During nervous system development, growing axons navigate towards their targets using signals from their environment. These signals may be biochemical or mechanical in nature; however, the role of mechanical cues in axon pathfinding in vivo, and the spatiotemporal dynamics of embryonic brain mechanics, are still largely uncharacterised. Here, I have identified a role for tissue mechanics in embryonic axon guidance in vivo, using retinal ganglion cell (RGC) axon outgrowth in the developing Xenopus laevis optic tract (OT) as a model system.

Using atomic force microscopy (AFM) to map brain stiffness in vivo, I found that embryonic Xenopus brain tissue was mechanically heterogeneous at both early and later stages of OT outgrowth, i.e. just before RGC axons make a stereotypical turn in the mid-diencephalon, and when they reach their target, respectively. The final path of RGC axon turning followed a clear mechanical gradient: by the later stage, tissue rostral to the OT had become stiffer than tissue caudal to it. This mid-diencephalic stiffness gradient was an intrinsic property of the underlying brain tissue, and correlated with local cell body densities (with higher density rostral to the OT and lower density caudal to it). Crucially, inhibiting cell proliferation in vivo during OT outgrowth abolished the stiffness gradient and reduced OT turning at the later stage.

Next, I developed a time-lapse AFM technique to track tissue stiffness and RGC axon behaviour simultaneously in vivo. Using this approach, I followed the evolution of the mid-diencephalic stiffness gradient during intermediate developmental stages, around the time when the OT’s caudal turn is initiated. The stiffness gradient was shallow pre-turn, but increased in magnitude during axon turning (mostly due to an increase in tissue stiffness rostral to the OT). This increase in stiffness gradient preceded the rise in OT turning angle, suggesting that the stiffness gradient is not caused by the invading axons. The observed rise in stiffness gradient correlated with stage-specific increases in local cell density, and was attenuated by blocking mitosis in vivo during time-lapse AFM measurements (which also reduced OT turning).

As final confirmation that brain stiffness contributes to RGC axon pathfinding, I disrupted mechanical gradients by artificially stiffening brain tissue in vivo. Importantly, global stiffening via application of transglutaminase eliminated the mid-diencephalic stiffness gradient by increasing tissue stiffness caudal of the OT, and reduced the OT turning angle. Sustained mechanical compression of small areas using an AFM probe stiffened brain locally and repelled RGC axons, consistent with the way they turned away from rapidly stiffening tissue regions during time-lapse AFM experiments. Taken together, these results are consistent with a function for tissue mechanics in axon pathfinding in vivo.