Scientists model brain growth on a chip

A team of Hopkins researchers has created a device that effectively mimics the complex biochemical environment present in the developing brain. The group, headed by Andre Levchenko at the Whiting School of Engineering, published their findings in the miniaturization-oriented journal Lab on a Chip.

The developing brain can reasonably be compared to Los Angeles rush hour at its worst. Billions of neurons each send out their sole axon - the long, tentacle-like arm of the neuron that makes connections with other cells - into a chaotic mess of tissue, protein and fat.

These axons, having traveled over relatively large distances and having passed by hundreds of other cells ripe for connection, usually end up where they're supposed to be and contact their proper targets.

Only within the last decade have scientists begun to understand how neurons accomplish such sophisticated migration. Neurons are known to receive a huge range of cues from the external environment.

These guidance molecules interact with the growth cone - the axon's enlarged tip - and modulate the direction and speed of its migration.

While guidance cue identity is important, perhaps more significant is guidance cue concentration. Cues are not homogenously distributed throughout the developing brain.

Rather, growth cones are thought to move toward areas where a guidance cue is highly concentrated and away from areas of low concentration, or vice versa.

Nonetheless, this is mostly theory. Studying the navigation of growth cones through concentration gradients in living tissue has been difficult; the developing brain's malleability hinders standardized research.

Researchers, including the Levchenko group, have thus looked to create an artificial environment that mimics that of the developing brain and allows for relatively easy quantification of growth cone migration.

Doing so, though, has been tricky. Current techniques are limited in that they permit researchers to study only one type of guidance-cue gradient at a time.

This is an oversimplification of the developing brain's true complexity. Indeed, many different kinds of gradients exist in the developing brain, and they likely interact in intricate ways.

With this in mind, the Hopkins group set out to construct an artificial environment that would allow them to study if and how two different kinds of gradients interact with one another to affect growth cone movement.

Specifically, the team chose to study brain-derived neurotrophic factor (BDNF) - a diffusible (i.e. freely-moving) molecule pumped out by other cells into the local environment - and laminin, a molecule that's linked to a network of mini-scaffolds outside the cell called the extracellular matrix.

Two gradients - one diffusible, the other surface-bound - is a much more physiologically realistic situation, but the researchers had another hurdle to overcome, this one involving the difficulty of growing neurons in a dish.

Previous attempts have been stymied by the effect of shear stress on the developing cells.

As growth cones move parallel to a surface, a certain amount of tangential force serves to slow them down. This comes both from the surface itself and the fluid they're being grown in.

The team solved the problem by growing axons within small depressions in the surface of their device. These "micro-wells" served to shield neurons from the dampening force.

The team then set out to prove the effectiveness and efficiency of their device. Over a series of experiments, they combined diffusible and surface-bound gradients in different ways in order to quantify how the developing neurons responded to different environmental conditions.

For example, neurons exposed to a gradient of surface-bound laminin but constant BDNF nearly always turned toward the area where laminin was most highly concentrated. In a uniform coating of laminin, however, growth cones turned away from areas high concentrations of the diffusible BDNF.

The most significant finding, however, came when the two gradients were presented to the neurons in the same direction (that is, lots of BDNF where there was lots of laminin).

If the growth cones were more sensitive to BDNF than to laminin, the researchers hypothesized, they would turn away from the high end of the gradient. If they were more sensitive to laminin, they would do the exact opposite.

In the end, though, the results showed no preference to one cue over the other. Conflicting signals caused the growth cones to turn randomly.

This proves that a single guidance cue (among which BDNF and laminin are but two) can elicit different responses from growth cones depending on the biochemical context of the local environment.

By allowing scientists to control the complexity of that environment, the Hopkins team's device will likely help elucidate details about nervous system development.