Sonic Hedgehog activates prostaglandin signaling to stabilize primary cilium length

Abstract

Sonic Hedgehog (SHH) is a driver of embryonic patterning that, when corrupted, triggers developmental disorders and cancers. SHH effector responses are organized through primary cilia (PC) that grow and retract with the cell cycle and in response to extracellular cues. Disruption of PC homeostasis corrupts SHH regulation, placing significant pressure on the pathway to maintain ciliary fitness. Mechanisms by which ciliary robustness is ensured in SHH-stimulated cells are not yet known. Herein, we reveal a crosstalk circuit induced by SHH activation of Phospholipase A2α that drives ciliary E-type prostanoid receptor 4 (EP4) signaling to ensure PC function and stabilize ciliary length. We demonstrate that blockade of SHH-EP4 crosstalk destabilizes PC cyclic AMP (cAMP) equilibrium, slows ciliary transport, reduces ciliary length, and attenuates SHH pathway induction. Accordingly, Ep4-/- mice display shortened neuroepithelial PC and altered SHH-dependent neuronal cell fate specification. Thus, SHH initiates coordination between distinct ciliary receptors to maintain PC function and length homeostasis for robust downstream signaling.

Figure 1.

PGE₂ generated downstream of SHH contributes to PC length control. (A and A′) PGE₂ was measured in IMCD3 (A) or WT and Abcc4^−/− NIH-3T3 (A′) cell culture media by ELISA. (A) Cells were pretreated with vehicle (DMSO), SMO inhibitor LDE225 (10 nM), or cPLA₂α inhibitor giripladib (GIRI, 2 µM), and then exposed to SHH or control conditioned media. PGE₂ was measured in cell culture supernatant ∼36 h after treatment. (A′) Cells were pretreated with vehicle (DMSO), LDE225 (10 nM), or COX2 inhibitor Celecoxib (Coxib,10 µM) and then stimulated with control or SHH-conditioned media. PGE₂ was measured in the supernatant ∼36 h after treatment. ELISA experiments were performed at least twice with three biological replicates per experiment. Pooled data are shown. (B) SMO PC signal intensity was determined for NIH-3T3 and Abcc4^−/− cells treated with control or SHH-conditioned media. (B′) NIH-3T3 and Abcc4^−/− cells were stimulated with control or SHH conditioned media for 18 h and induction of SHH targets Gli1 and Ptch1 was determined by qRT-PCR. Fold change over control conditioned media treatment was calculated using the 2^–∆∆Ct method. (C) IMCD3 cells were treated with SHH-conditioned media in the presence of GIRI (4 µM) or vehicle (DMSO). The PC axoneme is marked by acetylated α-tubulin (magenta). DAPI (blue) marks nuclei. Scale bar = 10 μm. (C′) Lengths of PC were quantified in IMCD3 cells treated with SHH-conditioned media in the absence or presence of GIRI. Average PC length was determined by measuring cilia of ≥100 cells/condition across three independent experiments. (D) IMCD3 cells were treated with GIRI (4 μM) in the presence of PGE₂ (40 µM) or vehicle. Scale bar = 10 μm. (D′) Quantification is shown in D′. For all experiments, statistical significance was calculated using one-way ANOVA. A P value of <0.05 was considered statistically significant with significance indicated as follows: *<0.05, **<0.01, ***<0.001, ****<0.0001, and ns, P > 0.05. Data are represented as mean ± SD.

Figure S1.

Further analysis of Abcc4^−/− cells and arachidonic acid effects on ciliary length and SMO ciliary trafficking. (A) Western blot of ABCC4 protein in control and Abcc4^−/− NIH-3T3 cells. Kinesin is the loading control. (B) Measurement of PGE₂ in conditioned media of NIH-3T3 control, Abcc4^−/−, and Abcc4^−/− cells transfected with ABCC4 expression vector. PGE₂ was measured in the supernatant ∼36 h after SHH or control-conditioned media treatment. The experiment was performed twice with three biological replicates per experiment. All data were pooled. For all experiments error bars indicate SD. (C) The indicated cell types were treated with control or SHH-conditioned media. The ciliary axoneme is marked by acetylated α-tubulin (magenta) and SMO is shown in green. DAPI (blue) marks nuclei. Scale = 5 μm. (D and D′) PC of ≥100 cells/condition was measured to determine the average ciliary length. Experiments were performed at least twice and all data were pooled. Statistical significance was determined using one-way ANOVA in D and a Student’s t test in D′. (E) IMCD3 cells were treated with control or SHH-conditioned media in the presence of vehicle or 4 µM GIRI plus increasing concentrations of arachidonic acid (12.5 or 25 µM). The ciliary axoneme is marked by acetylated α-tubulin (magenta) and DAPI (blue) marks the nuclei. Scale bar = 10 μm. (E′) Ciliary length quantification is shown in E′. (F) IMCD3 cells were treated with vehicle or GIRI (4 μM) in the absence or presence of SHH. GPR161 is magenta, SMO is red, ARL13B is green and marks PC. Scale bar = 3 μm. (F′) Quantification of ciliary signal intensity for GPR161, SMO, and ARL13B in cells treated with SHH, GIRI, or vehicle control. The experiment was performed twice with 30 cilia imaged per condition per experiment and all data were pooled. For all graphs, significance was determined using a one-way ANOVA. Significance is indicated as follows: *<0.05, **<0.01, ***<0.001, ****<0.0001, and ns, P > 0.05. Data are represented as mean ± SD. Source data are available for this figure: SourceData FS1.

Figure 2.

EP₄ is required for PC length homeostasis. (A) Endogenous EP₄ is shown in green and acetylated α-tubulin (magenta) marks the PC in an IMCD3 cell. Single channels are shown in grayscale. Scale bar = 5 µm. (A′) EP₄ localization was scored in IMCD3 cells. Approximately, 75 cells were analyzed across two experiments and all data were pooled. (B–C′) EP₄ localizes to PC in wild-type MEFs and the signal is lost in Ep4^−/− cells. Line scans indicate the degree of localization between EP₄ (green) and acetylated α-tubulin (magenta) in (B′) control and (C′) Ep4^−/− MEFs. Intensity profiles are presented as arbitrary units (AU). Scale bars = 5 μm. (D and D′) EP₄ inhibition shortens PC in IMCD3 (D) and NIH-3T3 (D′) cells. Cells were pretreated with vehicle (DMSO) or EP₄ inhibitor L161,982 (10 or 20 µM) for 2 h prior to the addition of SHH or control conditioned media. (E) IMCD3 cells were treated with control or SHH-conditioned media in the presence of vehicle or 20 µM L161,982 in media supplemented with increasing concentrations of PGE₂ (25 or 50 μM). At least 150 cilia/condition were measured over three experiments. Statistical significance was calculated using one-way ANOVA. For all experiments, a P value of <0.05 was considered statistically significant with significance indicated as follows: *<0.05, **<0.01, ***<0.001, ****<0.0001, and ns, P > 0.05. Data are represented as mean ± SD.

Figure S2.

Analysis of of Ep4^−/− MEFs. (A) Western blot showing expression level of EP₄ in control and Ep4^−/− MEFs. Kinesin is the loading control. (B and C) Confocal images of GPR161 in control and Ep4^−/− MEFs are shown. GPR161 is in cyan, ARL13B (red) marks primary cilia. Scale bar = 5 μm. (D and E) Confocal images of PTCH1-GFP are shown in control and Ep4^−/− MEFs following SHH stimulation. PTCH1-GFP is green and acetylated α-tubulin (magenta) marks PC. Scale bar = 5 μm. (F and F′) Quantification of ciliary signal intensity for (F) GPR161 and (F′) PTCH1. The experiment was performed twice with 20 cilia imaged per condition per experiment. Significance was determined using a Student’s t test. Significance is indicated as follows: *<0.05, **<0.01, ***<0.001. Source data are available for this figure: SourceData FS2.

Figure 3.

SHH signaling is attenuated by inhibition of cPLA₂α or EP₄. (A) SMO (red) enrichment in acetylated α-tubulin (magenta) marked PC was examined in IMCD3 cells. Cells were pretreated with vehicle (DMSO), GIRI (4 µM), or L161,982 (20 µM) for 18 h. Scale bar = 5 μm. (B) Average SMO ciliary signal intensity was quantified for each condition. Approximately, 50 cilia were analyzed across two experiments, and all data were pooled. (C–E) SMO (red) PC enrichment and ciliary length were examined in wild type and Ep4^−/− MEFs minus or plus EP₄ re-expression 18 h after treatment with control or SHH-conditioned media. (C) Cilia are marked by acetylated α-tubulin (magenta). Scale bar = 10 μm. Quantification of (D) average PC length and (E) fold change in SMO signal intensity are shown. Approximately, 50 cilia across two experiments were analyzed and all data were pooled. (F and F′) qRT-PCR analyses of Gli1 and Ptch1 expression in (F) IMCD3 and (F′) NIH-3T3 cells were performed. Cells were pretreated with vehicle (DMSO), GIRI (4 µM), or L161,982 (10 µM) for 2 h and then cultured in control or SHH-conditioned media plus inhibitor for 18 h. (G and G′) qRT-PCR analyses of Gli1 and Ptch1 expression in control and Ep4^−/− MEF cells were performed. Fold change in expression was determined using the 2^–∆∆Ct method. Average fold change was calculated across at least two independent experiments with three biological replicates per experiment. All data are pooled. For all experiments, statistical significance was determined using a one-way ANOVA. A P value of <0.05 was considered statistically significant. Significance is denoted as follows: *<0.05, **<0.01, ***<0.001, ****<0.0001, and ns, P > 0.05. Error bars indicate SD.

Figure 4.

Ciliary adenylyl cyclases (AC) contribute to PC length control and SHH-stimulated transcriptional activation. (A) A model for SHH and EP₄ effects on PC cAMP. (B) Ciliary cAMP was measured in control, Adcy3^−/−, Adcy5^−/−, Adcy6^−/−, and Ankmy2^−/− IMCD3 cells using the cADDis cAMP sensor. cADDis activity was recorded as cells were exposed to the vehicle (ethanol, black) or the AC activator forskolin (FSK, 100 μM, blue). Ciliary fluorescence intensity was recorded over 8 min in live cell imaging mode. An average of ∼5 cilia were recorded for each condition. The experiment was performed twice. A representative experiment is shown. Significance was determined by calculating the area under the curve followed by Student’s t-Test analysis. (C) Average ciliary lengths are shown for control, Adcy^−/−, and Ankmy2^−/− IMCD3 cells. Approximately 100 cilia across three experiments were measured and all data were pooled. (D and D′) Average ciliary lengths are shown for control, Adcy^−/−, and Ankmy2^−/− IMCD3 cells treated with (D) dBcAMP (100 μM) and (D′) SHH or SHH + PGE₂ (40 μM). Cells were pretreated with vehicle (DMSO) for 2 h prior to the addition of dBcAMP, SHH, or SHH + PGE₂. PC length was measured 18 h after stimulation for at least 150 cilia/condition over two independent experiments. Statistical significance was calculated using one-way ANOVA. All ciliary lengths were compared to vehicle-treated control IMCD3 cells in D and vehicle + SHH-treated IMCD3 cells in D′. (E) SMO ciliary signal intensity was determined in control, Adcy^−/−, and Ankmy2^−/− IMCD3 cells treated with control or SHH-conditioned media. Approximately 75 cilia across two experiments were analyzed and all data were pooled. (F and F′) qRT-PCR analyses of Gli1 and Ptch1 expression in control, Adcy^−/−, and Ankmy2^−/− IMCD3 cells were performed. Fold change in expression was determined using the 2^–∆∆Ct method. The average fold change was calculated across three independent experiments. Fold change over control conditioned media treatment was calculated using the 2^–∆∆Ct method. Significance was determined by one-way ANOVA. For all experiments, a P value of <0.05 was considered statistically significant. Significance is denoted as follows: *<0.05, **<0.01, ***<0.001, ****<0.0001, and ns, P > 0.05. Error bars indicate SD.

Figure S3.

Examination of ciliary AC localization and cAMP dynamics. (A) AC3 and AC5/6 localize to primary cilia in wild-type IMCD3 cells. ARL13B (magenta in merge and white in insets) marks primary cilia. Ciliary ACs are shown in green (merged) and in white (insets). Scale bar = 5 μm. (A′) Western blots for AC3, AC5, AC6, and ANKMY2 in lysates from control or pooled knockout IMCD3 cells. Kinesin is the loading control. (B and B′) Colocalization (M2 score) of (B) AC3 and (B′) AC5/6 with ARL13B in primary cilia of control, Adcy3^−/−, Adcy5^−/−, Adcy6^−/−, and Ankmy2^−/− cells. Statistical significance was determined using a one-way ANOVA. (C–D′) Time points from imaging of cADDis-expressing IMCD3 cells show changes in fluorescence intensity of a ciliary cAMP sensor (green) and stable fluorescence of the ciliary reference marker (magenta) following exposure to vehicle or forskolin (FSK, 100 µM). Increasing cAMP decreases green fluorescence. Scale bar = 2 μm. (D and D′) The relative cAMP shift is shown as the native (D) or (D′) inverse of the fluorescent intensity ratio of the cAMP sensor to reference marker. Inverse ratios are shown in the main figures to clearly illustrate cAMP increase. Significance was determined by calculating the area under the curve followed by Student’s t Test analysis. Significance is indicated as follows: *<0.05, **<0.01, ***<0.001, ****<0.0001, and ns, P > 0.05. Source data are available for this figure: SourceData FS3.

Figure S4.

Evaluation of PTCH1 and SMO ciliary localization in control and AC knockout cells. (A) Quantification of PTCH1-GFP ciliary signal intensity is shown for control, Adcy3^−/−, Adcy5^−/−, Adcy6^−/−, and Ankmy2^−/− IMCD3 cells. 24 h after transfection, cells were serum starved for 2 h prior to the addition of either control or SHH-conditioned media. PTCH1 ciliary intensity was measured after 18 h of stimulation. (B) SMO ciliary intensity was measured in control, Adcy^−/−, and Ankmy2^−/− IMCD3 cells following vehicle, SHH, or SHH + PGE₂ (40 μM) treatment. Cells were pretreated with vehicle (DMSO) for 2 h prior to the addition of SHH or SHH + PGE₂. SMO intensity was measured after 18 h of stimulation. SMO intensity fold change over vehicle control is shown. (C) Immunofluorescence of PC in control, Adcy3^−/−, Adcy5^−/−, Adcy6^−/−, and Ankmy2^−/− IMCD3 cells is shown following treatment with vehicle, SHH, or SHH + PGE₂. The ciliary axoneme is marked by ARL13B (magenta), SMO is green and DAPI (blue) marks the nucleus. Arrowheads mark the PC base. Scale bar = 5 μm. Significance was determined by A Student’s t Test or B one-way ANOVA. For all experiments, a P value of <0.05 was considered statistically significant. Significance is denoted as follows: *<0.05, **<0.01, ***<0.001, ****<0.0001, and ns, P > 0.05. Data are represented as mean intensities ± SD.

Figure S5.

Evaluation of primary cilium length and SHH transcriptonal output following ciliary cAMP modulation. (A) Adcy3, Adcy5, and Anmky2 knockdown in NIH-3T3 cells was validated by qRT-PCR ∼48 h after treatment with Adcy3, Adcy5, or Ankmy2 siRNA or scrambled control. Fold-change in expression was determined using the 2^–∆∆Ct method. Significance was determined using Student’s t Test. Significance is indicated as follows: *<0.05, **<0.01, ***<0.001, ****<0.0001, and ns, P > 0.05. (B) Cilia were measured for control, Adcy3, Adcy5, or Ankmy2 NIH-3T3 cells. Approximately 150 cilia were measured across two experiments, and all data were pooled. (C) Gli1 fold change was determined in control or SHH-stimulated NIH-3T3 cells following Adcy3, Adcy5, or Ankmy2 knockdown. Knockdown experiments were repeated 3 times with ∼75 primary cilia measured per experiment. Significance was determined by one-way ANOVA. (D) cADDis-expressing IMCD3 cells were pretreated overnight with vehicle or L161,982 (10 µM). The following morning, cADDis activity was monitored by live imaging to track the ciliary cAMP response as cells were treated with or without forskolin (FSK, 100 μM) for 1.5 min prior to addition of SAG (1 μM) or vehicle control. Ciliary fluorescence intensity was recorded over 8 min in live cell imaging mode. An average of ∼six cilia were recorded for each condition and the experiment was performed twice. A representative experiment is shown. (E) Schematic of optogenetic cAMP modulation in cilia. (F) Immunofluorescence imaging of IMCD3 cells expressing Cyto-bPAC-GFP or Cilia-bPAC-GFP (green) showed expression in the expected cell compartments. Acetylated α-tubulin marks cilia (magenta) and DAPI (blue) marks nuclei. Scale bars = 10 μm. (G) Average PC length was quantified in Cyto-bPAC and Cilia-bPAC expressing IMCD3 cells in the dark or following 3 h of blue light exposure. (G′) Total cellular cAMP was measured in Cyto-bPAC IMCD3 cells between resting and blue light stimulated conditions. Error bars indicate SD. Significance was determined by (G) one-way ANOVA and (G′) Student’s t test. (G″) PC of Cyto-bPAC or Cilia-bPAC expressing IMCD3 cells were imaged by confocal microscopy following 3 h of blue light exposure. Acetylated α-tubulin is magenta and DAPI is blue. Scale bar = 5 μm. (H and H′) qRT-PCR measurement of Gli1 expression in control, Cyto-bPAC or Cilia-bPAC expressing IMCD3 cells. Cells were stimulated overnight after 2 h of starvation with either (H) SHH conditioned media or (H′) SAG (100 nM) followed by either dark incubation or blue light treatment. Fold change over vehicle control was calculated using the 2^–∆∆Ct method replicates. The experiment was repeated at least twice with three biological replicates per experiment. All data were pooled. Significance was assessed using one-way ANOVA. For all experiments, significance is denoted as follows: *<0.05, **<0.01, ***<0.001, ****<0.0001, and ns, P > 0.05. Data are represented as mean ± SD.

Figure 5.

EP₄ stabilizes ciliary cAMP in SHH-stimulated cells to maintain anterograde IFT and promote SMO ciliary accumulation. (A) cADDis-expressing IMCD3 cells were pretreated overnight with vehicle or L161,982 (10 µM). The following morning, cADDis activity was monitored by live imaging to track the ciliary cAMP response as cells were treated with forskolin (FSK, 100 μM) for 1.5 min prior to the addition of the SMO agonist SAG (1 μM) or vehicle control. Fluorescence ciliary intensity was recorded over 10 min in live cell imaging mode. An average of ∼six cilia were recorded for each condition and the experiment was performed twice. A representative experiment is shown. (B) Anterograde IFT velocity was calculated in IMCD3 cells by tracking IFT88-GFP movement in the presence and absence of SHH, L161,982 (10 µM), or vehicle control. IFT velocity was calculated for 30 cilia per condition across four experiments. Velocity is shown as a violin plot with SD indicated. (C and D) Average ciliary length and SMO ciliary intensity were quantified in IMCD3-bPAC cells exposed to control or SHH-conditioned media in the absence or presence of 10 µM L161,982 in control or blue light-exposed cells. Significance was determined by one-way ANOVA. For all experiments, a P value of <0.05 was considered statistically significant. Significance is denoted as follows: *<0.05, **<0.01, ***<0.001, ****<0.0001, and ns, P > 0.05. Data are represented as mean ± SD. For all ciliary length and SMO ciliary intensity experiments, 50–100 cells per condition were analyzed over at least three independent experiments.

Figure 6.

EP₄ loss reduces PC length and alters SHH-dependent neural tube patterning in vivo. (A) Scanning electron micrographs of cardiac-level neural tube sections from E9.5 embryos of the indicated genotypes show apical membranes of cells lining the neural tube lumen. Scale bar = 2 μm. Images shown were taken from 27 somite-stage embryos. Cilia are pseudocolored in magenta. (B) The average PC length was calculated by measuring ∼30 cilia/section of cells lining the neural tube lumen. At least four/sections per embryo were analyzed across multiple embryos. (C) Cardiac level sections of E9.5 neural tubes from the indicated genotypes were immunostained for neural progenitor domain markers OLIG2 (magenta) and PAX6 (green). Images shown are from 29 somite-stage embryos. Scale bar = 100 μm. (C′) Mean expression domain areas of the indicated progenitor markers were measured and normalized to the overall neural tube area. At least four sections per E9.5/25–29 somite stage embryo per genotype were analyzed (gray dots). Pink (OLIG2) and green (PAX6) dots correspond to sections represented in C. (C″) Percent OLIG2 or PAX6-positive nuclei were counted by an automated method and normalized to the total DAPI stained nuclei count within neural tubes of each genotype. (D) Boxplots of the D/V positions of the indicated progenitor domains relative to neural tube length are shown. The floor plate midline is set to zero. Box tops indicate dorsal boundaries and box bottoms indicate ventral boundaries. Significance was determined by one-way ANOVA. Expression domain analyses were conducted on n = 4–5 embryos per genotype and 4–6 sections per embryo. All data were pooled. For all experiments, a P value of <0.05 was considered statistically significant. Significance is denoted as follows: *<0.05, **<0.01, ***<0.001, ****<0.0001, and ns, P > 0.05. Data are represented as mean ± SD.

Figure 7.

A model for SHH-to-EP4 signal crosstalk. (A) In the absence of SHH, the SHH receptor PTCH prevents ciliary accumulation and signaling of the GPCR SMO. Ciliary cAMP is maintained at a level sufficient to ensure anterograde IFT and ciliary length homeostasis through basal PGE₂ secretion and EP₄-Gα_s activation. Gα_s-coupled GPR161 is present in the PC. (B) SHH binding to PTCH leads to SMO derepression and GPR161 ciliary exit. SMO activates Gα_iβγ to increase the production of arachidonic acid by cPLA₂α. Arachidonic acid is metabolized to increase PGE₂ secretion for enhanced EP₄-Gα_s activation. This ensures ciliary cAMP levels remain sufficiently high in SHH-stimulated cells to maintain anterograde IFT and prevent ciliary shortening. (C) Inhibition of cPLA₂α or EP₄ reduces SHH-to-EP₄ signal crosstalk, slows IFT, reduces primary cilium length, and blunts SMO ciliary accumulation and high-level signaling.