2024 年 21 巻 3 号 論文ID: e210018
Visceral organs in vertebrates are arranged with left-right asymmetry; for example, the heart is located on the left side of the body. Cilia at the node of mouse early embryos play an essential role in determining this left-right asymmetry. Using information from the anteroposterior axis, motile cilia at the central region of the node generate leftward nodal flow. Immotile cilia at the periphery of the node mechanically sense the direction of leftward nodal flow in a manner dependent on the polarized localization of Pkd2, which is localized on the dorsal side of cilia. Therefore, only left-side cilia are activated by leftward nodal flow. This activation results in frequent calcium transients in the cilia via the Pkd2 channel, which leads to the degradation of Dand5 mRNA only at the left-side crown-cells. This process is the mechanism of initial determination of the left-side-specific signal. In this review, we provide an overview of initial left-right symmetry breaking that occurs at the node, focusing mainly on a recent biophysical study that revealed the function of nodal immotile cilia using advanced microscopic techniques, such as optical tweezers and super-resolution microscopy.
This review provides an overview of the mechanism of initial breaking of left-right symmetry, focusing mainly on a recent biophysical study. By combining advanced microscopic techniques, such as optical tweezers and super-resolution microscopy with biological methods, the recent study revealed the function of immotile (primary) cilia in the left-right organizer of the early embryo. Previous notable studies and the recent study provided insight into the elaborate mechanism of left-right determination based on information obtained from the dorsoventral and anteroposterior axes.
Visceral organs in vertebrate organisms are arranged in left-right asymmetry; for example, the heart is located on the left side of the body and the liver is on the right side of the body. This asymmetric arrangement of the visceral organs is defined by three axes: dorsoventral, anteroposterior, and left-right. During mouse development, initiation of the dorsoventral axis begins on embryonic day 5.5 (E5.5), with the anteroposterior axis established at around E6.5 [1]. Using these information, it is known that left-right symmetry is broken at the left-right organizer in the E7.5 embryo [2].
The embryonic node, referred to as the left-right organizer in mice, emerges at the anterior tip of the primitive streak on the midline of the embryo. The cells of the node possess monocilia, known as nodal cilia, consisting of motile cilia that are mainly located at the center of the nodal cavity and immotile cilia mainly in the peripheral region of the nodal cavity [3] (Figure 1A). Motile cilia generate leftward nodal flow and play an important role in establishing left-right asymmetry [4,5]. Subsequently, immotile cilia, referred to as the “cell’s antenna,” sense this flow. The direction of nodal flow determines the left-right asymmetry. When applying rightward artificial flow to the embryonic node using a custom-made chamber, the embryo shows reversal of left-right situs [6]. A recent study revealed that immotile cilia sense nodal flow and activate a left-side-specific signaling pathway via the cation channel Pkd2, which is localized in immotile cilia [7] (Figure 1A). Most recently, the authors found that nodal immotile cilia mechanically sense the direction of nodal flow and activate a left-side-specific signaling pathway [8].
When cilia are activated by nodal flow, intraciliary calcium transients occur via the Pkd2 channel in mouse [9]. A similar tendency in intraciliary calcium via the Pkd2 channel has been observed in zebrafish [10]. Intraciliary calcium signaling activates the endoplasmic reticulum (ER) via the inositol 1,4,5-trisphosphate pathway in crown cells, triggering frequent cytoplasmic calcium transients specifically in left-sided cells [9]. In particular, the ER localizes to the apical surface of nodal crown cells [9], which might be important for signal transduction from the cilia to the cell. Subsequently, calcium signaling activates the decay of Dand5 mRNA through the Bicc1-Ccr4 RNA degradation complex [11] (Figure 1A). Suppression of ciliary calcium transients delays Dand5 mRNA degradation, suggesting that ciliary calcium transients are responsible for Dand5 mRNA degradation [9]. Dand5, an antagonist of Nodal, leads to stable left-sided expression of Nodal when Dand5 mRNA is degraded only on the left side. Based on a mathematical model, interconnected feedback loops involving signals from flow (downstream of intraciliary calcium signaling), Dand5, and Wnt serve as a bistable switch for Dand5 expression [12]. This bistable switch is assumed to contribute to the robust establishment of left-right asymmetry. Subsequently, left-right asymmetric signals at the node are transmitted to the lateral plate mesoderm, resulting in the establishment of left side-specific Nodal expression in this region. During this process, Lefty1 is expressed at the midline of the embryo, where it acts as a midline barrier. According to a mathematical model, Nodal, Lefty1, and Lefty2 proteins constitute a self-enhancement and lateral inhibition system, contributing to the robust establishment of left-right asymmetry, particularly the exclusive expression of Nodal on the left side of the lateral plate mesoderm [13]. The initial bias, induced by leftward nodal flow and activated by nodal immotile cilia, robustly establishes left-right asymmetry through at least two feedback systems: the bistable switch and self-enhancement and lateral inhibition system. Based on recent studies, this review introduces a biophysical mechanism involved in the initiation of left-right symmetry breaking at the node using information from the anteroposterior and dorsoventral axes.
Motile cilia positioned at the center of the node exhibit clockwise rotation seen from the tip and are tilted toward the posterior side of the cell (Figure 1B). The cell surface has a no-slip boundary condition; therefore, as the cilia pass close to the cell surface, only weak rightward flow is generated [14]. Conversely, when cilia move away from the cell surface, they generate leftward flow. Hence, clockwise rotation and posterior tilting of nodal motile cilia are crucial for the generation of leftward nodal flow [5,15] (Figure 1B).
Posterior tilting of nodal cilia is generated based on information from the anteroposterior axis. Pit cells possess nodal motile cilia and exhibit planar cell polarity (PCP) along the anteroposterior axis [16] (Figure 1B). The PCP of pit cells is established through a concentration gradient of Wnt5 and its antagonist Sfrps, utilizing information from the anteroposterior axis [17] (Figure 1B). Utilizing this PCP along the anteroposterior axis, the basal body (base of cilia) and associated radial microtubule network (aster) are displaced toward the posterior side of pit cells by microtubules, kinesins, and actomyosin [18]. As the apical surface of pit cells exhibits a dome shape, posterior displacement of the basal body results in posterior tilting of the cilia [5,15]. Posterior tilting of the motile cilia is induced utilizing information from the anteroposterior axis through the PCP, resulting in the generation of leftward nodal flow. This mechanism generates leftward nodal flow using information from the anteroposterior axis.
How do nodal immotile cilia sense leftward nodal flow and activate the left-side-specific signal cascade? Whether immotile cilia function as mechanosensors or chemosensors remains controversial [19,20]. A previous study demonstrated that crown cell cilia, primarily comprising immotile cilia, function as flow sensors via cilia-localized Pkd2 channel, activating left-side-specific signals, such as Pitx2 [7]. However, activation of a single cilium by flow stimuli has not been observed. Cilia were not activated by an artificial flow from a glass pipette (note that a later study confirmed that the experimental medium conditions for mouse embryo culture were not suitable for evaluating ciliary function in the node) [9,19]. I and my colleagues demonstrated that nodal flow applies mechanical deformation to cilia, and artificial mechanical stimuli from optical tweezers are sufficient to activate immotile cilia [8].
We initially attempted to observe the flow-dependent passive motion of nodal immotile cilia. Mouse nodal immotile cilia expressing mNeonGreen were visualized using spinning disc confocal microscopy with deconvolution calculations (Figure 2A). This 3D high-resolution live imaging system enabled observation of the small bending motions of the cilium (Figure 2B). The heavy chain of dynein, which is responsible for ciliary beating, can be cleaved by UV radiation. An optical pathway that included an iris was designed to selectively irradiate pit cells (center of the node). Upon UV irradiation of the pit cells, all motile cilia stopped moving and leftward nodal flow was halted. We compared the 3D angles of the same cilia before (with nodal flow) and after (without nodal flow) UV irradiation using ellipsoidal fitting (Figure 2B). Our findings revealed that left-sided nodal immotile cilia exhibited ventral bending, whereas right-sided cilia exhibited dorsal bending in response to nodal flow (Figure 2B). The asymmetric bending of cilia can be explained by considering the geometry of the node. It is deduced that the nodal flow exerts a force perpendicular to the cilia, as the flow runs along the surface of the node, where immotile cilia protrude (Figure 2B; see numerical simulations in Fig. S5 of Ref [8]).
We evaluated whether this mechanical bending activated immotile cilia. We utilized optical tweezers [21,22], which can trap a small particle in medium and manipulate it using changes in the momentum of light rays, to manipulate a single cilium. A 3D manipulation system employing optical tweezers for single cilia was developed [23,24]. Using this system, we applied mechanical stimuli to nodal immotile cilia and measured downstream signaling cascades, including the calcium response (Figure 2C) and Dand5 mRNA degradation (Figure 2D).
In mouse nodal immotile cilia, left-sided cilia and their cytoplasm-specific calcium responses depend on leftward nodal flow. To assess whether these calcium responses could be triggered by artificial mechanical stimuli, we monitored calcium signals using transgenic mice expressing the calcium indicator GCaMP6 in the cilia and cytoplasm (Figure 2C). We utilized an iv/iv mutant mouse [25], which lacks nodal flow; thus, the cilia were not exposed to stimuli under this condition. The frequency of calcium transients in immotile cilia, mediated by the Pkd2 channel essential for left-right (L-R) determination, was significantly increased by mechanical stimuli applied to a single cilium (Figure 2C). Note that Pkd1l1 [26], as well as Pkd2, is involved in mechanosensation in nodal cilia, as it has been reported that Pkd2 and Pkd1 mediate mechanosensation in renal cilia [27], where they form a heteromeric complex that functions as a channel [28].
We evaluated whether Dand5 mRNA degradation was triggered by artificial mechanical stimuli. Dand5 mRNA degradation occurs only on the downstream side of nodal flow; ciliary calcium transients via the Pkd2 channel are necessary for this process. To monitor Dand5 mRNA degradation, we developed the “whole-cell FRAP” method [29], which was combined with a Dand5 3′UTR-dsVenus transgenic mouse [11]. Using this system with optical tweezers, we found that mRNA degradation occurred only in cells in which mechanical stimuli were applied to the cilium, whereas mRNA degradation did not occur when stimuli were applied to the cell body (Figure 2D). Further experiments revealed that mRNA degradation occurred via a pathway involving the Pkd2 channel (Figure 2D).
These results indicate that mechanical rather than chemical stimuli from nodal flow are sufficient to activate nodal immotile cilia and initiate downstream signals associated with L-R determination [8]. Recent studies have revealed that mechanosensing in nodal cilia is preserved in zebrafish as well as mice [30]. Therefore, these functions may be conserved in mammals and osteichthyes. However, cell migration is involved in L-R determination in chicks [31]. Thus, another mechanism not involving cilia may underlie L-R determination in chicks and reptiles [32].
We also examined the cause of activation of only left-sided cilia, despite both cilia undergoing mechanical bending induced by nodal flow. We revealed that nodal immotile cilia sense the bending direction; thus, only left-sided cilia were activated by leftward nodal flow.
First, we evaluated the ciliary localization of the Pkd2 channel, a transient receptor potential channel known to be involved in mechanosensory transduction [33], using stimulated emission depletion microscopy (Figure 3A). We found that Pkd2 was significantly localized toward the dorsal side of the immotile cilia along the midline-lateral direction (Figure 3A). Next, we evaluated membrane strain using 3D high-resolution images of cilia with and without nodal flow. Based on the intrinsic ciliary shape obtained from the ciliary shape without nodal flow, the ciliary membrane was modeled using approximately 2000 triangular elements. We then measured the deformation of each element in the image in the presence of nodal flow and calculated the membrane strain resulting from this deformation (Figure 3B). The results showed that membrane tension on the dorsal side of the L-side cilia increased, allowing the dorsally localized Pkd2 channel to be activated (Figure 3C). In contrast, membrane tension on the dorsal side of the right-side cilia decreased; therefore, right-side cilia were not activated by leftward nodal flow (Figure 3C). Particularly, the observed membrane tension of 1.6 mN/m on the dorsal side of left-side cilia was sufficient to activate a typical mechanosensing channel (Figure 3B). To evaluate how cilia sense the bending direction, we applied dorsal and ventral bending sequentially to the same cilium. We found that the cilia responded only to ventral bending with an increase in calcium frequency (Figure 3D).
Most recently, we found that BMP4 is responsible for generating an asymmetric distribution of Pkd2 [34]. Through imaging-based screening, we observed that excess BMP4 disrupted the correct distribution pattern of Pkd2, inducing a more polarized distribution toward the dorsal side of the nodal immotile cilia. Model calculations suggested that the BMP4 concentration gradient influenced the generation of the asymmetric Pkd2 distribution. Furthermore, excess BMP4 disrupted the calcium response triggered by mechanical stimuli to the cilia in crown cells. This result suggests a relationship between the correctly organized asymmetric distribution of Pkd2 in cilia and sensing of the bending direction of cilia [34].
We predicted the biophysical mechanism for generating L-R asymmetric signal cascades by utilizing information from the dorsal-ventral and anterior-posterior axes. Specifically, leftward nodal flow was generated using information from the anterior-posterior axis (Figure 1B), and an asymmetric Pkd2 channel distribution was established using information from the dorsal-ventral axis (Figure 3A). This model represents an elaborate biological mechanism for determining L-R asymmetry using only the anteroposterior and dorsoventral axes (Figure 3C). However, the asymmetric distribution of Pkd2 in immotile cilia has only been evaluated in mouse embryos, and the universality of this mechanism remains an open question for future research. Further investigation is needed to clarify the precise function of Pkd2, particularly regarding how BMP4 regulates Pkd2 distribution in cilia and how the Pkd1l1-Pkd2 heteromeric complex senses mechanical cues in cilia. The detailed mechanisms underlying the mechanosensation of Pkd2 still leave some questions unanswered [33]. Although the causal relationship between Pkd1l1 localization and initial L-R determination requires further evaluation, the left-side bias of Pkd1l1 localization in the node is interesting [35]. Notably, L-R asymmetric ciliary calcium transients are detected approximately three hours earlier (late head fold stage) than the observation of Pkd1l1 localization bias [9]. Regarding Dand5 mRNA degradation, based on the findings of Mizuno et al. [9], it is reasonable to deduce that an increase in cilia-depended cytoplasmic calcium frequency triggers the activation of mRNA degradation. Recent studies have reported that Bicc1 [11], ANKS3, and ANKS6 regulate mRNA degradation via the Ccr4-Not complex [36,37]; however, the detailed mechanism remains to be elucidated.
The author declare that he has no conflict of interest.
TAK conceptualized the topic of this article and wrote the manuscript.
The data generated or analyzed during the current study are available from the corresponding author upon reasonable request.
I thank H. Hamada (BDR, Riken), Y. Okada (The University of Tokyo and BDR, Riken), T. Ishikawa and T. Omori (Tohoku University), A. H. Iwane (BDR, Riken), T. Nishizaka (Gakushuin University), and members of the Hamada laboratory for their help in performing experiments and writing articles that are the basis of this review article.
This study was supported by the FOREST Program (Grant No. JPMJFR224N) of the Japan Science and Technology Agency (JST), Grant-in-Aid (grant nos. 21K15096, 24K18107, and 24H01270) from the Japan Society for the Promotion of Science (JSPS), RIKEN Special Postdoctoral Researcher Program, University of Tokyo Excellent Young Researcher Program, and Ultrastructure Research Fund from the Japanese Society of Microscopy.
