Modulation of TGF-beta signaling

Eunice Kennedy Shriver National Institute Of Child Health & Human Development

Investigators

Linked publications, trials & patents

Abstract

Bone morphogenetic proteins (BMPs) are potent secreted signaling factors that function at long- and short-range and elicit critical cellular responses during development and homeostasis. Long-range signaling is key to the formation and function of morphogen gradients which control cell fate and tissue allocation. Short-range signaling sculpts cellular junctions and has been implicated in the growth, development and homeostasis of synaptic junctions, such as Drosophila neuromuscular junction (NMJ). The fly NMJ is a glutamatergic synapse similar in composition and physiology to mammalian central synapses. The fact that individual NMJs can be reproducibly identified and are easily accessible for electrophysiological and optical analysis makes this genetic model system uniquely suited for in vivo studies on NMJ growth and synapse assembly, maturation and plasticity. In flies, BMPs shape NMJ development via multiple signaling pathways. At synaptic terminals, Gbb/ BMP7 binding to BMP receptors on the motor neuron membrane triggers canonical and non-canonical BMP signaling. Gbb is secreted from both muscle (retrograde signaling) and motor neurons (autocrine signaling). During canonical BMP signaling, the high-order BMP/BMPR (Gbb/Wit/Tkv/Sax) complexes are endocytosed and transported to the motor neuron soma where they phosphorylate Mad (Drosophila Smad) to regulate various transcriptional programs required for NMJ growth and function. During non-canonical signaling, Wit, the BMPRII, signals through LIMK1 to regulate synapse stability. Intriguingly, phosphorylated Smad (pMad) also accumulates at synaptic sites but the biological relevance of this phenomenon remained obscure. We found that this synaptic/local BMP signaling requires the BMPRs Wit, Tkv and Sax, but not Gbb. Also, synaptic pMad is selectively lost at synapses with reduced levels of postsynaptic ionotropic glutamate receptors (iGluRs): Mutants that lack a particular of glutamate receptor subtype, GluRIIA-containing type-A receptors, show complete loss of synaptic pMad signals but have normal canonical BMP signaling and normal expression of BMP target genes. More importantly, the accumulation of synaptic pMad followed the activity and not the net levels of GluRIIA-containing (type-A) iGluRs. Consequently, synaptic pMad appears to function as a local sensor for synaptic activity. Our studies indicate that synaptic pMad marks a completely novel BMP pathway that is genetically distinguishable from all other known BMP signaling cascades. Similar to canonical BMP pathways, local BMP signaling requires Mad and BMP receptors (Wit, Put and Sax); unlike canonical pathways, synaptic BMP signaling does not require Gbb/BMP7 but depends on postsynaptic type-A iGluRs. Also, nuclear and synaptic pMad are independently regulated: Genetic manipulations of type-A iGluRs induce proportional changes in synaptic pMad but have no effect on nuclear pMad. Conversely, overexpression of Mad in motor neurons induces strong accumulation of nuclear pMad but has no effect on the synaptic pMad levels. Super-resolution fluorescence microscopy studies indicate that the presynaptic pMad-positive domains distribute into thin discs of roughly 700nm diameter sandwiched in between the presynaptic active zones, the sites of neurotransmitter release and the postsynaptic iGluRs fields. The size and shape of the pMad-positive domains suggest that pMad associates with membrane-anchored complexes at the active zone. Since BMP signals are generally short lived, these domains likely represent pMad that, upon phosphorylation, remains associated with the BMP/BMPR kinase complexes at presynaptic sites. Genetic manipulations that selectively disrupt presynaptic pMad induce a reduction of type-A postsynaptic iGluRs, indicating that synaptic pMad functions to stabilize active type-A iGluRs at synaptic locations. Since motor neurons have limited levels of BMP receptors that are shared among different BMP signaling modalities, the neurons may use the local BMP signaling to monitor synapse activity then coordinate NMJ growth with synapse maturation and stabilization. How do postsynaptic glutamate receptors modulate presynaptic pMad and in turn are stabilized by it? Since synaptic pMad depends on active type-A iGluRs, we favor a model whereby Neto, an obligatory subunit of the iGluR channels, connects active postsynaptic type-A iGluRs with presynaptic BMP/BMPR complexes. Such trans-synaptic columns could offer a versatile means for relaying synapse activity status to the presynaptic neuron via fast conformational modifications. At the same time, these trans-synaptic nanocolumns may function as accumulation centers, holding the active type-A receptors at synaptic sites and facilitating interactions that stabilize the type-A receptors at postsynaptic densities. This positive feedback mechanism provides a molecular basis for understanding several key steps during synapse assembly, maturation and plasticity. For example, this mechanism can explain the positive feedback that promotes incorporation of type-A receptors at new synapses, then limits the type-A receptors accumulation during synapse maturation. This positive feedback mechanism can also explain the Hebbian mode of GluRIIA incorporation at PSDs: Local stimulation increases postsynaptic sensitivity by promoting synapse-specific recruitment/stabilization of type-A receptors, whereas type-A receptors are rapidly removed from synapses with reduced activity. At the molecular level, this positive feedback also requires that synaptic pMad remains associated with the BMP/BMPR complexes at the active zones. To learn more about the structural features of Mad that may influence its association with the BMPRs, we have collected most of existing Drosophila Mad alleles and compared them side-by side for their ability to sustain the canonical BMP signaling, marked by pMad accumulation in motor neuron nuclei, and the synaptic BMP signaling, marked by pMad accumulation at synaptic terminals. Within this comprehensive collection, we found that strong Mad alleles generally disrupt both synaptic and nuclear pMad accumulation, whereas moderate Mad alleles have a wider range of phenotypes and can differentially impact different BMP signaling modalities. In particular, Mad8, which has a single point mutation (S359L), showed drastically reduced synaptic pMad levels but only moderately diminished nuclear pMad signals. Through biochemical studies on Mad8 and other alleles, we identified a new molecular determinant for the Mad-Tkv interaction, the highly class conserved H2 helix. Genetic variants within the H2 helix have been uncovered in several patients with neuronal deficits or epithelial abnormalities suggesting that this motif may be critical for local BMP signaling and the integrity of the tight junctions throughout the animal kingdom. To examine how motor neurons (MNs) respond to different BMP signaling modalities or compensate for various synaptic perturbations, we are currently using droplet microfluidics (10x Chromium) methodology to compare MNs transcriptomes in control and various mutants. We have already developed a set of protocols to generate and analyze single cell RNAseq datasets for larval ventral nerve cords (VNCs), the fly equivalent of the spinal cord. We use un-supervised algorithms, reiterative processes and known specific transcripts to cluster these cells and identify different cell types. Based on sequences from more than 30,000 single cells (from multiple experimental sets), we are in the process of assembling an atlas of the larval VNC; it includes several glia subtypes, neuroblast and newborn neurons, motor neurons and interneurons. We will next examine how VNCs, in particular MNs, respond to synaptic perturbations or disruptions of various BMP signaling modalities.