Immunodepletion explored: biomarker discovery assisted by increased depletion

Unveiling biomarkers in clinical samples: disappointing results to date

The relentless search for biomarkers of disease has seen hundreds of reports of potential candidates identified from proteomics studies but there is a frustration in some quarters regarding the lack of progress thereafter. As pointed out by a team of researchers in the UK, there are "very few biomarkers with strong enough clinical performance to suggest that they will contribute to changing clinical practice."

That is not to suggest that the potent biomarkers do not exist, rather that they have not yet been unearthed. The discovery process for human serum or plasma is hindered by the presence of highly abundant proteins which mask many low abundant species that might be of value.

A number of strategies exist for bypassing the abundant proteins but one of the simpler techniques is immunodepletion. It is well established and several commercial columns are available for removing different numbers of proteins from biological fluids.

As pointed out by the aforementioned research team, there have been very few comprehensive studies on the power of immunodepletion columns. So, they have undertaken their own wide-ranging examination of the depth of serum proteome coverage, reproducibility and specificity of three different columns.

Immunoaffinity column comparison

Rosamonde Banks and colleagues from the Leeds Institute of Molecular Medicine at St. James's University Hospital, and Jenny Ho from Thermo Fisher Scientific, Hemel Hempstead, compared the Multiple Affinity Removal System MARS6 and MARS14 columns and the ProteoPrep20 column, which remove 6, 14 and 20 abundant proteins, respectively.

The columns were tested using pooled sera from 17 patients who had undergone renal transplant. The flow-through proteins representing depleted serum were analysed separately from the fraction containing the bound proteins.

The proteins in each fraction were analysed intact in a top-down approach by 2D differential gel electrophoresis using fluorescent dyes and image analysis. Any protein spots of interest were cut out for in-gel digestion with trypsin and LC/MS/MS analysis on a quadrupole-time-of-flight mass spectrometer.

For the bottom-up approach, the proteins in each fraction and from undepleted serum were in-solution digested with trypsin and analysed by LC/MS/MS on a high-resolution mass spectrometer.

In both cases, the peptides were searched against the International Protein Index human protein database, using at least two unique peptide identifications for protein confirmation. In some cases, one unique peptide was allowed to illustrate the depth of coverage.

Extensive immunodepletion with bottom-up approach is best for biomarkers

As expected, the amount of proteins in the flow-through fraction decreased with increasing depletion. The mean total recoveries were 11.2, 7.8 and 4.5% for the MARS6, MARS14 and Prot20 column, respectively, with corresponding CVs of 6, 10 and 16%.

The bulk reproducibilities of depletion were consistent at 25.3-28.4% and 28.5-36.2% for flow-through and bound fractions, respectively. However, the depletion efficiency was far from 100%. Mass spectrometry revealed the presence of unique peptides from all of the depleted proteins except for transferrin on MARS6.

For the bottom-up method, the total number of proteins identified in the depleted serum increased with the degree of immunodepletion. For identifications based on at least two peptides, the figures were 159, 234, 272 and 301 proteins in raw serum, MARS6, MARS14 and Prot20 depleted samples, respectively.

When the proteins identified using only one peptide were included, these figures rose to 266, 361, 405 and 442. This strategy is acceptable due to the high resolving power of the mass spectrometer which determines peptide masses accurately.

Although there was considerable overlap between the proteins identified across the four levels of depletion, the MARS14 and Prot20 samples each contained some unique proteins: 19 and 57 for two peptides and 25 and 60 when including one peptide.

Many proteins known to have low-to-moderate abundance in human serum were identified as proof of principle, including the known renal injury markers NGAL and FABP4. The number of identified cytoplasmic and membrane proteins rose with increasing immunodepletion at the expense of extracellular proteins.

It is one thing to be able to remove the abundant proteins from serum, but it is quite another to do it reproducibly. But all three columns gave acceptable reproducibilities, in line with other reports.

The performance of the top-down approach using 2D differential gel electrophoresis was less successful, despite the improved reproducibilities compared with the bottom-up method. The main problem was the detection of few proteins of interest in the increasingly depleted samples, many of the extra proteins revealed simply being additional isoforms.

So, the results "support the use of greater levels of immunodepletion (14 or 20 abundant proteins) in potentially revealing biomarker candidates in serum-based clinical discovery experiments when coupled with relatively high-throughput but high mass accuracy LC/MS/MS." The major difference between the columns was found moving from 6-14 depleted proteins, with smaller changes from 14-20.

The researchers regarded the experiments as more valuable than many published reports on immunodepletion because they were carried out on real pathological samples rather than healthy control samples.

Now, they are evaluating the potential of the 14-protein immunodepletion approach in carefully designed experiments which address specific clinical questions, to determine whether biomarkers can be discovered this way in practice.

The views represented in this article are solely those of the author and do not necessarily represent those of John Wiley and Sons, Ltd.