Establishing the left and right body plan during early embryonic development is a fundamental process, which hinges significantly on the presence of a specific structure known as the Lefty1+ midline barrier. This midline barrier plays a crucial role in guiding the development of various organs, each exhibiting left-right (LR) asymmetries. However, the brain stands out as it possesses the only documented organ-specific midline barrier, which tightly regulates the commissural axons in both the brain and spinal cord. This regulation is achieved through a variety of midline-localized guidance and repulsion signals, including fibroblast growth factors (FGFs), SLIT/ROBO signaling pathways, EFNB3, heparan sulfate proteoglycans, and the Rac-specific GTPase-activating protein -chimaerin (Cavalcante et al., 2002; Kullander et al., 2001; Brose et al., 1999; Kidd et al., 1998; Erskine et al., 2000; Neugebauer and Yost, 2014; Katori et al., 2017).

Moreover, the developing intestine similarly requires a mechanism to separate left and right cells, but intriguingly, it employs a distinct strategyutilizing an atypical basement membrane located at the midline. This basement membrane serves to segregate cells derived from the left and right coelomic epithelia, as these cells do not intermingle at the midline (Carmona et al., 2013). This separation is not just structural; it also appears to prevent the mixing of diffusible signals, which could significantly affect development. The extent of signal diffusion is highly dependent on the tissue context (Mller et al., 2013), with some morphogens, such as Nodal, capable of inducing effects at distances surpassing 200 m (Mller and Schier, 2011). Given that the HH19 DM (dorsal mesoderm) measures only about 150 m across, it raises questions regarding the potential for diffusible signals like bone morphogenetic proteins (BMPs) and TGF to traverse between the left and right DM.

To investigate this, we hypothesized that the midline might necessitate a barrier to effectively segregate these signals. Supporting this hypothesis, our research demonstrated that the midline restricts the diffusion of dextran migrating from right to left, indicating a similar obstruction to endogenous diffusible signals. Notably, dextran, with a modest molecular weight of 3 kDa, was unable to cross the midline, suggesting that proteins of average weightlike CXCL12 (10 kDa) and BMP4 (34 kDa)would also be impeded. Earlier findings indicated that even the relatively small drug AMD3100 (502.78 Da) could not migrate across the DM midline (Mahadevan et al., 2014). In our recent studies, we utilized a BODIPY-labeled variant of AMD3100 (Poty et al., 2015), which exhibited diffusion across the midline only when the basement membrane was intact. However, the absence of this membrane allowed BODIPY to diffuse freely across the entire DM width. The midlines integrity could not be compromised via Ntn4 overexpression, allowing for the potential exploration of selective methods to dismantle this basement membrane for a deeper understanding of its role in gut laterality.

This research contributes a new dimension to our understanding of the form and function of basement membranes. Typically, basement membranes are crucial barriers in both embryonic and adult tissues, often appearing as a single layer that undergirds polarized epithelial or endothelial cell layers, which are prominent in locations such as the intestines, surrounding blood vessels, or encasing muscle cells and adipocytes (Yurchenco, 2011). Notably, mutations leading to the absence of genes encoding basement membrane components can result in severe embryonic lethality and various postnatal pathologies (Bader et al., 2005; Miner et al., 2004; Smyth et al., 1999; Gatseva et al., 2019; Pozzi et al., 2017; Yao, 2017). However, the specific function of basement membranes in establishing LR asymmetry has not been previously documented.

The unique double membrane structure of the basement membrane we identified raises intriguing questions concerning its formation. Our findings indicated that mesenchymal cells of the DM do not produce this midline structure (Figure 4), nor is the notochord alone sufficient for its synthesis (Figure 4). Instead, we suspect that the endoderm may be responsible. When we electroporated the endoderm with GFP (green fluorescent protein), GFP-positive cells were not evident in the DM at later developmental stages, remaining confined to the endoderm (Figure 4FH). This observation suggests that the midline does not originate from epithelial-mesenchymal transition (EMT) of basement membrane-carrying endodermal cells. Additionally, endodermal cell death was ruled out as a possible explanation since no significant cell death was detected at the midline during critical developmental stages (data not shown).

Furthermore, the midline does not form in the same manner as other known double basement membranes, which typically arise from the convergence of the basal surfaces of two distinct tissues (Keeley and Sherwood, 2019; Pastor-Pareja, 2020). Classical examples of this include the kidney glomerulus, where epithelial podocytes interface with endothelial cells (Pastor-Pareja, 2020; Miner, 2012; Naylor et al., 2021), and the blood-brain barrier, which exists between endothelial cells, pericytes, and astrocytes (Keeley and Sherwood, 2019; Daneman and Prat, 2015). In stark contrast, the midline barrier of the DM features the apical surfaces of endodermal cells facing one another.

We propose that the midline is formed as the basement membrane remains in place while the endoderm descends ventrally during normal developmental processes, akin to a scar from the position of the endoderm in earlier stages (Figure 7). During initial development, the notochord is situated within the endoderm, covered by a shared basement membrane. It is only after a certain developmental stage that these two structures become distinctly separated by a complete basement membrane, indicating a strong connection between them (Fausett et al., 2014; Jurand, 1974). The medial migration of the aortae and coelomic epithelia may generate forces that push the notochord and endoderm apart, resulting in the remnant basement membrane at the DM midline. The resistance of both the endoderm and midline basement membranes to disruption by NTN4 further supports the idea that the endoderm is responsible for generating both types of membranes (Figure 7). This concept aligns with our current understanding that the basement membrane beneath the gut endoderm remains static while intestinal epithelial cells migrate from proliferative intestinal crypts to the tips of the villi over several days in adult development (Trier et al., 1990).

To explore this hypothesis, one potential approach could involve tracing the endodermal basement membrane through electroporation of a tagged basement membrane component, similar to recent techniques employed in Caenorhabditis elegans (Walser et al., 2017; Naegeli et al., 2017; Matsuo et al., 2019; Jayadev et al., 2019; Keeley et al., 2020; Jayadev et al., 2022) and Drosophila (Morin et al., 2001; Ramos-Lewis et al., 2018). Such methodologies are just beginning to be adapted for studies in mice (Tomer et al., 2022; Morgner et al., 2023) and have yet to be applied to chick embryos.

Figure 7 Download asset Open asset Model of endoderm descending hypothesis for midline formation. We hypothesize that as the endoderm moves ventrally and the distance between the notochord and endoderm grows, basement membrane from the endoderm may be left behind. This can be compared to a zipper where each side is the basement membrane underlying the endoderm, and when the zipper pull (tip of endoderm) moves downward, the basement membrane behind it pulls closer together.

The resistance of both the endodermal basement membrane and the midline against NTN4 disruption is especially noteworthy, as such resistance has not been documented previously. This observation implies the existence of an unknown factor or modification that protects or stabilizes the laminin network from NTN4 binding. Exploring the mechanisms that enhance the stability of basement membranes against disruptions could prove vital for future developmental studies and research related to diseases like cancer.

Interestingly, while Lefty1 expression is absent at the midline when this basement membrane is robust, it remains uncertain whether the expression of Lefty1 is essential for the formation or optimal function of this structure. Future research could involve the use of siRNA to inhibit Lefty1 expression at early developmental stages, followed by assessments of the midline structures integrity at HH19. Yet, similar to the challenges faced when disrupting global Nodal expression, differentiating between the direct effects of Lefty1 loss on subsequent midline cellular structures and the indirect consequences of free Nodal diffusion throughout early lateral mesoderm poses significant difficulties. Thus, a deeper investigation into the interplay between these two aspects of the midline barrier remains vital.

The midline barrier is distinctly characterized by its rapid degradation. We speculate that the breakdown of the DM midline barrier may be partially attributed to the stretching of the basement membrane as the notochord and endoderm become progressively separated due to embryo elongation (Figure 3J). Considering that the DM itself does not contribute basement membrane to the midline (Figure 4C), and the midline length increases swiftly (Figure 3J), it is plausible that the midline could be stretched until it reaches its ultimate tensile strength (reported to be between 0.5 and 3.8 MPa in other naturally occurring basement membranes) (Jain et al., 2022), resulting in structural failure. This suggests that the breakdown of the midline might be a passive outcome of embryo growth.

Conversely, we also consider the potential for an active breakdown mechanism for the midline. The turnover of stable basement membranes generally occurs over weeks (Trier et al., 1990; Decaris et al., 2014), yet the midline barrier appears to degrade within a mere 1224 hours. In other scenarios, basement membrane destruction can be extensive, such as through the secretion of matrix metalloproteinases during cancer metastasis (Miyoshi et al., 2004; Miyoshi et al., 2005), or more localized, as seen with immune or cancer cells via invadopodia prior to metastasis (Sekiguchi and Yamada, 2018; Santiago-Medina et al., 2015). Localized basement membrane breakdown is also essential for numerous developmental processes, including the formation of the mouth in deuterostome embryos (Dickinson and Sive, 2006). In this context, a basement membrane that seals off the digestive tract from the external environment specifically disintegrates to facilitate the emergence of the early mouth cavity (Dickinson and Sive, 2006). Additionally, a basement membrane divides the two halves of the embryonic brain and must be disrupted at the corpus callosum site to allow neuronal crossing for inter-hemisphere communication in the cerebrum (Hakanen and Salminen, 2015; Gobius et al., 2016). Such localized basement membrane dissolution is also critical for the fusion of the optic cup, where a contact-dependent breakdown occurs at the optic fissure (Torres et al., 1996; Barbieri et al., 2002). Congenital defects like coloboma, characterized by missing tissue within the eye, arise when optic fusion is halted (Patel and Sowden, 2019; ALSomiry et al., 2019). Importantly, in all these cases including the midline, the site of basement membrane breakdown is highly specific, with adjacent membranes remaining unaffected (Figure 3H). Collectively, these insights support the notion that the breakdown of the basement membrane in the DM may not simply be a passive mechanism but could be critical for the embryo in subsequent gut development or vascular patterning. As of now, we have not identified any matrix protease specific to the midline of the DM. Given that localized mechanisms of basement membrane breakdown are not well characterized in any of these contexts, there exists substantial potential for future exploration.

The DM midline may also have developmental implications extending beyond the intestine. It has been well-established that signals from axial structures within the embryo are pivotal for establishing LR asymmetry; however, pinpointing the exact midline structure or the midline signal responsible has proven challenging. For instance, proper development of the heart (Lohr et al., 1997; Chen et al., 1997), lungs (Arraf et al., 2016), and kidneys (James and Schultheiss, 2003) heavily depends on dorsal midline structures. Abnormalities resulting from the loss of these midline structures, such as the notochord, can lead to conditions like horseshoe kidney, wherein the kidneys remain closely aligned at the midline, fusing at their posterior ends (Natsis et al., 2014). This condition arises due to inadequate Shh signaling from the notochord, yet the downstream midline barrier that remains following this signaling has yet to be elucidated (Tripathi et al., 2010).

Moreover, the DM midline barrier could influence vascular patterning, including the fusion of the aorta. Initially, the aorta forms as two parallel tubes separated by an avascular zone (the midline area) (Garriock et al., 2010). Over time, these tubes fuse into a single vessel following an anterior-to-posterior wave until reaching the vitelline arteries (Figure 3figure supplement 3; Garriock et al., 2010). The timing of this fusion correlates with the fragmentation and disappearance of the midline barrier, indicating a potential relationship between these two processes. The correct timing of fusion relies on a carefully balanced interplay of factors including VEGF (Jadon et al., 2023), SHH (Vokes et al., 2004), and a wave of downregulation of BMP-inhibitory genes Chordin and Noggin originating from the notochord (Garriock et al., 2010; Reese et al., 2004; Sato, 2013). The precise mechanisms facilitating dorsal aorta fusion remain unclear, though evidence suggests that VEGF signaling may play a role in repositioning VE-cadherin away from cell-cell junctions (Jadon et al., 2023). This relocalization could be crucial for remodeling the aortic endothelium during fusion and potentially tied to the disruption of the DM midline barrier.

Midline structures are indispensable for the accurate development of laterality. While the notochord is undoubtedly involvedparticularly as a source of modulators for BMP and Hedgehog signalingit seems that the narrative does not end there. In several contexts, the specific midline barrier that exists downstream from notochord signals has yet to be defined. It is conceivable that the midline basement membrane of the DM holds significant importance, either for separating left and right signals or perhaps for binding signals (Pozzi et al., 2017) from the notochord to act as a buffer between the two sides. Additionally, the midline barrier may influence the rheology of the DM. Microindentation assessments conducted at HH21 reveal that the condensed left DM exhibits a significantly greater stiffness compared to the expanded right DM, indicating that the regulation of this stiffness is crucial for proper gut tilting (Sanketi et al., 2022). This raises the possibility that the midline barrier serves to segregate the stiffness-influencing components from each side, (for instance, covalently modified hyaluronic acid on the right [Sivakumar et al., 2018] and N-cadherin on the left [Kurpios et al., 2008]) and provides a structural wall for the right side to push against, facilitating the leftward swing of the gut tube.

It's noteworthy that mesenchymal cells from opposing sides do not traverse the midline to the opposite DM side. Additionally, during gut artery formation, a select group of vascular endothelial cells migrates from right to left, maintaining proximity to the dorsal tip of the endoderm without crossing the midline (Mahadevan et al., 2014). Moreover, while the midline basement membrane remains intact, N-cadherin expression is symmetric across the DM; however, it transitions to an asymmetric left-specific expression following midline disintegration (Kurpios et al., 2008). This suggests a shift in cellular segregation mechanisms within the DMwhile the midline is intact, the double basement membrane structure appears sufficient to maintain separation of left and right cells. Once the midline is compromised, a new cellular separation mechanism must be established to uphold asymmetric compartments, with N-cadherin playing a functional role in this context.

Looping patterns of the midgut are consistent among individuals yet vary across different species (Savin et al., 2011). While differences in growth rates between the gut tube and the DM have been identified as primary contributors to distinct gut looping patterns (Savin et al., 2011), mesentery-specific modifications, such as changes in the kinetics of asymmetry in the DM, may also influence these looping patterns, potentially adapting them to dietary needs and ecological niches. The presence, permeability, and timing of degradation of the midline basement membrane could offer evolution an additional mechanism for fine-tuning gut looping patterns among species. Our observations indicate this basement membrane is transiently present in the DM of veiled chameleon embryos from shortly after the 7-somite stage to the 29-somite stage (Figure 3figure supplement 4; Diaz et al., 2019), suggesting that this structure is conserved at least in reptiles and birds, with intriguing variations in degradation timing across species. In mammals, previous studies have documented midline laminin deposition within the mouse gut situated between the separating notochord and endoderm, further implying the conservation of a midline basement membrane among amniotes (Li et al., 2007; Hajduk et al., 2012).

In conclusion, we have identified a novel midline barrier within the gut mesentery characterized by an atypical double basement membrane that delineates the left and right sides while restricting the movement of diffusible signals and cells. This discovery occurs during a developmental stage when Lefty1 is no longer expressed at the midline. The DM midline offers an exciting opportunity to explore fundamental mechanisms of basement membrane formation and breakdown during vertebrate embryonic development, with significant implications for cancer metastasis research. We propose that this midline represents a distinctive strategy for the critical separation of left and right signals and cells, which is vital for establishing and maintaining LR asymmetry for healthy gut development.