January 1, 2019;
What are the roles of retinoids, other morphogens, and Hox genes in setting up the vertebrate body axis?
This article is concerned with the roles of retinoids and other known anterior
morphogens in setting up the embryonic vertebrate anterior
axis. The discussion is restricted to the very earliest events in setting up the anterior
axis (from blastula
stages in Xenopus embryos). In these earliest developmental stages, morphogen concentration gradients are not relevant for setting up this axis. It emerges that at these stages, the core patterning mechanism is timing: BMP-anti BMP mediated time space translation that regulates Hox temporal and spatial collinearities and Hox-Hox auto- and cross- regulation. The known anterior
morphogens and signaling pathways--retinoids, FGF''s, Cdx, Wnts, Gdf11
and others--interact with this core mechanism at and after space-time defined "decision points," leading to the separation of distinct axial domains. There are also other roles for signaling pathways. Besides the Hox regulated hindbrain/trunk part of the axis, there is a rostral part (including the anterior
part of the head
and the extreme anterior
domain [EAD]) that appears to be regulated by additional mechanisms. Key aspects of anterior
axial patterning, including: the nature of different phases in early patterning and in the whole process; the specificities of Hox action and of intercellular signaling; and the mechanisms of Hox temporal and spatial collinearities, are discussed in relation to the facts and hypotheses proposed above.
[+] show captions
Figure 1. Timing, axial patterning, and time-space translation. Above: The structure of the vertebrate A–P axis: domains with significant Hox genes and other markers. An unexpected element is introduced by the newly characterized extreme anterior domain (EAD), which makes the face. This is shown as the most anterior part of the straight axis. Actually, the anterior end of the dorsal A–P axis bends backward around to the ventral side of the embryo to face posteriorly like the handle of a walking stick (not shown). A. Head: anterior head (corresponding to telencephalon, diencephalon, mesencephalon). P. Head: posterior head (corresponding to anterior rhombencephalon, occipital somites). Neck: cervical somites, posterior rhombencephalon, Thorax: thoracic vertebrae, anterior spinal cord. Abdomen: Lumbar and sacral regions, spinal cord. Tail: coccygeal vertebrae, spinal cord. Above and below: Time space translation. A biological timer, represented by the clock face below, proceeds from 1 to 12 (red numbers). The timer starts with information needed for making the EAD, proceeds to the anterior head, then to posterior head, then to neck, then to thorax, then abdomen, then tail. The timer needs BMP to function, so occurs in tissues with high BMP (yellow/orange). Anti–BMP factors (blue) interact with the timer sequentially to freeze the identities of an A–P sequence of axial zones. In the axial sequence, the Hox genes are each both a component of the timer at their appropriate times and are sequentially involved in setting up the A–P sequence of axial zones. The genes involved in time space translation in the EAD‐head zones are unknown. The head and tail of the A–P timer are close together because of their representation on a clock face. No statement about molecular identities is intended.
Figure 2. Xenopus Hox sequences for axial cascade phenotypes relating to domain boundaries. Above: Wild type Hox sequence. Next down: Hox1 loss of function (LOF; all 3 Hox1 genes knocked down by morpholinos (MOs). The axis from Hox1 backward is compromised. The dotted line indicates there is still reduced residual expression for some posterior Hox genes. Next down: Hoxc6 LOF (MO). The axis from Hoxc6 backward is compromised/deleted. Next down: Hoxb9 gain of function (GOF) by ectopic expression of Hoxb9 in Hox‐free dorsalised embryos. A partial posterior axis is generated, starting with Hoxb9.
Figure 3. Hox genes, morphogens and boundaries between axial domains. (a) The neck‐thorax boundary and Hoxc6 expression. Chick has a long neck (light blue) and short thorax (mid blue; 14 and 7 somites, respectively). Mouse has a short neck (light blue) and long thorax (mid blue; 7 and 14 somites, respectively). In each case, the Hoxc6 anterior expression boundary (C6, marked in red) is at the neck/thorax boundary. Other vertebrates with different axial formulae (goose, Xenopus, zebrafish) show the same relationship. Hoxc6 seems to be a special gene (Burke, Nelson, Bruce, Morgan, & Tabin, 1995). (b) Ectopic expression of the Wnt inhibitor Dickkopf (Dkk‐1) in the presence of the anti‐BMP dorsaliser, noggin, causes Xenopus embryos to develop head structures only (eye red arrowed); the anterior head‐posterior head boundary is blocked (Glinka et al., 1998). (c) Ectopic expression of Hoxa10 in the mouse. Right: Normal mouse skeleton showing thoracic ribs (marked by red arrows). Left: Hoxa10 GOF skeleton, showing no ribs. The whole thorax has become abdominal (lumbar vertebrae; thorax‐abdomen boundary deleted; Carapuço, Nóvoa, Bobola, & Mallo, 2005). (d) Normal and Gdx8−/− mice. The normal mouse (left) has a tail (red arrow). The Gdx−/− mouse essentially does not. The abdomen‐tail boundary is blocked (Jurberg, Aires, Varela‐Lasheras, Novoa, & Mallo, 2013).
Figure 4. Growth factors and the axial domains. The anterior head (A. Head)/posterior head (P. Head) boundary is influenced by active retinoids/retinoic acid (RA) and Wnt (8 or 3A). Both turn posterior head on. The neck/thorax boundary is influenced by RA (thorax off) and FGF/Cdx (thorax on). The thorax/abdomen boundary is influenced by Wnt. The abdomen/tail boundary is influenced by GDF11 (tail on) and RA, which blocks the transition of abdomen to tail, that is, it changes tail to limbs or truncates the axis.
Intercellular signaling in Hox collinearity and TST BMP signaling (yellow arrows} is permissive for noncell autonomous Hox‐Hox PI signaling in NO mesodem (NO mesoderm cells are big blue oblongs). Hox expression occurs in each cell with different specificities. The shade of blue in the small rectangles (nucleus) indicates which specificity. (1) First blue cell to the left: Gbx2 specificity (mauve). Gbx is not known to be expressed in NO mesoderm, only in neurectoderm (NE). The mauve color represents the underlying A–P level specificity. The mauve nucleus emits a specific signal (mauve arrow) to cell 2. (2) Hox1 specificity, nucleus darkest blue shade. This emits a Hox1‐specificity signal (dark blue arrow) to cell 3 (which becomes Hox2 specificity (mid blue nucleus). (3) Cell 3 emits a Hox2‐specific signal to cell 4 (mid blue arrow), etc. The different colored blue arrows indicate that cells emit specific Hox signals to neighboring cells. The specificity of the signal is the same as the expression specificity of the emitting cell. The response is to develop the next more posterior specificity (PI). One signaling mechanism that could accomplish this is that Hox homeoproteins could be conveyed nonspecifically from cell to cell by Prochiantz transfer (i.e., if they are expressed, they are transferred; Joliot & Prochiantz, 2004) and that the specificity of the response is determined intracellularly by Hox transcription factor specificity in the nucleus. All NO mesoderm‐NO mesoderm interactions also require a yellow arrow (BMP). This determines the specificity of the interaction as PI. The Gbx2‐Hox1 transition also requires Wnt signaling (orange) and retinoid signaling (red). These enable the anterior head‐posterior head transition. The blue cells also signal vertically to light green cells (NE, neurectoderm). This represents specific Hox signaling (Au) from NO mesoderm to NE, copying the same specificity. In all cases, this also requires a dark green signal from neighboring dark green cells (Spemann Organizer: SO). These green BMP inhibitions specify this interaction as Au and not PI. Please note that the position of the dark green SO cells is not pictorially accurate. All that is indicated is that these cells are close to and thus in range of the NE. In the first case (copying Hox1 specificity), the copying also requires retinoid signaling (red arrow) as well as the specific Hox signal. This represents retinoid signaling at the anterior head–posterior head decision point. Red arrows are also shown for 'Hox4 NO mesoderm‐NE and for Hox3 NO mesoderm‐ Hox4 NO mesoderm to indicate the retinoid requirement for vertical signaling is not exclusively confined to the anterior head‐posterior head decision point. I note that vertebrate Hoxb4 and Hoxd4 expression at the Hox4 posterior head‐neck decision point is known to be depend directly on retinoid signalling. Mauve rectangle and arrow indicate the expression of and specific signaling from the last head determinant (Gbx). Orange arrow: Wnt signaling: Required for NOM mesodermal transition through the anterior head–posterior head decision point. The proposal is that Hox1 is turned on in the NO mesoderm by the last anterior head determinant and BMP, Wnt and retinoids, and is then signaled from NO mesoderm to NE via Hox, BMP inhibition and retinoids. The scheme below the figure is a color key for the signaling arrows