Wnt7b Activates Canonical Signaling in Epithelial and Vascular Smooth Muscle Cells through Interactions with Fzd1, Fzd10, and LRP5

Histological procedures.

Mouse embryos were fixed in 4% paraformaldehyde for 48 h and dehydrated through a series of increasing ethanol washes. Fixed embryos were embedded in paraffin, and 5 μm sections were affixed to slides, which were processed for radioactive in situ hybridization as previously described (45). Hybridized slides were coated with photographic emulsion and exposed for 7 to 10 days. Photographs were taken on a Nikon E600 microscope equipped with epifluorescence. In situ hybridization probes were generated by PCR from full-length cDNAs and contained T7 RNA polymerase binding sites in the antisense direction at the 3′ end. The oligonucleotides used to generate Fzd1-Fzd9 probes have been previously described (29). Oligonucleotides used to generate in situ probes for LRP5 and LRP6 were LRP5 sense (5′-GCGGAGTGAAGCTGGAGTCC-3′) and antisense (5′-GTAATACGACTCACTATAGGGCCTCGCATGGTGTGTGGAAGGG-3′) and LRP6 sense (5′-TACAAATGGCAAAGAGAATG-3′) and antisense (5′-GTAATACGACTCACTATAGGGCTTCCACATGGATTTGTAGCA-3′). Immunohistochemistry using the rabbit anti-Fzd10 antibody was performed as previously described (22). More details on histological procedures can be found at the University of Pennsylvania Molecular Cardiology Center web site: http://www.uphs.upenn.edu/mcrc/.

Plasmids and RT-PCR.

Plasmids containing mouse LRP5, LRP6, Fzd4, and Fzd7 cDNAs have been previously described (43). Expression vectors for Fzd1, Fzd4, Fzd7, and Fzd10 were generated in pCMVTag4A (Stratagene, Inc.) using reverse transcription-PCR (RT-PCR) from either mouse E14.5 embryonic lung RNA or the plasmids described above. Generation of the pcDNA3.1/Wnt7bHA expression vector was previously described (31). The DEP/dishevelled expression vector and the β-catenin/4145 expression plasmid in which the GSK-3β phosphorylation sites Thr 41 and Ser 45 are mutated to alanines have been previously described (31, 33).

Total RNA was extracted from 293 and PAC-1 cells using Trizol reagent. One microgram of DNase-treated total RNA was reverse transcribed using a commercially available kit (Invitrogen, Inc.). A total of 5% of the resulting cDNA from each cell line was subjected to PCR amplification using oligonucleotides specific for Fzd1 (sense, 5′-GGAGAGCTGTGCGTGGGCCAG-3′; antisense, 5′-AAACTTGTCGTTGCACACCAC-3′); Fzd4 (sense, 5′-CTTCACCGTGCTGACCTTCCTG-3′; antisense, 5′-TTGACCATTAGCCTTTCCAAC-3′); Fzd7 (sense, 5′-CATGCGTCGCTTCAGCTATCC-3′; antisense, 5′-CTGTCTTGGTGCCGTCGTGC-3′); and Fzd10 (sense, 5′-AGAAGCTCATGGTACGCATAG-3′; antisense, 5′-CACGCAGGCTGAAGGCTGGG-3′). The sizes of the resulting amplification products were as follows: Fzd1, 484 bp; Fzd4, 601 bp; Fzd7, 550 bp; and Fzd10, 430 bp.

Wnt reporter transfection assays.

To measure canonical Wnt signaling, the TOPFLASH and FOPFLASH reporter plasmids (0.5 μg) along with Wnt7b, Fzd, and LRP expression plasmids (0.5 μg) were transfected into 293, 3T3, PAC-1, or A7r5 cells using Fugene 6 as previously described (39). At 48 h after transfection, luciferase activity was measured using a commercially available kit (Promega) and compared to control Renilla luciferase activity.

JNK activity was measured using the PathDetect c-Jun kinase reporter system (Stratagene, Inc.). 293 and PAC-1 cells were transfected with the pFR-luc reporter plasmid, the pGAL4-c-Jun-N-term vector, and either pcDNA3/Wnt7b, pcDNA3/DEP, or pcDNA3 as a negative control. The pGAL4DBD was also used as a negative control. Luciferase activities were measured as described above.

Western blotting.

Cell extracts were made using an NE-PER extraction kit (Pierce, Inc.), and lysates were resolved in 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels, which were transferred to Immobilon polyvinylidene difluoride membranes. For determining c-Jun phosphorylation, the membranes were probed with the phospho-c-Jun antibody KM-1 (Santa Cruz Biotechnology, Inc.) at a 1:2,000 dilution. Total c-Jun protein expression was determined using the c-Jun (N) antibody (Santa Cruz Biotechnology, Inc.) at a 1:1,000 dilution. For the β-catenin Western blots, the membranes were probed with a mouse anti-β-catenin monoclonal antibody at a 1:2,000 dilution (BD Transduction Laboratories). Western blots were further processed as previously described to detect immunospecific bands (39).

Cell surface binding experiments.

To detect cell surface binding of Wnt7b to Fzd proteins, Wnt7b conditioned medium was generated from 293 cells transfected with the pcDNA3.1/Wnt7b-HA plasmid. At 72 h after transfection, conditioned medium was concentrated 20-fold in a Centricon-20 mini concentrator. 293 cells were transfected with expression plasmids containing full-length Fzd1, Fzd4, Fzd7, and Fzd10. At 48 h after transfection of Fzd expression plasmids, the 293 cells were incubated with 200 μl of concentrated Wnt7b conditioned medium for 1 h at room temperature. Cells were then washed three times with phosphate-buffered saline, fixed for 10 min with 4% paraformaldehyde, and then incubated with mouse monoclonal to the HA epitope (Convance, Inc.) for 1 h at room temperature. Cells were washed again and incubated with goat anti-mouse immunoglobulin G-fluorescein isothiocyanate for 1 h at room temperature, after which cells were washed again three times with phosphate-buffered saline and visualized under a fluorescent microscope with a fluorescein isothiocyanate filter.

Coimmunoprecipitation assays.

293 cells were transfected using Fugene 6 and the pcDNA3.1Wnt7b-HA expression plasmid and Fzd1, Fzd4, Fzd7, and Fzd10 expression plasmids. Cell lysates were generated using a lysis buffer containing 50 mM Tris, pH 7.5, 150 mM NaCl, 0.5% NP-40, 1 mM phenylmethylsulfonyl fluoride. Mouse anti-FLAG epitope tag antibody (M2; Sigma) was used to immunoprecipitate cell lysates, which were resolved on SDS-PAGE gels, blotted to Immobilon membranes (Millipore, Inc.), and subsequently probed with mouse anti-hemagglutinin (HA) tag antibodies to detect coimmunoprecipitated proteins on Western blots.

Expression of frizzled and LRP5/6 receptors during lung development.

Previous studies have demonstrated an important role for Wnt7b signaling in lung airway and vascular development (19, 31). To narrow our search for receptors used by Wnt7b in the lung, in situ hybridization was performed on mouse embryonic lung sections using probes for all 10 of the known Fzd cDNAs. Of these, Fzd1, 4, 7, and 10 were found to be expressed in the lung at significant levels from E12.5 through E16.5, the time period during which the lung undergoes extensive differentiation. Fzd1 and 7 were expressed in the developing lung mesenchyme and in presumptive bronchial smooth muscle as well as vascular smooth muscle tissues such as the aorta and the large pulmonary vessels (Fig. 1A to C and G to I). Fzd4 is expressed in lung mesenchymal cells, which include precursors to vascular endothelial cells in the lung (Fig. 1D to F). This expression pattern correlates with Fzd4's previous reported expression in endothelial cells (10, 41). Immunohistochemistry shows that Fzd10 is expressed primarily in distal airway epithelium during lung development, with expression decreasing with gestational age (Fig. 1J to L).

In situ hybridization for LRP5/6 expression revealed high levels of expression for both genes in the distal lung airway epithelium as early as E12.5 (Fig. 2A to F). Expression of LRP5/6 in airway epithelium decreased after E14.5, and both genes were expressed only at low levels in distal airway epithelium at E18.5 (Fig. 2E and F). LRP5 and LRP6 expression was noticed in large upper airways at E14.5 but by E18.5, only LRP6 expression was observed in upper airways (Fig. 2E to H). Expression of LRP5 was observed in developing blood vessels such as the dorsal aorta and pulmonary blood vessels (Fig. 2A to D, G, H, K, and L). Expression of LRP5 but not LRP6 was also observed at E14.5 in the outflow tract valves of the heart, where Wnt signaling has been demonstrated to play a key role in development (Fig. 2I and J) (11, 15, 21). These data suggest that LRP5 and LRP6 likely play both unique and redundant roles in Wnt signaling in pulmonary and cardiovascular development.

Wnt7b activates canonical Wnt signaling in a cell-specific manner.

To determine whether Wnt7b could activate the canonical Wnt pathway, 293 and 3T3 cells were transfected with the TOPFLASH and FOPFLASH reporter constructs along with a Wnt7b expression vector. 293 cells are a kidney epithelial cell line whereas 3T3 cells are a mesodermally derived mesenchymal cell line. Wnt7b was unable to activate the TOPFLASH reporter in either of these cells types while a phosphorylation mutant of β-catenin, β-catenin/4145, known to activate the canonical Wnt pathway, activated the TOPFLASH reporter robustly (Fig. 3A and B) (34). Our previous work demonstrated a role for Wnt7b activity in vascular smooth muscle integrity in the lung (31). To determine whether Wnt7b activated canonical signaling in vascular smooth muscle cells (VSMCs), we performed similar transfection studies with PAC-1 and A7r5 cell lines, which are derived from rat pulmonary vascular smooth muscle and aortic smooth muscle, respectively. Surprisingly, Wnt7b was able to activate the TOPFLASH reporter approximately eightfold in both of these cell lines, indicating that Wnt7b activates canonical Wnt signaling in a cell-specific manner (Fig. 3A).

Fzd expression was determined in 293 and PAC-1 cells to determine whether expression of these receptors could explain the differences in Wnt7b activation of canonical signaling (Fig. 3C). Fzd1 and Fzd7 are expressed in both 293 and PAC-1 cells. Fzd4 and Fzd10 are expressed at low levels in PAC-1 cells.

Wnt7b does not activate c-Jun kinase-dependent noncanonical Wnt signaling.

Wnt ligands can activate either canonical or noncanonical pathways, often in a cell context-dependent manner. Since Wnt7b could not activate canonical signaling in 293 or 3T3 cells, both of which are receptive to Wnt signaling (1, 14, 17, 25), we wanted to determine whether Wnt7b could activate the noncanonical planar cell polarity pathway. This pathway has been shown to activate c-Jun kinase (JNK), resulting in the phosphorylation of its downstream targets such as c-Jun (4, 20, 40, 44). To determine whether Wnt7b activated the JNK pathway, we used a transcriptional reporter system to measure activation of JNK by Wnt7b in 293 and PAC-1 cells. This assay measures the change in gene transcription through phosphorylation of the activation domain of c-Jun when fused to the GAL4 DNA binding domain (20). Wnt7b was unable to activate JNK while the Dishevelled-EGL10-Pleckstrin (DEP) domain of dishevelled, which has been shown to be critical for JNK mediated Wnt signaling (12, 32), significantly activated JNK in both 293 and PAC-1 cells (Fig. 4A). Cotransfection of LRP5 or LRP6 did not affect the ability of Wnt7b to activate JNK (Fig. 4A). Western blots demonstrated that expression of Wnt7b in 293 cells did not alter the level of c-Jun phosphorylation while expression of the DEP domain significantly increased c-Jun phosphorylation (Fig. 4B). These data indicate that Wnt7b is unable to activate JNK-dependent noncanonical Wnt signaling.

Wnt7b binds to cells expressing Fzd1 and Fzd10.

The ability of Wnt7b to activate canonical Wnt signaling in a cell-dependent manner, along with the cell-restricted pattern of Fzd expression, suggested that Wnt7b activity is likely regulated by the expression pattern and ability to interact with specific members of the Fzd family. To determine whether Wnt7b can bind to any of the four Fzds expressed in the lung during development, conditioned medium containing hemagglutinin (HA)-tagged Wnt7b (Fig. 5A) was incubated with 293 cells transfected with expression constructs encoding Fzd1, 4, 7, and 10. Equivalent expression of each of these four Fzd proteins in transfected 293 cells was confirmed by Western blotting (data not shown). After incubation and staining with an anti-HA epitope antibody, cells expressing Fzd1 and Fzd10 were found to bind Wnt7b at significant levels (Fig. 5D and G). However, cells expressing Fzd4 and Fzd7 did not bind Wnt7b at detectable levels (Fig. 5E and F). These data suggest that Wnt7b binds avidly to Fzd1 and Fzd10.

Wnt7b coimmunoprecipitates with Fzd1, Fzd4, and Fzd10.

Coimmunoprecipitation assays were performed to further assess the ability of Wnt7b to bind to Fzd1 and Fzd10. 293 cells were transfected with an HA-tagged Wnt7b expression vector along with a FLAG-tagged Fzd1, Fzd4, Fzd7, and Fzd10. Cell lysates were immunoprecipitated with anti-FLAG antibody, and coprecipitated proteins were detected on Western blots with the anti-HA epitope antibody. Interestingly, Wnt7b precipitates with Fzd1, Fzd4, and Fzd10 but not Fzd7 (Fig. 5H). These data show that Wnt7b binds efficiently to Fzd1 and Fzd10, supporting the findings from the above-described cell binding experiments. The ability of Wnt7b to interact with Fzd4 in a coimmunoprecipitation assay but not in a cell surface binding assay suggests either that there is additional specificity for ligand-receptor interactions that is conferred at the cell surface or that the cell surface binding assay is less sensitive than the coimmunoprecipitation assay.

Wnt7b activates canonical Wnt signaling in cooperation with Fzd1, Fzd10, and LRP5 in 293 cells.

Since the above-described data suggest that Fzd1 and Fzd10 were capable of binding Wnt7b at the cell surface, we sought to determine whether they could activate canonical Wnt signaling in cooperation with Wnt7b in a cell line not responsive to Wnt7b alone, such as the kidney epithelial cell line 293. 293 cells were transfected with expression vectors for Wnt7b and Fzd1, 4, 7, or 10. None of these combinations resulted in active signaling (Fig. 6A). LRP5/6 coreceptors, along with Fzd proteins, are required for canonical Wnt signaling (24, 33, 38). To determine whether LRP5 or LRP6 could cooperatively activate canonical Wnt signaling with Wnt7b, 293 cells were transfected with expression vectors for Wnt7b and LRP5 or LRP6. Data from these experiments demonstrate that LRP5 but not LRP6 can cooperatively activate canonical Wnt signaling with Wnt7b (Fig. 6B). These data suggest that 293 cells express an Fzd receptor at low levels or that an Fzd receptor binds to Wnt7b only weakly and that the receptor's activity is augmented by addition of LRP5. Expression of low levels of Fzd1 in 293 cells may help explain this finding (Fig. 3C).

To determine whether the addition of Fzds and LRP5 would further cooperatively activate Wnt signaling with Wnt7b, 293 cells were cotransfected with a combination of these expression vectors. As shown in Fig. 6C, addition of Fzd1 and Fzd10 resulted in an additional increase in Wnt signaling as measured by the TOPFLASH reporter. Western blot analyses were performed to determine whether this increase in TOPFLASH activity corresponded with an increase in steady-state levels of β-catenin, which occurs in canonical Wnt signaling. The Wnt7b/Fzd1/LRP5 and Wnt7b/Fzd10/LRP5 combinations significantly increased stabilized levels of β-catenin in transfected 293 cells while Wnt7b/Fzd4/LRP5 increased stabilized levels of β-catenin protein modestly. Interestingly, Wnt7b/Fzd1 and Wnt7b/Fzd10 caused an increase in stabilized levels of β-catenin even though Wnt activity as measured by TOPFLASH activation did not significantly change (Fig. 6A and D). This suggests that a threshold of β-catenin stabilization must be meet before Wnt7b activation of gene transcription is measurable by the TOPFLASH reporter plasmid. Together, these data demonstrate that Wnt7b cooperatively activates canonical Wnt signaling with LRP5 and Fzd1 and Fzd10.

Wnt7b activates canonical Wnt signaling in cooperation with Fzd1, Fzd10, and LRP5 in PAC-1 cells.

Wnt7b^lacZ −/− embryos exhibit defects in pulmonary vascular smooth muscle integrity (31). Since Wnt7b is expressed in the airway epithelium in the lung and not VSMCs or their mesenchymal precursors (31), it must act in a paracrine manner to regulate VSMC development. Previous reports have suggested that VSMCs are responsive to β-catenin/Wnt signaling (36). To determine whether Wnt7b signals through LRP5 and Fzd1 and Fzd10 in VSMCs, PAC-1 cells were transfected with expression plasmids encoding combinations of Wnt7b, LRP5, and Fzd1, 4, 7, and 10. As expected based on data in Fig. 2, Wnt7b activated the TOPFLASH reporter in PAC-1 cells (Fig. 7). However, Wnt7b plus LRP5 generated a significant increase in Wnt activity as measured by the TOPFLASH reporter (Fig. 7). Addition of Fzd1 and Fzd10 further increased Wnt activity over LRP5 alone (Fig. 7). These data demonstrate that Wnt7b activates canonical Wnt signaling in VSMCs cooperatively through LRP5 and Fzd1 and Fzd10.