Smoothened Promotes Glioblastoma Radiation Resistance Via Activating USP3-Mediated Claspin Deubiquitination

Glioblastoma (GBM) is one of the most aggressive and lethal human cancer types and is a major health issue worldwide (1, 2). Despite standardized treatment that prolongs survival, the prognosis of GBM patients remains poor, and the median survival time is only approximately 12 months (3, 4). Previous studies have established that GBM typically recurs within the initial radiation treatment course, which indicates that GBM cells are extremely radioresistant (5, 6). Therefore, identifying novel mechanisms underlying radiation resistance in GBM is imperative.

The Hedgehog (Hh) pathway performs crucial functions in cell proliferation, apoptosis, and self-renewal (7, 8). Smoothened (Smo) is a component of the Hh pathway that triggers the downstream activation of Gli transcription factors (9, 10). Gli transcription factors consist of three proteins, Gli1, Gli2, and Gli3, which are responsible for regulating downstream target genes (9).

The ATM–Chk2 and ATR–Chk1 protein kinases make up two main signaling pathways activated by DNA damage (11–13). ATM–Chk2 and ATR–Chk1 protein kinase activation participates in numerous cellular events, including DNA damage recognition and processing, cell-cycle arrest, and apoptosis (12, 14). Once loaded to DNA, Rad17, and Claspin collaborate to fully activate ATR. Claspin also binds to Chk1 and is essential for the ATR phosphorylation of Chk1, which acts as both a sensor and a mediator of DNA replication (11, 12).

Ubiquitination is a vital posttranslational modification. However, deubiquitination, the reverse process mediated by deubiquitinating enzymes (DUB), is an important regulatory mechanism in controlling protein proteasomal degradation (15, 16). The deubiquitination enzyme ubiquitin-specific protease 3 (USP3) has been identified as a chromatin modifier that stabilizes genome integrity (17, 18).

In this study, we identified that Smo mediated radiation resistance in GBMs via promoting USP3 transcription, leading to the activation of Claspin-dependent ATR–Chk1 signaling. Thus, Smo may be a novel potential therapeutic target for GBMs.

Primary GBM cell lines derived from GBM surgical specimens were maintained in primary serum-free cultures grown on laminin. All cell lines were validated by short-tandem repeat DNA fingerprinting using an AmpFlSTR kit (Applied Biosystems), according to the manufacturer's instructions. During experiments, all GBM cell lines were cultured in DMEM supplemented with 10% FBS.

Ten primary and the corresponding recurrent GBM patients who received radiotherapy (each pair was from the same patient) were all obtained from the Department of Neurosurgery, the First Affiliated Hospital of Nanjing Medical University. The histologic features of the tissues were independently examined independently by two neuropathologists according to the WHO criteria. The clinicopathologic and treatment characteristics of these patients are shown in Supplementary Table S1.

To quantify the mRNA expression of Smo, Claspin, and USP3, real-time PCR (qPCR) was carried out using SYBR Green (Applied Biosystems). The forward and reverse PCR primers for Smo were 5′-TGAAGGCTGCACGAATGAGG-3′ and 5′-CTTGGGGTTGTCTGTCCGAA-3′, the primers for Claspin were 5′-TGCCAAAGAAGCTACCGAGT-3′ and 5′-CCAGCATCTCATCCATAACACCT-3′, and the primers for USP3 were 5′-CAAGCTGGGACTGGTACAGAA-3′ and 5′-GCAGTGGTGCTTCCATTTACTT-3′. GAPDH was amplified as an internal control using the primers 5′-GAAGGTGAAGGTCGGAGTC-3′ and 5′-GAAGATGGTGATGGGATTTC-3′.

Methylation specific polymerase chain reaction (MSP) was performed with primers specific for either methylated or the modified unmethylated DNA. The forward and reverse PCR primers for unmethylated reaction were 5′-TTTGTGTTTTGATGTTTGTAGGTTTTTGT-3′ and 5′-AACTCCACACTCTTCCAAAAACAAAACA-3′, and primers for the methylated reaction were 5′-TTTCGACGTTCGTAGGTTTTCGC-3′ and 5′-GCACTCTTCCGAAAACGAAACG-3′.

IHC assays were performed as previously described (19).

Radiation effectiveness was assessed by a colony formation assay. Cells were plated in 6-well plates, irradiated (0, 2, 4, 6, and 8 Gy), and returned to the incubator for 12 to 15 days. Then, cell colonies (>50 cells) were stained with crystal violet and counted. The surviving fraction was calculated as mean colony number/ (cells inoculated × plating efficiency). The data were analyzed using GraphPad Prism 5.0 software, and survival curves were plotted according to the single-hit multitarget model SF = 1 − (1 − e^−D/D0)^N, where SF is the surviving fraction and D is the radiation dose.

Immunofluorescent staining was performed as described previously (20).

Cells were plated in 6-well plates, irradiated (0 or 4 Gy), and returned to the incubator. DNA damage repair was assessed by single-cell gel electrophoresis assay under alkaline conditions as previously described (20).

Cells were plated in 6-well plates, irradiated (0 or 4 Gy), and returned to the incubator for 12 hours. A flow cytometry assay was performed as described previously (21). The percentage of apoptotic cells was determined with Annexin V–PI/FITC staining. Cell-cycle progression was determined using propidium iodide staining.

A terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) apoptosis detection kit (Millipore) was used according to the manufacturer's instructions.

Antibodies against Smo, Gli1, Gli2, Rad17, Claspin, Plk1, ATR, p-ATR (Ser428), ATM, p-ATM (Ser1981), Chk1, p-Chk1 (Ser 317), p-Chk1 (Ser345), Chk2, p-Chk2 (T68; Abcam); HA (F-7), Myc (Santa Cruz Biotechnology); γ-H2AX, USP3, His, and β-Actin (Cell Signaling Technology) were used for Western blot analysis.

Coimmunoprecipitation (Co-IP) assays were performed as described previously (22).

A chromatin immunoprecipitation (Chip) assay was performed as described previously (23).

The reaction was carried out at 37°C for 2 hours in 20 μL reaction buffer (20 mmol/L Tris-HCl, pH 7.2, 5 mmol/L MgCl₂, 50 mmol/L NaCl, 1 mmol/L 2-mercaptoethanol, 10% glycerol) containing the following components: 50 mmol/L of ubiquitin WT, 100 nmol/L human E1, 2 μmol/L UbE2D3, 2 μmol/L HERC2, 2 mmol/L ATP, 5 μmol/L of GST-Claspin and His-Smo.

For subcutaneous xenograft studies, primary GBM2 cells (1 × 10⁶) were inoculated into 6-week-old nude mice. Tumor volume was calculated using the formula: V = 0.5 × length × width². After the tumor volume reached 50 mm³, the mice were irradiated with 0 or 2 Gy daily for 5 days. For orthotopic xenograft studies, GBM cells (5 × 10⁵) were injected intracranially into the striatum of nude mice. The mice were subjected to IVIS scanning on the ninth day to confirm tumor uptake before treatment with radiation. All animal experiments were conducted with the approval of the Nanjing Medical University Institutional Committee for Animal Research and in conformity with national guidelines for the care and use of laboratory animals.

All results were presented as the mean ± SEM or the mean ± SD as indicated of at least three independent experiments. Data were subjected to Student t test for pairwise comparison or ANOVA for multivariate analysis. Differences were considered to be statistically significant at P < 0.05.

GBM samples were collected with written informed consent from all patients. Human brain tumor specimens and the database were approved by the Institutional Review Board and the Ethics Committee of Nanjing Medical University.

The use of animals in this study was approved by Institutional Review Board and the Ethics Committee of Nanjing Medical University. The animals were administered treatment in accordance with the ethical standards formulated in the Helsinki Declaration and national guidelines and regulations.

The mRNA expression levels of Smo were significantly higher in recurrent samples than in primary samples (Fig. 1A). Higher Smo expression indicated poor prognosis of GBM patients after radiation treatment (Supplementary Fig. S1A). Next, we analyzed Smo levels by IHC and Western blot, and Smo was markedly higher in the recurrent GBM samples than in the primary GBM samples (Fig. 1B; Supplementary Fig. S1B and S1C). In addition, Smo was significantly higher in the GBM cells than normal human astrocytes (NHA), especially higher in recurrent GBM cells (Supplementary Fig. S1D). To control for cell line heterogeneity, we established 4 primary GBM cell lines from 2 paired primary and recurrent GBM samples. We next determined GBM-associated molecular markers, including the methylation status of the MGMT promoter and the mutation status of IDH 1/2, EGFRvIII, and TP53 for the paired primary and recurrent GBM samples. All of the molecules and their mutations were consistent between the original tumor and cell line pairs (Supplementary Table S1). MGMT promoter methylation is reported to be a predictive biomarker of radiation response (24). Next, we evaluated MGMT promoter methylation and detected the protein expression of MGMT in the 2 paired primary and recurrent GBM cell lines. No differences were detected between these cell line pairs (Fig. 1C and D), which excluded the possibility that radiation resistance in recurrent GBM is due to MGMT promoter methylation. The radiation effectiveness against primary and recurrent GBM cells was assessed by a colony formation assay. Recurrent GBM cells showed greater resistance to radiation than primary GBM cells (Fig. 1E). Additionally, to verify cell survival after radiation, primary and recurrent GBM cells were differentially labeled with fluorescent dyes and mixed in defined ratios (primary GBM cells: red, recurrent GBM cells: green). In the presence of IR, the percentage of recurrent GBM cells after IR significantly increased (Fig. 1F; Supplementary Fig. S1E). We examined primary and recurrent GBM cells in response to IR-induced DNA damage by using γ-H2AX and the bromodeoxyuridine (BrdUrd) comet assay. We evaluated the presence of γ-H2AX foci in primary and recurrent GBM cells (Supplementary Fig. S1F). Although the numbers of γ-H2AX foci were similar in primary and in recurrent cells within 30 minutes after IR, the number of γ-H2AX foci in recurrent cells began to decrease within 2 hours after IR, whereas primary cells had a constantly high number of γ-H2AX foci, even after 12 hours. Primary and recurrent cells were equally susceptible to DNA damage by IR initially, but the percentage of cells with comet tails decreased more rapidly in recurrent cells than in matched primary cells (Supplementary Fig. S1G). These data confirm that recurrent GBM cells have greater radioresistance than primary GBM cells in vitro. In vivo, we used primary and recurrent GBM2 cells to generate orthotopic xenograft tumors in nude mice. From the 10th to the 15th day, the tumor-bearing mice were irradiated with 0 or 2 Gy daily for 5 days. Tumor progression was monitored using in vivo luminescence imaging. Xenografts carrying recurrent GBM2 cells displayed lager tumors (Supplementary Fig. S1H). Together, these results showed that Smo is markedly upregulated in recurrent GBM cell lines and tissues, and recurrent GBM cells show greater radioresistance.

To investigate whether Smo is associated with radioresistance, we selected primary GBM cells (the surviving fraction after 2 Gy irradiation was less than 45%) to overexpress Smo. As expected, Smo was overexpressed in the two primary GBM cell lines (Fig. 2A). Then, we determined the impact of Smo overexpression on primary GBM cells survival after IR. As shown in Fig. 2B, Smo overexpression made primary GBM cells more resistant to IR. Next, we assessed the apoptosis rate in Smo-overexpressing cells or vector control cells in response to IR. Smo-overexpressing cells had significantly fewer apoptotic cells (Annexin V–PI/FITC or TUNEL staining) and corresponding lower cleaved caspase-3 levels than control cells (Fig. 2C–E; Supplementary Fig. S2A). These results demonstrate that Smo overexpression induces primary GBM cell resistance to IR and protects primary GBM cells from IR-induced apoptosis.

We then used an in vivo approach to examine whether Smo conferred radioresistance. We subcutaneously injected primary GBM1 cells overexpressing Smo or a vector control into nude mice. After GBM engraftment was established, the mice were treated with IR at 0 or 2 Gy daily for 5 days. Primary GBM1 cells overexpressing Smo resulted in significantly higher tumor volumes and weights in irradiated mice than cells overexpressing the vector control (Fig. 2F–H). Taken together, these results indicate that Smo overexpression contributes to the acquisition of radioresistance.

Because Smo overexpression could confer IR resistance in GBM cells, we determined whether inhibiting Smo may modify GBM cell radiosensitivity. We infected two recurrent GBM cell lines (the surviving fraction after 2 Gy irradiation was over 65%) with three independent luciferase-encoding Smo shRNAs or control shRNA. As shown in Fig. 3A and Supplementary Fig. S3A, altered Smo expression in recurrent GBM cells was confirmed by Western blotting. We chose sh-Smo-1 for subsequent experiments due to its silencing efficiency. Smo-depleted recurrent GBM cells were more sensitive to radiation than control cells (Fig. 3B; Supplementary Fig. S3B).

To investigate the impact of Smo inhibition on IR-induced DNA damage repair, we examined the profile of DNA damage at different time points via γ-H2AX. We evaluated the presence of γ-H2AX foci in sh-Smo-1 and shCtrl recurrent GBM cells (Fig. 3C; Supplementary Fig. S3C). Although the number of γ-H2AX foci was similar in sh-Smo-1 and shCtrl cells within 30 minutes after IR, the number of γ-H2AX foci in shCtrl cells began to decrease within 2 hours after IR, whereas sh-Smo-1 cells had a constantly high number of γ-H2AX foci, even after 12 hours. These results indicate that Smo inhibition delayed the repair of radiation-induced DNA damage. In addition, we determined the level of γ-H2AX by Western blotting to confirm this result (Fig. 3D; Supplementary Fig. S3D). Then, we compared the recovery of sh-Smo-1 and shCtrl cells in response to IR-induced DNA damage using a BrdUrd comet assay. sh-Smo-1 and shCtrl cells were equally susceptible to DNA damage to IR initially, but the percentage of cells with comet tails decreased more rapidly for shCtrl cells than for matched sh-Smo-1 cells (Fig. 3E; Supplementary Fig. S3E). These data indicate that sh-Smo-1 cells repaired the DNA damage less efficiently than shCtrl cells.

We next determined the effect of Smo knockdown on the cell cycle after IR in recurrent GBM cells (Fig. 3F; Supplementary Fig. S3F). Smo knockdown augmented IR-induced G₂–M arrest in recurrent GBM cells. Tumor cells with impaired DNA damage repair ability are more prone to apoptosis. Therefore, we examined the capacity of Smo knockdown to induce apoptosis in response to radiation. As shown in Fig. 3G and Supplementary Fig. S3G, a remarkable increase in the number of apoptotic cells was observed in sh-Smo-1 cells compared with shCtrl cells. We also observed a significant increase in the number of apoptotic cells (determined by TUNEL staining) in sh-Smo-1 cells compared with shCtrl cells (Fig. 3H; Supplementary Fig. S3H and S3I). In accordance with these results, cleaved caspase-3 levels were higher in sh-Smo-1 cells than in shCtrl cells (Fig. 3I).

To test the in vitro results in an in vivo setting, we used recurrent GBM2 cells to generate orthotopic xenograft tumors in nude mice. From the 10th to the 15th day, the tumor-bearing mice were irradiated with 0 or 2 Gy daily for 5 days. Tumor progression was monitored using in vivo luminescence imaging. In stark contrast, xenografts carrying Smo-depleted recurrent GBM2 cells displayed significantly suppressed tumor growth (Fig. 3J; Supplementary Fig. S4A). The survival curves showed that IR-treated Smo-depleted xenografts exhibited significantly increased survival compared with IR-treated control xenografts (Fig. 3K). Furthermore, we examined the expression of γ-H2AX by IHC (Supplementary Fig. S4B) and examined apoptosis by TUNEL staining (Supplementary Fig. S4C). These data suggest that Smo inhibition sensitizes recurrent GBM cells to IR.

Recent studies have shown that the ATR/Chk1 and ATM/Chk2 pathways are two primary kinase signaling cascades activated in response to DNA damage (12). Activation of ATR and ATM results in the phosphorylation of a diverse array of downstream targets that participate in numerous cellular events, including DNA damage recognition and processing, cell-cycle arrest at different points, and apoptosis (14). Chk1 and Chk2 are selectively phosphorylated and activated by ATR and ATM, respectively, which triggers a wide range of distinct downstream responses (11). This prompted us to determine whether Smo knockdown enhanced the radiation effect through the ATR/Chk1 and/or ATM/Chk2 signaling pathways. Smo knockdown in recurrent GBM cells led to a significant decrease in the phosphorylation of Chk1 at Ser317 and Ser345 and ATR at Ser428 with a concomitant increase in γ-H2AX levels after IR (Fig. 4A). However, Smo knockdown in recurrent GBM cells did not affect the ATM/Chk2 pathway (Supplementary Fig. S5A). Claspin, Rad 17, and Plk1 are essential proteins for the ATR–Chk1-dependent activation of the DNA replication checkpoint response (13). First, we assessed the effect of Smo knockdown on the levels of Claspin, Rad17, and Plk1. As shown in Fig. 4B, Smo knockdown in recurrent GBM cells decreased the levels of total Claspin after IR, whereas the protein levels of Rad17 and Plk1 were not altered. Claspin is phosphorylated by ATR and binds directly to Chk1, which is important for Chk1 activation. Second, we determined how Smo inhibition affected Claspin expression in response to radiation-induced DNA damage. We performed qPCR to measure Claspin mRNA levels in sh-Smo-1 or shCtrl cells after radiation. However, no significant difference in Claspin mRNA levels was observed between sh-Smo-1 and shCtrl cells (Fig. 4C). Therefore, Claspin expression was not regulated at the mRNA level by Smo knockdown. Previous studies have reported that Claspin degradation occurs through ubiquitination and proteasomal degradation. Thus, we next treated sh-Smo-1 or shCtrl cells with the proteasome inhibitor MG132 and radiation. As shown in Fig. 4D, MG132 restored the levels of Claspin in sh-Smo-1 and shCtrl recurrent GBM cells. These results indicated that Smo could regulate Claspin protein stability via a proteasome-dependent pathway. It is known that the ubiquitin-proteasome system (UPS) promotes the proteasomal degradation of target proteins. Thus, we asked whether Smo mediated radiation-induced Claspin proteolysis via the UPS. Sh-Smo-1 and shCtrl recurrent GBM cells were transiently cotransfected with plasmids encoding Myc-tagged ubiquitin and HA-tagged Claspin and then irradiated. Sh-Smo-1 cells showed significantly higher radiation-induced Claspin protein levels with dramatically less polyubiquitination than shCtrl cells (Fig. 4E). Additionally, Smo overexpression in primary GBM cells led to a significant increase in the phosphorylation of Chk1 at Ser317 and Ser345 and ATR at Ser428, as well as a decrease in γ-H2AX levels after IR (Supplementary Fig. S5B). Smo cells showed significantly lower radiation-induced Claspin protein levels with dramatically less polyubiquitination than vector cells (Fig. 4F). To determine whether Claspin itself is ubiquitinated, the bacteria-expressed and purified GST-Claspin and His-Smo proteins were incubated with E1, UbE2D3 (E2), HERC2(E3), and ubiquitin for 2 hours. The mixtures were subjected to GST pulldown and Western blot with anti-Myc antibody. In vitro ubiquitylation assays showed that ubiquitination was present on Claspin itself, and Smo could not directly decrease the ubiquitin of Claspin (Supplementary Fig. S5C). To rule out shRNA off-target effects, we showed that Smo add-back reverses the following three phenotypes due to the sh-Smo-1 knockdown: increased IR sensitivity, persistent γH2AX foci, and Claspin destabilization (Supplementary Fig. S5D–S5G). Finally, to further confirm whether Smo mediated the Claspin-mediated ATR/Chk1 pathway, we determined whether Claspin could rescue Sh-Smo-1 cells from radiation-induced DNA damage. An Myc-tagged Claspin overexpression construct was expressed in sh-Smo-1 and shCtrl recurrent GBM cells, which were then irradiated. Exogenous expression of Claspin rescued the ATR/Chk1 pathway (Fig. 4G). Together, these data indicate that Smo regulates the ATR–Chk1 signaling pathway by mediating Claspin polyubiquitination and proteasomal degradation.

Deubiquitinating enzymes (DUB) are a group of proteases that regulate ubiquitin-dependent pathways by cleaving ubiquitin-protein bonds (25). To identify potential DUBs responsible for Claspin degradation, we used a DUB siRNA library designed for effective and convenient siRNA to knock down the entire DUB family. DUB library screening revealed that siRNA-mediated inhibition of all DUB genes decreased Claspin protein levels. Among the 98 known human DUBs (Supplementary Table S2), the knockdown of only 5 (USP3, USP9X, USP9Y, USP20, and USP50) types of DUBs decreased Claspin protein levels by more than half (Fig. 5A; Supplementary Fig. S6A). A recent study showed that the Hedgehog (Hh) signaling pathway could promote DUB transcription (23). We asked whether Smo stabilizes Claspin by promoting DUB transcription. We detected the mRNA levels of USP3, USP9X, USP9Y, USP20, and USP50 in sh-Smo-1, shCtrl, Smo, and vector control HEK293 cells. Previous studies reported that USP7, USP20, and USP29 could stabilize Claspin level (26–28). We also detected the mRNA levels of USP7 and USP29 in sh-Smo-1, shCtrl, Smo, and vector control HEK293 cells. The results showed that Smo knockdown decreased the mRNA levels of only USP3, and Smo overexpression increased the mRNA levels of USP3 (Fig. 5B; Supplementary Fig. S7A and S7B). In addition, Smo affected the mRNA levels of USP3 through Gli 2 (Fig. 5C), but not Gli 1 and Gli 3 (Supplementary Fig. S7C and S7D). Transcription factors may bind to specific regions of promoters to regulate gene expression, and we found one potential Gli 2-binding site (Fig. 5D). Reporter gene assays showed that the USP3 promoter could promote reporter gene expression more effectively than the pGL3-Basic control vector (Fig. 5E). Then, a chromatin immunoprecipitation assay was performed to clarify the association of Gli 2 and the USP3 promoter region in recurrent GBM cells (Fig. 5F).

Then, we detected the expression of USP3 and Claspin in primary and recurrent GBM tissues and cells. We found that mRNA and protein levels of USP3 and protein levels of Claspin in recurrent GBM were higher than that in primary GBM. But the mRNA expression of Claspin was not changed (Supplementary Fig. S8A-S8C). We next examined whether USP3 could stabilize Claspin. We found that USP3-WT overexpression increased the protein levels of Claspin, but the catalytically inactive C168S mutant of USP3 failed to induce Claspin levels (Fig. 5G and H). USP3 knockdown in recurrent GBM cells decreased Claspin expression, and this effect was partially reversed by MG132 (Fig. 5I and J). Moreover, USP3 depletion promoted the degradation of Claspin in recurrent GBM cells treated with cycloheximide (Fig. 5K and L). Lastly, USP3 overexpression reverse sh-Smo–induced decreasing in Claspin, CHK1 phosphorylation, and γ-H2AX after IR (Supplementary Fig. S8D). So, Smo stabilizes Claspin via USP3.

We then examined the interaction between USP3 and Claspin. Co-IP experiments in HEK293T and recurrent GBM cells showed that USP3 and Claspin interact with each other (Fig. 6A and B). Confocal images showed colocalization of USP3 (green) and Claspin (red) in recurrent GBM cells (Fig. 6C). USP3 contains two conserved protein domains: a zinc-finger domain (ZnF) and a catalytic domain (UCH). To identify which USP3 domain mediated the interaction with Claspin, we generated two deletion mutants containing the ZnF and UCH domains of USP3 (Fig. 6D). We found that the UCH domain mediated the interaction with Claspin. In addition, we generated six truncated fragments spanning full-length Claspin as described in a previous study (28), and our domain mapping experiments revealed that the N-terminus of Claspin (1–330) mediated the interaction with USP3 (Fig. 6E). USP3 (UCH) showed an inhibitory effect on Claspin ISRE-luc activation, but USP3 (ZnF) did not have any inhibitory effect on Claspin-induced ISRE-luc activation (Supplementary Fig. S9A).

To determine whether USP3 deubiquitinates Claspin, HEK293 cells were cotransfected with HA-Claspin and Myc-Ub. Ubiquitinated Claspin was incubated with Flag-USP3 (WT or C168S) or control. As shown in Fig. 7A, USP3-WT, but not USP3-C168S, significantly reduced the ubiquitination of Claspin. In recurrent GBM cells, USP3 knockdown induced the ubiquitination of Claspin (Fig. 7B). Two major forms of polyubiquitin chains are known to be formed through distinct types of linkages (Lys48- or Lys63-linked chains). Lys48-linked ubiquitin chain consistently guides proteolysis by the proteasome (37), but the lysine 63-linked ubiquitination of the substrate protein indicates processing by the endosomal–lysosomal-dependent degradation pathway. Accordingly, we wondered which type of polyubiquitin modifications on Claspin protein was affected by USP3. USP3 significantly disassembled Lys48-linked polyubiquitylation of Claspin, but had no significant effect on the Lys63-linked polyubiquitylation of Claspin (Fig. 7C). Furthermore, we expressed an Lys48-resistant (Lys48R) form of ubiquitin in USP3 inhibition recurrent GBM cells and found that enforced expression of Lys48R ubiquitin reduced USP3 inhibition–induced Claspin downregulation (Fig. 7D). Together, these results demonstrate that USP3 deubiquitinates Claspin.

To leverage our results for clinical application, we used a small-molecule inhibitor of Smo, GDC-0449. GDC-0449 treatment induced recurrent GBM cells sensitive to IR but not to NHAs (Supplementary Fig. S10A). GDC-0449 treatment reduced the protein levels of Smo and Claspin, and induced he protein levels of γ-H2AX after IR in recurrent GBM cells (Supplementary Fig. S10B). GDC-0449 treatment, like Smo knockdown, promoted Claspin ubiquitination and degradation (Supplementary Fig. S10C). In vivo, GDC-0449 with IR treatment recurrent GBM2 cells displayed significantly suppressed tumor growth compared with IR treatment only (Supplementary Fig. S10D).

Radiotherapy is the most widely used modality in the treatment of GBM (29, 30). Nonetheless, GBM radioresistance is the main cause of local failure after radiotherapy (5, 6). Several studies indicated the influence of Hedgehog signaling to promote more aggressive disease and tumor resistance to therapeutics (31, 32). Smo plays a critical role in Hedgehog signaling, which activates Gli transcription factors. Additionally, we previously proved that Smo was a poor prognosis factor and a potential therapeutic target in glioma (33). In this study, we investigated the specific role of Smo in the modulation of GBM radioresistance. First, we detected paired primary and recurrent GBM patients who were treated with radiation and observed that Smo mRNA and protein expression were higher in recurrent GBMs than in primary GBMs. By correlating of Smo expression with survival for GBM patients after radiation treatment, we found that higher Smo expression indicated poor prognosis of GBM patients after radiation treatment. Additionally, recurrent GBM cells show greater resistance to radiation than primary GBM cells. Smo overexpression in primary GBM cells promoted acquired radiation resistance in vitro and in vivo. However, Smo inhibition sensitized recurrent GBM cells to radiation in vitro. These sensitizing effects are the result of the inhibition of DNA damage repair, leading to G₂ arrest and apoptosis. These data are further supported by patient-specific orthotopic GBM xenograft models.

ATR has been shown to be responsible for promoting G₂ arrest and Claspin degradation turns off the G₂ checkpoint (34). Of note, our data suggested G₂ arrest after IR in Smo depletion recurrent GBM cells. Considering that Smo, a cell-surface modulator, initiates an upstream signaling but might not solely depend on the Clasping–ATR axis here, we speculated the previously reported Gli1 inhibition (35, 36) may play a role in promoting G₂ arrest. Indeed, Smo inhibition decreased the expression of Gli family including both Gli1 and Gli2 in our cell models. When the expression of Gli1 was further regulated in recurrent GBM cells to evaluate the cell cycle, we found a significant G₂ arrest after IR in Gli1 depletion recurrent GBM cells. Conversely, G₂ arrest restored when we added back Gil1 (Supplementary Fig. S11A and S11B). Therefore, we have identified a mechanism by which the increase G₂ arrest was mainly regulated by Gil1. Thus, our findings uncover the importance of Smo in the radiation resistance of GBM and the applicability of Smo as a potential prognostic and predictive marker for radiation sensitivity in GBM.

DNA damage responses are controlled by two main signaling pathways in vertebrates, the ATM–Chk2 and ATR–Chk1 protein kinase pathways (37). ATM and ATR are two serine/threonine kinase members of a phosphoinositide 3-kinase–like family (38, 39). Chk1 and Chk2, which are selectively phosphorylated and activated by ATR and ATM, respectively, trigger a wide range of distinct downstream responses (40). Activated ATR–Chk1 and ATM–Chk2 regulate DNA damage repair and induce apoptosis (12). The present study showed that Smo knockdown could disrupt the activation of ATR/Chk1 but not ATM–Chk2 signaling in response to radiation-induced DNA damage, leading to inhibited DNA damage repair and increased apoptosis. Claspin is a critical protein for the ATR-dependent phosphorylation of Chk1. Previous studies have shown that Claspin is associated with DNA damage in human cells (41, 42). Claspin initiates DNA replication in human cells, while Claspin downregulation upregulates DNA damage (43). In this study, Smo mediated Claspin degradation via UPS-mediated proteolysis. We demonstrated that Smo knockdown could promote radiation-induced Claspin polyubiquitination and proteasomal degradation, resulting in Chk1 inactivation and increased radiosensitivity to GBM. However, Smo overexpression could decrease radiation-induced Claspin polyubiquitination and proteasomal degradation. Our findings demonstrate the connection between Smo and the regulation of Claspin proteolysis.

Smo is an important component of the Hh pathway, which triggers the activation of Gli transcription factors and then regulates downstream target genes (9). Protein ubiquitination is tightly controlled by ubiquitin ligases or their removal through deubiquitinating enzymes (44, 45). Deubiquitinases sever ubiquitin from substrates and stop ubiquitin-dependent signaling, and they are now known to be essential and specific components of the ubiquitin–proteasome pathway (46, 47). In the present study, we first determined the potential DUBs that could mediate Claspin degradation, and we selected 5 DUBs from the entire DUB family. A recent study reported that Gli 1 could promote USP48 transcription (23). Similar to Gli1, Gli2 is a key transcriptional target of Hh signaling. Indeed, the present study showed that Smo induced only USP3 expression through transactivation by Gli2. USP3 is a member of the DUB family that has been reported to regulate various cellular events through deubiquitinating various target proteins, such as RIG-I and p53 (17, 48). Here, we identify that USP3 removes ubiquitin conjugates from Claspin, which is turn suppresses Claspin's proteasome degradation. Ectopic expression of USP3 attenuated Claspin polyubiquitination and proteasomal degradation. Conversely, USP3 knockdown induces Claspin polyubiquitination and proteasomal degradation. More importantly, USP3 contains two domains: a ZnF domain and a UCH domain. The results show that only the UCH domain of USP3 can interact with Claspin but not the ZnF domain of USP3. Moreover, the N-terminus of Claspin (1–330) mediates the interaction with USP3. Thus, our findings show that Smo promotes radiation resistance in GBM through USP3-induced Claspin-dependent ATR–Chk1 signaling (Fig. 7E). GDC-0449 is an inhibitor of Smo and has entered clinical trial for glioma. In this study, we used GDC-0449 to assess the impact of pharmacologic inhibition of Smo in GBM. We found that GDC-0449 could promote Claspin ubiquitination and degradation, which lead to recurrent GBM cells sensitive to IR.

In summary, our study revealed that Smo is critical in modulating radiation resistance in GBMs. Smo regulates USP3 expression through transactivation by Gli2, leading to the activation of Claspin-dependent ATR–Chk1 signaling. Our results reveal a novel mechanism underlying acquired radiation resistance in GBM and have important implications in the development of treatment strategies for radiation-resistant GBMs.

All experiments were performed in triplicate. The data were subjected to Student t test for pairwise comparison or ANOVA for multivariate analysis. Differences were considered to be statistically significant at P < 0.05.

No potential conflicts of interest were disclosed.

All the data set and materials generated and/or analyzed during the current study are available.

Conception and design: Y. Tu, Y. Chen, J. Ji

Development of methodology: Y. Tu, Z. Chen, G. Sun, H. Chao, L. Fan, Y. You, J. Ji

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Tu, Z. Chen, C. Li, Y. You, J. Ji

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Tu, Z. Chen, Z. Bao, Y. You, J. Ji

Writing, review, and/or revision of the manuscript: Y. Tu, H. Chao, Y. Qu, Y. Chen, J. Ji

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Tu, J. Ji

Study supervision: P. Zhao, J. Ji

This study was supported by grants from the National Natural Science Foundation of China (81972153, 81772602, 91742105, and 91942309), Jiangsu Province's Natural Science Foundation (BK20160047), Jiangsu Province's Key Discipline of Medicine (XK201117), Jiangsu Province Medical Key Talent (ZDRCB2016002), Jiangsu Provincial Key Research Development Program of China (BE2018750), and the Priority Academic Program Development of Jiangsu Higher Education Institution.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.