Repression of smoothened by patched-dependent (pro-)vitamin D3 secretion

Abstract

The developmentally important hedgehog (Hh) pathway is activated by binding of Hh to patched (Ptch1), releasing smoothened (Smo) and the downstream transcription factor glioma associated (Gli) from inhibition. The mechanism behind Ptch1-dependent Smo inhibition remains unresolved. We now show that by mixing Ptch1-transfected and Ptch1 small interfering RNA-transfected cells with Gli reporter cells, Ptch1 is capable of non-cell autonomous repression of Smo. The magnitude of this non-cell autonomous repression of Smo activity was comparable to the fusion of Ptch1-transfected cell lines and Gli reporter cell lines, suggesting that it is the predominant mode of action. CHOD-PAP analysis of medium conditioned by Ptch1-transfected cells showed an elevated 3beta-hydroxysteroid content, which we hypothesized to mediate the Smo inhibition. Indeed, the inhibition of 3beta-hydroxysteroid synthesis impaired Ptch1 action on Smo, whereas adding the 3beta-hydroxysteroid (pro-)vitamin D3 to the medium effectively inhibited Gli activity. Vitamin D3 bound to Smo with high affinity in a cyclopamine-sensitive manner. Treating zebrafish embryos with vitamin D3 mimicked the smo(-/-) phenotype, confirming the inhibitory action in vivo. Hh activates its signalling cascade by inhibiting Ptch1-dependent secretion of the 3beta-hydroxysteroid (pro-)vitamin D3. This action not only explains the seemingly contradictory cause of Smith-Lemli-Opitz syndrome (SLOS), but also establishes Hh as a unique morphogen, because binding of Hh on one cell is capable of activating Hh-dependent signalling cascades on other cells.

Figure 1. Cholesterol Synthesis and Possible Modes of Action of Ptch1

(A) Shown is a schematic representation of cholesterol synthesis. Boxed are lathosterol and 7-DHC, cholesterol precursors known to accumulate in lathosterolosis and SLOS patients, respectively. Mutations in or genetic loss of sterol C5-desaturase (Sc5d) cause stacking of lathosterol, whereas dysfunction of 7-DHC reductase (Dchr7) or the addition of its synthetic inhibitor AY-9944 causes accumulation of 7-DHC. The conversion of 7-DHC to vitamin D3 (cholecalciferol) is mediated by UV light. Statins, like pravastatin, inhibit HMG-CoA reductase, the enzyme that forms mevalonate. (B) Shown are three possible models for the inhibitory action of the Hh receptor Ptch1 on Smo. (1) A cell-autonomous mode of action, in which direct binding of Ptch1 inhibits Smo. (2) An intracellular inhibitory action, mediated by direct binding of Ptch1 to Smo. (3) The model in which Ptch1 pumps an inhibitory small molecule that is capable of Smo-repression intercellularly (as well as intracellularly; not shown).

Figure 2. Confirmation of Functionality of Constructs and Model System Used

(A) Transfection of Ptch1 expression construct in C3H/10T1/2 fibroblasts increased Ptch1 expression over basal expression, as seen on Western blot. Actin levels remained unaltered; cells were lysed 24 h post transfection. (B) C3H/10T1/2 cells are sensitive to Hh pathway components as indicated by Gli reporter activity when pathway components are expressed: Smo increased Hh pathway activity as determined by Gli reporter luciferase assay. Cotransfection of Ptch1 suppressed Smo-induced Gli activation. Transfection of Ptch1 in the absence of Smo overexpression did not decrease Gli activity below control levels. Shh stimulation (1 μg/ml for 6 h, 16 h post transfection) and transfection of a Gli1 expression construct showed highest reporter activity, as expected. Addition of 1 μg/ml Shh-blocking antibody 5E1 reduced Shh-mediated activation of Gli reporter activity. The Ptch1-insensitive mutant SmoM2 showed a high basal activity that could not be diminished by cotransfecting Ptch1, as expected. Cells were transfected and lysed after 24 h. Values shown are relative light unit (RLU) values corrected for an internal CMV Renilla standard, expressed as percent increase relative to vector (pcDNA 3.1–) transfection or control stimulation. Depicted is the mean ± SEM. ( n = 4; *, p < 0.05; **, p < 0.01). (C) Shh concentration in medium is below the detection limit (5 ng/ml) of Western blotting. Medium was spiked with decreasing concentrations of recombinant Shh, and blotted along with a 4× concentrated medium sample obtained from C3H/10T1/2 fibroblasts (incubated for 16 h at a volume-to-surface ratio identical to the mix-and-match and medium transfer experiments).

Figure 3. Ptch1 Inhibits Smo Intercellularly

(A) Detaching and mixing of two populations of fluorescently labelled cells resulted in a homogenous mixture with intimate cell–cell contacts. The procedure is summarized in the left panel. Original magnification is 400×. Intercellular capacity of Ptch1 to inhibit Smo was assessed by mixing reporter cells overexpressing Smo and a Gli reporter with Ptch1-overexpressing donor cells. The observed decrease in Gli reporter activity could not be inhibited by the addition of 1 μg/ml 5E1 Shh-blocking antibody. Exogenously added Shh (1 μg/ml) was capable of abolishing the Ptch1-mediated inhibition. Transfecting donor cells with Ptch1 siRNA reversed the inhibitory effect (blue bars). MDA-MB-231 breast carcinoma cells (hatched bars) showed a response similar to the C3H/10T1/2 fibroblasts'. Donor cells were transfected, washed (16 h post transfection), mixed with reporter cells, and lysed after 6 to 8 h. Shown are RLU values corrected for an internal CMV Renilla luciferase control, expressed as percent differences relative to control-transfected donor cells (vector DNA or scrambled control siRNA). ( n ≥ 4; *, p < 0.05; ***, p < 0.005). Depicted is the mean ± SEM. (B) Fusion of reporter cells with Ptch1 overexpressing donor cells showed a decrease in Gli reporter activity comparable in magnitude to the inhibition observed in mixed cells. Right panel: treating fluorescently labelled cells with PEG1500 4 h after mixing resulted in multiple nuclei per cell and composite cell colours, indicating successful cell fusion. Magnification is 1,000×. Statistics and incubation times are as in (A). (C) To assess the specificity of Ptch1 acting on Smo, reporter cells expressing the Ptch1-insensitive mutant SmoM2 (as indicated on the y-axis) were mixed with donor cells. The loss of inhibition indicates the specificity of Ptch1 acting through Smo in the mix-and-match setup. Statistics and incubation times are as in (A). (D) Shown is a model for the proposed intercellular action of Ptch1 inhibition following the results obtained in the mix-and-match experiments.

Figure 4. Ptch1 Secretes Smo-Inhibitory 3β-Hydroxysteroids

(A) The left panel shows a schematic representation of the medium transfer experiments, in which medium was incubated on (transfected) donor cells and subsequently transferred to reporter cells. Right panel: intercellular Smo inhibition by Ptch1 is carried by the medium. Medium conditioned by control-, Ptch1-, or Ptch1 siRNA-transfected (C3H/10T1/2, solid bars, and MDA-MB-231, hatched bars) cells was transferred to Gli reporter cells. The same inhibitory action for Ptch1 was found as in the mix-and-match experiments. Neither 5E1 blocking antibody nor recombinant Shh addition to the reporter cells could diminish the inhibitory potential of the conditioned media. Medium conditioned on wild-type (Ptch1 ^+/+) MEFs showed a pronounced inhibitory effect on reporter cells as compared with Ptch1 knockout (Ptch1 ^–/–) MEF-conditioned medium. Gli activity in reporter cells transfected with SmoM2 was not inhibited by Ptch1 ^+/+ MEF-conditioned medium. In the absence of serum, Ptch1 transfectant conditioned medium did not contain the inhibitory activity. Incubation times are as in Figure 3B. Values shown are RLU values corrected for an internal CMV Renilla standard and expressed as percent difference to control-transfected donor cells. Depicted is the mean ± SEM. ( n ≥ 4; *, p < 0.05; **, p < 0.01; medium Ptch1-transfected compared with medium Ptch1 siRNA-transfected). (B) Vector, Ptch1, and Ptch1 siRNA transfectant-conditioned media investigated with FPLC-coupled CHOD-PAP analysis shows loading of 3β-hydroxysteroids on lipoproteins by control cells. Medium conditioned by Ptch1-transfected cells shows a strong increase in 3β-hydroxysteroids, mainly in the LDL fraction. Ptch1 siRNA transfection abolishes 3β-hydroxysteroid loading on LDL. The inset shows the lipid standard FCS containing VLDL, LDL, and HDL. Medium was incubated on cells for 16 h. Shown are typical profiles. (C) Shown is quantification of the LDL peaks, expressed as mM. LDL-C shows the Ptch1-induced increase, a reduction by HMG-CoA reductase inhibitor pravastatin treatment, and the abolition by Ptch1 siRNA transfection of 3β-hydroxysteroid loading on LDL ( n = 3; *, p < 0.05; **, p < 0.01). (D) Pravastatin-inhibited Ptch1 action on Smo-driven Gli reporter activity, whereas the addition of the cholesterol precursor mevalonate enhanced this inhibition. Cells were transfected, and, after 16 h, pravastatin or mevalonate was added for 6 to 8 h. Depicted is the mean ± SEM. ( n = 4; *, p < 0.05; **, p < 0.01). (E) Hh-induced endocytosis is not inhibited by 1 mM pravastatin ( n = 24; ***, p < 0.005). Cells were stimulated with 1 μg/ml Shh for 1 h, and preincubated with 1 mM pravastatin or control for 6 h.

Figure 5. Differentially Modulated Ptch1 Action in Dhcr7 ^–/– and Sc5d ^–/– MEFs

(A) Basal Gli reporter activity measured in MEFs either deficient for Sc5d or Dhcr7 shows that the stacking of 7-DHC (in the Dhcr7 ^–/– MEFs) lowers Gli activity in the mutant cells compared with lathosterol stacking in the Sc5d ^–/– MEFs. Transfer of the conditioned media to reporter cells (indicated as “MEFs as medium donors”) shows that the inhibitory potential of the Dhcr7 ^–/– MEFs is medium-borne. Media were incubated on donor MEFs for 16 h and transferred to reporter cells for 6 to 8 h. ( n ≥ 4; *, p < 0.05; ***, p < 0.005). (B) Medium transfer of Ptch1, Ptch1 siRNA, or control-transfected Sc5d ^–/– and Dhcr7 ^–/– MEFs shows that cells stacking lathosterol but lacking 7-DHC ( Sc5d ^–/– MEFs) are not able to translate different Ptch1 levels to an inhibitory action on reporter cells. The Dhcr7 ^–/– cells were able to properly convey an inhibitory signal when transfected with Ptch1 or, inversely, show a diminished inhibitory potential upon Ptch1 siRNA transfection. UVB treatment of Ptch1 transfectant medium (2 h) amplified the inhibitory action. Media incubations and statistics are as in (A).

Figure 6. Analysis of Vitamin D3 as a Specific Smo Antagonist

(A) Shown are the Gli reporter inhibitions by 7-DHC, AY-9944, a Dhcr7 inhibitor, vitamin D3, and a combination of the latter two. Also shown is the inhibition that can be conferred by the well-known Smo antagonist cyclopamine and the previously observed effects for Ptch1 transfectant conditioned medium as well as Ptch1 cotransfection (data taken from Figure 4B and Figure 2B, respectively). Cells were stimulated with the various compounds for 6 h (16 h past transfection), except for the cotransfection, which was lysed after 24 h ( n ≥ 4). (B) A range of concentrations of vitamin D3 was tested for inhibition of Gli reporter activity (also in SmoM2-transfected cells), MTT viability assay, and N-terminal repressor form Gli3 levels. Cells were stimulated with vitamin D3 for 6 h before lysis. For the reporter assays, RLU values were corrected for an internal CMV Renilla standard and expressed as percent difference to control stimulated cells ( n = 4). The MTT assays shown were raw absorption data expressed as percent difference to control stimulated ( n = 8); MTT agent was added for 3 h. N-terminal repressor form Gli3 was determined by quantifying ECL signal from Western blot and corrected for β-actin ( n = 3). Depicted is the mean ± SEM. (C) Gli specificity of the inhibitory effect of 10 μM vitamin D3 was assayed by using a panel of luciferase reporter constructs, one of which was a mGli reporter (mGli-LUC). Cells were stimulated for 6 h with vitamin D3, and the luciferase activity was assayed; values were calculated from RLU values corrected for an internal CMV Renilla standard and expressed as percent difference to control stimulated cells ( n ≥ 4). Depicted is the mean ± SEM. (D) Different cell types share the inhibitory response to vitamin D3 on Gli activity. C3H/10T1/2 and MDA-MD-231 cells were used earlier in the mix-and-match and medium transfer experiments. Ptch1 ^–/– MEFs served as a control for any possible effects of vitamin D3 on Ptch1 function. Cells were stimulated for 6 h, and the luciferase activity was assayed,; values are RLU values corrected for an internal CMV Renilla standard and expressed as percent difference to control stimulated cells ( n = 4). Depicted is the mean ± SEM. (E) Proposed mechanism of vitamin D3 action. First panel: in the presence of Hh, Ptch1 is inactive and does not inhibit Smo. The Hh pathway is active, and Gli activity can be measured in a reporter assay. Second panel: in the absence of Hh, Ptch1 is active and uses vitamin D3 to inhibit Smo. The pathway is inactive, and low Gli activity is measured. Third panel: in the presence of Hh as well as exogenous D3, Smo is inhibited independently of Ptch1, and Hh can no longer elicit a Gli response. (F) Confluent Shh-LIGHT II stable Gli reporter transfectants were stimulated with 10 μM vitamin D3 and/or 200 ng/ml Shh overnight in the presence of 0.5% FCS. Luciferase activity was assayed; values are RLU values corrected for an internal CMV Renilla standard and expressed as percent difference to control stimulated cells (0 μM vitamin D3; 0 ng/ml Shh). In the presence of vitamin D3, Shh is no longer able to induce reporter activity. n = 4; depicted is the mean ± SEM.

Figure 7. Scatchard Analysis of Vitamin D3 Using Smo-Expressing P. pastoris

(A) 24 h methanol induction (buffered methanol-complex medium; BMMY) of Smo-transformed P. pastoris yielded successful Smo expression as shown by Western blot analysis with an antibody directed against Smo. (B) Scatchard analysis of heterologous competition with cyclopamine representing the binding of [3H]-D3 to Smo-expressing (solid circles, MeOH induced) and nonexpressing (open circles, noninduced) yeast strains. The solid line indicates the induced high-affinity binding, whereas the dotted line indicates nonspecific low-affinity binding. Scatchard analysis and fitting was performed as described earlier [ 53].

Figure 8. Effects of Vitamin D3 on Developing Zebrafish Embryos

(A) Fertilized zebrafish eggs were dechorionated and placed in buffer containing 1.2 mg/ml sonicated vitamin D3, starting the continuous incubation between the 32- and 64-cell stages. At the different developmental stages (as indicated), the embryos were analysed, revealing persistent somite malformations, a ventrally curved body, poorly developed myoseptum, an aberrant extension of the yolk tube, a prominent dorsal midbrain, and a reduced trunk. The wild-type Engrailed staining (4D9) of muscle pioneers and surrounding cells in control-treated animals was not observed upon vitamin D3 treatment. ISH for ptc1 mRNA showed a strongly reduced signal in vitamin D3–treated animals compared with controls. Staining with the F59 antibody revealed that slow muscle fibres were disturbed in number and orientation in the treated animals. (B) At 18 hpf, embryos were photographed in detail and somite angles were determined (as shown in top panel). The measurement was taken of an individual somite between dorsal and ventral portions of the vertical myoseptum for corresponding somites between the wild-type and vitamin D3–treated embryos. Shown are the mean angles (±SEM; n = 3) of wild-type control- and vitamin D3–treated embryos. C) Detail of 4-dpf embryo. Note the slightly enlarged pericardial cavity in vitamin D3–treated animals. The orientation of all images is a lateral view, anterior to the left.