Meis2 is expressed in subclasses of DRG cutaneous mechanoreceptive neurons in mouse embryos (A-B).

A) ISH for Meis2 mRNA showed expression in a subpopulation of DRG sensory neurons at embryonic stages E11.5, E14.5 and E18.5, at P0 and at adult stages. Dashed lines delineate the DRG. Scale bar=50µm. B) IF for Meis2 (red) and c-Maf, Ntrk2 or Ntrk3 (blue) at P7 following injection of cholera toxin B subunit (CTB in green) in the skin of new-born pups. Note that Meis2+/CTB+ retro-traced sensory neurons co-expressed c-Maf, Ntrk2 or Ntrk3 (arrows). Scale bar=50µm. We estimated that 30.5 ± 3.5% (mean ± SEM; n = 3) of Meis2-positive neurons co- expressed Ntrk2, and that 39.5 ± 5.4% co-expressed Ntrk3. Conversely, Meis2 was co-expressed in 53.6 ± 9.4% of Ntrk2-positive neurons, and in 78.5 ± 5.0% of Ntrk3-positive neurons. Meis2 expression depends on target-derived signals (C-D). C) Representative images of Meis2 mRNA expression (blue or pseudo-colored in red) and islet1 (green) in DRGs of Hamburger-Hamilton stage (HH) 36 chick embryos on the ablated and contralateral sides. Box plots showing the number of Islet1+/Meis2+ DRG neurons per section at stage HH36 following limb bud ablation. For Islet1-positive neurons, the contralateral side was considered as 100% of neuron per section. For Meis2-positive neurons, values represent the percentage of Meis2+ over Islet1+ neurons. D) Representative images of Meis2 mRNA expression (blue or pseudo-colored in red) and islet1 (green) in DRGs of HH29 chick embryos on the ablated and contralateral sides. Box plots showing the quantification of Islet1+/Meis2+ neurons number per section at stage HH29 on the contralateral and ablated sides. Arrow heads point at remaining Meis2-positive VL neurons. Dashed lines encircle the DRGs. **p≤0.005 *** p ≤ 0. 0 0; 0n5s= not significant following Student t-test. n=3 chick embryos. Scale bar=100µm. Altered touch perception in Meis2 m u t a n t m i c e (E)E-BoHx) p. lots showing the responses following application of Von Frey filaments of different forces. Isl1+/Cre::Meis2LoxP/LoxP mice exhibited a significantly reduced sensitivity to the 0.16, 0.4 and 0.6 g Von Frey filaments but not to higher forces filaments compared to WT and Isl1+/Cre littermates. * p≤0.05; ** p≤0.005; *** p≤0.001 following Kruskal-Wallis statistical analysis. F) Box plots showing the dynamic touch responses when the hind paw palms of individual mice were stroked with a tapered cotton-swab. Analysis showed that Isl1+/Cre::Meis2LoxP/LoxP mice were less responsive to the stimulus than WT and Isl1+/Cre littermates. *** p≤0.0001 following a one-way Anova statistical analysis. G) Box plots indicating that the latency to the first signs of aversive behavior in the hot plate test is similar in all groups of mice. WT, n=19; Isl1+/Cre, n=16; Isl1+/Cre::Meis2LoxP/LoxP, n=9. H) Box plots showing the number of bouts when a sticky paper tape was applied on the back skin of mice. Analysis indicated a significant decrease in the number of bouts in Isl1+/Cre::Meis2LoxP/LoxP mice compared to WT and Isl1+/Cre littermates. * p≤0.05 following a one-way Anova statistical analysis.

Meis2 is dispensable for LTMR neurons survival and specification.

A) Box plots showing that the DRGs volumes along the rostro-caudal axis are similar in E16.5 WT and Isl1Cre/+::Meis2LoxP/LoxP embryos. IF for Ntrk2 or Ntrk3 (red) and Islet1 (blue) and box plots analysis indicating that the percentage of Ntrk2+ and Ntrk3+ neurons are not affected in E16.5 Isl1Cre/+::Meis2LoxP/LoxP. Dashed lines encircle the DRGs. n=4; n.s. = not significant. Scale bar = 20µm. B) Representative images showing IF for Ntrk1, Ret, Ntrk2, Ntrk3 and Maf in E18.5 WT and Isl1Cre/+::Meis2LoxP/LoxP DRGs. Box plots showing that the number of Ret+, Ntrk2+, Ntrk3+ and Maf+ LTMR neurons and of Ntrk1+ nociceptive neurons are similar in E18.5 WT and Isl1Cre/+::Meis2LoxP/LoxP DRGs. n=3; n.s. = not significant. Scale bar = 100µm. C) Representatives images showing IF for Ntrk2 or Ntrk3 (green) with Pvalb or Maf (red) in P0 WT and Wnt1Cre::Meis2LoxP/LoxP DRGs. Box plots showing that the number of Ntrk2+ and Ntrk3+ neurons are unchanged in P0 WT and Wnt1Cre::Meis2LoxP/LoxP DRGs. n=3, n.s. = not significant. Scale bar = 20µm.

Meis2 inactivation dysregulates genes linked to neuronal projections and synaptogenesis.

A) Venn diagram comparing the number of DEGs between each genotype (n=3; p<0.05). This comparison identified 10 DEGs genes that were differentially expressed compared to both control genotypes (WT or Isl1+/Cre embryos), and a total of 140 genes that were differentially expressed in Meis2 mutant compared to either WT or Isl1+/Cre embryos. B) Gene ontology analysis for the 3 paired- analysis (WT vs Isl1+/Cre::Meis2LoxP/LoxP; Isl1+/Cre vs Isl1+/Cre::Meis2LoxP/LoxP and WT vs Isl1+/Cre) datasets using DAVID and the full RNAseq gene list as background. Graphs shows the comparison of the Fold Enrichment and the −log10(p value) of selected significant (p<0.05) GO or KEGG_PATHWAY terms associated to synapse and neurons projections whatever the number of genes. Blue dotted line indicate a p value of 0.05. Note that following David analysis gene ontology terms associated to synapse and neurons projections were overrepresented in the WT vs Isl1+/Cre::Meis2LoxP/LoxP and the Isl1+/Cre v s I s+l/C1re::Meis2LoxP/LoxP datasets compared to the WT vs Isl1+/Cre dataset. C) Heat maps showing the DEGs related to the GO terms synapse, neuron projection including dendrite, and protocadherin. D) Representative images showing a strong overall deficit of Nefh+ (red) sensory projections innervating the dermal papillae in the hairy and the foot pads in the glabrous skin of P0 Wnt1Cre::Meis2LoxP/LoxP neonates forepaw compared to WT littermates. Dashed lines delineate the hair follicle and the epidermis. Scale bar = 50µm.

Meis2 gene inactivation compromised Merkel cells innervation in the glabrous skin and increased Slowly-Adapting Mechanoreceptor (SAM) vibration threshold.

A) Graph showing the percentage of tap units among all recorded Aβ fibers in the nerve-skin preparation both in the hairy and glabrous skins. The number of tap units over the number of recorded fibers are indicated. Note that Tap-units are only present in both the hairy and glabrous skin of adult Isl1+/Cre::Meis2LoxP/LoxP mice but not in WT littermates. B) In the hairy and glabrous skins, SAMs in Isl1+/Cre::Meis2LoxP/LoxP mice (n = 22 from 6 mice) had significantly increased vibration threshold compared to WT mice (n = 29 from 6 mice), but normal firing activity to a 25-Hz vibration. Trace shows the stimulation applied to the skin and red squares indicate the time frame during when the number of spikes was calculated. C) SAM responses to a ramp of 50 Hz vibration with increasing amplitude are similar in WT, Isl1+/Cre and Isl1+/Cre::Meis2LoxP/LoxP mice. SAM responses to ramp stimuli and their static force responses were also identical in the different genotypes. Fibers from WT and Isl1+/Cre mice (n=5) displayed similar responses. * p≤0.05; ** p≤0.005. Traces show the applied stimulus and red squares the time frame during which the parameters below were measured. D) Confocal images of Nefh+ innervation (green) of CK8+ Merkel cells (red) in the forepaw glabrous skin of Isl1Cre/+and Isl1Cre/+::Meis2LoxP/LoxP adult mice. Dotted white squares indicate the close-up on CK8+ Merkel cells. Note the lack of Nefh+ fibers innervating Merkel cells in mutant mice. White arrows point at contact between NF200+ fibers and CK8+ Merkel cells. Scale bar = 10µm. The box plot indicates the percentage of Merkel cells in contact with Nefh+ fibers. n=4. * p≤0.05 in Mann-Whitney test. E) Confocal images of Nefh+ innervation (green) and CK8+ Merkel cells (red) of guard hairs in the hairy back skin of Isl1Cre/+and Isl1Cre/+::Meis2LoxP/LoxP adult mice. Dotted white squares indicate the close-up on CK8+ Merkel cells with apparently normal Nefh+ innervation. White arrows point at contacts between Nefh+ fibers and CK8+ Merkel cells. Scale bar = 10µm. The box plot indicates the percentage of Merkel cells contacted by Nefh+ fibers. n=4. F) Representative images of whole mount staining for CK8 in the hairy back skin of WT and Isl1Cre/+::Meis2LoxP/LoxP E18.5 embryos showing no difference in the number of touch dome between genotypes. Box plots show the number of touch domes per surface area and the number of Merkel cells per touch dome. No significant differences were found between both genotype in Mann-Whitney test. n=5 (WT) and 4 (Isl1+/Cre::Meis2LoxP/LoxP).

Meis2 gene inactivation affects Meissner corpuscles morphology.

A) Representative images showing S100β+ Meissner corpuscles (red) and their innervation by Nefh+ fibers (green) in the glabrous skin of WT and Isl1+/Cre::Meis2LoxP/LoxP adult mice. Scale bar=10µm. The box plot shows the average number of times Nefh+ fibers cross the midline of the Meissner corpuscles. Dashed blue lines indicate the Meissner corpuscle midline. Blue arrow heads indicate sites where Nefh+ fibers cross this midline. B) RAMs of the glabrous skin exhibited similar vibration threshold and firing activity to a 25-Hz vibration in WT (n = 16 from 4 mice) and Isl1+/Cre::Meis2LoxP/LoxP mice (n = 21 from 6 mice). Glabrous RAMs showed a non-significant decrease in firing activity to a ramp of 50 Hz vibration with increasing amplitude in Isl1+/Cre::Meis2LoxP/LoxP compared to WT littermates, but their response to ramp stimuli was similar in both genotypes. Traces indicate the type of stimulation and red squares the time frame during which the number of spikes was calculated. *** p≤0.001; Student’s t-test.

Meis2 gene inactivation affects hair follicle innervation and RAM fibers electrophysiological responses in the hairy skin.

A) Representative images of whole mount immunostaining for Nefh+ sensory projections (green) in the hairy skin of adult WT and Isl1+/Cre::Meis2LoxP/LoxP embryos counterstained with S100β (red) to highlight terminal Schwann cells decorating the periphery of hair follicles. Scale bar=100µm. B) Box plots showing the quantification for the number of branch points in the innervation network and the number of innervated hair follicles. n=3; * p≤0.05. C) RAMs in the hairy skin of Isl1Cre::Meis2LoxP/LoxP mice (n = 24 from 3 mice) exhibited significantly increased vibration threshold and reduced firing activity to a 25-Hz vibration compared to WT mice (n = 20 from 3 mice). RAMs in the hairy skin of Isl1+/Cre::Meis2LoxP/LoxP mice also showed a reduced firing activity in response to a ramp of 50 Hz vibration with increasing amplitude compared to WT and Isl1+/Cre animals. Fibers recorded from Isl1+/Cre mice (n=5) showed similar responses than those recorded from WT mice. * p≤0.05; ** p≤0.005.