Laterally orienting C. elegans using geometry at microscale for high-throughput visual screens in neurodegeneration and neuronal development studies

Abstract

C. elegans is an excellent model system for studying neuroscience using genetics because of its relatively simple nervous system, sequenced genome, and the availability of a large number of transgenic and mutant strains. Recently, microfluidic devices have been used for high-throughput genetic screens, replacing traditional methods of manually handling C. elegans. However, the orientation of nematodes within microfluidic devices is random and often not conducive to inspection, hindering visual analysis and overall throughput. In addition, while previous studies have utilized methods to bias head and tail orientation, none of the existing techniques allow for orientation along the dorso-ventral body axis. Here, we present the design of a simple and robust method for passively orienting worms into lateral body positions in microfluidic devices to facilitate inspection of morphological features with specific dorso-ventral alignments. Using this technique, we can position animals into lateral orientations with up to 84% efficiency, compared to 21% using existing methods. We isolated six mutants with neuronal development or neurodegenerative defects, showing that our technology can be used for on-chip analysis and high-throughput visual screens.

Figure 1. Microfluidic device to passively orient C. elegans for visual screening.

(A) Device used for orienting, imaging, and sorting animals. Flow layer is shown in green with black text, valve control layer in red with white text. LD is loading valve, MT is channel for mutant output, and WT is channel (black) and valve (white) for wild-type output. IMG is imaging valve. Not labeled are the mutant valve (right of mutant channel) and flush channel (green area above red box). Input and flush channel fluid flow are controlled off-chip. Imaging area is indicated by red dashed box. Scale bar is 800 µm. (B) Device schematic of valve state during worm loading. Worm is driven into imaging area using positive input pressure to induce fluid flow. Loading valve is not actuated, allowing fluid flow. Imaging valve is actuated to prevent worm from exiting imaging area. Wild-type valve is not actuated to assist in worm loading and to provide an exit should an animal slip past the imaging valve. Mutant valve is actuated to prevent animals from entering mutant output. (C) Device schematic of valve state during analysis. All valves are actuated and input pressure is cut off halting fluid flow. (D) Device schematic of valve state during worm sorting. Example for wild-type sorting is shown. Worm is driven out of the imaging area by positive pressure from the flush channel. Loading valve is actuated to prevent any other animals from entering the imaging area. Imaging valve is not actuated to allow worm exit. Wild-type valve is not actuated to allow worm exit into wild-type output. Mutant valve is actuated to prevent animals from entering mutant output. Mutant and wild-type valve state is reversed when mutant sorting is performed. Fluid flow direction indicated by white arrows and is not proportional to size of arrow. Non-filled and filled red boxes indicate non-actuated and actuated valves respectively.

Figure 2. Effect of curved channel on animal orientation.

(A) Zoomed in view of imaging area shown by dashed red box from Figure 1A. White arrow indicates device radius of curvature (RoC) from arc center to outer edge. Dashed line indicates length (L) between loading and imaging valves. Scale bar is 200 µm. (B) Frequency of lateral nematode orientation for various channel geometries with standard error of proportion. Triple asterisk indicates statistical significance compared to straight channel designs (p<0.001 determined using chi-squared test). S represents straight channel, while remaining labels indicate the 105 µm, 125 µm, and 145 µm RoC devices respectively. L for all devices is 700 µm. (C) Nematode oriented laterally in curved channel device (both nerve cords visible). Commissures present on different focal plane are obscured. (D) 3-D model of animal body section and microscope objective (viewpoint reference) showing nerve cord placement for a laterally oriented animal. (E) Nematode in a non-lateral body orientation as observed when loading animals into straight channel devices; DNC not visible due to animal orientation. Animal is within the same field of view as seen in panel C. Arrowheads for images (C) and (E) indicate ventral nerve cord (VNC) determined by placement of VD and DD motor neuron cell bodies. Arrow indicates dorsal nerve cord (DNC). Scale bars are 100 µm. Transgene marker for all fluorescent images is juIs76(Punc-25::GFP). (F) 3-D model of animal body section and microscope objective for non-lateral oriented animal. Model diagrams (D) and (F) not drawn to scale. Red lines illustrate dorso-ventral axis. Green spheres represent DD and VD neuron cell bodies.

Figure 3. Morphology of wild-type and defective D-type motor neuron commissures.

(A) Full body image of adult worm. Scale bar is 25 µm. (B) Expanded view of area within dashed box in panel A. Ventral nerve cord (VNC) is distinguished by presence of neuronal cell bodies. Dorsal nerve cord labeled DNC. Scale bar is 25 µm. (C) Representative image of panel B. Black arrows label neuronal commissures while black arrowheads identify neuronal cell bodies. Respective phenotypes characterized by: (D) commissure never reaching dorsal nerve cord; (E) commissure running along the opposite side of the animal's body (left, colored yellow); (F) absent commissure; (G) additional commissure present; (H) two commissures crossing over each other; (I) two commissures entering the dorsal nerve cord or leaving the ventral nerve cord together, they may also partially fasciculate; (J) bifurcating commissure; (K) neighboring commissures joined by a process; (L) neurite with length less than half nematode width; (M) an absence of GFP expression along either dorsal or ventral nerve cords; (N) break in GFP expression in a commissure. Scale bar is 25 µm. Representative image of phenotype shown beneath each photo accompanied with illustrative phenotype symbol. (O) Proportion of animals in a population with at least one incidence of each independent defect. Error bars represent standard error of proportion.

Figure 4. Adult phenotype characterization.

(A–G) Proportion of animals in a population with at least one incidence of each of the independent defects: guidance, handedness, missing, and gap for wild-type (juIs76) and newly isolated mutants. Asterisks indicate statistical significance of phenotype compared to wild-type. Single, double, and triple asterisks indicate p<0.05, p<0.01, and p<0.001, respectively. Pound signs indicate statistical significance of phenotype compared between a073 and a074. Double and triple pound signs indicate p<0.01 and p<0.001, respectively. Statistical significance determined using chi-squared test. Number of animals examined for each allele labeled in graph. Error bars represent standard error of proportion.

Figure 5. L1 phenotype characterization.

(A–G) Proportion of animals in a population with at least one incidence of each of the independent defects: guidance, handedness, and missing defects seen for wild-type (juIs76) and newly isolated mutants. Asterisks indicate statistical significance of phenotype compared to wild-type. Single, double, and triple asterisks indicate p<0.05, p<0.01, and p<0.001 respectively. Statistical significance determined using chi-squared test. Number of animals examined for each allele labeled in graph. Error bars represent standard error of proportion.

Figure 6. Cell specific characterization for L1 animals.

(A–G) Penetrance of defects in each DD neuron for wild-type (juIs76) and isolated alleles. Asterisks indicate statistical significance of combined phenotypes per cell compared to wild-type. Single, double, and triple asterisks indicate p<0.05, p<0.01, and p<0.001, respectively. Statistical significance determined using chi-squared test. Number of animals examined for each allele labeled in graph. Error bars represent standard error of proportion for sum of phenotypes per cell.