Larger Profiling Studies
“The field of affinity proteomics is still limited by the availability of high-quality validated antibodies,” says Jochen Schwenk, Ph.D., associate professor, Science for Life Laboratory, KTH—Royal Institute of Technology. His team relies on antibodies from the Human Protein Atlas.
The goal of the Human Protein Atlas is to systematically explore the human proteome using an affinity-based approach. At the time this article was written, the Atlas contained 15,598 antibodies targeting 12,238 genes. Antibodies are developed in a high-throughput manner against recombinant protein epitope signature tags, selected based on their low homology to other proteins.
“To retrieve these epitopes for protein profiling, we heat-denature biological samples in a controlled manner. The heat treatment exposes the epitopes that may be otherwise hidden by native protein conformations. This makes recognition of some protein targets more efficient,” says Dr. Schwenk, who also gave a presentation at the Select Biosciences conference in Edinburgh. “This feature enables us to use crude unfractionated blood and urine samples for biomarker discovery.
“We generally utilize bead arrays, where antibodies are coupled onto beads with distinctive color codes. Bead arrays have notable advantages over planar arrays. Beads can be mixed to achieve any desired combination of antibodies. The capture process can be analyzed by flow cytometry with dual laser, where one of the lasers reads the bead identity, and another analyzes interaction with captured protein target.”
This format proved to be highly sensitive, with only 10 µL of plasma required for analysis of 10,000 target proteins, according to Dr. Schwenk. The process lends itself well to larger-scaled profiling studies within disease proteomics, with hundreds of thousands of assays performed in a single day.
The team recently described human fibulin-1 as a candidate marker for renal impairment. High-throughput profiling of samples from patients with renal impairment and from controls revealed significant differences in levels of fibulin-1. The team is validating this marker using other clinical assays.
“Antibodies derived by Human Protein Atlas can also be used for high-throughput biomarker analysis via immunohistochemistry (IHC) on tissue microarrays,” states William Gallagher, Ph.D., associate professor of cancer biology, school of biomolecular and biomedical science, University College Dublin, and CSO of OncoMark.
OncoMark is focused on tissue-based biomarkers as targets for drug development. Starting with primary tumor resections represented in an array format, Dr. Gallagher’s academic group developed a signature consisting of three proteins that predict the probability of cancer recurrence.
“As we can analyze hundreds of tissue samples at the same time, we can fast-track development of biomarkers,” continues Dr. Gallagher, who provided a detailed description of his research at the Edinburgh meeting.
“For example, we found that patients’ outcome, and potentially their response to chemotherapy, is regulated in a significant way by immune cell infiltration into tumors. Moreover, we were able to identify specific macrophage populations that may present novel targets for antitumor therapies.”
The key enabling technology is OncoMark’s IHC-MARK software, which can recognize the specific morphological features of tumor cells and automatically quantify levels of biomarkers in these malignant cells. OncoMark used this approach to investigate the roles of novel candidate biomarkers in the progression of malignant melanoma for a European Union-funded program, Target-Melanoma.
The technology has also proven to be valuable in a more traditional diagnostic application. The single most important assessment for any breast cancer patient is the hormone receptor levels, determined by immunocytochemistry. The response to tamoxifen therapy is predicted by the percentage of cells positive for the estrogen receptor. Actual scoring of positive cells is very subjective and can lead to a large number of false positives and false negatives.
IHC-MARK’s objective scoring accurately predicted response to hormonal therapy in a randomized controlled trial involving over 500 patients, Dr. Gallagher reports. OncoMark is currently combining the detection capabilities of IHC-MARK with novel biomarker content, particularly indicators of response to both classical chemotherapy and molecularly targeted agents.
Due to the complexity of the human proteome, large dynamic range in protein concentrations, and changes of protein abundance due to minor physiological stimuli, processing of samples for proteomic analysis is a challenging task.
“Over a period of 10 years, we have systematically developed methods for protein extraction from tissues that provide maximum protein solubilization and preserve protein native conformation,” says Joerg Hoheisel, Ph.D., head of functional genome analysis, German Cancer Research Center.
“We found that depletion of high-abundance proteins and other types of fractionation led to co-depletion of minor proteins and, therefore, introduced a strong bias in protein representation,” explains Dr. Hoheisel, one of the key presenters at the U.S. HUPO Conference on the Future of Proteomics in March in San Francisco.
“Our goal was to create a set of reproducible conditions for protein analysis by antibody arrays using unfractionated body liquids and tissue samples. An entirely new protein purification and hybridization strategy had to be worked out for the antibody arrays.”
Once the appropriate conditions were optimized, Dr. Hoheisel’s group was able to establish distinct pancreatic cancer signatures by analyzing nearly 900 samples. Another signature forecasts the recurrence of bladder cancer.
Similar to cDNA microarrays, antibody arrays are analyzed by two-color fluorescent assay. Moreover, in certain conditions only one dye molecule binds one protein molecule. The team adapted already existing single-molecule detection techniques to visualize individual affinity capture events.
Selective illumination excites fluorophores only in a small region of the specimen immediately adjacent to the glass-water interface. Such a readout allows for easier discrimination between fluorescence arising from captured molecules and background caused by light scattering at the surface.
Using this approach, as few as 600 molecules per spot could be detected and counted. The team applied this method to develop a proof-of-concept diagnosis of tuberculosis based on capture and detection of lipoarabinomannan, a polysaccharide marker of tuberculosis. Current tuberculosis diagnosis is not sensitive enough to detect the infection at early stages when the treatment may be the most effective.
With the rapid re-emergence of tuberculosis worldwide, a reliable and sensitive diagnosis is of critical importance for disease control. Antibody capture combined with fluorescence detection seems to be well positioned to address this growing need.