September 1, 2010 (Vol. 30, No. 15)
K. John John Morrow Jr. Ph.D. President Newport Biotech
Better Diagnosis, Prognosis, and Drug Targeting Are among Potential Benefits
The pace of biomarker development is accelerating as investigators report new studies on cancer, diabetes, Alzheimer disease, and other conditions in which the evaluation and isolation of workable markers is prominently featured. CHI’s “ADAPT” meeting, to be held later this month, will profile various new strategies being developed by both the academic and private sectors.
Wei Zheng, Ph.D., leader of the R&D immunoassay group at EMD Chemicals, is overseeing a program to develop biomarker immunoassays to monitor drug-induced toxicity, including kidney damage.
Although immunohistochemistry has traditionally been looked upon as the gold standard for these studies, this approach is slow and cannot be adapted to multiplexing. “One of the principle reasons for drugs failing during development is because of organ toxicity,” says Dr. Zheng. “This results in proteins being liberated into the serum and urine in abnormal amounts. These proteins can serve as biomarkers of adverse response to drugs, as well as disease states.”
Through collaborative programs with Rules-Based Medicine (RBM), the EMD group has released panels for the profiling of human renal impairment and renal toxicity. These urinary biomarker based products fit the FDA and EMEA guidelines for assessment of drug-induced kidney damage in rats.
Although the FDA has not yet approved kidney toxicity biomarker tests, the industry is clearly moving in that direction, according to Dr. Zheng. “We see this trend, not only for drug-induced organ damage but also for biomarkers for disease states affecting organ function.”
The group recently performed a screen for potential protein biomarkers in relation to kidney toxicity/damage on a set of urine and plasma samples from patients with documented renal damage. Multiplexed immunoassays were used to quantify protein analytes and standard blood and urine chemistries were also measured.
Additionally, Dr. Zheng is directing efforts to move forward with the multiplexed analysis of organ and cellular toxicity. Diseases thought to involve compromised oxidative phosphorylation include diabetes, Parkinson and Alzheimer diseases, cancer, and the aging process itself.
Good biomarkers allow Dr. Zheng to follow the mantra, “fail early, fail fast.” With robust, multiplexible biomarkers, EMD can detect bad drugs early and kill them before they move into costly large animal studies and clinical trials. “Recognizing the severe liability that toxicity presents, we can modify the structure of the candidate molecule and then rapidly reassess its performance.”
Scientists at Oncogene Science a division of Siemens Healthcare Diagnostics, are also focused on biomarkers. “We are working on a number of antibody-based tests for various cancers, including a test for the Ca-9 CAIX protein, also referred to as carbonic anhydrase,” Walter Carney, Ph.D., head of the division, states.
CAIX is a transmembrane protein that is overexpressed in a number of cancers, and, like Herceptin and the Her-2 gene, can serve as an effective and specific marker for both diagnostic and therapeutic purposes. It is liberated into the circulation in proportion to the tumor burden.
Dr. Carney and his colleagues are evaluating patients after tumor removal for the presence of the Ca-9 CAIX protein. If the levels of the protein in serum increase over time, this suggests that not all the tumor cells were removed and the tumor has metastasized.
Dr. Carney and his team have developed both an immunohistochemistry and an ELISA test that could be used as companion diagnostics in clinical trials of CAIX-targeted drugs.
The ELISA for the Ca-9 CAIX protein will be used in conjunction with Wilex’ Rencarex®, which is currently in a Phase III trial as an adjuvant therapy for non-metastatic clear cell renal cancer.
Additionally, Oncogene Science has in its portfolio an FDA-approved test for the Her-2 marker. Originally approved for Her-2/Neu-positive breast cancer, its indications have been expanded over time, and was approved for the treatment of gastric cancer last year. It is normally present on breast cancer epithelia but overexpressed in some breast cancer tumors.
“Our products are designed to be used in conjunction with targeted therapies,” says Dr. Carney. “We are working with companies that are developing technology around proteins that are overexpressed in cancerous tissues and can be both diagnostic and therapeutic targets.”
The long-term goal of these studies is to develop individualized therapies, tailored for the patient. Since the therapies are expensive, accurate diagnostics are critical to avoid wasting resources on patients who clearly will not respond (or could be harmed) by the particular drug.
“At this time the rate of response to antibody-based therapies may be very poor, as they are often employed late in the course of the disease, and patients are in such a debilitated state that they lack the capacity to react positively to the treatment,” Dr. Carney explains.
Nanoscale Real-Time Proteomics
Stanford University School of Medicine researchers, working with Cell BioSciences, have developed a nanofluidic proteomic immunoassay that measures protein charge, similar to immunoblots, mass spectrometry, or flow cytometry. But unlike these platforms, this approach can measure the amount of individual isoforms, specifically, phosphorylated molecules.
“We have developed a nanoscale device for protein measurement, which I believe could be useful for clinical analysis,” says Dean W. Felsher, M.D., Ph.D., associate professor at Stanford University School of Medicine.
Critical oncogenic transformations involving the activation of the signal-related kinases ERK-1 and ERK-2 can now be followed with ease. “The fact that we measure nanoquantities with accuracy means that we can interrogate proteomic profiles in clinical patients, by drawing tiny needle aspirates from tumors over the course of time,” he explains.
“This allows us to observe the evolution of tumor cells and their response to therapy from a baseline of the normal tissue as a standard of comparison.”
According to Dr. Felsher, 20 cells is a large enough sample to obtain a detailed description. The technology is easy to automate, which allows the inclusion of hundreds of assays. Contrasting this technology platform with proteomic analysis using microarrays, Dr. Felsher notes that the latter is not yet workable for revealing reliable markers.
“Microarray studies are not always consistent, and what works in one person’s hands doesn’t always appear to work for another investigator,” he argues. “This means that reliable clinical assays may not be as easily developed using this approach.”
Dr. Felsher and his group published a description of this technology in Nature Medicine. “We demonstrated that we could take a set of human lymphomas and distinguish them from both normal tissue and other tumor types. We can quantify changes in total protein, protein activation, and relative abundance of specific phospho-isoforms from leukemia and lymphoma patients receiving targeted therapy. Even with very small numbers of cells, we are able to show that the results are consistent, and our sample is a random profile of the tumor.”
Splice Variant Peptides
“Aberrations in alternative splicing may generate much of the variation we see in cancer cells,” says Gilbert Omenn, Ph.D., director of the center for computational medicine and bioinformatics at the University of Michigan School of Medicine. Dr. Omenn and his colleague, Rajasree Menon, are using this variability as a key to new biomarker identification.
It is becoming evident that splice variants play a significant role in the properties of cancer cells, including initiation, progression, cell motility, invasiveness, and metastasis. “Of course, one could spend a lifetime fully characterizing the consequences of variation in a single protein by classical biochemical methods,” says Dr. Omenn.
Alternative splicing occurs through multiple mechanisms when the exons or coding regions of the DNA transcribe mRNA, generating initiation sites and connecting exons in protein products. Their translation into protein can result in numerous protein isoforms, and these isoforms may reflect a diseased or cancerous state.
Regulatory elements within the DNA are responsible for selecting different alternatives; thus the splice variants are tempting targets for exploitation as biomarkers.
Despite the many questions raised by these observations, splice variation in tumor material has not been widely studied. Cancer cells are known for their tremendous variability, which allows them to grow rapidly, metastasize, and develop resistance to anticancer drugs.
Dr. Omenn and his collaborators used mass spec data to interrogate a custom-built database of all potential mRNA sequences to find alternative splice variants. When they compared normal and malignant mammary gland tissue from a mouse model of Her2/Neu human breast cancers, they identified a vast number (608) of splice variant proteins, of which peptides from 216 were found only in the tumor sample.
“These novel and known alternative splice isoforms are detectable both in tumor specimens and in plasma and represent potential biomarker candidates,” Dr. Omenn adds.
Dr. Omenn’s observations and those of his colleague Lewis Cantley, Ph.D., have also shed light on the origins of the classic Warburg effect, the shift to anaerobic glycolysis in tumor cells. The novel splice variant M2, of muscle pyruvate kinase, is observed in embryonic and tumor tissue. It is associated with this shift, the result of the expression of a peptide splice variant sequence.
It is remarkable how many different areas of the life sciences are tied into the phenomenon of splice variation. The changes in the genetic material can be much greater than point mutations, which have been traditionally considered to be the prime source of genetic variability.
“We now have powerful methods available to uncover a whole new category of variation,” Dr. Omenn says. “High-throughput RNA sequencing and proteomics will be complementary in discovery studies of splice variants.”
Splice variation may play an important role in rapid evolutionary changes, of the sort discussed by Susumu Ohno and Stephen J. Gould decades ago. They, and other evolutionary biologists, argued that gene duplication, combined with rapid variability, could fuel major evolutionary jumps.
At the time, the molecular mechanisms of variation were poorly understood, but today the tools are available to rigorously evaluate the role of splice variation and other contributors to evolutionary change.
“Biomarkers derived from studies of splice variants, could, in the future, be exploited both for diagnosis and prognosis and for drug targeting of biological networks, in situations such as the Her-2/Neu breast cancers,” Dr. Omenn says.
The peptidome should be a fertile field for isolation of new biomarkers, but progress has been slow in this area. This situation may now be changing, according to Paul Tempst, Ph.D., professor and director of the Protein Center at the Memorial Sloan-Kettering Cancer Center.
“By correlating the proteolytic patterns with disease groups and controls, we have shown that exopeptidase activities contribute to the generation of not only cancer-specific but also cancer type specific serum peptides. So there is a direct link between peptide marker profiles of disease and differential protease activity.” For this reason Dr. Tempst argues that “the patterns we describe may have value as surrogate markers for detection and classification of cancer.”
To investigate this avenue, Dr. Tempst and his colleagues have followed the relationship between exopeptidase activities and metastatic disease.
“We monitored controlled, de novo peptide breakdown in large numbers of biological samples using mass spectrometry, with relative quantitation of the metabolites,” Dr. Tempst explains. This entailed the use of magnetic, reverse-phase beads for analyte capture and a MALDI-TOF MS read-out.
“In biomarker discovery programs, functional proteomics is usually not pursued,” says Dr. Tempst. “For putative biomarkers, one may observe no difference in quantitative levels of proteins, while at the same time, there may be substantial differences in enzymatic activity.”
In a preliminary prostate cancer study, the team found a significant difference in activity levels of exopeptidases in serum from patients with metastatic prostate cancer as compared to primary tumor-bearing individuals and normal healthy controls. However, there were no differences in amounts of the target protein, and this potential biomarker would have been missed if quantitative levels of protein had been the only criterion of selection.
Ironically, many studies of correlations of enzyme activity and cancerous states were carried out long ago, in the 1950s and 1960s. Acidic phosphatase activity was used as a blood-based marker for prostate cancer long before the PSA test was put into general use in the 1980s.
“These older observations on enzymes, including aminopeptidases caught our attention and convinced us that we were on solid ground,” Dr. Tempst states.
It is frequently stated that “practical fusion energy is 30 years in the future and always will be.” The same might be said of functional, practical biomarkers that can pass muster with the FDA. But splice variation represents a new handle on this vexing problem. It appears that we are seeing the emergence of a new approach that may finally yield definitive diagnostic tests, detectable in serum and urine samples.