|Select image to enlarge|
Hypothetical mechanisms for ensuring a single bristle per proneural cluster.
In the Lateral Inhibition Model (above), the SOP (sensory organ precursor) prevents neighbors from becoming SOPs . Arcs denote inhibitory signals that could be transmitted by diffusion or contact. Contact between nonadjacent cells would require extensions like the filopodia seen on moth scale cells .)
In the Mutual Inhibition Model (below), every proneural cell inhibits adjacent cells, and the SOP can be chosen later [1563, 3022, 3270]. Uncommitted proneural cells are shaded, the SOP is black, and cells that succumb to inhibition (and become epidermal) are white. Degrees of shading reflect levels of AS-C proteins.
a, b. Supposed interactions at different stages. Cubes are individual cells, and inscribed circles are nuclei. a. Initially all cells are equivalent in the extent to which they emit (Delta ligand) and receive (Notch receptor) inhibitory signals, but this balance is unstable because each cell's AS-C (achaete-scute Complex) feeds back on itself in a loop that traverses the other cell . The feedback is positive because the loop has two negative steps. (Arrows denote activation; cross-barred (|) lines denote repression.) Any asymmetry will be amplified, so that if one cell is slightly more inhibited, its AS-C and Delta activities will wane, as will its ability to inhibit. Amplification is damped by a separate loop that uses the EGFR pathway (not shown) . Other factors may also prevent cells from stably expressing both Notch and Delta [990, 2839]. b. Eventually this cell shuts OFF its AS-C and becomes epidermal. Conversely, the cell that is less inhibited keeps increasing its AS-C activity, and at some point crosses a threshold that triggers an autocatalytic loop which keeps the AS-C switched ON. By this stage the 'winner' has purged its Notch receptors, and the 'loser' has purged its Delta ligands.
c, d. Two of the possible outcomes for a 5-by-5 array of cells undergoing mutual inhibition [1614, 1805, 4269]. (Another is parallel stripes.) The intended outcome -- a single SOP among 25 cells -- is actually impossible. One way out of this dilemma is to imagine that the inhibitory arc of the loop is functional, but the excitatory arc is disabled. In that case, cells would 'damp' one another but would not become SOPs based on how severely they themselves are damped. Another means of biasing the SOP decision would then be needed -- perhaps involving underlying prepattern factors or an overlay of Extramacrochaetae protein .
This montage is adapted from [1458, 3866, 4115] (a, b) and [1181, 1614, 1889, 4269] (c, d). N.B.: In the wing, Delta-Notch signaling manages to create 3-cell wide stripes that become veins  by a feedback mechanism that seems similar  but may not be (cf. Ch. 6). Thus, the same circuit may be used in different ways in different contexts [444, 627, 1457, 2018]. Odd results in the eye have raised questions, however, about the circuit's versatility . . . and validity . See also App. 7.