ABSTRACT
The correct positioning of joints in the vertebrate skeleton is not well understood. Mutations in connexin43 (cx43) cause the short segment phenotype of the zebrafish short fin (sofb123) mutant. We have shown that Cx43 suppresses evx1 expression, a transcription factor required for joint formation. Here, we provide novel insights into how Cx43 influences evx1 transcription. First, we find that Simplet (Smp) knockdown recapitulates the sofb123 phenotypes of reduced regenerate length and reduced segment length, and we find evidence for synergy between cx43 and smp. Moreover, knockdown of Smp increases the evx1 expression, similar to cx43 knockdown. Previous studies have shown that Smp is required for the nuclear localization of β-catenin. Indeed, β-catenin activity is required for segment length, and is reduced in both sofb123 mutants and following Smp knockdown in regenerating fins. We further show that blocking canonical Wnt signaling results in a synergistic reduction in segment length in sofb123/+ heterozygotes. Together, our findings suggest that both Smp and β-catenin function in a common molecular pathway with cx43 to influence both evx1 expression and joint location.
INTRODUCTION
The correct placement of joints is required for skeletal flexibility and functionality. However, the mechanisms by which joints are correctly positioned in the skeleton are poorly understood. We use the zebrafish regenerating caudal fin as the model system in which to address this fundamental issue. The caudal fin is composed of 16-18 fin rays, and each fin ray is composed of bony segments flanked by joints. The main advantages of this system include that the fin is a rich source of joints, that the amputated fin regenerates rapidly and that the early stages of joint morphogenesis in the fin ray resemble the early stages of mammalian joint formation (Sims et al., 2009).
The fin rays are made of two hemirays of bone matrix surrounding a central mesenchyme of undifferentiated fibroblasts. The dividing cells are located in the distal mesenchyme, whereas the differentiating osteoblasts are located in the lateral mesenchyme. Actinotrichia are produced at the distal ends of each fin ray. Actinotrichia are collagen-like fibrils where the osteoblasts align and secrete the bone matrix, or lepidotrichia (Becerra et al., 1983). Osteoblasts and joint-forming cells are derived from common precursor cells (Tu and Johnson, 2011), located distally and laterally. We refer to this population of cells collectively as the ‘skeletal precursor cells’. The fin regenerates in the proximal-to-distal direction, with new segments continually added to the distal end of the fin ray. Therefore, the newer joints are always located distal to the older joints. This system permits the study of joint morphogenesis over time (Sims et al., 2009; Dardis et al., 2017). Joint initiation occurs when a single row of joint-forming cells aligns at the site of the presumptive joint. As the joint matures, joint-forming cells appear to move apart, forming two rows of cuboidal cells. These cells resemble the interzone cells of mammalian synovial joints because of their appearance at the site of the future joint and by their apparent ability to contribute to joint morphogenesis (Pacifici et al., 2006). Thus, the appearance of the interzone indicates that a joint will be produced at a specified location, and the cells of the interzone help to build the physical joint. It is unclear when the molecular pathways for determining joint location are mechanistically separable from the molecular pathways for building the joint. We refer to pathways that influence joint initiation as regulating joint location. Alternatively, if the affected pathway is unclear we instead use the more general term ‘joint formation’.
Previous studies in the zebrafish fin length mutants short fin and another long fin (sofb123 and alfdty86) revealed that the gap junction protein connexin 43 (Cx43) plays important roles in determining the location of joints. Hypomorphic mutations in cx43 cause the sofb123 phenotypes of shorter fin length, reduced cell proliferation and shorter segment length (Iovine et al., 2005). Cx43-knockdown (KD) also recapitulates the sofb123 phenotype (Hoptak-Solga et al., 2008). The alfdty86mutant, in contrast, exhibits stochastic joint failure and longer segments, on average (Sims et al., 2009). Interestingly, although the mutation causing the alfdty86 phenotypes is not in the cx43 gene (Perathoner et al., 2014), alfdty86 mutants exhibit increased cx43 mRNA expression. Moreover, Cx43 KD in alf rescues the irregular segment length phenotype (Sims et al., 2009). Together, these studies provide evidence that Cx43 inhibits joint formation in zebrafish fin. Indeed, more recent studies have identified a transient reduction in cx43 expression coincident with joint initiation (Dardis et al., 2017). From these and other studies, we suggest that the short segment phenotype of sofb123 is due to premature joint initiation, whereas the long segment phenotype of alfdty86 is due to a failure of joint initiation. Thus, for Cx43-dependent phenotypes, segment length is a proxy for joint location.
The transcription factor evx1 is required for joint formation (Schulte et al., 2011) and low levels of evx1 expression precede joint initiation (Dardis et al., 2017). Interestingly, expression of evx1 is premature in sofb123 and sporadic in alfdty86 (Ton and Iovine, 2013b), consistent with our model for segment length phenotypes in these mutants. Moreover, Cx43 KD rescues evx1 expression in alfdty86 mutants (Ton and Iovine, 2013b). Therefore, we suggest that Cx43 influences segment length by influencing the timing of evx1 expression, which in turn influences the location of the next joint. Cx43 is expressed both in the medial mesenchyme and in the distal-most joint-forming cells (Sims et al., 2009). However, we have found that Cx43 function in the medial mesenchyme, and not in the lateral cells, is capable of rescuing the short segment phenotype of sofb123 mutants (Dardis et al., 2017). Therefore, our current model is that Cx43 function in the medial mesenchyme influences cell fate decisions in the lateral skeletal precursor cells. High cx43 expression favors differentiation of bone-forming cells (i.e. low evx1 and joint pathway off), whereas transient reductions in cx43 expression favors differentiation of joint-forming cells (i.e. increased evx1 and joint pathway on) (Dardis et al., 2017).
The findings in this article contribute to our understanding of how Cx43 influences evx1 expression and cell fate decisions in the lateral skeletal precursor cells. We find that Simplet (Smp) and β-catenin signaling mediates Cx43-dependent expression of evx1. Smp, or Fam53b, is a protein of 407 amino acids containing a 14-3-3 binding motif and a nuclear localization signal (NLS). Prior studies suggest that Smp regulates cell proliferation during early developmental stages of Medaka development (Thermes et al., 2006) and in regenerating fins (Kizil et al., 2009). More recently, Smp has been shown to influence the nuclear localization of β-catenin in zebrafish embryos (Kizil et al., 2014). Here, we show that smp is expressed downstream of Cx43, and that Smp is required for segment length and evx1 expression. We similarly find that β-catenin is required for segment length and evx1 expression, and that β-catenin signaling is reduced in both sofb123 and in Smp KD regenerating fins. These findings provide compelling evidence that Smp and β-catenin function downstream of Cx43 to influence joint location through the regulation of evx1 expression.
RESULTS
smp and cx43 function in a common pathway to regulate joint location
Smp has been shown to regulate cell proliferation during zebrafish caudal fin regeneration (Kizil et al., 2009). Expression of smp is observed in the distal epidermis, and also in both the distal and lateral mesenchyme (Kizil et al., 2009), and could contribute to joint formation. To test this possibility, we completed morpholino (MO)-mediated gene knockdown (KD) of Smp using one of the two morpholinos previously validated for Smp KD (i.e. Smp-1, Exon3–Intron3 splice antisense MO; Kizil et al., 2009). We first demonstrated that this MO appropriately targets smp mRNA by interfering with correct splicing (Fig. 1). This MO is predicted to cause skipping of exon 3, which we observed in Smp-MO-treated fins but not in standard control MO (STD-MO)-treated fins.
The Smp-KD phenotypes include significantly reduced segment length and regenerate length. (A) Regenerate length is reduced in Smp-MO-treated fins compared with STD-MO-treated fins. All fins were amputated at the 50% level. The amputation plane is indicated (black line). The black arrow indicates the distance from the amputation plane to the distal tip of the 3rd fin ray. Graph shows significant reduction in percentage similarity in Smp-MO- and STD-MO-injected fins (*P<0.05, Student's t-test, two-tailed and unpaired). Three independent trials were performed (n=25 total per treatment). The data did not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent s.e.m. (B) Calcein-stained fin rays show reduced segment length in Smp-MO-treated fins compared with STD-MO-treated fins. Double-headed arrows indicate the first completed segment following the amputation plane (white line). The inset shows a higher magnification of the segments (arrows indicate joints). Graph reveals significant reduction in percentage similarity in Smp-MO- and STD-MO-injected fins (*P<0.05, Student's t-test, two-tailed and unpaired). Three independent trials were performed (n=25 total per treatment). The data did not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent s.e.m. (C) Schematic representation of zebrafish smp pre-mRNA. The exons are shown in gray boxes and the introns are drawn as black lines. Positions of the MO and of the primers are indicated. The MO is predicted to cause skipping of exon 3 (55 bp). (D) Results of the RT-PCR analysis using C1-C2 primer set (396 bp product is predicted). (E) Results of the RT-PCR analysis using P1-P2 primer set reveals reduced full-length product and Smp-MO-dependent presence of the shorter product. The P1-P2 primer pair amplified product (380 bp) in lane 3 marked with an asterisk is caused by the skipping of exon 3 compared with lane 2, where exon 3 was included (435 bp). Scale bar: 50 μm.
The Smp-KD phenotypes include significantly reduced segment length and regenerate length. (A) Regenerate length is reduced in Smp-MO-treated fins compared with STD-MO-treated fins. All fins were amputated at the 50% level. The amputation plane is indicated (black line). The black arrow indicates the distance from the amputation plane to the distal tip of the 3rd fin ray. Graph shows significant reduction in percentage similarity in Smp-MO- and STD-MO-injected fins (*P<0.05, Student's t-test, two-tailed and unpaired). Three independent trials were performed (n=25 total per treatment). The data did not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent s.e.m. (B) Calcein-stained fin rays show reduced segment length in Smp-MO-treated fins compared with STD-MO-treated fins. Double-headed arrows indicate the first completed segment following the amputation plane (white line). The inset shows a higher magnification of the segments (arrows indicate joints). Graph reveals significant reduction in percentage similarity in Smp-MO- and STD-MO-injected fins (*P<0.05, Student's t-test, two-tailed and unpaired). Three independent trials were performed (n=25 total per treatment). The data did not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent s.e.m. (C) Schematic representation of zebrafish smp pre-mRNA. The exons are shown in gray boxes and the introns are drawn as black lines. Positions of the MO and of the primers are indicated. The MO is predicted to cause skipping of exon 3 (55 bp). (D) Results of the RT-PCR analysis using C1-C2 primer set (396 bp product is predicted). (E) Results of the RT-PCR analysis using P1-P2 primer set reveals reduced full-length product and Smp-MO-dependent presence of the shorter product. The P1-P2 primer pair amplified product (380 bp) in lane 3 marked with an asterisk is caused by the skipping of exon 3 compared with lane 2, where exon 3 was included (435 bp). Scale bar: 50 μm.
We next tested for skeletal phenotypes during regeneration. We injected the Smp-MO into half of the fin rays of wild-type regenerating fin at 72 h post-amputation (hpa), leaving the other half as an internal control. Alternatively, STD-MO was similarly injected. After injection, the whole fin was electroporated to induce cellular uptake of the fluorescein-labeled MO. Successful injection/electroporation was determined by observing fluorescein signal at 24 h post-electroporation (hpe) (Thummel and Iovine, 2017). Regenerate length and segment length were evaluated at 4 dpe (i.e. 7 dpa), and measurements were taken from the 3rd fin ray, which we use as a standard (Iovine and Johnson, 2000). We use the percentage similarity method to determine whether the gene-targeting MO has an effect (i.e. Banerji et al., 2017; Govindan et al., 2016). With this method, we take the ratio of the injected side over the uninjected side and multiply by 100, and we compare these values for the Smp-MO- and for the STD-MO-treated fins (values close to 100% indicate little effect of the MO; values with low similarity indicate that the MO had an effect). This method reduces the effect of fin to fin variation. Importantly, we observed significant reduction in regenerate length in Smp-KD compared with the STD-MO (Fig. 1), recapitulating the published phenotype for Smp-KD (Kizil et al., 2009). We further showed that Smp-KD leads to reduced segment length (Fig. 1), similar to sofb123 mutants and to Cx43-KD.
To test whether smp is expressed downstream of cx43, we performed whole-mount in situ hybridization in 5 dpa wild-type and sofb123 regenerating fins. We found that smp expression is reduced in sofb123 compared with wild type (Fig. 2). This finding was confirmed through qRT-PCR (Table 1 and Fig. S1). Together, these findings suggest that smp acts downstream of cx43 to influence segment length.
smp and cx43 function in a common pathway to influence evx1 expression. (A) Wild-type and (B) sofb123 fins were amputated at 50% and permitted to regenerate for 87 h. At 87 hpa, the fins were harvested and processed for smp whole-mount in situ hybridization. Asterisks represent the expression domains of smp in each of the three fin rays. Three independent trials were performed (n=18 per genotype). (C) Smp-MO concentrations of 1 mM (n=6), 0.75 mM (n=6) and 0.5 mM (n=11) were used to identify the subthreshold concentration. The 0.5 mM concentration of Smp-MO was selected as the subthreshold dose, as there is no significant decrease in the segment length. The data did not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent s.e.m. Student's t-test was performed to test for significance (two-tailed and unpaired, *P<0.05). (D) Synergistic effects of the 0.5 mM dose of Smp-MO with sofb123 heterozygotes (n=10) are revealed compared with either wild-type fins (n=11) injected with 0.5 mM Smp-MO or with sofb123/+ heterozygotes injected with STD-MO (n=7). The data do not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent s.e.m. Student's t-test was performed to test for significance (two-tailed and unpaired, *P<0.05). (E,F) Whole-mount in situ hybridization shows that the frequency of evx1 expression is increased in Smp-MO fins (n=21 fins) compared with the STD-MO fins (n=22 fins) (three biological replicates). All fin rays across the fins were injected with Smp-MO or STD-MO at 72 hpa and harvested at 87 hpa. Plus indicates fin rays positive for evx1; minus indicates fin rays negative for evx1. (G) The Smp-MO fins show an increased frequency of evx1-positive fin rays compared with STD-MO injected fins. The data did not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent s.e.m. Student's t-test was performed to test for significance (two-tailed and unpaired, *P<0.05). Scale bar: 50 μm.
smp and cx43 function in a common pathway to influence evx1 expression. (A) Wild-type and (B) sofb123 fins were amputated at 50% and permitted to regenerate for 87 h. At 87 hpa, the fins were harvested and processed for smp whole-mount in situ hybridization. Asterisks represent the expression domains of smp in each of the three fin rays. Three independent trials were performed (n=18 per genotype). (C) Smp-MO concentrations of 1 mM (n=6), 0.75 mM (n=6) and 0.5 mM (n=11) were used to identify the subthreshold concentration. The 0.5 mM concentration of Smp-MO was selected as the subthreshold dose, as there is no significant decrease in the segment length. The data did not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent s.e.m. Student's t-test was performed to test for significance (two-tailed and unpaired, *P<0.05). (D) Synergistic effects of the 0.5 mM dose of Smp-MO with sofb123 heterozygotes (n=10) are revealed compared with either wild-type fins (n=11) injected with 0.5 mM Smp-MO or with sofb123/+ heterozygotes injected with STD-MO (n=7). The data do not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent s.e.m. Student's t-test was performed to test for significance (two-tailed and unpaired, *P<0.05). (E,F) Whole-mount in situ hybridization shows that the frequency of evx1 expression is increased in Smp-MO fins (n=21 fins) compared with the STD-MO fins (n=22 fins) (three biological replicates). All fin rays across the fins were injected with Smp-MO or STD-MO at 72 hpa and harvested at 87 hpa. Plus indicates fin rays positive for evx1; minus indicates fin rays negative for evx1. (G) The Smp-MO fins show an increased frequency of evx1-positive fin rays compared with STD-MO injected fins. The data did not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent s.e.m. Student's t-test was performed to test for significance (two-tailed and unpaired, *P<0.05). Scale bar: 50 μm.
Based only on the observed differences in gene expression, it is difficult to distinguish between smp expression being reduced within all cells in sofb123 regenerating fins, or sofb123 exhibiting fewer smp-positive cells. To test whether smp and cx43 work in the same pathway to regulate joint formation, we tested for synergy of the two gene products. First, we identified the subthreshold concentration of the Smp-MO by injecting/electroporating morpholino concentrations of 1 mM (typical dose), 0.75 mM and 0.5 mM in 3 dpa regenerating fins, and measuring segment length at 7 dpa. We identified the 0.5 mM concentration of Smp-MO as the subthreshold dose (Fig. 2). We next completed KD using this dose of MO in sofb123/+ heterozygous fins, as sofb123 is recessive and therefore represents a subthreshold activity of cx43. If smp and cx43 function in a common pathway, we expected to find a significant decrease in segment length in sofb123/+ fins injected with 0.5 mM Smp-MO. Indeed, we found that there is a significant decrease in the segment length in sofb123/+ fins injected with 0.5 mM Smp-MO compared with either the STD-MO injected in sofb123/+ fins or with 0.5 mM Smp-MO injected into wild-type fins (Fig. 2). These findings provide evidence for synergy, supporting the conclusion that smp and cx43 function together in a common molecular pathway to regulate joint formation.
Previous studies have shown that Cx43 suppresses evx1 (Dardis et al., 2017). Expression of evx1 is observed in a discrete row of joint-forming cells at the distal ends of the regenerating fin in many, but not typically all, fin rays (Ton and Iovine, 2013b). Therefore, one way to evaluate changes in evx1 expression is to calculate the frequency of evx1-positive fin rays across regenerating fins (Ton and Iovine, 2013b; Dardis et al., 2017). An increase in the percentage of evx1-positive fin rays reflects abrogation of evx1 suppression. To investigate whether smp similarly inhibits evx1 expression, we monitored the percentage of evx1-positive fin rays in fins treated with Smp-MO compared with STD-MO. The percentage of evx1-positive fin rays increased in the Smp-KD fins compared with the STD-MO fins (Fig. 2), strongly suggesting that smp suppresses evx1. We validated this result through qRT-PCR, which also showed increased evx1 expression in Smp-KD fins compared with the fins treated with STD-MO (Table 1 and Fig. S1). Together, the reduced expression of smp in sofb123 regenerating fins, the evidence for synergy between cx43 and smp, and the fact that both cx43 and smp negatively regulate evx1 expression in skeletal precursor cells, strongly suggests that cx43 and smp act in a common pathway to influence joint location.
Active levels of β-catenin are reduced in both sofb123 and Smp-KD regenerating fins
It was previously shown that Smp influences β-catenin nuclear localization during zebrafish development (Kizil et al., 2014), and that β-catenin signaling contributes to regenerate length (Wehner et al., 2014). Furthermore, β-catenin activity and protein localize to the distal-most blastema as well as to the lateral skeletal precursor cells (Wehner et al., 2014; Stewart et al., 2014). Therefore, we tested whether β-catenin signaling is involved in joint formation using two independent inhibitors of canonical Wnt signaling. IWR-1 interacts with and stabilizes axin directly (Chen et al., 2009). XAV939 binds and inhibits Tankyrase enzymes that inhibit axin, and therefore similarly leads to axin stabilization (Huang et al., 2009). Both IWR-1 and XAV939 enhance β-catenin degradation (Huang et al., 2009; Chen et al., 2009). To test whether these inhibitors influence joint formation, we amputated wild-type fins at 50%. At 72 h post-amputation (hpa), fish were treated with either 10 μM IWR-1 (versus DMSO alone) or with 5 μM XAV939 (versus DMSO alone) for 3 days (changing the water and treatment each day). We observed a significant reduction in both regenerate length and in segment length using each drug (Fig. 3). The axin2 gene is a direct target of canonical Wnt signaling, and is often used as a readout of β-catenin signaling (Jho et al., 2002). Therefore, effects on β-catenin signaling following drug treatment were confirmed by demonstrating reduced expression of the target gene axin2 (Fig. 4). These findings suggest that, in addition to influencing cell proliferation, β-catenin also inhibits joint formation.
β-Catenin regulates regenerate length and segment length. (A,B,E,F) Wild-type fish at 72 hpa were treated with IWR-1 (10 µM) compared with DMSO for 3 days (n=13 per treatment) or with XAV939 (5 µM) compared with DMSO for 3 days (n=16 per treatment). The same fins were used to measure regenerate length and segment length. Regenerate length was reduced in drug-treated fins compared with DMSO-treated fins. (C,D,G,H) Segment length was reduced in drug-treated fins compared with DMSO-treated fins. The amputation plane is identified by the horizontal line, and the double-headed arrows identify regenerate length in the top panels and segment length in the bottom panels. Insets in C and G show the measured segment at a higher magnification (arrowheads indicate joints). There is a significant reduction in regenerate length (B,F) and segment length (D,H). The data in B,D,F,H do not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent s.e.m. Student's t-test was performed to test for significance (two-tailed and unpaired, *P<0.05). Scale bar: 50 μm.
β-Catenin regulates regenerate length and segment length. (A,B,E,F) Wild-type fish at 72 hpa were treated with IWR-1 (10 µM) compared with DMSO for 3 days (n=13 per treatment) or with XAV939 (5 µM) compared with DMSO for 3 days (n=16 per treatment). The same fins were used to measure regenerate length and segment length. Regenerate length was reduced in drug-treated fins compared with DMSO-treated fins. (C,D,G,H) Segment length was reduced in drug-treated fins compared with DMSO-treated fins. The amputation plane is identified by the horizontal line, and the double-headed arrows identify regenerate length in the top panels and segment length in the bottom panels. Insets in C and G show the measured segment at a higher magnification (arrowheads indicate joints). There is a significant reduction in regenerate length (B,F) and segment length (D,H). The data in B,D,F,H do not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent s.e.m. Student's t-test was performed to test for significance (two-tailed and unpaired, *P<0.05). Scale bar: 50 μm.
Active β-catenin requires Smp and Cx43. (A) Immunoblots of wild-type fin lysates compared with sofb123 fin lysates, and STD-MO fin lysates compared with Smp-MO fin lysates. Immunoblots were probed with active (non-phosphorylated) anti β-catenin antibody and with Lamin a/c antibody. (B) Graph showing the average relative densities of the β-catenin bands between experimental and control samples from three biological replicates. The amount of active β-catenin is reduced by 52% in sofb123 lysates, and by 37% in Smp-MO lysates. Error bar represents s.e.m. (C-H) axin2 expression by in situ hybridization in wild-type versus sof 5 dpa regenerating fins (C,D, n=10 for each); DMSO versus IWR-1-treated 5 dpa regenerating fins (E,F, n=10 for each); DMSO versus XAV939-treated 5 dpa regenerating fins (G,H, n=10 for each). Asterisks represent the expression domains of axin2 in each of three fin rays. The black line represents the amputation plane. Scale bar: 50 μm.
Active β-catenin requires Smp and Cx43. (A) Immunoblots of wild-type fin lysates compared with sofb123 fin lysates, and STD-MO fin lysates compared with Smp-MO fin lysates. Immunoblots were probed with active (non-phosphorylated) anti β-catenin antibody and with Lamin a/c antibody. (B) Graph showing the average relative densities of the β-catenin bands between experimental and control samples from three biological replicates. The amount of active β-catenin is reduced by 52% in sofb123 lysates, and by 37% in Smp-MO lysates. Error bar represents s.e.m. (C-H) axin2 expression by in situ hybridization in wild-type versus sof 5 dpa regenerating fins (C,D, n=10 for each); DMSO versus IWR-1-treated 5 dpa regenerating fins (E,F, n=10 for each); DMSO versus XAV939-treated 5 dpa regenerating fins (G,H, n=10 for each). Asterisks represent the expression domains of axin2 in each of three fin rays. The black line represents the amputation plane. Scale bar: 50 μm.
Next we tested whether β-catenin activity is reduced in sofb123 and in Smp-KD regenerating fins. The active form of β-catenin is the non-phosphorylated form, which can be detected with a specific antibody (Moorer et al., 2017). We first examined the amount of active β-catenin using immunoblotting. We found that the amount of active β-catenin is decreased in sofb123 lysates by about 52%±18 compared with wild-type lysates, and, similarly, in Smp-KD fins lysates by about 37%±8 compared with STD-MO-treated fins (Fig. 4).
To confirm that active β-catenin signaling is reduced in sofb123, we monitored axin2 expression. Importantly, axin2 expression is decreased in sofb123 compared with wild type by both in situ hybridization (Fig. 4) and qRT-PCR (Table 1 and Fig. S1). We similarly found reduced axin2 expression in Smp-KD fins compared with the STD-MO-treated fins (Table 1 and Fig. S1). These findings support the conclusion that β-catenin signaling occurs downstream of both Cx43 and Smp.
β-Catenin and Cx43 regulate joint location through a common molecular pathway
Next we tested whether β-catenin and Cx43 function in the same pathway by testing for synergy between sofb123/+ and IWR-1. We first identified 8 µM IWR-1 as the subthreshold dose (Fig. 5). We next tested this concentration in sofb123/+ heterozygous fish. We observed a significantly reduced segment length in sofb123/+ treated with 8 μM IWR1 compared with either the 8 μM IWR-1 dose in wild type (i.e. the subthreshold dose) or with DMSO alone in sofb123/+ (i.e. demonstrating DMSO alone had no effect) (Fig. 5). These data provide evidence that canonical Wnt signaling acts in a common pathway with Cx43 to influence joint formation.
β-Catenin and Cx43 function in a common molecular pathway. (A,B) Different concentrations of IWR-1 were used to identify the subthreshold dose on both regenerate length and segment length (8 µM concentration was selected for both) (n=13 fins per treatment). (C,D) Synergistic effects were identified for 8 µM IWR-1 in sofb123/+ heterozygotes (n=14) compared with either 8 µM IWR-1 in wild-type (n=17) or to DMSO alone in sofb123/+ heterozygotes (n=13) on both regenerate length and segment length. The data did not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent the s.e.m. Student's t-test was performed to test for significance (two-tailed and unpaired, *P<0.05).
β-Catenin and Cx43 function in a common molecular pathway. (A,B) Different concentrations of IWR-1 were used to identify the subthreshold dose on both regenerate length and segment length (8 µM concentration was selected for both) (n=13 fins per treatment). (C,D) Synergistic effects were identified for 8 µM IWR-1 in sofb123/+ heterozygotes (n=14) compared with either 8 µM IWR-1 in wild-type (n=17) or to DMSO alone in sofb123/+ heterozygotes (n=13) on both regenerate length and segment length. The data did not differ significantly from normality (Shapiro-Wilk's test, P>0.05). Error bars represent the s.e.m. Student's t-test was performed to test for significance (two-tailed and unpaired, *P<0.05).
Finally, we further tested whether β-catenin signaling influences evx1 expression. We measured levels of evx1 via qRT-PCR in IWR-1-treated fins compared with DMSO-treated fins. We found that IWR-1-treated fins indeed have higher levels of evx1 expression compared with DMSO-treated fins (Table 1 and Fig. S1). Together with our previous findings, these results suggest that Cx43, Smp and β-catenin act in a common pathway to inhibit evx1 expression and thereby regulate joint location.
DISCUSSION
We are interested in providing insights into the molecular mechanism of joint formation and, more specifically, on how joint location is selected. In our previous studies, we explored cellular changes that occur during joint morphogenesis (Sims et al., 2009), we identified a molecular pathway occurring downstream of Cx43 that influences joint formation (Ton and Iovine, 2012; Ton and Iovine, 2013b), we found that Cx43 function in the medial fibroblasts is responsible for joint formation (Dardis et al., 2017) and we showed that Cx43 suppresses both evx1 expression and joint initiation (Ton and Iovine, 2013b; Dardis et al., 2017). Here, we provide new insights into how Cx43 suppresses evx1 expression. First, we showed that smp is downstream of cx43 and functions synergistically with cx43 to suppress evx1 in the joint formation pathway. Smp has been shown to influence the nuclear localization of β-catenin in zebrafish embryos (Kizil et al., 2014). Therefore, we tested whether β-catenin influences segment length, and showed that β-catenin functions in a common pathway with Cx43 to influence joint location and evx1 expression. Furthermore, we showed that active-β-catenin, which translocates to the nucleus to transcribe Wnt target genes, is reduced in both sofb123 and Smp-MO lysates. Reduced axin2 expression in sofb123 and Smp-MO-treated fins confirmed that β-catenin signaling is reduced when either Cx43 or Smp function is reduced. From these findings, we propose a model that Cx43 activity influences β-catenin activity through regulation of smp expression in the skeletal precursor cells (Fig. 6). The smp gene is expressed more broadly than the skeletal precursor cells (Kizil et al., 2009). However, the β-catenin activity that contributes to the fin skeleton is found mainly in the lateral mesenchyme (Stewart et al., 2014; Wehner et al., 2014), and evx1 expression is restricted to a subset of the skeletal precursor cells (Ton and Iovine, 2013b). Thus, the most parsimonious explanation of our results is that the Cx43-dependent upregulation of smp facilitates an increase in nuclear β-catenin (i.e. via the Smp-NLS; Kizil et al., 2014), and suppression of evx1 expression in skeletal precursor cells. The transient decrease in cx43 expression coincident with joint initiation (described by Dardis et al., 2017) thereby relieves the Cx43/Smp/β-catenin inhibition on evx1 and permits joint formation.
Model for the role of Cx43 during joint formation. Cx43 expression in the medial mesenchyme influences joint formation in the lateral skeletal precursor cells (see Dardis et al., 2017). Here, we show that Cx43 influences smp expression, which in turn influences β-catenin signaling in the skeletal precursor cells. We propose that Smp/β-catenin suppress evx1. Thus, high levels of Cx43 favor bone formation (and inhibit evx1), whereas reduced levels of Cx43 permit evx1 expression by relieving Smp/β-catenin inhibition.
Model for the role of Cx43 during joint formation. Cx43 expression in the medial mesenchyme influences joint formation in the lateral skeletal precursor cells (see Dardis et al., 2017). Here, we show that Cx43 influences smp expression, which in turn influences β-catenin signaling in the skeletal precursor cells. We propose that Smp/β-catenin suppress evx1. Thus, high levels of Cx43 favor bone formation (and inhibit evx1), whereas reduced levels of Cx43 permit evx1 expression by relieving Smp/β-catenin inhibition.
Findings by others have shown that Wnt/β-catenin signaling is responsible for regulating blastemal cell proliferation (Wehner et al., 2014; Stoick-Cooper et al., 2007), for patterning the epidermis (Wehner et al., 2014), for maintaining the population of pre-osteoblasts (Stewart et al., 2014) and for promoting osteoblast differentiation via BMPs (Stewart et al., 2014). Our findings provide additional support for the role of Wnt/β-catenin during cell proliferation and in promoting osteoblast differentiation, and we further demonstrate that Wnt/β-catenin influences joint location. Thus, we extend previous studies by demonstrating that Wnt/β-catenin influences a cell-fate decision (i.e. between the osteoblast fate versus the joint-forming cell fate) in the skeletal precursor cells, favoring the osteoblast cell fate by inhibiting evx1 expression. We further provide evidence that this function of β-catenin occurs downstream of Cx43 and Smp. BMP signaling was also reported to occur downstream of Wnt/β-catenin, promoting continued differentiation of the osteoblast lineage (Stewart et al., 2014). It will be of interest to determine whether BMP signaling similarly inhibits the differentiation of joint-forming cells.
This model of Cx43 and β-catenin cooperating to inhibit joint formation is also supported in other model systems. In primary rat BMSCs, Cx43 increases with osteoblast differentiation, and KD suppresses both differentiation and β-catenin expression/activation (Lin et al., 2018). Further, Cx43 has been shown to promote β-catenin signaling in developing rabbit (Liu et al., 2016) and mouse (Moorer et al., 2017) skeletons, although the mechanism(s) of this regulation may not be transcriptional. For example, in the latter study the Cx43 C terminus was shown to serve as a docking platform for signaling molecules, and appears to be required for the full activity of such signaling pathways, including β-catenin (Moorer et al., 2017). Direct interactions between the Cx43 C terminus and β-catenin have been revealed (Spagnol et al., 2018). Moreover, Wnt/β-catenin signaling is absent in the cells of the mammalian joint interzone, and reducing β-catenin in these early skeletal elements activity similarly expands joint cell markers (Yamagami et al., 2009). These findings are consistent with β-catenin serving as a conserved negative signal for joint formation. Interestingly, others have also found that β-catenin signaling may promote cx43 transcription (Xia et al., 2010; Mureli et al., 2013), suggesting the possibility of positive feedback.
Several questions remain unanswered. For example, as Cx43 function in the medial fibroblasts is responsible for regulating the joint cell fate in the lateral skeletal precursor cells (Dardis et al., 2017), how is smp expression regulated by Cx43? Possible answers include that heterotypic gap junction channels are formed between Cx43 and another unidentified connexin in the skeletal precursor cells, and/or that Cx43-positive cells release a secreted growth factor that stimulates changes in gene expression in the skeletal precursor cells (also discussed by Ton and Iovine, 2013a). The mechanism for how Smp/β-catenin suppresses evx1 also requires elucidation. β-Catenin (alone, or possibly with Smp) may function as a direct repressor of evx1. β-Catenin has been shown to bind to the Evx1 promoter during gastrulation in the mouse (Funa et al., 2015), and this interaction may be conserved. Alternatively, β-catenin (i.e. ±Smp) may instead increase the expression of a repressor that, in turn, negatively regulates evx1. Importantly, a recent study identified a candidate transcription factor, hoxa13a, that may act upstream of evx1 during the specification of joint-forming cells (McMillan et al., 2018). It will be of interest to discover the direct targets of nuclear β-catenin during this process. Future studies will address these and other issues regarding the mechanism of Cx43-dependent inhibition of joint formation.
MATERIALS AND METHODS
Fish maintenance
Zebrafish (Danio rerio) were maintained in circulating water system built by Aquatic Habitats (now Pentair). Zebrafish were kept at 27-28°C in a 14:10 h light:dark period (Westerfield, 1993). The quality of the fish tank water was monitored and dosed to maintain conductivity (400-600 mS) and pH (6.95-7.30). Research was performed according to the IACUC for Lehigh University (protocol #187, 3/7/2017 approval). Food was provided to the zebrafish tanks three times daily. Every day, brine shrimp (hatched from INVE artemia cysts) was fed once and flake food twice (Aquatox AX5) supplemented with 7.5% micropellets (Hikari), 7.5% Golden Pearl (300-500 μm, Brine Shrimp Direct) and 5% Cyclo-Peeze (Argent) (Banerji et al., 2016).
Zebrafish strains
C32 and sofb123 were used (Iovine and Johnson, 2000). We used an equal number of males and females, aged between 6 months and 1 year. For amputation, fish were anesthetized in 0.1% tricaine solution and their caudal fin rays amputated to the 50% level. Regenerating fins at the indicated time points were harvested and fixed in 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) overnight at 4°C. After fixation, fins were dehydrated in 100% methanol and stored at −20°C.
Morpholino-mediated gene knockdown
All morpholinos (MOs) used in the experiments were fluorescein-tagged and purchased from Gene Tools. The MOs were reconstituted in sterile water to 1 mM. The smp-1 MO has been described previously (Kizil et al., 2009). The standard control MO was used as a negative control. Microinjection and electroporation procedures were carried out as described previously (Banerji et al., 2016; Thummel and Iovine, 2017). Briefly, caudal fins were amputated at the 50% level. At 3 days post-amputation (3 dpa), fish were anesthetized and MOs were injected using a Narishige IM 300 Microinjector. Approximately 50 nl of MO was injected per ray into either the dorsal or ventral side of the regenerating fin tissue (the first five or six bony fin rays), keeping the other side uninjected as the internal control. Immediately after injection, both sides of the caudal fin were electroporated using a CUY21 Square Wave electroporator (Protech International) (Banerji et al., 2016). The following parameters were used during electroporation: ten 50 ms pulses of 15 V with a 1 s pause between pulses. After one day post-electroporation (1 dpe), which is equivalent to 4 dpa, the injected side of the fins were evaluated by fluorescence using a Nikon Eclipse 80i Microscope (Diagnostic Instruments) to confirm MO uptake. The MO injected fins were evaluated for regenerate length (4 dpe/7 dpa), segment length (4 dpe/7 dpa), in situ hybridization, protein levels by western blot and RNA levels by qRT-PCR.
Regenerate length, segment length and statistics
Fins were calcein stained (Du et al., 2001) before measuring regenerate length and segment length. Briefly, fish were permitted to swim for 15 min in 0.2% calcein (pH 7) at room temperature, and then returned to fresh system water for 10 min. The fish were anesthetized by using tricaine and imaged using Nikon Eclipse 80i Microscope equipped with a SPOT-RTKE digital camera (Diagnostic Instruments) and SPOT software (Diagnostic Instruments). Measurement of regenerate length and segment length were taken on the 3rd fin ray from the ventral- or dorsal-most lobe of the caudal fin, as previously established (Iovine and Johnson, 2000). Measurements were analyzed using Image Pro software. Measurements for regenerate length were considered from the amputation plane to the distal tip of the 3rd fin ray. For segment length, the distance between the first two flanking joints formed following amputation plane was measured on the 3rd fin ray. For each experiment, 7-10 fish were used per trial and at least three independent trials were performed. The SPSS was used to test all data sets for normality. First, skewness and kurtosis values indicated the data did not differ significantly from normality (i.e. z-values were within the range of ±1.96). Second, we completed Shapiro-Wilk's tests (P>0.05) (Razali and Wah, 2011). Student's t-test (two-tailed, unpaired) was used to test for statistical differences (P<0.05).
Inhibition of β-catenin activity
IWR-1 was dissolved in DMSO, and this solution was diluted to the appropriate concentration (i.e. 8 or 10 µM) in 500 ml of fish water. XAV939 was dissolved in DMSO and diluted to 5 µM in 500 ml of fish water. Fish with 3 dpa regenerating fins were treated with either IWR-1, XAV939 or DMSO alone at the same concentration as the treatment (with water replacement every 24 h as needed). At the end of the treatment, the skeleton was stained with calcein. Measurements for regenerate length and segment length were collected as described above. Three trials were carried out using five to seven fish per trial.
qRT-PCR analysis
Trizol reagent (Gibco) was used to extract total RNA from a minimum of 10 fins per replicate. Regenerates were harvested with a scalpel under a dissection microscope to ensure only regenerating tissue was collected. For making cDNA, 1 mg of total RNA was reverse transcribed with SuperScript III reverse transcriptase (Invitrogen) using an oligo (dT) primer. The primers (5 mM) for simplet, keratin (Sims et al., 2009), evx1 (Dardis et al., 2017) and axin2 used for qRT-PCR analysis are listed in Table S1. Samples from three independent wild-type, sofb123, Smp-KD, standard-control-MO, IWR-1-treated and DMSO-treated RNA samples were prepared. Analyses were performed using the Rotor-Gene 6000 (Corbette Research) and the average cycle number (CT) was determined for each amplicon. Keratin was used as an internal control (Sims et al., 2009). The delta CT (ΔCT) values represent expression levels normalized to keratin values (Banerji et al., 2016). ΔΔCT values represent the relative level of gene expression. The fold difference was determined using the ΔΔCT method (2−ΔΔCT) as described previously (Ton and Iovine, 2013b). Standard deviation was calculated using the comparative method described in User Bulletin 2 # ABI PRISM 7700 Sequence Detection System (assets.thermofisher.com/TFS-Assets/LSG/manuals/cms_040980.pdf). Student's t-tests (two-tailed, unpaired) were performed to test for statistical significance (P<0.05).
In situ hybridization
Antisense digoxigenin-labeled probes were generated as described previously (Kizil et al., 2009 for smp; Iovine et al., 2005 for cx43; and Ton and Iovine, 2013b for evx1). Whole-mount in situ hybridization was performed as described previously (Sims et al., 2009). To evaluate the relative level of gene expression, whole-mount in situ hybridization was completed on four fins in each of three independent experiments.
Preparation of protein lysates and immunoblotting
Fin lysates were prepared in triplicate as described in the user manual for the Thermo Scientific nuclear and cytoplasmic extraction kit (NE-PER). Fins were amputated at 50% (10 fins for wild type and sof) and permitted to regenerate for 3 dpa. For KD analyses, 3 dpa regenerating fins (WT) were injected with either Smp-MO or STD-MO. Fins were harvested at 4 dpa and the fin regenerate tissue was homogenized by a tissue homogenizer (Bio-Gen, PRO 200) at high speed (3) for 5 s with 10 s cooling intervals in the CER buffer (Thermo Scientific). Homogenized samples were centrifuged at 200 g for 10 min at 4°C and supernatant protein levels normalized according to Bradford assays. Anti-non-phosphorylated (i.e. active) β-catenin antibody (Cell Signaling, 8814, rabbit mAb used at 1:1000) was used to detect active β-catenin. Lamin a/c (Cell Signaling, 4777S, mouse mAB used at 1:1000) was used as a loading control. Fluorescent secondary antibodies were used for detection (anti-rabbit Alexa-488 at 1:2000 and anti-mouse Alexa 647) in conjunction with Image Lab software (Bio-Rad).
ImageJ software was used to measure the band intensities. Relative pixel densities of gel bands were measured using a gel analysis tool in ImageJ software as described (Banerji et al., 2017). The density of each band was obtained as the area under the curve. For relative density calculation, the density of the active β catenin, lamin a/c bands for the sof and Smp-MO lysates was first normalized against the density of the active β catenin, lamin a/c bands from the control samples (i.e. wild type or STD-MO). Relative pixel density was calculated as the ratio of active β catenin and lamin a/c bands. The percentage reduction was calculated by subtracting the relative pixel density from 1 and multiplying by 100. The average percentage reduction with standard deviation is reported in the text. Student's t-test (two-tailed, unpaired) was performed on the relative densities to test for statistical significance (P<0.05).
Acknowledgements
The authors thank Annie Sanchez for help with axin2 in situ hybridization, Rebecca Bowman for care of the zebrafish colony, members of the Iovine lab and the Lehigh University Biological Science Department for supporting this research.
Footnotes
Author contributions
Conceptualization: S.B., M.K.I.; Methodology: S.B., D.G.; Formal analysis: S.B., D.G.; Writing - original draft: S.B.; Writing - review & editing: S.B., D.G., M.K.I.; Supervision: M.K.I.; Project administration: M.K.I.; Funding acquisition: M.K.I.
Funding
This work was supported by the National Institutes of Health grant R15-HD080507 to M.K.I. Deposited in PMC for release after 12 months.
References
Competing interests
The authors declare no competing or financial interests.