Gremlin1 controls digit numbers and identities by spatial regulation of two early limb bud mesenchymal progenitor populations

The developing vertebrate limb bud is an excellent model to study the mechanisms that provide the vertebrate embryo with robustness during pattern formation and organogenesis. We have previously shown that the core of this signaling system consists of SHH/GREM1/AER-FGF feedback signaling. The BMP antagonist GREM1 is a key node for the control of patterning, proliferation, and survival of limb mesenchymal progenitors (LMPs). Inactivation of Grem1 in mice limb bud results in skeletal defects, including a severe reduction in digit numbers. Whole limb RNA-sequencing (RNA-seq) for several distinct mutations in Grem1-enhancers that reduce Grem1 expression revealed that genotypes with a visible digit phenotype, have many differentially expressed genes in key developmental pathways while mutants with no apparent phenotype are able to compensate the Grem1 deficiency. Amongst the mutants with a phenotype, specific deletion of multiple enhancers has shown to alter the spatial expression of Grem1 in such a way that it causes the loss of one digit, resulting in stable tetradactyly. With the use of specific recombinant Cre alleles and lineage tracing, we determined that the missing digit is d2.
In order to understand the digit loss on a molecular and cellular level, single cell RNA-seq analysis of wildtype and two distinct tetradactyl mutants (E1C5Δ/Δ, E1C8Δ/Δ) was performed at E10.75 (37 – 39 somites), which revealed molecular signatures of two LMP populations. In the mutants, both populations exhibit altered cell numbers and levels of expression, compared to the wildtype. RNA-Fluorescence in situ hybridization (RNA-FISH) analysis of key signature markers for these progenitor populations confirmed a reduction of the most distal population (dLMP) at the expense of an expansion of the peripheral one (pLMP). This result points to a problem in specification of the cells that will give rise to the handplate, including the digits. The reduction in dLMPs detected at E10.75 persists during progression of limb bud outgrowth and onset of handplate development (E10.75-E11.25, 37-43 somites) showing that mutants fail to catch up to wildtypes and eventually end up with one digit less. To gain insight into how these populations behave in the context of digit loss/gain, mutants that are models for human malformations were analyzed, namely Tbx3Δ/Δ, ShhΔ/Δ and Gli3Δ/Δ. Genetic inactivation of Shh disrupts Grem1 expression and results in digit agenesis such that only a rudimentary condensation forms. In contrast to wildtype forelimb buds, the pLMP signature genes are expressed uniformly in the peripheral mesenchyme of ShhΔ/Δ forelimb buds at E10.75, whereas expression of dLMP marker genes is not detected. By E10.75, Grem1 expression is expanded anteriorly in Gli3Δ/Δ forelimb buds. Concurrent with this anterior expansion, the anterior bias in pLMPs is lost, while markers of dLMPs are also anteriorly expanded. This analysis of Gli3Δ/Δ limb buds shows that this early peripheral distal-anterior expansion of dLMP signature foreshadows anterior digit duplications, while loss of the anterior bias in pLMP signatures correlates with the loss of anterior digit identities and asymmetry. Furthermore, single cell analysis and RNA-FISH were done on mutant forelimb buds without Grem1 expression (Grem NULL) in comparison to wildtype. dLMPs were severely reduced with some of their makers depleted while pLMPs were increased and had a symmetric expression, corroborating the conclusions of previous analysis.
Tetradactyl mice exhibit a paraxonic skeletal phenotype, resembling that of pigs at corresponding developmental stages. We further investigated gene expression patterns by analyzing mRNA levels of key progenitor population markers. This analysis revealed similarities in the expression of pLMPs but also identified an expanded dLMP domain. Taken together, our findings provide new insights into the molecular mechanisms of digit specification and show how the balance of two distinct progenitor populations can shift the limb axis asymmetry and influence digit numbers and identities.