Analysis of Primary Cilium Expression and Hedgehog Pathway Activation in Mesothelioma Throws Back Its Complex Biology

Simple Summary Primary cilium (PC) is a solitary organelle protruding from the cellular membrane of almost all mammalian cells. It transduces signals from external to intracellular compartments regulating many physiological pathways. Dysfunction in PC can lead to a number of pathologic conditions including cancer. Among the pathways transduced by PC, the Hedgehog (Hh) signaling is the focus of recent clinical studies. Malignant pleural mesothelioma (MPM) is a cancer of the membranes covering the lungs with poor prognosis and limited therapeutic options. Although Hh is known to be over-activated in MPM, it responds poorly to smoothened (SMO)/Hh inhibitors. We investigated for the first time the status of PC in MPM tissues, demonstrating intra- and inter-heterogeneity in its expression. We also correlated the presence of PC with activation of the Hh pathway, providing uncovered evidence of the co-existence of a PC-independent regulation of the Hh signaling in MPM. Abstract The primary cilium (PC) is a sensory organelle present on the cell surface, modulating the activity of many pathways. Dysfunctions in the PC lead to different pathologic conditions including cancer. Hedgehog signaling (Hh) is regulated by PC and the loss of its control has been observed in many cancers, including mesothelioma. Malignant pleural mesothelioma (MPM) is a fatal cancer of the pleural membranes with poor therapeutic options. Recently, overexpression of the Hh transcriptional activator GL1 has been demonstrated to be associated with poor overall survival (OS) in MPM. However, unlike other cancers, the response to G-protein-coupled receptor smoothened (SMO)/Hh inhibitors is poor, mainly attributable to the lack of markers for patient stratification. For all these reasons, and in particular for the role of PC in the regulation of Hh, we investigated for the first time the status of PC in MPM tissues, demonstrating intra- and inter-heterogeneity in its expression. We also correlated the presence of PC with the activation of the Hh pathway, providing uncovered evidence of a PC-independent regulation of the Hh signaling in MPM. Our study contributes to the understanding MPM heterogeneity, thus helping to identify patients who might benefit from Hh inhibitors.


Introduction
Primary cilium (PC) is a microtubule-based non-motile sensory organelle located on the surface of all mammalian cells that, similar to an antenna, gathers extracellular therapy is a combination of platinum agents plus folate antimetabolites with the optional addiction of Bevacizumab [30]. Recently, nivolumab plus ipilimumab was introduced for the treatment of selected unresectable mesothelioma [31]. However, due to the lack of markers for patient stratification, personalized therapies remain an urgent need.
Many studies have shown that the Hh pathway is deregulated in most MPM [32][33][34] and that high levels of downstream transcriptional targets GLI1 and SMO are correlated with a poor prognosis [33,35]. However, the expression of Hh components in MPM is not correlated with the expression of their upstream stimuli Shh and SMO [32] and there is a lack of direct evidence linking GLI activity to Hh. Accordingly, the direct inhibition of GLI1 in MPM demonstrated greater efficacy then SMO-i [32,[35][36][37]. Unfortunately, no direct-acting GLI1 inhibitors have already entered clinical trials [38].
Despite these results, only two in vitro studies have investigated the presence of PC in MPM as a marker of SMO/Hh activation [37,39,40]. This, therefore, remains an important issue for the selection of patients who may respond to SMO inhibitors and for understanding the mechanisms underlying resistance to this class of drugs. In fact, unlike other cancers, mutations in the components of the pathway are rare in MPM and cannot explain its resistance to treatments [31,38].
In light of the good clinical results achieved with SMO-i in other tumor types, and since MPM is a highly heterogeneous cancer, we believe that this class of compounds can be repurposed to treat a subset of MPM. However, as with other targeted therapies, proper predictive biomarkers are needed.
On these premises, here we investigated the presence of PC in MPM tissues and in a panel of primary cell lines, correlating its presence with the Hh activation. Our finding demonstrated a high heterogeneity in PC expression, and its loss is a frequent event in the more aggressive histotypes.
Our study provided uncovered evidence that the Hh pathway in MPM can also be activated by non-canonical/PC-independent mechanisms explaining, at least in part, the lack of response of MPM to SMO/Hh inhibitors upstream of GLI1.
Furthermore, our analysis adds new insights into the role of PC expression in cancer. Although our findings need to be confirmed in a larger cohort, this study identified a new cellular marker of MPM heterogeneity that could be correlated to MPM progression, providing another puzzle piece in understanding the complexity of MPM biology, in an attempt to offer personalized treatments for this tumor in the near future.

Cell Lines and Culture Conditions
Mesothelioma cell lines were isolated from patients who underwent multiple thoracoscopic biopsies at the Thoracic Surgery Unit of the Siena university Hospital (Siena, Italy) prior and/or after chemotherapy, between 2016 and 2022. All specimens analyzed were from patients diagnosed for pleural mesothelioma with their written informed consent. Two patients with pleurisy and two with reactive mesothelial hyperplasia were included in the non-cancer group. Human investigations were performed with approval from the Research Ethics Committee (Comitato Etico Regione Toscana-Area Vasta Sud Est) approval (#CCMESOLUNG, #Mi-PP). The study conformed to the standards of the Declaration of Helsinki. All cases were classified as mesothelioma according to the 2021 morphologic and immunophenotypic WHO classification [27].
Tissues and pleural effusions were processed as previously described [41]. Briefly, the samples were transported to the laboratory for primary cell culturing within 30 min of collection. The solid tissues were minced into small pieces, 1-3 mm, and then incubated in complete medium supplemented with collagenase type I from Clostridium histolyticum (Thermo Fisher Scientific, Waltham, MA, USA, Cat #17100017) at 200 U/mL concentration for 1 h to digest collagen and release tumor cells. Pleural effusions were centrifuged at 400× g and washed twice in PBS. Macrophages, red blood cells and lymphocytes were the The mesothelial origin of patients-derived cultures was assessed by immunohistochemical analysis (IHC) for a panel of markers (WT-1, α-SMA, CD31, CD34). The cell lines were also analyzed by transmission electron microscopy (TEM) for the presence of microvilli on the cellular membrane. Briefly, the cells were fixed in 2.5% cacodylate-buffered glutaraldehyde, post-fixed in buffered 1% osmium tetroxide, dehydrated, and embedded in epoxy resin. Ultrathin sections were double-stained with uranyl-acetate replacement stain (UAR) and lead citrate and examined with a Philips 208S Transmission Electron Microscope.
Clinical information on patients (age, gender, diagnosis, and treatment) from whom the cell lines have been established are summarized in Table S1. Non-tumoral mesothelial cell lines were established from the pleural effusion of non-oncologic patients for whom the final diagnosis was negative for MPM.
Cells were then incubated overnight at 37 • C, 5% CO 2 . After 24 h, the medium was changed, and cells were left in culture for three days. At 72 h post-transduction, cells were incubated in their regular growth medium containing Neomycin (10 µg/m) to select for stable hTERT transduced cells. Selection was carried out until the cultures were devoid of non-resistant cells (<14 days) and surviving cells were further expanded in standard medium and routinely passaged.
Non-malignant primary mesothelial cells HMC12 and HMC13 were grown with the addition of 20 ng/mL hr-EGF (Sigma-Aldrich, ST. Louis, MO, USA).
All cell lines were grown at 37 • C in a humidified atmosphere with 5% CO 2 , passaged every 3-5 days, and routinely tested for mycoplasma contamination using MycoSPY ® -PCR Mycoplasma Test Kit (BIONTEX Laboratories, München, Germany) according to the manufacturer's instructions.
Unspecific signal was evaluated for each antibody using a control condition without primary antibody and with a non-specific antibody. Images were acquired using the Zeiss Axio microscope (Zeiss Laboratories, White Plains, NY, USA). The proportion of ciliated cells was determined across multiple fields of view for each condition.

Real-Time Reverse Transcriptase-Quantitative PCR (RT-qPCR)
Total RNA was isolated from cell lines using the RNeasy Mini kit (Qiagen, Hilden, Germany Cat #74106) as previously described [42]. RNA concentration was determined using a NanoDrop™ ND-1000 (Thermo Fisher). Complementary DNA was synthesized from 500 ng of RNA using the iScript cDNA Synthesis Kit (Bio-rad, Hercules, CA, USA Cat #1708891BUN) and amplified in the LightCycler™ instrument (Roche Applied Sciences) using SsoAdvanced™ Universal SYBR ® Green Supermix (Bio-rad, Cat #1725274) according to the manufacturer's instructions.
The primers used were from Bio-Rad: GLI1 Assay IDqHsaCED0043346, PTCH1 Assay ID qHsaCED0001809, C-MYC Assay ID qHsaCID00012921. The housekeeping glyceraldehyde-3-phosphate dehydrogenase (GAPDH Assay ID qHsaCED0038674) gene was used to normalize the expression of the genes of interest. Gene expression levels were calculated by the 2 −∆∆Ct method [43].

Immunohistochemical Analysis
Formalin-fixed, paraffin-embedded (FFPE) tumor specimens were obtained from the Section of Pathology of the Siena University Hospital, Siena, Italy. Immunohistochemistry staining was performed on 4-µm-thick paraffin sections adjacent to those that had been stained for Haematoxylin & Eosin (H&E). Clinical information about mesothelioma patients is summarized in Table S2.
Frequency of cilia was determined by dividing the number of ciliated cells by the total number of nuclei. MPM with PC frequency less than 15% were considered as negative.
The assessment of nuclear GLI1 expression levels included the staining intensity and the percentage of stained cells. The staining intensity was scored as 0 = no staining, 1 = low expression, 2 = moderate expression, 3 = strong expression.
Statistical analysis of the correlation between GLI1 expression was performed using the IBM SPSS Statistics Software, version 25 for Mac. Pearson's correlation coefficient was used to measures the statistical relationship, or association, between two continuous variables. Kendall's tau-b correlation coefficient was used for nonparametric measure of the strength and direction of association that exists between two variables measured on at least an ordinal scale.

TCGA Database Analysis
To investigate the association between GLI1 gene expression level and MPM histotype, we performed a differential expression analysis of the single GLI1 gene in the biphasic and epithelioid histotypes of MPM samples. For this analysis, from The Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/; accessed on 28 March 2022), a subset of 19 publicly available samples' data are retrieved. We selected common features (male sex and tumor stage 3) in order to harmonize the dataset and thus make it more suitable for differential expression analysis using DESeq2. Then, "htseq_count" files (10 epithelial and 9 biphasic) obtained from RNA sequencing experiments were extracted by download and differential expression analysis was performed with the R tool (v4.2.1, https://www.Rproject.org/; accessed on 26 August 2022). Using the PlotCounts function, it was possible to investigate differential expression (up or down-regulation of the gene) and significance at the individual gene level. Differential expression is calculated by the DESeq2 package based on the Log2FoldChange value: if Log2FoldChange is <0, the gene is down-regulated; if Log2FoldChange > 0 it will be up-regulated. In the case where Log2FoldChange = 0, the gene will not be differentially expressed [44]. Statistical significance is calculated by the DESeq2 package by estimating the mean and standard deviation in the two groups of samples compared for each gene. From this initial analysis, the algorithm derives p-values, which are then corrected with multiple tests to obtain p-value-adjusted (padj). Statistical significance is then estimated from the padj for each gene; the lower it is, the higher the significance will be (padj < 0.05). In conclusion, the differential and significative genes were screened using the following criteria: Log2FoldChange = 0 and adjusted p-values (padj) < 0.05.

Malignant Pleural Mesothelioma (MPM) Shows an Heterogeneous Expression of Primary Cilium (PC)
In order to investigate the status of PC in MPM, we performed an exploratory analysis on 28 FFPE pleural tissues (24 mesothelioma and 4 non-oncological specimens).
IHC analysis showed PC expression in seven MPM samples (29.1%) and loss in 13 (54.2%). The latter group includes all the biphasic (n = 4) and the desmoplastic (n = 1) MPM. The correlation between cilium expression and histology (epitheliomorphic vs. biphasic/desmoplastic mesothelioma) is statistically significant (p = 0.020). Four tissues showed heterogeneous intra-tumor expression (16.7%) ( Figure 1). Among patients for whom biopsies were available prior to and after chemotherapy (n = 4; 8 = samples), only in one patient did we observe a change in PC expression, with a heterogeneous presence of PC prior chemotherapy, and loss after the treatment (data not shown). However, it remains to be ascertained whether this effect is due to a differential response to chemotherapy between PC-positive and PC-negative cells.   All the non-neoplastic cases retain PC. The results are summarized in Table 1.  as in the tissue of origin, except for MMP1 which derived from a specimen with an intraheterogeneous expression of PC. We hypothesize that, during the generation of MMP1 cell line, the cells lacking PC preferentially adapted to in-vitro culture.
Regarding long term cultures, MSTO-211H cells lose PC, while NCI-H2052 cells express it. The results were summarized in Table 2. Representative images of selected PCpositive and -negative primary MPM cell lines and PC-positive mesothelial cells are shown in Figure 2.

Cell Lines Lacking PC Had Hh/GLI1 Pathway Activated
The Hh is one of the most important oncogenic signaling starting from PC, and one of the top five deregulated pathways in MPM [45] (Supplementary Table S3). For the poor response of MPM to SMO/Hh inhibitors, by qRT-PCR and IF analysis, we explored the activation of Hh in PC-negative MPM cells (MMP1, MMP4 and MSTO-211H), compared to HMC7 mesothelial cells. We also included in the analysis PC-positive MPM cell lines (MMP18, MMP21, MMP23 and NCI-H2052) and normal mesothelial cells LP-9.
We analyzed the expression of bona fide Hh/GLI transcription targets GLI1, PTCH1 and c-MYC. As shown in Figure 3 and in Table S4, PC-negative cells (gray bars) showed upregulation of all the analyzed genes, even though the GLI1 transcript in MMP1 did not reach statistical significance. Conversely, in all the PC-positive cells (green bars), the levels of GLI1, PTCH1, and c-MYC are comparable to the those of normal mesothelial cells, with the exception of MMP23, which showed a statistically significant upregulation of GLI1, and NCI-H2052, which showed a very significant upregulation of c-MYC. However, it should be emphasized that both MSTO-211H and NCI-H2052 cell lines have amplification of the c-MYC gene (8q24.1), which is both a target and an activator of GLI1 [46].   To confirm these results, we performed IF analysis for GLI1 in primary MPM cells, to assess its nuclear localization. Contrary to qRT-PCR analysis, and thus regardless of PC expression, all MPM cell lines showed nuclear GLI1, except for MMP21 in which GLI1 was undetectable, similar to HMC7 mesothelial cells. Representative images of nuclear GLI1 IF in PC-positive and PC-negative cells are shown in Figure 4.
To confirm these results, we performed IF analysis for GLI1 in primary MPM cells, to assess its nuclear localization. Contrary to qRT-PCR analysis, and thus regardless of PC expression, all MPM cell lines showed nuclear GLI1, except for MMP21 in which GLI1 was undetectable, similar to HMC7 mesothelial cells. Representative images of nuclear GLI1 IF in PC-positive and PC-negative cells are shown in Figure 4. It is to note that the MMP18 cell line did not show significant upregulation of Hhrelated genes by qRT-PCR analysis. However, from the IF analysis, these cells showed the nuclear localization of GLI1, indicating that the transcript levels alone are not a sufficient marker of Hh activation.

GLI1 in MPM Tissues Is Activated Independently from PC Expression
To confirm the results obtained in MPM cell lines, we analyzed GLI1 activation in nineteen MPM (five PC-positive and fourteen PC-negative), according to the availability of tissue specimens.
Positive IHC staining of nuclear GLI1 was observed in twelve out of nineteen MPM samples (63.2%). Among these, three are PC-positive (25%), and nine are PC-negative (75%).
Seven MPM samples (36.8%) did not show nuclear GLI1 staining (Score = 0); of these, two are PC-positive (28.6%), five are PC-negative (71.4%). It is to note that the MMP18 cell line did not show significant upregulation of Hhrelated genes by qRT-PCR analysis. However, from the IF analysis, these cells showed the nuclear localization of GLI1, indicating that the transcript levels alone are not a sufficient marker of Hh activation.

GLI1 in MPM Tissues Is Activated Independently from PC Expression
To confirm the results obtained in MPM cell lines, we analyzed GLI1 activation in nineteen MPM (five PC-positive and fourteen PC-negative), according to the availability of tissue specimens.
Positive IHC staining of nuclear GLI1 was observed in twelve out of nineteen MPM samples (63.2%). Among these, three are PC-positive (25%), and nine are PC-negative (75%).
Representative images are shown in Figure 5. The results are summarized in Table 3.
Representative images are shown in Figure 5. The results are summarized in Table 3.

GLI1 Status Does Not Correlated with MPM Histotype
Finally, using the TCGA database, we investigate if GLI1 is differentially expressed in epithelioid mesothelioma compared to biphasic MPM. We selected a subset of 19 samples with common characteristics (male sex and tumor stage 3).
From this analysis, the GLI1 gene (Ensembl ID: ENSG00000111087.8) has Log2FoldChange < 0, indicating that the gene is downregulated in epithelial tissues, but it does not appear to be significantly differentially expressed between epithelial and biphasic tissue as padj = 0.81 (padj > 0.05) ( Figure 6).  Table 3. IHC analysis of GLI1 expression in MPM. n = sample size.

GLI1 Status Does Not Correlated with MPM Histotype
Finally, using the TCGA database, we investigate if GLI1 is differentially expressed in epithelioid mesothelioma compared to biphasic MPM. We selected a subset of 19 samples with common characteristics (male sex and tumor stage 3).
From this analysis, the GLI1 gene (Ensembl ID: ENSG00000111087.8) has Log2FoldChange < 0, indicating that the gene is downregulated in epithelial tissues, but it does not appear to be significantly differentially expressed between epithelial and biphasic tissue as padj = 0.81 (padj > 0.05) ( Figure 6).
According to this analysis, we did not find a statistically significant correlation between nuclear GLI1 and MPM histology (Epithelioids vs. biphasic/desmoplastic MPM) (p = 0.845). Figure 6. Differential expression analysis of GLI1 gene between biphasic and epithelioid MPM. The analysis was drawn using the Deseq2-package, PlotCounts. The GLI1 gene has Log2FoldChange < 0, indicating that the gene is down-regulated in epithelial tissues, but it is not significantly differentially expressed since padj = 0.81. Y axis: normalized count; X axis: Histological subtype. Epithelial = yes; non-epithelial = no.
According to this analysis, we did not find a statistically significant correlation between nuclear GLI1 and MPM histology (Epithelioids vs. biphasic/desmoplastic MPM) (p = 0.845).

Discussion
The PC is a conserved organelle protruding from the cellular membrane of all mammalian cells, involved in the regulation of many important pathways. An exhaustive understanding of all its functions is still incompletely defined and is an evolving field of research.
In cancer, PC loss is related to both pro-and anti-tumor functions that are tissue-and cancer-specific. Restoring PC functions represents an intriguing target in the field of personalized medicine.
Hh signaling is involved tissue development and homeostasis and is regulated through activators and repressors enriched in the cilium, and Hh deregulation is a feature of many cancers.
MPM is a tumor of the pleural membranes with a dismal prognosis and limited therapeutic option. For its high intra-and inter-heterogeneity, MPM cannot be considered a single tumor type, therefore requires patients-tailored therapies and reliable markers for patient stratification.
Due to the limited progress made in the last decades in MPM treatment, the repositioning of drugs already approved in the clinical practice for other tumor types represents a turning point in accelerating the introduction of novel treatments for this devastating disease [47,48]. In this regard, in the light of the encouraging results obtained with SMO/Hh inhibitors in different clinical trials, and since no direct-acting GLI1 inhibitors have entered clinical trials [38], the use of SMO-i in MPM has been investigated by many authors [37,49,50]. Figure 6. Differential expression analysis of GLI1 gene between biphasic and epithelioid MPM. The analysis was drawn using the Deseq2-package, PlotCounts. The GLI1 gene has Log2FoldChange < 0, indicating that the gene is down-regulated in epithelial tissues, but it is not significantly differentially expressed since padj = 0.81. Y axis: normalized count; X axis: Histological subtype. Epithelial = yes; non-epithelial = no.

Discussion
The PC is a conserved organelle protruding from the cellular membrane of all mammalian cells, involved in the regulation of many important pathways. An exhaustive understanding of all its functions is still incompletely defined and is an evolving field of research.
In cancer, PC loss is related to both pro-and anti-tumor functions that are tissueand cancer-specific. Restoring PC functions represents an intriguing target in the field of personalized medicine.
Hh signaling is involved tissue development and homeostasis and is regulated through activators and repressors enriched in the cilium, and Hh deregulation is a feature of many cancers.
MPM is a tumor of the pleural membranes with a dismal prognosis and limited therapeutic option. For its high intra-and inter-heterogeneity, MPM cannot be considered a single tumor type, therefore requires patients-tailored therapies and reliable markers for patient stratification.
Due to the limited progress made in the last decades in MPM treatment, the repositioning of drugs already approved in the clinical practice for other tumor types represents a turning point in accelerating the introduction of novel treatments for this devastating disease [47,48]. In this regard, in the light of the encouraging results obtained with SMO/Hh inhibitors in different clinical trials, and since no direct-acting GLI1 inhibitors have entered clinical trials [38], the use of SMO-i in MPM has been investigated by many authors [37,49,50].
However, despite Hh being one of the top five deregulated pathways in MPM [45], and that GLI1 and SMO levels are correlated with a poor prognosis, the response of MPM to SMO/Hh-i is poor. While numerous in-vitro studies support a therapeutic value for SMO/Hh-i in MPM, limited information remains available from clinical studies. Two phase I solid-tumor trials, overall including five mesothelioma patients treated with the SMO inhibitors vismodegib (GDC-0449) or sonidegib (LDE225), reported no clinical benefit in MPM [49,51]. Only one case report described a durable response with vismodegib in a MPM presenting a loss-of-function mutation in PTCH1 [50]. However, mutations affecting the components that control the Hh are relatively rare in MPM compared to other SMO/HH-i resistant tumors [34,52] and cannot be considered a general mechanism of resistance.
Due to the role of Hh pathway in MPM, and the poor response to SMO/Hh inhibitors in this tumor, here, we hypothesized that Hh can be controlled by both SMO-dependent and SMO-independent mechanisms.
On these premises, and since PC is indispensable for the canonical activation of Hh pathway, here we have investigated the correlation between the presence of PC and the Hh activation in MPM tissues specimens, in non-tumoral pleural tissues, and in a panel of primary MPM cell lines.
While non-malignant mesothelial tissues retained PC, it is lost in 75% of MPM, mainly related to more aggressive phenotypes. Accordingly, the correlation between PC expression and histology is statistically significant. This observation is in line with a recent transcriptomic analysis showing increased expression of PC-related genes in epithelioid mesothelioma compared to the biphasic phenotype [53].
In agreement with IHC analysis, we observed the presence of PC in four non-tumoral mesothelial cell lines, and in four out of eight epithelioid MPM cell lines.
Then, we investigated the correlation between PC positivity and nuclear localization of GLI1 in tissues and cell cultures. Among 24 MPM specimens analyzed, 63.2% had nuclear GLI1 and, of these, 75% were PC-negative. Statistical analysis shows that the correlation between nuclear GLI1 and cilium expression is not significant supporting our hypothesis that the Hh pathway can be activated in MPM through non-canonical PC-independent pathways. We obtained similar results in primary MPM cell lines in which the absence of PC did not preclude the Hh activation.
It should be noted that although qRT-PCR analysis in PC-positive cell lines did not show high transcript levels of all Hh components, in two cell lines we observed a nuclear localization of GLI1. Although this may seem contradictory, it is in agreement with a recent analysis demonstrating that the expression levels of Hh components cannot be markers of Hh ligand-dependent activation of the pathway [54].
Furthermore, in agreement with recent studies demonstrating the synergistic effects of the simultaneous inhibition of SMO and GLI1 [32,55], it is conceivable that in PC-positive MPM the activation of the Hh signaling can be fueled by both canonical and non-canonical pathways. This could explain the poor response of MPM to SMO-i and also sustain the importance of Hh/GLI1 for MPM growth.
These results should be interpreted in the light of several limitations. First, the sample size does not allow drawing robust conclusions about the prognostic significance of PC frequency in MPM. Furthermore, in a larger series, a stratification of the morphological parameters, including nuclear pleomorphism, mitotic count, necrosis, architectural pattern, and expression of molecular markers, such as BAP1, CDKN2A, and MTAP, could provide useful information for the validation of these preliminary results.
It also remains to identify the non-canonical pathways that support Hh/GLI1 activation in PC-negative MPM, and their weight in the ligand-dependent regulation of the Hh signaling.
In summary, the results of this study indicate that PC loss is a frequent event in MPM, likely associated with a more aggressive phenotype, and that the Hh pathway in MPM can be activated both canonically and non-canonically. Since there is a lack of correlation between the SMO-i response and the expression of Shh and SMO upstream of the pathway [32], we sustain the importance of analyzing the presence PC, together with nuclear GLI1, as a predictive marker of response to SMO-i.
This study can help to rationalize the repositioning of Hh pathway inhibitors for MPM therapy, taking into account the high heterogeneity of this tumor.

Conclusions
The present pilot study provides the first evidence that PC loss occurs frequently in MPM and does not preclude the activation of the Hh cascade. Our findings lay the foundation for future studies in the context of MPM, highlighting once again the need of patient-tailored therapies for this cancer.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/cancers14215216/s1, Table S1: Clinical information about mesothelial and mesothelioma patient-derived cell lines; Table S2: Clinical information about non-malignant and mesothelioma FFPE tissues; Table S3: Top five pathways deregulated in mesothelioma; Table S4: qRT-PCR analysis of GLI1, PTCH1 and c-MYC genes; Figure S1: TEM images of primary mesothelioma cells.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author upon reasonable request.