NEW YORK (GenomeWeb) – University of Victoria researchers have developed a MALDI mass spec assay for measuring levels of the proteins AKT1 and AKT2 in breast and colorectal cancer tissue.
Described in a paper published last week in Analytical Chemistry, the assay provides a potential tool for identifying cancer patients likely to respond to treatment with inhibitors targeting the PI3K/AKT/mTOR pathway and demonstrates the suitability of MALDI mass spec for measuring low-abundance proteins in small clinical samples, said Christoph Borchers, head of the U Vic-Genome BC Proteomics Centre and senior author on the paper.
Borchers added that from a clinical perspective the assay is also notable in that it was developed on Bruker's Microflex instrument, which is used in the company's MALDI Biotyper clinical microbiology platform. That platform has won broad uptake from hospital and other clinical labs, which positions it well for additional clinical proteomics assays, Borchers suggested.
"The Microflex is not the most sensitive MALDI instrument, but we wanted to develop something that ideally we could implement in the clinical chemistry lab, and so we wanted to have an instrument that is already accepted in the clinic," he said.
Borchers and his colleagues developed the assay in response to a request from the pharma firm AstraZeneca, which like many drug companies is developing AKT inhibitors for cancer treatment, as well as inhibitors targeting other members of the PI3K/AKT/mTOR pathway, which is dysregulated in several cancer types.
Typically, patients are selected for treatment with these inhibitors based on genomic data, but, Borchers said, while the proteins PI3K and PTEN often harbor mutations, AKT itself is not usually mutated in cancer patients. Immunohistochemistry is commonly used for measuring expression of cancer markers at the protein level, but the technique can suffer from non-specificity of antibodies and is only semi-quantitative.
Mass spec, on the other hand, has the potential to offer highly accurate and specific clinical measurements of cancer protein markers, and Borchers and his colleagues have been active in developing targeted protein assays using multiple-reaction mass spec on triple quadrupole instruments, which is the mass spec approach most commonly used in the clinic.
For the AKT work, however, the researchers chose to use MALDI, believing that the technique's more streamlined sample prep would make for a simpler, higher-throughput assay more amenable to clinical use.
"The question from pharma was, can you build an assay that determines both the expression level and phosphorylation level of [AKT]," he said. "And we wanted to have a very robust and fully automatable [assay]."
To reach the required level of sensitivity using conventional MRM mass spec methods, the researchers would have had to use nanoflow liquid chromatography for their upfront separations, which is a notoriously finicky technology.
"So we said, 'Let's try to get away from MRM,'" Borchers said. MALDI, on the other hand, "is very robust and can be fully automated," he said, "and so that is how we started."
Sensitivity was also a challenge for MALDI, however, especially given the small size of the clinical samples Borchers and his team plan to work with.
"We need to analyze fine needle biopsy samples, which are in the range of 100 micrograms to maybe 300 micrograms of total protein," he said. "So that is not much, and then if you are looking for something that is not well expressed, you need in the end sensitivity in the attomole range."
To get there, the researchers applied a variety of techniques including upfront antibody-based enrichment of the target peptides along with a wash step of the MALDI matrix spots that upped sensitivity by roughly ten-fold.
In all, development of the assay took around two years, Borchers said. "It took quite a while to optimize, but it is working now and it is very robust and we are achieving in many cases [coefficients of variation] of below 5 percent."
The Analytical Chemistry paper covered only development of an assay to AKT1 and AKT2 expression levels, but Borcher said his team has since developed it further to include information on phosphorylation levels of these proteins, which is linked to their activity.
By measuring the proteins before and after a dephosphorylation step, they can assess the levels of phosphorylated versus non-phosphorylated AKT.
"We have applied this to real samples and analyzing tumor tissue and adjacent normal tissue we have seen that phosphorylation levels in AKT jumps from, say, 5 percent [in the normal tissue] to 40 percent in the cancerous region," Borchers said.
They have also expanded the assay to include PTEN and PI3K, he said.
Having demonstrated the assay's performance and sensitivity, the researchers will now apply it to the pharma questions that originally inspired the development work. The plan is to use the assay to measure protein expression and phosphorylation levels in patients undergoing treatment with AKT inhibitors and to correlate this data with clinical outcomes.
"These [patient] samples are very precious, and so we had to first show pharma that the assay was solid," Borchers said.
He also plans to use the assay to guide therapy in patients at McGill University's Jewish General Hospital, with which he is affiliated.
"I am working with [JGH's] Segal Cancer Centre, where we have many cancer patients whose tumors are analyzed on the genomics level, and now we want to add proteomics in hopes that we can better determine the best possible therapies," Borchers said. "One idea in particular is to look at the PI3K/AKT/mTOR pathway and the expression and phosphorylation levels of these proteins and hopefully draw some correlation between [that] and clinical outcomes along with the genomics data."
He added that he and his colleagues have several grant proposals to support this work under review.