Partial truncation of the C-terminal domain of PTCH1 in cancer promotes tumourigenesis by non-canonical activation of a GLI-PI3K loop
Begoña Caballero-Ruiz, Rosa Bordone, Sonia Coni, Danai S. Gkotsi, Eva Gonzalez, Gianluca Canettieri, Natalia A. Riobo-Del Galdo
TL;DR
Truncated PTCH1 in colon cancer promotes tumor growth through a new GLI-PI3K pathway, offering potential drug targets.
Contribution
Discovery of a non-canonical GLI-PI3K loop activated by PTCH1 C-tail truncation in colon cancer.
Findings
Truncated PTCH1 enhances tumor growth and proliferation in colon cancer cells.
GLI1 and GLI2 are upregulated via a Smoothened-independent mechanism.
PTCH1 truncation activates PI3K and cancer-related pathways like EGFR and Ras.
Abstract
Loss of function mutations of the Hedgehog receptor PTCH1 are oncogenic drivers in some skin and brain cancers. We recently reported mutations in exons encoding the C-terminal tail of PTCH1 in colon cancer, which result in premature truncation but do not impair canonical Hedgehog signalling. In this study, we show that colon cancer cells engineered by CRISPR/Cas9 to express endogenous truncated PTCH1 have enhanced proliferation, colony formation, anchorage-independent growth and form larger tumours in vivo than isogenic cells expressing wild-type PTCH1. Analysis of the mechanisms underlying this growth advantage revealed profound transcriptional changes and unexpectedly, upregulation of GLI1 and GLI2 by a Smoothened-independent route, which proved to be necessary for the proliferative advantage. Furthermore, we found that truncation of PTCH1 C-tail upregulated several cancer-related…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6- —https://doi.org/10.13039/501100000268RCUK | Biotechnology and Biological Sciences Research Council (BBSRC)
- —https://doi.org/10.13039/501100004271Sapienza Università di Roma (Sapienza University of Rome)
- —Sapienza University of Rome doctoral fellowship
- —https://doi.org/10.13039/501100005010Associazione Italiana per la Ricerca sul Cancro (Italian Association for Cancer Research)
- —https://doi.org/10.13039/501100003407Ministero dell'Istruzione, dell'Università e della Ricerca (Ministry of Education, University and Research)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsHedgehog Signaling Pathway Studies · Wnt/β-catenin signaling in development and cancer · Axon Guidance and Neuronal Signaling
Introduction
The Hedgehog (Hh) pathway plays an essential role in embryonic development, tissue regeneration, cell fate and stem cell renewal [1–3]. Canonical Hh signalling is triggered by binding of one of the three Hh ligands (Sonic, SHH; Indian, IHH; or Desert, DHH) to the receptor Patched1 (PTCH1), leading to derepression of Smoothened (SMO) and activation of the GLI-family of transcription factors. The presence of Hh ligands is sensed by the constitutively expressed GLI2 and GLI3 isoforms, which then induce expression of GLI1, PTCH1, and other cell type-specific genes [3]. Hh binding to PTCH1 also inhibits non-canonical PTCH1 intrinsic functions that are independent of SMO/GLI signalling, such as proapoptotic signalling and inhibition of autophagy [4–6]. These functions of PTCH1 are mediated by its C-terminal domain (CTD), a 273 amino acids-long intrinsically disordered region that acts as a protein-protein interaction hub for ubiquitin E3 ligases, DRAL/TUCAN-1/procaspase-9, and ATG101 [5, 7, 8]. However, most of the CTD is dispensable for canonical Hh signalling via SMO and GLI transcription factors [9, 10].
Excessive activation of canonical Hh signalling is common in many cancers, with loss-of-function mutations in PTCH1 and gain-of-function mutations in SMO driving tumourigenesis in a Hh ligand-independent manner in most basal cell carcinomas of the skin and the SHH-type medulloblastoma [11]. Upregulation of SHH and IHH without pathway mutations is a common finding in carcinomas and adenocarcinomas from the GI tract [12–14]. In addition, multiple oncogenic pathways can lead to activation or potentiation of GLI transcriptional activity. Indeed, there is increasing evidence that regulation of GLI proteins in cancer might also occur through non-canonical signalling pathways, reducing therapeutic efficacy of SMO antagonists [15].
In a previous study we reported that indel mutations in exons 22-23 encoding the CTD of PTCH1 in 5-13% human colon, stomach, and endometrial cancers result in loss of PTCH1 interaction with ATG101 [16]. The mutational hotspots S1203fs, R1308fs and Y1316fs, enriched in right-sided MSI+ colon cancer, result in premature stop codons. We reported that these frameshift mutations lead to loss of interaction with ATG101, increasing both basal and induced autophagy, and promote survival under nutritional stress [16]. Based on those findings, we hypothesised that partial truncation of PTCH1 reduces its tumour suppressor activity. In this study, we demonstrate that homozygous CTD truncation of endogenous PTCH1 in colon cancer cells engineered by CRISPR/Cas9 have a proliferative advantage in vitro and in vivo compared to isogenic cells expressing wild type PTCH1. The increased fitness is provided by a SMO-independent, PI3K-dependent upregulation of GLI1 and other profound transcriptome changes. In summary, we report for the first time that PTCH1 CTD truncations found in cancer underlie a more aggressive phenotype in vivo and in vitro in a manner that is insensitive to SMO inhibitors but sensitive to PI3K inhibitors.
Material and methods
Cell lines and cultures
Scrambled (SCR) and PTCH1 CTD mutant SW620 cells (C9 and C15) were previously described [16]. LoVo (CCL-229) and SW480 (CCL-228) were purchased from ATCC (Manassas, VA, USA). Ptch1^−/−^ mouse embryonic fibroblasts were a gift from Dr. Matthew Scott (Stanford University). All cell lines were maintained in Dulbecco’s modified Eagle medium (DMEM, high glucose, Sigma-Aldrich) supplemented with 10% Foetal Bovine Serum (FBS) (GIBCO), 1 mM penicillin-streptomycin (Sigma-Aldrich) and 1 mM glutamine (Sigma-Aldrich) in a humidified incubator at 37 °C and 5% CO_2_.
Treatment with inhibitors
Cells were exposed to KAAD-Cyclopamine (Biovision #1910-50), Erlotinib (Selleckchem #S7786), Cetuximab (C225) (TargetMoI #T9905), GANT61 (Stratech Scientific Ltd #A1615), SD208 (Selleckchem #S7624), Trametinib (Selleckchem #S2673), UO-126 (Selleckchem #S1102), BKM120 (Selleckchem #S2247), or LY294002 (Thermo Fisher #PHZ1144) dissolved in DMSO. Treatments were performed in complete growth medium for the indicated time and concentration.
Animal studies
For the xenograft studies, animals (n = 6 per experimental group) were randomly assigned to receive subcutaneous injections of 2 × 10^6^ SCR cells (SCR group) or C9 cells (C9 group). Tumor cells were resuspended in 70 μl phosphate-buffered saline (PBS) and mixed 1:1 with Matrigel (Corning, #354248) before implantation into both flanks of 5-week-old female athymic nude mice (CD1, Charles River Laboratories). Once tumors reached a volume of 100 mm^3^, their size was measured every 3 days using a caliper without blinding. Animals were euthanized after 3 weeks, and all tumor masses were excised. The experiment was independently repeated twice.
Data are presented as mean ± standard deviation for each experimental group and time point. For each time point, a two-tailed Student’s t-test was used to compare groups, and p-values < 0.05 and < 0.01 were considered statistically significant and are indicated as * and **, respectively. Sample size was determined a priori using G*Power software.
All animal procedures complied with the European Community Council Directive 2010/63/EU and were approved by the local Ethical Committee for Animal Experiments of Sapienza University of Rome (Authorization no. 877/2016-PR).
Immunohistochemistry
Freshly excised xenografts were fixed in 4% paraformaldehyde for 24 h at 4*°*C, paraffin embedded and sectioned at 5 µm using a RM2245 microtome (Leica Biosystems, Wetzlar, Germany). Antigen was retrieved using 1X Tris-EDTA, pH 9.0 (Himedia #ML087). Sections were incubated with Gli1 (Cell Signaling Technology # 2643S) and Ki67 (Pierce #MA5-14520) antibodies followed by biotinylated goat anti-polyvalent secondary antibody and streptavidin-conjugated horseradish peroxidase and visualised with DAB using the Mouse-to-Mouse Kit (ScyTek Laboratories #MTM001). Images were captured using a ZEISS Axiocam 208 colour (ZEISS Group, Oberkochen, Germany) on a ZEISS Axiolab 5 light microscope (ZEISS Group).
Colony formation assays
Cells (1 × 10^3^ cells/well) were seeded into 6-well plates in complete DMEM and 24 h later treated with inhibitors or DMSO. After 10 days, cells were fixed with 100% methanol, stained with 0.005% crystal violet (Sigma-Aldrich #V5265), and colonies were counted.
Soft agar colony formation assay
A cell suspension containing 5 × 10^3^ cells was added to each well of a 6-well plate in soft agar (ThermoScientific #R0801) as previously described [17]. Once the colonies were visible (typically after 15 days), colonies were stained with 0.005% crystal violet solution for counting.
Proliferation assay
Cells (4 × 10^4^/well) were seeded in 24-well plates and incubated overnight at 37 °C. After 24 h, cells were trypsinised and live cells were counted (time = 0) by the trypan blue exclusion method. Cells were then treated with the different inhibitors at the concentrations indicated in the text. Every 24 h cells were trypsinised and counted following the same method.
For WST-1 proliferation assay, 2 × 10^3^ cells/well were seeded in 96-well plate in 100 μl and left for 24 h before being subjected to different treatments. At each time point, 10 μl of WST-1 reagent (Roche Applied Sciences #05015944001) was added to each well and incubated for 0.5- 4 h at 37 °C, 5% CO_2_. Then, the plate was shaken for 1 min and absorbance measured using a microplate reader at 420–480 nm.
EdU incorporation assay
The Click-iT EdU Alexa Fluor 488 HCS Assay (Invitrogen, #C10350) was performed following the manufacturer’s protocol to assess cell proliferation. Specifically, cells were supplemented with EdU for 6 h prior to fixation with 3.7% formaldehyde in phosphate-buffered saline (PBS). Permeabilization was achieved using 0.5% Triton X-100 in PBS. The Click-iT reaction cocktail was prepared as specified by the manufacturer and incubated for 30 min. Finally, nuclei were counterstained by adding 1X Hoechst 33342 solution for 10 min. Confocal images were acquired using a ZEISS LSM 980 with Airyscan 2 laser-scanning confocal microscope, employing a 63× Oil immersion objective to count the percentage of EdU+ nuclei.
Transient transfections
For transfection assays, cells (8 × 10^4^ /well) were seeded in a 12-well plate. On the following day, cells were transfected with Lipofectamine 2000 (Thermo Fisher #11668019) using 1.5 μl transfection reagent and 0.38 μg of DNA in Opti-MEM medium (Thermo Fisher # 31985-062). Before transfection, cells were washed once with PBS and the culture medium was replaced with Opti-MEM. The lipid-DNA complexes were then added to the cells. Cells were incubated with the transfection mixture for 6 h at 37 °C. Then, the media was replaced with complete growth medium lacking antibiotics. The following day, medium containing antibiotics was added to the cells. At the indicated time points, cells were trypsinised, counted following the trypan blue exclusion method and lysed for protein analysis. The plasmids encoding Akt1 (T308A and S473A) and myr-Akt1 were gifts of J. Testa (Fox Chase Cancer Center, Philadelphia). PKA-CQR was a gift from S. McKnight (University of Washington, Seattle). pcDNA3.1+ (Invitrogen) was used as empty control plasmid.
For transfections with Gli1 K518R and Gli1(2-413)-VP16 [18] cells (2 × 10^5^/cm^2^) were seeded in 21-cm² culture dishes and transfected following the method previously described. After 48 h from transfection, cells were then detached and a small aliquot was collected and processed for Western blot analysis to assess mutant overexpression, while the remaining cells were reseeded at 2 × 10^4^/well in 24-well plate. Cell number was quantified 72 h later using trypan blue exclusion method.
Western blotting
Cells or tissues were lysed in denaturing buffer containing 50 mM Tris-HCl, 2% SDS, 10% Glycerol, 10 mM Na_4_P_2_O_7_, 100 mM NaF, 6 M Urea, 10 mM EDTA or by direct lysis in RIPA buffer (Thermo Scientific # 89900) containing protease and phosphatase inhibitor cocktails for western blotting (Thermo Scientific #87785 # 78420). Protein extracts were quantified with the Pierce BCA protein assay kit (ThermoScientific # 23227) and resolved by SDS-PAGE. Separated proteins were transferred to a nitrocellulose (Perkin Elmer # NBA085C001EA) or PVDF membrane (BioRad #1620177), blocked in 5% non-fat dried milk in Tris-buffered saline with 0.05% Tween 20 (TBST) (Sigma #P7949), and incubated overnight with primary antibodies at 4°C at the specified concentrations. Next day, membranes were washed in TBST and incubated with HRP-conjugated secondary antibodies at RT for 1 h and developed using WesternBright ECL (Advansta #K-12045-D50) in a ChemiDoc imaging system (BioRad) using the Image Lab software (BioRad). Signal intensity was quantified by ImageJ software and normalized to loading control indicated in the text.
Antibodies
The following antibodies were used: EGFR (Cell Signaling Technology #2256, WB 1:1,000), AKT (Cell Signaling Technology #4691, WB 1:1,000), phospho- AKT (Ser473) (D9E) XP (Cell Signaling Technology #5012, WB 1:1,000), p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (Cell Signaling Technology #4695, WB 1:1,000), phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (Cell Signaling Technology #4376, WB 1:1,000), Gli1 (Cell Signaling Technology #2643, WB 1:1,000), Phospho-PKA substrate (Cell Signaling Technology #9624, WB 1:1,000), GAPDH-HRP (Proteintech #HRP-60004, WB 1:3,000), Vinculin (Santa Cruz #sc-73614, WB 1:10,000), Poly-ADP ribose Polymerase (Cell Signaling Technology #9542, WB 1:2,000), anti VP16 (Santa Cruz #sc7545, WB 1:1,000),HA-Tag (Invitrogen #26183, WB 1:10,000), goat anti-rabbit IgG HRP (BioRad #172-1019, WB 1:3,000), goat anti-mouse IgG (H L)-HRP conjugate (Bethyl #A90-116P, WB 1:3,000).
Quantitative Real-Time PCR (qPCR) assay
Total mRNA was isolated with RNeasy Mini Kit (QI-AGEN #74104) or by Trizol method. Synthesis of cDNA was performed using 1 µg RNA with the Sensifast cDNA synthesis kit (Bioline #BIO-65054) or iScript cDNA Synthesis Kit (BioRad #1708897) using hexa-random primers. Real-time quantitative PCR was performed using 2 µL of cDNA and target-specific primers (Table S1) and the SsoFast EvaGreen Supermix (BioRad #1725201) or SensiFast Sybr Lo-Rox Mix (Bioline #BIO-94020) in a CFX Connect Real-Time PCR System (BioRad) or ViiA 7 Real-Time PCR System (Applied Biosystems). Amplification was quantified using the ∆∆Ct method, where ΔCt is the difference of the Ct value between the target gene minus the Ct value of the control gene (GAPDH), and the second Δ is the difference between a PTCH1 CTD mutant clone and the SCR control.
RNAseq library construction, quality control and sequencing
Total RNA from the 3 cell clones in three consecutive cell passages was isolated with RNeasy Mini Kit (QIAGEN #74104). After fragmentation, cDNA was synthesised using hexa-random primers, followed by adaptor ligation, size selection, amplification and purification. The library was checked with Qubit and real-time PCR for quantification and bioanalyzer for size distribution detection. Quantified libraries were sequenced on Illumina PE150 platforms to generate paired-end reads at Novogene.
RNA-seq data analysis
Raw data in fastq format were cleaned using fastp software, aligned to the GRCh38 human reference genome using Hisat2 v2.0.5., the number of reads mapped to each gene quantified with featureCounts v1.5.0-p3, followed by calculation of FPKM (Fragments Per Kilobase of transcript sequence per Millions base pairs sequenced) per gene. Differential expression analysis was performed using the DESeq2 R package (1.20.0). The resulting P-values were adjusted using the Benjamini and Hochberg’s approach for controlling the false discovery rate (FDR). Genes with an adjusted P-value <=0.05 found by DESeq2 were assigned as differentially expressed.
Enrichment analysis of differentially expressed genes
Gene Ontology (GO) enrichment analysis of differentially expressed genes (DEG) was implemented by the clusterProfiler R package. GO terms with corrected P-value < 0.05 were considered significantly enriched. We tested the statistical enrichment of DEGs in KEGG, Reactome, Disease Ontology (DO) and DisGeNET pathways databases using clusterProfiler R package, with a P < 0.05 cutoff.
Gene set enrichment analysis (GSEA)
The DEGs between the 3 cell lines were ranked according to the degree of differential expression and the P-value. They were compared to predefined Gene Sets to determine if they were enriched at the top or bottom of the list using a local version of the GSEA analysis tool (http://www.broadinstitute.org/gsea/index.jsp).
Lentiviral-mediated shRNA knockdown
shRNA knockdown of GLI1 was performed through lentiviral infection. Lentiviruses were generated in HEK293 cells transfected with pLKO.1 vectors encoding shGLI1 (sh-hGLI1-85: 5’ CCGGCCTGATTATCTTCCTTCAGAATCTAGATTCTGAAGGAAGATAATCAGGTTTTTG, and sh-hGLI1-87: 5’ CCGGCCAAACGCTATACAGATCCTATCTAGATAGGATCTGTATAGCGTTTG GTTTTTG) or a scrambled sequence together with 15 μg of pCMV-R 8.74, and 10 μg of pMDG with the calcium phosphate method. After 24 h, the medium was changed and then collected at 48 h and stored at −80 °C. Control shRNA (scrambled, #SHC002) was obtained by Sigma-Aldrich. Viral titers were determined by qPCR as previously described [19]. For lentiviral transduction, 4 × 10^5^ cells were seeded overnight in 60 mm dishes. The day after, 1 ml of a 1:1 Sh-Gli1-85 and 87 virus-containing media was added to cells with 5 μg/ml polybrene (Sigma-Aldrich #H9268) and cells were incubated for 72 h. Knockdown efficiency was monitored by western blotting and qPCR.
Statistical analysis
Statistical analysis and generation of graphs was performed using GraphPad Prism 9 (GraphPad Software, La Jolla, California, USA, https://www.graphpad.com/). Data were analyzed by Student’s t-test to analyse the significant difference between two groups. One-way ANOVA analysis was selected to analyse at least three different groups when the samples showed normal distribution and equal variance, followed by post-hoc Tukey’s test for multiple comparisons. Figure legends indicate statistical details of experiments.
Results
PTCH1 CTD truncation enhances tumourigenesis in vitro and in vivo
To investigate the biological consequences of cancer-associated PTCH1 frameshift mutations that result in premature truncation of the CTD in the 1203-1316 region, we engineered indel mutations at position D1222 of PTCH1 in the colon cancer cell line SW620 by CRISPR/Cas9 (Fig. S1A). SW620 cells contain mutations in KRAS and TP53 (homozygous RAS(G12V); heterozygous TP53(R273H)/TP53(P309S)) [20]. We selected two single-cell derived clones harbouring different frameshift-causing mutations (C9 and C15) alongside the parental cells containing wild-type PTCH1 (SCR) to lower the risk of picking up a non-specific phenotype (Fig. S1B, C) [16, 21]. To investigate phenotypic changes caused by PTCH1 CTD truncation, we first examined the proliferative capacity of the C9 and C15 cells in vitro. Both PTCH1 mutant clones exhibited accelerated growth, reflected as an increase in cell number of ~2-fold over the parental SCR cells after 96 h (Fig. 1A). C9 and C15 cells exhibited increased proliferation, as determined by Edu incorporation (Fig. 1B) and reduced apoptosis, assessed by cleavage of Poly-ADP ribose polymerase (PARP) (Fig. 1C). In agreement, C9 and C15 cells showed a much higher colony forming capacity than SCR cells, with an average 13.3-fold (C9, p < 0.001) and 11-fold (C15, p < 0.01) increase (Fig. 1D). We also evaluated the anchorage-independent growth of the cells in soft agar, reflecting resistance to anoikis. The number of anchorage-independent colonies formed by C9 and C15 cells was increased by 8-fold (p < 0.0001) and 4-fold (p < 0.0001), respectively, compared to SCR cells (Fig. 1E). Therefore, SW620 cancer cells with PTCH1 CTD truncations have a much higher proliferative and survival capacity in vitro irrespective of their surrounding environment. Importantly, C9 and C15 cells expressed similar or greater levels of PTCH1 at the transcript level, ruling out nonsense-mediated decay (Fig. S1D). Therefore, the phenotype of PTCH1 CTD mutant cells is unlikely to be caused by functional PTCH1 deficiency.Fig. 1. Isogenic colon cancer cells with PTCH1 CTD truncation have a proliferative advantage in vitro and in vivo.A Change in cell number of SCR, C9 and C15 cells over time in complete medium (n = 3). B EdU incorporation, quantified as % of total nuclei, in SCR, C9 and C15 over time in complete medium (n = 3). C Representative immunoblot of full-length PARP and cleaved PARP in SCR, C9 and C15 cells. Expression of vinculin was used a a loading control. D Colony formation ability of SCR, C9 and C15 during 10 days of culture. Top. Representative images at the end of the experiment. Bottom: number of colonies per 1000 cells plated at day 10 (n = 3). E Anchorage-independent growth of SCR, C9 and C15 cultured in soft agar for 15 days. Top: representative images at the end of the experiment. Bottom: number of colonies per 5000 cells plated at day 15 (n = 6). F Changes in tumour volume over time of xenograft tumours after injection of SCR or C9 cells in the flanks of nude mice. G Weight of xenografts from SCR or C9 cells at the end of the experiment (n = 10). H Representative images of xenograft mice and isolated tumours at the end of the experiment. Representative images for IHC staining of Ki67 in sections of xenograft masses. All quantitative data represent the mean ± SEM of independent biological repeats. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Next, we grafted SCR or C9 cells into the flanks of athymic nude mice to evaluate the effect of PTCH1 CTD truncations on tumour growth in vivo over 21 days. C9 cells xenografts grew significantly faster compared to SCR control tumours. The growth advantage of C9 cells was evident from day 6 onwards, showing a 183% increase in tumour volume (p < 0.05) and a 211% increase in weight (p < 0.001) compared to tumours formed by SCR parental cells at day 21 (Fig. 1F, G). Consistently, Ki67 staining was increased in C9 tumours compared to SCR xenografts (Fig. 1H).
The results of altered oncogenic properties of isogenic colorectal cancer cells that differ in the mutational status of the PTCH1 CTD in vitro and in vivo strongly indicates that partial truncation of the PTCH1 CTD confers a growth advantage and a more malignant phenotype.
Non-canonical activation of the GLI transcription factors mediates the growth advantage of cancer cells with truncation in the PTCH1 CTD
We next set to investigate the mechanism by which truncation of the PTCH1 CTD provides an oncogenic advantage. We first investigated if truncation of the PTCH1 CTD results in activation of canonical Hh signalling. We tested this using Ptch1^-/-^ mouse embryonic fibroblasts, in which loss of Ptch1 results in constitutive activation of SMO and high Gli-transcriptional activity (Fig. S2A). In agreement with previous work [9], introduction of PTCH1 CTD truncation mutants lacking increasingly longer regions of the CTD have comparable activity to wild-type PTCH1 to inhibit basal Gli-luciferase activity in Ptch1^-/-^ MEFs (Fig. S2B). In addition, point mutations and small deletions in the CTD regions 1195-1998 and 1216-1235 do not affect canonical Hh signalling (H. Ollerton, F. Cross, NA Riobo-Del Galdo, unpublished data). Combined with the absence of nonsense-mediated decay of the mutant PTCH1 transcripts, this suggests that the cancer-associated truncations of the CTD do not promote oncogenic properties by derepression of SMO. To confirm this, we determined the expression level of GLI1, a GLI-target gene used as a hallmark of canonical Hh signalling. Unexpectedly, both C9 and C15 cells showed a strong expression of GLI1 at the protein level compared to SCR control cells (Fig. 2A). Furthermore, not only GLI1 but also GLI2, a constitutively expressed mediator of canonical Hh signalling, were transcriptionally upregulated between 6 and 9-fold in C9 and C15 cells (Fig. 2A). In agreement with the findings in cultured cells, C9-derived tumours xenograft masses showed strong GLI1 signal by immunohistochemistry compared to SCR-derived tumours (Fig. 2B).Fig. 2PTCH1 CTD truncation leads to non-canonical upregulation of GLI transcription factors.A Top: GLI1 protein levels in SCR, C9 and C15 cells. GAPDH is shown as loading control. Bottom: expression of PTCH1, GLI1 and GLI2 by qPCR. Expression levels were normalized to GAPDH mRNA and represented as fold change relative to control SCR cells (n = 3). B Representative images for IHC staining of GLI1 and negative control (NoAb- no primary antibody) in sections of xenograft tumours generated with SCR and C9 cells. C Effect of increasing concentrations of GANT61 on viability of SCR, C9 and C15 cells after 72 h (n = 3). The inset displays the calculated IC_50_ in each cell line. D Effect of 0.5 μM KAAD-cyclopamine (KAAD-CP), 5 μM GANT61 or vehicle (DMSO) on the viability of SCR, C9 and C15 cells after 72 h (n = 3). E Effect of 5 μM GANT61 or vehicle control (DMSO) on the colony formation ability of SCR, C9 and C15 cells upon culture for 10 days (representative images at the end of the experiment). F Number of colonies per 1000 plated cells at day 10 in the same conditions as (E) (n = 3). G GLI1 protein levels in SCR, C9 and C15 cells transduced with lentiviruses encoding control scrambled (shScr) or GLI1 targeting (shGLI1) shRNAs. H Viability of SCR, C9 and C15 cells following 72 h of transduction with shScr-lenti or shGLI1-lenti in complete medium. Data are expressed as % of shScr (n = 3). I Change in cell number of SCR cells in complete medium 48 h after transfection with empty vector, Gli1 K512R or Gli1(2-413)-VP16, compared to untransfected C9 and C15 cells (n = 3). J Representative immunoblot depicting Gli1 protein levels following ectopic expression of Flag–Gli1-K518R and/or Gli1(2-413)–VP16 in SCR cells used in the proliferation assay shown in (I). The top blot was developed with a mix of anti-Gli1 and anti-VP16 antibodies. β-actin was used as loading control. All quantitative data represent the mean ± SEM of independent biological repeats (indicated n), performed in technical replicates. ns not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Given the unexpected upregulation of GLI1 in PTCH1 CTD mutant cells, we performed a series of experiments to determine if increased GLI signalling was implicated in their augmented oncogenic properties. Treatment of the cell lines with the dual GLI1/GLI2 inhibitor GANT61 blocked proliferation in SCR, C9 and C15 cells in a dose-dependent manner, although C9 and C15 cells were more sensitive to lower concentrations of GANT61 (Figs. 2C and S3). Importantly, the SMO inhibitor KAAD-cyclopamine did not affect cell viability of any of the cell lines, while GANT61 had a strong inhibitory effect (Figs. 2D and S3), which ruled out the possibility that GLI1 upregulation in PTCH1 CTD cells is caused by activation of canonical Hh signalling. In addition, GANT61 dramatically reduced colony formation of C9 and C15 clones to the level of SCR cells (Fig. 2E, F). To confirm the specific effect of GANT61, we stably knocked down GLI1 expression using an shGLI1-encoding lentivirus, while using a scrambled shScr as control (Fig. 2G). In agreement with the cytotoxic effect of GANT61, silencing of GLI1 reduced viability in C9 and C15 cells but not in SCR cells (Fig. 2H), suggesting that GLI1 upregulation is essential for the growth advantage imparted by truncation of the PTCH1 CTD. Next, we tested whether expression of a constitutively active form of GLI1 in SCR cells was sufficient to impart a proliferative advantage. Two different active GLI1 mutants were transiently transfected into SCR cells: an acetylation-deficient mutant (GLI1 K518R) and a Gli1-VP16 fusion protein which render GLI1 constitutively active [18]. Both active GLI1 mutants increased proliferation of SCR cells to a similar level of C9 and C15 cells (Fig. 2I. Expression of the plasmids was verified by Western blot analysis (Fig. 2J). These results suggest that GLI1 upregulation in CTD mutant clones is exclusively through non-canonical pathways independent of SMO, and that GLI upregulation is essential for their more aggressive cancer behaviour. Finally, we investigated if colon cancer cell lines that naturally acquired a PTCH1 CTD truncation are highly dependent on GLI1, like the CRISPR/Cas9 engineered C9 and C15 clones. Screening of the Cancer Cell Line Encyclopedia database (Broad Institute) identified indel mutations in the PTCH1 CTD in LoVo and SW480 cells. Culture of both cell lines in the presence of low GANT61 concentrations (5 μM) strongly reduced cell numbers over time (Fig. 3A, C). Stable knockdown of GLI1 using a lenti-shGLI1 vector also abolished cell proliferation in LoVo and SW480 cells (Fig. 3B, D). Altogether, these data indicate that mutations that lead to partial truncation of PTCH1 CTD confer a proliferative advantage through a non-canonical upregulation of GLI1.Fig. 3. Colorectal cancer cells with natural PTCH1 CTD indels are dependent on GLI1.A Change in LoVo cell number over time cultured in the presence of DMSO (control) or 5 μM GANT61 (n = 3). B Change in LoVo cell number over time following transduction with shScr-lenti or shGLI1-lenti (n = 3). C Change in SW480 cell number over time cultured in the presence of DMSO (control) or 5 μM GANT61 (n = 3). D Change in SW480 cell number over time following transduction with shScr-lenti or shGLI1-lenti (n = 3). All quantitative data represent the mean ± SEM of independent biological repeats (indicated n), performed in technical replicates. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Truncation of the PTCH1 CTD induces vast transcriptome changes and upregulation of multiple oncogenic pathways
To explore the mechanisms underlying the oncogenic role of PTCH1 CTD mutations, we performed transcriptome analysis in C15, C9 cells and parental cells (SCR). RNA-seq was performed in each cell line on three consecutive passages to obtain meaningful biological repeats. Data analysis revealed a significant overlap of a set of genes only expressed in C9 and C15 cells (1396 genes) but not in SCR cells (Fig. 4A), while each clone had ~400 unique genes that were not considered for further analysis. Pearson correlation analysis confirmed the similarity between C9 and C15, with the largest differences observed between C9 and SCR cells (Fig. 4B). Therefore, pooled read counts from C9 and C15 (renamed as “PTCH1 mut” group) was compared to SCR cells (renamed “WT”) to identify genes and pathways dysregulated in cells with PTCH1 CTD truncations, revealing upregulation of 3,619 genes and downregulation of 1,705 genes (Fig. 4C and Tables S2 and S3). Comparison of C9 vs. C15 transcriptomes showed differential expression of only 111 genes (Fig. S4 and Table S4), indicating a very similar transcriptome in C9 and C15 cells, despite carrying different PTCH1 CTD indels, and a divergence of the isogenic WT cells.Fig. 4PTCH1 CTD mutant clones have transcriptional upregulation of cancer-related pathways.A Venn diagram showing gene co-expression among SCR, C9 and C15 clones (n = 3 biological repeats). B Pearson correlation of gene count reads between each biological repeat of the SCR, C9, C15 groups. C Volcano plot of statistically significantly expressed genes (-log10(Padj)>1.301) between both C9 and C15 clones (“PTCH1 mut”) and SCR cells (“WT”). D Most upregulated pathways in PTCH1 mut vs. WT cells by KEEG analysis. The size of the dot represents the gene count in that category and the colour reflects the Padj value. E cAMP production in basal conditions (growth medium supplemented with DMSO) or after 10 min stimulation with 3 μM forskolin (FSK) in SCR, C9 and C15 cells (n = 3). F Representative immunoblot of phosphorylated-PKA substrates in SCR, C9 and C15 cells 48 h after transfection with empty vector (pcDNA3.1) or a plasmid encoding a constitutively active PKA (CA-PKA). GAPDH was used as a loading control. G Change in cell number of SCR, C9 and C15 48 h after transfection with empty vector (control) or a plasmid encoding a constitutively active PKA (CA-PKA) (n = 3). All quantitative data represent the mean ± SEM of independent biological repeats (indicated n), performed in technical replicates. ns not significant; **P < 0.01; ****P < 0.0001.
KEGG pathway enrichment analysis of the differentially expressed genes (DEGs) between PTCH1 mut vs. WT cells showed increase in gene signatures of cancer-related pathways, including MAPK signalling pathway, cAMP signalling pathway, Ras signalling pathway, and Basal Cell Carcinoma (Fig. 4D). Gene set enrichment analysis (GSEA) of ranked gene lists also revealed enrichment of general pathways in cancer, including Ras, MAPK, HIF1 and basal cell carcinoma, as well as pathways regulating pluripotency, cAMP and calcium signalling (Fig. S5).
Table 1 shows selected upregulated genes in both C9 and C15 clones involved in key pathways in tumourigenesis by RNA-seq. In agreement with the increased level of GLI1 in C9 and C15 cells, RNA-seq revealed increased GLI1 and GLI2 levels as well as some key GLI target genes (HHIP and PTCH2); however, no change in Hh ligands or SMO were observed and the negative regulator GLI3 was downregulated 19-fold (Table S3). Amongst the most upregulated genes were genes belonging to the epidermal growth factor receptor (EGFR) signalling pathway: EGFR and several of its ligands, components of the TGFβ signalling pathway and regulators of cAMP-dependent kinase signalling (Table 1). These transcriptomic changes suggest that multiple pathways could mediate the non-canonical upregulation of GLI1 and GLI2 by activation of GLI positive modulators and/or inhibition of negative regulators. For instance, cAMP-dependent kinase (PKA) is the main negative regulator of all GLI proteins. It phosphorylates all GLI isoforms, targeting GLI1 for proteasomal degradation and GLI2 and GLI3 for processing into transcriptional repressors [1]. The upregulation of several cAMP phosphodiesterases in PTCH1 mutant cells led us to examine cAMP and PKA activity level. As suggested by the RNA-seq changes, PTCH1 CTD mutant cells produced much lower cAMP in response to forskolin (~65–80% of SCR cells) (Fig. 4E), suggesting a reduced PKA activity. Phosphorylation of PKA substrates, containing the RRX(S/T) motif, was diminished in C9 and C15 cells compared to SCR cells (Fig. 4F). Based on this, we hypothesised that reduced PKA activity mediates their proliferative advantage through activation of GLI-dependent signalling. Transfection of a constitutively active PKA α subunit (PKA-CQR) increased phosphorylation of PKA substrates (Fig. 4F) and specifically reduced proliferation of C9 and C15 cells (Fig. 4G), supporting the notion that transcriptomic changes in cAMP signalling pathways promote proliferation of PTCH1 CTD mutant cells.Table 1. Selected upregulated genes in C9 and C15 cells compared to SCR cells.Gene symbolGene nameLog2FoldPadjustHedgehog signalling pathwayHHIPHedgehog interacting protein3.590.0031PTCH2Patched 21.779.37 × 10^-5^DISP1Dispatched RND transporter family member 11.762.32 × 10^-7^GLI1GLI family zinc finger 13.431.39 × 10^-8^GLI2GLI family zinc finger 23.101.38 × 10^-8^MOSMOModulator of smoothened1.331.99 × 10^-5^HHATHedgehog acyltransferase3.833.25 × 10^-12^EGFR signalling pathwayEGFREpidermal growth factor receptor3.941.05 × 10^-18^AREGAmphiregulin4.376.65 × 10^-8^EREGEpiregulin3.839.15 × 10^-15^NRG1Neuregulin7.690.02EGFEpidermal growth factor4.130.018BTCBetacellulin3.263.31 ×10^-12^Ras signalling pathwayRASGRP2RAS guanyl releasing protein 23.836.75 × 10^-22^RASSF2Ras association domain family member 23.351.23 × 10^-17^RASD2RASD family member 23.062.35 × 10^-12^MRASmuscle RAS oncogene homolog2.995.81 × 10^-13^RRASRAS related2.729.39 × 10^-9^RASGEF1ARasGEF domain family member 1A2.583.05 × 10^-12^RASGEF1CRasGEF domain family member 1C2.559.14 × 10^-5^RASGRP3RAS guanyl releasing protein 32.530.0033RASGRP4RAS guanyl releasing protein 42.140.0030RASGRP1RAS guanyl releasing protein 11.955.75 × 10^-5^GAB1GRB2 associated binding protein 11.965.08 × 10^-6^GAREM2GRB2 associated regulator of MAPK1 subtype 21.71GAREM1GRB2 associated regulator of MAPK1 subtype 11.282.88 × 10^−6^GAB3GRB2 associated binding protein 35.100.0016TGFβ signalling pathwaySMAD3SMAD family member 31.494.96 × 10^−^^7^SMAD2SMAD family member 21.054.46 × 10^−6^SMAD4SMAD family member 43.872.28 × 10^−22^PKA signalling pathwayPRKAR2BProtein kinase cAMP-dependent type II regulatory subunit beta3.686.73 × 10^−12^PRKACBProtein kinase cAMP-activated catalytic subunit beta2.219.55 × 10^−19^AKAP3A-kinase anchoring protein 32.241.18 × 10^−6^AKAP11A-kinase anchoring protein 111.111.26 × 10^−^^5^AKAP12A-kinase anchoring protein 123.834.27 × 10^−17^PDE4CPhosphodiesterase 4C3.720.044PDE9APhosphodiesterase 9A1.638.59 × 10^−^^6^PDE5APhosphodiesterase 5A1.190.0052PDE4APhosphodiesterase 4A1.105.57 × 10^−6^PDE4BPhosphodiesterase 4B6.189.41 × 10^−52^PDE7BPhosphodiesterase 7B5.820.00011PDE10APhosphodiesterase 10A4.580.0094PDE6APhosphodiesterase 6A4.350.0175PDE1CPhosphodiesterase 1C4.270.021PDE2APhosphodiesterase 2A4.181.09 × 10^−6^
PTCH1 CTD truncation leads to upregulation of the EGFR signalling pathway
Consistent with the RNA-seq data, qRT-PCR confirmed the mRNA a ~9 -fold and ~6 -fold upregulation of EGFR in C9 and C15 compared to SCR, and of its ligands AREG (~16-fold and ~14-fold) and EREG (~11 -fold and ~11-fold), respectively (Fig. 5A, top). EREG and AREG are commonly overexpressed in CRC at the transcriptional and protein levels and have been proposed as predictors of response to anti-EGFR therapy in RAS WT CRC [21–23]. Their levels were not significantly reduced by the GLI1/2 inhibitor GANT61 (Fig. S6), suggesting that they are not a consequence of non-canonical activation of GLI transcription. The upregulation of EGFR, AREG and EREG mRNA was maintained in explanted xenografts from C9 cells compared to SCR cells (Fig. 5A, bottom). C9 and C15 cells had increased basal ERK1/2 and Akt phosphorylation level compared SCR cells under serum starvation conditions (Fig. 5B). The EGFR inhibitor erlotinib did not affect cell proliferation of C9 and C15, unlike of SCR cells (Fig. 5C). In addition, erlotinib reduced plating efficiency in colony formation assays in C9 and C15 clones by 61% and 55%, respectively; however, the PTCH1 CTD mutant clones still formed more colonies than the PTCH1 WT cells (Fig. 5D). The partial effect of erlotinib on SCR cells is not surprising since SW620 is a KRAS mutated cell line, and thus upstream activation of EGFR signalling is expected to have only a partial effect on signalling downstream of RAS. To confirm this observation, we treated the cells with cetuximab, a blocking anti-EGFR antibody. Like erlotinib, cetuximab treatment failed to reduce proliferation of SCR, C9 and C15 cells (Fig. 5E). Instead, inhibition of MEK with trametinib reduced ERK1/2 phosphorylation (Fig. 5F) and proliferation of SCR, C9 and C15 cells in a dose dependent manner (Figs. 5G and S7). A second MEK inhibitor, UO126, also showed a strong antiproliferative effect in all groups (Fig. S7). Therefore, PTCH1 CTD mutant cells do not show a higher sensitivity to MEK inhibitors, suggesting that additional mechanisms support their proliferative advantage.Fig. 5. Upregulation of EGFR-MAPK signalling in PTCH1 CTD mutant cells.A Top: EREG, AREG, and EGFR mRNA expression in SCR, C9, and C15 cells by qPCR. Expression levels were normalized to GAPDH mRNA and represented as fold change relative to control SCR cells (n = 3). Bottom: expression of the same genes in pooled RNA from xenografts derived from SCR and C9 cells. B Representative western blot of EGFR, phospho-Akt, phospho-ERK1/2, total Akt, total ERK1/2, and GADPH in SCR, C9, and C15 cells. C Effect of 10 μM Erlotinib or vehicle control on cell proliferation of SCR, C9 and C15 cells over time (n = 3). D. Effect of 10 μM Erlotinib or vehicle control (DMSO) on the colony formation ability of SCR, C9 and C15 cells upon culture for 10 days. Left: representative images at the end of the experiment. Right: number of colonies per 1,000 plated cells at day 10 (n = 3). E Effect of 10 μg/ml Cetuximab or a control IgG on the cell number of SCR, C9 and C15 cells after 72 h treatment (n = 3). F Effect of 30 nM Trametinib or DMSO vehicle on phospho-ERK1/2 levels in SCR, C9 and C15 cells. G Dose dependent reduction of cell proliferation in SCR, C9, and C15 cells treated with Trametinib for 72 h (n = 3). All quantitative data represent the mean ± SEM of independent biological repeats (indicated n), performed in technical replicates. ns = not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Increased tumourigenesis of PTCH1 CTD truncated cells is dependent on hyperactivation of PI3K
Next, we tested the role of PI3K/Akt signalling in the enhanced proliferation of C9 and C15 cells. A low concentration of the PI3K inhibitor BKM-120 (buparlisib) decreased P-Akt (Fig. 6A) and specifically reduced the proliferation rate of both PTCH1 CTD mutant clones without affecting proliferation of isogenic cells expressing WT PTCH1 (Fig. 6B). BKM-120 IC_50_ was lower in C9 and C15 clones compared to SCR cells, in agreement with their increased sensitivity to low concentrations (Fig. 6C). Colony formation was also reduced dose-dependently by BKM-120 in PTCH1 mutant clones (Fig. 6D, E). Given the central role of GLI1 in the proliferative advantage of PTCH1 CTD mutant cells, we investigated if normalisation of the cell’s behaviour by PI3K inhibition was mediated by changes in GLI1. BKM-120 treatment reduced GLI1 mRNA levels (Fig. 6F), suggesting that PI3K signalling is partly required for upregulation of GLI1 downstream of the PTCH1 CTD mutation. A second PI3K inhibitor, LY294002, also caused a reduction in GLI1 expression (Fig. S8). We have previously reported that GLI activation requires PI3K signalling through Akt, and that the latter opposes PKA-mediated phosphorylation and degradation of GLI2 [24]. Therefore, we tested whether a dominant negative Akt1 mutant (Akt1 S473A/T308A) could mimic the effect of PI3K inhibition. Transfection of SCR, C9 and C15 cells with dnAkt specifically abolished the enhanced proliferation of C9 and C15 cells (Fig. 6G). Interestingly, lysates of cells with lentiviral-mediated knockdown of GLI1 from Fig. 2 showed a strongly reduced P-Akt signal and, to a lower extent, P-ERK1/2 levels (Fig. 6H), suggesting that sustained PI3K/Akt activation in PTCH1 CTD mutant cells is also propelled by GLI1, leading to a positive feedback loop (Fig. 6I). Notwithstanding, increasing Akt signalling by expression of a constitutively active Akt1 mutant (myr-Akt1) was insufficient to rescue proliferation in cells treated with the Gli1/2 inhibitor GANT61 (Fig. S9). These observations indicate that the increased tumourigenic capacity of PTCH1 CTD truncated cells is reliant on increased GLI1 expression and activity, which is maintained by hyperactivation of PI3K/Akt pathway.Fig. 6A positive PI3K/Akt-GLI loop in PTCH1 mutant cells underlies their growth advantage.A Effect of 2 μM BKM-120 or DMSO vehicle on phospho-Akt(S473) and total Akt levels in SCR, C9 and C15 cells. GAPDH was used as a loading control. B Effect of 0.5 μM BKM-120 or DMSO (control) on cell number of SCR, C9 and C15 at 72 h (n = 3). C Reduction of viability in SCR, C9, and C15 cells treated with increasing concentrations of BKM-120 for 72 h (n = 3). D Effect of 0.25, 0.5 and 1 μM BKM-120 or vehicle control (DMSO) on the colony formation ability of SCR, C9 and C15 cells upon culture for 10 days (representative images at the end of the experiment). E Quantification of number of colonies formed per 1,000 plated SCR, C9 and C15 cells treated with DMSO or 0.5-1 μM BKM120 for 10 days (n = 3). F Effect of 0.5 μM BKM-120 or vehicle control on GLI1 mRNA levels in SCR, C9 and C15 cells by qPCR. Values are normalised to GAPDH expression and represent the ΔΔCt (n = 3). G Top: Representative immunoblot of dominant negative Akt (HA-tag) and total Akt (endogenous plus transfected) in SCR, C9 and C15 cells 48 h after transfection with empty vector (pcDNA3,1) or dnAkt (S473A/T308A) (dnAkt). Bottom: Change in cell number of SCR, C9 and C15 48 h after transfection with empty vector (control) or a plasmid encoding dnAkt (n = 3). H Protein levels of GLI1, phospho-Akt, total Akt, phospho-ERK1/2, total ERK1/2 and GAPDH in SCR, C9 and C15 cells transduced with lentiviruses encoding control scrambled (shScr) or GLI1 targeting (shGLI1) shRNA. I Proposed model of a positive feedback loop between GLI1 and PI3K triggered by PTCH1 CTD truncations. Dotted lines indicate potentially indirect mechanisms. All quantitative data represent the mean ± SEM of independent biological repeats (indicated n), performed in technical replicates. ns = not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Discussion
The canonical Hedgehog (Hh) signalling pathway, driving activation of SMO and GLI1, underlies tumourigenesis of basal cell carcinoma and Shh-type medulloblastoma; however, its role in CRC and other cancers of epithelial origin remains controversial. While CRC patients with increased expression of GLI1 have decreased overall survival [25], the pharmacological SMO inhibitor vismodegib did not show additional benefits to standard treatment in patients with metastatic colon cancer [24]. Here, we link frameshift mutations in the PTCH1 gene that result in early truncation of the PTCH1 CTD in CRC with SMO-independent upregulation of GLI1. In this study, we show that CRC cells engineered to endogenously express PTCH1 with a truncated CTD, found in ~4% of CRC patients, are characterized by increased proliferation, reduced apoptosis and increased anchorage-independent growth in vitro and form larger tumours in vivo compared to isogenic cells expressing WT PTCH1. PTCH1 CTD mutant cells show non-canonical upregulation of GLI1, which was surprising as the resulting truncations in the PTCH1 CTD did not impair SMO control. GLI1 upregulation is a hallmark of increased transcriptional activity of the GLI transcription factor family. The other family members, GLI2 and GLI3, are constitutively expressed and activated in response to diverse stimuli, mainly by Hh binding to PTCH1 leading to SMO activation, but also by crosstalk with other pathways in a SMO-independent manner. When GLI2 and GLI3 are activated, they induce GLI1 expression by binding to GLI-responsive elements in its promoter. Phosphorylation of GLI1 and GLI2 by PKA triggers additional post-translational modifications that mark them for proteasomal degradation, while phosphorylation of GLI3 leads to partial processing into a transcriptional repressor [1, 26]. We propose that the reduced cAMP concentrations and lower PKA activity observed in PTCH1 CTD mutant cells contributes to the increased GLI activity.
We have previously shown that GLI transcriptional activity is stimulated by crosstalk with the MAPK and PI3K signalling pathways [26, 27]. RNA-seq revealed vast transcriptome changes due to premature truncation of PTCH1, which pointed towards increased MAPK and PI3K/Akt signalling, confirmed at the protein level as increased ERK1/2 and Akt phosphorylation, despite all cells being KRAS mutated. However, our results indicate that the growth advantage of the PTCH1 mutant cells is primarily driven by increased PI3K/Akt signalling, acting through increasing GLI1 activity and expression. This feature explains the insensitivity to KAAD-cyclopamine that we observed and predicts lack of response to vismodegib and other FDA-approved SMO inhibitors, as previously reported [28]. Moreover, the findings suggest a potential vulnerability of PTCH1 CTD mutant CRC cases to buparlisib or other PI3K inhibitors.
The results also showed that GLI1 upregulation is essential for maintaining the more aggressive phenotype of PTCH1 CTD mutant cancer cells in vitro and in vivo. Unlike approaches that target SMO, inhibition or silencing of GLI effectively blocked proliferation and colony formation capacity of PTCH1 CTD truncated CRC cells. Moreover, blocking GLI1 strongly impaired doubling of LoVo and SW480 CRC cell lines, in which the PTCH1 CTD mutations are naturally occurring. Interestingly, inhibition or silencing of GLI1 also reduced P-Akt levels in PTCH1 CTD mutant cells. Together with the loss of proliferative advantage in PTCH1 CTD mutant cells by PI3K inhibition, we propose that upregulation of GLI1 increased oncogenic properties in these cells through activation of a GLI1/PI3K positive signalling loop.
PTCH1 CTD mutations are much more prevalent, if not exclusive, in colon cancers derived from the ascending and transversal colon, characterised by higher incidence of BRAF mutations and lower incidence of KRAS mutations than left-sided colon cancers. Amongst the most upregulated genes in PTCH1 CTD mutant cells were EREG and AREG, which are predictive biomarkers for anti-EGFR therapy responsiveness in RAS WT CRC patients [22, 23, 28]. However, upregulation of EGFR, AREG and EREG in cells carrying CTD-truncated PTCH1 did not translate into increased sensitivity to the EGFR inhibitors erlotinib and cetuximab nor to the MEK inhibitors trametinib and UO126, likely due to a compensatory increase in survival through activation of the PI3K-GLI1 loop. This suggests that the PTCH1 mutational status could be a predictor of responsiveness to anti-EGFR therapy and anti-PI3K/Akt therapeutics. Therefore, we propose that PTCH1 sequencing could be beneficial for stratification of patients with right-sided colon cancer to guide the choice of better targeted therapeutic interventions.
In conclusion, this study supports a pathogenic role of frameshift mutations in exons 22-23 of PTCH1, encoding the CTD. Our data provide mechanistic insights of the effect of PTCH1 mutations in CRC and identify therapeutic vulnerabilities.
Supplementary information
Table S1 Table S2 Table S3 Table S4 Supplementary figures