Our laboratory uses chemical and genetic approaches to explore the molecular mechanisms that underlie embryonic patterning and tumorigenesis. Much of our work has focused on the Hedgehog (Hh) pathway, which contributes to digit specification, neural tube differentiation, hair follicle cycling, and a variety of other developmental processes. Dysregulated Hh pathway activation can also lead to cancer in children and adults, including basal cell carcinoma, medulloblastoma, and rhabdomyosarcoma. Our laboratory has discovered and/or mechanistically characterized several Hedgehog pathway antagonists and agonists, including the cyclopia-inducing natural product cyclopamine, the SANTs (Smoothened ANTagonists), SAG (Smoothened AGonist), purmorphamine, robotnikinin, JK184, and the HPIs (Hh Pathway Inhibitors) (Figure 1). Many of these compounds directly target canonical Hedgehog pathway components, whereas others disrupt cellular processes required for Hh signal transduction. For example, HPI-4 (ciliobrevin A) is the first specific inhibitor of cytoplasmic dyneins.
We have also conducted genome-scale cDNA overexpression, siRNA knockdown, and CRISPR knockout screens to identify new Hh pathway modulators. Among our discoveries is the role of ARHGAP36 in Gli transcription factor activation. This atypical member of the Rho GAP family regulates motor neuron specification, and it is dysregulated in certain subtypes of medulloblastoma and basal cell carcinoma. ARHGAP36 acts at least in part by suppressing protein kinase A (PKA), a negative regulator of Gli proteins (Figure 2), and we have used high-throughput mutagenesis and mass spectrometry-based proteomics to study the molecular mechanisms that regulate and transduce ARHGAP36 activity.
More recently, we identified first-in-class inhibitors of aldehyde dehydrogenase 1B1 (ALDH1B1), a mitochondrial enzyme that is upregulated in intestinal and pancreatic progenitors and promotes the oncogenic transformation of these tissues (Figure 3). We have used these chemical tools and CRISPR mutagenesis to substantiate a role for ALDH1B1 in colorectal cancer growth, obtaining evidence that this enzyme is required for the survival of stem-like cell populations. We also solved the first x-ray crystal structures of ALDH1B1 to uncover the molecular basis for inhibitor binding, and we employed chemical and genetic perturbations to elucidate the transcriptional programs that are activated by ALDH1B1. We are now investigating the mechanisms by which ALDH1B1 promotes tumor growth and developing ALDH1B1 antagonists with optimized potency, selectivity, and pharmacological properties. In addition to these studies, we are pursuing small-molecule modulators of other ALDH1 isoforms that promote cancer stem cell survival.
While many aspects of developmental signaling can be studied in cultured cells, model organisms are required to gain a comprehensive understanding of how these pathways regulate embryonic patterning, tissue regeneration, and tumorigenesis. Our laboratory uses the zebrafish as a model of vertebrate biology, taking advantage of its facile aquaculture and husbandry, rapid development, optical transparency during embryogenesis, compatibility with forward-genetic approaches, and accessibility to chemical perturbations. Realizing the full potential of zebrafish in biomedical research, however, will require new technologies for manipulating and visualizing the genetic programs within this model organism.
Toward this goal, we have developed caged morpholinos that enable light-or enzyme-controlled gene silencing in zebrafish embryos (Figure 4). With these reagents, we can inhibit in vivo gene function with spatial and temporal precision, creating the chemical equivalent of genetically mosaic organisms. We have also integrated caged morpholinos (MOs), photoactivatable fluorophores, fluorescence-activated cell sorting, and RNA profiling technologies to investigate how transcription factors dynamically regulate embryonic patterning (Figure 5). Our findings have demonstrated a remarkable degree of functional plasticity during No tail-a (Ntla)-dependent notochord development and elucidated the mechanisms by which Ntla and T-box gene 16 (Tbx16/Spadetail) regulate medial floor plate formation. We have also characterized the Tbx16-dependent transcriptome that drives paraxial mesoderm development, revealing an unexpected role for this transcription factor in hox gene activation (Figure 6).
Optogenetic tools are an emerging interest in our laboratory, particularly those that enable light-dependent of developmental signaling pathways. Our work in this area has primarily focused on light-oxygen-voltage (LOV) domains, compact flavin-binding photoreceptors that undergo conformational changes in response to blue light. For example, we have used a photoactivatable transcription factor (GAVPO) to control gene expression and cell ablation in zebrafish embryos (Figure 7). GAVPO is composed of the VIVID LOV domain, GAL4 DNA-binding domain, and p65 activation domain, and it forms a transcriptionally active dimer in a light-dependent manner. We are currently using LOV domains from various microbial and plant species to engineer new optogenetic variants of tissue-patterning genes.
Finally, we have developed methods for imaging photoluminescent lanthanide chelates in vivo. These metal complexes have large Stokes shifts, narrow emission lines, and long-lived luminescence that can be differentiated from background autofluorescence. Lanthanide-based probes are widely used for ultrasensitive photometric detection (e.g., in microplate assays); however, their application in optical microscopy has been relatively limited. In collaboration with Pehr Harbury, we have integrated Q-switched laser (QSL) transreflected illumination, luminescence resonance energy transfer, and time-resolved microscopy to achieve ultrasensitive lanthanide imaging (Figure 8). This new approach (QSL trLRET) is compatible with both live and fixed zebrafish embryos, and our laboratory is now developing lanthanide-based probes for visualizing vertebrate development at the molecular level.
Our research group is also interested in the germ cell biology that precedes embryonic development. We recently identified homeodomain-interacting protein kinase 4 (HIPK4) as an essential driver of spermiogenesis, the process by which spermatids differentiate into mature spermatozoa. Hipk4 knockout male mice are completely sterile but otherwise appear have normal physiology. Our studies show that HIPK4 is predominantly expressed in round and elongating spermatids and required for spermatid head shaping. This remarkable morphological transformation involves external Sertoli cell-generated forces and the acroplaxome, an F-actin-scaffolded plate that circumscribes the anterior pole of spermatid nucleus (Figure 9). Our findings support a role for HIPK4 in remodeling the acroplaxome during spermatid differentiation, and we are now seeking to identify the direct substrates and downstream effectors of this kinase. We are concurrently pursuing the development of HIPK4-specific inhibitors, with the hope that these small-molecule probes could lead to new strategies for reversible, non-hormonal male contraception.