ABSTRACT
In the present study we have investigated the molecular causes of the absence of digit 1 in the Hoxa13 mutant and why the absence of Hoxa13 protein, whose expression spans the entire autopod, specifically impacts the anterior-most digit. We show that in the absence of Hoxa13, the expression of Hoxd13 does not extend into the anterior mesoderm consequently leaving the presumptive territory of digit1 devoid of distal Hox expression and providing an explanation for the agenesis of digit 1. We provide compelling evidence that the lack of Hoxd13 transcription in the anterior mesoderm is due to increased Gli3R activity, in turn resulting from the loss of transcriptional repression exerted by Hoxa13 on Gli3. Our results are compatible with a mutual transcriptional repression between Gli3 and Hox13 genes that determines the anterior-posterior asymmetry of the autopod.
INTRODUCTION
The developing vertebrate limb has long proved as an excellent system for studying the mechanisms involved in pattern regulation and morphogenesis (Zeller et al., 2009) (Zuniga, 2015). Many of the genes important for limb patterning have been identified but little is known about the mechanistic implementation of gene expression patterns into specific morphological traits.
Among the genes essential for the outgrowth and patterning of the tetrapod limb are the Hox genes (Zakany and Duboule, 2007). Members of the HoxA and HoxD clusters display complex and dynamic patterns of expression during limb development that contribute to the organization of limb morphology (Spitz et al., 2001; Tarchini and Duboule, 2006; Zakany and Duboule, 2007).
A large body of work has established that the expression of Hoxd genes takes place in two successive phases. The first phase occurs in the emerging limb bud, principally involves expression of Hoxd8 to Hoxd11, and correlates with the specification of the upper-arm (stylopod) and forearm (zeugopod) morphology (Tarchini and Duboule, 2006; Woltering and Duboule, 2010). The second phase of Hoxd gene expression occurs in the hand plate, mainly involves Hoxd10 to Hoxd13, and is associated with the morphology of the hand (autopod) (Kmita et al., 2002; Spitz et al., 2001; Woltering and Duboule, 2010). The domains of expression corresponding to each of the two phases of expression are clearly separated by a transversal band of tissue, at the prospective wrist/ankle level, devoid of Hoxd transcripts (Woltering and Duboule, 2010).
Recent investigations have revealed that these precise patterns of expression rely on complex transcriptional regulation that involves multiple long-range enhancers located within the flanking topologically associating domain (TAD), regions of the chromatin with a discrete three-dimensional architecture in which internal interactions are favored (Andrey et al., 2013; Dixon et al., 2012; Lonfat et al., 2014). First phase transcription relies on enhancers located telomeric to the cluster, in the so-called T-DOM region, while the second phase transcription relies on enhancers located centromeric to the cluster, in the so-called C-DOM region (Andrey et al., 2013; Montavon and Duboule, 2013; Montavon et al., 2011). Hoxa13 has recently emerged as a major regulator of the switch between these two types of transcriptional regulation terminating the T-DOM dependent proximal regulation and re-enforcing the C-DOM distal regulation (Beccari et al., 2016; Sheth et al., 2016).
The HoxA cluster is the other Hox cluster involved in instructing limb morphology with 5’ Hoxa genes sequentially activated in the distal limb bud (Boulet and Capecchi, 2004; Davis et al., 1995; Fromental-Ramain et al., 1996; Kmita et al., 2005). Similarly to Hoxd genes, the transcription of Hoxa genes depends on remote enhancers scattered over the genomic landscape upstream of the cluster (Berlivet et al., 2013). Interestingly, the eventual activation of Hoxa13 in autopod progenitor cells abrogates Hoxa11 expression, generating mutually exclusive Hoxa11-Hoxa13 domains of expression that define the two distal segments of the limb, the zeugopod and the autopod respectively (Tabin and Wolpert, 2007). The activation of Hoxa13, together with Hoxd13, in the autopod progenitors controls expression of Hoxa11 sense and antisense transcription in a negative and positive way, respectively (Kherdjemil et al., 2016; Sheth et al., 2014; Woltering et al., 2014). Thus, Hoxa13 is a major regulator of both Hoxd and Hoxa gene expression.
Another major regulator of Hoxd gene transcription is Gli3, the principal transducer of Sonic hedgehog (Shh) signaling during limb development (Hui and Joyner, 1993; Litingtung et al., 2002; Lopez-Rios, 2016; Schimmang et al., 1992; te Welscher et al., 2002). In the absence of Shh, Gli3 is processed to a short form that acts as a strong transcriptional repressor (GLI3R) (Wang et al., 2000). Initially Hox proteins contribute to activate Shh transcription (Capellini et al., 2006; Kmita et al., 2005) but then Shh function is essential for the second phase of Hoxd expression by relieving the Gli3R repression (Lewandowski et al., 2015; Vokes et al., 2008). In the absence of Gli3, as in the extratoes (Xt) spontaneous mutation in mice, the autopod is characterized by a prominent uniform anterior-posterior (AP) expansion of Hoxd expression without a noticeable change in Shh expression (Zuniga and Zeller, 1999).
In addition to their temporospatial transcriptional control, the function of Hox products can also be modulated by interaction with co-factors. Some Hox products have been shown to interact with other DNA-binding co-factors including Smad5, Gli3 and Hand2 (Chen et al., 2004; Galli et al., 2010; Williams et al., 2005). For example, through physical binding, Hoxd12 can sequester Gli3 repressor (Gli3R) and even at high levels convert it from a transcriptional repressor to a transcriptional activator contributing in this way to the regulation of digit patterning (Chen et al., 2004). Likewise, proteins of the Hox13 paralogous group can modulate Bmp and TGFβ/Activin signaling activity through protein-protein interaction with Smads (Williams et al., 2005) and associate with Hand2 to activate Shh transcription (Galli et al., 2010).
Hoxa13 is the only member of the 39 mammalian Hox genes whose deletion is embryonic lethal as it is required for proper placental function (Scotti and Kmita, 2012; Shaut et al., 2008). Here we have focused on the study of the Hoxa13-/- null limb phenotype which is restricted to the autopod and characterized by the absence of the anterior-most digit 1, syndactyly and brachydactyly (Burke et al., 1995; Fromental-Ramain et al., 1996; Goodman et al., 2000; Innis et al., 2002; Mitsubuchi and Endo, 2006; Perez et al., 2010; Stadler et al., 2001). Because of the importance of digit 1, the thumb in the human hand, we asked why the absence of Hoxa13 protein, whose expression spans the entire autopod, specifically impacts the anterior-most digit. In contrast, although Hox paralogs often display considerable functional overlap, selective Hoxd13 loss does not impair digit 1 formation (Dolle et al., 1993). We report that in the absence of Hoxa13, the expression of Hoxd13 does not extend into the anterior mesoderm consequently leaving the presumptive territory of digit1 devoid of distal Hox expression, a circumstance that is considered sufficient to prevent digit condensation (Fromental-Ramain et al., 1996; Sheth et al., 2012). We also show that the lack of Hoxd13 transcription in the anterior mesoderm of Hoxa13 mutants correlates with increased Gli3R activity resulting from the loss of the transcriptional repression of Gli3 exerted by Hoxa13. Our results are compatible with a mutual antagonism between Gli3 and Hox13 genes that determines the anterior-posterior asymmetry of the autopod.
RESULTS
A gene dosage effect of Hoxa13 in digit 1 morphology
Hoxa13-/- mutants die at mid gestation, usually between E12.5 and E14.5, due to placental and vascular defects (Fromental-Ramain et al., 1996; Scotti and Kmita, 2012; Shaut et al., 2008; Stadler et al., 2001). Although it has been reported that a small percentage of Hoxa13 homozygous mutants survive to adulthood in the C57BL/6J genetic background (Perez et al., 2010), no homozygous pup was born in our colony despite being maintained in this genetic background. Indeed, the oldest homozygous embryos that we recovered were at embryonic day 16.5 (E16.5). At this stage, the typical Hoxa13 null limb phenotype, consisting predominantly of absence of digit 1 and syndactyly, was prominent (Fig. 1A; (Fromental-Ramain et al., 1996; Stadler et al., 2001)).
A careful inspection to determine the onset of the limb phenotype showed that the external aspect and shape of the mutant limb was indistinguishable from normal up to E11-E11.5. This was expected since activation of Hoxa13 normally occurs at E10.5. By E12.5, a flattening of the anterior border of the mutant autopod became progressively conspicuous reflecting the failure of digit 1 to form (Fig. 1B (Fromental-Ramain et al., 1996)). Accordingly, the expression of Sox9, the best marker of chondroprogenitors and differentiated chondrocytes, remained diffuse in the prospective digit 1 area without evidence of a digital condensation (arrowhead in Fig. 1B). In heterozygous embryos, the condensation corresponding to digit 1 was less well defined than in wild type littermates (arrow in Fig. 1B) supporting a gene dosage effect for Hoxa13. We also noticed that adult heterozygotes showed, in addition to the already reported partial syndactyly and claw alterations (Fromental-Ramain et al., 1996), a mild but consistent hypoplasia of the first digit, particularly conspicuous in the proximal phalanx of the hindlimb (33% reduction compared with wild type n=6; Fig. 1C). This trait was already observed in newborn heterozygotes (arrow in Fig. 1A).
Altered Hox code expression in the anterior limb bud mesoderm of Hoxa13-/- mutants
The hallmark of digit 1 is the expression of Hoxd13 but not the other 5’Hoxd genes Hoxd10, Hoxd11 and Hoxd12. This unique combination of Hoxd products is achieved during the second phase of Hoxd transcription when the expression of Hoxd13 spreads into the anterior mesoderm of the handplate surpassing the anterior limit of Hoxd11 and Hoxd12 domains (Kmita et al., 2002; Montavon et al., 2008). Interestingly, it has recently been shown that Hoxa13 plays a pivotal role in the transition from phase one to phase two of Hoxd gene expression (Beccari et al., 2016; Sheth et al., 2016; Sheth et al., 2014). In the absence of Hoxa13, the first phase of expression of Hoxd11 is abnormally prolonged and, consequently, the gap between the two phases becomes distally displaced (Fig. 2A). Since the prolongation of the first phase is more marked anteriorly, the prospective digit 1 cells now reside within the first phase domain of Hoxd11 expression (arrow in Fig. 2A and (Sheth et al., 2016; Sheth et al., 2014)). This also occurs with the first phase of the more 3’Hoxd genes Hoxd4, Hoxd9 and Hoxd10, that extends over the presumptive digit 1 cells and also with Hoxa11, a gene typical of the zeugopod (arrows in Supplementary Fig.1;(Sheth et al., 2014)). In contrast, no gross perturbation of the normal Hoxd12 expression, which extends to the anterior digit 2 border during the second phase of Hox gene expression, was detected in Hoxa13 null autopods (Fig. 2B). Interestingly, Hoxd13 transcription failed to extend into the most anterior mesoderm and remained in a domain similar to that of Hoxd12 (arrowhead in Fig. 2C). Thus, digit 1 progenitors in Hoxa13 mutants express an altered Hox code that corresponds to the zeugopod, rather than to the autopod, as it includes Hoxa11 and the first phase of Hoxd genes but lacks the characteristic expression of Hoxd13. The fact that the prospective digit 1 cells in Hoxa13 null embryos are devoid of both Hoxa13 and Hoxd13 may account for the loss of digit 1 in the Hoxa13 null limb, as mice lacking both Hoxa13 and Hoxd13 exhibit a profound loss of chondrogenic condensations throughout the autopod (Fromental-Ramain et al., 1996; Sheth et al., 2012).
To further explore the expression of genes characteristic of digit 1 in Hoxa13 mutants, we also analyzed the expression of Hand1 (Fig. 2D; (Fernandez-Teran et al., 2003)). In contrast to wild type littermates, Hand1 expression was not observed in digit 1 in Hoxa13 mutants confirming the absence of digit 1 specification (arrowhead in Fig. 2D). No differences in the pattern of expression of other markers of the anterior mesoderm, Tbx2, Tbx3, Alx4, the expression of which primarily occurs at proximal level, were observed in Hoxa13 null limb buds compared with wild type littermates (Supplementary Fig. 2). This indicates that the gene expression perturbations are specific of the anterior autopod mesoderm (Knosp et al., 2007).
No major disturbance of signaling centers in the absence of Hoxa13
Since the formation of the digits also depends on the activity of the apical ectodermal ridge (AER) and of the zone of polarizing activity (ZPA), and given the interaction of Hox genes with these major signaling centers of the limb bud (Galli et al., 2010; Kmita et al., 2005; Sheth et al., 2013), we decided to investigate the state of the AER and ZPA in Hoxa13 mutants. In situ hybridization at E11.5 showed a downregulation of Fgf8 in the most anterior AER of homozygous limb buds (arrow in Fig. 3). To assess a possible impact on FGF signal reception, we assayed for the expression of Dusp6 (formerly Mkp3) and Sprouty4 (Spry4), genes considered sensitive readouts of FGF signaling (Minowada et al., 1999; Smith et al., 2006). Reduced Spry4, but not Dusp6, indicated a minor downregulation of FGF signaling in the territory of digit 1 suggesting that Spry4 may be a more sensitive readout of FGF signaling than Dusp6. In situ hybridization for Shh and its major targets Ptc1, Gli1 (Ahn and Joyner, 2004; Harfe et al., 2004) and Hand2 (te Welscher et al., 2002a), failed to reveal any difference between Hoxa13 mutants and littermates control limb buds (Fig. 3).
Growth during early limb development is supported by a positive regulatory feedback loop established between Shh and AER-FGFs (Bastida et al., 2009; Benazet et al., 2009; Laufer et al., 1994; Niswander et al., 1994; Scherz et al., 2004; Verheyden and Sun, 2008). A crucial component of this regulatory feedback loop is Gremlin1 (Grem1), a Bmp antagonist and Shh target gene, responsible for AER maintenance (Khokha et al., 2003; Michos et al., 2004). Interestingly, the expression of Grem1 in the Hoxa13 mutant autopod did not propagate into the anterior mesoderm, as occurs in control limb buds, but remained more posteriorly restricted (Fig. 3). It is likely that the downregulation of Fgf8 in the anterior AER is secondary to the lack of Grem1 in the anterior mesoderm. Since Shh expression and signaling are normal, the anterior downregulation of Grem1 must depend on the alteration of other transcriptional regulators, a plausible candidate being Gli3 (Vokes et al., 2008). Since Grem1 is primarily regulated by release from repression by Gli3R function, we next wanted to examine the state of Gli3 in Hoxa13 mutants.
Increased Gli3R activity in Hoxa13-/- anterior mesoderm
Our results show that Hoxa13 is required, directly or indirectly, for the normal anterior spread of Hoxd13 in digit 1 territory. However, we have previously shown that in the double Hoxa13;Gli3 mutant Hoxd13 is uniformly expressed at high level all along the anterior-posterior extension of the handplate ((Sheth et al., 2012); Supplementary Fig. 3). Indeed, in the absence of Gli3 the second phase of expression of all 5’Hoxd genes occurs symmetrically all along the anterior-posterior axis of the handplate providing a similar Hox code and presumably a similar amount of Hox products to all digits (Supplementary Fig. 3; (Litingtung et al., 2002; Montavon et al., 2008; te Welscher et al., 2002b)). This suggests that in the absence of Gli3, Hoxa13 is no longer necessary for Hoxd13 transcription in the anterior mesoderm and raises the possibility that Hoxa13 function is required to counteract or modulate Gli3R activity. Therefore, we decided to investigate the expression and activity of Gli3R in Hoxa13 mutants.
Because Gli3 activity depends on posttranscriptional processing, as a first step to evaluate Gli3R activity we explored the expression of its main target genes. RNA in situ hybridization showed distal expansion of the domains of expression of Pax9 and EphA3 two target genes activated by GLI3R (McGlinn et al., 2005), at E11.5 and E12.5 (red arrowheads in Fig. 4A-B). Bmp4, a gene whose expression positively correlates with the level of Gli3R (Bastida et al., 2004), was more robustly detected and in an extended domain in the anterior mesoderm (red arrowhead in Fig. 4C). In addition, the GLI3R repressed target gene Jag1 (McGlinn et al., 2005) was absent from the anterior mesoderm of Hoxa13 mutants at E11.5 and E12.5 (green arrowheads in Fig. 4D). Overall, the modifications in these expression patterns are consistent with higher Gli3R activity than normal in the anterior mesoderm of the Hoxa13 mutant. We next wanted to investigate whether this excess occurs at transcriptional or posttranscriptional level.
To quantify the level of Gli3 mRNA in the anterior mesoderm of mutants versus wild type, we performed RT-PCR at E11.75. For this we dissected the anterior mesoderm corresponding to digit 1, as depicted in Fig. 4E, the region in which we observed altered gene expressions. Our results showed that the expression of Gli3 in the Hoxa13 mutant anterior mesoderm was 1.7 fold higher than in wild type. We also quantified the level of expression of Hoxd13 by RT-PCR that showed a decrease of 90% confirming our in situ hybridization results (Fig. 2C).
To explore Gli3 processing in the mutant, we also used dissected anterior mesoderm fragments of E11.5 forelimbs and analyzed the level of Gli3R by Western Blot (Fig. 4F). Our analysis confirmed a slightly higher level of Gli3R in the anterior limb mesoderm of Hoxa13 mutant limb buds.
Altered Gli3 pattern of expression in the absence of Hoxa13
Having shown that the mRNA and protein level as well as the activity of Gli3R, are all increased in the anterior mesoderm, we examined the overall pattern of Gli3 expression in the limb bud. Analysis of Gli3 expression by in situ hybridization showed a dynamic pattern that has not been previously appreciated. Because of the similarity with the pattern of expression of 5’Hoxd genes, we describe it as evolving in two phases (Fig. 5A). As previously described, at E10.5 Gli3 expression occurred in most of the limb mesoderm except for the most posterior part where Shh is expressed (Ahn and Joyner, 2004; Benazet et al., 2009) although the level of expression was less intense in the central mesoderm. By E11, Gli3 expression in the distal mesoderm became progressively confined to the anterior autopod. By E11.5 a second domain of expression started in the autopod roughly overlapping digit 4 primordium and progressively spanned all the digit primordia (Fig. 5A). Therefore, by E12.5, two domains of expression were clearly distinct and separated by a gap of tissue devoid of transcripts. The proximal domain, remnant of the first phase of expression, remained as a transverse band at the zeugopod-autopod boundary. The distal domain, the second phase of expression, overlapped the distal digit plate and became progressively confined to the interdigital tissues at E12.5 as the digit condensations differentiate (Fig. 5A). By E13.5 the majority of Gli3 distal expression concentrated in the joints (Fig. 5A).
In the absence of Hoxa13, the pattern of Gli3 expression was unaffected up to E10.5 (Fig. 5). However as early as E11, the dynamics of Gli3 expression were dramatically altered; the downregulation of the first phase of expression was delayed and incomplete, never reaching the digit 1 region (red arrowheads, Fig. 5A). As a consequence, digit 1 territory remained within the first phase of expression of Gli3 even at later stages, while the second distal domain was restricted to the posterior digits (digits 2-4). A schematic representation of the dynamic expression pattern of Gli3 and the changes observed in Hoxa13 mutants is shown in Fig. 5B.
This result uncovered a previously unidentified hierarchical effect of Hoxa13 in Gli3 regulation. Interestingly, in E12.5 Hoxa13 mutants the distribution of Gli3 transcripts was similar to that of Hoxd10-11 (compare to Fig. 2 and Supplementary Fig. 2) raising the intriguing possibility that Hoxa13 regulates Gli3 and 5’Hoxd genes in a complementary manner. As mentioned in the introduction, Hoxd13 cooperates with Hoxa13 in regulating the switch from the first to the second phase of Hoxd gene expression (Beccari et al., 2016; Sheth et al., 2016; Sheth et al., 2014), therefore, we wanted to examine whether 5’Hoxd genes also played a role in the regulation of Gli3 expression. In this regard we note that the downregulation of the first phase of Gli3 transcription correlated with the anterior spread of Hoxd13 domain. This can be clearly appreciated when the two limbs of the same embryo are hybridized one for Hoxd13 and the other for Gli3 and compared (Fig. 5C, and schemes within). The anterior boundary of Hoxd13 domain coincides with the posterior domain of first phase Gli3 expression. The comparison between Gli3 and Hoxa13 patterns of expression showed overlapping at anterior level that becomes attenuated with time (Fig. 5C).
To start exploring this possibility, we screened the Gli3 genomic landscape for Hoxa13 and Hoxd13 binding sites using the published data sets of Hoxa13 and Hoxd13 in limb buds, as well as changes in the chromatin state and transcriptome between wild type and Hoxa13;Hoxd13 double mutants (Sheth et al., 2016). The examination of the Gli3 locus identified several Hox13 binding regions upstream of the transcriptional start site. Some of these binding regions were also enriched in H3K27ac marks (highlighted in Fig. 6A). Curiously, two peaks (highlighted in yellow in Fig. 6A) overlapped with two previously reported VISTA enhancers (hs1586 and hs1179; (Osterwalder et al., 2018)) with activity in the limb. This result is compatible with a direct regulation of Gli3 transcription by Hoxa13 and Hoxd13. Our previous in situ and RT-PCR analyses indicated that this regulation was negative and this is confirmed by the increase in Gli3 transcription (1.7 fold, FDR<= 0.05) that occurs in Hoxa13;Hoxd13 double mutant autopods, as can be observed in the RNA-seq profiles shown in the bottom tracks in Fig. 6A (Sheth et al., 2016).
A negative regulation of Gli3 expression by distal Hox genes was confirmed by the changes in expression of Gli3 in Hoxd11-13 mutants and Hoxa13;Hoxd11-13 double mutants (Fig. 6B). In agreement with previous results showing that the absence of Hoxd11-13 had no major impact on Gli3 mRNA and protein expression (Huang et al., 2016) our analysis additionally confirmed that the pattern of Gli3 expression was normal in this mutant (Fig. 6B). However, the study of the allelic series showed that the removal of one functional allele of Hoxd11-13 from the Hoxa13 homozygous had a stronger impact on Gli3 expression than the removal of Hoxa13 alone (Fig. 6B). Finally, in the total absence of distal Hox products (Hoxa13;Hoxd11-13 double homozygous mutants), Gli3 expression spanned the whole autopod except the posterior border where it is never expressed, again strikingly reproducing the pattern previously reported for Hoxd10-11 genes in this double mutant (Fig. 6B; (Beccari et al., 2016; Sheth et al., 2014; Woltering et al., 2014)). Interestingly, we have also observed increased Hoxa13 mRNA expression (unpublished RNAseq, E12.5 autopod, 1.6-fold, FDR=0.01) and protein levels (about 7-fold increase) and protein levels when Hoxd11-13 gene function was removed (Suppl. Fig. 4). Given the proven redundant function between Hoxa13 and Hoxd13, this increase in Hoxa13 could compensate for the loss of Hoxd11-13 and explain why the removal of Hoxd11-13 has no effect on Gli3 transcription (Fig. 6B) or on digit 1 formation.
Together, these results support a dose dependent function of Hoxa13 and 5’Hoxd gene products in regulation of Gli3 expression mainly controlling the downregulation of its first phase of expression.
Gli3R repression of Hand2 operates at later stages
In the emerging limb bud, anterior-posterior patterning is established by the mutual antagonisms between Hand2 and Gli3 (te Welscher et al., 2002a). This antagonism is based on Gli3R directly repressing Hand2 in the anterior limb bud mesoderm (Vokes et al., 2008) and Hand2 repressing Gli3 transcription directly and through Tbx3 (Osterwalder et al., 2014). It is known that the requirement of Hand2 to repress Gli3 transcription is transient, as the removal of Hand2 after the onset of Shh expression, has no consequences (Galli et al., 2010). However, it remains unknown whether Gli3 repression of Hand2 is also operating at later stages. The analysis of Hand2 expression in Hoxa13;Hoxd11-13 compound mutants offers a good opportunity to investigate this question. We found that the Hand2 domain of expression strongly correlated inversely with that of Gli3 (Fig. 7). This observation is compatible with Gli3R continuously repressing Hand2 transcription, even at later stages while the early repression of Gli3 transcription by Hand2 seems to be taken over by Hox13 products at later stages.
DISCUSSION
In this study we have used a Hoxa13 null mutant allele to further analyze the contribution of Hox genes to distal limb bud development. In particular, we have investigated the mechanisms underlying the loss of digit 1 in Hoxa13 null mice. The thumb is the last digit to form (Frobisch et al., 2007) and the one with higher risk of developmental disruption. More than 1000 syndromes in the Online Mendelian Inheritance in Man (OMIM) database have hypoplastic thumbs (Oberg, 2014).
The absence of Hoxd13 expression in digit 1 progenitor cells explains the lack of digit 1 in Hoxa13 mutants
Our results show that in the absence of Hoxa13, the domain of expression of Hoxd13 is similar to that of Hoxd12, with an anterior limit coincident with the anterior border of digit 2 and therefore, with no evidence of the so-called reverse colinearity (Montavon et al., 2008; Nelson et al., 1996). This is in agreement with the crucial role of Hoxa13 in promoting the second phase of Hoxd gene expression and reveals that Hoxa13 is continuously required for this function (Beccari et al., 2016; Ros, 2016; Sheth et al., 2016; Sheth et al., 2014).
In the Hoxa13 mutant, the lack of anterior propagation of Hoxd13 expression leaves digit 1 territory devoid of Hox13 paralogues, a situation equivalent to Hoxa13;Hoxd11-13 or Hoxa13;Hoxd13 double mutants and that is considered sufficient to preclude the formation of the digit condensations (Fromental-Ramain et al., 1996; Sheth et al., 2012; Zakany et al., 1997). Digit patterning, the generation of a periodic digit-interdigit pattern, is under the control of a Turing-type or reaction-diffusion mechanism in which 5’Hox genes modulate, in a dose dependent manner, the digit spacing (Sheth et al., 2012). The current model predicts that in the total absence of distal Hox genes (Hoxa13, Hoxd11-13), the area of digit patterning is strongly reduced and no distinct digital condensations form, as occurs in the digit 1 territory in the Hoxa13 mutant.
Loss of Hoxa13 function is associated with a gain in Gli3R activity in the anterior mesoderm
Although Hoxa13 seems to play a role in the anterior spread of Hoxd13, this function is not required in the absence of Gli3. Actually, in the absence of Gli3, regardless of whether or not Hoxa13 is present, the second phase of expression of 5’Hoxd genes uniformly spans the AP axis of the autopod (Litingtung et al., 2002; te Welscher et al., 2002b). Consequently, the prospective digit 1 progenitors (those located at the anterior border) form a digit with posterior identity because they express a combination of Hox products that includes Hoxd12 and Hoxd11 in addition to Hoxd13, which can be interpreted as a “transformation” or “posteriorization” of digit 1 identity.
Therefore, we reasoned that during normal development, Hoxa13 could modulate the repressor function of Gli3R to permit a fully realized second phase of Hoxd13 expression. It has been suggested that because of its higher transcriptional efficiency, Hoxd13 seems to be less sensitive than Hoxd12 and Hoxd11 to repression by Gli3R, and therefore is the only 5’Hoxd member normally expressed in digit 1, the area of maximum GLI3R level (Montavon et al., 2008). Thus, Hoxa13 could potentially attenuate Gli3R levels sufficiently to permit the spread of Hoxd13 but not the other 5’Hoxd genes.
Supporting this hypothesis, we provide compelling evidence of increased Gli3R activity in the anterior mesoderm of the Hoxa13 mutants. First, the expression of bona fide Gli3R activated targets, such as Pax9 and EphA3 (McGlinn et al., 2005), are clearly upregulated while repressed targets, such as Jag1 (McGlinn et al., 2005) and Grem1 (Vokes et al., 2008) are downregulated in the anterior mutant mesoderm. The failure of Grem1 to propagate to the anterior mesoderm also explains the mild downregulation of Fgf8 at the anterior border, which could secondarily contribute to the Hoxa13 phenotype. Second, we also show that both Gli3 mRNA transcription and Gli3R protein levels are elevated in the anterior mutant mesoderm as determined by RT-PCR, RNA-seq and immunoblotting, respectively.
At first glimpse it may seem contradictory that digit 1 is lost because of excess of Gli3R as this is the only digit that reportedly forms in the hindlimb with high levels of Gli3R such as in the Shh and Ozd mutants (Chiang et al., 2001; Kraus et al., 2001; Ros et al., 2003). However, it should be noted that, as in the hindlimb, the small growth that occurs in the Shh null forelimb is also accompanied by expression of both Hoxa13 and Hoxd13. Due to the massive cell death that occurs in both Shh and Ozd mutant limb buds, a cell lineage study would be required to determine the origin of the progenitors of the single rudimentary digit that forms.
Hoxa13 dependent regulation of Gli3 transcription and mutual transcriptional repression between Gli3 and distal Hox genes
Interestingly, analysis of the Gli3 expression pattern by in situ hybridization unexpectedly showed that it was highly altered in the absence of Hoxa13 pointing to a role for Hox proteins in the transcriptional regulation of Gli3. Our analysis uncovered a previously unappreciated and highly dynamic pattern of Gli3 expression that, due to similarity with that of some 5’Hoxd genes, we describe as evolving in two phases. After an initial wave of expression in the early limb bud mesoderm except for the most posterior Shh-expressing area, Gli3 transcription becomes progressively downregulated from posterior to anterior until becoming confined to a proximal band at the zeugopod-autopod boundary. A second wave of expression is gradually established along the distal digital plate. As a consequence, two separate domains of Gli3 expression are clearly seen in the E12.5 autopod separated by a band of tissue devoid of transcripts that corresponds to the wrist.
The dynamic pattern of Gli3 transcription is highly altered in the absence of Hoxa13 as the downregulation of the first wave of expression is delayed and incomplete remaining over digit 1 territory so that when the second phase of expression is established, the gap between the two phases lies between prospective digit 1 and 2. This altered pattern of expression is identical to that of Hoxd10 and Hoxd11 in the absence of Hoxa13. Indeed, the dynamics of the downregulation of the Gli3 first phase is totally coincident with the progression of the second phase of Hoxd13 genes, both in the wild type bud and in Hoxa13;Hoxd11-13 allelic series, pointing to a mutual transcriptional repression between Gli3R and Hoxd genes and uncovering an additional level of interaction between Hox genes and the Shh/Gli3 pathway (Fig. 8). This is supported by the presence of several Hox13 binding sites in the Gli3 genomic landscape the function of which definitely deserves further investigation.
No evidence of Hoxa13-Gli3 physical interaction
We have also considered the possibility that Hoxa13 interacts with Gli3R at the protein level, as has been shown for Hoxd12 and Hoxd13 (Chen et al., 2004), and that this binding modifies GLI3R activity either by sequestering it or transforming its repressor activity into an activator. However, coimmunoprecipitation (CoIP) of E 11.5 limb bud lysates using antibodies specific to HOX13 and GLI3 (Chen et al., 2004; Knosp et al., 2004; Wen et al., 2010) didn’t detect such interaction (authors unpublished results). It is noteworthy that a recent screen based on affinity purification of endogenous protein and mass spectrometry (RIME, Mohammed et al., 2016) did not identify GLI3 as an interacting partner of HOXA13 (Marie Kmita, personal communication) supporting the conclusion that the gain in Gli3R activity observed in the anterior autopod of Hoxa13 mutants is primarily the result of the transcriptional control exerted by Hox13 proteins.
Summary
It is well known that Gli3R regulates the transcription of the 5’Hox genes in the anterior mesoderm in a dose-dependent manner (Litingtung et al., 2002; te Welscher et al., 2002b); Supplementary Fig. 2). Gli3R repressor activity may be mediated by direct binding of Gli3R to the enhancers in the regulatory Hoxd landscape (Vokes et al., 2008). Here we provide compelling evidence for the implication of Hoxa13 in the attenuation of Gli3R expression required for Hoxd13 transcriptional spread into the most anterior mesoderm (reverse collinearity). Hoxa13 together with Hoxd13 repress Gli3 transcription establishing a mutual antagonism necessary for the correct anterior-posterior asymmetry of the autopod. The acquisition of the repressive activity of distal Hox products on Gli3 transcription may have been determinant in the reduction of the anterior skeletal elements and digits that gradually occurred during the fin-to-limb transition (Sheth et al., 2012; Onimaru et al., 2015)
MATERIAL AND METHODS
Embryos
The Hoxa13 (Fromental-Ramain et al., 1996), HoxdDel(11–13) (Zakany and Duboule, 1996) and Gli3 (Gli3XtJ Jackson allele; (Hui and Joyner, 1993)) mutant lines were kindly provided by Pierre Chambon, Denis Duboule and Rolf Zeller, respectively. The Hoxa13 line was maintained on a C57BL/6 genetic background and the Gli3 line on a mixed CD1 and C57BL/6 genetic background. Genotyping was performed using tail biopsies or embryonic membranes according to previously published reports. Embryos of the desired embryonic embryonic day (E) were obtained by cesarean section. All animal procedures were conducted accordingly to the EU regulations and 3R principles and reviewed and approved by the Bioethics Committee of the University of Cantabria, and according to the ethical guidelines of the Institutional Animal Care and Use Committee (IACUC) at NCI-Frederick under protocol #ASP-12-405.
Skeletal Preparations and in situ hybridization
Whole-mount skeletal preparations were performed by staining with Alcian blue 8GX (Sigma Aldrich) and Alizarin red S (Sigma Aldrich) following standard protocols. Briefly, the specimens were fixed in 95% ethanol, skinned and eviscerated before staining, cleared in a series of KOH solutions and stored in glycerol. Whole mount in situ hybridization was performed according to standard procedures using digoxigenin labeled antisense riboprobes. At least 2 embryos per stage and genotype were analyzed. The probes used were Sox9, Hoxd13, Hoxd12, Hoxd11, Hoxd10, Hoxd9, Hoxd4, Hoxa11, Gli3, Pax9, Jagg1, EphA3, Hand1, Hand2, Bmp4, Shh, Gli1, Ptc1, Dusp6, Spry4, Grem1 and Fgf8.
Western Blot
For immunoblot (Western blot) analysis, dissected autopod regions of E11.5 embryos were lysed with ice-cold RIPA buffer. 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis was used to resolve Gli3-190 from Gli3-83 protein, and the mouse monoclonal Gli3 clone 6F5 anti-Gli3-N antibody was used (kindly provided by Dr. Scales at Genentech; Wen et al., 2010). βactin (mouse monoclonal C4SC-4778, Santa Cruz) was assessed as control for normalization. Three independent experiments were performed.
Quantitative real-time PCR (qRT-PCR)
The region of digit 1 (Fig. 4E) was dissected in cold RNAse-free PBS from E11.75 wild type and Hoxa13 null embryos. Total RNA was extracted with RNeasy® Plus Micro Kit (Qiagen) and 50 ng of total RNA was reverse transcribed to produce first-strand cDNA with iScript™ cDNA Synthesis kit (Bio-Rad) using standard conditions. qRT-PCR was carried out on an Applied Biosystems StepOnePlus™ using SYBR Green Supermix (Bio-Rad) and the data were analyzed using the StepOne(tm) software. The primers used to amplify Gli3 and Hoxd13 were previously described in (Huang et al., 2016). Relative transcript levels were normalized to Vimentin (Huang et al., 2016). Four biological replicates were analyzed for each genotype, with at least two technical replicates for each sample. The expression levels of mutant samples were calculated relative to wild-type controls (average set to 100%). The significance of all differences was assessed using Student-test, being statistically significant when p-0.05. GraphPad Prism5.0 (LaJolla, CA) was used for graph and statistics analysis. Histogram bars represent the average expression values after normalization to Vimentin (standard deviation shown as error bars).
Funding
This research was supported by the Spanish Ministry of Science, Innovation and Universities Grant (BFU2017-88265-P) to MAR, the Center for Cancer Research, National Cancer Institute, NIH (SM, intramural Research Program) to SM and Shriners Hospitals for Children Basic Research Grant (85140-POR) to HSS.
Author contributions
MF Bastida: Conceptualization; Investigation; Validation; Writing—review and editing R Perez-Gomez: Formal analysis; Investigation; Methodology
A Trofka: Formal analysis; Investigation; Methodology
R Sheth: Conceptualization; Formal analysis; Investigation; Writing—review and editing
HS Stadler: Formal analysis; Investigation; Funding acquisition; Writing—review and editing
S Mackem: Conceptualization; Formal analysis; Investigation; Funding acquisition; Writing—review and editing
MA Ros: Conceptualization; Formal analysis; Investigation; Funding acquisition; Supervision; Project administration; Writing—original draft
Supplementary Figure 1
Altered Hoxd gene expression in Hoxa13 mutant limb buds.
Forelimb autopods of wild type and Hoxa13 homozygous mutants hybridized with Hoxd4, Hoxd9, Hoxd10 and Hoxa11 at E 12.5. The arrows point to altered expression patterns in homozygous mutants.
Supplementary Figure 2
Expression of markers of the anterior mesoderm in Hoxa13 mutant limb buds.
Pattern of expression of Tbx2 (A), Tbx3 (B) and Alx4 (C) in E11.5 forelimb buds of wild type and Hoxa13 homozygous mutants. No obvious modification in the expression patterns is observed in the mutant.
Supplementary Figure 3
Hox gene expression in limb buds of the Hoxa13;Gli3 allelic series.
E12.5 forelimb buds are shown with the genotype indicated at the top and the hybridization probe on the left. Note that in the absence of Gli3, disregarding the presence or not of Hoxa13, Hoxd13 expression spans the entire autopod.
Supplementary Figure 4
Hoxa13 protein levels increase with reduction in Hoxd11-13 dosage.
1% SDS lysates of dissected distal limb bud (E12.5 autopod region) from wildtype and Hoxd11-13 mutants were electrophoresed, blotted, and probed with affinity-purified polyclonal anti-Hoxd13 antibody which recognizes both Hoxa13 and Hoxd13 (Chen et al., 2004), and with anti-Vinculin (1:1,000, Sigma # V9264). Bands were visualized with fluorescent secondary antibodies and quantified using the Odyssey Li-Cor system. Hox13 fluorescence signals were normalized to Vinculin, and three independent samples were analyzed for each genotype. Significance of differences were determined using the two-tailed, Student’s t-test. An approximately 7-fold increase in Hoxa13 expression in Hoxd11-13-/- limb buds compared with wildtype was significant at p=0.01 and a 2-fold increase in Hoxd11-13+/- was significant at p=0.05.
Acknowledgments
We thank Genentech for the 6F5GLI3 antibody and Laura Galán, Mar Rodriguez and Víctor Campa for excellent technical assistance. We are most grateful to Berta Casar, Lorena Agudo and Endika Haro for helpful discussions.