Imala Alwis – Senior Research Fellow and Microscopy Manager, Thrombosis Research Group, Charles Perkins Centre, University of Sydney
The Heart Research Institute at the University of Sydney is made up of several research groups. This includes the thrombosis research group, which mainly focuses on platelet biology and treating strokes by breaking up blood clots in a safe and efficient manner – preventing them from blocking blood vessels in the brain.
All research groups in the institute require microscopy, and many of our systems are bespoke setups focused towards answering specific research questions. One such system is used to visualise microfluidics devices which we have developed to mimic blood vessels. This in vitro model is formed of channels moulded from PDMS polymer, and we apply platelets isolated from human blood donors. Platelets can easily activate during the isolation procedure, with the disc shape forming spindle like structures to facilitate coagulation and clot formation, and our platelet isolation protocols have been carefully optimised over decades to avoid this.
The device itself can overlay a glass coverslip treated with matrix protein combinations present on blood vessels, either in healthy tissue, or in the context of vascular injury – as platelets stick to extracellular matrix proteins released when a blood vessel is damaged. Blood samples can be pumped through the channels and we can study how the blood reacts to the matrix. From a clinical perspective, we can investigate different samples from patients with diabetes or blood diseases, and see how their blood reacts in this clean assay system. From a biological perspective, a variety of knockout mice also allows us to interrogate different vascular pathways.
The microscope setup to visualise the microfluidic devices is a customised epifluorescence system, fitted with two cameras with a splitter. For studying platelet dynamics (i.e., the events that happen once platelets become activated), we get the best of both worlds by visualising fluorescence and the 3D morphology with differential interference (DIC) microscopy. An sCMOS acquires images from the fluorescence channel, where we can label lipid proteins to provide insights into morphology of the platelets from a lipid-protein perspective. With the EMCCD camera, we can capture DIC images for studying platelet morphology and identify the stage of activation.
Illumination was previously achieved with a large and heavy arc lamp, but we decided to look at upgrading this when one of the bulbs wore out. Testing the pE-300white with different cameras, we were pleased to see the setup achieves the temporal resolution required to capture the fast events happening in our microfluidics environment.
We chose the pE-300white based on its ability to switch very quickly between DIC and fluorescence channels, driven through software. Additional benefits include the small footprint, and of course the light generated which is considerably brighter compared to the lamp previously used. When studying a process that is stable for an extended period of time, for example in vascular injury, we could be following clot formation continuously in the microfluidic device for 10 minutes or more. It is important to avoid photobleaching during this time-lapse, and with so much light generated from the pE-300white, we can afford to reduce this or use faster exposure rates for fluorescence imaging.
Some studies only require fluorescence information, and we can label different components of a blood clot, including platelets, fibrin, red blood cells and white blood cells to understand the process in more detail. In these cases, we pick two channels (without nuclei, platelets do not require DAPI staining in the UV channel), and we can acquire fast sequential imaging of two channels due to the quick channel switching of the Illumination System.
The system works well for our research, and different research groups use it for different studies, as although we are in different fields of blood research and answering different questions, we all depend on the microfluidics model.