Skip to main content
Premium Trial:

Request an Annual Quote

Q&A: ISB, UBC Team Develops HTP Microfluidic-Based Live-Cell Imaging Technology


Researchers at the Institute for Systems Biology and the University of British Columbia have developed a fully automated microfluidic platform that enables them to conduct live-cell experiments and manipulate images and on-chip microenvironments in high throughput.

The platform, described online this week in the Proceedings of the National Academy of Sciences, combines programmable on-chip mixing and perfusion with high-throughput image acquisition and processing to perform 256 time-lapse live-cell imaging experiments simultaneously.

Nonadherent cells are captured in an array of 2,048 microfluidic cell traps that can image eight different genotypes over 12 hours in response to as many as 32 unique sequences of stimulation. In total, the researchers found that the technology generated 49,000 images per run.

In the PNAS paper, the investigators reported how they used 12 different devices to carry out more than 3,000 live-cell imaging experiments to investigate the mating pheromone response in Saccharomyces cerevisiae under combined genetic perturbations and changing environmental conditions.

They analyzed 11 deletion mutants, and observed distinct thresholds for morphological switching and new dynamic phenotypes that were not seen under static conditions.

Tim Galitski, an associate professor at the Institute for Systems Biology and a corresponding author on the paper, spoke with CBA News this week about the device and its potential application in drug discovery. Below is an edited transcript of that interview.

Can you give me a little background on this work?

A lot of biological problems involve the dynamics of the system. But most of the scientific methods and scientific platforms that we have in use are similar to what I would call a steady state approach, where you add a reagent or you make a genetic change, and then a variety of things happen, and it arrives at some new state, and you make some observations and measurements about it.

But what happens in between is important to the understanding of how the system works or how you might intervene in it. In addition, I think that many of the things that one wants to understand are inherently dynamic. For example, noise resistance — Can a cell make a decision to respond in a certain way if, say, a brief encounter with some stimulus occurs or not, or does it filter it out? Those things are inherently time dependent.

These are the kinds of things that motivated the project from a biological perspective — to be able to do experiments where we carefully control the microenvironment of cells, in terms of what chemical[s] they are exposed to, and to be able to change that at will over time, and to observe the cells at essentially all times during that process, over many hours.

So we have been using this technology to make advances in the field of microfluidics.

We are using soft polymer-based microfluidics. Currently, a major transition is going on in the design of these microfluidic circuits and devices. It is analogous to what happened decades ago when electronic circuits went mostly from analog to digital, with the invention of, say, a transistor.

The digital nature of electronic circuits enables large integrated circuits to do lots of things on the same chip, and do them in very high throughput.

A similar sort of thing is happening in microfluidics, and our work really takes that to the next level. With these true-sealing microvalves and these soft polymer devices, it is analogous to a transistor, because it enables one to control the flow of liquids or cells in the chip [in a way that is] very on-off.

[ pagebreak ]

What we have developed is a high-throughput, integrated device design that does things like mixing. A multiplexer perfuses an array of cell chambers with media that are mixed on the chip in a computer-controlled way.

Another thing that is new here is that the various studies that are out there so far are not really fully automated or computer controlled. With this large-scale, integrated design, everything about our experiments is fully automated and computer controlled. It is programmed into the computer ahead of time, and then it unfolds over a period of hours. That includes things like the data collection as well, via microscope. So the microscope that is imaging the cells at both visible and fluorescent wavelengths is computer controlled, and is moving through things over a period of about a day, collecting all this data.

Was high-throughput imaging and microenvironment manipulation, such as you have described here, possible before?

Not on this scale; not at this high a throughput. People have previously done these kinds of experiments, where they control the microenvironment of the cells. That has actually been around for some years. The difference is one or a few experiments versus hundreds of experiments. It is just a question of putting more experiments on a chip.

That is why I was emphasizing before the analog-to-digital transition. The limitation has not been just the ability to put more experiments on the chip, it's the ability to have it be controllable, and to have it work robustly over, say, a day of experimentation. So having this highly integrated circuit design enables us to automate the entire process and achieve high-throughput and individualized microenvironment control in hundreds of experiments simultaneously.

How would this technology be applicable in drug discovery?

It's the ability to take compounds of any nature, and to program how cells of different genotypes in all these hundreds of experiments are being exposed to what combination of compounds, at what concentration, and for what duration of time. So one can systematically probe not only panels of compounds, but combinations of those compounds that differ in their concentration, and differ over time. This is important in the dynamic aspects of biology that I mentioned earlier.

The cells are trapped in these perfusion chambers, in these hundreds of experiments sitting on the chip. The chip can take in inputs, which are whatever you would like, really. And you can program the computer that controls the device to deliver those compounds however you want — whatever compounds you want, in what combinations you want, at whatever concentrations you want, and at whatever times you want, in all of those different experiments on the chip.

How exactly does this technology work?

One component of the platform is this soft polymer device that comprises a few layers of this plastic called PDMS. Those are made by a computer-automated design, so we designed the device in much the same way an engineer would design an electronic circuit. So we designed the perfusion chambers, the flow channels, and the control channel.

The basis of this design is that there are essentially two levels: there is a flow layer, where the cells and chemicals are, and there is a control layer on top of it. The control layer comprises lines that you pressurize or do not pressurize in a computer-controlled way. The control lines are laying across the flow lines. When you pressurize one of the control lines, it is like stepping on a garden hose with your foot. You cut off the flow.

That is the transistor element, the flow is on-off, on-off. We designed this device, and we used photolithography to make a mold from that, using a technique very similar to that used to make electronic circuits. This mold has the channels etched in it, and then we just put this plastic polymer on top of it. And it eventually acquires an image, essentially, of that mold.

Then we assemble the device. They are rather small, usually about 1 cm by a couple of centimeters, or something like that. They fit on a microscope slide. In fact, we mount them on a microscope slide.

[ pagebreak ]

We put pins in that microscope slide, through which we deliver, say cells or chemicals, or to which we hook up pressurized lines to control pressure in the valve gates. And this whole contraption fits on a microscope stage. The microscope is computer controlled, so it is moving and auto focusing and collecting images systematically through the device, both in visible and fluorescent wavelengths.

We collect visible images to look at the morphology of the cells over time, and we collect fluorescent images because in our experiments, we incorporated a fluorescent gene expression reporter for a certain pathway. Controlling, say, the flow of liquids into the device are some solenoids that are controlled by a computer card that is hooked up to a computer.

The computer is controlling simultaneously the flow of liquids, chemicals, and cells in the device, and at the same time, it is controlling the movements, auto focusing, and image acquisitions of the microscope.

What is the next step in this work?

We will be making more use of the high-throughput aspects of the technology to perform combinatorial genetics, and adding the dimension of dynamic chemical control, which we have already achieved in this set of experiments.

We also need to make further improvements in the area of tracking individual cells over time. So we are working on ways to immobilize cells in gels within the chip so that it makes it easier for them to be tracked as individual cells over time.

The Scan

Germline-Targeting HIV Vaccine Shows Promise in Phase I Trial

A National Institutes of Health-led team reports in Science that a broadly neutralizing antibody HIV vaccine induced bnAb precursors in 97 percent of those given the vaccine.

Study Uncovers Genetic Mutation in Childhood Glaucoma

A study in the Journal of Clinical Investigation ties a heterozygous missense variant in thrombospondin 1 to childhood glaucoma.

Gene Co-Expression Database for Humans, Model Organisms Gets Update

GeneFriends has been updated to include gene and transcript co-expression networks based on RNA-seq data from 46,475 human and 34,322 mouse samples, a new paper in Nucleic Acids Research says.

New Study Investigates Genomics of Fanconi Anemia Repair Pathway in Cancer

A Rockefeller University team reports in Nature that FA repair deficiency leads to structural variants that can contribute to genomic instability.