Noninvasive Assessment of Traumatic Brain Injury by LC/MS/MS Determination of 8 Urinary Metabolites.
The paper I'll discuss in this post is this one: Simultaneous Determination of Eight Urinary Metabolites by HPLC-MS/MS for Noninvasive Assessment of Traumatic Brain Injury (Shi et al., J. Am. Soc. Mass Spectrom. 2020, 31, 9, 1910–1917).
When your Doctor orders blood or urine tests to obtain information about your state of health, the general term for the analyte being measured is a "biomarker." Blood levels of glucose, or A1C, for example are biomarkers for diabetes, cholesterol, a biomarker for heart health. Quietly, without much fanfare, there has been a revolution going on in the determination of biomarkers with increasing sensitivity. To some extent, this revolution - at least originally - was involved in something known as "competitive binding assays" or "ligand binding assays." When I was a kid I used to make reagents for these assays that were labeled with radioactivity, "RIA" reagents, "radioimmunoassay" reagents; almost no one uses this technology anymore, it has been supplanted by ELISA, enzyme-linked immunosorbent assays, in which the radiation detected is of much lower energy than is provided by radioactive materials: It is the radiation we know as "visible light." I expect that within a decade, much of the ELISA testing will go the way of RIA, supplanted by mass spectrometry, or at least in tandem with mass spectrometry.
Mass spectrometry, in general, falls into two classes, "triple quads" which provide high sensitivity for the "work a day" world - including commercial medical laboratories, and in research, high resolution (where resolution refers tp mass accuracy) mass spectrometry, which can be of several types, "orbitrap," "time of flight" as well as the much rarer but incredibly accurate "Fourier Transform Ion Cyclotron Resonance Mass Spectrometry (FT-ICRMS).
A recent trend has been to approach combining high resolution mass specs with high sensitivity; this trend is only going to continue.
In the absence of validated biomarker analysis, diagnosis can involve considerable risk. Personally, I have an unusual EKG. Since I was a young man, doctors have often interpreted my EKG as implying that I have had - or am having - a heart attack. I've actually been hospitalized several times because of this, and once was briefly admitted to ICU, despite the fact that I have not had a heart attack. The second time I went through this process, my physician recommended that I have an angiogram, which involves inserting a catheter directly into the heart. While the test is very cool - you're on a local anesthetic and can actually watch the chambers of your beating heart on a television screen along with the doctor - it also is fairly risky. I had to sign all sorts of releases stating that I was aware that the test could either severely injury me or kill me. (There is a non-invasive biomarker test, "cardiac enzymes," troponins, that can indicate an active heart attack - but the damage done by previous heart attacks - I've had this test a number of times.
The situation with respect to the brain is even more complicated. Although in recent years many technologies have been developed for brain imagining, since the brain is an organ where many problems, for example Alzheimer's, actually take place on the cellular or molecular level.
The instrument utilized in this paper is certainly not a "state of the art" instrument, but it is a solid instrument that was widely in use ten or fifteen years ago, and was very popular, bordering on something of a scientific "cult" - the Sciex API 4000, a triple quad instrument. This lab, at the University of Missouri is probably not funded enough to drop half a million dollars regularly on the latest and greatest mass spectrometer every two or three years. Nevertheless, it's good work, which can lead people - some of whom will have more advanced instrumentation - to have something on which to build.
From the introduction:
Molecular biomarkers can provide additional biological information about concussive injuries that can complement conventional neurological assessments and medical imaging. Prior efforts to identify TBI biomarkers have noticed novel associations between several proteins found in cerebrospinal fluid (CSF) and serum to clinical outcomes in patients with TBI.(9) For example, glial S100β, which is involved in low affinity calcium binding in astrocytes, has been connected to astrocyte stress and death.(10) Glial fibrillary acidic protein (GFAP) has also been extensively studied owing to its function as an intermediate filament protein in astrocytes, where increased serum levels have been associated with astrocyte damage.(11,12) Additionally, serum levels of neuronal, specifically axonal, damage indicators, including neuron-specific enolase,(13) cleaved tau protein,(14) and ubiquitin C-terminal hydrolase,(15,16) have been correlated with poor clinical outcomes.
It is possible to obtain "CSF" for medical tests - and it is also possible to detect proteins by mass spec, either intact, with high resolution MS or by digestion with triple quads - but obtaining CSF is a somewhat involved and possibly traumatic enterprise. Urine is easier to get, obviously.
The authors rationalize their choice of biomarkers - all of which are small largely endogenous molecules. (Endogenous molecules are somewhat more difficult to analyze than molecules - such as drugs - that do not occur naturally in the body simply because it is difficult to procure blanks lacking these molecules but also containing possible interfering species.
....(17,22−24) Furthermore, excess oxidative stress occurs as a result of mitochondrial damage and altered oxidative metabolism, which has prompted interest to explore oxidative stress metabolites, most notably N-acetylaspartic acid (NAA)(25,26) and F2α-isoprostane,(27,28) as potential TBI biomarkers. This complex pathophysiological cascade invokes distinct metabolic processes that may therefore be probed to assess the extent and nature of the injury.(17,18,29) We have therefore proposed a novel combination of these interesting metabolites, shown in Figure 1, alongside several established protein biomarkers, to study potential patterns and differences in TBI samples. This panel, composed of a combination of excitatory neurotransmitters, glycolytic intermediates, and oxidative stress indicators, represents broad and diversified potential biomarkers of neurological processes involved in TBI...
...While cerebrospinal fluid has conventionally served as the gold standard for brain metabolite quantitation, owing to its direct contact with the extracellular matrix of the brain,(30) urine and serum present minimally invasive specimen types that can be readily obtained in the field. In particular, urine offers several advantages including large sample volumes, fewer sample pretreatment requirements, and noninvasive sample collection requiring no medical expertise. However, and to the best of our knowledge, a comprehensive panel of TBI biomarkers has yet to be developed in urine... ...In this study, we have therefore proposed a new technique for simultaneous determination of eight chemically diverse metabolites in urine without intensive sample preparation using high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS). This novel approach may provide timely supplemental characterization of traumatic brain injuries in the field...
Here are the biomarkers and their structures as reported in the paper, in one case incorrectly:
The caption:
The structure drawn for 5-hydroxindole acetic acid is missing the "5-hydroxy" and thus may generate a correction note in a subsequent issue of the journal. Probably the structures were drawn by a student.
What is noticable is that many of these molecules are generally ubiquitous in cells: Pyruvate is a constituent of the citric acid (Kreb's) cycle which takes place in mitochondria. Glutamic acid is an amino acid found in practically every protein - the monosodium salt is the famous "MSG" which upsets certain people eating Chinese food. Methionine sulfoxide is a derivative of methionine, a common amino acid in many proteins, wherein the methionine is oxidized to its sulfoxide as part of normal physiological processes, for example methylation reactions.
The simplicity of these molecules makes their analysis significantly more challenging than it is for complex molecules. This said, the F2α-isoprostane is challenging because it belongs to a class of lipid molecules known as prostaglandins, themselves a subset of molecules called "eicosanoids," the metabolic products of highly unsaturated acids. (These molecules are important to inflammatory pathways involved in healing and in immune responses.) A problem with these classes of molecules is that different molecules, with different roles can have the same molecular weight - we call these "isobaric" molecules - with the result that since a mass spectrum depends on mass signals, and thus can give ambiguous and indeed, wrong results. (This possibility is not discussed in the paper, but presumably it was addressed in the laboratory.)
Here is the chromatogram of the analysis, which is actually in two dimensions, one being a mass related dimension ("mass transitions" in fragmentation, and the other in time, specifically retention time, the time it takes for the molecule to emerge from the separating chromatography column. The first is using reference standards:
The caption:
Figure 2. Representative overlaid XIC chromatogram of the eight metabolite standards prepared at 500 μg/L in synthetic urine.
The second is a representative chromatogram:
The caption:
Here are tables of the method performance:
Here the authors report the LOD, limit of detection, rather than the more useful LOQ, limit of quantitation, but the reported signal to noise ratio is decent.
I have seen prostaglandin analysis that have LOQ's that are two orders of magnitude lower than what is reported here as an LOD, but this was very challenging to accomplish and involved far more advanced instrumentation than a Sciex API 4000.
Strictly put, the spike recovery for the prostaglandin, its accuracy, is out of range for a regulatory method, which may or may not be involved in interference, but it's not bad overall.
My snottiness aside, the method is pretty decent within its limits - it's an academic, not a regulatory method - and has a reasonably short run time for a method measuring 8 analytes simultaneously.
The conclusion is a little bit overstated, but not so much as to be entirely unreasonable:
It is notable that urine - despite its apparent simplicity - can be a difficult matrix for mass spec for a number of reasons, and the authors have done a very workman like job within the limits of their equipment base. It's nice work.
I trust you will have a safe, healthy, and enjoyable weekend.