A simple microfluidic device and time lapse imaging of embryonic stem cells

One thing that is overlooked in our general understanding of cell biology is how dynamic everything is. During school, you are taught that cells are a tiny round blobs full of even tinier factories performing various pre-defined functions. These cells are each of a certain type and shape and combine to form greater structures that do higher-level functions, such as transmit information over long distances or contract to generate force. When I see cells I see something that is much different. They do not form single defined shapes; they are dynamic and ever changing. They grow and shrink, change shape and move around. They reach out with small arm-like protrusions. They react and respond to their environment. They even can change their identity entirely.

We as humans are a fascinating example of all of these characteristics. The body begins as a single cell that divides to create many more of itself. As these cells divide they communicate with each other, change into other cell types, and move and organize themselves. Eventually they become a person; made up of trillions of cells categorized into over 200 types. Embryonic stem cells represent the earliest cells from which the body forms and are thus considered pluripotent, meaning that they have the potential to become any cell type in the body. Mouse embryonic stem cells (mESCs) are similar to their human counterparts in many regards and are commonly used as model to study these characteristics.

mESCs can be grown in the lab indefinitely. They divide about once every 16 hours and when they get too crowded, we take them and split them onto new dishes. Then, when needed, we can change their environment in ways that lead them to become other cell types, such as heart muscle or the neurons that make up the brain and spinal cord.

Using the Foldscope, my iPhone, a small microfluidic chamber for them to grow in, and an app that allowed me to snap one image every minute or two, I was able to watch them in action.

Fabrication of the microfluidic device

To image with the Foldscope, I used a microfluidic device that creates a protected enclosure to put cells in and that can be mounted on the Foldscope. The sealed chamber affords space in all directions for cells to proliferate and to be bathed in their required nutrient rich media.

The chamber is created out of a PDMS (polydimethylsiloxane) polymer (10 parts pre-polymer mixed thoroughly with 1 part curing reagent) that is degassed and poured onto a micro-fabricated negative relief of the cell culture chamber (20 mm length x 16 mm width x 60 µm height with small inlet and outlet channels to allow entry of cells and media exchange). The PDMS is then solidified by baking at 80C for 2 hours, pealed off of the mold, and attached to a glass slide through a process called oxygen plasma cleaning. (The roof of this design happened to also have 200 µm diameter pillars that served as nice way to generate a rough scale bar).


Seeding cells in the culture chamber

Cells attach to a class of proteins that together make up what is called the extracellular matrix (ECM). The ECM sticks to and coats the surrounding area, and provides a structural support for the cells to grip onto. mESCs are best cultured when attach to gelatin coated surfaces. To coat the chambers I incubated overnight at 4C with a 0.2% gelatin solution.

The next day I rinsed out the excess gelatin, and flowed in fresh culture media (composed of a complex mixture of various nutrients and growth factors). I then added 40 µL of cells to the inlet, suspended in the same media at a concentration of 4 million cells/mL, and gently tilted the device to allow gravity to guide the cells to their desired location. Next, I placed parafilm over the inlet and outlet channel openings to prevent evaporation and incubated at 37C for two hours to allow the cells to adhere to the gelatin-coated surface.



After the cells attached I sealed the inlet and outlet with scotch tape, inverted the device, inserted it into the Foldscope, and was ready to image. I sealed the Foldscope with my phone attached in a sterile container and placed it in an incubator set at 37C.

mouse Embryonic Stem Cells

The first video was shot over an hour and 40 minutes with a photo taken every two minutes.

The second video below was shot over 40 minutes with a photo taken every minute, and with slightly higher magnification. What is fascinating in this next video to me is watching what I thought was a single cell, just below and to the left of the center, jump apart into two cells and then come back together.


6 Comments Add yours

  1. mherring says:

    Wow Josh! Very nice time lapse! What was the height of the sample in the well? Do you think the two cells that jumped apart and together had any preference for stacking the way that they did? Or do you think it was one cell?
    It’s cool to see the extent of the cell adhesion; the cells near the edges of your first video almost look like water droplets. Were you able to see much of the cellular processes?

    Great device and data!

  2. Manu Prakash says:

    Absolutely wonderful post @Josh. Thanks for sharing your expertise of making frugal microfluidic devices. Lovely contrast on the crawling cells. I had never seen stem cells; and now I have..

    How did you calculate the scale bar. In the future; we should have a simple phone app that automagically puts cake bars 🙂

    Wonderful – wonderful. You are quickly rising to Foldscope super user level 🙂


  3. Josh G says:

    @Marie The height of the sample probably varied from 5-30 microns tall, after a few days the stem cell colonies grow to be about 50-60 µm in height. I’m not sure about those two cells, could it possibly be a cell division? With this many cells in view and over that amount of time there must be divisions happening. These cells favor being as tightly packed together as possible, and as they continue to divide their cytoplasms get smaller and smaller. This could explain why they seem to be the same size before and after. Other than divisions, Im not sure about other processes. They do look like water droplets! Especially when they come in contact; I wonder if they deform so that they can maximize their area of contact with each other.

    @Manu For the still image with the scale bar I measured one of the 200 µm diameter pillars in imagej to generate very rough pixel to µm conversion. An app for scale bars would be great!

    Also, does anyone know how to convert mp4 or .mov files to a format that can be edited in imagej? Or of a plugin that would work for this?


  4. Manu Prakash says:

    @Josh: on video processing via imageJ – of things collected on an iPhone; see an earlier post of mine here: https://microcosmos.foldscope.com/?p=12198

    Most of the time Fiji can import movies from cellphones.


  5. laksiyer says:

    This is absolutely brilliant. Now how would a differentiating cell look under a similar setup?

  6. damontighe says:

    Wow! Any more recent work on this ?

Leave a Reply