High-Throughput Behavioral Aging and Lifespan Assays Using the Lifespan Machine

In our lab, we study the physiological and molecular determinants of aging and lifespan, and we do so in a quantitative manner. For this, we use the nematode C.elegans as a fast-aging, genetically-tractable model, and we use it with a combination of tools such as the lifespan machine. Reproducibility is an important issue in aging research.

To overcome this, large population of animals are often required for measuring robust changes in lifespan and therefore, this process really benefits from automated high throughput protocols. Lots of people have used the lifespan machine for many different purposes and are really happy about that. For our own applications, the two most significant findings are, first in 2016, our discovery of temporal scaling as a common outcome of many interventions in aging.

And second, in 2022, our discovery that healthspan and lifespan are determined by distinct underlying physical declines. The lifespan machine adapts commonly used agar plate protocols to collect survival and behavioral aging data from large populations of nematodes by automation and high frequency of image acquisition. It eliminates the daily tedious process of by-hand survival assays and gathers previously unattainable datasets on nematode behavior.

To begin, ensure the scanner plates are ready with nematodes loaded. In general, about 40 animals and 200 microliters of bacterial lawn per plate are optimal conditions for a lifespan lawn experiment. Start the experimental setup by applying an anti-fog glass cleaner on both sides of the panes of glass that will support the scanner plates.

Then, apply a protective hydrophobic glass treatment on the side of the glass that will be in contact with the scanner plates. Wait five to 10 minutes and remove any residue on the glass pane. Apply 70%ethanol to disinfect the side of the glass that will be in contact with the scanner plates.

Leave for one to two minutes and then remove the ethanol with a cloth or paper towel. to load the plates onto the scanners, place the autoclaved rubber mats on top of the treated plate support glass. Take the scanner plates used for imaging with loaded nematodes and remove the lid.

Place the plates on rubber mat locations facing the glass surface. Unplug the scanner fans to protect the experimenter's fingers during plate loading. Gently slide the plates and the glass sheets supporting them onto the surface of the scanner.

Reactivate the scanner fans and confirm that the front and side fans are powered. Turn on the scanners at this point. Before starting the experiment, perform the essential pre-image acquisition steps, including obtaining a preview capture image of the entire scanner surface.

Using the web interface, find the image acquisition section of the main page and follow the link called Capture Devices and Image Servers. In the opened box, click on Search for New Devices and ensure that each scanner attached to the server appears in the image capture devices box. Select the check box corresponding to each scanner containing newly loaded plates, and then click on the Request Preview Capture button.

Next on the worm browser, select File, Open Image, and then choose the desired image to open each preview capture image. For each image, click to select the columns corresponding to regions containing plates. Once all the images are defined, select Image Acquisition, Define Scan Areas, save selected scan areas to disc, and then choose the desired location to export the region specifications to disc.

Assemble a file that contains the experiment name, the physical locations of each column on the scanners, the total duration of the experiment and the image capture frequency. Save this as a text and an XML file. On the worm browser, click on Image Acquisition, Submit Experiment Schedule, and choose the generated XML file.

The worm browser will ask whether to output a summary of the schedule or run the experiment. Generate a summary file. Verify the dates of scheduled captures, the location, name, and number of scanners in the summary.

Again, load the XML file for the experiment schedule as demonstrated. The worm browser will prompt a second time whether to output a summary of the schedule or to run the experiment. This time, choose to run the experiment.

The lifespan machine will now work autonomously until the final scans, specified on the experiment schedule, are completed. After the experiment has ended, and all image and postage acquisition steps have been performed, manually validate the automated results by using the worm browser. On the worm browser menu, click on File, Select Current Experiment, and choose the desired experiment.

To generate the storyboard, click on Validation, Browse Entire Experiment, and then select Immediately After Each Worm's Death. Right click on the object's image to exclude non-worm objects from the storyboard. To exclude all objects on a page simultaneously hold the control key and right click on any object.

Right click twice to re-include an excluded object in the analysis. To save the annotations made during the storyboard annotation, click on the Save button. Left click once on any object of the storyboard to open a new window that displays detailed time series information about that object.

To manually annotate the death times, left click on the bottom bar at the point corresponding to the time of death. Use the space bar and right or left arrows of the keyboard to move through the timeframes or click directly on the bar at the desired timeframe. Manually annotate the contraction and expansion events by right clicking on the bottom bar at the desired timeframe.

See the survival curves on the left side and the scatter plot comparing the time of vigorous movement cessation to the time of death for each individual on the right side of the storyboard. To generate the death times for an experiment on the worm browser, select Data Files, Death Times, and then click on Generate Death Times For Current Experiment. The effect of manual data validation on survival curves is shown here.

The initial survival curves, produced before storyboard annotation, were distorted by improper inclusion of the non-worm objects. After excluding non-worm objects during manual storyboard annotation, the resulting survival curves displayed significantly improved resolution. The graphical representation demonstrates the estimated average remaining lifespan across different steps of the post-image acquisition analysis and its correlation with worm lifespan.

It was found that manual annotation of death times in behavioral aging analysis is especially important in longer living worms. In this instance, those exposed to the lowest concentrations of oxen. In conclusion, the lifespan machine automated death time estimation is sufficiently accurate for survival assays.

While behavioral aging assays can benefit from manual annotation.