Gastrulation and axis formation: organizer signaling and Hox gene patterning

Intuition Beginner

After cleavage produces a hollow ball of cells (the blastula), the embryo faces its most dramatic transformation: gastrulation. During gastrulation, cells rearrange themselves through coordinated movements to form three distinct layers. These three germ layers are the foundations from which every organ in the body is built.

The three germ layers are:

• Ectoderm (outer layer): becomes skin, hair, nails, and the entire nervous system (brain, spinal cord, nerves).
• Mesoderm (middle layer): becomes muscle, bone, cartilage, blood, the heart, blood vessels, and kidneys.
• Endoderm (inner layer): becomes the lining of the gut, lungs, liver, pancreas, and other internal organs.

Gastrulation also establishes the body's main axes: head-to-tail (anterior-posterior), back-to-belly (dorsal-ventral), and left-to-right. Once the axes are in place, each body segment needs to know what structure to become. This is the job of Hox genes -- a set of genes that act like a zip code system. Each Hox gene is expressed in a specific region along the head-to-tail axis and tells the cells in that region what to build: "you are the thorax, make ribs" or "you are the lumbar region, make vertebrae but no ribs."

In 1924, Hans Spemann and Hilde Mangold discovered that a small group of cells at the dorsal lip of the blastopore in newt embryos could organize the formation of an entire secondary body axis when transplanted. This region became known as the Spemann organizer -- a signalling centre that tells neighbouring cells where they are and what they should become.

Visual Beginner

Germ layer	Major derivatives
Ectoderm	Epidermis, hair, nails, nervous system (brain, spinal cord, peripheral nerves), sensory epithelia, tooth enamel, adrenal medulla, anterior pituitary
Mesoderm	Skeletal muscle, cardiac muscle, smooth muscle, bone, cartilage, connective tissue, blood, blood vessels, heart, kidney, gonads, dermis, notochord
Endoderm	Gut lining (epithelium), liver, pancreas, lungs (epithelial lining), thyroid, parathyroid, thymus, bladder lining, auditory tube

Worked example Beginner

The Spemann-Mangold experiment (1924) is the classic demonstration of an organizing centre in development:

1. Setup. Working with newt embryos (Triturus), Spemann and Mangold identified a small pigmented region on the dorsal side of the blastopore lip in early gastrulae.

2. Transplant. They excised the dorsal lip from a donor embryo (pigmented, dark cells) and grafted it onto the ventral side of a host embryo (lighter cells).

3. Result. The host embryo developed not one but two body axes. A second neural tube, notochord, and somites formed on the ventral side, producing a twinned embryo. Cell lineage tracing showed that the secondary axis was composed primarily of host cells, not donor cells -- the grafted organizer had induced the surrounding host tissue to reorganize into a new axis.

4. Conclusion. The dorsal lip of the blastopore is an organizer: a signalling centre that can instruct neighbouring cells to change their fate and adopt new positional identities. The organizer does not build the secondary axis itself; it recruits and reprogrammes host tissue.

This experiment established the principle of embryonic induction -- one group of cells can instruct another group to adopt a particular fate through chemical signals.

Check your understanding Beginner

Formal definition Intermediate+

Gastrulation across model organisms

The specific cell movements of gastrulation vary across species, but the outcome is conserved: a triploblastic embryo with three germ layers in their correct spatial relationships.

Sea urchin (Strongylocentrotus purpuratus). Primary mesenchyme cells at the vegetal pole undergo ingression -- they detach from the epithelial wall and migrate into the blastocoel as individual mesenchymal cells. The vegetal plate then invaginates inward, forming the archenteron (primitive gut). Secondary mesenchyme cells at the tip of the archenteron use filopodia to pull the gut toward the animal pole. The sea urchin gastrulome is the most completely mapped gene regulatory network in any developmental system (Davidson et al., 2002).

Drosophila. The fruit fly embryo is a syncytium (no cell membranes between nuclei) during early development. Gastrulation begins when mesodermal cells along the ventral midline invaginate to form the ventral furrow, which internalizes to produce the mesodermal layer. Endoderm invaginates at the anterior and posterior ends (the anterior and posterior midgut invaginations). Because the embryo is syncytial, early patterning by morphogen gradients (Bicoid, Dorsal) occurs through protein diffusion in a shared cytoplasm.

Xenopus. The frog embryo has moderate yolk content concentrated in the vegetal hemisphere. Gastrulation involves coordinated involution of marginal zone cells through the dorsal lip of the blastopore. Surface cells roll inward over the blastopore lip and spread along the inner surface of the blastocoel roof, forming the mesodermal and endodermal layers. Simultaneously, the animal pole ectoderm expands by epiboly (spreading of an epithelial sheet) to cover the entire embryo. The dorsal mesoderm undergoes convergent extension -- it narrows mediolaterally and lengthens anterior-posteriorly, elongating the body axis.

Chick and mouse. In amniotes, the blastopore is replaced by the primitive streak, a groove along the posterior midline of the epiblast. Epiblast cells undergo epithelial-to-mesenchymal transition (EMT) at the streak, ingressing between the epiblast and hypoblast layers to form mesoderm and definitive endoderm. The anterior end of the streak (Hensen's node in the chick, the node in the mouse) is functionally equivalent to the Spemann organizer. The node expresses the same BMP antagonists (Chordin, Noggin) and gives rise to the notochord.

Cell movements in gastrulation

The repertoire of cell movements during gastrulation includes:

• Invagination: local buckling of an epithelial sheet inward (sea urchin vegetal plate, Drosophila ventral furrow).
• Involution: rolling of a sheet of cells over a lip or edge (Xenopus blastopore lip).
• Ingression: detachment of individual cells from an epithelium and migration into the interior (sea urchin primary mesenchyme, chick primitive streak).
• Epiboly: spreading of a surface epithelium to cover a larger area (Xenopus animal cap ectoderm, zebrafish yolk syncytial layer).
• Convergent extension: narrowing and lengthening of a tissue through mediolateral cell intercalation (Xenopus dorsal mesoderm, vertebrate body axis elongation).
• Delamination: splitting of one cell layer into two (mammalian hypoblast formation from the inner cell mass).

The Spemann organizer: molecular mechanisms

The Spemann organizer (dorsal lip of the blastopore in amphibians; Hensen's node/the node in amniotes) specifies dorsal-anterior fates by secreting antagonists of ventralizing signals. The organizer's strategy is predominantly inhibition of inhibition:

Organizer signal	Target inhibited	Pathway blocked	Effect
Chordin	BMP2/4	BMP-Smad1/5/8	Neural induction, dorsal mesoderm
Noggin	BMP2/4/7	BMP-Smad1/5/8	Neural induction
Follistatin	Activin/BMP	TGF-beta family	Neural induction
Dickkopf (Dkk)	LRP5/6 co-receptor	Wnt/beta-catenin	Head formation, anterior identity
Cerberus	BMP, Wnt, Nodal	Multiple	Head induction
Lefty	Nodal	Nodal-Smad2/3	Restriction of Nodal to left side (LR axis)
SFRP (Frzb)	Wnt	Wnt/beta-catenin	Anterior neural protection

The default model of neural induction (proposed by Hemmati-Brivanlou and Melton, 1997) holds that ectoderm autonomously tends toward neural fate unless inhibited by BMP signalling. The organizer secretes BMP antagonists (Chordin, Noggin, Follistatin) that block BMP, thereby permitting neural differentiation. This was demonstrated experimentally by culturing amphibian ectoderm with BMP inhibitors: neural tissue formed in the absence of any organizer-derived signals.

Nodal signalling is required for mesoderm induction and organizer formation. In Xenopus, maternal mRNA for the TGF-beta family ligand Vg1 is localized to the vegetal cortex. After fertilization, cortical rotation (driven by microtubule arrays) transports Dishevelled, GSK3-binding protein (GBP), and beta-catenin to the future dorsal side. Stabilized beta-catenin activates transcription of Siamois and Twin (paired-like homeobox genes), which in turn activate the Nodal-related genes Xnr1, Xnr2, Xnr4 and the transcription factor Goosecoid -- defining the organizer.

Anterior-posterior axis: Hox genes

Hox genes are a family of conserved transcription factors containing a 60-amino-acid homeodomain that binds DNA. They exhibit two forms of colinearity:

1. Spatial colinearity: the order of Hox genes on the chromosome matches their expression domains along the anterior-posterior axis. Genes at the 3' end of the cluster are expressed most anteriorly; genes at the 5' end are expressed most posteriorly.
2. Temporal colinearity: genes at the 3' end are activated first; genes at the 5' end are activated later, as development proceeds.

In Drosophila, a single Hox cluster (the Antennapedia complex and Bithorax complex) contains 8 Hox genes. In mammals, four Hox clusters (HoxA, HoxB, HoxC, HoxD, on chromosomes 7, 17, 12, and 2) contain 39 genes total, arising from two rounds of whole-genome duplication in early vertebrate evolution.

Hox genes operate by posterior prevalence: a more posteriorly expressed Hox gene suppresses or overrides the effect of more anterior Hox genes. This creates sharp boundaries of segment identity. For example, Hoxc6 is expressed in the thoracic region and specifies rib-bearing vertebrae; its expression boundary defines the cervical-thoracic transition. In mice, loss of Hoxc6 causes cervical vertebrae to form ribs (a homeotic transformation), because the anterior genes that normally specify cervical identity now act without correction from Hoxc6.

Retinoic acid (RA) is a key morphogen that regulates Hox gene activation. RA is produced in the posterior mesoderm and forms a gradient (high posterior, low anterior). RA activates 3' Hox genes at lower concentrations and 5' Hox genes at higher concentrations, contributing to the spatial colinearity of expression. Excess RA during pregnancy can cause homeotic transformations and birth defects, because it shifts Hox gene expression boundaries.

Dorsal-ventral axis: BMP gradient

The dorsal-ventral axis is patterned by a gradient of BMP (Bone Morphogenetic Protein) signalling. In Xenopus:

• The organizer on the dorsal side secretes BMP antagonists (Chordin, Noggin, Follistatin), creating a zone of low BMP signalling.
• Ventral mesoderm produces BMP2/4, creating a zone of high BMP signalling.
• Between these extremes, intermediate BMP concentrations specify intermediate fates.

The BMP gradient specifies mesodermal subtypes:

BMP level	Dorsal-ventral position	Mesodermal fate
Low	Dorsal	Notochord, prechordal plate (axial mesoderm)
Low-moderate	Dorsolateral	Somites (paraxial mesoderm)
Moderate	Lateral	Kidney, gonads (intermediate mesoderm)
High	Ventral	Blood, blood vessels (lateral plate mesoderm)

In Drosophila, the DV axis is patterned by the Dorsal protein (an NF-kB homolog), which forms a nuclear concentration gradient (high ventral, low dorsal). High Dorsal specifies mesoderm; moderate Dorsal specifies neurogenic ectoderm; low Dorsal specifies dorsal ectoderm. The insect DV axis is inverted relative to vertebrates (the ventral side of a fly corresponds to the dorsal side of a vertebrate), but the underlying logic of a morphogen gradient specifying tissue types along the axis is conserved.

Left-right axis

The left-right axis is the last to be established and is critical for the asymmetric placement of organs (heart on the left, liver on the right, spleen on the left). The molecular pathway involves:

1. Ciliary rotation at the embryonic node (or its equivalent). Motile cilia rotate clockwise, generating a leftward flow of extraembryonic fluid (nodal flow). This was demonstrated in mice with mutations in the axonemal dynein gene Kif3a or Lrd (left-right dynein): nodal flow is abolished, and organ laterality is randomized.

2. Asymmetric Nodal expression. Nodal flow is sensed by immotile mechanosensory cilia on the left side of the node, or carries a morphogen to the left side. This activates Nodal and Lefty2 expression specifically on the left side of the lateral plate mesoderm.

3. Pitx2 activation. Left-sided Nodal signalling activates the transcription factor Pitx2 on the left side. Pitx2 then drives asymmetric organ morphogenesis: leftward looping of the heart tube, leftward stomach rotation, asymmetric lung lobation.

Counterexamples to common slips

• Gastrulation is not a single movement but a coordinated set of cell behaviours (invagination, involution, ingression, epiboly, convergent extension) that vary across species while achieving the same three-layer outcome.
• The organizer does not build structures; it signals to neighbouring cells to change their fate. The secondary axis in the Spemann-Mangold experiment is made of host cells, not donor cells.
• Hox genes do not specify the segments themselves (that is the job of segmentation genes); they specify the identity of already-existing segments.
• BMP is not simply a "bone-inducing" molecule. In development, BMP signalling is a conserved dorsoventral patterning signal whose role in bone formation is a specialized function in vertebrates.
• The three germ layers do not correspond exactly to "outer, middle, and inner" after gastrulation is complete. Endoderm forms internal organs but also the epithelial lining of the gut tube, which is topologically external (continuous with the outside world at mouth and anus).

Key theorem with proof Intermediate+

Theorem (BMP gradient domain boundaries). Consider a one-dimensional dorsal-ventral axis of length $L$, with BMP produced at the ventral end (position $x=L$) at rate $\rho$ and degraded at rate $\mu$. The organizer at the dorsal end (position $x=0$) secretes a BMP antagonist at rate $\alpha$. In the simplest model, the antagonist acts as a sink for BMP with rate constant $k$. If free BMP concentration satisfies the steady-state equation

$$D\frac{d^2[B]}{d{x}^2}-\mu [B]-k[A](x)\cdot [B]=-\rho \delta (x-L),$$

and the antagonist $[A](x)$ decays exponentially from its source: $[A](x)=A_0{e}^{-x/{\lambda }_A}$, then the free BMP concentration defines distinct dorsoventral domains. For the case where $k[A](x)\ll \mu$ (weak antagonism approximation), the BMP gradient is approximately

$$[B](x)\approx B_0\left(e^{-(L-x)/{\lambda }_B}-\frac{k{A}_0{\lambda }_A^2}{D}{\int }_0^{(L-x)/{\lambda }_A}{e}^{-s}{e}^{-{\lambda }_A\sqrt{\mu /D}\cdot s}ds\right),$$

where ${\lambda }_B=\sqrt{D/\mu }$ is the BMP decay length. A tissue fate $F_i$ specified at BMP threshold $T_i$ occupies the domain $x>x_i$ where $[B](x)>T_i$. The domain widths depend on both the BMP source strength $B_0$ and the antagonist source strength $A_0$.

Proof sketch. In the absence of the antagonist ($k=0$), the BMP gradient is the standard exponential decay from the ventral source: $[B](x)=B_0{e}^{-(L-x)/{\lambda }_B}$. Introducing the antagonist adds a spatially varying degradation term $k[A](x)$. When the antagonist effect is treated as a perturbation, the correction to the free BMP concentration is negative and largest near the dorsal end (where antagonist concentration is highest), reducing BMP levels dorsally. This steepens the effective BMP gradient, sharpening the boundary between dorsal and ventral fates.

The domain boundary for a fate specified at threshold $T_i$ is found by solving $[B](x_i)=T_i$. In the unperturbed case, $x_i=L-{\lambda }_B\ln (B_0/T_i)$. The antagonist shifts this boundary ventrally (toward higher $x$), because BMP is reduced near the dorsal end. The shift is proportional to the antagonist strength $A_0$. This explains the experimental observation that overexpression of Chordin (a BMP antagonist) dorsalizes the embryo: increasing $A_0$ expands dorsal fates at the expense of ventral fates, shifting all domain boundaries ventrally. $\square$

Bridge. The BMP gradient model illustrates how two opposing sources (a morphogen source and an antagonist source) interact to produce a steeper, more robust gradient than either alone. This principle of "sharpening by antagonism" recurs throughout development: the Spemann organizer sharpens the DV axis, the anterior organizer (Cerberus, Dkk) sharpens the AP axis, and Lefty sharpens the LR axis. In each case, the antagonist is produced at one end, opposes the morphogen, and creates a zone of low signalling that defines a specific tissue fate. The quantitative framework extends directly to the reaction-diffusion systems analysed in the Master tier.

Exercises Intermediate+

Advanced treatment Master

Morphogen gradient models: French flag and beyond

The French flag model (Wolpert, 1969) formalized how a continuous morphogen gradient is converted into discrete tissue domains by concentration-threshold responses. In its simplest form, a morphogen $M$ produced at a source diffuses through tissue and decays, establishing an exponential gradient $M(x)=M_0{e}^{-x/\lambda }$. Cells read the local concentration and activate different gene programmes above different thresholds ($T_1>T_2>T_3$), producing three domains (analogous to the blue, white, and red of the French flag).

The French flag model makes specific quantitative predictions:

1. Robustness of intermediate domains. The width of an intermediate domain between thresholds $T_1$ and $T_2$ is $\Delta x=\lambda \ln (T_1/T_2)$, which depends only on the threshold ratio, not on the source concentration $M_0$. Intermediate fates are therefore buffered against fluctuations in morphogen production.

2. Scaling. If the gradient length scale $\lambda$ scales with the domain size $L$ (e.g., through regulation of the diffusion coefficient or degradation rate), the pattern scales proportionally. This scaling problem has been studied extensively in the Drosophila embryo, where the Bicoid gradient adjusts its decay length in proportion to embryo length across related species.

3. Boundary sharpness. The sharpness of a fate boundary depends on the steepness of the gradient ($dM/dx$) relative to the threshold-reading precision. Steeper gradients produce sharper boundaries. Hill-type cooperative responses ($M^n/(M^n+K^n)$ with Hill coefficient $n>1$) sharpen the threshold readout.

The French flag model applies directly to the BMP gradient in DV patterning and the retinoic acid gradient in AP patterning. However, not all patterning events fit this model. Many embryonic patterns arise from self-organizing reaction-diffusion systems (Turing patterns), where no preformed gradient is required.

Reaction-diffusion systems and pattern formation

As developed in 18.11.01, Turing (1952) showed that two interacting chemicals with different diffusion rates can spontaneously generate spatial patterns from a homogeneous initial state. The Gierer-Meinhardt activator-inhibitor system provides the canonical formulation. In gastrulation, reaction-diffusion dynamics contribute to:

• Somite segmentation: the segmentation clock -- an oscillating gene expression cascade involving Notch-Delta signalling, Hes/Her genes, and Lunatic fringe -- generates periodic stripes of gene expression that prefigure somite boundaries. The clock oscillates with a period of ~90 minutes (in the chick), and the wavefront of oscillation sweeps posteriorly as the body axis elongates. Each oscillation cycle deposits one pair of somites. The clock-and-wavefront model (Cooke and Zeeman, 1976) combines a posteriorly moving determination front (set by FGF/Wnt gradients) with the oscillating clock to segment the presomitic mesoderm.

• Limb patterning: the Sonic hedgehog (Shh) - Gremlin activator-inhibitor system in the limb bud generates the periodic pattern of digits. Shh activates its own expression (activator) and induces Gremlin (inhibitor), which diffuses further and blocks Shh signalling. Loss of Gremlin produces polydactyly (extra digits).

Hox gene evolution: cluster duplication and regulatory diversification

The Hox gene cluster has undergone dramatic evolutionary change:

• Ancestral condition: A single Hox cluster, as seen in amphioxus (Branchiostoma, a cephalochordate), with ~14 genes. This likely represents the ancestral chordate condition.

• Vertebrate whole-genome duplications: Two rounds of whole-genome duplication (the "2R hypothesis") at the base of vertebrate evolution produced four Hox clusters (HoxA, B, C, D). Teleost fish underwent a third duplication ("3R"), producing up to eight clusters (though some were subsequently lost).

• Cluster degeneration: After duplication, genes are redundant and can be lost without penalty. The mammalian HoxA cluster retains 11 genes, HoxB retains 10, HoxC retains 9, and HoxD retains 9. Some duplicated genes acquired new functions (neofunctionalization), while others subdivided the original gene's expression domain (subfunctionalization).

• Non-coding regulatory evolution: Changes in the cis-regulatory elements of Hox genes are the primary mechanism by which Hox clusters evolve to produce morphological diversity. The HoxD cluster in vertebrates, for example, is regulated by a global control region (GCR) and a late-phase regulatory landscape that together orchestrate the complex spatial expression of multiple HoxD genes during limb development. Evolutionary changes in these regulatory elements account for differences in limb morphology between mammals, birds, and snakes.

• Cluster fragmentation and loss: In some lineages, the Hox cluster has partially or completely disintegrated. The tunicate Oikopleura dioica has dispersed Hox genes, and the nematode Caenorhabditis elegans has lost the clustered arrangement entirely. These cases demonstrate that colinearity is not an absolute requirement for Hox gene function, although it is conserved in the vast majority of bilaterians.

Homeotic mutations in Drosophila

The genetic dissection of Hox gene function in Drosophila produced some of the most iconic results in developmental biology:

• Antennapedia (Antp): loss-of-function mutations cause antennae to form in place of legs on the second thoracic segment (legs are the default fate of thoracic segments; Antp is normally repressed in the head by the anterior Hox genes Labial and Proboscipedia). Gain-of-function mutations (ectopic expression of Antp in the head) cause legs to form in place of antennae -- a spectacular homeotic transformation.

• Bithorax (Ubx, abd-A, Abd-B): the Bithorax complex controls posterior segment identity. Loss of Ubx transforms the third thoracic segment (T3, which normally bears the halteres -- balancing organs) into a second copy of T2 (which bears the wings), producing a four-winged fly. This was Ed Lewis's life work and earned him the Nobel Prize in 1995 (shared with Nusslein-Volhard and Wieschaus).

• Comb mutations: Scr (Sex combs reduced) specifies the first thoracic segment. Ectopic expression in the second thoracic segment produces sex combs on the second pair of legs.

These mutations demonstrate that Hox genes control identity, not formation. The segments still form; they simply adopt the wrong identity. A bithorax fly has the correct number of segments but the wrong structures on one of them.

Gastrulation defects in human development

Disruption of gastrulation produces severe, often lethal, congenital malformations:

• Neural tube defects (NTDs): Failure of neurulation (which follows gastrulation) produces spina bifida (incomplete closure of the posterior neural tube) and anencephaly (failure of anterior neural tube closure). The molecular basis involves folate metabolism: folate is required for one-carbon metabolism and methylation of genes involved in neural tube closure. Folate supplementation (400 microg/day) before and during early pregnancy reduces NTD incidence by ~70%. Mouse models show that mutations in the Planar Cell Polarity pathway (Vangl2/Looptail, Celsr1/Crash) disrupt convergent extension and cause NTDs.

• Caudal regression syndrome: Defective gastrulation at the posterior end of the primitive streak can cause sacral agenesis (absence of sacral vertebrae) and lower spinal cord malformations. Associated with maternal diabetes (hyperglycaemia disrupts the primitive streak) and mutations in Van Gogh-like 2 (Vangl2). Severity ranges from isolated sacral vertebral defects to sirenomelia (fusion of the lower limbs, historically called "mermaid syndrome").

• Laterality disorders: Failure of the left-right axis to form correctly produces heterotaxy (randomized organ situs), situs inversus totalis (complete mirror-image reversal, often asymptomatic), and situs ambiguus (partial reversal, with complex cardiac and visceral malformations). Approximately 1 in 8,000 live births.

• Conjoined twinning: Incomplete separation of the inner cell mass or abnormal organizer duplication can produce conjoined twins. The classification (thoracopagus, omphalopagus, craniopagus, pygopagus) reflects the site of fusion, which corresponds to the region where the body axis failed to separate. Conjoined twinning is thought to arise from a single zygote that develops two primitive streaks or two organizer regions on a single embryonic disc, producing two body axes that remain partially fused.

Situs inversus and primary ciliary dyskinesia

Situs inversus totalis (complete mirror-image reversal of all organs) occurs in approximately 1 in 10,000 individuals and is usually asymptomatic -- the organs function normally in their reversed positions. However, situs inversus is a hallmark of primary ciliary dyskinesia (PCD), also known as Kartagener syndrome when situs inversus is present. PCD is caused by mutations in genes encoding axonemal dynein arms (DNAH5, DNAI1) or other ciliary components. The immotile cilia fail to generate nodal flow during gastrulation, randomizing left-right patterning. Approximately 50% of PCD patients have situs inversus (the other 50% have normal situs, reflecting the random outcome of the LR decision).

PCD patients also suffer from chronic respiratory infections (immotile respiratory cilia fail to clear mucus), chronic sinusitis, otitis media, and male infertility (immotile sperm flagella). The link between cilia, LR axis determination, and respiratory disease is one of the most striking examples of a shared molecular mechanism underlying apparently unrelated clinical conditions.

Twinning and axis formation

Twinning provides insight into when and how the body axes are established:

• Dizygotic (fraternal) twins: two separate oocytes are fertilized by two separate sperm. Each zygote develops its own axis independently. No shared axes or organizer.

• Monozygotic (identical) twins: a single zygote splits at various stages. The timing determines the relationship between the twins' axes:

  • Day 1--4 (before blastocyst formation): splitting occurs before axis specification. Each twin develops independently with separate chorions and amnions (dichorionic, diamniotic). The twins may have opposite handedness, suggesting independent LR axis specification.
  • Day 4--8 (early blastocyst): splitting occurs after blastocyst formation but before the primitive streak. Twins share a chorion but have separate amnions (monochorionic, diamniotic).
  • Day 8--12 (post-ICM): splitting occurs after ICM formation. Twins share both chorion and amnion (monochorionic, monoamniotic). Higher risk of cord entanglement.
  • Day 13+: incomplete splitting produces conjoined twins (see above). The axes are partially shared, with fusion at the site where the split was incomplete.

The phenomenon of mirror-image twinning (approximately 25% of monozygotic twins show opposite handedness, and a subset show mirror-image asymmetries such as hair whorl direction or dental patterns) may reflect the influence of the LR axis on morphological laterality.

Stem cell models of gastrulation: gastruloids

Gastruloids are three-dimensional aggregates of mouse or human pluripotent stem cells (PSCs) that self-organize to recapitulate key events of gastrulation in vitro. First described by van den Brink and colleagues (2014), gastruloids are generated by:

1. Aggregating ~300 mouse PSCs in low-adhesion U-bottom plates.
2. Treating with the Wnt activator CHIR99021 for 24 hours (mimicking the Wnt signal that initiates gastrulation).
3. Withdrawal of CHIR, allowing self-organization.

Over 3--5 days, gastruloids undergo symmetry breaking, establish AP polarity (marked by Bra/T expression at the posterior and Otx2 at the anterior), elongate through convergent extension, and generate cells of all three germ layers in spatially organized domains. Remarkably, gastruloids recapitulate the Hox gene colinearity seen in vivo, with Hox genes activated sequentially from anterior to posterior along the elongating structure.

Gastruloids have several advantages over traditional embryo models:

• They are accessible to live imaging, drug perturbation, and genetic manipulation.
• They avoid the ethical constraints associated with culturing human embryos beyond 14 days.
• They can be generated in large numbers for high-throughput screening.
• They capture self-organization dynamics that cannot be studied in two-dimensional cultures.

Current limitations include the absence of extraembryonic tissues (no placenta, no yolk sac), the lack of a functional cardiovascular system, and incomplete organogenesis. Despite these limitations, gastruloids have been used to model the effects of teratogens on gastrulation, to study human-specific developmental events inaccessible in model organisms, and to test the sufficiency of specific signalling pathways for axis formation.

More recently, embryo models assembled from combinations of embryonic and extraembryonic stem cells (Sozen et al., 2019; Amadei et al., 2022) have recapitulated implantation, lumen formation, and early gastrulation-like events, extending the gastruloid concept to structures that more closely resemble the post-implantation embryo. These models raise new ethical questions about how closely an in vitro structure must resemble an embryo before it warrants equivalent moral consideration.

Connections Master

• Embryology and morphogenesis 18.11.01. This unit deepens the gastrulation overview presented in 18.11.01, providing detailed treatment of the cell movements (ingression, involution, epiboly, convergent extension), the molecular mechanisms of the Spemann organizer, and the Hox gene system. The morphogen gradient models (French flag, Turing) introduced in 18.11.01 are applied here to specific axis-formation events.

• Gene regulation 17.06.01. Hox genes are transcription factors that bind to enhancers and regulate downstream gene cascades. The spatial colinearity of Hox expression is controlled at the level of chromatin opening and enhancer activation -- direct applications of the gene-regulation mechanisms covered in the gene-regulation unit. The concept of cis-regulatory evolution (changes in enhancer sequences rather than protein-coding sequences) as the primary driver of morphological evolution is a key bridge between gene regulation and evolutionary biology.

• Mendelian genetics 19.01.01. Homeotic mutations (Antennapedia, bithorax) follow Mendelian inheritance patterns and were among the first developmental mutations to be genetically mapped. The Nusslein-Volhard and Wieschaus saturation screen used classical genetics to systematically identify segmentation genes, demonstrating that forward genetics can dissect developmental pathways.

• Evolution and natural selection 19.03.01. Hox gene cluster duplication in vertebrate evolution, the conservation of Hox gene function across bilaterians, and the evolution of cis-regulatory elements are central topics in evolutionary developmental biology ("evo-devo"). The deep homology of body-patterning mechanisms -- the same Hox genes pattern the body axis in flies, worms, and humans -- is one of the strongest lines of evidence for common descent.

• Immunology 18.10.01. The Dorsal protein in Drosophila (which patterns the DV axis) is homologous to the NF-kB transcription factor that controls immune gene expression in vertebrates. This is a striking example of molecular co-option: a signalling pathway that evolved for developmental patterning was later recruited for immune defence.

• Fertilization and early development 18.09.03 pending. The blastocyst (described in 18.09.03 pending) is the starting point for mammalian gastrulation. The inner cell mass gives rise to the epiblast, which undergoes gastrulation through the primitive streak. The primitive streak forms at the posterior end of the epiblast approximately 14--16 days post-fertilization in humans.

Historical & philosophical context Master

The Spemann-Mangold organizer experiment (1924) is one of the most important experiments in the history of biology. Hans Spemann had already demonstrated the concept of embryonic induction in 1901, when he showed that the lens of the eye forms only when the optic vesicle contacts the overlying ectoderm. The 1924 experiment, performed with his graduate student Hilde Mangold, went further: it showed that a specific group of cells could organize an entire body axis.

Hilde Mangold was 26 years old when the paper was published. She died later that year in an explosion caused by a faulty gas heater in her apartment building. Spemann received the Nobel Prize in 1935 and acknowledged Mangold's contribution, but the prize was awarded to him alone (Nobel Prizes are not awarded posthumously). Mangold's thesis, which contained the complete experimental results, was published posthumously. The original experiment was reconstructed by Horst Freytag and colleagues in 2003, confirming that Mangold's results were reproducible and that the donor organizer tissue had indeed been correctly identified.

The molecular identity of the organizer's signals was not uncovered until the 1990s. In 1994, Smith and Harland isolated Noggin from Xenopus and showed it could dorsalize ventral mesoderm. Chordin was identified by De Robertis and colleagues in 1994 as a BMP-binding protein expressed in the organizer. Follistatin was shown to bind Activin and BMP by Hemmati-Brivanlou's group in 1995. The "default model" of neural induction (BMP inhibition is necessary and sufficient for neural fate) was proposed by Hemmati-Brivanlou and Melton in 1997 and confirmed by the triple knockout of Chordin, Noggin, and Follistatin in mice, which eliminates neural tissue entirely.

The discovery of Hox genes followed a parallel trajectory. Edward B. Lewis spent decades studying the Bithorax complex in Drosophila, mapping mutations that transformed one body segment into another. His 1978 paper "A gene complex controlling segmentation in Drosophila" proposed that the genes were arranged colinearly on the chromosome and that they acted sequentially to specify segment identity. When the homeobox was cloned by McGinnis, Levine, and colleagues in 1984, it was found in essentially the same arrangement in vertebrates -- a result that astonished the field and launched the discipline of evolutionary developmental biology. The conservation of the homeobox across all bilaterian animals is one of the most dramatic examples of deep homology in biology.

The left-right axis was the last to be understood at the molecular level. The key insight came from Nonaka et al. (1998), who showed that motile cilia in the mouse embryonic node generate a leftward fluid flow that initiates LR asymmetry. This connected a longstanding embryological observation (the asymmetry of the heart loop) to a cellular structure (the cilium) and a disease (primary ciliary dyskinesia/Kartagener syndrome, described by Manes Kartagener in 1933 as the triad of situs inversus, chronic sinusitis, and bronchiectasis). The realization that a single organelle -- the cilium -- links embryonic patterning, respiratory defence, and sperm motility is a striking example of how understanding developmental mechanisms can illuminate apparently unrelated clinical phenomena.

Philosophically, the study of gastrulation and axis formation has driven the transition from descriptive embryology to mechanistic developmental biology. Spemann and Mangold's experiment transformed the organizer from a descriptive concept ("the dorsal lip has special properties") to an experimentally testable hypothesis ("specific cells secrete signals that reorganize their neighbours"). The subsequent molecular identification of Chordin, Noggin, and other organizer signals completed the reduction of a classic embryological phenomenon to defined molecules and pathways. Yet the organizer retains its conceptual power because it captures an emergent property -- self-organization of the body plan -- that cannot be predicted from the properties of any single molecule. The organizer exists not in any one gene or protein but in the spatial relationship between a signalling centre and its responsive neighbours.

Bibliography Master

1. Spemann, H. & Mangold, H., "Uber Induktion von Embryonalanlagen durch Implantation artfremder Organisatoren", Roux' Arch. Entw. Mech. 100 (1924), 599--638.

2. Wolpert, L., "Positional information and the spatial pattern of cellular differentiation", J. Theor. Biol. 25 (1969), 1--47.

3. Lewis, E. B., "A gene complex controlling segmentation in Drosophila", Nature 276 (1978), 565--570.

4. Nusslein-Volhard, C. & Wieschaus, E., "Mutations affecting segment number and polarity in Drosophila", Nature 287 (1980), 795--801.

5. McGinnis, W., Levine, M. S., Hafen, E., Kuroiwa, A. & Gehring, W. J., "A conserved DNA sequence in homoeotic genes of the Drosophila Antennapedia and bithorax complexes", Nature 308 (1984), 428--433.

6. Smith, W. C. & Harland, R. M., "Expression cloning of noggin, a new dorsalizing factor localized to the Spemann organizer in Xenopus embryos", Cell 70 (1992), 829--840.

7. Sasai, Y., Lu, B., Steinbeisser, H., Geissert, D., Gont, L. K. & De Robertis, E. M., "Xenopus chordin: a novel dorsalizing factor activated by organizer-specific homeobox genes", Cell 79 (1994), 779--790.

8. Hemmati-Brivanlou, A. & Melton, D. A., "Vertebrate neural induction", Annu. Rev. Neurosci. 20 (1997), 43--60.

9. Nonaka, S., Tanaka, Y., Okada, Y., Takeda, S., Harada, A., Kanai, Y., Kido, M. & Hirokawa, N., "Randomization of left-right asymmetry due to loss of nodal cilia generating leftward flow of extraembryonic fluid in mice lacking KIF3B motor protein", Cell 95 (1998), 829--837.

10. Cooke, J. & Zeeman, E. C., "A clock and wavefront model for control of the number of repeated structures during animal morphogenesis", J. Theor. Biol. 58 (1976), 455--476.

11. Gilbert, S. F., Developmental Biology, 12th ed. (Sinauer, 2020).

12. Sadava, D., Hillis, D. M., Heller, H. C. & Berenbaum, M. R., Life: The Science of Biology, 13th ed. (Sinauer, 2023).

13. Sherwood, L., Human Physiology: From Cells to Systems, 9th ed. (Cengage, 2016).

14. Davidson, E. H. et al., "A genomic regulatory network for development", Science 295 (2002), 1669--1678.

15. van den Brink, S. C. et al., "Symmetry breaking, germ layer specification and axial organisation in aggregates of mouse embryonic stem cells", Development 141 (2014), 4231--4242.

16. Turing, A. M., "The chemical basis of morphogenesis", Phil. Trans. Roy. Soc. B 237 (1952), 37--72.

17. Gierer, A. & Meinhardt, H., "A theory of biological pattern formation", Kybernetik 12 (1972), 30--39.

18. Afzelius, B. A., "A human syndrome caused by immotile cilia", Science 193 (1976), 317--319.

19. Amadei, G. et al., "Inducible stem-cell-derived embryos model human post-implantation development", Nature 610 (2022), 615--623.