Photo: Shutterstock
Disease and treatment

Chemical fingerprinting can benefit severely ill people

In the future, medicine will not just be tailored so that it matches the genetic profile of an individual. In the not-too-distant future, in the comfort of our homes, we will be able to test such things as our saliva, skin swabs, faecal samples, breath and urine and get information on our illnesses and whether our diet is good for us. A researcher from the United States is developing an analysis pipeline that can map the chemical and biological profile of everyday objects, including plants, our offices, phones and even our skin. The computational advances to study the chemistry and microbiome are already being used to understand disease.

When she stood up, her pulse increased from 55 to 188 beats per minute in no time. This 39-year-old woman, who was otherwise fit and functioned well, could therefore do nothing other than rest. She had once been close to death when her heart failed. The doctors had done everything possible and were not able to find the reason – but Pieter Dorrestein did.

“We were lucky that she had participated in a microbiome research study a few years before, and this meant that we had older samples and could examine her profile before and afterwards. We could therefore use these samples to obtain a picture of the chemistry and microbes. With our informatics pipeline that is being developed, we aimed to understand what had changed. The doctors believed it was a heart problem, but the major changes were lipids, proteolysis that we see with inflammation and cell death as well as Chlamydophila pneumoniae,” explains professor Pieter C. Dorrestein from Skaggs School of Pharmacy at the University of California, San Diego.

The bacteria had not only attacked the woman’s heart but also her whole body including her lungs and brain. Since the bacteria is intracellular, it is hard to detect and can paralyse cells, and the researchers believed that this was exactly what happened when the woman stood up. Luckily, such bacterial infection can be treated with ordinary antibiotics, and the once severely ill woman has now almost fully recovered.

Thousands of peaks

Rapid response precision microbiome is the initiative that Pieter Dorrestein and his colleague Rob Knight have given to this approach, which provides research-grade multi-omics understanding of samples to end-users. The analysis pipeline, in which biologists, chemists and computer experts work together, aims to make routine analysis possible in less than 48 hours. The pipeline is then used to draw a chemical profile and microbiome profile.

Mass spectrometry is used to analyse molecules of the microbiome. After the samples being examined are treated chemically, they are collided with helium to break apart the molecules into smaller charged fragments. To separate the fragments, they are accelerated through a magnetic field. The greater the mass and the lower the charge, the further the molecules travel. The researchers therefore end up with a mass spectrum of all the molecules in the test as a function of their mass-to-charge ratio.

“Then the real detective work begins. A single bacterium contains thousands of molecules and thus has at least as many mass spectrum peaks. Each bacterium has a very specific molecular profile – a sort of unique barcode – but when we look at a complete microbiome, all the barcodes are jumbled together. Deciphering precisely which barcode or bacteria the individual microbiome contains is therefore difficult and extremely time-consuming.”

Tomatoes in 3D

Although researchers can currently only decipher about 2% of the spectrum peaks, this was still all that was needed to solve the mysterious infection of the 39-year-old woman. The potential is thus enormous if the remaining 98% of the peaks can be successfully deciphered, and this is also the goal of the system in the future.

“We have created the Global Natural Products Social Molecular Networking website, on which all researchers in the field can deposit their knowledge about the molecules to which the individual peaks refer. In this way, we can very gradually fill in the necessary knowledge so we can get better and better at deciphering the enormous mass spectra.”

The research is even more remarkable because Pieter Dorrestein’s team has decided to combine the mass spectrometry method with 3D scanning and imaging. The researchers recently scanned a tomato plant. And are currently exploring how microbial communities on plants and their chemistry impact the health of plants. Then the researchers cut the 3D scan into smaller pieces and analysed each fragment using mass spectrometry.

“This gives us a picture of the diverse types of interactions that occur on and in a plant depending on whether the part is a leaf, a stem or a tuber. For example, we hope that our research will indicate which microbial interactions take place in the tuber. If we can determine this, we may also be able to influence them to grow better under drought-like conditions by simply adding certain bacteria that can help them.”

Still a long way to go

The researchers also used a similar approach on both humans and mice. Because, unlike plants, they cannot be cut into pieces, the research used skin swabs or biopsies. Nevertheless, the results were just as remarkable. Since 90% of the cells that make up our bodies are actually microorganisms, bacteria undoubtedly play a much greater role in our lives than we thought.

“Microbial communities are unique. For example, the natural bacteria on our skin are very important for our immune system, and if we remove them by, for example, washing our hands excessively, we risk throwing our immune system out of balance. Our technique can naturally also be used to detect when new bacteria invade our natural bacterial flora. In some cases this is not good for us, as shown by the case of the 39-year-old woman but on other cases it may be beneficial. However that was an example where we could explain what was happening but that is not always the case. In most microbiome studies and if we want to routinely apply such approaches, we need to understand what is a beneficial microbial community, what is a benign and what is bad from a microbiome but we believe the readout of the microbiome chemistry will enable us to figure this out in the years to come”

Furthermore, while the proof-of-concept has been demonstrated, the technique is still too demanding for routine clinical applications and other emergencies. For example, the analysis of the 39-year-old woman took 21 people for 2 days working 24/7. Now this ability demonstrated that it is possible and are now working to streamline the process so that it can become routine in the future, perhaps to the point that it becomes as easy as taking pictures using a cell phone.

In May 2017, Pieter Dorrestein gave a lecture at the Copenhagen Bioscience Conference "Data-driven Biotechnology - Bench, Bioreactor and Bedside" supported by the Novo Nordisk Foundation. “Mass spectrometry based molecular 3D-cartography of plant metabolites” was published in Frontiers in Plant Science in March 2017, and “Antimicrobials from human skin commensal bacteria protect against Staphylococcus aureus and are deficient in atopic dermatitis” was published in Science Translational Medicine in February 2017.

Pieter C. Dorrestein
Our work aims to develop new mass spectrometry based methods to understand the chemistry of microbes, our microbiome and their ecological niche. In short, we develop tools that translate the chemical language between cells. This research requires the understanding of (microbial) genomics, proteomics, imaging mass spectrometry, genome mining, enzymology, small molecules structure elucidation, bioactivity screening, antibiotic resistance and an understanding of small molecule structure elucidation methods. The collaborative mass spectrometry innovation center that he directs is well equipped and now has twelve mass spectrometers, that are used in the studies to investigate capture cellular chatter (e.g. metabolic exchange), metabolomics, metabolism and to develop methods to characterize natural products. These tools are used to defining the spatial distribution of natural products in 2D, 3D and in some cases real-time. Areas of recent research directions are capturing mass spectrometry knowledge to understand the microbiome, non invasive drug metabolism monitoring, informatics of metabolomics, microbe-microbe, microbe-immune cells, microbe-host, stem cell-cancer cell interactions and diseased vs. non-disease model organisms and the development of strategies for mass spectrometry based genome mining and to detect and structurally characterize metabolites through crowd source annotation of molecular information on the Global Natural Products Social Molecular Networking at http://gnps.ucsd.edu through the NIH supported center for computational mass spectrometry that is co-developed with Nuno Bandeira. A more detailed biography can be found in this Nature article http://www.nature.com/news/the-man-who-can-map-the-chemicals-all-over-your-body-1.20035