Would you prefer multiple reaction monitoring or antibodies with your biomarker validation?

The past several years have seen the proteomics community working fervently to discover proteins that can be used as biomarkers of diseases such as cancer, neurological disorders and others. While the scientific community can debate the relative success of this continuing endeavor, the lack of clinically useful protein biomarkers that have been reported from these efforts has not prevented investigators from looking forward toward developing methods that are necessary to verify the clinical utility of a biomarker. As wonderfully described by Steve Carr, the four phases of producing a novel protein biomarker are discovery, qualification, verification and validation [1]. One of the most striking features as the process moves from discovery to validation is the inverse relationship between the number of analytes and samples that are measured. While the discovery phase may require only tens of samples to be scanned for thousands of analytes, the validation phase will require only a few analytes to be measured in thousands of samples. During the validation phase, clinical assay development may proceed concurrently.

Mass spectrometry (MS), combined with a variety of sample preparation and separation techniques, has been the driving force in the biomarker discovery phase [2–4]. The question is how big a role MS will play in the qualification, verification and validation phases. Conventional thinking suggests that antibodies (or other types of affinity reagents) provide the most expedient means of validating the usefulness of a potential biomarker. Antibody-based tests such as immunohistochemistry (IHC) and ELISA are routinely performed, and a wide range of scientists feel comfortable in the results obtained using these methods. There is, however, another resurgent method that is being touted as an alternative to antibody-based tests, multiple reaction monitoring (MRM)-MS. This method uses triple quadrupole MS to specifically target and quantify proteins of interest in a complex mixture. Ions corresponding to the mass of the biomolecule of interest are isolated within the mass spectrometer. The molecule is then fragmented and a selection of these products is monitored and used for identification and quantitation purposes. The MRM-MS approach provides exquisite specificity and does not suffer from issues related to cross-reactivity with other molecules. If designed properly, only the targeted molecule will be measured by the mass spectrometer. In addition, the method does not require a highly specific antibody (or antibodies) and permits absolute quantitation through the addition of known quantities of isotopically labeled molecules to the mixture. The increasing popularity of MRM-MS was evident at the past 2008 meeting of the American Society of Mass Spectrometry (ASMS) by the number of oral and poster presentations that highlighted this method.

The question that one has to answer is: will MRM-MS replace antibody-based testing in the validation of biomarkers? Based on many of the presentations at the past ASMS meeting, one would think that that is the case. However, a number of presentations simply highlighted the advantages of MRM-MS by focusing on the disadvantages of using antibody-based methods. These disadvantages included the lack of suitable, high-specificity antibodies and the cost in both time and dollars in generating an antibody when one is not available. The MRM-MS approach was presented as being reagent-independent (except for stable-isotope labeled internal standards), can be used for any molecule that is observable using MS, and can monitor several different molecules within a single analysis.

While many of these characteristics are true, MRM-MS has a number of its own challenges if it is to supplant antibody-based methods as the tools of choice for biomarker validation. One major issue with conducting MRM-MS is its sensitivity. The MS signal produced by a peptide or protein is negatively affected by the complexity of the sample. Serum, plasma and tissue proteomes that are utilized in biomarker discovery are extremely complex, require fractionation and separation, and inhibit the ability to measure any one specific protein. Improving sensitivity requires the development of a method for enriching the specific target molecules prior to MRM-MS analysis. If a biomarker panel is being developed, the enrichment method needs to be applicable to all of the targeted proteins. Regardless of the enrichment strategy chosen, this step introduces experimental variability and further decreases throughput. Antibody-based methods do not require prefractionation or enrichment steps; therefore, throughput is maintained without compromising experimental variability.

While many of the attributes discussed earlier are applicable to the debate of MRM-MS versus antibodies, they must be considered in the bigger picture of biomarker discovery, qualification, verification and validation. Much like students do not graduate from kindergarten and go directly to college; biomarkers do not go from discovery directly to validation. MRM-MS is considered a high-throughput method for qualification and verification since it can measure several potentially useful biomarkers in a single analysis. However, MRM-MS does not have the throughput necessary to meet the demands of a validation test where thousands of samples must be analyzed. This lack of throughput is due to the inability to analyze samples concurrently, something that is trivial for IHC or ELISA.

Another reason that is given for using MRM-MS that needs to be carefully dissected is the lack of suitable antibodies targeted against potential biomarkers. While this situation may be true today, it may not be true in the near future because of the number of global initiatives aimed at developing highly specific antibodies against every human protein. For example, the Human Proteome Organization’s Human Antibody Initiative aims ‘to produce a comprehensive catalog of validated antibodies against human proteins’ [101]. The Swedish Human Proteome Resource is currently focused on producing ‘quality-assured antibodies to all nonredundant human proteins’ [102]. Part of the National Cancer Institute Clinical Proteomic Technologies for Cancer program is the production of affinity reagents to assist proteomic investigators [103]. Collectively, all of these efforts should generate a major increase in the availability of useful antibodies (and possibly other types of affinity reagents) for producing validated biomarkers. It is important for investigators who are developing MRM-MS methods for the eventual validation of biomarkers to take a global view of this research area and be aware of other initiatives that may impact or intersect future goals.

Let us consider the scenario of all things being equal; that is a straightforward MRM-MS method and highly specific antibodies are available to measure a potential biomarker. Which one should an investigator use? Taking all these parameters into consideration, antibody-based tests are still the method of choice. The primary reasons are throughput and the fact that MS is not presently able to meet the accuracy and precision required by the US FDA for a routine clinical test. Will the accuracy and precision of MS increase? Certainly. Mass spectrometers are continually improving at an exponential pace and as more MRM-MS studies are conducted the ability to measure molecules with the degree of precision and accuracy required by the FDA will be achieved. The two tougher challenges are throughput and operator comfort level. As far as operator comfort, while MS is becoming more common, its prevalence within clinical centers does not begin to approach the level of IHC or ELISA. As mentioned earlier, IHC and ELISA can be conducted on numerous samples concurrently, and if throughput needs to be increased, little effort or expense is required. MS, however, is limited to the analysis of one sample at a time, and the only way of substantially increasing throughput is to install multiple mass spectrometers, which represents a major capital expense.

Conclusions & five-year view

It is going to be critical for reagents to become available and methods developed to move biomarkers from the discovery phase through the qualification and verification stages and finally onto becoming a validated biomarker. These reagents and tools are critical because proteomics is creating a glut of ‘potential’ biomarkers that do not get properly vetted because the time and cost required to produce a validated biomarker are intimidating and/or exorbitant to most laboratories. Neither MRM-MS nor antibody-based tests presently provide the complete solution. The choice of which method to employ for biomarker qualification, verification and validation will depend on the availability of highly specific antibodies against the biomarker(s) of interest. With the higher comfortability factor that antibody-based tests enjoy compared with MRM-MS methods and the expected increase in the availability of these reagents through the efforts of groups such as the Human Proteome Organization Human Antibody Initiative and Swedish Human Proteome Resource, tests such as ELISA and IHC will continue to dominate biomarker validation for the near future.

Acknowledgements

The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the United States Government.

Financial & competing interests disclosure

This project has been funded in whole or in part with federal funds from the National Cancer Institute, National Institutes of Health, under Contract N01-CO-12400. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this editorial manuscript.