The Beat Goes On

By James Badham, UCSB Engineering

UC Santa Barbara bioengineering professor Beth Pruitt has spent years studying cardiomyocytes, the cells responsible for human-heart contractions, first as a professor at Stanford University and, since 2018, at UCSB, where she is the inaugural chair of the new Bioengineering Department, which she was instrumental in establishing. Now, researchers in Pruitt’s lab, in collaboration with colleagues at Stanford, have developed software to enable high-throughput observations of the contractile dynamics of individual cardiomyocytes derived from human induced pluripotent stem cells (hiPSC_CMs). The work appeared in a paper titled Tracking single derived cardiomyocyte contractile function using CONTRAX [Tracking of Contractile Function over Time], an efficient pipeline for traction-force measurement,” published in the June 26 issue of the journal Nature Communications. 

“I developed CONTRAX to make it easier to perform the type of experiments I was interested in,” says first author, Gaspard Pardon,who as a postdoc in the Pruitt lab led implementation of CONTRAX. Using the software, eventually I could attend our lab meeting while data acquisition was going on automatically, saving lots of time that I could dedicate to data analysis. Without CONTRAX, I could not have generated the amount of data we published here during my entire postdoc.”
 
Current assays of cardiomyocyte contractile function in response to perturbations that include drugs, environmental changes, and disease mutations are often mostly qualitative, lack standards applied across this field of research, and generally fail to offer the scalability and throughput necessary for larger-scale studies that would have increased statistical power and, therefore, relevance, the authors note. 

Recent advances involving the use of high-speed video and traction force microscopy (TFM) have made it possible to quantify total contractile force in cells; however acquiring and analyzing TFM data remains a time-consuming process requiring significant expertise. For instance, it takes several hours to acquire video recordings for each experiment, a process lengthened by the need to manually select cells of interest to record. Additionally, data processing places high demands on both computing power and user involvement, given the size of the datasets resulting from the ten seconds of high-frame-rate video needed to capture each CM contraction.  

To do such work previously, Pruitt says, “A user would sit in front of a microscope and, using only their own training, scroll around the dish to make the best guess as to which cell they should video-record. They would record that, then find another cell, and repeat. Working quickly, they could do maybe twenty or thirty cells in an hour.”
 
Because of that low achievable throughput, which limits the ability both to measure enough of the highly heterogeneous single cells to generate relevant results and ensure that enough cells survive during a long-lasting experiment, making it possible to track changes in the same cells over time, Pruitt says, “The potential of such single-cell-resolution measurements has been largely untapped.”

“As with any science or medicine, the more observations you can make, the better your statistics and the power of observation will be,” Pruitt says, summing up. “This software allows us to increase the throughput of our observations of individual cells’ responses to various stimuli. That, in turn, makes it possible to determine mean population statistics — how these cells respond, on average — and so to begin to isolate the signal from the noise.” 

CONTRAX, which the authors describe as a “versatile, streamlined, open-access pipeline,” makes possible “quantitative tracking of the contractile dynamics of thousands of single hiPSC-CM over time,” at a very much increased rate of throughput. That, in turn, makes it possible to “reveal converging maturation patterns, quantifiable drug response, and significant deficiencies in hiPSC-CMs that carry disease mutations. 

The following three software modules comprise CONTRAX:

Module 1 provides parameter-based cell identification based on user specifications for, say, certain shape changes or other characteristics. “If you want to count only cells that are larger than one hundred microns, or only cells that glow green, you provide that input, and the software shows you only those cells and takes images of them. From this, CONTRAX will then create a position list — a list of cells that should be measured and their position. 

Module 2 is for the more time-intensive part of the study. “All of the microscopy for these studies is done outside the typical environment in which cells are maintained,” Pruitt explains. “If they are out of that environment for too long, something will be out of equilibrium for them, and they will start to change. For instance, they are sensitive to too much ultraviolet light, which they are exposed to during the process. They’re also sensitive to fluctuations in temperature, which is harder to maintain on a microscope than in an incubator. So, you want to work as quickly as you can. For our heart-muscle cells, we need to see about ten seconds of video per cell in each fluorescent channel that we look at, so, that ends up being about half a minute per cell, which limits how many cells you can collect data from before they become ‘unhappy’ being out of the incubator. This module allows us to do that in a more streamlined fashion and take video of only the cells that met our criteria in the first module. We can then analyze contractions via traction force microscopy.”

Module 3 is used to stitch together video images to quantify mechanical function and create a timed trace from the heartbeat in a dish, important in trying to figure out the peak force, and contraction and relaxation velocities. From there, Pruitt says, “We can determine if there are differences in response to our changing parameters between the different experimental treatments.”

The Pruitt lab used CONTRAX to follow the same cells at multiple time points over twenty days to learn which media promotes “better” cardiomyocyte viability and function, and to quantify differences between multiple cells that have and do not have disease mutations, and to compare the contractility of cells before and after drug treatments. They continue to use it in ongoing studies of the effects and mechanisms of additional mutations and treatments, and hope that the open-source methods and code will help other researchers to increase their throughput as well.