Life in the Fat Lane

Meeting the Challenge

Lipidomics has come a long way – but there are still many analytical challenges ahead.

The word “lipidomics” did not even exist 20 years ago. Back then, only a limited number of lipid classes and species were analyzed, and rarely in connection with other biomolecules and biological backgrounds. Since then, the field has changed dramatically; nowadays, the application of lipidomic analysis in various fields is increasingly prevalent.

Why have lipids and lipidomics taken on greater prominence? First, because they play crucial roles in human biology. Numerous lipids are present in every cell in our body and have a huge influence on health and disease: they form the lipid bilayer that surrounds the cell and intracellular compartments, they are involved in energy storage and cell signaling. Consequently, there has been a serious drive to develop methods for lipid analysis, which – together with rapid innovation in the area of mass spectrometry – has led to the rapid growth of the lipidomic field.

I first began analyzing lipids in 1995, with an initial focus on monitoring biodiesel production – specifically, the transesterification of triacylglycerols to methyl esters of fatty acids. We had recently had an LC-MS system installed (the first in the Czech Republic – a single quadrupole from Waters!). Starting with the analysis of triacylglycerols, we developed analytical methods for all existing types of isomers up to triacylglycerol enantiomers, extending our interest to phospholipids, sphingolipids, and other lipid categories to build a more comprehensive picture of the lipidome. From there, we decided to look at the application of lipidomic analysis for cancer biomarker research, which remains our main focus today.

Lipidomic analysis is very complex. There are now dedicated proteomic databases and standardized approaches for structural elucidation, whereas such tools are not yet available in lipidomics. In lipidomic analyses, there are too many sources of variability – different lipid categories, classes, subclasses, and many types of isomerisms inside classes – including fatty acyl chain lengths, the double bond number, positions, geometry, positional isomers, and enantiomers. The theoretical complexity is extremely high and we are still not sure why the body needs such a huge diversity of lipids.

Methods in the madness

There are three major approaches in lipidomic analysis (1)(2)(3). The first involves direct infusion (often called shotgun): the lipidomic extract is infused directly to the mass spectrometer at a low flow rate, so all required scans can be performed using either a low-resolution (precursor ion and neutral loss scans with triple quadrupole or Q-LIT instruments) or high-resolution (QTOF, Orbitrap or, more rarely, ICR) approach.

The second area is MS coupled to chromatographic techniques, which brings obvious advantages, including the chance to separate various types of isomeric lipids. Chromatography offers a wide variety of different separation mechanisms, but by far the most common mode is reversed-phase LC, which provides an excellent separation selectivity for isomers differing in the hydrophobic part of the molecule. Silver-ion chromatography is a special mode that uses embedded silver ions in the stationary phase, which interacts with double bonds in lipids or other molecules. Increasing numbers of double bonds mean stronger interactions, making it possible to separate lipids differing only in the double bond position or cis/trans configuration. These methods provide the selectivity for species separation. 

Alternatively, lipidomic class separation can be used, such as hydrophilic interaction chromatography (HILIC) and normal-phase chromatography. Lipid species within one class coelute with the lipid class internal standard, which makes the quantitation more robust. In fact, there is a small partial separation even inside classes, but this only affects quantitation to a very limited extent. Reversed-phase LC can also be used for quantitation, but careful attention should be paid to the selection of an adequate number of internal standards to cover the whole retention window, and to full method validation, including the determination of the matrix effect. In principle, any MS method can be used for quantitation, provided that all requirements of quantitative analysis are followed; method reliability is clearly demonstrated using validation parameters, quality control samples and cross-validation; and the quantitative data obtained are in agreement with other laboratories.

Desorption ionization MS and imaging MS is the third area. Matrix-assisted laser desorption/ionization (MALDI) is the most common ionization technique, but desorption electrospray ionization (DESI) and other desorption ambient techniques are also now available. A typical application is MS imaging of tissues to show the distribution of lipids and other biomolecules within tissues and organs. There are some limitations in terms of quantitation (the signal in desorption ionization MS is not as stable as atmospheric pressure ionization techniques) but semi-quantitation can also be achieved (4).

In future, I anticipate further implementation of ultra-high-resolution mass spectrometry (UHRMS), ultra-high-performance liquid chromatography (UHPLC) and ultrahigh-performance supercritical fluid chromatography (UHPSFC) in common practice. In my group, we are fans of UHPSFC-MS coupling for high-throughput lipidomic quantitation (5); it is robust enough for high-throughput lipidomic analysis, and the sensitivity for less polar lipid classes is much higher than any other available techniques. We expect to see more groups implement this relatively new approach in the near future.

From analytical method to clinical practice

The next step for the field of lipidomics is to gain a better understanding of the biology behind the changes observed in analytical experiments. For this, we urgently need experienced biologists to collaborate with analytical chemists. Many biologists consider the lipid class as a whole, but it’s important to note that – even in the same lipid class – some lipids are upregulated while others are downregulated. Therefore, we need biologists who are able to interpret observed changes lipid-by-lipid.

Our own group’s focus for the next few years is quite clear. We have obtained really exciting results in the area of lipid cancer biomarkers, and we are working hard to get the methodology fit for implementation in real clinical practice for early cancer screening and monitoring of treatment progress.

One challenge we have to be particularly aware of is the biological variability among individuals. To achieve acceptable accuracy for lipidomic bodily fluid analysis, full optimization of the whole methodology is required, starting with sample collection, storage, transport, sample preparation, analysis, and data processing up to statistical evaluation. With so many steps, there is a risk of introducing artefacts and errors in addition to the biological variability – so we have to be as rigid as possible in terms of method optimization, validation, and quality control. Currently, it is possible to differentiate cancer patients from healthy volunteers – but the analytical methodology must be as robust as possible.

Our ultimate goal is to convince a strategic partner to move the methodology from the academic lab into real clinical practice. To do that, we need experts from various fields (clinicians, analytical chemists, statisticians, and biologists) – and finding a common language among these experts is another challenge ahead of us!

Michal Holčapek is Head of the Lipidomics group at University of Pardubice, Czech Republic.