Introduction
Malignant mesothelioma (MM) is a rare and aggressive tumor which mainly originates from pleural and peritoneal mesothelial cells [
1,
2]. MM is histologically classified in three main subtypes: the epithelioid subtype, which has the best prognosis among the three, the sarcomatoid subtype, with the worst prognosis, and the biphasic or mixed subtype, with both epithelioid and sarcomatoid features [
3]. Untreated pleural MM patients have a median survival time of 6 months, and the majority of patients die within 24 months after diagnosis [
1,
2]. The median overall survival of pleural MM with a single chemotherapeutic agent is 7–8 months and only few drugs have a response rate of 15–20% [
2]. The current standard therapy for MM is the combination of pemetrexed and cisplatin chemotherapy with a response rate of ~ 40% [
1,
4]. Anti-angiogenic therapy was shown to enhance overall survival when added to the first line therapy [
1,
5]. It has been recently shown that the application of hyperthermic intraperitoneal chemotherapy (HIPEC) and cytoreductive surgery increased MM patients’ survival in particular for peritoneal MM [
6,
7]. However, the therapeutic strategies for the treatment of pleural MM are referred to as ‘life-extending treatments’ [
8]. Recent findings on the pathogenesis of MM have emphasized the importance of tumor suppressor gene alterations for sustaining aberrant signaling pathways which promote the uncontrolled growth of mesothelial cells [
9‐
12]. In agreement with new findings, with awareness of the resistance of MM to conventional therapies and of the poor patient survival following traditional chemotherapy, novel molecular targeted therapies for MM treatment have been identified [
2,
12‐
14]. Novel therapeutic approaches include inhibitors of mTOR, folate, receptor tyrosine kinases, ciclooxygenase and angiogenesis, synthetic lethal treatment, miRNA replacement, oncoviral therapies, and immunotherapy alone or in combination with chemotherapy [
2,
8,
14,
15]. Among the other aberrantly activated signaling pathways in MM, it has been reported that the EGF receptor (EGFR) is overexpressed in approximately 60% of human pleural MMs, but is not expressed in normal pleura [
16,
17]. In addition, at least one ErbB family member was found to be expressed in 88% of tumors. ErbB receptor expression was strongly dependent upon histologic subtype, with highest expression in epithelioid tumors [
17,
18]. It was found that MMs expressed EGFR (79.2%), ErbB4 (49.0%) and HER2 (6.3%), but lacked ErbB3 [
18]. Recently, a patient with pleural MM harboring both G719C and S768I EGFR mutations was successfully treated with Afatinib (AFA), a second-generation EGFR/HER tyrosine kinase inhibitor (TKI) [
19]. However, other clinical studies using EGFR TKIs in MM have not reported clinical efficacy [
17,
20,
21]. Mechanisms of resistance to EGFR inhibition by TKIs could be due to a simultaneous activation of alternative signaling pathways and rare mutations of EGFR in MM [
22,
23]. On the other hand, it has been suggested that when chemotherapy options have been exhausted, the use of EGFR TKIs is indicated in patients with wild-type EGFR tumors [
24]. Finally, it was reported that AFA induced apoptosis in non-small cell lung carcinoma (NSCLC) cells without EGFR mutation [
25].
The Hedgehog (Hh)/Glioma-associated oncogene (GLI) pathway is a complex signaling pathway which carries out critical functions in vertebrate embryogenesis and adult tissue homeostasis [
26,
27]. There are three Hh homologs in vertebrates: Sonic Hedgehog (Shh), Indian Hedgehog (Ihh) and Desert Hedgehog (Dhh) [
28]. GLI family of zinc-finger transcription factors and Smoothened (SMO) are signal transducers of the Hh pathway frequently aberrantly activated in tumors and MMs [
29‐
31]. After ligand binding, SMO enters the cilium and transduces the Hh signal, by activating the cytoplasmic GLI transcription factors. GLI1 and GLI2 stimulate the function of Shh-GLI1/2 while GLI3 antagonizes them [
28,
32]. High GLI1 and SMO expression levels and aberrant activation were associated with poor survival in patients with pleural MM [
30,
33]. SMO and GLI-related inhibitors have shown anti-cancer properties both in vitro and in vivo and when employed in clinical trials [
31]. However, the resistance to Hh signaling pathway inhibitors remains a drawback to overcome [
31].
Given these premises, the aim of this study was to evaluate the combined effect of inhibitors of the Hh- (GANT-61) and ErbB receptors (AFA)-mediated signaling pathways on the growth of MM both in vitro and in vivo. We demonstrate that combined treatment with both inhibitors is more effective than the single treatments in reducing the growth of MM. Therefore, the combined use of two drugs capable of counteracting the activation of two different aberrantly activated signaling pathways could be a useful tool to reduce the growth of MM cells.
Discussion
MM is a rare orphan aggressive neoplasia with low survival rates [
1,
2]. ErbB receptors and Hh signaling pathways are deregulated in MM [
16‐
18,
57]. Thus, molecules involved in these signaling pathways could be used to develop targeted therapy approaches.
AFA is a second-generation TKI, which irreversibly binds and inhibits the tyrosine kinase activity of members of the ErbB family including EGFR, ErbB2/HER2 and ErbB4. AFA targets both mutated and wild-type EGFR [
58]. AFA has been employed in several clinical trials in the last years for the treatment of NSCLC, head and neck squamous cell carcinoma, breast cancer, colorectal cancer, brain cancer, prostate cancer, gastric cancer. Upon TKIs treatment acquired resistance inevitably occurs in cancer patients and remains a biological challenge [
59‐
64].
Mechanisms of resistance to EGFR inhibition by TKIs could be due to a simultaneous activation of alternative signaling pathways [
22,
23].
GANT is a small molecule with inhibitory activity on the transcription factors GLI1 and GLI2, showing a high specificity for the Hh-mediated transduction pathway. Although several inhibitors of the Hh pathway targeting SMO are presently available (e.g. Vismodegib, Sonidegib), we chose to employ the GLI inhibitor GANT, since, as compared to SMO, GLI proteins have been indicated as more potent therapeutic targets in different tumors, including MM [
65,
66]. Indeed, it is known that GLI can be activated also via SMO-independent non-canonical mechanisms involving cross-talks between different oncogenic signaling pathways, among which the ErbB pathway, leading to resistance to SMO inhibitors [
27,
32,
66]. GANT has been shown to exert an anti-cancer activity in vitro and in vivo on different types of cancer, by inhibiting the expression of both GLI and Ptch [
27,
67‐
70].
The effect of AFA was previously analyzed in MM cells alone or in combination with other drugs. Specifically, the effect of AFA was evaluated in association with crizotinib, a drug used in NSCLC and capable of inhibiting MET, Alk and ROS-1 kinases. Another study on MM cells evaluated the synergistic use of AFA and trastuzumab (anti-HER2 monoclonal antibody). An increase in the antibody-dependent cellular cytotoxicity (ADCC) mechanism has been observed, with promising anticancer effects [
71]. In addition, GANT is reported to induce apoptosis following oxidative stress in MM cells in vitro [
72].
Accumulating evidence indicates that a complex interplay can occur between ErbB receptors and Hh signaling. Moreover, cooperation of EGFR signaling with Hh/GLI was demonstrated to promote cancer cells transformation and proliferation [
32]. To date, there are no studies reporting the combined use of AFA and GANT for cancer treatment.
Accordingly, the aim of this study was to evaluate the in vitro effects of AFA and GANT, used alone or in combination, on growth, cell cycle regulation, activation of cell death and autophagy, modulation of molecules involved in signaling pathways in three human MM cell lines, having different histotype (MM-B1, biphasic phenotype; MM-F1, fibromatous phenotype; H-Meso-1, epithelioid phenotype), and in a murine epithelioid MM cell line (#40a). To our knowledge, this is the first report analyzing the effect of the combination of AFA and GANT on MM cells.
Our results show that treatment with AFA and GANT, used alone and in combination, significantly inhibited the growth of MM cells in a dose- and time-dependent manner. The growth inhibition induced by the combined treatment AFA + GANT was superior to the effect obtained employing single treatments.
When analyzing the effect of single treatments on cells growth and death and on the induction of apoptosis, it was observed that the cell lines most sensitive to the drugs were H-Meso-1 and MM-B1, with epithelioid and biphasic histotypes, respectively. The fibromatous cell line (MM-F1) was less sensitive to single treatments when analyzing the same parameters. The behavior of the mouse epithelioid cell line #40a treated with single drugs was similar to that of the H-Meso-1 cell line. As regards the effect of the two single treatments on cell proliferation, it must be considered that AFA was certainly more powerful than GANT in its anticancer effects on MM. However, it must be highlighted that when the two drugs were used at the highest concentrations the combination with GANT further increased the AFA-induced inhibition on cell proliferation. Based on the analysis of drug interaction, the combination of AFA and GANT had synergistic effects in reducing MM cell survival depending on both the dose and cell line tested. The synergistic effect was more evident in MM-B1 and MM-F1 cell lines when the two drugs were used at high concentrations. When analyzing the effects of the combined treatment on the human H-Meso-1 and mouse #40a cell lines, the overall effect was instead less than additive, but still superior to that obtained with AFA alone. Our findings support the use of the Hh/GLI pathway inhibitor in combination with the ErbB receptors pathway inhibitor for the treatment of MM. Worthy of note, based on the results obtained on the MM-F1 cell line, which was the less sensitive to treatment with AFA alone, the combined treatment may be beneficial to those patients that are poorly responsive to the single treatment with an ErbB receptors pathway inhibitor.
To corroborate the in vitro findings, C57BL/6 mice were i.p. inoculated with syngeneic MM cells (#40a) and i.p. treated with AFA, GANT, or AFA + GANT. Our results show that AFA, GANT, and AFA + GANT treatments were able to significantly interfere with the in vivo tumor growth of mouse MM cells transplanted into the peritoneum compared to control mice. Moreover, AFA, GANT, and AFA + GANT were able to induce a significant increase in the mean survival and a reduction of the tumor volume compared to control mice. Remarkably AFA + GANT were more effective in increasing the mean survival and in reducing the abdominal circumference of mice compared to GANT and AFA alone.
In order to understand whether and how GANT treatment could interfere with the signaling pathways targeted by AFA, we analyzed their effects alone or in combination on the activation of autophagy and signaling pathways mediated by ErbB receptors and Hh. From these studies it emerges that some of the molecular events induced by the drug treatments were similar across the different cell lines, whereas other effects were more variable and cell-line dependent.
For instance, on the whole AFA and GANT appeared to inhibit autophagy on all the cell lines tested, when used alone or in combination. Indeed, the single and combined treatments induced a concurrent increase of the autophagosome marker LC3-II and the selective autophagy substrate p62 in H-Meso-1, MM-B1 and #40a cell lines, consistent with an inhibition of the autophagic flux. Similar effects were obtained with GANT and AFA + GANT also in MM-F1 cells, where however the AFA single treatment had no effect on LC3-II levels. On the other hand, AFA was still able to increase the levels of p62 in MM-F1 cells, indicating that an inhibition of autophagy may be induced by AFA also in this cell line. According to these findings, the increased antitumor effect exerted by the combined treatments might be due to an impaired ability of the cells to cope with cellular stress via autophagy. By the way, the functional significance of Beclin-1 fluctuations induced by the single drugs appears controversial, since AFA and GANT had variable, cell line-dependent effects on Beclin-1 levels when used alone. On the other hand, the combined treatment decreased Beclin-1 in every cell line except MM-F1.
Next, the effect of AFA on signaling mediated by ErbB receptors was analyzed. AFA downregulated ErbB2 levels in all cell lines and EGFR levels in H-Meso-1. Consistent with the effect exerted on ErbB receptors, AFA inhibited the activation of the downstream effectors ERK1 and ERK2, which primarily transduce proliferative signals [
32], and p54 JNK in all human and mouse cell lines. Of note, GANT potentiated the inhibitory effect of AFA on ERK1 and ERK2 phosphorylation in H-Meso-1 cells and on ERK1 in MM-B1 cells. On the other hand, as regards p38, the effect of AFA and GANT were opposite or similar, depending on the cell line. AFA decreased p38 phosphorylation in MM-B1 and #40a cells, whereas GANT had an opposite effect. However, the effect of AFA was prominent, and the addition of GANT was not able to modify the level of p38 phosphorylation. Conversely, both AFA and GANT promoted p38 phosphorylation in H-Meso-1 cells. The p38 MAPK plays a dual role as a promoter of cell death or survival depending on the type of stimulus and cell [
73‐
75]. Similarly, it has been reported that JNK has a dual role in regulating both apoptosis and survival of cancer cells [
74‐
76]. Several studies have demonstrated that p46 JNK and p54 JNK exhibit opposite functions in the regulation of cell survival and tumor development. In particular, some studies have reported that p46 JNK triggers a death signal, whereas p54 JNK induces cell survival. However, other studies have reported an opposite effect. Thus, the main regulator of cell survival between p46 and p54 JNK remains to be determined [
76]. In addition, a crosstalk between p38 and JNK in regulating autophagy and apoptosis induced by DNA damage has been reported [
77]. In our study, AFA alone potently inhibited the activity of p54 JNK in all cell lines. However, the addition of GANT to AFA appeared to affect p54 JNK phosphorylation in a variable, and once more cell line-dependent manner.
As far as the effect of the treatments on Akt are concerned, a more consistent trend was instead observed across the different cell lines. In fact, AFA inhibited Akt expression in all cell lines tested with the exception of MM-F1. However, the combined AFA + GANT treatment significantly reduced the expression of Akt in the MM-F1 cell line where AFA alone did not affect this kinase expression. In addition, Akt expression was significantly decreased with the combined treatment vs AFA alone in H-Meso-1 and in the mouse #40a cell line. Therefore, as compared with the decrease of Akt obtained with the single agents, the combination of AFA + GANT was able to further reduce the expression of this pro-survival kinase in three out of four MM cell lines.
The decrease of Akt expression by AFA plus GANT appeared to inhibit pro-survival signals and induce apoptosis in MM cells. Accordingly, we found that AFA increased the Bax/Bcl-2 ratio in H-Meso-1, MM-B1, #40a cell lines but not in MM-F1 cell line. It is worth noting that AFA was able to increase apoptosis in the MM-F1 cell line and this effect was potentiated by GANT. The activation of apoptosis in AFA and AFA + GANT-treated MM-F1 cells was corroborated by the increase in γH2AX.
It is well known that PI3K/Akt can be regulated by EGFR/ErbB2 receptors and by Hh/GLI signaling pathways [
32]. In turn, PI3K/Akt regulate the activation of GLI1 in melanoma and other cancer cells. Thus, Akt is essential for GLI-dependent activation of Hh signaling [
78‐
80]. In addition, EGFR/ErbB2 signaling synergizes with GLI1/2 to selectively induce the transcription of target genes through stimulation of RAS/RAF/MEK/ERK signaling [
81].
Concerning Hh signaling, we observed that GANT reduced GLI1, GLI2 and Ptch1 mRNA levels in H-Meso-1 and MM-F1 cells. However, we could not detect a similar inhibitory effect in MM-B1 cells, probably due to the low basal Hh pathway activity in this cell line. AFA decreased GLI2 mRNA levels, as well, but it increased GLI1 mRNA levels in all the cell lines analyzed, and Ptch1 mRNA levels in MM-F1 cells. This finding suggests that the inhibition of the EGFR pathway can induce the activation of alternative pathways in MM cells such as the Hh/GLI1 pathway, that could be activated most likely through a non-canonical pathway.
In turn, it has been demonstrated that activation of the Hh pathway can promote resistance to EGFR inhibitors, possibly via the induction of tumor cell epithelial-to-mesenchymal transition [
82]. Accordingly, in the Hh-dependent cell lines H-Meso-1 and MM-F1, AFA treatment appears to trigger an “escape” mechanism through the overactivation of Hh/GLI1. In this respect, the combination with a GLI inhibitor may be beneficial, since it may allow to compensate for the stimulatory effect of AFA on Hh/GLI1 signaling. Indeed, the addition of GANT to AFA was able to counteract the AFA-mediated induction of GLI1 and Ptch1 expression. Of note, the reduction of GLI1 mRNA levels can explain the additive inhibitory effect exerted by the combination of GANT and AFA on Akt expression, since both EGFR/ErbB2 receptors and Hh/GLI signaling pathways can mediate the inhibition of this kinase.
Current multimodal therapies including surgery, chemotherapy and radiotherapy, have improved the quality of life of MM patients, but the prognosis for this tumor still remains poor.
Preclinical studies and clinical trials have addressed the therapeutic efficacy of many new antitumor agents on MM. Intracavitary drug administration has also been investigated as a mean of improving the effectiveness of different agents via local delivery at the tumor site [
15].
Among the angiogenesis-targeting agents, the anti-VEGF monoclonal antibody Bevacizumab has been shown to improve overall survival in MM patients when combined with standard chemotherapy [
15]. Conversely, other anti-angiogenic agents, including Axitinib (an anti VEGFR TKI), Sorafenib (a TKI targeting VEGFR2/3, platelet-derived growth factor receptor (PDGFR) and rapidly accelerated fibrosarcoma (RAF)/c-KIT), and Imatinib mesylate (a TKI targeting BCR-ABL, c-KIT, and PDGFR) failed to show clinical efficacy on MM [
83‐
87].
The MEK and p110β/PI3K inhibitors Seleumetinib and AZD8186, currently being tested in clinical trials on different cancer types, have been reported to increase survival and display a low toxicity profile in a murine model of sarcomatoid MM [
88]. A strong EGFR activation associated with HER2, HER3, Axl, and MET co-activation, mediated mainly by receptor heterodimerization and autocrine-paracrine loops induced by the expression of their cognate ligands, has been demonstrated in patients with peritoneal MM. These results support the possible use of different targeted therapy approaches to inhibit the pathways activated by these receptors [
89].
The combination of different targeted therapy and immunotherapy approach, using immune checkpoint blockade regimens can potentially deliver new opportunities to improve anti-cancer treatments for MM patients [
90].
In the present study we provide evidence that the combined use of two drugs capable of counteracting the activation of two different, aberrantly activated signaling pathways (ErbB receptors and Hh) could be a useful tool to reduce the growth of MM cells.