Pattern formation in the nervous system, from the amphibian neural plate to the mammalian neocortex. Interpretation of inductive signals. Function of the SHH-Gli signaling pathway in the brain and in tumorigenesis.

The main focus of research in the laboratory is to understand how cells respond to positional information provided by inducing signals. The response to such signaling can then be either differentiation or proliferation. Within this general framework we have two major projects that involve work with frogs or mice, according to experimental needs. The two main projects at this time are: 1- How cells in the developing neural plate and brain respond to Sonic hedgehog-Gli signaling. 2- What are the molecular mechanisms underlying the parcellation of the mammalian neocortex.

Within the first project, we are studying many aspects of Gli protein function. We have cloned the Gli genes from frog (Xenopus laevis) embryos and analyzed their expression in the neural plate and neural tube. We have compared this to their expression in mice and initiated experiments in frogs given their ease of manipulation. Gli1 is expressed in midline neural plate cells and then in ventral neural tube cells, is a target of Shh signaling and induces ventral differentiation. In contrast, Gli2 and Gli3 are widely expressed in the neural plate and restricted mostly dorsally within the neural tube. Gli2 is also a target of Shh, suggesting a complex regulation of its expression, whereas Gli3 and Shh have a mutually antagonistic relationship. We have recently shown that Gli proteins have a strong neurogenic activity and that they appear to induce distinct classes of neurons, raising the possibility that a combinatorial Gli code is involved in patterning the CNS along the dorsoventral axis. We also recently showed that Zic2, a Gli-related protein, is expressed in defined zones within the neural plate and acts as a vertebrate prepattern gene. Zic2 counteracts the neurogenic role of Glis in the neural plate, thus defining neurogenic domains.

A parallel line of work has shown that Gli1 can induce epidermal tumors in frog epidermis when ectopically expressed. This was the first demonstration of the tumorigenic potential of Gli1 in vivo. In addition, we have shown that all human sporadic basal cell carcinomas tested express GLI1, providing a molecular framework for the cause of this most common skin tumor. Moreover, The Gli and Shh genes are expressed in the developing hair follicle, showing that the cellular framework for this tumor is the developing follicle. We predicted that Gli1 activation, as a last element in the Shh signaling pathway, is a sensitive detector of Shh signaling and sporadic BCC formation caused by any one of a variety of potential mutations. Consistent with our prediction, a recent paper has shown that a sporadic BCC appeared to be caused by a mutation in Smoothened, a component of the Shh receptor complex. Interestingly, there is evidence that medulloblastomas, childhood cancers of the cerebellum, also arise from deregulated Shh signaling and we have extended our study of BCCs into CNS tumors. In this context, we have recently shown that SHH controls the size and patterning of the cerebellum as it is a required endogenous mitogen for granule cell precursors in the external germinal layer. Here, it regulates Gli1 expression and drives proliferation of precursor cells in the outer EGL. These findings not only provide a clear view of how precursors for the most abundant neuronal type in the brain proliferate, but also they provide a cellular basis for medulloblastomas. We propose that any one of many possible mutations that will lead to the activation, proximally or distally, of the SHH-GLI pathway will then result in maintained GLI function and the initiation of cerebellar tumors. GLI proteins are thus prime targets for disease detection and control. Ongoing studies focus on the idea that the SHH-GLI pathway controls the development of the post-natal dorsal brain and that its deregulation leads to tumor initiation in diverse areas of the brain where stem cells are found.

Together, these findings beg the question of how the different Gli proteins act to dictate cell fate in neural and skin precursors and what is the biochemistry of these proteins. If it is like that of their Drosophila counterpart, Cubitus interruptus (Ci), it is likely to be quite interesting. Ci acts as a positive activator of hedgehog target genes but it is cleaved to form a smaller repressor. This cleavage, however, is inhibited by hedgehog signaling. We are now addressing how Gli proteins interact and if, and how, they are processed since these issues are critical to understand Gli protein function and dysfunction. A major effort in the lab is to obtain specific reagents with which to detect or alter Gli activity, such as specific monoclonal antibodies and viral constructs encoding dominant-negative Gli forms. In addition, we have now undertaken a transgenic approach to better understand Gli function in mice during normal development and in brain tumorigenesis.

The second project mainly involves a large molecular screen for genes differentially expressed within the prospective neocortex, the dorsal telencephalon, at stages (E10-13) before thalamic innervation takes place. We have already isolated a number of interesting gens the expression of which is regionalized and we are studying how this expression evolves. It is possible also that the function of the encoded proteins will play a role in regionalization and we are beginning to characterize them functionally. In parallel with this effort, we have discovered that the homeobox gene Otx2 is selectively expressed in the occipital, visual, neocortex of postnatal mice and rats. Using this expression we have investigated where the information for cortical regionalization resides. A series of experiments using in vitro explants show that signals intrinsic to the cortical neuroepithelium contain spatial information, extending previous results on somatosensory cortex patterning. In addition, our results show that the thalamus is unlikely to contain patterning information but that it may be involved in coordinating the coherent development of cortex and thalamus.

In collaboration with different laboratories here at NYU, in New York and abroad we continue to work on the SHH-GLI pathway, focusing mostly on their role in neural stem and progenitor cells, as well as on how the early progenitors in the brain, that give rise to the neocortex, acquire regional character.