Measuring Lipids in the Blood Can Predict Disease Risk Decades Before Onset
Advancements in personalized medicine have led to the identification of many novel disease biomarkers. Such biomarkers can be used to diagnose and monitor diseases, predict patient responses to specific therapies and to identify those that, while perhaps healthy right now, might be at risk of developing a disease in the future.
The high-throughput and low-cost nature of modern next-generation sequencing (NGS) technologies led to a recent influx of genetic-based biomarkers in clinical diagnostics. However, our DNA code does not always reflect our exact physiological state in each moment. Beyond the cell’s nucleus, a biological “soup” of molecules are continuously synthesized and degraded. Identifying and quantifying these molecules can provide a more accurate representation of our health – or disease – status.
Lipids, a class of biomolecules including fatty acids, vitamins, monoglycerides and phospholipids, play central roles in functions such as cell signaling and energy storage. Their concentrations fluctuate in response to stimuli such as food consumption, exercise and disease. Research is therefore exploring the lipidome – the collection of lipids that exist within a cell, tissue or organism at a specific moment in time – as a potential non-invasive biomarker source, as lipids can be extracted from blood samples.
A challenge in modern medicine is that many patients present to their doctor once they have started to experience symptoms of a disease. Diagnosis, therefore, often comes once the disease has progressed. A goal of personalized medicine is to be able to monitor health status such that disease can be prevented.
A new study published in PLOS Biology, led by Professor Chris Lauber, demonstrated that lipidomic profiling can be used to predict the risk of developing Type 2 diabetes (T2D) and cardiovascular disease (CVD) years before disease onset. Lauber is part of the Lipotype team, a company that offers shotgun lipidomics analysis using cutting-edge mass spectrometry (MS).
Currently, risk assessment for both T2D and CVD involves utilizing patient history and measuring high- and low-density cholesterol. Lauber and colleagues hypothesized that there are likely hundreds of other lipids in the blood that contribute to disease risk. In their study, they analyzed blood data from over 4,000 healthy, middle-aged Swedish male and female individuals that were involved in a longitudinal study occurring between 1991–1994, with subsequent follow-up until 2015.
From baseline blood assessments obtained in the 90s, 184 lipids were analyzed via MS, and machine learning generated risk scores for T2D and CVD for the 23-years of subsequent follow up. These risk scores were then used to stratify patients into six categories, ranging from low to high disease risk. A lipidomics risk score resulted in a 168% increased incidence rate in the high-risk group for T2D, and a 77% decreased incidence rate in the lowest risk group. For CVD, lipidomics predicted an 84% increase in incidence rate in the high-risk group, and 53% decrease in the lowest risk group.
In contrast, the longitudinal case study data showed that, by 2015, 13.8% of the study participants had developed T2D, while 22% had developed CVD. “Our results demonstrated that a subset of individuals at high risk for developing T2D or CVD can be identified years before disease incidence,” the authors write in the paper.
Technology Networks recently interviewed Lauber to discuss the study and its wider implications for biomarker discovery in more detail.
Molly Campbell (MC): For our readers that may not be familiar, what are lipids? Why are they useful to study as potential biomarkers for diseases?
Chris Lauber (CL): Lipids are a diverse group of small molecules. Glycerolipids, sterol lipids, sphingolipids, phospholipids – there are thousands of distinct lipids. Each lipid is chemically unique, and they have many different biological functions.
One biological role is to form the matrix of cell membranes, where they support a variety of vital functions. But lipids also function as energy storage, and serve as hormones, essential nutrients or cellular signaling molecules too. More and more we find functions related to individual lipid molecules.
In humans, there are thousands of different lipids that are linked to health and disease. Their molecular interplay with genes, proteins and metabolites influences our well-being. This interplay is strongly affected by the cellular supply of each of these lipids.
“Healthy” means all lipids are perfectly balanced. Each cell must constantly adjust its lipid metabolism to achieve this goal. A disordered lipid metabolism poses a direct threat to health and life. It can fuel the development and even cause severe diseases. Lipids can thus serve as biomarkers to monitor the health of an individual.
MC: Can you discuss how a “lipidomics profile” is created?
CL: Testing selected lipid molecules provides only a glimpse at the state of health. Measuring all lipids at once paints the big picture, a full lipidome. The technical solution to this is lipidomics, where thousands of lipids from biological samples are analyzed at once
The lipids are shot into a mass spectrometer. Bioinformatics solutions and bio-statistical methods convert the mass spectrometer results into powerful charts and graphs. These extended lipidomics profiles tell us all about the lipids and lipid metabolism of the original biological sample.
MC: This study focused on 184 lipids – can you discuss why these lipids were analyzed, and provide further detail on specific lipids that were particularly important to focus on?
CL: Our lipidomics analysis covers a wide range of lipids from phospholipids and sphingolipids to sterol lipids, glycerolipids and more – in total more than 4200 lipids. This means that the coverage includes those lipids that are part of the traditional lipid panel, and thus already known to be linked to cardiovascular disease and diabetes, e.g., triglycerides and cholesterol.
Yet, the lipidomics measurement of a biological sample covers the full lipidome and thus the data set includes further lipids and also more details about the lipids. As we analyzed the samples, we acquired detailed lipidomics profiles containing hundreds of lipids. We then proceeded with the data analysis and selected only the most reliable lipid candidates. This set of most reliable candidates included the 184 lipids we focused on in this work.
MC: Can you tell us more about the methods used in this study?
CL: In the 90s, blood samples of about 4,000 study participants were sampled and frozen. Over the course of the next 23 years, the study participants were followed up with and incidents of multiple diseases were recorded, e.g., CVD and T2D. We analyzed these samples to acquire their lipidomes, i.e., their detailed lipid profiles. We excluded participants with prevalent T2D or CVD at the time of sampling and performed statistical tests to calculate the risk for developing T2D and CVD for the remaining lipid profiles. We created predictive models based on the full lipidomes.
MC: How could this study be used to inform personalized medicine practices?
CL: In principle, this study can be used to calculate the individual risk for T2D or CVD from the lipidome of a person. It is a first step in the direction of personalized medical practices.
MC: Are there any limitations to the work that you wish to highlight?
CL: This study is based on a Swedish cohort and thus will need to be independently checked in other countries and ethnicities.
MC: What are your next steps in this research space?
CL: We want to move from research towards an assay that can be used in medical practice.
Professor Chris Lauber was speaking to Molly Campbell, Senior Science Writer.