When asked to give a “historical perspective” to this special issue on live cell imaging, my mind flew to a time and place where optics and electronics came together allowing new ways to visualize cell dynamics. About 25 years ago I unpacked lab equipment in Whitman 210—a small room at the Marine Biological Laboratory (MBL) in Woods Hole, MA, under the guidance of my postdoctoral advisor, J. P. Trinkaus. When I came to work with him, “Trink” had spent almost 50 summers at the MBL, studying cell behavior in the embryos of a local killifish, Fundulus heteroclitus. My goal was to study a population of cells that roamed the killifish yolk sac, a system that seemed tailor-made for in vivo imaging due to the large size and transparency of the embryos. Our application for lab space that summer described experiments using a new inverted microscope equipped with differential interference contrast optics and high numerical aperture objectives. A new microscope was, of course, cause for great excitement, but it was the attachment of a video camera that was transformative.
Today it is taken for granted that a laboratory full of people can simultaneously watch the behavior of live cells, viewing a monitor of some kind. But before the commercialization of video cameras for microscopy, studying cells was an individual endeavor. Even with a collaborator in the same room, you had to take turns looking through the eyepieces, describing out loud what you were seeing. Since Trink's eyes were set much farther apart than mine, we constantly fought about the separation distance of the binocular eyepieces. Using sketches to orient each other, we would point to our drawings and send each other back to the scope again and again. Our conversations continued days after the embryo had been on the stage of the microscope, once we had developed negatives and printed test strips of the 35 mm images taken during the experiment.
In the early 1980s, two labs with strong ties to the MBL demonstrated how controlling a video signal increased the resolution and contrast of a microscope image (Allen et al.,1981; Inoué,1981). Previously invisible phenomena could be seen in living cells, and our relationships to our microscopes changed. It was still true that the scope itself needed to be in perfect alignment, that light paths and Wollaston prisms were adjusted to obtain the best possible image, that it was worth spending the money for the best objective lenses. Now, however, the electronic signal from the video camera could be manipulated, and by controlling functions such as gain and black level we effectively increased the brightness and contrast of the image. As we turned our chairs to focus on the monitor, details of killifish cell behavior came into view. We could not believe that we were watching the dynamics of cell–cell connections (see Fig. 1) or previously transparent surface folds (see Fig. 2) with such clarity.
The conversion of biological phenomena into electronic signals also opened new ways to store data, and the enthusiastic home electronics market meant that every lab could afford videocassette recorders. About this same time the security industry marketed time-lapse VCRs for cameras in banks and police stations, as an inexpensive way to store weeks of surveillance data. As we unpacked the boxes for our recorders we ran to borrow cables from the electrophysiologists down the hall, learning about BNC connectors and banana plugs. I remember buying our first pack of 10 VHS videotapes and lining them up near our brand-new recorders, knowing that hours of live cell behavior could be recorded, and every experiment easily archived.
Cells and embryos develop on their own time and transformations over many hours and days needed techniques other than real-time imaging. Before the advent of video, time-lapse cinematography was used for the study of cell dynamics. With 16-mm ciné cameras controlled by intervalometers, frames were captured at regular time intervals, compressing hours of behavior into minutes of amazing (grainy, jumpy) sequences. Experiments were planned carefully, film was sent off for developing, and everyone in the lab learned to thread finicky movie projectors. I was proud of my ability to repair old films, cutting and splicing worn segments with little pieces of sticky tape. Video changed all of this, and with a time-lapse VCR we could review the experiment simply by hitting the rewind and play buttons, adjusting the time compression as new patterns emerged. Killifish deep cells were seen to migrate directionally in response to a wound in the overlying epithelium (see Fig. 3) and to analyze these films, we taped pieces of acetate to the monitor, tracing cell outlines and noting frame times. Thermal video printers made this easier, but most analysis still happened by hand.
When I began filming, live cell imaging was predominantly in black and white. Color video cameras were available, but their higher cost and lower resolution kept them out of many labs. The 1984 Nobel Prize for the discovery of monoclonal antibodies celebrated a technique that allowed the labeling of specific cells based on molecular markers, and we were excited by the yellow-green of fluorescein and the red of rhodamine in our eyepieces and photographs. But it was the development of monochrome video cameras able to detect fluorescent signals coupled with the growing availability of vital fluorescent stains that allowed the recording of subcellular dynamics in living cells. We saw the fluorescent world in color when looking through the microscope, but on video monitors it was shades of grey. Fluorescent resources for vital labeling were coming too quickly to keep track of. Companies such as Molecular Probes had newsletters every few months describing new dyes for long-term staining of cell membranes or organelle-specific markers. Now phenomena known only from fixed material could be studied in intact embryos. As the surface cells of the killifish embryo spread to cover the enormous yolk, these tight epithelial cells undergo active rearrangement, extending basal protrusions that underlap neighboring cells. Trink and others (such as Ray Keller) had studied this phenomenon with scanning electron microscopy. By using a new fluorescent lipid that partitioned into plasma membranes and pushing the capabilities of an early silicon-intensified-target camera, videos of these exquisitely thin and dynamic cell extensions were obtained (see Fig. 4).
This was a time of movie data, and we all wanted to share what we were capturing. Tuesday night Cell Motility talks at the MBL attracted an audience ready to be astounded each week by data revealed by new video technologies. Sliding microtubules, intracellular calcium dynamics, mitotic spindles, and sperm acrosomal reactions delighted us. Movie nights at the annual ASCB meetings were, for many of us, the highlight of the trip. A convention hall would darken, the video projector would start, and live-cell sequences would tower over us on huge screens. We would crowd around a speaker after the session, trading bitnet addresses and phone numbers so the conversation could continue, and ask to get a copy of the film. This entailed dubbing the video with a second VCR and mailing it off in a padded envelope. In my collection are tapes whose curling labels in a wide range of handwritings represent the generosity of our community.
When I started teaching as a young assistant professor, I channeled some of my excitement for video data into sharing sequences with my undergraduate students. With ties to the MBL and larger cell biological communities, I became an informal clearinghouse for video sequences that were the most useful in teaching, and soon I copied and mailed tapes around the country. This led me to approach the Society for Developmental Biology to see if there was interest in a teaching video, and within a few years I had published “A Dozen Eggs: Time-Lapse Microscopy of Normal Development.” By this time my lab was equipped with more than a microscope, video camera and time-lapse VCR. The technical revolutions kept on coming, and by the late 1980s image acquisition and processing packages were on laboratory computers. VCRs were put away, as new modes of image storage came to the fore, and we did not often stop to look back. The first time I used the “background subtraction” function on my Image-1 system I was hooked on digital manipulation. Soon we moved from video cameras to confocal lasers and photomultiplier tubes—the science drove the questions, but the technology opened windows we never thought possible. Now we can watch molecular dynamics of living killifish cells as they rearrange in the early movements of gastrulation (Supporting Information 1, 2).
The top shelves in my lab are filled with old data in formats that reflect the evolution of live cell imaging. Sixteen-mm films sit next to VHS tapes, optical memory discs crowd boxes of floppies (large and small format), and a short stack of 100 MB zip drives are mixed in with CDs. Tucked into cabinets are Kodak carousels filled with slides, sleeves of negatives I developed myself, and a treasure trove of brittle Chartpak letters once used to label figures. That images and movies now live on hard drives, and can be shared with the click of a mouse, is astonishing. The revolutions of Green Fluorescent Protein, molecular genetics, and new imaging modalities continually change the landscape for live cell imaging, and I cannot wait to see what is next revealed.