Flow Cytometry Uncovers Predictive Biomarker for Severe COVID

Using image-based flow cytometry, researchers at the Technical University of Munich (TUM) identified a biomarker for predicting how serious a COVID-19 infection will be. They also developed a rapid test for assessing a patient’s risk of severe COVID-19 infection. The ability to determine the potential gravity of a COVID infection will help physicians determine the most effective treatment for these patients at an early stage of the disease, leading to better patient management.

The human body produces a series of immune responses when infected with the virus that causes COVID-19. One response is the formation of cell aggregates in the bloodstream. To understand the significance of these aggregates in COVID-19 disease progression, the researchers investigated the number and components of cell aggregates in SARS-CoV-2-infected blood using phase imaging flow cytometry.

A research team from the Technical University of Munich (TUM) used a method from image-based flow cytometry to rapidly analyze the interactions between large numbers of blood cells. The team identified a predictive biomarker for the risk of serious infection in COVID-19 patients. Courtesy of iStockphot.com/iLexx.

The researchers rapidly analyzed the interactions between large numbers of blood cells. They found that blood from patients with severe COVID-19 was associated with higher numbers and a specific composition of cell aggregates.

To establish a method that would enable high-throughput measurement of cell aggregate concentration and composition of cell aggregates, the researchers combined a quantitative phase imaging method, also known as digital holographic microscopy, with a microfluidic chip and a customized image analysis tool, thereby creating a label-free, high-throughput imaging flow cytometer.

The use of digital holographic microscopy allowed for the label-free differentiation of blood cells due to the sufficient contrast of the corresponding phase images. This enabled measurement to be performed without the need for time-consuming sample preparation.

The researchers optimized their setup to perform micro-aggregate measurements at low shear rates mimicking blood vessel flow conditions. Despite the low shear rates, they achieved a high throughput of 500 to 2000 cells per second. They limited the time from blood draw to measurement to <30 min to ensure the quality of the blood samples and avoid any effects from sample aging that could lead to aggregate artifacts.

This approach enabled the researchers to resolve the concentration of platelet and platelet-leukocyte aggregates, the composition of platelet aggregates, and the size distribution of the contributing platelets. Using image-based flow cytometry, just a few drops of blood were needed to count thousands of blood cells and their aggregates within seconds.

According to study lead Oliver Hayden, a professor at TUM, the method removes the need for treatment and marking of the samples and allows direct investigation with standardized methods without aggregation losses from high shearing forces.

The team used microfluidic polymethyl methacrylate (PMMA) chips with a channel diameter of 500 µm, a height of 50 µm, and a length of 5000 µm for high throughput measurements with precise blood cell alignment in a sub-monolayer. To ensure that >90% of the cells were in focus, the researchers combined viscoelastic and hydrodynamic focusing methods.

Sub-monolayers of blood cells were measured in a microfluidic channel using a transmission digital holographic microscope. The researchers generated interferograms by using off-axis holography combined with a double-shearing interferometric approach.

The image analysis pipeline comprised three steps: preprocessing of the phase images, segmentation, and analysis of detected cells and aggregates.

The researchers created a synthetic data set of aggregates by composing multiple classified single-cell images together to form cell aggregates. Using the synthetic data set, they trained the neural network on 200,000 cell images. They evaluated the performance of the trained network on both computer-generated aggregates and manually labeled images.

Oliver Hayden, a professor of biomedical electronics, led the research team that discovered the connection between platelet aggregates and severity of COVID-19. Hayden develops new methods for in vitro diagnostic and biomedical issues as part of his research. Courtesy of Andreas Heddergott /TUM.

The last step in the processing pipeline was the analysis of the detected cells and aggregates. Based on the segmentation and classification results, image patches were categorized as single cells or cell aggregates. The number of cells in an aggregate, as well as their cell classes and their morphological parameters, were saved for each patch, allowing an in-depth analysis of single components and whole aggregates at the same time.

To the best of the team’s knowledge, this study is the first to show that quantitatively detecting blood cell aggregates in COVID-19 by label-free digital holographic microscopy could provide a way to assess and predict the disease severity in critical care patients at the point-of-care. The researchers observed a high concentration of cell aggregates in all COVID-19 patients in the study who were admitted to intensive care.

This diagnostic method has the potential to identify high-risk patients at an early stage and improve their care. It could potentially support risk stratification and prevent complications in other medical disorders where immune cells are shown to aggregate. “In the future, this cost-effective method could help quantify interactions between the coagulation system and the immune system,” Hayden said.

The interdisciplinary team of engineers and medical researchers now plan to transfer what they have learned to other diseases. They believe that the method described here could also function with cardiovascular disease and cancers.

The research was published in Communications Medicine (www.doi.org/10.1038/s43856-023-00395-6).