August 2015, Vol 4, No 4
New Methods for Surveillance and Preclinical Modeling Highlight Cancer Biomarker Research
Several cancer biomarkers are in preclinical development and on the verge of commercialization, including an inexpensive method to transport blood that preserves RNA stability and a mouse avatar model for translational cancer, said presenters at PMO Live 2015.
In the case of the patient-derived xenograft (PDX), or mouse avatar, the PDX model can be used to explore the boundaries of therapeutics when such is not possible in cancer patients.
Jerald Radich, MD, spoke of novel techniques for cancer diagnosis, monitoring, and treatment guidance. One is the use of a novel 7-gene score to stratify prognosis in older patients with acute myeloid leukemia (AML) (J Clin Oncol. 2014;32:548-556). Expression of the 7 genes (CD34, RHOC, SCRN1, F2RL1, FAM92A1, MIR155HG, and VWA8) with promoter differentially methylated regions correlated with 3-year disease-free and overall survival.
In looking at different validation sets in different study groups, “there’s a stunning difference in expression of these genes, and the number that are aberrant, and survival,” said Radich, Director of the Molecular Oncology Laboratory, Fred Hutchinson Cancer Research Center, Seattle, WA. The same risk stratification also applies to younger AML patients, he said.
Next-generation DNA sequencing can generate hundreds of billions of nucleotides of DNA sequence, but an error rate of about 1 per 1000 results in numerous sequencing mistakes. Duplex sequencing (Proc Natl Acad Sci U S A. 2012;109:14508-14513) is a novel method that removes errors by separately amplifying and sequencing the 2 DNA strands, explained Radich. “The error rate of duplex sequencing is about 1 in 107,” he said. “This technology is now being used to develop multiplex assays for all of the AML mutations, with the idea that you can do a panel with this kind of duplex sequencing and be able to use it in lieu of even flow cytometry to determine minimal residual disease.”
Access to such advanced technology, however, is lacking in low-resource settings, said Radich. Disease monitoring can be accomplished by “home brew” polymerase chain reaction (PCR) at local institutions, but this requires highly technical training and expensive machines; by user-friendly PCR technology such as Cepheid, which uses expensive machines and cartridges; or by shipping blood to standardized centers, which can also be expensive and risks compromising samples.
An alternate inexpensive method of monitoring patients with chronic myelogenous leukemia (CML) uses direct spotting of 4 50-µL samples of peripheral blood onto a paper template (dried blood spots), which can be mailed anywhere in the world without specialized shipping requirements, for RNA extraction and Bcr-Abl testing. Mutations in the Bcr-Abl kinase domain may confer resistance to tyrosine kinase inhibitors in patients with CML. Dried blood spots shipped from Australia by mail preserved RNA stability, with a concordance of 0.97 to fresh peripheral blood specimens, allowing for highly accurate Bcr-Abl detection, he said.
Finally, Radich described an equipment-free gene amplification technique known as loop-mediated isothermal amplification, which provides rapid amplification (<1 hour) under isothermal conditions employing 6 primers spanning 8 distinct sequences of a target gene. The technology has the potential for diagnosis and surveillance in low-resource countries.
Philip C. Mack, PhD, followed with a discussion of the merits and advantages of a PDX system that is “fast being recognized as potentially the preeminent preclinical modeling system for cancer therapeutics.”
A PDX is a piece of human tumor that has been implanted into a mouse without any intervening cell culturing involved, making it fully representative of a human tumor, said Mack, Director of Molecular Pharmacology, University of California Davis Comprehensive Cancer Center.
The technology enabling PDX is the NSG mouse, which is profoundly immune deficient, lacking mature T cells, B cells, and functional natural killer cells and consequently has a high take rate for xenotransplantation. When a human tumor fragment is implanted into the mouse, the tumor grows and is implanted into a new generation of mice. The model is considered established upon tumor expansion at this point.
More than 130 of these models have been successfully implanted in various tissue types, and the principal driver mutation of the patient’s tumor has been preserved, he said.
The PDX program has 2 goals:
- A generalized preclinical modeling system for the purpose of testing therapeutic strategies
- Patient-specific predictive modeling (avatars).
The timing of the 2 goals differs in that the preclinical modeling system can be used to evaluate new drugs in a reasonable time frame, whereas use as a predictive avatar is more difficult and expensive.
“We’ve been taking our PDX models and using them for therapeutic development,” said Mack. A cancer patient with an accessible tumor undergoes excision or biopsy on which molecular profiling is done at the same time the tumor is implanted into NSG mouse models. The patient is treated with a molecularly selected regimen, whereas the mice are treated with innovative regimens also guided by molecular profiling, and outcomes from the 2 are compared.
“We use PDX models to explore the boundaries of therapeutics that we simply cannot do with cases,” he said. “We can try those combinations that cannot be used in clinical trials because they’re represented by such small subsets of patients that it’s not feasible to do. We can treat cohorts in the mice and look at their tumor kinetics. In addition, it’s very important to do tumor pharmacodynamics to see what the drug is actually doing in the tumor and whether or not we are inhibiting the target,” said Mack.
In addition, the effect of the drug on alternate pathways and feedback loops can be gauged.
Research into molecular biomarkers is bringing powerful insights to the identification of disease with considerable potential for prevention and treatment, but current regulatory paradigms may not be adequate to address this new potential, said Andrew Stainthorpe, PhD, at PMO Live 2015. Stainthorpe, head of the National Health Services Research Development [ Read More ]
Molecular classification of diffuse large B-cell lymphoma (DLBCL), the most common type of non-Hodgkin lymphoma, is beginning to influence treatment selection and hopefully will improve outcomes, according to presenters at an education session. Cell-of-origin (COO) identification by genetic analysis has enabled classification of DLBCL into 2 subtypes that have different [ Read More ]