 percent of low- and intermediate-grade tumors but is much less prevalent in high-grade tumors. “We don’t understand this distinction,” Dr. Hunt told CAP TODAY.
Nonetheless, the translocation may be useful for grading mucoepidermoid carcinoma, which is essential to prognosis and thus extent of therapy. FISH and RT-PCR are both options to detect the translocation. “What is most practical in routine paraffin sections will probably be FISH-based assays,” Dr. Hunt says. She does not do this test, explaining that it is “at an early stage and only being done at a few places and in the research settings.”
Papillary thyroid carcinoma (PTCa), Dr. Hunt noted, is a common tumor and has “an excellent prognosis” with appropriate treatment, including surgery and postoperative radioactive iodine. However, a small subgroup of patients will have aggressive or radioactiveiodine-resistant PTCa that is more difficult to treat. Recent discovery of the oncogene BRAF in some PTCa samples may provide future therapeutic targets for treating these particular tumors. “The literature does suggest that tumors with BRAF mutations are associated with more aggressive features,” such as extrathyroidal extension and lymph node metastases, Dr. Hunt says. “There is not yet total agreement on whether BRAF is associated with poor survival, though.”
Current therapy for PTCa rests on the twin pillars of total thyroidectomy and radioactive iodine ablation with followup monitoring. In the future, Dr. Hunt hypothesized, patients with PTCa may be assessed with a panel of markers, such as BRAF and RAS, and treatment for radioactiveiodine-resistant tumors might include molecular targeted therapy, such as inhibitors of RAF and related proteins. She said this approach is “not far off”: Trials of targeted therapy in thyroid cancer are ongoing.
Detection of the BRAF mutation in PTCa is most commonly done via sequencing to detect the mutant gene directly. Dr. Hunt says the laboratory at Massachusetts General does not perform this test on routine cases, but it has it available using direct gene sequencing on an automated capillary electrophoresis platform. Despite the $100,000 to $250,000 price tag of sequencing instruments, she called sequencing “a practical approach to look for many different oncogenes, even in high-volume private practice laboratories.”
Seeking oncogenes may become routine in molecular laboratories, particularly oncogenes in the common tumor-associated RAS-RAF-MEK-ERK-MAP kinase pathway. “This is a tumor pathway that will become increasingly important to us in the practice of pathology,” she says. “As pathologists, we need to know about this pathway because mutations in this pathway are present in a variety of different human carcinomas.”
In contrast to papillary thyroid carcinoma being caused by oncogenes, parathyroid cancer is associated with a tumor suppressor gene, HRPT2. This gene was first discovered because of its association with Hyperparathyroidism-Jaw tumor syndrome, which features parathyroid cysts and carcinomas and fibro-osseous lesions of the jaws. Sporadic parathyroid carcinomas also appear to harbor mutations in this gene. “After mutations in the gene were discovered to be the basis of the syndrome, it was found also to be involved in sporadic cases of parathyroid carcinoma,” Dr. Hunt says (Shattuck TM, et al. N Engl J Med. 2003;349: 1722–1729). Dr. Hunt described parathyroid carcinoma as “a rare but challenging tumor to diagnose.” While the genetic lesion can be sought directly by FISH or PCR, most cases will also have loss of the protein product, parafibromin, which can be measured by IHC. Though all of these different assays have pros and cons, Dr. Hunt’s advice is to “use what you have in your lab.”
Finally, HPV has been found in certain morphologic subtypes of oral squamous cell carcinoma (SCCa)—in the oropharynx rather than the oral cavity and with basaloid and nonkeratinizing SCCa. When found, viral presence denotes better prognosis. Detection of HPV can be done by PCR or in situ hybridization. “PCR is not a very practical tool” in this situation, Dr. Hunt says, because it may be overly sensitive and because the virus cannot be easily localized to the tumor. She prefers ISH, which shows the cellular localization and can even potentially identify integrated virus and copy number. Dr. Hunt cites several remaining unanswered questions: “Is HPV subtyping necessary? Which cases should have routine HPV testing? And, can p16 be a reliable surrogate marker for HPV?”
William Check is a medical writer in Wilmette, Ill.