sign in sign up
<<
Home , Sonic hedgehog

Journal of Biology BioMed Central

Research news Agonizing Jonathan B Weitzman

Published: 6 November 2002 Journal of Biology 2002, 1:7 The electronic version of this article is the complete one and can be found online at http://jbiol.com/content/1/2/7 © 2002 BioMed Central Ltd ISSN 1475-4924

An approach using ‘chemical genetics’ has identified small-molecule agonists of the Hedgehog signaling pathway that may lead the way to drugs for chronic degenerative diseases.

It is a rare treat when a drug discovery , and it was subsequently events in a range of developing tissues program teaches us something about found to control patterning of struc- (see the ‘Background’ box) [2,3]. the biology of the process that it tures such as the eye and the abdominal Recently, signaling by Hh has been attempts to modulate. The paper by cuticle. In mammals there are three hh shown to be important for patterning Maria Frank-Kamenetsky and col- homologs, called , of the , where it promotes leagues in this issue of the Journal of Indian Hedgehog and Desert Hedgehog the proliferation of granule neuron Biology [1] presents a compelling (Shh, Ihh and Dhh, respectively), which precursors. A link between Hh proteins example of how the search for thera- have been implicated in patterning and stem- has raised peutics can provide powerful experi- mental tools and insights into fundamental biology, blurring the dis- tinction between applied and basic The bottom line research. By characterizing a small group of chemically similar agonists of • Frank-Kamenetsky and colleagues designed a cell-based, high-throughput the Hedgehog signaling pathway, assay to screen 140,000 compounds to find modulators of the Hedgehog Frank-Kamenetsky et al. have been able signaling pathway. to propose a new model for how the component of the • They identified a small group of related synthetic non-peptidyl molecules Hedgehog- complex works, that can act as agonists of Hedgehog signals at nanomolar concentra- and to hint at the existence of natural- tions, having previously identified small-molecule Hedgehog antagonists. ligand agonists of the Hedgehog signaling pathway (see ‘The bottom • A range of in vitro and in vivo assays were used to show that the line’ box for a summary of their work). agonists can be used as drugs to overcome Hedgehog-signaling defects and to promote cell proliferation. Hedgehog history Signaling by the Hedgehog (Hh) • The action of the agonist compounds is independent of the Hedgehog family of secreted proteins plays a ligand and the inhibitory receptor Patched. Further characterization central role in regulating cell differenti- revealed that the agonists bind directly to the receptor Smoothened. ation and tissue patterning during development [2]. The hedgehog gene • These data give new insights into the nature of Hedgehog-Smoothened (hh) was first identified by virtue of its signaling and raise the possibility of analogous endogenous modulators role in the specification of positional of Hh signaling. identity during embryonic

Journal of Biology 2002, 1:7 7.2 Journal of Biology 2002, Volume 1, Issue 2, Article 7 Weitzman http://jbiol.com/content/1/2/7

the enticing possibility that modulating a proto-oncogene and activates signal- (Johns Hopkins University School of Hh signaling might be relevant for the ing in response to Hh ligand. The Smo Medicine, Maryland, USA). “The previ- clinical management of certain degener- protein is a seven-transmembrane ous models are not tenable,” and he ative diseases. Indeed, it was recently receptor that resembles conventional G- suggests that alternative models of demonstrated that Shh might be effec- protein-coupled receptors (Figure 1a). receptor function must be considered. tive in treating peripheral nerve damage It appears that Ptc inhibits Smo, or degenerative brain disorders such as although the precise mechanisms are The power of chemical Parkinson’s disease [4,5]. Perhaps not unclear. Hh stimulation relieves Smo genetics surprisingly given its role in develop- from inhibition by Ptc, leading to the Frank-Kamenetsky et al. [1] chose to ment, misregulated Hh signaling has generation of an intracellular signal use chemical genetics in an attempt to also been implicated in cancer [3]. that culminates in a nuclear transcrip- identify compounds that could inter- Specifically, and basal tional response (Figure 1b). When Ptc fere with the inhibition of Smo by Ptc, cell carcinoma (BCC) are associated is removed the pathway is constitu- or could activate Smo independent of with inappropriate activation of Hh sig- tively ‘on’, independent of the Hh Ptc, or that might act downstream of naling [6,7]. These observations moti- ligand, whereas certain mutations in Smo. In addition to making effective vated Frank-Kamenetsky and colleagues Smo can activate Hh signaling, bypass- drug candidates, these molecules might to search for small-molecule modula- ing Ptc regulation. A heteromeric also help to illuminate the complex tors of the Hh pathway, with the hope receptor model has been proposed, in mechanisms underlying signaling by that antagonists and agonists might be which Hh interacts directly with Ptc Hh, Ptc, and Smo (see the ‘Behind the used as drugs to treat proliferative or and thereby affects the interaction of scenes’ box for more of the background degenerative diseases, and that small Ptc with Smo [8]. “But things look to the work). molecules might prove more amenable much more complicated than we had The high-throughput screen was to pharmacological delivery than the earlier thought,” says Philip Beachy elegantly simple. First, Frank-Kamenetsky Hh-family proteins themselves.

Sending the signal Hedgehog signaling challenges the Background way we normally think about pathways. Biology is full • Mammals have three Hedgehog (Hh) proteins (Sonic Hedgehog, of examples of extracellular ligands Indian Hedgehog and Desert Hedgehog) that are processed to that bind to specific cell-surface recep- generate functional extracellular ligands. Two receptors are involved in tors, initiating a cascade of biochemi- Hh signaling: Patched (Ptc) is a negative regulator of the Hh-triggered cal events (often involving protein signaling pathway and Smoothened (Smo) is a positive activator kinases) that leads to the activation of (Hh, Ptc and Smo were all originally named for the cuticular pheno- a and the induc- types of mutant Drosophila larvae). When Hh ligands bind to Ptc they tion of a set of effector genes. But relieve the negative inhibition and Smoothened initiates a signaling Hedgehog signaling is not so simple. cascade that results in the activation of nuclear transcription factors of Even the ligand is complicated: the the Gli family that regulate effector gene transcription (see Figure 1). Hh proteins undergo unusual process- ing and cleavage to generate an extra- • Hh signaling has been implicated in a wide range of developmental cellular -linked peptide processes. Small-molecule modulators of Hh signaling are that serves as the signaling ligand [2]. potential candidates as therapeutic drugs to treat human diseases – And the receptor component is far molecules that mimic Hh signaling (agonists) might be used to treat from understood. degenerative disorders such as Parkinson’s disease, whereas The cellular response to Hh is con- blockers (antagonists) could be used as drugs against certain types trolled by two transmembrane pro- of cancer. teins, Patched (Ptc) and Smoothened (Smo). The Ptc protein weaves across • Chemical genetics uses synthetic small molecules to dissect cellular the cell membrane twelve times and functions such as signal transduction pathways. By analogy with loss- resembles some transmembrane chan- and gain-of-function mutations in genetics, the functional interactions nels. It acts as a negative regulator of between chemical inducers and inhibitors is used to define the the Hh signal and has been defined as hierarchical relationship between protein components of the pathway. a tumor-suppressor. In contrast, Smo is

Journal of Biology 2002, 1:7 http://jbiol.com/content/1/2/7 Journal of Biology 2002, Volume 1, Issue 2, Article 7 Weitzman 7.3

(a) (b) researchers then turned to an in vivo Hedgehog ligands model, feeding the compounds to pregnant mice and following the Dhh Shh Ihh effects on the of embryos lacking Shh or Smo. The treated embryos displayed activated Hh sig- naling, demonstrating that the com- pounds were not toxic and could cross the placental barrier. The devel- Membrane opmental defects of Shh-/- embryos were rescued by treatment with the Cytoplasm agonist but the compound had no Patched Smoothened effect on Hh signaling in the absence of Smo. Nucleus “Once we knew that the agonists were targeting Smo, we wanted to Gli investigate whether they bound to Smo directly and how they activated Effector gene Hh signaling,” says Porter. The cell line that was created for the screen served Figure 1 as a useful tool to test the effects of The Hedgehog signaling pathway. (a) In the absence of Hedgehog ligands, Patched protein inhibits known antagonists on the function of the activity of Smoothened, which resembles G-protein-coupled receptors. (b) On activation by a the agonist. An anti-Hh blocking anti- ligand of the Hedgehog family, the inhibition of Smoothened by Patched is relieved and Smoothened body had no effect, so the agonist is freed to trigger an intracellular signaling cascade; this ultimately leads to transcriptional regulation by transcription factors of the Gli family. must work downstream of the Hh-Ptc interaction. But the agonist was blocked by antagonists that work at the level of Smo or further down- et al. identified a cell line that responds effective at eliciting a cellular response, stream. “We have similar conclusions,” well to Hh stimulation. The introduc- affecting cells when applied in the says Beachy, whose group used photo- tion of a reporter gene (encoding the nanomolar range. “Getting more affinity labeling and cross-linking firefly luciferase protein) that was potent compounds was essential if we experiments to show that small- turned on by Hh signaling permitted were to figure out where the agonists molecule agonists and antagonists screening for compounds that block or were acting” recalls Jefferey Porter who bind directly to Smo. induce signaling by monitoring the headed the team at Curis, Inc. Next, for Frank-Kamenetsky et al., it expression of the luciferase protein Then the cell biologists began was time for some careful biochemistry. (using a simple luminescence test). again, studying the effects of the ago- Analysis of the expression of fusion They had previously used such a screen nists on the proliferation of primary proteins of Ptc or Smo showed that, to isolate antagonists of Hh signaling, neonatal cerebellar granule neuron unlike the Hh ligand itself, the agonist and had demonstrated the effective- precursors. They monitored the incor- had no effect on the stability of the Ptc ness of some antagonists as potential poration of tritiated thymidine into protein. In contrast, both Hh and the anti-tumor drugs to treat BCC [7]. cultured rat neurons (as a marker of agonist could increase the stability of Their new screen of 140,000 synthetic DNA synthesis and hence prolifera- the Smo receptor. Immunoprecipita- compounds led to the discovery of a tion) and were pleased to see that the tion experiments with radiolabeled few candidate agonists that could stim- agonists were as effective in this assay agonist showed that the agonist must ulate the reporter gene and mimic as the Hh protein itself. An assay using bind directly to Smo receptors, and Hh activity. an explant of embryonic neural plate that Hh-signaling inhibitors compete Once the cell-based screen was com- was used to confirm that the agonists with the agonist for binding. Pharmo- pleted, the chemists took over, synthe- could induce dose-dependent gene kinetic analysis provided evidence for sizing 300 derivative molecules until expression in neural precursors, just as a single binding site competition they found a few compounds that were the Shh protein does. model. Finally, Frank-Kamenetsky et related to the previous ‘best’ agonist Having established the effects of al. exploited an oncogenic, constitu- but that were a thousand times more the agonists in culture assays, the tively active mutant form of Smo

Journal of Biology 2002, 1:7 7.4 Journal of Biology 2002, Volume 1, Issue 2, Article 7 Weitzman http://jbiol.com/content/1/2/7

(Smoact); the agonists bound equally According to this model, active and or antagonists at independent sites. well to mutant and wild-type forms, inactive conformations of Smo are Furthermore, the model predicts that whereas the antagonist bound less well selected by the binding of agonists Smo binds to a novel effector molecule to the mutant form.

Learning lessons from drugs Frank-Kamenetsky et al. have demon- Behind the scenes strated convincingly that high-through- put chemical genetics can be used to Journal of Biology asked Jeff Porter, group leader at Curis Inc., to comment isolate modulators of a developmental on the background to the project to search for small-molecule modulators signaling pathway. The agonists that of Hedgehog signaling. they have generated are efficient and apparently non-toxic mimics of Hh What prompted you to set up a screen for agonists of the signals and are promising candidates Hedgehog pathway? for drugs for regenerative medicine. There was evidence that manipulating Hedgehog signaling might be useful The authors work at the biotechnology in a therapeutic context to treat degenerative neurological diseases. There company Curis Inc. (Cambridge, were promising data using modified Hedgehog ligands in animal disease USA), so finding new drugs is obvi- models, but we felt that a would be a more effective thera- ously their primary objective. But the peutic. We had previously set up the cell-based assay to screen for antago- agonists also provide useful tools for nists of Hedgehog signaling and successfully isolated inhibitors that could probing the complex Hh-Ptc-Smo inhibit tumor growth. So we adapted the assay to find Hedgehog agonists. signaling pathway. “A lot of our biological insight is What was your initial reaction to the results, and how were they driven by having specific chemical received by others? inhibitors, and many drugs are used This was a kind of ‘black box’ screen - we were looking for a change in a as tools to dissect signaling systems as biological readout, in contrast to traditional pharmacological screens that a substitute for genetic studies,” says focus on a particular target. So, we couldn’t be sure what type of molecule Arnon Rosenthal (Rinat Neuroscience we would get; we could have predicted that we’d find inhibitors of Corp., Palo Alto, USA). Beachy agrees: Patched or stabilizers of Gli. I was surprised and excited when we realized “these compounds, first the plant- that our agonists were targeting Smoothened. Once we had nailed down derived inhibitor and the target, there was a lot of interest in these compounds and what they more recently the agonists, have really can teach us about Hedgehog signaling. helped us to understand Smoothened function.” Recent work from Beachy’s How long did the project take? lab [9,10], demonstrating that Ptc We began the screen in late 1999. It took a few months to get the initial suppresses Smo in catalytic manner, compounds, and then we began the process of making slow improve- has led to speculation that Ptc may ments by chemical modification. Without the improvements we couldn’t function as a transporter protein have figured out what was going on. We had to establish the techniques pumping natural small-molecule Smo and acquire the reagents necessary to characterize the agonists in detail. modulators across the cell membrane. The potent compound derivatives were also key to the success of the Rosenthal adds that “good basic cell-free binding assays - they look simple but the binding experiments research always leads to better medi- were very tricky. cines, since the more we understand about the mechanisms operating in What are the next steps? the body, the better able we are to We want to figure out the mechanisms of Smo regulation and signal trans- modulate them rationally.” duction - how Ptc talks to Smo and how Smo talks to Gli. And, of course, The data presented by Frank- we are also testing these molecules as potential therapeutics in animal Kamenetsky et al. [1] led them to models of disorders of the . The preclinical results propose a new model for Hh signaling are promising: the compounds show low toxicity, they can be administered based on the classic ‘ternary complex orally and cross the blood-brain barrier. These compounds might also be model’ that was developed to describe useful in ex vivo therapies to stimulate stem cell proliferation. I am optimistic ligand-induced conformational changes that Hedgehog agonists will be tested in human trials in the future. of G-protein-coupled receptors [11].

Journal of Biology 2002, 1:7 http://jbiol.com/content/1/2/7 Journal of Biology 2002, Volume 1, Issue 2, Article 7 Weitzman 7.5

and it raises the possibility that excitement of ‘real’ scientific discovery Chen JK, Cooper MK, Taipale J, Olson JM, Beachy PA: Medulloblastoma endogenous ligands, analogous to the with the satisfaction of isolating growth inhibition by Hedgehog newly found agonist, may naturally compounds that might be used for pathway blockade. Science 2002, regulate Smo activity. tomorrow’s therapies. 297:1559-1561. 7. Williams JA, Guicherit OM, Zaharian BI, Cheryll Tickle (University of Dundee, Xu Y, Chai L, Gatchalian C, Porter JA, UK) studies the role of Hh in develop- References Rubin LL, Wang FY: Identification of ment and finds the possibility that there 1. Frank-Kamenetsky M, Zhang XM, novel inhibitors of the hedgehog sig- Bottega S, Guicherit O, Wichterle H, naling pathway: effects on basal cell are endogenous small-molecule agonists Dudek H, Bumcrot D, Wang FY, Jones F, carcinoma-like lesions. Proc Natl Acad that interact with Smoothened particu- Shulok J, Rubin LL Porter JA: Small Sci USA, in press. larly intriguing. “This would mean that molecule modulators of Hedgehog 8. Stone DM, Hynes M, Armanini M, signaling: identification and charac- Swanson TA, Gu Q, Johnson RL, Scott we are missing a whole layer of control terization of Smoothened agonists MP, Pennica D, Goddard A, Phillips H, et of Hedgehog signaling in our current and antagonists. J Biol 2002, 1:10. al.: The tumor-suppressor gene models,” she notes. Porter adds, “It is 2. Ingham PW, McMahon AP: Hedgehog patched encodes a candidate recep- signaling in animal development; tor for Sonic Hedgehog. Nature 1996, tempting to speculate that endogenous paradigms and principles. Genes Dev 384:129-134. molecules act directly on Smo, bypass- 2001, 15:3059-3087. 9. Chen JK, Taipale J, Young KE, Maiti T, 3. Taipale J, Beachy PA: The Hedgehog Beachy PA: Small molecule modula- ing Ptc. That is consistent with what we and Wnt signaling pathways in tion of Smoothened activity. Proc know about how G-protein-coupled cancer. Nature 2001, 411:349-354. Natl Acad Sci USA 2002, 99:14071-14076. receptors work. But at the moment, it’s 4. Pepinsky RB, Shapiro RI, Wang S, 10. Taipale J, Cooper MK, Maiti T, Beachy Chakraborty A, Gill A, Lepage DJ, Wen PA: Patched acts catalytically to sup- pure speculation.” D, Rayhorn P, Horan GS, Taylor FR, et al.: press the activity of Smoothened. The study by Frank-Kamenetsky et Long-acting forms of Sonic hedge- Nature 2002, 418:892-896. al. [1] exploits a dazzling variety of hog with improved pharmocokinetic 11. De Lean A, Stadel JM, Lefkowitz RJ: A and pharmacodynamic properties ternary complex model explains the experimental techniques to illustrate are efficacious in a nerve injury agonist-specific binding properties the path from high-throughput screen- model. J Pharm Sci 2002, 91:371-387. of the adenylate cyclase-coupled ing to compound characterization. 5. Tsuboi K, Shults CW: Intrastriatal beta-adrenergic receptor. J Biol Chem injection of sonic hedgehog reduces 1980, 255:7108-7117. Biochemists, cell biologists, pharm- behavioural impairment in a rat acologists and chemists have come model of Parkinson’s disease. Exp together to demonstrate effectively that Neurol 2002, 173:95-104. Jonathan B Weitzman is a scientist and science 6. Berman DM, Karhadkar SS, Hallahan AR, writer based in Paris, France. ‘drug discovery’ can combine all the Pritchard JI, Eberhart CG, Watkins DN, E-mail: [email protected]

Journal of Biology 2002, 1:7