Fig 1.
Microfluidic device for parallelized low-density worm culture.
(A) The microfluidic chip is designed to have two arrays of 50 growth arenas. Flow through the system enters in region 1, a perfusion inlet on one side of the chip connected to a distribution network that ends in an open space upstream of the loading chambers. Flow then passes through the loading chambers (region 2) and enter the growth arenas (region 3). Flow exits each of the growth arenas through a filter (region 4) and reconnects to a single outflow network (region 5). The exit ports at the end of region 5 are located on the opposite side of the chip from the perfusion inlets to avoid shadowing by overhanging tubing. (B) Each arena includes a 2.2 mm x ~0.9 mm “crawl-zone” of 0.2 mm diameter pillars spaced 0.1 mm apart (see [12]) (region 3.2). Upstream of the pillar array is a “swim zone” (region 3.1) that stretches 1.1 mm from entry point to the first set of pillars. The combination of an entry plug and flow keeps loaded animals in the arenas. (C) Automated loading is facilitated by holding chambers with 0.028 mm wide constriction points that enable the parallel pre-loading of 50 animals which are then moved into the chambers with a short burst of increased pressure (adapted from [10]).
Fig 2.
Scalable pressure drive and distributed imaging system.
(A) Pressurized air was split through a manifold with ball valves to 1-gallon air tanks. Air from the 1-gallon tanks was regulated using a pressure-relieving micro-regulator to 1–5 psi. The regulated air was then passed through a t-split to two 1-liter bottles with custom fabricated lids that had seven 1.5 mm OD stainless steel stems that passed through the lid. One stem was used to connect the pressurized air to the air in the bottle, while the remaining six stems were connected internally to tubing that reached ~50 mm from the bottom of the bottle to allow for a stir bar, and externally to 1.5 mm ID tubing that ran to the profusion inlet ports of the microfluidic chips (see Fig 1A). (B) Microfluidic chips on 50 mm x 75 mm glass slides were mounted in the imaging zone approximately 12 mm from the edge of the scanner, with three chips on each side. (C) Image capture scheduling was performed using a Linux implementation of the C. elegans Lifespan Machine software [4] and images were stored on a local network attached storage device. Image processing was performed using custom MATLAB code (see Materials and methods) run on a Windows 7 desktop.
Fig 3.
Scanner temperature stability.
(A) Temperature probes were mounted on glass slides as PDMS chips are and arrayed to cover the entire surface of the scanner. The outermost columns of probes lie in the imaging zone where microfluidic chips would be mounted. (B) Temperature data from a mock run shows little variability in temperature over time. The cycling of temperature that is observed corresponds to HVAC cycling in the temperature-controlled room not with the imaging cycle of the scanner. (C) Measurements on different scanners on different days are consistent.
Fig 4.
(A) Images of N2 animal at 1, 30, 35, and 47hrs showing the characteristic “bag of worms” death. (B) Time to death for animals in M9 buffer (Number of animals = 244 (N2), 327 (fog-2(q71)), 263 (mated fog-2(q71))). *** signifies p<0.0001. The N2 versus mated fog-2 comparison is not significantly different, with a p = 0.87. (C) N2 Lifespan curves in M9 from day 1 of adulthood. Each solid line is one of six different sample lanes representing between 34 and 49 individuals. The black dotted line represents the curve of all N2 animals. No significant differences are seen between replicates. (D) Lifespan curves in M9 from day 1 of adulthood. Log Rank p<0.0001.
Fig 5.
(A) Images of a representative fog-2(q71) animal at 1, 12, and 24hrs showing the characteristic rod-like death. (B) Lifespan curves for fog-2(q71) animals in M9 buffer supplemented with increasing concentrations of H2O2 (Number of animals = 379 (0.25 mM), 3518 (0.5 mM), 276 (0.75 mM), 93 (1 mM)). (C) Power fit of median lifespan versus concentration of H2O2 yields an equation (y = 10.503x-1.394) with an R2 of 0.9976. (D) Comparison of fog-2(q71) and fog-2(q71) daf-16(mgDf47) animals at 0.25 mM H2O2 (Number of fog-2 daf-16(mgDf47) animals is 498). Log-Rank p<0.0001.
Fig 6.
(A) Images of a representative fog-2(q71) animal at 1, 12, and 24hrs showing the characteristic squat worm phenotype at NaCl application, and recovery over time. (B) Lifespan curves for fog-2(q71) animals in M9 buffer supplemented with increasing concentrations of NaCl (Number of animals = 297 (300 mM), 792 (400 mM), 508 (500 mM)). Log-Rank p<0.0001. (C) Comparison of fog-2(q71) and fog-2(q71) daf-16(mgDf47) animals at 500 mM NaCl (Number of fog-2(q71) daf-16(mgDf47) animals is 450). Log-Rank p = 0.0005.