Feature: From proteomics to biomarkers

By Susan Williamson

Read part I of top down proteomics. Read part II of top down proteomics.

One interesting observation Professor Neil Kelleher’s lab has made has involved the detection of post-translational dynamics of cancer cells in response to intrinsic DNA damage. “We damage the DNA in cells and watch them respond, and over one to four days the cells decide what to do,” says Kelleher.

“The cells can make one of three choices: they can repair their DNA and live; they can enter into apoptosis and die; or they can undergo senescence and live but stop replicating.”

Kelleher thinks they have identified protein isoforms that correlate with each of these decisions the cells make. “Our separation platform involves sIEF [isoelectric focusing in solution], a tubular form of polyacrylamide gel electrophoresis [called GELFrEE for short], and reversed phase liquid chromatography [RPLC], which fractionate proteins according to charge, mass and hydrophobicity, respectively,” says Kelleher.

“With this we have achieved the highest levels of throughput with record-breaking proteome coverage across a range of protein physiochemical properties.”

An analysis of HeLa S3 cells using this 3D platform resulted in about 50 fractions and molecular weights up to 100 kDa in a total separation time of 3.5 hours. These fractions were further separated using capillary-RPLC coupled online to a Fourier-transform mass spectrometer for the direct analysis of intact proteins by tandem mass spectrometry.

Kelleher said the protein isoforms identified are simple in concept and should be more closely related to cellular state or organismal phenotype. For example, if a cell were to decide to continue dividing after undergoing DNA damage, it would produce a particular protein isoform that could potentially serve as a biomarker to specifically detect the status of cancerous cells upon treatment with an anti-cancer drug.

He thinks intact proteins have the potential to form much better biomarkers for disease than the biomarkers that are currently available.

“Take the prostate specific antigen (PSA) as an example,” he says. “This is a complex protein with many different isoforms and glycoforms etc, and the antibody used in the PSA test recognises only one piece of this protein.”

Not surprisingly the PSA test is relatively unreliable as far as diagnosing prostate cancer goes. A recent meta-analysis of six trials on PSA screening showed that screening resulted in a 46 per cent increased likelihood of men being diagnosed; however, this diagnosis had no significant impact on all-cause or disease-specific mortality.

“If you could measure all the different PSA isoforms, this would correlate more tightly to the disease phenotype,” he mused.

Kelleher was recently elected to the council of the Human Proteome Organisation (HUPO), the international scientific organisation that represents proteomics.

He maintains that the top down approach should be part of the human proteome project and is looking forward to generating more awareness about this technique on an international platform. He is also looking forward to some summer sun down under in February after a sleety winter in Chicago.