In preprints: oxygen and NFκB signals shift the timing of hindlimb formation

Broadly conserved developmental programmes have been challenged with modifications throughout evolution, generating phenotypic diversification. A big part of this diversification arises from changes in the timing, order and speed in the formation of tissues, which together are included under the umbrella term heterochrony (Gould et al., 1977; Smith, 2002). Despite their relevance from an evolutionary developmental biology (evo-devo) perspective, the molecular mechanisms behind heterochrony remain unsolved, partly due to the practical challenges of cross-species comparisons.

Limb development in tetrapods has often been used as a system to examine the mechanisms underlying similarities and differences in morphogenetic events, because a vast diversity can emerge from the limb's highly conserved and well-characterised developmental programmes. This system allows comparisons between species but also within individuals, by contrasting forelimb and hindlimb development. The differences between forelimb and hindlimb are not only characterised by their position along the anterior-posterior axis but also by their timing of formation and growth. The relative time sequence of forelimb and hindlimb formation varies by class. Aves develop both limbs synchronously, Amphibia develop hindlimbs first, and Mammalia develop forelimbs first, with marsupials showing the most extreme forelimb prioritisation (Bininda-Emonds et al., 2007; Chew et al., 2014).

In a recent preprint, Zhu and colleagues compare chicken and mouse limb development to investigate the mechanisms behind limb heterochrony (Zhu et al., 2024 preprint). Two T-box transcription factors, Tbx5 and Tbx4, are known key players in the initiation of forelimb- and hindlimb-forming fields, respectively (Agarwal et al., 2003; Logan et al., 1998). Using RNA fluorescent in situ hybridization chain reaction (HCR RNA-FISH), the authors quantified Tbx5 and Tbx4 transcripts to determine their expression dynamics across a time course in chicken and mouse embryos. While the expression dynamics of both factors were almost synchronous in chicken embryos, Tbx4 expression in the mouse hindlimb was delayed by around 18 h relative to Tbx5 in the mouse forelimb. However, upstream regulators of Tbx4, including Pitx1 and Isl1, were expressed synchronously with Tbx5, suggesting that a delay in the expression of known upstream factors did not cause the delay in Tbx4 expression.

Species-specific cis-regulatory elements can control the timing of expression of a particular gene. For example, a marsupial-specific sequence promotes the early activation of Sox9 in the opossum neural crest (Wakamatsu and Suzuki, 2019). Zhu and colleagues investigated the influence of cis-regulatory elements by identifying chicken Tbx4 enhancers and comparing their activity to known mouse Tbx4 enhancers. Using elegant enhancer-swap experiments to test the activity of mouse enhancers in the chicken embryo, and vice versa, the authors confirmed that these elements were functionally equivalent, being activated at the expected time and place in both species. These experiments ruled out the possibility that sequence variation between enhancers would be responsible for the temporal shift in Tbx4, but different factors binding to those elements could still drive it. To identify differentially expressed factors that could act as Tbx4 regulators, the authors performed forelimb and hindlimb bulk RNA-seq analyses in mouse and chicken embryos at the time of forelimb and hindlimb initiation. These analyses identified cRel (Rel), an effector of the NFκB signalling pathway, as a potential Tbx4 repressor. Overexpression of cRel in chicken embryos resulted in a reduction in hindlimb bud size and lower Tbx4 expression. Conversely, cRel knockout mouse embryos showed scarce but earlier expression of Tbx4 in the hindlimb field. The authors reasoned that the low penetrance could be due to the compensatory role of other NFκB effectors such as RelA (Rela).

Zhu et al. then analysed oxygen levels as a potential upstream regulator of NFκB signalling based on a link in other cellular contexts (Jiang et al., 2015; Yin and Juurlink, 2000). Furthermore, they argued that oxygen conditions are different during chicken (normoxia) and early mouse (hypoxia) development. Once the mouse placenta is functional, the embryo has access to higher oxygen levels, and this correlates with the timing of hindlimb formation. These observations prompted the authors to test whether oxygen levels could influence the timing of hindlimb initiation. They collected mouse embryos at embryonic day (E) 6.75 and cultured them ex utero under normoxic (20% oxygen) and hypoxic (5% oxygen) conditions for 1.5 days (ending the culture at the equivalent of ∼E8.25; i.e. before normal Tbx4 expression takes place in vivo). Remarkably, Tbx4 expression was initiated in mouse embryos cultured in normoxic conditions, but not in those cultured in hypoxic conditions. Concomitantly, cRel expression levels were higher in embryos cultured in hypoxic conditions, sustaining its repressor role.

These interesting observations support the idea that hypoxic conditions could delay the expression of Tbx4 in the mouse hindlimb and raise more questions about the direct or indirect link between hypoxia, NFκB signalling and the Tbx4 enhancers. Moreover, this work opens up an exciting evolutionary perspective where one could investigate the timing of hindlimb formation in other mammals in which the maturation of the placenta occurs at different time points or in those where the placenta capacity, and thus the oxygen supply, is limited. Similarly, comparisons between oviparous and viviparous lizards could reveal whether the presence of a placenta delays hindlimb induction in closely related species. Interestingly, a small delay in hindlimb outgrowth can be inferred from the description of developmental hallmarks in Eremias multiocellata, a viviparous lizard, although some oviparous lizards also develop forelimbs slightly earlier (Lin et al., 2021).

An interesting observation of Zhu and colleagues is that, when cRel was knocked out in mouse, the low-penetrant premature expression of Tbx4 always happened on one side of the embryo. This result is reminiscent of the left-specific defects that a moderate loss of function of Tbx5 – or of its upstream regulator retinoic acid – has been shown to generate in mouse embryos (Niederreither et al., 2002; Sulaiman et al., 2016). It remains to be determined whether this indicates that retinoic acid had an ancestral role in controlling limb induction, which was superseded in the hindlimb when the hypoxia/NFκB axis took over.

HIF1α, a hypoxia response transcription factor, is needed later for early chondrogenesis in the limb bud (Provot et al., 2007). Therefore, oxygen levels could regulate different steps in limb morphogenesis; however, it is unclear if and how oxygen, HIF and NFκB factors coordinate the progression of these programmes.