Single-Cell Metabolomics: Still a Ways to Go

Single-cell metabolomics combines two cutting-edge technologies: metabolomics and single-cell analysis, and thus entails the strengths and shortcomings of both.

Unlike genes, whose chemical concentrations in eukaryotic cells is predictable (and, through amplification methods, expandable), the concentration dynamic range of metabolites may be as high as 10¹³. This presents many needle-in-a-haystack scenarios where scientists look for species at picomolar concentrations in the presence of others at millimolar concentrations. The other issue is the vanishingly tiny sample size, which strains the sensitivity capabilities of modern instruments. Then there is the separate issue of isolating single cells.

Luckily, the rationale for analyzing individual cells—as opposed to tissue homogenates of many cells—is the same whether the analysis mode is proteomics, genomics, or metabolomics. Homogenates are certainly easier to study, as analysis of even a needle biopsy involves billions of cells. However, the analysis of bulk tissue homogenates only provides an average answer for, say, glutamate. The answer “glutamate 0.03 mM” is sufficient and satisfactory for many studies but it says nothing about relative concentrations of cell populations. ‘Omic profiling of single cells can provide a census or map of cells behaving in specific ways, e.g., in their output of glutamate.

Chemical diversity

For these reasons single-cell metabolomics remains highly experimental, and lags behind proteomics and genomics at the single-cell level.

“Metabolites are chemically very much more diverse than nucleic acids, and there are no simple ways to amplify them as is possible with RNA or DNA by (RT)PCR,” says Professor Andrew N. Lane, in the Department of Toxicology and Cancer Biology at the University of Kentucky.

The chemical noise background is high for metabolite analysis by mass spectrometry (MS), so one is often quantifying species at close to noise or trace levels, Lane tells Biocompare.

“Consider, for example, a single epithelial cell having a total volume of 2 pL (and thus a cytoplasmic volume closer to 1 pL). A high intracellular concentration of ATP or NAD⁺, say at 1 mM, represents just one or two femtomoles [10^-15 mole, or about 1 billion molecules] of analyte per cell. But many metabolites are present at only around 1 μM, or 1–2 attomole [10^-18 mole, or about 1 million molecules] per cell. Signaling molecules may be present at even lower concentrations, in the nanomolar range, so that is a very, very small number of molecules per cell. And, if the analyte is uncharged, ionization efficiency losses in the MS come into play.”

Such losses, even for mild ionization methods such as electrospray, may be higher than 90%.

“And no extraction technique can guarantee you will get 100% of the sample,” Lane explains. “That is why much of the single-cell metabolomics thus far reported has focused on a relatively small number of abundant compounds, and has also required dissociation of tissue into individual cells, thereby losing critical spatial information.”

Where metabolic labeling with stable isotopes is required, sensitivity suffers more as each analyte may be spread into several isotopomers.

Another analytical technique for metabolomics worth noting is nuclear magnetic resonance (NMR), which provides structural information as well as identity and quantitation, but, according to Lane, NMR is not very sensitive. “Under ideal conditions, using the latest in NMR probes and high magnetic field strength, the minimum concentration of analyte in the sensing region must be of the order 0.5 nanomole, which is six orders of magnitude higher than what is present in individual cells.”

That is why single-cell NMR is not yet feasible, even for isolated cells except, perhaps, for very large oocytes. “With hyperpolarization, one might reach picomolar sensitivity for a brief period, maybe for two minutes, but NMR does not have the resolving power to achieve single-cell resolution in a tissue or organ.”

Dilution and interferences

Most chemical analyses of complex, living systems involve some type of sample preparation to either concentrate target analyte, remove interfering species, or both. Purification is not such a great idea for metabolomics, at least not for single cells.

“Since any separation, for example capillary electrophoresis or liquid chromatography, involves dilution, single-cell metabolomics experiments avoid separation,” says Professor Zhibo Yang in the Department of Chemistry and Biochemistry at the University of Oklahoma. “In our studies we use microscale probes to extract cellular contents directly and inject them into the MS.”

On the other hand, separation reduces matrix effects, whereby co-existing molecules hinder the ionization and detection of target molecules, and thus increase ionization efficiency. When this is desirable Yang suggests applying either capillary electrophoresis or ion-mobility spectrometry before ionization. “IMS is a post-ionization technique occurring inside the MS,” Yang explains. “It separates molecules based on the collision between ions and helium gas molecules. For ions with the same mass/charge ratio, compact ions travel faster. IMS also provides structure information.”

Be prepared

Cell preparation and choice of MS ionization method may be as important for obtaining consistent single-cell metabolomic results as the analyte concentration.

“Coupling separation to single-cell experiments requires access to cell-handling systems that are not standard equipment in MS labs,” Yang says. “Commonly, multiple steps such as cell identification, capture, cellular content extraction, and sample ionization will be required. It is also critical to avoid contamination and losses by, for example, minimizing the use of containers. In addition, a high-efficiency ionization source, such as a nanoESI based method, at low flowrate, is preferred for better detection sensitivity.”

Compared with MALDI, another gentle ionization method, nanoESI provides good ionization efficiency and relatively cleaner background within the low mass range characteristic of metabolites.

“In terms of the spectrometer, high-resolution instruments such as Orbitrap, FT-ICR, and some TOF instruments, are preferred because they provide the ability to detect the many complex species existing within a cell. Tandem MS is challenging because of limited sample. Other methods, such as traditional LC/MS of cell lysates, CE, and IMS, provide complementary information.”

Whether the target is a gene, a protein, or a metabolite, single-cell analysis is not for the faint hearted. The quantities are tiny and the workflows are prone to both systematic and random errors. “Single-cell MS studies require more experimental skills compared with most common MS studies,” says Yang. To this requirement one can add cell-preparation methods and data processing to make sense of it all. While experimenters continue to push the boundaries of what is possible for single cells, the eventual commercialization of diagnostics or even “single-cell metabolomics kits” will probably be limited to just a few high-abundance metabolites.