The specifics of control matrices: How far to go?

Written by Robert MacNeill, Envigo

robert macneil (pic)Robert MacNeill received his Bachelor’s degree with Honors in Chemistry from Heriot-Watt University then his MSc in Analytical Chemistry from the University of Huddersfield (Both UK). Robert is also a Chartered Chemist and Member of the Royal Society of Chemistry. With 20 years of experience in all aspects of quantitative bioanalytical LC–MS/MS method development, eleven of these years heading method development activities within HLS/Envigo, and a regular author and peer reviewer for the journal Bioanalysis and Expert Panelist on Bioanalysis Zone.

In his current role Robert coordinates all LC–MS method development and associated training, takes the lead in keeping abreast of innovation and technological development in the industry, coordinating in-house research projects and technical writing.


In this installment of Robert MacNeill’s (Envigo) column, Robert gives his thoughts about some pertinent aspects of control matrices.  

It has been becoming more commonplace and noticeable, at least on the CRO side of the industry, that bioanalytical LC–MS assays are no longer presumed to involve simply the species’ matrix as control, for the purposes of producing double/single blanks, and for spiking calibrant and QC samples. For instance, until recently, mouse plasma could have been harvested from males and females, indeed pooled as such and the strain was not typically regarded as important for control purposes.

The contemporary angle when carrying out a study on, for example, a specific strain of a species is that the control matrix for analysis must also be entirely from the same strain. This is easily justifiable on the surface, but is it really necessary?

The composition of the matrix is a pivotal aspect of bioanalysis, in that it is what we need to understand and come to terms with, such as how to keep an analyte freely solubilized within the matrix, free of binding to the container surface, and ultimately how to efficiently extract said analyte from the matrix without co-extracting a significant amount of interferences.

There are properties of matrices that are obvious and readily investigated. Hemolysis, for instance, is a phenomenon that changes a blood-derived matrix to something markedly different, and comes with implications for selectivity and stability. Yet, hemolyzed samples are analyzed alongside regular non-hemolyzed samples after the method is demonstrated to be steadfast as regards maintaining accuracy and precision for these samples.

Hyperlipidemia, for human blood-based samples, is another visually clear (perhaps unclear is more apt) phenomenon that alters the matrix and for which a method must be proven rugged enough to maintain data integrity in the analysis. Indeed, clinical samples can be notoriously challenging, bringing never-seen-before interferences, viscosity changes and other curiosities that can really challenge a validated method. This is partly what brings such tremendous value to incurred sample reanalysis, when it comes to testing a method’s mettle.

One very pertinent component of incurred post-oral samples is the metabolite content. Even with the selectivity of the various modes of MS at our fingertips, their presence can prove deleterious to a method’s performance, and selectivity with regard to known or probable metabolites should be established in method development. Often there is a plethora of metabolic possibilities, both phase I and phase II, and there is almost inevitably a certain abundance that can efficiently revert to the parent compound when exposed to the thermal conditions of the ion source. There, of course, one must rely on the preceding chromatography for selectivity.

Also, it is not just at elevated temperatures where such ex vivo conversion can take place. Transesterification, for instance, readily occurs at ambient temperature. Without overdoing the metabolite discussion, suffice to say that this generally metabolite-infested matrix can be said to be an important analytical departure from the relatively simple control.

Matrix aging is something that is less obvious than the likes of hemolysis but nonetheless a very real phenomenon, and a danger to analytical integrity of a matrix as certain endogenous components change concentration over time, and degradants are created, in essence creating a different matrix. This realization has led us to ideally store control matrices at -80°C, whereas historically it has been -20°C.

Then we have the effect of the anticoagulant used for blood, plasma or serum assays. This came into the limelight recently in the bioanalytical community, and was discussed and debated at length. A large-scale collaborative experiment was eventually performed, with the convincing outcome that the selection of anticoagulant counter-ion, at least, for a validated LC method with tandem MS detection does not significantly affect the accuracy of the resultant data. In what amounts to a necessary and largely uncontrolled adulteration of samples, and with the negligible outcome on data reliability, this must offer more strength to the notion that generally insisting on a specific strain or gender is largely a futile exercise. A good method deals with the presence of an anticoagulant, just as it deals with a host of other potential interferences that can vary enormously in concentration.

To sum up, the logic behind the kind of decision that brings only a specific strain or gender of control matrix is readily understandable. I would endorse it too, where program/study design practicalities and costs are not prohibitive, in order to acquire data with the greatest reliability. However, it is considerations like the aforementioned that, to me, swamp the validity of doing everything one can to stick to a specific gender or strain for control matrix.

There are too many other factors at work, and if it is required to do this so that the method works for accuracy and precision, it raises a question mark about the method integrity. The most important theme has to be the reliability of the method itself and how it fares under all possible challenging circumstances within the realm of species and matrix. The more observed variation in the control matrix composition accompanied by no adverse effect on performance, the better.

This article is part of Robert MacNeill’s (Envigo) quarterly column for Bioanalysis Zone which focuses on quantitative method design. You can read past installments of the column here

Click here for more information on Envigo