In bioanalysis, working within a regulated scientific discipline doesn't mean we shouldn't question the processes we use. Discussing and challenging established practices helps us understand the guidance, SOPs, and workflows we follow. Critical thinking also helps identify weaknesses and opportunities to improve our science, especially as we address rapidly changing drug design and disease investigations.
A recent discussion in our lab highlighted an example, focused on the extracted sample stability experiment, also known as post-extracted stability (PSS). This experiment is specific to methods that employ sample pretreatment before bioanalytical measurement, common in chromatographic assays like liquid chromatography mass spectrometry (LC-MS). The premise is that once a sample has been 'cleaned up' by processes such as protein precipitation, liquid/liquid extraction, or solid-phase extraction, the analyte stability needs to be confirmed in the modified matrix. However, this experiment has a history of controversy.
To understand the issue, we need to consider what happens to our analyte of interest and how we plan to measure it quantitatively. Our objective is to measure the analyte concentration in the original sample matrix but stability events can impact our assay at any point. Of course, with chromatographic assays we have the incredibly useful internal standard (IS) and the benefit of using the analyte/IS response ratio to measure concentration. However, herein lies the challenge with the PSS experiment. The IS introduces an unwanted variable in the PSS experiment since it may, or may not, have the same stability as our analyte of interest.
The commonly accepted approach to establishing PSS involves measuring extracted quality control (QC) samples stored for the stability period and storage conditions (e.g., autosampler or refrigerator temperature) against a freshly-prepared calibrator standard curve. The problem is that each stability QC is spiked with IS before extraction. If the IS has the same stability as the analyte, it masks any instability of the analyte. This is particularly relevant to the use of a stable-isotope labelled (SIL) IS. Conversely, if the IS has different stability, it may erroneously impact the analyte stability measurement. Ironically, this experiment intended to measure stability is not a definitive stability experiment.
The European Bioanalytical Forum (EBF) has highlighted this flaw in several presentations available online. Despite the critique, the prevalent use of SIL ISs for LC-MS methods, where any IS instability is effectively masked, has enabled practical operation. There are arguments against the value of the PSS experiment altogether, particularly compared to the run-reinjection reproducibility experiment (also a requirement of bioanalytical method validation), but as of today it remains a requirement of BMV Guidance. This raises the question, "Is there a better way to measure analyte stability?" If there is, we must decouple the IS stability from the stability of the analyte.
Our bioanalytical laboratory is piloting an alternative approach to PSS in method development while still executing the traditional experiment in method validation. The alternative approach involves storing stability samples without IS and adding the IS only before LC-MS analysis. The workflow is as follows:
- Extract 12 replicates of stability QC (in matrix) without IS spiked.
- Post-extraction, add a fixed concentration/volume of IS to the first 6 of the 12 QC replicates.
- Assay ONLY the IS-spiked stability QC wells (first 6 wells) = Result A set.
- Store the plate for the required PSS time (e.g., 72 hours at 4°C).
- Spike the stored stability QC wells with the same IS (same concentration/volume as Step 2).
- Assay the 6 replicates (second set of 6 from the original 12 QC replicates) = Result B set.
- Compare Result B set to Result A set directly based on area ratios only, not regressing against a standard curve.
Spiking with IS prior to LC-MS analysis compensates for any instrument variance but does not compromise the stability assessment. However, this approach isn't perfect. We lack the quantitative normalizing effect of the IS throughout the experiment, but that's the point. The IS spiking solution must be stable over the time-period of the experiment. It's also critical that quantitative spiking of IS in steps 2 and 5 is identical. There is also an assumption that we are using an appropriate concentration and volume of IS for spiking but that is likely assessed in prior method development experiments. We can anticipate that the increasing use of robotic pipetting on liquid handler platforms will help ensure the needed consistency with this approach as we move forward.
To date the approach described has proven practical and valuable. In terms of value, it primarily helps the bioanalyst obtain a handle on the chemistry of the molecule of interest. Instability (even if compensated by the IS) can have deleterious effect on assay sensitivity and the proposed approach gives opportunity in method development to address true stability events in an extracted sample. I believe this perspective on the bioanalytical method is one that can be overlooked when the traditional PSS experiment, using a SIL IS, passes acceptance criteria. Ultimately, I still maintain that the run-reinjection reproducibility is the definitive practical test but conducting this rudimentary stability experiment in method development can add value and ensures regulatory compliance too.
I invite feedback and thoughts on this topic. While the established PSS experiment is ingrained in bioanalytical practice, open discussion on such matters is valuable.
