Excitable motility

December 21, 2010 § 2 Comments

We’ve talked before about the puzzle of how cells like neutrophils figure out how to follow a shallow gradient of attractive chemicals.  In a recent paper (Xiong et al, 2010.  Cells navigate with a local-excitation, global-inhibition excitable network.  PNAS, PMID 20864631) the Devreotes and Iglesias labs describe a new model of how chemotaxis might work.  It’s a very pretty model for a couple of reasons: it embodies a brand new shiny idea about how ameboid cells manage to be so efficient about moving in the direction of a gradient; it rationalizes a whole series of observations about what mutations in different pathways do to cellular motility and direction-finding; and it points the way to future experiments.

The problem of how ameboid cells (such as neutrophils, or Dictyostelium) do such a good job of identifying and responding to shallow concentration gradients of certain chemicals (chemoattractants) has puzzled biologists for decades.  Mathematicians and modelers have more recently joined the ranks of the puzzled.  A huge amount of information has been amassed about which proteins are involved in the response, and lots of movies have been taken of cells responding to gradients.  One of the observations that is key for this paper is that cells do move around even in the absence of a gradient.  A cell sitting in a uniform environment sends out exploratory projections (pseudopodia) in random directions, which cause the cell to move in a kind of random walk.  When a gradient is added, the orientation of the pseudopodia becomes biased, orienting the cell “up” the gradient.  The pseudopodia look similar in both cases, what changes is just the direction in which they’re pointing.  So the process that makes the cell move seems to be separate from, though influenced by, the process that senses the direction of the gradient.

There are many mysteries here.  For one, the process of directional sensing seems to be impossibly sensitive.  It appears that cells can detect a 1% difference in concentration of chemoattractant between the front and the back of the cell, or less.  Somehow, this directional sensing leads to spatial reorganization of the downstream signaling molecules so that they are unevenly distributed around the perimeter of the cell; for example, activation of PI3-kinase happens at the front of the cell, while the PI phosphatase, PTEN, gets moved to the back.  Cells are also highly sensitive to changes in the direction of a gradient, for example a neutrophil will change direction if the bacterium it’s chasing has moved away.  And if you expose a cell to a uniform increase in the concentration of a chemoattractant — no gradient this time, just an increase in stimulation from all directions — it will round up (“cringe”) and then gradually settle down again, and begin to send out pseudopodia in all directions as if nothing had happened.  On the molecular level you can see that about half of the signaling pathways involved in directional sensing and motility shoot up in response to the chemoattractant increase, while the other half go down, just as suddenly; as the cell recovers from its cringe response, they all go back to normal.  This adaptation response is essential to allow the cell to respond to a wide range of gradients, but is hard to understand.

There have been many efforts to model different aspects of these behaviors, none of which have been particularly satisfying.  One of the more influential ideas is that gradient sensing relies on a combination of local excitation with global inhibition, which amplifies the external signal and makes a shallow gradient easier to detect.  The basic idea here is that the chemoattractant stimulus activates two factors, an inhibitor, I, and an excitor, E.  The inhibitor is an integrator of information: it responds to the average number of chemoattractant receptors that are activated over the whole cell.  The excitor is a local response, and its level depends on the number of receptors that are active nearby.  This could be achieved if I is long-lived and E turns over quickly, for example.  Both I and E influence the activity of a third factor, the response regulator, RR: E increases its activity, and I inhibits it.

This little circuit neatly explains both directional sensing and adaptation.  If the signal is the same all over the cell, then E and I are activated to the same degree (though there may be some transient imbalance between them, leading to responses such as the “cringe”) and the steady state level of RR will be the same no matter what the concentration of chemoattractant.  But if the chemoattractant signal is coming only from one direction, then in the area of the cell facing the “up” direction E will be higher than I and the concentration of RR on that side of the cell will go up.  The problem with this model is that it doesn’t explain the extraordinary sensitivity of the directional sensing.  Other types of models do a better job of amplifying the sensitivity of the network, but don’t do so well on adaptation or the ability to follow rapid changes.

The new idea described here is that there is a second circuit linked to this local excitation—global inhibition module, which is “excitable”.  This hypothetical circuit (in its simplest form) involves two components, X and Y.  X activates itself, and also activates Y; Y shuts X down.  Let’s focus on this circuit on its own for a moment, and let’s assume that the level of X is somehow connected to the extension of pseudopodia.  The exact behavior of X will depend on many factors — how quickly X is made and degraded, how quickly Y levels respond to X levels, and so on — but often, the behavior will be some variation on the following: any time a little bit of X is made, a runaway reaction starts that makes a lot more.  After a while, Y catches up, and then X gets shut down to a basal level.  Then Y goes away too, and the whole cycle can start again.  Where the X came from in the first place doesn’t matter: it could be induced by the E/I/RR circuit, or it could be from spontaneous fluctuations.  In the absence of chemoattractant, fluctuations in the level of X (which would be amplified as described above) could lead to the generation of pseudopodia at random points around the periphery of the cell.  When Y kicks in, the pseudopod collapses.  This could explain the “resting” behavior of cells rather neatly.

When you connect the two circuits together, things get even more interesting.  Remember that in the E/I/RR network, E depends on local receptor occupancy while I depends on the average receptor occupancy across the whole cell.  If the chemoattractant is given all over the cell equally, then the increase in E and I is uniform too.  E and I don’t necessarily increase at the same rate — in this model, E goes up faster than I, leading to a brief excitation all over the cell — but in the end they go up by the same amount, so RR (controlled by the balance between E and I) comes back to the same level as it started from.  But when the chemoattractant signal is directional, the balance between E and I is broken — there is more E on the up side of the gradient, and less E on the down side, but the level of I is the same everywhere.  So, since the level of RR depends on both E and I, RR is higher than the original steady-state RR level on the up side (more E, and uniform I) and lower on the down side (less E, and still the same amount of I).  If RR drives X, the consequence is that X is strongly activated on the up side of the gradient, and inhibited on the down side.  This could explain how directional sensing is so heavily amplified, and how it leads to “upwards” movement.  Note also that as the cell crawls up the gradient, nothing much changes; as long as the level of receptor occupancy stays higher on the up side, the level of RR will stay high, and X will go on being stimulated.  But if the gradient changes direction, so will the stimulation.

This is getting to be a too-long post and I’ll have to stop soon.  I want to mention, though, that Xiong et al. did some careful simulations and also some experimental work to convince themselves (and you, dear reader) that there was no obvious contradiction between this model and what is already known.  There’s a good deal of discussion of the interpretation of this model in terms of the known biochemistry, which anyone seriously interested in this area should read.  In particular, there’s an interesting discussion of how the known classes of mutants in Dictyostelium and neutrophils can be interpreted in the framework of this model.  There are clues to which biochemical components may be part of the X and Y activities here, but Xiong et al. point out that the X/Y circuit is probably considerably more complex than a simple two-component loop.  If this idea turns out to be useful, which I think it will, there will be much work to do to identify and test candidate X and Y’s.  But  the conceptual framework offered here goes further than ever before to rationalize the main observations in studies of motility and directional sensing; it seems very much worthy of further study.

Xiong Y, Huang CH, Iglesias PA, & Devreotes PN (2010). Cells navigate with a local-excitation, global-inhibition-biased excitable network. Proceedings of the National Academy of Sciences of the United States of America, 107 (40), 17079-86 PMID: 20864631

Tagged: , , ,

§ 2 Responses to Excitable motility

  • Bobak Hadidi says:

    Very cool!

    Although, there is the model’s parametric equivalence to a third of an elephant (I suppose more accurately a fuzzy elephant); von Neumann and Occam aside, I am looking forward to their work on extensions to the model for inclusion of polarity phenomena.

    I also recently came across the following article, which may be of interest, and which takes another approach at this system, seeking a theoretical model generalizing two qualitative chemotactic behaviors: “stable single-axis” (straight) and “split and choice” (zig-zag).

    http://stke.sciencemag.org/cgi/content/abstract/3/152/ra89

    Perhaps some sort of integrated model is in order (LEGI-Biased Excitable Network driving the MS-MCRD?).

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

What’s this?

You are currently reading Excitable motility at It Takes 30.

meta

%d bloggers like this: