A Potent Antagonist of Smoothened in Hedgehog Signaling for Epilepsy

Epilepsy is one of the common encephalopathies caused by sudden abnormal discharges of neurons in the brain. About 30% of patients with epilepsy are insensitive and refractory to existing antiseizure medications. The sonic hedgehog signaling pathway is essential to the development and homeostasis of brain. Aberrant sonic hedgehog signaling is increased in refractory epileptic lesions and may involve the etiology of epilepsy. Thus, new inhibitors of Smoothened, a key signal transducer of this signaling pathway are urgently need for refractory epilepsy. We have established a high-throughput screening platform and discovered several active small molecules targeting Smoothened including TT22. Here we show that the novel Smoothened inhibitor TT22 could block the translocation of βarrestin2-GFP to Smoothened, reduce the accumulation of Smoothened on primary cilia, displace Bodipy-cyclopamine binding to Smoothened, and inhibit the expression of downstream Gli transcription factor. Moreover, TT22 inhibits the abnormal seizure-like activity in neurons. Furthermore, we demonstrated that FDA-approved Smoothened inhibitor GDC-0449 and LDE-225 are able to inhibit abnormal seizure-like activity in neurons. Thus, our study suggests that targeting the sonic hedgehog signaling with new small-molecule Smoothened inhibitors might provide a potential new therapeutic avenue for refractory epilepsy.


Introduction
The evolutionarily conserved sonic hedgehog (Shh) signaling pathways play an important role in the development and homeostasis of the central nervous system and Smoothened (Smo) protein is an essential receptor in the Shh signaling pathway [1,2]. The Shh pathway is initiated by the binding of Shh ligands to the twelve-transmembrane receptor patched on the cell membrane, releasing its inhibitory effect on the seven-transmembrane Smo receptor. Activated Smo initiates the translocation of downstream Gli transcription factor to the nucleus and regulates target gene expression [3,4]. It has been reported that the abnormal activation of Shh signaling pathway plays an important role in the development of many types of cancers, including basal cell carcinoma (BCC), medulloblastoma (MB), and other solid tumors [5,6]. Smo antagonists GDC-0449/vismodegib and LDE-225/sonidegib were approved as drugs for BCC and MB in 2012 and 2015, respectively [7][8][9].
Our previous studies have shown that constitutively activated Smo has the ability to recruit β-arrrestin2 proteins to the plasma membrane [2].  Figure 1C). The aggregation of βarr2-GFP could be restored again when cyclopamine and a well-known Smo agonist SAG were present at the same time ( Figure 1D). The green aggregates are quantitated in Figure 1E.
Accordingly, we established a high-throughput screening method for Smo antagonists, screened a small molecule compound library and discovered that several small molecules including TT22 could inhibit the formation of intracellular βarr2-GFP aggregates ( Figure 2A). The IC 50 of TT22, LDE-225 and GDC-0449 were 4.75 nM, 26.78 nM, 44.55 nM, respectively ( Figure 2B).   Accordingly, we established a high-throughput screening method for Smo antagonists, screened a small molecule compound library and discovered that several small molecules including TT22 could inhibit the formation of intracellular βarr2-GFP aggregates ( Figure 2A). The IC50 of TT22, LDE-225 and GDC-0449 were 4.75 nM, 26.78 nM, 44.55 nM, respectively ( Figure 2B). To further characterize TT22, we demonstrated that TT22 can block the translocation of βarr2-GFP using cyclopamine as a control ( Figure 3A-C). The uniform distribution of βarr2-GFP induced by TT22 in cytoplasm could be reversed by SAG, which further confirmed the TT22 as a Smo inhibitor ( Figure 3D). The structure of TT22 is shown in Figure 4A. It was synthesized through the route described in Figure 4B. To further characterize TT22, we demonstrated that TT22 can block the translocation of βarr2-GFP using cyclopamine as a control ( Figure 3A-C). The uniform distribution of βarr2-GFP induced by TT22 in cytoplasm could be reversed by SAG, which further confirmed the TT22 as a Smo inhibitor ( Figure 3D). The structure of TT22 is shown in Figure  4A. It was synthesized through the route described in Figure 4B.

TT22 Is a Competitive Antagonist of Smo
We investigated whether TT22 could compete with Bodipy-cyclopamine, which has been used to assess the binding affinity of Smo ligand, to bind Smo [25]. We found that the known Smo antagonists GDC-0449 and TT22 could block the binding of 5 nM Bodipy-cyclopamine to Smo in a dose-dependent manner ( Figure 5). These results indicate that TT22 can compete with Bodipy-cyclopamine to bind to the Smo receptor.

TT22 Is a Competitive Antagonist of Smo
We investigated whether TT22 could compete with Bodipy-cyclopamine, which has been used to assess the binding affinity of Smo ligand, to bind Smo [25]. We found that

TT22 Is a Competitive Antagonist of Smo
We investigated whether TT22 could compete with Bodipy-cyclopamine, which has been used to assess the binding affinity of Smo ligand, to bind Smo [25]. We found that the known Smo antagonists GDC-0449 and TT22 could block the binding of 5 nM Bodipy-cyclopamine to Smo in a dose-dependent manner ( Figure 5). These results indicate that TT22 can compete with Bodipy-cyclopamine to bind to the Smo receptor.

TT22 Blocks the Aggregation of Smo on the Primary Cilia
One of the key steps to activate the Shh pathway is the localization and aggregation of Smo on the primary cilia, so we evaluated the effect of TT22 on the localization of Smo using methods as described previously [26][27][28]. We used DMSO, 5 µM GDC-0449, and 5 µM TT22 treated PTCH −/− MEF cells for 24 h, and then immunofluorescence staining with anti-Smoothened and ARL13B antibodies. As expected, compared with the DMSO control group, TT22 and GDC-0449 could inhibit the aggregation of Smo on the primary cilia ( Figure 6).

TT22 Blocks the Aggregation of Smo on the Primary Cilia
One of the key steps to activate the Shh pathway is the localization and aggregation of Smo on the primary cilia, so we evaluated the effect of TT22 on the localization of Smo using methods as described previously [26][27][28]. We used DMSO, 5 μM GDC-0449, and 5 μM TT22 treated PTCH -/-MEF cells for 24 h, and then immunofluorescence staining with anti-Smoothened and ARL13B antibodies. As expected, compared with the DMSO control group, TT22 and GDC-0449 could inhibit the aggregation of Smo on the primary cilia ( Figure 6).

TT22 Inhibits Shh Signaling
Gli1 is the downstream target gene of the Shh signaling pathway [29]. In order to assess the inhibitory effect of TT22 on the downstream of Shh, NIH3T3 cells were treated with N-Shh conditioned medium (Shh-CM) or Smo agonist SAG to stimulate the expres-

TT22 Inhibits Shh Signaling
Gli1 is the downstream target gene of the Shh signaling pathway [29]. In order to assess the inhibitory effect of TT22 on the downstream of Shh, NIH3T3 cells were treated with N-Shh conditioned medium (Shh-CM) or Smo agonist SAG to stimulate the expression of Gli1. We firstly prepared Shh-CM and explored the optimal effective concentration of Shh-CM and SAG for Gli expression through a gradient concentration treatment, which was 20% of CM and 100 nM SAG (Figures 7A and 8A). We observed that 20% Shh-CM or 100 nM SAG stimulated NIH3T3 cells treated with TT22 (0.01, 0.  rum-starved NIH3T3 cells were treated with Shh-CM for 24 h, then analyzed for mRNA levels of Gli1. (B,C) Serum-starved NIH3T3 cells were induced by Shh-CM with DMSO (vehicle), with an indicated concentration of TT22, 1 μM LDE-225 (LDE-1) or 1 μM GDC-0449 (GDC-1) for 24 h. Cells were analyzed for mRNA levels (B) and the protein levels (C) of Gli1. (D) Quantitation of Gli1 protein levels normalized to Tubulin loading control measured the dose-dependent inhibition of the Hh pathway activity by TT22 upon Shh stimulation. All data are means ± SEM (Student's t-test and one-way ANOVA). *** p < 0.001.

TT22 Inhibits the Abnormal Seizure-like Activity in Cultured Neurons
To evaluate the effect of Smo inhibitors on electrophysiological responses of neurons, we used proconvulsant solution containing Mg 2+ -free/K + -high (0 Mg 2+ /8 K + ) to record seizure-like activity. The experiment scheme is shown in Figure 9A. The treatment of TT22, GDC-0449, and LDE-225 did not affect the resting membrane potential of neurons as shown ( Figure 9B). Neurons treated with indicated small molecules and the seizure-like activity are measured using methods as described in the Materials and Methods section. TT22, GDC-0449, and LDE-225 inhibited the seizure-like activity ( Figure 9C). TT22 as well as GDC-0449 and LDE-225 can reduce the frequency of seizure-like activity ( Figure 9D). Furthermore, the recorded data of 10-20 min drug treatment showed that TT22 with a concentration of 1 µM or 10 µM as well as GDC-0449, and LDE-225 treatment inhibited the seizure-like activity ( Figure 9E). These results indicate that TT22, GDC-0449, and LDE-225 could effectively inhibit abnormal seizure-like activity in neurons by suppressing Shh signaling. The number of seizure-like activities between the 10th and 20th minute was quantitated. All data are means ± SEM (one-way ANOVA and two-way ANOVA). * p < 0.05 ** p < 0.01 *** p < 0.001.

Discussion
Epilepsy is one of the common chronic recurrent diseases of the brain caused by highly synchronized discharges of neurons [30]. The World Health Organization reports that there are about 70 million patients with epilepsy all over the world [13]. Most of antiseizure medications (ASMs) are mainly anticonvulsant, which directly act on the ion channel to rapidly reduce the excitability of the neurons during attack to control the seizure [31]. ASMs widely used are phenobarbital, phenytoin, and ethosuximide which The number of seizure-like activities between the 10th and 20th minute was quantitated. All data are means ± SEM (one-way ANOVA and two-way ANOVA). * p < 0.05 ** p < 0.01 *** p < 0.001.

Discussion
Epilepsy is one of the common chronic recurrent diseases of the brain caused by highly synchronized discharges of neurons [30]. The World Health Organization reports that there are about 70 million patients with epilepsy all over the world [13]. Most of antiseizure medications (ASMs) are mainly anticonvulsant, which directly act on the ion channel to rapidly reduce the excitability of the neurons during attack to control the seizure [31]. ASMs widely used are phenobarbital, phenytoin, and ethosuximide which belong to three categories through inhibiting high-frequency neuronal discharges by blocking voltage-dependent sodium channels, blocking T-type calcium channels (T-type VGCC), and enhancing γ-aminobutyric acid type A receptor (GABA A receptor) mediated postsynaptic inhibition [32]. However, about 30% of patients are insensitive to the existing drug treatment, which is referred to as refractory epilepsy. Only surgery is suitable for refractory epilepsy patients [14]. There is a medical need for the treatment of refractory epilepsy, and finding a new treatment for it is still a challenge.
Shh signaling is important for brain development and homeostasis. Its activity is abnormally increased in advanced BCC, MB, and other cancers as well as refractory temporal lobe epilepsy patients [33,34]. For patients with refractory epilepsy, it has been reported that Shh expression increased in the epileptogenic zone [20]. The abnormal increase of neuronal activity leads to the release of Shh, and regulates glutamate-mediated functions which may contribute to the occurrence and development of refractory epilepsy [24]. These data indicate that Smo in Shh signaling could be a new therapeutic target for refractory epilepsy.
We discovered that the small molecule TT22 obtained from the high-throughput screening can inhibit Shh signaling. Furthermore, TT22, GDC-0449, and LDE-225 can effectively inhibit seizure-like activity in neurons and significantly reduce the frequency and increase the interictal stage. It has been reported that the effects of Smo inhibitors on inhibition abnormal epileptiform discharge in neurons within 30 min of compound treatments suggests that this process is mediated by noncanonical Shh signaling [35,36]. In the postnatal rodent hippocampus, the noncanonical Shh signaling plays an important role during early postnatal neuronal circuit construction and synaptic plasticity involving intracellular Ca 2+ signaling and the BDNF-TrkB signaling pathway as well as has a role in regulating GABAergic transmission [37,38]. The underlying mechanisms will be intriguing to study in the future.
Previously we have identified 0025A as a novel Smo inhibitor. The IC50 of TT22 and 0025A are 4.75 nM and 1.7 nM, respectively. However, the TT22 displays more activities for the inhibition of abnormal seizure-like activity in neurons. Further studies of the structure-activity relationship (SAR) analysis are needed to understand the mechanism of action of drugs (MOA) for clinical development between these two compounds. In addition, the therapeutic effect of TT22 as well as GDC-0449 and LDE-225 on epilepsy should be further studied in vivo to explore whether these inhibitors decrease epilepsy symptoms. Our study provides a potential new treatment by targeting the Shh signaling pathway associated with refractory epilepsy with a novel Smo inhibitor TT22 which can effectively inhibit seizure-like activity in cultured neurons. We will evaluate the effect of TT22 on the epileptiform discharges and epileptic activities in vitro and on animal models in future studies. Overall, TT22 as well as FDA-approved Smo inhibitors (GDC-0449 and LDE-225) could be developed as new antiseizure medications through targeting Shh signaling pathways associated with refractory epilepsy. These studies could provide a basis for further clinical trials and may provide a new therapy for refractory epilepsy.
In summary, we identify TT22 as a novel antagonist of Smo. Most importantly, TT22 as well as GDC-0449 and LDE-225 effectively inhibit seizure-like activity ( Figure 10). These findings provide a new avenue for targeting Shh signaling associated refractory epilepsy and lay out a basis for future clinical trials.

N-Shh Conditioned Medium
The pRK5-ShhN plasmid was transfected into 293FT cells using Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA). Twenty-four hours later, the medium was replaced with fresh DMEM supplemented with 2% FBS. The culture medium was harvested at 24 h and 48 h after the medium change, then combined and centrifuged for 10 min at 1000 rpm. The supernatant was considered as N-Shh conditioned medium (Shh-CM) and stored at −20 °C until used.

Cell Culture, Plasmids, and Western Blotting
The 293FT, NIH3T3 and U2OS cells were obtained from Dr. Zhiheng Xu. The PTCH−/− MEF cells were provided by Dr. Steven Y. Cheng. The U2OS, NIH3T3, 293FT and PTCH−/− MEF cells were cultured in DMEM (Gibco, Waltham, MA, USA) containing a 10% fetal bovine serum (Hyclone, Logan, UT, USA). To detect the effect of TT22 on Gli1 expression, NIH3T3 cells were seeded into a 12-well plate at the density of 2 × 10 5 per well and cultured at 37 °C. When the cells were at 100% confluence, they were starved in DMEM supplemented with 0.5% FBS for 1 h at 37 ℃, then treated with 20% Shh-CM or 100 nM SAG with different concentrations of indicated compounds for 24 h at 37 °C. After treatment, the cells were lysed by RIPA, and centrifuged at 12,000 rpm for 10 min at 4 °C. Cell lysates were harvested and detected by Western blotting with the following primary antibodies: Gli1 (CST, 2534s, 1:1000, Boston, MA, USA); and Tubulin (CST, 3873 s, 1:2000).

Immunofluorescence Staining
For cell staining, Ptch −/− MEF cells were cultured and seeded on coverslips. Then, 4% (w/v) paraformaldehyde/phosphate-buffered saline (PBS) was used to fix cells at room temperature for 20 min. After fixation, the cells were washed three times with PBS, blocked with 5% (w/v) bovine serum albumin (BSA)/PBS at room temperature for 1 h, and incubated with a primary antibody in 5% BSA overnight at 4 °C. Then, the cells were washed and treated with secondary antibodies conjugated with Alexa Fluor 488 or

N-Shh Conditioned Medium
The pRK5-ShhN plasmid was transfected into 293FT cells using Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA). Twenty-four hours later, the medium was replaced with fresh DMEM supplemented with 2% FBS. The culture medium was harvested at 24 h and 48 h after the medium change, then combined and centrifuged for 10 min at 1000 rpm. The supernatant was considered as N-Shh conditioned medium (Shh-CM) and stored at −20 • C until used.

Cell Culture, Plasmids, and Western Blotting
The 293FT, NIH3T3 and U2OS cells were obtained from Dr. Zhiheng Xu. The PTCH−/− MEF cells were provided by Dr. Steven Y. Cheng. The U2OS, NIH3T3, 293FT and PTCH−/− MEF cells were cultured in DMEM (Gibco, Waltham, MA, USA) containing a 10% fetal bovine serum (Hyclone, Logan, UT, USA). To detect the effect of TT22 on Gli1 expression, NIH3T3 cells were seeded into a 12-well plate at the density of 2 × 10 5 per well and cultured at 37 • C. When the cells were at 100% confluence, they were starved in DMEM supplemented with 0.5% FBS for 1 h at 37°C, then treated with 20% Shh-CM or 100 nM SAG with different concentrations of indicated compounds for 24 h at 37 • C. After treatment, the cells were lysed by RIPA, and centrifuged at 12,000 rpm for 10 min at 4 • C. Cell lysates were harvested and detected by Western blotting with the following primary antibodies: Gli1 (CST, 2534s, 1:1000, Boston, MA, USA); and Tubulin (CST, 3873 s, 1:2000).

Immunofluorescence Staining
For cell staining, Ptch −/− MEF cells were cultured and seeded on coverslips. Then, 4% (w/v) paraformaldehyde/phosphate-buffered saline (PBS) was used to fix cells at room temperature for 20 min. After fixation, the cells were washed three times with PBS, blocked with 5% (w/v) bovine serum albumin (BSA)/PBS at room temperature for 1 h, and incubated with a primary antibody in 5% BSA overnight at 4 • C. Then, the cells were washed and treated with secondary antibodies conjugated with Alexa Fluor 488 or 568 dyes in 5% BSA at room temperature for 1 h. The images were achieved through LSM710 Zeiss confocal microscope and analyzed with Image J software. The following primary antibodies were applied for immunostaining: ARL 13B (Proteintech, Wuhan, China, 17711-1-AP, 1:100); and Smoothened (Santa Cruz, CA, USA, sc-166685, 1:100).

Primary Rat Hippocampal Neuron Culture
The hippocampus of an SD rat brain at around embryonic day 18 was dissected and digested with 0.05% trypsin containing Dnase I (1:200) for 5-10 min. After terminating the digestion, the cell suspension was gently pipetted to single cell as soon as possible, filtered with a 70-mesh cell sieve, and centrifuged at 1000 rpm for 4 min. Then, fresh DMEM medium was added to resuspend the cells and mixed by gentle pipetting. After counting, suspended cells were seeded at a density of 5 × 10 4 /cm 2 onto coverslips (Fisherbrand, Waltham, MA, USA) which had been precoated by poly-D-lysine (20 µg/mL, Sigma-Aldrich, St. Louis, MO, USA) at room temperature for 1 h, and cultured in neurobasal medium supplemented with 10% B-27 and 0.5 mM gluta-max at 37 • C for 10-14 days before use.

RNA Isolation, Reverse Transcription, and Real-Time PCR
Total RNA was isolated using the TRIzol reagent (Thermo, Waltham, MA, USA) according to the manufacturer's recommended procedures, and reverse transcription was performed with PrimeScript™ RT reagent Kit (TaKaRa, Osaka, Japan). Real-time PCR was carried out according to the instructions of the manufacturer. The expression of Gli1 mRNA level was normalized to Actin and determined by real-time PCR according to the 2 −∆∆Ct method. The primers used to amplify specific regions for real-time PCR were mouse Gli1, F: 5 -CTCAAACTGCCCAGCTTA ACCC-3 , R: 5 -TGCGGCTGACTGTGTAAGCAGA-3 ; mouse Actin, F: 5 -GCAAGTGCTTCTAGGCGGAC-3 , and R: 5 -AAGAAAGGGTGTAAA-ACGCAG C-3 .

Bodipy-Cyclopamine Binding Assay
Human flag-tagged Smo WT was transfected into HEK293 cells. The transfected cells were digested by trypsin 24 h later and washed in phenol-red free DMEM containing 0.5% FBS. Then, 4% (w/v) paraformaldehyde/PBS was used to fix cells at room temperature for 10 min. After fixation, the cells were washed with phenol-red free DMEM containing 0.5% FBS and incubated with 5 nM Bodipy-cyclopamine and a range of concentration (10 −9 -10 −5 M) of indicated compounds at 37 • C for 2 h. Following incubation, the cells were washed, and the fluorescent signals analyzed by flow cytometry.
Resting membrane potential (V m ) was determined when I = 0. The proconvulsant solution containing Mg 2+ -free/ K + -high (0 Mg 2+ /8 K + ) was used to record seizure-like activity [39][40][41]. In the current-mode, continuous recordings were performed for 30 min. The seizure-like activity was divided into two forms: the first was a large depolarization shift with ≥10 mV depolarization and ≥300 ms in duration and at least five action potentials; the second was an action potential featured with a significant increase in amplitude above the baseline and high frequency of signals. We carried out the events described, as above, when the cell attained the stable epileptiform state between 10 and 20 min after we had transferred the cells into proconvulsant solution. Cohorts of cells included in the "Vehicle" group were pretreated with an equal dose of vehicle solution (dimethyl sulfoxide, DMSO), while the "Drug" groups were subjected to different drugs which needed to be tested with varied concentrations for 30 min prior to patch recording.

Statistical Analysis
All data were acquired from more than two independent experiments and analyzed using Prism software (GraphPad Software, Inc., San Diego, CA, USA). An unpaired Student's t test was performed for two-sample comparisons. One-way ANOVA was applied for multiple comparisons, followed by a p value adjustment with the Bonferroni method, and a two-way ANOVA was performed for multiple comparisons with two independent variables. All values were expressed as means ± SEM, and * p < 0.05 was judged to be statistically significant.

Conclusions
The new Smoothened inhibitor TT22 as well as the GDC-0449 and LDE-225 may provide a new therapy targeting sonic hedgehog signaling for refractory epilepsy.