The HH pathway remarkably induces tumor proliferation (41, 42). HH signaling regulates INSM1 interactions with N-myc to promote SCLC cell growth in a Gli-dependent manner (43). The E3 ligase (HERC4) and liver kinase B1 (LKB1) inhibit cell proliferation in lung cancer by inactivating the HH pathway. Mechanistically, HERC4 has been shown to negatively regulate the HH pathway by destabilizing the Smo protein (44), while the mechanisms through which the HH pathway is inhibited by LKB1 have not been elucidated (45).

The HH pathway is involved in cell invasion and migration of multiple tumors (46). SAM- and SH3-domain containing 1 (SASH1) inhibit hepatocellular carcinoma invasion and migration by inactivating the HH and PI3K/AKT pathways in vitro and in vivo (47). In vitro, transmembrane 107 protein (TMEM107) inhibited epithelial-mesenchymal transition (EMT) and invasion by negatively regulating HH signaling in NSCLC (48). Besides, Gli1, the downstream gene of the HH pathway, promotes NSCLC cell invasion and metastasis in vitro and in vivo by inducing the EMT process. The downregulation of Gli1 significantly inhibited tumor growth and enhanced E-cadherin expressions (49). Interestingly, the KRAS mutation activated the Kras/YY1/ZNF322A/SHH axis by triggering the expression of downstream genes, including YY1, ZNF322A, and SHH, which promoted angiogenesis in NSCLC in vitro and in vivo (16). The overexpression of phosphatidylethanolamine‐binding protein 4 (PEBP4) promotes NSCLC cell proliferation and EMT by regulating the HH pathway (50) while fibroblast activation protein α (FAPα) promoted LSCC cell growth, adhesion, and migration in vitro via the PI3K and HH pathways (51).

Hedgehog-interacting proteins (HHIP) negatively regulate the HH pathway by binding SHH proteins (52–54). HHIP serve as tumor suppressors by significantly inhibiting lung cancer cell proliferation, migration, and invasion (53). Similarly, overexpression of signal peptide-CUB-EGF domain-containing protein 2 (SCUBE2) inhibits NSCLC cell proliferation and invasion by modulating the HH pathway. The SCUBE2-mediated inhibition of the HH pathway can be reversed by recombinant SHH proteins (55). GATA-6 impairs the metastatic abilities of LAC cells (56). Moreover, GATA-6 suppresses cell proliferation and migration in LSCC by inhibiting the expressions of SHH at transcriptional levels (57). Additionally, miR‐520b enhances NSCLC cell proliferation and metastasis by activating the SPOP‐Gli2/3 axis (58). Hematopoietic pre-B-cell leukemia transcription factor (PBX)-interacting protein (HPIP) is a nucleo-cytoplasmic shuttling protein (59) that promotes NSCLC cell proliferation, invasion, and migration in vitro and tumor growth in vivo by interfering with the HH pathway (54).

In an A549 cell line, tumor suppressor candidate 3 silencing (TUSC3) inhibited growth, proliferation and induced apoptosis as well as radiation sensitivity (60). In contrast, TUSC3 overexpression in NSCLC promoted cell proliferation, migration, and invasion in vitro and in vivo, possibly by involving HH signaling (61). Leucine zipper transcription factor-like 1 (LZTFL1) inhibits lung cancer tumorigenesis, at least partly by inhibiting HH signaling pathway to maintain epithelial cell differentiation (62). Figure 4 shows HH signaling interaction networks in lung cancer patients.

Drug resistance is one of the leading causes of tumor treatment failure (63), while hyperactivation of the HH pathway is frequently described in many drugs resistant malignant tumors. The HH pathway promotes cell proliferation as well as invasion in LAC with acquired drug resistance to EGFR-TKIs, partially by acting on HGF and MET signaling (64). Della Corte et al. (65) investigated the mechanisms in acquired resistance to EFGR-TKI during NSCLC treatment. The HH pathway was found to induce drug resistance by regulating the EMT process. Additionally, the active HH pathway was shown to induce EMT and upregulated ABCG2 in EGFR-TKI-resistant NSCLC patients, while the opposite result was observed when the HH pathway was inhibited (66).

MiRNA disorders are prevalent in many malignant tumors, including lung cancer (67, 68). miR-182-5p is the direct target of Gli2, the negative feedback axis of miR-182-5p/Gli2 regulates cell cisplatin resistance in LAC (69).

A small proportion of tumor cells are regarded as CSCs that play a vital role in tumor occurrence, development and recurrence (70, 71). Aberrantly HH signaling is common in CSCs (72, 73). SHH-positive (SHH⁺) cells in NSCLC, expressing an uncleaved full-length SHH protein, exhibited drug resistance and CSC characteristics. A high abundance of SHH⁺ cells is a biomarker for worse prognostic outcomes for NSCLC patients (74). Low-folate (LF) is associated with malignant progression of lung cancer. Functionally, the LF microenvironment increases CSCs-like potential, partially through the HH pathway (75). Fibroblast growth factor receptor 1 (FGFR1) amplification in NSCLC is particularly prevalent, especially in LSCC (76, 77), where it stimulates the expressions of Gli2 through ERK pathway activation. The FGFR1/Gli2 axis upregulates the expression levels of CSCs markers (78). Aberrantly activated Notch and HH pathways induce CSCs phenotypes. Notch-Hedgehog positive tumor cells mediate immune evasion by enhancing the functions of regulatory T cells (79).

The tumor microenvironment, a local homeostatic environment composed of tumor cells, stromal cells (including fibroblasts, immune and inflammatory cells) as well as the extracellular matrix, provides the necessary materials for tumor initiation and progression (80). Forkhead box F1(FoxF1) is associated with poor prognostic outcomes in some lung cancer subtypes. FoxF1, a downstream effector of HH signaling, stimulates the secretions of HGF and FGF2 by lung cancer fibroblasts, promoting lung cancer cell growth and migration in vivo (81).

NSCLC cells activate HH signaling in lung fibroblasts in a paracrine manner to stimulate the production of pro-angiogenic and metastatic factors, which induce fibroblast proliferation, invasion and collagen deposition, leading to lung cancer progression (82). Similarly, tumor cell-derived HH ligands stimulate perivascular stromal cells to secrete VEGF-A, promoting tumor angiogenesis in vitro and in vivo (83). CAFs-derived SHH ligands activate HH signaling in NSCLC cells in a paracrine manner, upregulating the expressions of EMT-related genes such as α-SMA and FAP, and enhancing the migration abilities of lung cancer cells (84). CAFs remodel actin cytoskeleton-induced EMT process in an Smo-dependent manner, enhancing the insensitivity of lung cancer cells to EGFR-TKIs (85).

TAMs receiving HH ligands secreted by tumor cells activate KLF4 and STAT3, downstream transcription factors of HH signaling, which drive TAMs M2 polarization, reduce CD8+ T cells, and inhibit their functions, thereby promoting malignant progression in various tumors (86, 87). By co-culturing A549 cells with THP-1-derived macrophages, A549 cells stimulated THP-1-derived macrophages toward TAMs M2 polarization. In contrast, THP-1-derived macrophages upregulate the expressions of stemness-related genes, including Sox2 and NANOG through HH, STAT3 and Notch signaling, increasing the stemness characteristics of lung cancer CSCs (88).

In conclusion, CAFs and TAMs interact with tumor cells to induce a pro-tumorigenic microenvironment, and HH signaling in the tumor microenvironment regulates CAFs, TAMs, and tumor cells in an autocrine and paracrine manner to increase stemness characteristics and promote the EMT process. Targeting HH signaling is a promising strategy to inhibit the formation of the tumor microenvironment.

Through autoproteolytic cleavage, full-length SHH proteins generate 25 kDa SHH-C and 19 kDa SHH-N. Then, SHH-N undergoes C-terminal cholesterylation and N-terminal palmitoylation to mature into the SHH-N ligand (22). HHat mediates N-terminal palmitoylation of SHH-N and is essential for SHH-N ligand maturation. HHat inhibitors inhibit SHH-mediated activation of canonical and noncanonical HH pathways by blocking SHH-N ligand palmitoylation. RU-SKI43 (IC₅₀: 0.85 μM), a small molecule inhibitor, exhibited HHat inhibitory effects in vitro and in vivo (115). RU-SKI-201 (IC₅₀: 0.73 μM), optimized based on RU-SKI43, overcame the off-target toxicity of RU-SKI 43 and specifically inhibited HH signaling in multiple tumor cell lines (116). Although both small molecule inhibitors exhibit potent anti-HH signaling activities, their oral availability and pharmacological toxicities should be evaluated at the animal level.

The inhibition of HH proteins induces all Ptch-induced pro-apoptotic pathways (117). 5E1, a monoclonal antibody to SHH-N, has been widely used in biological experiments. Interestingly, most of the binding sites of 5E1 on SHH-N ligands overlap with HHIP (118). Similar to 5E1, Tolani et al. developed a new therapeutic antibody that targets the full-length SHH protein and SHH-N ligand to inhibit tumor growth (119). Owens et al. identified HL2-m5 (IC₅₀: 230 nM), a macrocyclic peptide inhibitor with a high affinity for SHH ligands, which inhibited HH signaling in vitro without significant pharmacological toxicities (120). This compound should be further validated via animal experiments to assess its oral availability and safety.

Smo inhibitor resistance is the leading cause of progression in most HH-dependent tumor patients and is mainly attributed to mutations in the drug binding pocket of the Smo protein (121). In a phase I trial, patients with locally metastatic and progressing BCC previously treated with, or not with HH inhibitors exhibited good clinical responses to LY2940680 (Taladegib) and were well tolerated (122). Cyclopamine and its derivatives from natural compounds reverse HH signaling activation by targeting Smo, thereby limiting HH-dependent tumor cell growth (123). As a semisynthetic cyclopamine analogue, IPI-926 (Saridegib) addresses the issues of low potency and poor aqueous solubility of cyclopamine. Significantly, it inhibited tumor progression in HH-dependent medulloblastoma (MB) mice models with no obvious toxicity (124). IPI-926 has certain inhibitory activities against Smo^D473H (Smo mutant), which is resistant to GDC-0449 (Vismodegib). When combined with gemcitabine, IPI-926 was shown to enhance anti-tumor activities of gemcitabine by increasing tumor vascular densities, which prolonged the survival outcomes of mice with pancreatic ductal adenocarcinoma (125). In a clinical phase I study involving basal cell carcinoma patients, despite differences in chemical structures between IPI-926 and Vismodegib, both exhibited similar sites of action on the Smo protein and shared same resistance mechanisms (126).

The FDA approved itraconazole for antifungal treatment, and its safety as well as side effects are well understood (111, 127). Compared to Vismodegib, Itraconazole inhibits Smo through different mechanisms. Itraconazole exerts significant anticancer activities in vivo, even targeting Vismodegib-resistant Smo^D477G (110). Similar to itraconazole, posaconazole exerts anti-tumor activities by targeting Smo. Interestingly, posaconazole without triazole retains its antagonistic effects on HH signaling and disrupts the inhibitory effect on Cyp3A4 (128). The des-triazole derivatives, based on the posaconazole scaffold, were effective at inhibiting HH signaling and were well tolerated (128).

The presence of an oxysterol binding site in each of the seven transmembrane domains and extracellular cysteine-rich domain (CRD) of the Smo protein is necessary for HH signaling activation (129, 130). Cholesterol molecules can be transferred from the cytoplasm to oxysterol sites on the seven transmembrane regions or extracellular CRD via hydrophobic channels on Smo proteins, mediating Smo activation (129, 130). Therefore, the oxysterol binding site on the Smo protein is a potential HH signaling target. There is a need to determine the mechanisms through which the Smo protein integrates the information conveyed by the two oxysterol binding sites to determine the activities of HH signaling in vivo. As a Bcl-2 inhibitor, ABT-199 is approved by the FDA for treatment of chronic lymphocytic leukemia (ALL). ABT-199 (IC₅₀: 215.9 nM) targets the extracellular CRD of Smo while acting as a potential competitive inhibitor of oxysterol, which inactivates HH signaling (131). In addition, ABT-199 was shown to target the drug binding pocket mutant and/or constitutively activated Smo proteins with excellent pharmacological properties and safety, therefore, its effects in patients with HH inhibitor resistance should be evaluated (131).

Multi-targeted HH inhibitors act on multiple components of HH signaling and/or inhibit multiple noncanonical HH pathways, exhibiting potent inhibitory anti-HH signaling effects. Lospinoso Severini et al. identified compound 22 (IC₅₀: 7.1 μM), which contains an isoflavone scaffold, to be a multi-target HH inhibitor that inhibits MB cell growth by targeting Smo and Gli proteins (132). As a semisynthetic oxysterol analogue, Oxy210 (IC₅₀: 1.83 μM) is a dual inhibitor of TGFβ and HH signaling, with excellent pharmacological properties, but should be further evaluated at the animal level (133). Li et al. and Zhu et al. determined that HH-13, HH-20 (IC₅₀: <0.2 μM), and L-4 (IC₅₀: 2.33 nM) act as potential Smo inhibitors by targeting Smo^D473H, respectively (134–136). Moreover, L-4 showed high efficiency, good tolerability, and high oral bioavailability in ICR mice (134).

In conclusion, reliable HH inhibitors from FDA-approved drugs, such as itraconazole and ABT-199 have been screened and their toxic effects as well as side effects evaluated. The Oxysterol site of the Smo protein is essential for activation of HH signaling, and the development of Smo inhibitors, based on the Oxysterol site, can overcome Smo acquired resistance, with great potential for clinical applications.

Gli inhibitors target Smo downstream genes to overcome Smo acquired resistance and inhibit Gli-dependent noncanonical pathways, thereby achieving better clinical efficacies and improving patient clinical outcomes. GANT58 and GANT61 (IC₅₀: 5 μM) are widely used in biological experiments by targeting the Gli1 transcription factor (137). Notably, GANT61 is more active against HH signaling than GANT58, which is mainly attributed to its ability to interfere with the binding of Gli1 to DNA (137). The Gli-activated but Smo-independent pathway is activated in some types of LSCC, including PI3K/AKT and RAS-MEK signaling. Therefore, GANT61 is more potent than Vismodegib in Gli-dependent LSCC patients (138).

Arsenic trioxide (ATO) (IC₅₀: 2.7 μM), approved by the FDA for acute promyelocytic leukemia (APL), targeted the Gli1 transcription factor and exhibited significant cytotoxicity in mice models of MB and Ewing sarcoma, supporting further pharmacological evaluation in Gli1-dependent malignancies (139). ATO kills SCLC CSCs by downregulating CSCs-related genes, such as Sox2 and c-Myc at least in part by targeting HH signaling (140). The lack of efficacy of ATO in relapsed SCLC patients and SCLC patient-derived xenografts (PDX) mice in phase II clinical trials may be attributed to low proportions of SCLC CSCs and intermittent dosing regimens (141). The therapeutic benefits of ATO in combination with chemotherapeutic agents and daily continuous dosing regimens in SCLC patients should be investigated (141). Due to differences in study methods and different cell lines, it should be determined if the primary target of ATO is Gli1 or Gli2 (142). Pyrvinium (IC₅₀: 10nM) is an FDA-approved casein kinase-1a (CK1a) agonist with high efficacies. Moreover, CK1a can also act as a negative regulator of HH signaling (143). Mechanistically, CK1a downregulates HH target genes by modulating their stability via phosphorylation of Gli transcriptional factors (143).

Studies have found that VDR signaling negatively regulates Gli1 expression through crosstalks with HH signaling (144–148). Vitamin D3-based derivative compounds 16, 21, and 22 modulated Gli1 proteins in vitro by targeting VDR signaling and HH signaling (149). However, the mechanisms through which VDR regulates HH signaling have not been elucidated, which requires further investigations.

Bromodomain (BRD), a class of conserved protein domains that specifically recognize acetylated lysine in histones, promote the enrichment of chromatin remodeling factors and transcription factors to specific gene transcription sites by binding acetylated lysine, altering the activities of RNA polymerase II, and regulating gene expressions (150, 151). Liu et al. developed non-selective BET inhibitors, compounds 25 and 35 (IC₅₀: < 1nM) by further optimization of the BRD4 inhibitor, ABBV-075, which inhibited Gli protein expressions and overcame resistant Smo mutations in MB mice without apparent toxicity (152). Compared to compound 35, compound 25 exhibits stronger efficacies and better safety, however, it should be optimized to improve its pharmacokinetic parameters (152).

In patients with recurrent metastatic BCC, Itraconazole in combination with ATO (ATO-ITRA) reduced Gli1 mRNA levels by 75% from baseline levels (153). Three of the five patients treated with an ATO-ITRA regimen were stable for three months, but did not achieve tumor shrinkage, which was attributed to sequential dosing as well as lower doses. Higher doses and/or daily continuous dosing regimens should be used to evaluate the clinical efficacy of the ATO-ITRA regimen (153).

In conclusion, Gli inhibitors overcome Smo acquired resistance and inhibit Gli activatation via the Smo-independent noncanonical pathway, with robust anti-tumor activities in Gli-dependent tumors. Gli inhibitors (including ATO and GANT61) combined with Smo inhibitors (including Itraconazole and Taladegib) or chemotherapeutic agents represent a promising strategy for patients with HH-dependent tumors ( Figure 5 ).

Autophagy, a process of cellular self-degradation, removes damaged or redundant proteins as well as organelles. It plays an important role in maintaining intracellular homeostasis as well as tumorigenesis (158–161). Autophagy is a double-edged sword that promotes and inhibits tumor development (162–164). Lei et al. reported that Jervine triggers autophagy and increases the expressions of cleaved caspase 3, leading to NSCLC cell apoptosis by acting on AKT/mTOR and HH pathways (165).

The additional natural products that block cell proliferation include Evodiamine, Sulforaphane, and Cordyceps militaris. Evodiamine inhibits cell proliferation by modulation of the AKT/NF-κB and HH pathways in NSCLC (166). Cordyceps militaris induces NSCLC cell apoptosis by blocking Gli1 nuclear translocation via inhibiting the expressions of tectonic protein 3 (TCTN3) (167). In the PC9 cell line, Sulforaphane reversed lung cancer cell resistance to gefitinib by regulating HH signaling (168).

As a traditional Chinese medicine, Scutellariabarbata D. Don extraction (SBE)has been used in clinical management for many years because of its reliable and effective anti-cancer activities (169, 170). SBE downregulates the downstream components of HH signaling, leading to tumor cell cycle arrest and increasing lung cancer cell sensitivity to cisplatin (171).

Tumor angiogenesis is closely associated with multiple malignant behaviors of lung cancer, such as proliferation, invasion, and metastasis. Inhibition of tumor angiogenesis is of great clinical significance. However, angiogenesis inhibitors that target VEGF/VEGFR2 did not significantly improve the five-year survival rate for cancer patients (172, 173). Therefore, there is a need to develop novel angiogenesis inhibitors that lack toxic effects. Pristimerin, a compound isolated from Celastrus aculeatus Merr, exhibits multiple pharmacological activities, including anti-inflammatory, anti-microbial, anti-tumor, and anti-peroxidative effects (174, 175). In NSCLC, Pristimerin affects the maturation of tumor angiogenesis by regulating the translocation of its downstream gene (Gli1) to the nucleus through HH signaling (176).

Triptonide, a natural compound separated from Tripterygium wilfordii Hook, exhibits potent anti-inflammatory (177) and anti-tumor efficacies (178, 179). It has been shown to inhibit the proliferation, invasion, as well as migration of lung cancer cells and eliminates stem-like signatures by reducing the expression levels of Gli1 at both transcriptional and translational levels via inhibiting Gli1 promoter activities (180).

Curcumin is a potent compound isolated from the rootstalk of Curcuma longa (181, 182). It was shown to effectively eliminate various cancer stem cells and reduce their CSCs-like characteristics (183–185). Curcumin has powerful inhibitory effects on lung cancer CSCs in vitro. Mechanistically, Curcumin downregulates the expression levels of CSCs markers, including CD133, CD44, ALDHA1, Nanog and Oct4, by inhibiting the HH and Wnt/β-catenin pathways (186). In addition, SBE suppresses stemness-like phenotypes in NSCLC by targeting the Sox2/Smo/Gli1 positive feedback loop (187).