Keeping life in focus isn’t easy for researchers studying living cells, and the greatest part of the problem isn’t that the cells are moving.

Whether capturing images over time or trying to record events in multiple locations in rapid succession, maintaining a sharp focus is one of the investigator’s most thorny challenges. Axial fluctuation over minutes, hours, or days is called focus drift, and occurs independent of living specimens’ movements. Similar problems face researchers who rapidly capture sequential images from multiple well plates, do image tiling, or implement other high-throughput or repeat-focus applications. Not only are the specimens often at slightly different focal planes, but the experimental environment and setup itself can contribute to additional focus problems as the apparatus moves from one image area to the next.

The issues at hand
There are several system characteristics that can contribute to focus problems:

Gravity. A fully loaded objective turret places weight on the focus mechanism, which in turn may cause focal drift. This can be partially solved by placing fewer objectives on the turret.

Immersion media. Many microscope objectives require oil, water or glycerin immersion media, but their viscosity, refractive index, surface tension, evaporation rate, and other properties can change or vary as the experiment continues, causing the focal plane to shift. Some researchers use a gasket that forms a seal between the objective and the coverslip to help address this issue.

Temperature changes. Thermal drift can be caused by different rates of expansion and contraction in the specimen vessels and microscope components. Environmental temperature shifts can also be at fault. A variation of just one degree Celsius may be enough to cause drift to a completely different part of the cell. Insulating environmental enclosures are helpful; they should enclose not only the specimen but also the entire microscope system to be effective. One other issue is that objective heaters and stage-top incubators are not always compatible with immersion-dependent objectives.

Vibration. Not only can sound, airflow or imperceptible motion in the environment— such as that caused by elevators, foot traffic, or doors opening— affect focus, but electromechanical elements within the experimental set-up itself can also cause the loss of focus over time. Vibration reduction systems including isolation tables and separate flooring can help resolve this problem, but may not provide a complete solution.

Minute focal plane differences. Multi-well plate applications rarely have a consistent focal point among all the wells in an array, and similar limitations apply virtually any time that multiple points of interest are imaged during a single experiment.

Software solutions
DIC time—lapse observation including (top) HeLa cell proliferation over 0-36 hours; (bottom) cell division captured every 15 minutes over 0-3 hours; all images captured using the Nikon Perfect Focus system and CFI Plan Fluor 40x, NA 0.6 objective. Images: Nikon. Click to enlarge.
Software routines that rely on edge detection have been engineered into a variety of scientific imaging packages to help correct focus. These are particularly useful in fluorescence imaging because of the high degree of contrast between a fluorescently labeled specimen and its background. However, to calculate focus, the system must first capture several images of the fluorescent specimen, increasing the negative effects of photobleaching and phototoxicity with each exposure. Even more important, software-based routines may take as long as six seconds to implement, making them much too slow for high-throughput applications and frequent-exposure time lapse experiments. Finally, despite recent improvements, most contrast-based software systems remain subject to reporting false positive edge detections.

Focusing on TIRF
Over the last year or two, several companies have introduced hardware-based solutions to help ensure sharp focus for imaging. All have in common the premise that experimental live cells are not free-floating. Cells are usually attached to the coverslip via a glycoprotein such as laminin. Thus, detection of the external coverslip face by measuring reflectance using a weak infrared (IR) laser allows the mathematical calculation of the location of the cell and determination of the correct focal plane.

Applied Scientific Instrumentation, Eugene, Ore., offers a system that measures the focus position using the reflectance of the IR beam via total internal reflection from the coverslip. The solution works well for investigators doing time-lapse total internal reflection fluorescence (TIRF) imaging with high numerical aperture objectives, a particularly thorny application for holding focus. ASI’s closed-loop CRIFF system, developed in conjunction with researcher Clare Waterman-Storer of the Scripps Research Institute in San Diego, easily retrofits to the photo port of most research microscopes, and shares the optical path being used for the experiment itself.

According to the company’s chief scientist Gary Rondeau, the ASI system is also now in use in a French laboratory where it is being used in high-throughput applications. The main issue yet to be resolved, according to Rondeau, is ease of alignment when installing the system. TIRF applications are also usually limited to oil immersion optics with NAs of 1.4 or higher, as objective NAs must be higher than the refractive index of live cells (around 1.33-1.38) in order for reflectance measurement to occur.

Exploiting natural light
Nikon, Melville, N.Y., and Olympus, Center Valley, Pa., have both introduced systems that take advantage of the natural reflection of light from the coverslip. Olympus’s Zero Drift system operates by comparing the experimental image with a reference image created by reflectance; a CCD measures the reference image’s intensity. Based on the intensity information acquired, the system compensates for thermal and mechanical variation, then offsets to a user-defined value of Z, providing sharp focus for the specimen.

Kymography is a way of recording and measuring motion, most commonly used to show undulations of muscle activity or to show seismic activity during earthquakes. Here, kymography is used to show relative focus stability with and without Zero Drift focus compensation. Time-lapse images of sub-resolution microspheres were captured with their corresponding kymographs (column to left of each image). The beads were imaged via total internal reflection (TIRF) with Zero Drift compensation enabled (A) and disabled (B). Images: Olympus America Inc.
With Nikon’s system, when the user focuses on a plane of interest and activates the Perfect Focus system, an infrared light-emitting diods (LED) source is projected through the objective identifying the surface of the coverslip (the interface of glass and medium in immersion applications or glass and air in dry applications), and continuously feeds the position data back to the focusing mechanism to maintain precise focusing. These systems can correct for drift in milliseconds—a distinct advantage over protracted software-based autofocus routines. Software can synchronize the process with peripherals such as wavelength switchers or motorized stages.

“People underestimate how important focus compensation systems are,” says Alexey Khodjakov, a research scientist and associate professor at the Wadsworth Center in Albany, N.Y., “not only for imaging over the long term, but also for imaging multiple adjacent fields. They expand the investigator’s ability to stay with the coverslip while moving the lens.” Khodjakov’s own work involves imaging nine adjacent fields to create multi-image mosaics using high magnification objectives.

Recently, the Nikon and Olympus systems have become compatible with confocal microscopy, opening up a new avenue of research exploration. The Olympus Zero Drift IX81-ZDC can be integrated with the company’s FV1000 laser scanning confocal microscope. The system is in use, for instance, in the lab of Christine Lavoie, Canada Research Chair in Cellular Pharmacology at the Univ. of Sherbrooke in Québec, Canada. Nikon too has integrated its Perfect Focus system with its Confocal EZ-C1 software for the C1 confocal products, as well as its NIS-Elements Software for the LiveScan SFC confocal product, for use in long-term time course studies involving confocal microscopy.

Distinct systems
While all of the focus offset systems currently available are designed to measure reflectance from the coverslip, they have a number of differences. First, there are substantial price variations among the various companies’ offerings, so scientists have considerable flexibility in configuring a system to their needs.

Another important area of differentiation is the capacity of various systems to handle complex imaging of multiple features at different x, y, and z locations within the confines of a single experiment. Still another variation involves how frequently the various systems “retune” focus. Nikon offers continuous focus tracking of the coverslip interface with a separate imaging offset light path while the Olympus system autofocuses at the instant each image is captured.

Defeating the problem of focus drift is a key consideration to be taken into account during configuration of an experimental system designed for either high-throughput or time-lapse imaging. The problem is even more acute when using high-NA objectives where the narrow depth of focus requires very tight focal specifications. It is also a considerable challenge when imaging a complex sequence of x, y, and z points within one specimen or over a series of specimens. Thermostatic controls, environmental and incubator chambers, lighter-weight components, isolation tables, and the new focus drift compensation systems each have a place in making the sharpest images possible.

—Ilene K. Semiatin
Ilene Semiatin is a freelance writer based in White Plains, N.Y.