A limb bud morphogen bites the dust

Summary

Hedgehog signaling has traditionally been considered a morphogen for digits. In this issue of Developmental Cell, Zhu et al. (2022) show that a brief exposure to Sonic hedgehog is sufficient for digit specification, suggesting that it is not acting as a direct morphogen but rather as an initiator of this process.

In classic embryology studies, scientists found that a small group of cells in the posterior limb bud could induce mirror image digit duplications when transplanted to the opposite side of another limb bud. The factor from these cells was secreted and generated more posterior digit fates in a dose-dependent fashion, leading to the proposal that it was a digit morphogen (Tickle et al., 1975). This was subsequently identified as Sonic hedgehog (Shh), which is expressed in the same posterior group of cells and is required for digit development (Chiang et al., 1996; Riddle et al., 1993). However, the notion that SHH was acting as a direct morphogen for digit development was complicated by studies showing that digit formation still occurred in a Hedgehog-free environment (Litingtung et al., 2002; teWelscher et al., 2002). As these limb buds were polydactylous, with undefined digit identities, they did not rule out a role for SHH as a morphogen but they highlighted the additional roles for Shh in regulating cell survival and growth (Lopez-Rios et al., 2014). These dual roles have made it experimentally challenging to determine the specific properties of Shh in mediating digit identity. In a substantial advance, Zhu et al. report a strategy for separating digit specification from growth enabling them to discover important new properties in Hedgehog-mediated digit patterning (Zhu et al., 2022).

The loss of digits in Shh^−/− limb buds is primarily caused by high levels of apoptosis. This means that even if digits were initially specified, the cells would subsequently undergo cell death, masking any earlier role in specifying digits. To circumvent this, Dr. Susan Mackem’s group used a conditional knockout approach that enabled them to simultaneously delete Shh and the pro-apoptotic factors (Bax and Bak). Using a limb bud-specific CreER that enabled temporal deletions with the drug tamoxifen, they found that conditional knockout of Shh prior to when it was initially expressed resulted in a loss of all except for one digit, replicating the reported Shh^−/− phenotype (Chiang et al., 1996). However, when Shh expression was deleted approximately 3 hours later, after the onset of Shh expression, there was a substantial rescue of digit development, with most limb buds expressing 4 digits in the Bax;Bak mutant background. There was no further rescue when Shh expression was deleted at a later timepoint, indicating that in the absence of apoptosis, Shh activity is only required for digit formation during a startlingly narrow time window of 2–3 hours. The brevity of this time-window precludes a previously proposed model for the temporal integration of Hedgehog response in which Hedgehog dosage is integrated over a prolonged period of time to generate more posterior digit identities (Harfe et al., 2004). Most significantly, it defines a critical period for examining limb bud responses to Hedgehog signaling.

If Shh was acting as a direct morphogen during this critical period, then HH-responsive cells should be found throughout the domains generating HH-responsive digits, which prior to this study were thought to be digit numbers 2–5. Using the pathway target genes Gli1 and Ptch1 in combination with fate-mapping approaches, they found that Hedgehog responsiveness was restricted to the portion of the limb bud forming the most posterior digits (digits 4 and 5). Since Shh is also required for generating the other digits during this period, the authors interpret this as indicating that Shh is only indirectly specifying the remaining digits. However, there is a caveat to this interpretation in that Gli1 and Ptch1 are incomplete readouts of HH activity as they are activated by the higher level Hedgehog transcriptional responses that are mediated by GLI transcriptional activators. The majority of Hedgehog target genes in limb buds, exemplified by Hand2, are activated upon a loss of GLI transcriptional repression, a process that might be triggered by lower levels of Hedgehog signaling than GLI activation (Lewandowski et al., 2015; teWelscher et al., 2002). Unfortunately, the expression patterns and responses of this class of target genes are poorly characterized, especially at this early time-point, making it challenging to directly measure low-level Hedgehog pathway responses. Instead, the authors utilized several alternative approaches to show that overall GLI3 repressor activity is not altered in their mutant background. They conclude that Hedgehog signaling does not directly signal to digits 2 and 3 at this timepoint. Instead, they suggest that they are indirectly patterned by an unknown secondary ‘relay’ signal. Based on the transient 3 hour duration of this signal, this is a reasonable assumption, although it does not formally rule out the possibility that low levels of HH signaling regulate specification of digits 2 and 3 through GLI de-repression.

As a single digit resembling digit 1 remains in Shh^−/− limb buds, it has long been assumed that this is digit 1, the most anterior digit that presumably does not require Hedgehog signaling. However, using fate-mapping, the authors found that in mutants, this digit is actually composed of surviving posterior cells. They attempted to rescue the Shh^−/− phenotype using a cell-autonomous method of activating the Hedgehog pathway in the descendants of Shh-expressing cells. In doing so, they compellingly demonstrate that digit 1 is rescued in the absence of direct Hedgehog activation. Together, their results indicate that all 5 digits genetically require Hedgehog signaling. Yet only the most posterior digits 4 and 5 have evidence of direct Hedgehog pathway response in the critical time window for digit specification. Therefore, rather than acting as a morphogen, they propose that SHH acts as a trigger to activate an unknown secondary signal(s) that then specify digits 1–3 (Figure 1).

Figure 1.

A transient pulse of Hedgehog signaling specifies digits

A model to depict the findings from Zhu et al.: a 2–3 hour exposure to Sonic Hedgehog (SHH) is sufficient for nearly normal digit (‘d’) specification. Digit 4,5 progenitors are directly exposed to Hedgehog signaling. Digits 1–3 genetically require Hedgehog signaling for their specification but this likely (digits 2,3) and definitely (digit 1) occurs through a secondary signal initiated by Hedgehog signaling. After this time, sustained Hedgehog signaling remains critical for limb bud growth but no longer contributes to digit specification.

This study raises a number of key questions. It is unclear how long cells might remain responsive to digit induction by Hedgehog signaling at later time-points after the 2–3 hour interval that is sufficient for digit induction. The rescue experiment did not result in the formation of digits 2 and 3, as expected, perhaps because the later activation of Hedgehog signaling in this model bypassed a critical time window for digit specification.

Recent work has shown that a comparable early time-point coincides with the activation of the GLI transcriptional repressors that serve to regulate transcriptional responses in the absence of Hedgehog signaling (Lex et al., 2022). Perhaps limb bud cells at this stage are primed to have heightened responses to Hedgehog signaling, enabling digit specification during what is likely a highly transient period of competence. The biggest unanswered question of all is the composition of the secondary signal that generates digits 1–3. The identity of the precise timepoint for its activation makes this a realistic goal for future studies.