Transformation of a neural activation and patterning model

Abstract

The activation and transformation model of vertebrate nervous system formation posits that neural tissue is initially induced, or activated, with anterior forebrain character. Once established, a subset is then transformed into the more posterior midbrain, hindbrain, and spinal cord by signals emanating from the posterior of the embryo. This has been a predominant model in the field for decades. In the June issue of EMBO Reports, Polevoy and colleagues evaluate the role of signals thought to act as the neural transforming factors during Xenopus development, and find that while these signals are consistent with the activation transformation model during brain patterning, they do not fit the model with respect to spinal cord formation [1]. This work, along with other recent studies on the origin of the spinal cord, necessitates an updated model of vertebrate nervous system formation, where spinal cord induction and patterning is distinct from that of the brain.

Developmental biologists have long been fascinated with the formation of the vertebrate nervous system, and several models have been proposed to describe the induction and subsequent patterning of neural tissue. One of the most prominent is called the activation and transformation model, first described by Nieuwkoop in 1954 2. Naïve ectodermal explants from the axolotl salamander grafted into different regions of neural tissue along the anterior–posterior axis of host embryos would first exhibit neural induction with an anterior identity (activation), followed by patterning into more posterior fates depending on the axial level of the graft (transformation). Thus, the activation and transformation model proposes that all neural tissue is first induced as anterior (forebrain) and subsequently transformed to more posterior fates such as hindbrain and spinal cord (Fig 1A).

Figure 1. An updated model of neural induction and patterning.

(A) The historical activation and transformation model, here shown in a Xenopus neural plate. Neural tissue is induced or activated from ectoderm via BMP inhibition, and subsequently patterned or transformed by a Wnt signaling gradient. (B) The work by Polevoy et al shows that neural induction and patterning of the brain follows the traditional activation and transformation model with respect to BMP and Wnt signaling, but that the spinal cord does not fit this model. In this neural tissue, BMP signaling promotes spinal cord through FGF activation, and low Wnt signaling does not inhibit spinal cord formation. (C) A lineage model of the central nervous system showing signaling inputs based on this work and the work of others in the NMP field, (D) which shows that in vertebrates, here shown in a mouse embryo, brain and spinal cord have separate origins. FB, forebrain; MB, midbrain; HB, hindbrain; SC, spinal cord; NMPs, neuromesodermal progenitors.

Several decades later, the molecular revolution in developmental biology allowed the discovery of some of the molecules involved in neural induction and patterning, and to a certain degree, these fit with the activation and transformation model (Fig 1A). The first of these, the BMP antagonists, could induce anterior neural tissue in naïve ectoderm 3. Later discoveries showed that the combination of BMP antagonists with posteriorly localized signals, such as FGF, canonical Wnt, or retinoic acid, could transform anterior neural tissue into more posterior fates 3. On the other hand, separate studies showed that the proposed transforming factor FGF can act as a direct neural inducer, independent of BMP antagonism, raising the possibility that posterior neural fates like the spinal cord are induced independently of anterior neural tissue 4, 5, 6. This alternative model supports classic embryology experiments performed by Mangold suggesting separate organizing and inductive events specifying anterior and posterior tissues independently 3.

The possibility that posterior neural tissue is induced without passing through an anterior intermediate state was further supported by the discovery of neuromesodermal progenitors (NMPs). NMPs reside in the posterior‐most domain of vertebrate embryos and continuously make a germ‐layer decision to form neural ectoderm (spinal cord) or mesoderm (primarily somites) well after gastrulation 7. The continuous allocation of spinal cord after brain specification runs contrary to a model where the neural plate is first established and later patterned along the anterior–posterior axis. Nevertheless, it remained formally possible that newly generated spinal cord cells coming from the NMPs first passed through an anterior forebrain intermediate before becoming spinal cord. This notion was conclusively laid to rest by a recent study from Metzis et al, which showed cells first adopt an axial position and then undergo neural fate acquisition appropriate to that axial level, and in so doing spinal cord cells do not pass through an anterior neural (brain) intermediate state 8.

The manuscript by Polevoy et al in this issue revisits some of the signaling factors that have been implicated in the activation and transformation model, including BMP, FGF, and canonical Wnt signaling 1. They identify new roles for these factors that also run contrary to the activation and transformation model. The authors show that elevated BMP signaling, which is thought to inhibit neural induction, can promote posterior neural (spinal cord) development in an FGF‐dependent fashion. This suggests opposing roles of BMP signaling in anterior vs posterior neurogenesis, with BMP antagonism inducing brain and BMP activation promoting spinal cord, again supporting independent embryonic origins of the brain and spinal cord. Polevoy et al also tested the role of canonical Wnt signaling during neural transformation. Inhibition of Wnt signaling caused a shift from hindbrain to forebrain patterning, but did not eliminate spinal cord. This result fits well with the NMP model of spinal cord development, as Wnt signaling plays an essential role in mediating the spinal cord vs mesoderm fate decision, where Wnt inhibition promotes the spinal cord fate 9. Thus, Wnt signaling acts as a transforming factor, but only within the brain (Fig 1B–D).

From the Polevoy and Metzis studies, along with the growing NMP literature, a new model of neural induction and patterning is emerging. In this model, the spinal cord and brain have separate embryonic origins, reflecting the differential signaling requirements for induction and patterning of these two regions of the nervous system. Brain formation follows a more traditional activation and transformation model, whereas the spinal cord is induced independently and sequentially from an NMP population during body axis elongation. Separate embryonic origins of the vertebrate brain and spinal cord support an evolutionary model of nervous system development. In this model, distinct apical and blastoporal nervous systems evolved in pre‐bilaterian metazoans, which later merged into the single continuous nervous system of bilaterians 10. Such evolutionary models will become further testable as our knowledge of the molecular basis of the vertebrate nervous system development grows.

EMBO Reports (2019) 20: e48060