The geometry of photopolymerized topography influences neurite pathfinding by directing growth cone morphology and migration

Abstract

Objective. Cochlear implants provide auditory perception to those with severe to profound sensorineural hearing loss: however, the quality of sound perceived by users does not approximate natural hearing. This limitation is due in part to the large physical gap between the stimulating electrodes and their target neurons. Therefore, directing the controlled outgrowth of processes from spiral ganglion neurons (SGNs) into close proximity to the electrode array could provide significantly increased hearing function.Approach.For this objective to be properly designed and implemented, the ability and limits of SGN neurites to be guided must first be determined. In this work, we engineer precise topographical microfeatures with angle turn challenges of various geometries to study SGN pathfinding and use live imaging to better understand how neurite growth is guided by these cues.Main Results.We find that the geometry of the angled microfeatures determines the ability of neurites to navigate the angled microfeature turns. SGN neurite pathfinding fidelity is increased by 20%-70% through minor increases in microfeature amplitude (depth) and by 25% if the angle of the patterned turn is made obtuse. Further, we see that dorsal root ganglion neuron growth cones change their morphology and migration to become more elongated within microfeatures. Our observations also indicate complexities in studying neurite turning. First, as the growth cone pathfinds in response to the various cues, the associated neurite often reorients across the angle topographical microfeatures. Additionally, neurite branching is observed in response to topographical guidance cues, most frequently when turning decisions are most uncertain.Significance.Overall, the multi-angle channel micropatterned substrate is a versatile and efficient system to assess neurite turning and pathfinding in response to topographical cues. These findings represent fundamental principles of neurite pathfinding that will be essential to consider for the design of 3D systems aiming to guide neurite growthin vivo.

Keywords: dorsal root ganglion neuron; growth cone; neurite guidance; pathfinding; photopolymerization; spiral ganglion neuron.

Figure 1.

Schematic of photopatterning process and pattern characterization. (A) Schematic of photopolymerizing micropatterned substrates. A monomer solution (yellow) is added to a silane coupled glass cover slip (grey). Then the photomask (black outlined pattern) is placed on top of the solution before exposing the system to UV (ultraviolet, 355 nm) light. The monomer polymerizes to form a solid substrate and the photomask is removed. Schematic is representative and not to scale. (B) The photomask (black). This pattern repeats over the photomask to create the topographically micropatterned substrate. (C) A depth, color-coded confocal microscopy image of multi-angled microfeatures patterned in the substrate surface. Scale bars = 100 μm.

Figure 2.

Feature geometry determines the ability of SGNs to navigate complex angle microfeature cues. (A)–(C). Representative epifluorescence images of spiral ganglion neuron (SGN) neurites encountering a 120° angle microfeatures across the amplitude conditions (A) 2 μm (green), (B) 4 μm (white), and (C) 8 μm (purple). The SGNs were labeled with anti-NF200 antibodies. D. Scatter plots of the distance that SGN neurites follow a microfeature once encountering that feature. Median ±95% CI shown in black. n for each sub-condition ranges from 30 to 123 neurons traced. Unpatterned represents the shape of an angled microfeature superimposed onto a flat substrate. Since the data are not normally distributed and there are two sets of independent variables (channel and amplitude), a two-way ANOVA on ranks was conducted. This analysis shows that both microfeature amplitude and angle of turn affect the length that SGN neurites follow the microfeatures, with length increasing with deeper microfeatures and more obtuse turns. p < 0.01. Scale bars = 50 μm.

Figure 3.

Microfeature amplitude promotes the ability of SGN neurites to turn. (A)–(C). Representative images of spiral ganglion neuron (SGN) neurites encountering a 90° angle microfeature across the amplitude conditions (A) 2 μm (green) failing to turn, (B) 4 μm (white) aligning across the microfeature, and (C) 8 μm (purple) making a turn. The SGNs were labeled with anti-NF200 antibodies. (D) Proportion of SGN neurites that successfully navigate a given microfeature turn, n for each sub-condition ranges from 20 to 64 neurite encounters with a turn (supplementary table 1). Since the dependent variable is a proportion and there are two sets of independent variables (channel and amplitude), a multinomial logistic regression was used to assess the effect of channel geometry. This analysis shows that increasing microfeature amplitude improves the ability of SGN neurites to navigate turns. Since no effect was seen from turn angle, data were compiled by channel amplitude. The overall data were compared via Chi-square test with follow-up z-tests, indicating similarly that turning proportion increases with channel amplitude. Error bars represent ± standard error for a proportion for all data. p < 0.05. Scale bars = 50 μm.

Figure 4.

Growth cones are more likely to remain in deeper microfeatures. (A), (B) Representative epifluorescence images from supplemental video 1(a) in which a replated dorsal root ganglion neuron (rDRGN) growth cone encounters and remains in an 8 μm amplitude, 120° angle microfeature through the course of 1.5 h of imaging. (C), (D) Representative epifluorescence images from supplemental video 1(b) in which a rDRGN growth cone encounters and exits a 4 μm amplitude, 120° angle microfeature during 1.5 h of imaging. (E) Percent of rDRGN growth cones that remained in the microfeature throughout the 1.5 h recording. Data are proportions so a Z-test was conducted, which shows that a greater proportion of growth cones remain in the microfeature in the 8 μm amplitude condition. n = 11 and 19 growth cones imaged. Error bars represent ± standard error for a proportion. p < 0.01. Scale bars = 20 μm.

Figure 5.

Growth cones exhibit distinct morphology and behavior on topographically micropatterned substrate. (A) Depth, color-coded confocal image of repeating rows of ridges and grooves substrate used for this experiment (3 μm amplitude and 10 μm periodicity). (B) Representative image of a replated dorsal root ganglion neuron (rDRGN) growth cone grown on an unpatterned substrate labeling with anti-NF200 antibodies and phalloidin (actin). (C) Representative image of rDRGN growth cone grown on the micropatterned substrate in (A). (D) rDRGN growth cone shape was approximated as a spheroid and its prolate ellipticity was calculated (supplementary equation (1)). Data are normally distributed and a t-test shows growth cones on the patterned substrate were more prolate. Error bars represent ±SEM, p < 0.05. (E) The major axis of this spheroid was found and the angle difference between the major axis and the neurite shaft was measured. Data are not normally distributed thus a Mann–Whitney shows that this angle was smaller for the neurons on the patterned substrate. Error bars represent 95% confidence interval. p < 0.001. n = 19 and 23 growth cones measured for both graphs. Scale bars = 5 μm.

Figure 6.

Neurite shafts remain in microfeatures at similar rates when navigating turns regardless of feature amplitude. (A) Representative epifluorescence image of a spiral ganglion neuron (SGN) neurite unable to hold position around a turn through a 4 μm amplitude, 60° angle microfeature. (B) Representative image of SGN neurite holding position around two turns through an 8 μm amplitude, 60° angle microfeature. Red circles indicate turns where the neurite shafts do not hold the turn while blue arrows show turns where neurites that do hold the turn. (C) The fraction of neurites that hold the position in the microfeature after successfully navigating a turn (supplementary table 3). Data are proportions from three treatment groups, thus a Chi square test was done which suggested no significant difference in treatment groups (p = 0.24). n = 49, 45, and 39 neurites observed to successfully make the microfeature turn. Error bars represent ± standard error for a proportion. Scale bars = 50 μm.

Figure 7.

Neurite branching is most likely when neurite turning result is uncertain. (A) spiral ganglion neuron (SGN) neurite turning in response to an 8 μm amplitude, 120° angle microfeature, representing an ‘easy turn’ toward the far right of the graph. (B) Epifluorescence image of SGN neurite exiting a 2 μm amplitude microfeature represents a ‘challenging turn’ towards the left of the graph. (C) SGN neurite branching to both follow and exit the microfeature in a 4 μm amplitude, 60° angle feature, representing an intermediate challenge where branching was seen to be most likely. (D) For each type of neurite encounter (angle of approach and channel amplitude), the proportion of neurites which successfully turns and that branched was determined. Here, we plotted neurite branching as a function of neurites successfully turning to assess their relationship. We sought to compare the relationship between these variables. Thus, a second order polynomial was generated to fit the data. We then compared how well the regression fit the data with an F-test and found r ² = 0.54 which equates to p < 0.01. Scale bars = 50 μm.