Planar cell polarity breaks the bilateral symmetry by controlling ciliary positioning

Abstract

Defining the three body axes is a central event of vertebrate morphogenesis. Establishment of left-right (L-R) asymmetry in development follows the determination of dorsal-ventral (D-V) and anterior-posterior (A-P) body axes^1,2, though the molecular mechanism underlying precise L-R symmetry breaking in reference to the other two axes is still poorly understood. Here by removing both Vangl1 and Vangl2, the two mouse homologues of a Drosophila core planar cell polarity (PCP) gene Van Gogh (Vang), we have uncovered a previously unappreciated function of PCP in initial breaking of lateral symmetry. The leftward nodal flow across the posterior notochord (PNC, also referred to as “the node”) has been identified as the earliest event in the de novo formation of L-R asymmetry^3,4.We found that PCP is essential in interpreting the A-P patterning information and linking it to L-R asymmetry. In the absence of Vangl1 and Vangl2, cilia are positioned randomly around the center of the PNC cells and nodal flow is turbulent, which results in disrupted L-R asymmetry. Importantly, PCP in mouse, unlike what has been implicated in other vertebrate species, is not required for ciliogenesis, cilium motility, Sonic hedgehog (Shh) signaling or apical docking of basal bodies in ciliated tracheal epithelial cells. Our data suggest that PCP acts earlier than the unidirectional nodal flow during bilateral symmetry breaking in vertebrates and provide insight into the functional mechanism of PCP in organizing the vertebrate tissues in development.

Identifying the mechanism underlying the initial bilateral symmetry breaking has been an overriding goal in understanding how nature is able to produce organisms with consistent and distinct left and right anatomies. In mice, bilateral symmetry is broken by a leftward flow across a pit-like, teardrop-shaped PNC generated by posteriorly localized motile cilia^3–6. However, an earlier asymmetric event that positions the cilia to the posterior sides of PNC cells in response to the A-P patterning signals remains to be identified. PCP, originally refers to asymmetry of epithelial cells along an axis orthogonal to the apical-basal axis, is an evolutionarily conserved mechanism by which positional information is transmitted and maintained among and within cells⁷. A group of core PCP proteins including the transmembrane protein Vang have been identified in Drosophila⁸. While mutants of core PCP genes have thus far failed to show L-R asymmetry defects, it is likely due to functional redundancy among members of a core PCP component. To overcome this and uncover important regulatory functions of PCP, we inactivated both Vangl1 and Vangl2, the two mouse homologues of the Drosophila Vang.

Vangl1 and Vangl2 were broadly expressed during L-R patterning (Supplementary Fig. 1). Importantly, prior to appearance of ciliated PNC cells, Vangl1 and Vangl2 protein levels were too low to be detected. However, after E7.75, both Vangl1 and Vangl2 protein levels were selectively and progressively upregulated in the PNC cells and asymmetrical localization of Vangl1 and Vangl2 protein was observed along the anterior-posterior (A-P) axis (Supplementary Fig. 2a–c), indicating that the PNC cells are indeed polarized and PCP may be required by the PNC cells to regulate early embryonic development. The Vangl1 and Vangl2 localization patterns are also consistent with the role of PCP in regulating convergent extension (CE) of neurulating vertebrate embryos⁹. In the sensory epithelia of the cochlea, Vangl1 and Vangl2 demonstrated similar asymmetrical localization pattern in both hair cells and supporting cells (Supplementary Fig. 2d, e). These data suggest that Vangl1 and Vangl2 play redundant roles in PCP, which should be severely disrupted by removing both Vangl genes.

A bona fide null allele of Vangl1 (Vangl1^gt) was generated using a gene trap ES cell line (Supplementary Fig. 3). We also generated a floxed Vangl2 allele, Vangl2^f, from which the null Vangl2^Δ allele was derived (Supplementary Fig. 3). Vangl1 and Vangl2 regulated PCP in a gene dose-dependent manner (Fig. 1 and Supplementary Fig. 4). In the cochlea of the Vangl2^Δ/Δ mutant, asymmetrical Vangl1 localization was still detected (Fig. 1a). The Vangl2^Δ/Δ; Vangl1^gt/gt embryos showed the most severe developmental defects compared to embryos with other genotypes, particularly when sensory hair cell polarity and CE were examined to assess PCP. Thus, Vangl1 and Vangl2 function of regulating PCP was removed in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos.

Figure 1. The Vangl2^Δ/Δ; Vangl1^gt/gt embryos exhibit the most severe polarity defects.

a, Asymmetrical localization of Vangl1 protein (green) in the sensory hair cells of the Vangl2^Δ/Δ cochlea at E18.5. b, The sensory hair cell polarity at E18.5 is shown by the orientation of actin-based stereocilium bundles (red) and the microtubule-based kinocilia (green). c, Quantification of misoriented hair cells in embryos with different genotypes. Data are means ±SD from 3 samples. The orientation defects in the cochlea ranged from few misoriented cells (Vangl2^Δ/+;Vangl1^gt/gt) to several affected cells (Vangl2^Δ/Δ, Vangl2^Lp/+;Vangl1^gt/gt and Vangl2 ^Lp̃Lp) to almost complete loss of polarity (Vangl2^Δ/Δ; Vangl1^gt/gt).

The Vangl2^Δ/Δ; Vangl1^gt/gt embryos exhibited multiple laterality defects, which were not detected in embryos with other genotypes (Fig. 2a). The Vangl2^Δ/+; Vangl1^gt/+ mice were fertile, morphologically normal and served as controls in our study. The Vangl2^Δ/+; Vangl1^gt/+ and Vangl2^Δ/Δ; Vangl1^gt/+ embryos always turned to the right side and their hearts correctly looped to the right (n>50), but in all of the Vangl2^Δ/Δ; Vangl1^gt/gt embryos assayed (n>10), both embryonic turning and heart looping failed. In addition, the normal mouse lung has five lobes, four on the right and one on the left. In the Vangl2^Δ/Δ; Vangl1^gt/gt embryos, however, mild left lung isomerism was observed with three lobes on the right side (4/9). Since PCP in visceral organs may also control later stage morphogenesis, to further assess the early L-R asymmetry defects in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos, we examined the components in the Nodal signal cascade that are asymmetrically expressed and required for L-R patterning¹⁰. In control embryos, Nodal was first expressed symmetrically around the PNC (somite 0–1). Then between somite 2–7 stages, Nodal expression on the left side of the PNC became progressively stronger and its expression in the left lateral plate mesoderm (LPM) was also detected^11,12. In the Vangl2Δ/Δ; Vangl1^gt/gt embryos, however, Nodal expression around the PNC was reduced (6/11) (Fig. 2b, c). Strikingly, more robust Nodal expression on the right side of the PNC was also observed in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos (5/11) (Fig. 2b, c). In the LPM of Vangl2^Δ/Δ; Vangl1^gt/gt embryos, reversed (1/8) or bilateral (3/8) Nodal expression sites were observed (Fig. 2c). Controlled by Nodal signaling, expression of Lefty1, Lefty2 and Pitx2 was only observed on the left side of the control embryos^10,13,14,15. In the double mutant embryos, Lefty1 and 2 expression sites (1/2) were reversed (Fig. 2d) and bilateral Pitx2 expression (2/7) was found (Fig. 2e). These results demonstrate that PCP is required upstream of Nodal to pattern the L-R axis. The incomplete penetrance of the laterality defects in the Vangl1/2 mutant embryo is possibly due to other input(s) to the PCP pathway¹⁶.

Figure 2. Laterality defects in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos.

a, Failure of embryonic turning and heart (marked by Nkx2.5 expression) looping in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos. The Vangl2^Δ/Δ; Vangl1^gt/gt lung has three lobes (outlined) instead of four on the right side. R: right; L: left. b, Perinodal Nodal expression is enhanced on the right side of the Vangl2^Δ/Δ; Vangl1^gt/gt embryo (arrow). c, Ventral views of Nodal expression in LPM (upper panel) and around the node (lower panel, arrows). Weak perinodal Nodal expression was outlined. d, Lefty1 expression (orange arrow) on the left side of prospective floor plate and Lefty2 expression (white arrow) in the LPM were shown in ventral views (upper panel) and transverse sections (lower panel). e, Ventral views of Pitx2 expression.

Since PCP has been implicated in ciliogenesis¹⁷, we examined the morphology and movement of cilia in PNC cells. Cilia are long, thin microtubule-based projections from basal bodies of most mammalian cells¹⁸. The PNC cell has a single motile cilium³, which is essential for establishing leftward fluid flow across the ventral surface of the PNC. When examined by scanning electron microscopy and immunofluorescent labeling, cilia appeared normal in the PNC of the Vangl2^Δ/Δ; Vangl1^gt/gt embryos (n=4 in each case) at E8.0 (Fig. 3a, b). Thus, the L-R asymmetry defects in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos are not secondary to defects in ciliogenesis. The leftward nodal flow is enabled by the chirality of the cilium, as it rotates in a clockwise manner³, and the posterior ciliary localization in the PNC cell, which results in a posterior tilt of the cilium and its rotation axis^5,6. Since it has been postulated that the posterior positioning of the basal body might be controlled by PCP⁶, we examined basal body localization in the PNC cells of the Vangl mutant embryos (Fig. 3c, d). The basal bodies were mostly localized posteriorly in both Vangl2^Δ/+; Vangl1^gt/+ and Vangl2^Δ/Δ; Vangl1^gt/+ embryos. However, in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos, the PNC cell density was slightly decreased (Supplementary Fig. 5a) and the number of posteriorly localized basal bodies was significantly reduced. Most basal bodies remained around the center and anteriorly localized basal bodies were increased, resulting in an even distribution of basal bodies on the anterior and posterior sides of the PNC cells in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos (Fig. 3d and supplementary Fig. 5b, c). Such loss of basal body posterior localization in the Vangl mutant is not a result of developmental delay (Supplemental Fig. 5c). Thus, posterior positioning of basal bodies in the PNC cells is indeed controlled by PCP, in the absence of which, basal bodies are mostly localized around the center. Cilia in the PNC emerge from the center of each cell at late streak (LS) stage (~E 7.25). By early head fold (EHF) stage (E7.75), most cilia are located in the posterior side of the PNC cells^6,19. As positioning of the kinocilium in sensory hair cells is also an event downstream of PCP²⁰, ciliary positioning may be a common readout of PCP.

Figure 3. Randomized ciliary positioning in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos.

a, Scanning electron microscopy of the PNC showing normal cilia at E8.0. b, Normal cilia (acetylated tubulin, red) in the PNC of the Vangl2^Δ/Δ; Vangl1^gt/gt embryo at E8.0. c, Basal 16 bodies (γ–tubulin, green) and cell boundaries (F-actin, red) of PNC cells at somite 0 stage. White lines indicate the middle position along the A-P axis of each PNC cell. A: anterior; P: posterior. d, Relative positions of basal bodies to the midline (along the A-P axis) of the PNC cells were plotted. Posteriorly biased basal body positioning was lost in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos.

To test whether altered ciliary positioning in the PNC cells affect function, we examined nodal flow (Fig. 4a and Supplementary movie 1, 2). In control embryos, fast and smooth leftward bead movement across the PNC was observed. In the Vangl2^Δ/Δ; Vangl1^gt/gt embryos, bead movement was still biased leftward, though beads were often trapped in swirling vortices and requires more time to cross the PNC. While the nodal flow of control embryos generated straight and parallel bead paths, in the mutant, bead paths contained “knots”, abrupt turns and were often crossed with each other (2/2). When observed directly, the cilia in the PNC of the Vangl2^Δ/Δ; Vangl1^gt/gt embryo were beating at a speed and direction similar to that of the control (Fig. 4b and Supplementary movie 3, 4). The abnormal nodal flow in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos also disrupted the left-biased perinodal rise of intracellular calcium in control embryos²¹ (Fig. 4c, d). Incidence of bilateral rise or reduction of intracellular calcium was increased in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos. The ectopic calcium signal on the right side was weaker and the domain was narrower and shorter. Taken together, unidirectionality of nodal flow was significantly compromised in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos. Thus, we have identified PCP as the first step in breaking L-R symmetry by transmitting the A-P positional information to the PNC cells to enable posterior ciliary positioning (Fig. 4e).

Figure 4. Vangl1 and Vangl2 are required for normal leftward nodal flow.

a, Nodal flow shown by tracing latex bead movement across the PNC with lines in different colors (Supplementary movie 1, 2). A: anterior; P: posterior; L: left; R: right. b, Ciliary beating was recorded by live imaging (Supplementary movie 3, 4) and the mean beating frequency was calculated from at least 10 nodal cilia of 3 embryos with indicated genotypes. Data are means ±SD. c, Intracellular calcium release around the node, intensity scales are shown in the lower panels. d, Distribution of embryos (percentage and number) according to calcium patterns. e, Schematic diagram showing the role of Vangl1 and Vangl2 in breaking the L-R symmetry. I. A mouse embryo at E8.0. (II) Enlarged distal view of (I). (III) Normal nodal flow generated by posteriorly positioned cilia. A side view of ciliated cells are shown in the lower panel. (IV) In the Vangl2^Δ/Δ; Vangl1^gt/gt embryo, randomly positioned cilia around the center of the PNC cells resulted in turbulent nodal flow.

Our data show that PCP is not required for ciliogenesis in mouse (Fig. 1b and supplementary Fig. 6c–f). In addition, Vangl1 and Vangl2 were not colocalized with cilia in any of the cells we had examined including the PNC cells, sensory hair cells, tracheal epithelial cells and floor plate cells (Supplementary Figs. 2a, d and 6a, b). In the Vangl2^Δ/Δ; Vangl1^gt/gt embryos at E18.5, there were no defects in ciliogenesis or docking of the basal bodies along the apical-basal axis of the multiciliated tracheal epithelial cells (Supplementary Fig. 6c, d). Normal Hedgehog signaling in the Vangl2^Δ/Δ; Vangl1^gt/gt embryos and cells provides further evidence that primary cilia are functionally normal (Supplementary Figs. 6g, h and 7).

To promote a robust break in symmetry, vertebrates have evolved mechanisms of early asymmetry initiation followed by asymmetry amplification and stabilization²². As PCP is a major conserved molecular mechanism by which directional cellular and tissue behaviors are controlled through reorganizing cytoskeleton, it could lead to both cilia-dependent and cilia-independent regulation of L-R asymmetry by controlling asymmetrical ciliary localization in ciliated PNC or cell movement around the node in chick which lacks ciliated PNC^5,6,23. Involvement of PCP in regulating L-R asymmetry has been implicated before since mutations in Inversin (Inv), Bicaudal C and Seahorse lead to L-R asymmetry defects upstream of Nodal expression and these genes can interact with a core PCP gene Dishevelled (Dvl)^24–26. Dvl has recently been found to regulate ciliary localization in the PNC and implicated to regulate L-R patterning²⁷. However, it is unlikely that Inv acts mainly through PCP since Inv, but not Vangl1/2, is localized to PNC cilia and the phenotypes of the Inv and Vangl1/2 mutants are different in many aspects. As abnormal L-R patterning is also associated with multiple congenital diseases²⁸, the availability of Vangl1 and Vangl2 conditional double mutants has opened a door to investigate the roles of PCP in many fundamental aspects of vertebrate development and diseases.

Note added in proof: While this work was in press, two independent studies^29,30 were published, both also found that planar cell polarity mediated by Vangl2 is required for orienting and positioning motile cilia.

Methods Summary

Scanning electron microscopic analysis

Embryos were dissected at E7.75 and fixed in 2% glutaraldehyde, 4% paraformaldehyde, 0.1M sodium cacodylate, pH 7.4 at 4 °C overnight. Standard procedures were followed to prepare samples. Samples were examined by a Hitachi S-570 scanning electron microscope.

Analysis of calcium signaling

Embryos were dissected at E7.75–E8.0 and analyzed according to the established procedure²¹.

Video Microscopy

Embryos were dissected at E7.75–E8.0 in DMEM with 20% FBS. The node was placed facing down in a chamber with a cover slip on the bottom. Fluospheres (Invitrogen) were added into the medium to visualize the direction of nodal flow. Wide-field images were collected using a Personal DeltaVision system (Applied Precision Inc, Issaquah, WA, USA) mounted on an inverted Olympus IX71 microscope with an oil immersion PlanApo N 60×/1.42 objective lens. Images were acquired using a CoolSNAP ES2 camera with 2×2 binning and 128 pixels ×128 pixels imaging field.

Methods

Generation of and genotyping mouse lines

Generation of the floxed Vangl2 allele is described in Supplementary Fig. 1e. Vangl2^Δ was generated by deleting exon 4 and 5 using the Sox2-Cre line³¹. Vangl1^gt allele was generated using a Baygenomic gene trap ES cell line (XL802) (Supplementary Fig. 1a). Genotyping was done by PCR using genomic DNA prepared from yolk sac or tail. Vangl2P1 (5′-CCATCGATGCACATGTGGTATC-3′) and Vangl2P2 (5′-GACACTGTCCTCCATGTCCTTG-3′) were used to genotype the Vangl2^f allele and Vangl2⁺ allele. Vangl2P1 and Vangl2P3 (5′-GCTATGACCAGCTACCTGAAGTC-3′) were used to genotype the Vangl2^Δ allele. Vangl1P1 (5′-CCCTGGCTTTCTTGTTGGTC-3′) and Vangl1P3 (5′-GGGCTGGCTCTTGGAGTCAT-3′) were used to genotype the Vangl1⁺ allele. Vangl1P1 and Vangl1P2 (5′-CCTGGGGTTCGTGTCCTACA-3′) were used to genotype the Vangl1^gt allele.

Whole mount in situ hybridization and LacZ staining

Whole mount in situ hybridization and LacZ staining were performed as described previously³². Embryos were cryosectioned at 20 µm.

Immunofluorescence and antibodies

Embryos were fixed in 4% paraformaldehyde for 30 min at 4 °C, washed in PBS and incubated overnight in 30% sucrose/PBS. Embryos were embedded in OCT compound (Tissue-Tek) and sectioned at 14 µm. Sections were permeabilized with 0.5% Triton X-100 in PBS and blocked in 3% bovine serum albumin, 0.1% Triton X-100 in PBS for 1 hour. Sections were incubated with primary antibodies at 4°C overnight. Secondary antibodies were applied at room temperature for 1 hour. Primary antibodies used in immunofluorescence include mouse anti-Pax6, Pax7, Nkx2.2, Lim1/2, Nkx6.1, Hnf3β and Islet-1 (Developmental Studies Hybridoma Bank, 1:50), rabbit anti-Olig2 (Abcam, 1:100), mouse anti-acetylated-tubulin (Sigma, 1:5,000), goat anti-γ-tubulin (Santa Cruz, 1:200), goat anti-Vangl2 (Santa Cruz, 1:100), rabbit anti-Vangl2 (gift of Dr. Mathew Kelley’s lab, 1:50) and goat anti-E-Cadherin (R&D, 1:500). Vangl1 antibodies were generated against Vangl1 (1-110 aa) polypeptides fused to glutathione S-transferase (GST) in rabbit. Anti-Vangl1 antibodies were purified using the immunizing peptide coupled to Affigel (Bio-Rad). Alexa Fluor 488 and 568 of donkey anti-mouse, rabbit and goat secondary antibodies were from Invitrogen. Confocal images were acquired using a Zeiss LSM 510 NLO microscope. Excitation wavelengths of 488, 561 and 740 nm were used for detection of secondary antibodies conjugated by Alexa Fluor 488, 568 and DAPI respectively.

Whole-mount Immunostaining of cochlea or early E6 – E8.5 embryos

Dissected cochlea or embryos were fixed in 4% PFA for 30 min at 4 °C. Cochlea or embryos were permeabilized with 0.5% Triton X-100 in PBS for 5 min and blocked in PBS with 3% BSA, 0.1% Triton X-100 at room temperature for 1 hour, followed by overnight incubation with primary antibodies at 4°C. Cochlea or embryos were washed in PBS with 0.1% Triton X-100 at room temperature for 1 h, followed by incubation with fluorescent secondary antibodies and Alexa Fluor-594-conjugated phalloidin (Invitrogen 1:200) at room temperature for 1 hour. The actin-based stereocilium bundles were shown by phalloidin staining whereas the microtubule-based kinocilia or PNC cilia were shown by staining using the mouse anti-acetylated-tubulin antibody (Sigma, 1:5,000). After washed with 0.1% Triton X-100 in PBS at room temperature for 1 hr, cochlea or the PNC region were flat mounted. Confocal images were acquired using a Zeiss LSM510 confocal microscope.

Immunolot Analysis

Embryonic tissues or cells were lysed in RIPA buffer. Cell lysates were analyzed by Immunoblot. Primary antibodies used include goat anti-Vangl2 (1:2000), rabbit anti-Vangl1(1:1000), rabbit anti-Gli3 (gift from Dr. Susan Mackem, 1:5000).

Analysis of calcium signaling

Embryos were dissected at E7.75–E8.0 and incubated with the fluorescent calcium indicator Fluo-4 (Invitrogen) at a concentration of 10 µM in DMEM with 20% FBS for 20 minutes at 37 °C. Embryos were then washed for 3 times in DMEM with 20% FBS and placed node facing up in a silicon chamber under a cover slip. Embryos were incubated at 37 °C for 20 minutes before confocal imaging. Confocal images were acquired using a Zeiss LSM510 confocal microscope. Excitation wavelength of 488 was used to detect Fluo-4. Embryos were genotyped after imaging. Data were analyzed with Image-Pro 6.3 software. High intensity was shown in red and low intensity in blue. Intensity scales of the calcium signal were plotted from left to right side of the node.

Video Microscopy

Embryos were dissected at E7.75–E8.0 in DMEM with 20% FBS. The node was placed facing down in a chamber with a cover slip bottom. For the nodal flow assay, Fluospheres (Invitrogen) were added into the medium to visualize the direction of nodal flow. Embryos were genotyped after imaging. Wide-field images were collected using a Personal DeltaVision system (Applied Precision Inc, Issaquah, WA, USA) mounted on an inverted Olympus IX71 microscope with an oil immersion PlanApo N 60×/1.42 objective lens. All images were acquired using a CoolSNAP ES2 camera with 2×2 binning and 128 pixels × 128 pixels imaging field. Videos were created using Applied Precision’s SoftWoRx software package version 4.0.0. Individual bead movement was traced and shown as a line with different color. The mean cilia beating frequency was calculated from at least 10 nodal cilia of embryos with indicated genotypes at E8.0. Data are means ±SD. The p value is calculated according to the student’s t-test.