Medulloblastoma Drugs in Development: Current Leads, Trials and Drawbacks

Abstract

Medulloblastoma (MB) is the most common malignant brain tumor in children. Current treatment for MB includes surgical resection, radiotherapy and chemotherapy. Despite significant progress in its management, a portion of children relapse and tumor recurrence carries a poor prognosis. Based on their molecular and clinical characteristics, MB patients are clinically classified into four groups: Wnt, Hh, Group 3, and Group 4. With our increased understanding of relevant molecular pathways disrupted in MB, the development of targeted therapies for MB has also increased. Targeted drugs have shown unique privileges over traditional cytotoxic therapies in balancing efficacy and toxicity, with many of them approved and widely used clinically. The aim of this review is to present the recent progress on targeted chemotherapies for the treatment of all classes of MB.

Graphical Abstract

1. Introduction

Medulloblastoma (MB) is the most common malignant brain tumor in children.¹ Patients with MB suffer from increased intracranial pressure and cerebellar dysfunction. Common symptoms of MB may inlude headaches, nausea, vomiting, dizziness, and difficulty walking. Current treatment for MB includes surgical resection, radiotherapy and chemotherapy.¹ Despite significant progress in MB management, more than 25% of patients experience cerebella mutism, dysarthria and neurocognitive disorders after surgical resection. Adjuvant radiotherapy and non-specific chemotherapy like cisplatin (CIS), vincristine (VCR), etoposide (ETP), and temozolomide (TMZ) can cause severe side effects and secondary tumors.² Based on molecular and clinical characteristics, MB patients are classified into four groups: Wnt, Hedgehog (Hh), Group 3 (Grp3), and Group 4 (Grp4).³ Selectively targeting each subgroup’s abberantly regulated genes/proteins is a primary strategy for anti-MB drugs currently in development.⁴ Among all four subgroups, the Hh-subgroup is the most studied and well-defined MB subgroup. Clinically, Hh-MB represents the most common molecular subgroup in both infants (<3 years of age) and adults (>17 years of age).¹ Adult Hh-MBs are characterized by a higher prevalence of Hh pathway-associated mutations to infants. Many small molecule inhibitors targeting the somatic mutations or gene amplifications of the Hh-signaling pathway have been evaluated extensively in preclinical and clinical trials and several compounds have been approved for clinical use.⁵ Wnt-MB patients are primarily composed of older children with a median age of diagnosis around 11 years. Wnt-MB tumors are usually of classic histology and are infrequently metastatic at diagnosis.¹ Patients with Wnt-MB are largely homogeneous between individuals, with regard to genome-wide expression signatures and methylation profiles.¹ This group has the best outcomes and prognosis among all four groups with 5-year survival rates above 95%.⁶ Grp3- and Grp4-MBs share some molecular and biological similarities, and 30–40% patients of these groups are metastatic at diagnosis. Grp3-MB occurs during infancy and childhood and is rarely seen in adults (>18 years of age), whereas Grp4-MB occurs across all age groups.¹ Unlike Wnt- and Hh-MBs, which have been identified by mutations within these specific pathways, the molecular causes of Grp3- and Grp4-MBs have not been fully characterized.

The signaling pathways involved in these subtypes differ greatly from each other. Developing subtype-based molecularly targeted therapies with reduced toxicity has been a primary approach for anti-cancer drug development during the past several decades.^7–9 Improved knowledge of aberrant signaling pathways in MB has provided novel emerging therapeutic targets for drug development, including epigenetic modulators and protein kinases. Inhibitors of these pathways have shown therapeutic potential in preclinical and/or clinical models. The aim of this paper is to present the advances of molecularly targeted modulators currently undergoing preclinical or clinical evaluation as potential anti-MB agents. Key pathways and molecular nodes involved in this review are summarized in Figure 1; compounds referred to this review are described and summarized in Tables 1 and 2.

Figure 1. Illustration of proposed MB related proteins/pathways described in this review. (i) Conventional MB-related signaling pathways.

Components of canonical Wnt signaling pathway: FZD (frizzled), transmembrane receptor for Wnt ligand; DSH (dishevelled), transduces the Wnt signal to the cytoplasm; APC (adenomatosis polyposis coli), promotes the degradation of β-catenin; GSK3β (glycogen synthase kinase 3β), constitutively active protein kinase targeting β-catenin for degradation; Axin, component of β-catenin destruction complex; β-catenin, key mediator of canonical Wnt-signaling pathway. Components of canonical Hh-signaling pathway: Ptch (patched), transmembrane receptor for Hh ligands; Smo (smoothened), transmembrane receptor associated with Ptch and transducing Hh signal; Sufu (suppressor of fused homologue), negative regulator of Hh signaling; Gli (glioma-associated oncogene homologue), key transcription factor downstream of Hh signaling; Gli^A, active form of Gli that binds to DNA and serves as a transcriptional activator. (ii) Emerging proteins/pathways. Components of Sema3-Nrp signaling: Sema3 (Semaphorin 3), a member of a secreted ligand family; Nrps (neuropilins), transmembrane receptors for Sema3; PDE4D (phosphodiesterase 4D), translocating to cell membrane when recruited by Nrps; PKA (protein kinase A), cross-talk with Hh signaling through Gli. LSD1 (lysine demethylase 1), physically interacting with methylated lysines and Gfi1. HDAC (histone deacetylase) and BET (bromodomain and extraterminal domain), physically interacting with acetylated lysines. FACT (facilitates chromatin transcription), a histone chaperone to promote H2A/H2B dimer dissociation from the nucleosome and allow RNA polymerase II transcription on chromatin. PI3K (phosphoinositide 3-kinase), controlling Akt phosphorylation and activation status; Akt (protein kinase B), controlling mTOR phosphorylation and activation status; mTOR (mammalian target of rapamycin), key regulator downstream of PI3K-signaling. CDK (cyclin-dependent kinase), controlling the phosphorylation status of retinoblastoma. CK2 (casein kinase 2), involving p53 and MAPK phosphorylation status. Chk1 (checkpoint kinase 1), controlling Cdc25A phosphorylation status. ROCK (Rho-associated protein kinase), key regulator of Rho/ROCK-signaling, controlling MYPT1, MLC and LIMK1/2 phosphorylation status. Wee1, controlling Chk1/2 phosphorylation status. Src, controlling a wide range of protein phosphorylation, including PI3K. DDX3, a member of RNA helicase. PARP (poly(ADP-ribose) polymerase), interacting with single and double-strand DNA breaks. Sp1 (specific protein 1 transcription factor), binding to GC-boxes, CACCC-boxes and basic transcription elements. HMGCR (β-hydroxy β-methylglutaryl-CoA reductase), a key component of the mevalonate pathway that envolving a number of essential cellular functions. ABC (ATP-binding cassette) transporter, responsible for the translocation of multiple substrates across membranes. COX-2 (cyclooxygenases 2), envolves in tumor cells apoptosis and cell growth. DNA polymerases, catalyzes the synthesis of DNA molecules from nucleoside triphosphates. PP2A (protein phosphatase 2A), a serine/threonine phosphatase enhances anticancer immunity. mGPD (mitochondrial glycerophosphate dehydrogenase), a component of the glycerophosphate shuttle.

Table 1.

Anti-MB drugs and candidates functioning through conventional MB-related targets (2010 to present).^a

Name	Structure	Mode of action	BBB permeable	In vitro model used	In vivo model used	Development Statusd	Reference
Pyrvinium		CK1α agonism	Yes	Hh-Light2 cellsa NIH3T3, SmoD473H cells and Sufu−/− MEFsa	Ptch+/− MB allografts	Preclinical	17
Fenretinide		Wnt3a/β-cateni n antagonism	Yes	Daoy and ONS-76 cellse	Daoy or ONS-76 xenografts	Preclinical	19
Vismodegib (GDC-0449)		Smo antagonism	Yes	C3H10T1/2 MEFsa	Ptch+/− MB allografts	Phase II	27,28
Sonidegib (NVP-LDE225)		Smo antagonism	Yes	mSmo and hSmob	Ptch+/−; p53+/− MB subcutaneous/orthotopic allografts	Phase II	29
Glasdegib (PF-04449913)		Smo antagonism	Yes	C3H10T1/2 MEFsa	Ptch+/−; p53+/− or Ptch+/−; p53−/− MB allografts	Preclinical	30,31
Taladegib (LY2940680)		Smo antagonism	Yes	C3H10T1/2 MEFsa SmoD473H transfected NIH3T3 cellsa Daoy cellse	Ptch+/−; p53−/− transgenic mice	Phase I (Withdrawn)	32,33
NVP-LEQ506		Smo antagonism	Yes	mSmo and hSmob SmoD473H transfected NIH3T3 cellsa	Ptch+/−; Hic−/− MB allografts	Phase I	34
MK-4101		Smo antagonism	Yes	Hh-Light2 cellsa Ptch+/− mice-derived cellse SmoD477G cellsa	Ptch+/− MB allografts Ptch+/− mice bearing primary MB	Preclinical	35
12b		Smo antagonism	NA	Hh-Light2 cellsa	Ptch+/−; p53−/− MB allografts	Preclinical	36
TAK-441		Smo antagonism	NA	Hh-Light2 cellsa SmoD473H transfected cellsa	Ptch+/−; p53−/− MB allografts	Phase I (Discontinued)	37–39
Itraconazole		Smo antagonism	Limited	Hh-Light2 cellsa ASZ001 cellsc	Ptch+/−; p53−/− MB allografts SmoD477G MB allografts	Preclinical	40
Arsenic trioxide	As2O3	Gli antagonism	Yes	Daoy and D556 cellse	ND2:SmoA1 transgenic mice Ptch+/−; p53−/− MB allografts SmoD477G MB allografts	Phase II	47,48
MK-5710		Smo antagonism	NA	Hh-Light2 cellsa	Ptch+/− MB allografts	Preclinical	49,50
PF-5274857		Smo antagonism	Yes	hSmo and Hh-dependent MEFsa	Ptch+/−; p53+/− MB allografts	Preclinical	51
9d		Smo antagonism	NA	hSmo and mSmoa	Ptch+/− MB allografts	Preclinical	52
14f		Smo antagonism	NA	hSmo and mSmoa	Ptch+/− MB allografts	Preclinical	52
5		Smo antagonism	NA	SmoWT and SmoD473H transfected NIH3T3 cellsa	Ptch++/−; p53−/−; SmoD477G MB allografts	Preclinical	53
24		Smo antagonism	Yes	Hh-Light2 cellsa	Ptch+/− MB subcutaneous xenografts	Preclinical	54
L-4		Smo antagonism	NA	Hh-Light2 cellsa SmoD473H transfected cellsa	Ptch++/−; p53−/− MB allografts	Preclinical	56
ANTA XV		Smo antagonism	Yes	TM3-Gli-Luc shift assay hSmo bindinga	Ptch++/−; p53−/− MB allografts	Preclinical	57
43		Smo antagonism	NA	hSmo and mSmo bindinga	Ptch+/−; p53−/− MB allografts	Preclinical	58
27		Smo antagonism	NA	hSmo and mSmo bindinga	NOD-Scid mice	Preclinical	59
10e		Smo antagonism	NA	Hh-Light2 cellsa Ptch+/−; p53−/−mice-derived MB cellse	Ptch+/−; p53−/− MB allografts	Preclinical	60
23b		Smo antagonism	NA	Hh-Light2 cellsa Ptch+/−; p53−/− mice-derived MB cellse	Ptch+/−; p53−/− MB allografts	Preclinical	61
65		Smo antagonism	NA	Hh-Light2 cellsa	Ptch+/−; p53−/− MB allografts	Preclinical	62
CAT3		Smo antagonism	Yes	Daoy cellse	Daoy orthotopic xenografts	Preclinical	63
E-1		Gli antagonism	NA	Sufu−/− MEFsa	Ptch+/−; p53−/− MB allografts	Preclinical	64
BRD9526		Gli downregulating	NA	Hh-Light2 cellsa	NA	Preliminary	65
29a		Gli downregulating	NA	Hh-Light2 cellsa	Ptch+/−; p53−/− MB allografts	Preclinical	66

^(a)Gli-Luc reporter assay

^(b)BODIPY-Cyc competitive binding assay

^(c)RT-qPCR assay monitoring Gli1 mRNA expression

^(d)Development status against MB unless otherwise denoted

^(e)Cell viability/proliferation assay; NA = no available data.

Table 2.

Anti-MB drugs and candidates functioning through epigenetic, phosphorylation, and other emerging targets (2010 to present). ^a

Name	Structure	Mode of action	BBB permeable	In vitro model used	In vivo model used	Development Statusd	Reference
Panobinostat (LBH589)		HDAC antagonism	Yes	UW228, UW426 and Med8A cellse	UW426-effLuc MB leptomeningeal seeding model	Phase I	72
Mocetinostat		HDAC antagonism	Yes	Hh-Light2, Sufu−/− MEFsc Med1-MB cellsd	Primary Hh-MB allograft Ptch−/− allografts	Preclinical	73
Dacinostat		HDAC antagonism	Yes	Daoy and D283 cellsd	Daoy xenografts	Preclinical	75
Quisinostat		HDAC antagonism	Yes	Daoy and D283 cellsd	Daoy xenografts	Preclinical	75
JQ1		BET-bromo domain antagonism	Yes	Med1-MB and SmoWT-MB cellsc	Med1-MB allografts SmoWT-MB and SmoD477G-MB allografts MYC-amplified Grp3-MB xenografts	Preclinical	77–79
Birabresib		BET-bromodomain antagonism	Yes	HD-MB03, Daoy, ONS-76 and UW228d	MYC-amplified HD-MB03 xenografts	Preclinical	80
I-BET151		BET-bromodomain antagonism	Yes	Hh-Light2 cellsa Sufu−/− MEFsc	Ptct+/− MB allografts	Preclinical	81
GSK-LSD1		LSD1 antagonism	Yes	MG cellsd	MG flank allografts	Preclinical	85
CBL0137		FACT antagonism	Yes	D425, D458 and HD-MB03 cellsd	HD-MB03 flank/intracranial xenografts	Preclinical	87
Buparlisib		PI3K antagonism	Yes	Twelve MB cellsd	Daoy xenografts	Preclinical	93
Temsirolimus (TEM)		mTOR antagonism	Yes	Daoy cellsd	Daoy xenografts	Phase I	94–96
Milciclib		CDK2 antagonism	Yes	GTML2, MB002, Sd425 and D283 cellsd	GTML2 orthotopic xenografts MB002 xenografts	Preclinical	100
Palbociclib		CDK4/6 antagonism	Yes	Med-211FH cellsd	Med-211FH or Med-1712FH subcutaneous xenografts	Phase II	105
Silmitasertib		CK2 antagonism	Yes	MB21, MB55 and MB53 cellsc,e	Ptch+/−; Tpr53−/−; SmoD477G MB allografts	Phase I/II	106
Prexasertib		CHK1 antagonism	Yes	Grp3-MB cellsf	orthotropic Grp3-MB xenografts	Phase I	109,110
RKI-1447		ROCK antagonism	NA	Nine MB cellsd	D425 xenografts	Preclinical	114
Adavosertib		WEE1 kinase antagonism	Limited	Daoy and UW228 cellsd	Daoy xenografts	Phase I/II	115,116
S29		Src kinase antagonism	NA	Daoy and D283-MED cellsd	Daoy xenografts	Preclinical	118
Dasatinib		Src kinase antagonism	Yes	Daoy and D556 cellsd	PS125 Hh-MB allografts	Phase I	120
Roflumilast		PDE4D antagonism	Yes	NIH3T3 cellsa	SmoD477G-MB allografts	Preclinical	124
RK-33		DDX3 antagonism	NA	Daoy and UW228 cellsd	Daoy xenografts	Preclinical	125
Veliparib		PARP antagonism	Yes	NA	Pediatric MB orthotopic xenografts	Phase I	127–129
Tolfenamic acid		Sp1 degradation	Yes	Daoy and D283 cellsd	D283 subcutaneous xenografts	Preclinical	132
Lovastatin		HMG-CoA reductase antagonism	Yes	Daoy and D283 cellsd	Daoy subcutaneously/orthotopic xenografts	Preclinical	133,134
Simvastatin		HMG-CoA reductase antagonism	Yes	Daoy, D283, D341 and Ptck+/− mice detived cell lined	Ptch+/− CB17/SCID subcutaneous allografts	Preclinical	135,136
Verapamil		ABC transporter antagonism	Yes	TE671and Daoyd	Daoy xenografts	Preclinical	138,139
Celecoxib		COX-2 antagonism	Yes	D234-Med and D283-Mede,f	D283-Med xenografts MB-DP orthotopic xenografts	Phase II	143
Ganciclovir		DNA polymerases antagonism	Yes	D234, D283-Med and UW228–3 cellse,f	D283-Med xenografts	Preclinical	141,144
Fingolimod		PP2A agonism	Yes	D341, D384 and D425-Med	D425-Med xenografts	Preclinical	146
Phenformin		mGPD antagonism	Yes	Med1-MBd	Med1-MB allografts	Preclinical	148
AEE788		VEGFR1/2 and HER1/2 antagonism	Yes	Daoy and D283 cellsd	Daoy, DaoyPt or DaoyHER2 xenografts	Phase I (Discontinued)	149,151
Lapatinib		EGFR and HER2 antagonism	Limited	Daoy, D283, and VC312 cellsd	Daoy xenografts	Phase II (Lack of efficacy)	152,153
Nilotinib		Multi-target tyrosine kinases and Smo antagonism	Yes	MB-PDX and Daoy cellsd	MB-PDX xenografts	Preclinical	155
Foretinib		Multi-target tyrosine kinases antagonism	Yes	Daoy cellsd	Daoy and ONS76 subcutaneous xenografts Daoy orthotopic xenografts Transgenic model of primary Hh-MB	Preclinical	158
Piperlongumine (PL)		ROS generating Raf-1 and CDK2 antagonism	Yes	D425 and D458 cellsd	D458 Subcutaneous xenografts D425 intracranial xenografts	Preclinical	159,160
Alsterpaullone (ALP)		CDK1, CDK5 and GSK3P antagonism	Yes	D425 and D458 cellsd	D458 Subcutaneous xenografts D425 intracranial xenografts	Preclinical	159,160
Pazopanib		Multi-target tyrosine kinases antagonism	Yes	Daoy, MEB-Med-8A, D283 Med, D341 Med cellsd	MEB-Med-8A orthotopic xenografts	Phase I	161
AT-9283		Multi-target kinases antagonism	NA	Daoy and D556 cellsd	NA	Phase I	164,165
Digoxin		Erk and Akt modulating	Yes	Med8A and D283 cellsd	ICb-2555MB and ICb-1078MB orthotopic PDX models	Preclinical	166
Disulfiram		ROS generating MAPK agonism NF-κB antagonism	Yes	Daoy cellsd	Daoy intracranial xenografts	Preclinical	167–169
Mebendazole		Hh and VEGFR2 antagonism	Yes	Hh-Light2 and C3H10T1/2 cellsa Daoy and SmoD477G MEFs cellsc,e	Daoy orthotopic xenografts Ptch+/−, p53−/−MB allografts D425 xenografts	Phase I/II	170,171

^(a)Gli-Luc reporter assay

^(b)BODIPY-Cyc competitive binding assay

^(c)RT-qPCR assay monitoring Gli1 mRNA expression

^(d)Development status against MB unless otherwise denoted

^(e)Cell viability/proliferation assay

^(f)Colony-forming assay. NA = no available data.

2. Wnt antagonists

The Wnt signaling pathway is a highly conserved signaling cascade that is essential for embryonic patterning and development.¹⁰ Approximately 10% of MB patients are characterized by Wnt dysregulation and classified in Group 1.¹ Of these, 85–90% harbor somatic activating mutations in exon 3 of CTNNB1 (encoding β-catenin). Wnt-MB patients have the best outcome for patients less than 16 years of age as over 95% have a survival rate greater than five years.¹⁰ The Wnt-MB subgroup is identified by the β-catenin nucleopositive immunophenotype, suggesting that nuclear β-catenin accumulation is a marker of favorable outcome in MB.¹¹ In normal cells, the cytoplasmic β-catenin levels are maintained through phosphorylation of β-catenin by a cytoplasmic destruction complex consisting of adenomatous polyposis coli (APC), axin, casein kinase 1α (CK1α), and glycogen synthase kinase (GSK3β). β-catenin is degraded by the ubiquitin proteasome complex. Mutations in the phosphorylation site lead to β-catenin stabilization, which leads to its translocation to the nucleus where it interacts with the TCF/LEF complex and causes up-regulation of Wnt target genes, such as Axin2, CCND1, MYC, and Survivin.

Substantial efforts have been made to develop therapeutic agents that modulate the Wnt pathway and many of these are currently being investigated as potential antitumor therapies in preclinical and/or clinical trials.^12,13 However, it is still unclear whether targeting the Wnt signaling cascade can provide significant clinical benefits as no Wnt modulators have been approved as antitumor agents.¹³ Moreover, despite a direct relationship between Wnt/β-catenin activation and MB development, no Wnt-targeting compound under clinical trials has been tested against Wnt-dependent MB. Wnt inhibitors have proven efficacious in preclinical models of colon, breast and liver cancer, very few have been profiled in MB-bearing animal models.^12,13 By contrast, it has been noted that Wnt activation is involved in neural regeneration in the central nervous system (CNS) after injury.¹⁴ Combined with the recent suggestions of an anti-tumorigenic role for Wnt activation in Hh-MB,¹⁵ and Grp3/Grp4-MB samples,¹⁶ Wnt/β-catenin agonism has been postulated as an anti-cancer strategy. More clinical studies on Wnt antagonist/agonists are needed to better outline the therapeutic potential of Wnt-signaling manipulation in MB treatment.

2.1. Pyrvinium

Using Xenopus laevis egg extract to screen for compounds that both stabilize Axin and promote β-catenin turnover, Thorne et al identified an FDA-approved anthelmintic drug, pyrvinium, as a potent inhibitor of Wnt signaling (IC₅₀ = 10 nM).¹⁷ CK1α was identified as the critical cellular target of pyrvinium and these studies showed that pyrvinium acts as an allosteric activator of this protein kinase. In vitro and cellular binding studies demonstrated that pyrvinium avidly binds to CK1a (K_d = 1 nM).¹⁷ As CK1a is also involved in Hh signaling and implicated as a negative regulator, a later study proved that pyrvinium indeed attenuated Hh signaling.¹⁸ In Hh-dependent cells, pyrvinium inhibited Hh activity with an IC₅₀ of 10 nM. Pyrvinium has also been validated in vivo in a well-accepted Hh-driven mouse tumor model. Pyrvinium (0.8 mg/kg × 11d, QOD, sq.) dramatically reduced tumor growth in Ptch^+/− MB allografts.¹⁸ Of note, in the pyrvinium-treated tumors, the expression of Hh target genes were decreased relative to the control. However, Wnt target genes, which are not hyperactivated in this subtype of MB, were not different between pyrvinium and vehicle group.¹⁸

2.2. Fenretinide

Fenretinide, a synthetic analogue of all-trans retinoic acid, has emerged as a promising cancer chemopreventive and chemotherapeutic agent for various cancers.¹⁹ Investigation of the effects of fenretinide on the immortalized human MB cell lines Daoy and ONS-76 resulted in GI₅₀ values of 5–10 μM and 2–10 μM, respectively.¹⁹ Fenretinide induced down-regulation of Wnt3a expression levels in both Daoy and ONS-76 cells, which correlated with downregulation of its downstream factors (Axin-1, GSκ3β and β-catenin) in Wnt pathway. Significant tumor growth inhibition was observed in Daoy and ONS-76 xenografts (12 mg/kg × 26d, QOD, ip.). Of note, it has been suggested that the in vivo efficacy of fenretinide involved multiple targets and mechanisms, including apoptosis, cell cycle, reactive oxygen species (ROS), and stemness, rather than solely Wnt-pathway inhibition.

3. Hh antagonists

The Hh pathway is a signaling cascade that plays a crucial role in cell growth and tissue patterning.²⁰ Canonical Hh signaling is initiated by two transmembrane proteins, Patched (Ptch) and Smoothened (Smo). In the absence of an Hh ligand, the 12-transmembrane domain receptor Ptch resides on the cell surface and suppresses activation of the 7-transmembrane domain receptor Smo. In this state, Suppressor of Fused (Sufu) serves as a negative regulator of pathway signaling by promoting phosphorylation and truncation of the Gli proteins, which ultimately act as repressors of Hh responsive genes. When an Hh ligand binds to Ptch at the cell surface, its suppression of Smo is relieved. Active Smo drives disruption of the Sufu/Gli heteroprotein complex and allows for full-length Gli to migrate to the nucleus where it induces the transcription of Hh-related genes, such as Ptch1, Gli1/2, TERT, and DDX3X.²¹ In healthy adults, the Hh pathway is typically quiescent; however, constitutive activation of Hh signaling has been observed in various types of cancers, including leukemia, basal cell carcinoma (BCC) and MB.²² Clinically, the Hh-dependent MB patients comprise approximately 30% of all MBs.²³ Among which, Ptch mutations are the most common MB driver.²⁴

A primary consideration for Hh pathway inhibitors as anti-cancer agents has been the rapid rise of resistance to Hh inhibitors that function through direct Smo antagonism. This resistance is most often associated with mutant forms of Smo that are not susceptible to inhibition with these small molecules. Several clinically relevant drug-resistant Smo mutants have been identified in patients receiving Smo antagonists for Hh-dependent MB and BCC.^20,21 Point mutations in the primary small molecule Smo binding site reduce compound affinity and efficacy.

Since the natural steroidal alkaloid cyclopamine (Cyc) was identified as the first Smo antagonist to block Hh signaling, strategies to suppress the development of Hh-dependent MB at different levels of signal transduction have been widely explored, including designing Ptch receptor modulators, developing Smo antagonists, or synthesizing Gli binders that directly or indirectly interfere with DNA binding.^25,26 Several of these have been approved and are marketed drugs for the treatment of advanced BCC, while many more Hh pathway inhibitors are in various stages of preclinical or clinical evaluation. For this review, we will focus on Hh pathway inhibitors that have been explored against Hh-dependent MB.

3.1. Vismodegib, Sonidegib and Glasdegib

Vismodegib (GDC-0449, Genentech, Inc.) is an Hh pathway inhibitor that blocks signal transduction through direct Smo binding.²⁷ Vismodegib demonstrates potent Hh inhibition with an IC₅₀ value of 13 nM in Gli luciferase reporter assays in C3H10T1/2 mouse embryonic fibroblasts (MEFs). Oral gavage of vismodegib (12.5 mg/kg × 8d, BID, po.) produced complete regression in tumor growth in Ptch^+/−-derived MB allografts.²⁷ The approval of vismodegib for the treatment of metastatic BCC or locally advanced BCC in 2012 represented the first Hh pathway inhibitor approved by the FDA.²⁸ Genentech also conducted several phase II studies in patients with operable nodular BCC, keratocystic odontogenic tumors (KCOT), basal cell nevus syndrome (BCNS), and MB. Of note, one trial evaluating vismodegib in combination with TMZ in patients with recurrent or refractory Hh-dependent MB has been terminated, due to lack of success at the first stage of phase II (http://clinicaltrials.gov).

Sonidegib (NVP-LDE225, Novartis, Inc.) was the second Hh pathway inhibitor approved by the FDA. As a Smo antagonist, sonidegib demonstrated IC₅₀ values of 1.3 nM and 2.5 nM against mouse and human Hh signaling, respectively.²⁹ Sonidegib (20 mg/kg × 13d, QD, po.) significantly inhibited tumor growth, corresponding to a T/C value (treated to control value) of 83% in a Ptch^+/−; p53^−/− subcutaneous allograft model of Hh-dependent MB.²⁹ Sonidegib (40 mg/kg × 4d, BID, po.) also proved centrally efficacious in an orthotopic Ptch^+/−; p53^−/− MB model.²⁹ The initial indication for sonidegib was adult patients with locally advanced BCC that recurred following surgery or radiation therapy, or patients that were not candidates for surgery/radiation. Additionally, experiments to evaluate efficacy against new indications, such as recurrent or refractory MB, advanced or metastatic hepatocellular carcinoma, relapsed acute leukemia, and in combination with other drugs for the treatment of certain solid tumors are also under investigation (http://clinicaltrials.gov).

The N-phenylbenzamide glasdegib (PF-04449913, Pfizer, Inc.) inhibits Hh signaling with an IC₅₀ value of 5 nM from a Gli Luc reporter assay in C3H10T1/2 cells.³⁰ As an anti-MB agent, glasdegib is still at the preclinical stage. Treatment of Ptch^+/−; p53^+/− or Ptch^+/−; p53^−/− MB allografts with glasdegib produced potent dose-dependent inhibition of Hh-pathway activity resulting in stable tumor regression.³¹ Glasdegib has been extensively evaluated in hematologic malignancies. In 2018, glasdegib was approved, in combination with low-dose cytarabine, for the treatment of newly diagnosed acute myeloid leukemia (AML) in adults who are above 75 years old or who have comorbidities that preclude use of intensive induction chemotherapy. Other studies against broader indications, such as chronic graft-versus-host disease, glioblastoma multiforme (GBM), myelodysplastic syndrome and chronic myelomonocytic leukemia were launched (http://clinicaltrials.gov).

3.2. Taladegib

Taladegib (LY2940680, Eli Lilly Inc.) is a Smo antagonist that showed potent anti-Hh activity with IC₅₀ values ranging from 4.56 to 7.64 nM.³² Taladegib also retains inhibitory activity against a mutant form of Smo that is resistant to other Smo antagonists (Smo^D473H) with an IC₅₀ value of 400 nM.³³ Taladegib potently inhibited Daoy cell growth in vitro, with an IC₅₀ value of 0.79 μM. In Ptch^+/−; p53^−/− transgenic mice, oral administration of taladegib produced remarkable efficacy and significantly improved overall survival.³³ In addition, taladegib inhibited Hh-regulated gene expression in the subcutaneous xenograft tumor stroma and produced significant anti-tumor activity.³³ Clinical studies in treating patients with localized esophageal or gastroesophageal junction cancer have advanced to phase I/II. Other studies in participants with advanced cancers, including CNS tumors, have also been launched. Of note, a phase I study in pediatric MB or rhabdomyosarcoma has been withdrawn (http://clinicaltrials.gov).

3.3. NVP-LEQ506

NVP-LEQ506 is a Smo antagonist developed by Novartis that had as its primary development goal to maintain anti-Hh activity against vismodegib-resistant Smo mutants. NVP-LEQ506 retained high potency (IC₅₀ = 96 nM) in C3H10T1/2 cells transfected with Smo^D473H, which emphasizes its potential.³⁴ NVP-LEQ506 also displayed promising blood–brain barrier (BBB) penetration with a brain/plasma AUC_0–∞ ratio of 0.69 after a single dose of 20 mg/kg in mice.³⁴ Oral administration of NVP-LEQ506 (40 mg/kg or 10 mg/kg), achieved comparable tumor regression to sonidegib at lower dosage in a Ptch^+/−; Hic^−/− MB allograft model in both mouse and rats.³⁴ This compound recently completed a phase I clinical trial for safety, tolerability, and PK/PD research in patients with recurrent or refractory MB (http://clinicaltrials.gov).

3.4. MK-4101

MK-4101 is a novel Smo antagonist that is structurally unique to previous antagonists described above.³⁵ MK-4101 inhibited Hh signaling with an IC₅₀ value of 1.5 μM in the Hh-Light2 cells. In vitro, MK-4101 inhibited the proliferation of MB cells derived from neonatally irradiated Ptch^+/− mice with an IC₅₀ of 0.3 μM. MK-4101 penetrates the BBB and has significant CNS exposure with a brain AUC/plasma AUC ratio of 0.6. A benefit of its structural uniqueness, MK-4101 retained activity against vismodegib-resistant mutant Smo (Smo^D477G).³⁵ In vivo, MK-4101 showed dose-dependent tumor growth inhibition (40 and 80 mg/kg, QD, po.) and tumor regression at the highest dosage (80 mg/kg × 24d, BID, po.) in CD1 nude female mice bearing Ptch^+/− MB allografts. From a longer term in vivo efficacy study, MK-4101 completely inhibited tumor growth and prevented tumor relapse after treatment termination, strongly suggesting that the tumor had been durably eradicated.³⁵ Lastly, MK-4101 (80 mg/kg, BID, po.) treatment significantly improved survival in neonatally irradiated Ptch^+/− mice bearing primary MB.³⁵

3.5. TAK-441

A high throughput screen in a cell-based Gli-Luc reporter assay identified compounds with a thieno[3,2-c]quinoline-4-one scaffold as Hh-pathway inhibitors with IC₅₀ values in the nanomolar range.³⁶ After several rounds of SAR study, 12b was identified as a potent Hh pathway inhibitor (IC₅₀ = 4.6 nM) with good metabolic stability. In an efficacy study in the Ptch^+/−; p53^−/− MB allograft model, 12b (6.25 mg/kg × 14d, BID, po.) resulted in almost complete suppression of tumor growth (T/C = 3%).

A follow-up study of 12b at higher doses (100 mg/kg) revealed a low C_max value of 3.63 μg/mL, which is likely due to poor solubility of this compound.^37,38 Efforts to improve solubility by reducing the aromatic system of 12b led to the discovery of TAK-441 (Takeda Pharmaceutical). In vitro, TAK-441 potently inhibited Hh activity with an IC₅₀ value of 4.4 nM in a Gli-Luc reporter assay and 79 nM in Smo^D473H mutant cells.³⁷ Of note, the C_max value of TAK-441 (21.5 μg/mL) was greatly improved as compared to 12b when dosed at 100 mg/kg. TAK-441 (25 mg/kg × 14d, QD, po.) completely inhibited tumor growth with T/C value of 1%.³⁸ Although it entered clinical trials, its development was discontinued after phase I studies due to pipeline prioritization.³⁹

3.6. Itraconazole analogues

Itraconazole (ITZ), an FDA-approved antifungal agent, was identified as an Hh pathway inhibitor.⁴⁰ Strong evidence suggested that ITZ binds and directly inhibits Smo with IC₅₀ values of 800 nM (Gli_Luc reporter assay) and 140 nM (RT-qPCR assay monitoring Gli1 expression).^40,41 ITZ retained potent Hh inhibitory activity in vitro and in vivo in the presence of vismodegib-resistant Smo mutants.⁴⁰ Oral gavage of ITZ (100 mg/kg × 21d, BID, po.) largely suppressed the growth of allografted MB from a Ptch^+/− p53^−/− mouse model. Combined treatment with ITZ (100 mg/kg/day) and cyclopamine (25 mg/kg × 21d, BID, po.) produced a complete inhibition of tumor growth.⁴⁰ A collection of clinical experiments are currently ongoing or in the recruitment stage to evaluate the potential of ITZ as an anti-tumor agent (http://clinicaltrials.gov).

Despite the clinical efficacious of ITZ as an antifungal agent, several of its overall features are not inherently druglike. The molecular weight of ITZ (706 g/mol) is above the threshold typically targeted for small-molecule therapeutics (500 g/mol), and its aqueous solubility is notoriously poor (0.8 μM in PBS, pH 7.4).⁴² The presence of the BBB makes it more difficult for ITZ as an anti-MB candidate. Multiple formulations,^43,44 and chemical modifications^41,42,45,46 of ITZ have been developed and evaluated to improve its oral bioavailability and overall PK profile. Recent studies suggested that the triazole around the dioxolane region,^41,46 and the triazolone moiety of ITZ^42,45 were unnecessary for its potent anti-Hh activity. Getting rid of those redundant fractions led to the discovery of analogues with improved in vitro profiles. Further experiments to verify their in vivo efficacy in murine models of Hh-dependent MB are currently ongoing.

3.7. Arsenic trioxide

Arsenic trioxide (ATO), an FDA-approved drug for the treatment of acute promyelocytic leukemia (APL), was identified as a novel Hh pathway antagonist.⁴⁷ ATO displaces 4’,5’-bis(1,3,2-dithioarsolan-2-yl)fluorescein-(1,2-ethanedithio l)2 (FlAsH-EDT2), a green fluorescent-labeled organoarsenical from binding to recombinant Gli1 protein, suggesting ATO inhibits Hh signaling at the level of Gli1. ATO showed potent anti-proliferative activities against two different MB cells, sharing an IC₅₀ value of 0.4 μM in both Daoy and D556 cells.⁴⁷ ATO (0.15 mg/kg × 2w, TIW, ip.) increased survival of ND2:SmoA1 transgenic mice, which is a model of constitutively active Smo-dependent cancer.⁴⁷ In Ptch^+/− p53^−/− MB allografts, intraperitoneal administration of ATO (7.5 mg/kg × 9d, QD, ip.) significantly inhibited tumor growth. ATO also inhibited tumor growth in Smo^D477G MB at a comparable level to its inhibition of Smo^wt MB.⁴⁸ The effectiveness of ATO as a single agent or combination therapy for the treatment of advanced MB or other solid tumors is currently being evaluated in phase II trials. (http://clinicaltrails.gov).

3.8. MK-5710

A series of bicyclic hydantoin Smo antagonists was identified through a cell-based Gli-Luc reporter screen.⁴⁹ After extensive SAR studies, MK-5710 was discovered as an extremely potent Smo antagonist with an IC₅₀ value of 17 nM in Hh-Light2 cells, and a binding affinity for Smo of 22 nM.^49,50 MK-5710 inhibited proliferation of the MB cell line with a CC₅₀ value of 0.4 nM. Administration of MK-5710 in Ptch^+/− MB allografts at multiple doses (40, 80, 120 and 160 mg/kg × 9d, BID, po.) gave rise to sustained tumor regression. Analysis of tumor samples following biopsy revealed MK-5710 led to a significant time-dependent downregulation of Gli1 (77%, 4 hr post-administration) and >85% pathway inhibition 8 hr following administration (80 mg/kg).

3.9. PF-5274857

PF-5274857 (Pfizer Inc.) is a potent Smo antagonist (IC₅₀ = 5.8 nM against hSMO) that inhibits pathway activity in Hh-dependent MEFs with an IC₅₀ of 2.7 nM.⁵¹ Oral administration of PF-5274857 in Ptch^+/−; p53^+/− MB allografts resulted in dose-dependent inhibition of tumor growth and induced tumor regression (doses ≥ 10 mg/kg × 6d, QD, po.). PF-5274857 administration (30 mg/kg) in Ptch^+/−; p53^−/− MB allografts inhibited Gli1 expression in both tumor and skin tissues in a time-dependent fashion. Moreover, PF-5274857 was reported to have good BBB penetrating ability in rats and mice.

3.10. N-Phenylbenzamide analogues

Researchers at Genentech developed second generation N-phenylbenzamide analogues incorporating the 2-pyridyl biphenyl amide scaffold common to vismodegib.⁵² After extensive SAR research, analogues with polar groups in the para-position of the aryl amide ring optimized potency, displayed minimal CYP inhibition, and possessed good exposure in rats. Two compounds (9d and 14f) potently inhibited Hh signaling as measured by Gli1 down-regulation through a Gli-Luc reporter assay (IC₅₀ values = 22 nM and 4.0 nM, respectively). Both compounds demonstrated high exposure in mice, which translated into excellent in vivo activity in a Ptch^+/− MB allograft model that was similar (9d, 5 mg/kg × 20d, BID) or better (14f, 5 mg/kg × 20d, BID) than vismodegib. Unfortunately, neither compound retained significant activity against the Smo^D473H mutant.⁵²

To identify compounds that retained activity against the drug-resistant forms of mutant Smo, researchers at Genentech screened a panel of Smo antagonists that had not previously advanced past the preclinical stage.⁵³ Of these, compound 5 showed a terminal half-life (t_1/2) of 22 hours and displayed promising activity against both Smo^WT (IC₅₀ value of 300 nM) and Smo^D473H (IC₅₀ value of 700 nM) in the Gli-Luc reporter assay. The IC₅₀ values of compound 5 against endogenous human and mouse Smo are 8 nM and 27 nM, respectively. In a murine model of subcutaneous allografts of the Ptch^+/−; p53^−/−; Smo^D477G MB tumor line SG274, oral treatment with compound 5 (100 mg/kg × 11d, QD, po.) resulted in significant tumor regression. Tumor growth inhibition was accompanied by down-regulation of Gli1 levels, indicating that compound 5 can suppress Hh signaling mediated by vismodegib-resistant Smo in vivo.⁵³

Using a scaffold hopping strategy, a series of tetrahydropyrido[4,3-d]pyrimidine derivatives were developed as effective Smo antagonists.⁵⁴ Representative analogue 24 demonstrated superior activity compared to vismodegib in the NIH3T3-GRE-Luc reporter gene assay, with an IC₅₀ value of 7.5 nM. In a subcutaneous xenograft model of murine Hh-dependent Ptch^+/− MB, compound 24 (50 mg/kg × 14d, QD, po.) inhibited tumor growth.

By replacing the 4-methylamino-piperidine linker of taladegib with a pyrrolidin-3-amine moiety and the 1-methyl-1H-pyrazol group with a benzene ring, Zhang et al. generated a series of novel Smo antagonists to address the adverse effects and drug resistance of current Hh antagonists.^55,56 Of these compounds, L-4 exhibited an IC₅₀ value of 2.3 nM in Hh-Light2 cells. L-4 also significantly suppressed Hh pathway activity driven by the Smo^D473H mutant. Orally administered L-4 exhibited dose-dependent anti-tumor efficacy in the Ptch^+/−; p53^−/− MB allograft model (20 mg/kg × 13d, QD, po.).

3.11. Arylphthalazine analogues

A cell-based screen conducted at Novartis identified 1-amino-4-benzylphthalazines analogues, which share structural similarity to NVP-LEQ506, as micromolar inhibitors of the Hh-pathway.⁵⁷ Subsequent SAR studies resulted in the identification of Anta XV, which exhibited low nanomolar activity against Hh signaling in both TM3-Gli-Luc IC₅₀ shift assay (IC₅₀ value of 2.7 nM) and mouse and human Smo membrane filter binding assays (with IC₅₀ values of 5 and 8 nM, respectively). Anta XV displayed a good PK profile in mice, including good oral bioavailability (F = 73% under 5 mg/kg single dose) and demonstrated promising BBB penetration. Treatment with Anta XV in a Ptch^+/−; p53^−/− MB allograft model afforded dose-dependent reduction in tumor growth, with full regression observed at 40 mg/kg (× 8d, BID, po.).

Researchers at Amgen discovered a series of 1-piperazinyl-4-arylphthalazines as potent Smo antagonists.⁵⁸ Follow-up SAR studies to address the stability issues of the hit compound led to the discovery of compound 43, which was potent in both mSmo and hSmo binding assays (IC₅₀ values of 2.0 and 2.8 nM, respectively). Oral administration of 43 (10 mg/kg × 6d, QD, po.) in Ptch^+/−; p53^−/− MB allograft mice resulted in full tumor regression. Further evaluation of compound 43 and related compounds revealed a pregnane X receptor (PXR) activation liability, raising the potential for this series of compounds in transactivation of a variety of drug metabolizing enzymes and subsequent drug–drug interactions.⁵⁹ Continuing SAR studies designed to mitigate the PXR activation liability led to the discovery of compound 27,⁵⁹ which inhibited Hh signaling in vitro with IC₅₀ values of 0.7 and 1 nM against mSmo and hSmo, respectively. Favorable PK properties were observed in rat and mouse models. Compound 27 further proved efficacious at disrupting Hh signaling in NOD-Scid mice. Reduction of Gli1 expression in the skin of mice treated with 27 was observed, with an EC₅₀ around 0.46 μM.

Based on the structure of Anta XV, Dong et al. reported a series of arylphthalazine analogues focused on modifications to the piperazine linker of Anta XV.⁶⁰ Compound 10e with a piperidin-4-amine linker inhibited Hh signaling with an IC₅₀ value of 0.58 nM in Gli-Luc reporter assay. Compound 10e possessed potent anti-tumor activities against Ptch^+/−; p53^−/− derived MB cells in vitro (IC₅₀ = 38 nM). In vivo studies of 10e (40mg/kg × 17d, BID, po.) in Ptch^+/−; p53^−/− MB allografts reduced tumor growth. A similar design strategy to taladegib led to the discovery of compound 23b, which exhibited potent Hh inhibition in the Gli-Luc reporter assay and anti-proliferative activity against Ptch^+/−; p53^−/− derived MB cells, with IC₅₀ values of 0.17 and 24.6 nM, respectively.⁶¹ Compound 23b (50 mg/kg × 13d, BID, ig.) reduced tumor growth of Ptch^+/−; p53^−/− MB allografts; however, both 10e and 23b were inferior to the parent compounds Anta XV and taladegib in vivo.

3.12. Artemisinin analogues

The natural product artemisinin and its derivatives have been widely used in a clinical setting for the treatment of malaria. Zhang et al. synthesized a collection of artemisinin analogues by incorporating 4-chloro-3-(pyridin-2-yl)aniline, a well-established Smo-targeting bullet, to the C-10 position of artemisinin.⁶² Most of those analogues were able to inhibit Hh-signaling in the Gli-Luc reporter assay. Of these, compound 65 was identified as the most potent analogue, with an IC₅₀ value of 9.53 nM. Significant tumor growth inhibition was observed following treatment of Ptch^+/−; p53^−/− MB allografts with compound 65 (50 mg/kg × 13d, BID, ip.).

3.13. Deoxytylophorinine analogues

PF403, a metabolite of the bioactive natural product 13a-(S)-deoxytylophorinine, demonstrated potent anti-proliferative activity across multiple cancer cell lines. Mechanistic studies revealed that PF403 preferentially binds Smo in a similar manner to vismodegib and reduced Gli1 translocation to the nucleus by promoting interactions between Sufu, protein kinase A (PKA), and Gli1.⁶³ It exhibited potent anti-proliferative activity with an IC₅₀ of 0.013 nM against Daoy cells. However, its potent in vitro activity do not translate to in vivo efficacy, most probably due to poor PK properties.⁶³ Through incorporating a tert-butylcarbonyl group to the C-3 hydroxyl of PF403, CAT3 was developed as a suitable prodrug. PF403 can be found in both plasma and brain as soon as 5 min after oral gavage of CAT3 (10 mg/kg), with C_max of 37.76 ng/mL in the brain tissue. CAT3 (12 mg/kg × 10d, QD, po.) demonstrated potent tumor growth inhibition in a Daoy orthotopic xenograft model.

3.14. Phenylimidazole iminium analogues

A series of phenylimidazole iminium analogues have been recently identified as Hh pathway inhibitors.⁶⁴ Studies in genetically engineered cell lines suggested A-1 is functioning downstream of Sufu, which sets it apart from other Smo antagonists under development. Further research showed that A-1 is a specific inhibitor of Gli1 with IC₅₀ value of 1.42 μM in Sufu^−/− MEFs. SAR efforts led to the identification of compound E-1, which showed improved anti-Gli1 activity of 0.14 μM. In vivo studies of E-1 (2 mg/kg × 5d, BID, ip.) in Ptch^+/−; p53^−/− MB allograft mice reduced tumor volume by half compared to control at day 5.

3.15. 1,5-Oxazocin-6-one analogues

Through a Gli-Luc reporter assay-based screen in Hh-Light2 cells, BRD9526 was identified as an Hh pathway inhibitor (IC₅₀ = 60 nM).⁶⁵ Follow-up studies of BRD9526, including structural modification and mechanism-of-action evaluation, were carried out because of its structural novelty. From the extensive SAR studies, compound 29a was identified as a potent Hh inhibitor, with an IC₅₀ value of 23 nM in the Gli-Luc reporter assay.⁶⁶ Mechanism of action studies in genetically engineered cell lines indicated that 29a inhibited the Hh signaling pathway by suppressing the expression of the transcriptional factors Gli rather than by interrupting the Gli/DNA binding. Compound 29a (50 mg/kg × 15d, BID, ip.) inhibited proliferation of MB cells and inhibited tumor growth in Ptch^+/−; p53^−/− MB allograft mice. An in vivo PD study revealed that 29a effectively suppressed Gli1 expression in tumor cells 4 h post-administration of a single 50 mg/kg dose.

4. Epigenetic related targets

Epigenetic dysregulation is a common feature of cancers.⁶⁷ Nine epigenetic agents are currently approved for standard-of-care treatment in the U.S.: two DNA methyltransferase (DNMT) inhibitors, four histone deacetylases (HDACs) inhibitors, two socitrate dehydrogenase (IDH) inhibitors, and most recently, an inhibitor of the histone methyltransferase enhancer of zeste homolog 2 (EZH2).⁶⁸ Indications of these drugs include cutaneous/peripheral T-cell lymphoma, acute myeloid leukemia and follicular lymphoma.

Approximately 25% of MB patients are classified as Grp3-MB, predominantly among infants and children, with a peak diagnosis from 3 to 5 years old.⁶⁹ Grp4-MB represents approximately 40% of all MB patients and is the most prevalent MB subtype. Most Grp4-MB patients are identified between the ages of 3 and 16.⁶⁹ Grp3- and Grp4-MBs share several molecular similarities and often relate to poor prognosis and metastases. While distinct differences between these two subgroups are not well-defined, it has been acknowledged that MYC amplification is a common genetic feature of Grp3-MB while MYCN and cyclin-dependent kinase 6 (CDK6) amplifications are commonly reported in Grp4-MB.^24,69 In the last decade, many epigenetic modulators have demonstrated promise in treating MB, especially Grp3-MB. We highlight here the latest molecules in preclinical and clinical stages as potential MB treatments of targets implicated in either Grp3- or Grp4-MB.

4.1. HDAC antagonist

HDACs promote chromatin condensation and abrogate gene transcription related to cell cycle regulation and cell differentiation. Upregulation of several HDAC isoforms identified in MB contributes to aberrant cell cycle progression.⁷⁰ HDAC inhibitors are among the most studied class of epigenetic modulators and several have been approved for clinical use as anti-cancer agents. Multiple preclinical studies are evaluating the effects of these inhibitors on MB; however, few have advanced to clinical trials for this form of cancer.⁷¹

Panobinostat (LBH589), a pan-HDAC inhibitor, potently inhibited cell viability against various MB cell lines, with IC₅₀ values of 0.054 μM, 0.067 μM, and 0.046 μM in UW228, UW426 and Med8A cells, respectively.⁷² Panobinostat also significantly suppressed migration and adhesion while inducing cell cycle arrest and apoptosis in the above three MB cell lines. It also reduced MB leptomeningeal seeding, a strong negative prognostic factor of the disease, in a UW426-effLuc MB leptomeningeal seeding model.⁷² A modest survival enhancement was observed in the panobinostat-treated group (median survival of 25 days) over the control group (median survival of 21.5 days).⁷² The anti-MB effect of panobinostat was largely through suppression of MB leptomeningeal seeding without eliminating bulk tumors. An early phase I study in children and adults with recurrent MB was proposed in 2020 and is currently recruiting patients (http://clinicaltrials.gov).

Mocetinostat (MGCD0103) is an inhibitor of class I HDACs (including HDAC1/2/3/8) and prevents the growth of different tumors in preclinical and clinical settings.⁷³ Previous studies demonstrated that overexpression of HDAC1/2 were observed in Hh-MB. Hh-MB cells lacking HDAC1/2 grew more slowly, suggesting the requirement of the two deacetylases for Hh-dependent tumor growth.⁷⁴ In multiple Hh-dependent cellular models, mocetinostat decreased Gli1 expression.⁷³ Mocetinostat also suppressed Med1-MB cell (generated from a spontaneous tumor arising in a Ptch^+/−; lacZ mouse) growth in vitro and in vivo. In a primary Hh-MB allograft, mocetinostat (170 mg/kg × 20d, QD, po.) completely abrogated tumor growth. In a spontaneous intracranial Hh-MB model, mocetinostat repressed Hh-mediated transcription and induced apoptosis in tumor cells, suggesting its ability to cross the BBB. Oral gavage of mocetinostat in Ptch^−/− mice significantly increased their survival. A mechanistic study demonstrated that mocetinostat increased endogenous Gli1 acetylation, specifically at K518, which resulted in cell growth inhibition of Gli1-driven MB cells.

A cell-based anti-proliferative screen (~13K compounds) in Daoy cells to identify new drug candidates for MB identified two HDAC inhibitors dacinostat and quisinostat as superior anti-proliferative agents (IC₅₀ < 100 nM).⁷⁵ Both compounds also demonstrated potent anti-proliferative activity (IC₅₀ < 200 nM) against the MB cell line D283 and induced apoptosis and G₂/M arrest. Both dacinostat and quisinostat (20 mg/kg × 18d, QOD, ip.) significantly suppressed tumor growth in Daoy xenografts; moreover, both compounds reduced the expression of MYC, suggesting their potential uses in Grp3-MB.

4.2. BET-bromodomain antagonist

The BET (bromodomain and extraterminal domain) proteins function as epigenetic readers and transcriptional activators by recognizing acetylated lysine and chromatin through their bromodomains (BRD). The relationships between BRD2/3/4 and CNS tumors including MB have been extensively explored and well-reviewed.⁷⁶ Several BET inhibitors with validated in vivo efficacy against MB have been characterized.

JQ1, a small-molecule inhibitor targeting BRD4, was known to suppress tumorigenesis in preclinical models of hematological malignancies and neuroblastoma. JQ1 treatment down-regulated Gli1 expression, supporting an effect of BET inhibition on Gli promoters.⁷⁷ In Med1-MB and Smo^WT-MB cells, JQ1 treatment decreased cell viability with IC₅₀ values ranging from 50 to 150 nM.² In Med1-MB allografts, JQ1 (50 mg/kg × 21d, QD, ip.) significantly reduced tumor size and increased overall survival.⁷⁷ Similar results were also observed in Smo^WT-MB and Smo^D477G-MB allografts following JQ1 administration (50 mg/kg × 14d, QD, ip.).⁷⁷ MB patients with tumors that overexpress MYC or harbor a MYC oncogene amplification have an extremely poor prognosis.⁹ JQ1 treatment down-regulated MYC expression and resulted in transcriptional de-regulation of MYC targets in various MB cell lines.⁷⁸ In a xenograft model of MYC-amplified MB, tumor growth was significantly reduced following JQ1 dosing (50 mg/kg × 14d, QD, ip.), which correlated with down-regulated MYC expression at both mRNA and protein levels.⁷⁸ Since Grp3-MB patients demonstrate the worst survival outcomes of any subgroup, these preclinical results against Grp3-MB3 suggest the promise of JQ1 as a candidate to treat these high-risk MB patients.⁷⁹

Birabresib (OTX-015, MK-8628), a JQ1 derivative, is a potent BET inhibitor targeting BRD2/3/4 with IC₅₀ values from 92 to 112 nM.⁸⁰ Birabresib decreased cell viability in HD-MB03, Daoy, ONS-76 and UW228 MB cell lines in vitro, with GI₅₀ values between approximately 160 and 450 nM.⁸⁰ Similar to JQ1 treatment, MYC protein and mRNA expression was repressed after birabresib treatment. In Grp3-MYC-amplified MB03 xenografts, oral administration of birabresib (25 mg/kg × 17d, QD, po.) led to a significant decrease in tumor growth. Combination studies of birabresib and volasertib, a Polo-like kinase 1 inhibitor, resulted in synergism between the two compounds in vitro in a MYC-driven MB.

I-BET151 is another well-studied BET inhibitor with submicromolar activities against BRD2/3/4. Recent studies suggested I-BET151 attenuates Hh activity in multiple Hh-dependent cellular contexts with IC₅₀ values in the 30–100 nM.⁸¹ Knockdown of Brd4 resulted in a phenotype comparable to I-BET151 treatment, suggesting that BRD4 is a regulator of the Hh pathway. I-BET151 attenuated the expression of these pivotal Hh signaling components in Sufu^−/− MEFs. It also suppressed the growth of Hh-MB cells in vitro and in vivo. In Ptch^+/− MB allografts, I-BET151 (30 mg/kg × 6d, QD, ip.) significantly inhibited tumor growth and down-regulated Gli1 expression in the tumor cells.

4.3. LSD1 antagonist

As noted earlier, MYC is generally acknowledged as a key driver of Grp3-MB since the majority of Grp3-MB patients exhibit amplification of the MYC oncogene or over-expression of MYC protein. Interestingly, recent studies revealed MYC amplification alone is not sufficient to promote tumor growth.^82,83 An integrative genomics approach to identify additional drivers for Grp3- and Grp4-MB identified growth factor independent 1 (Gfi1) and growth factor independent 1B (Gfi1b) as prominent MB oncogenes and prevalent drivers of these two groups of MB.⁸⁴ Follow-up studies showed that Gfi1 expression is required for MB tumor maintenance and described a critical role for lysine demethylase 1 (LSD1) in mediating its oncogenic effects in MB.⁸⁵ Pharmacological inhibition of LSD1 by the known inhibitor GSK-LSD1 significantly inhibited proliferation of MYC + Gfi1^WT (MG) cells in vitro (IC₅₀ = 0.05 nM). In flank allografts of these MG tumors, GSK-LSD1 (10 mg/kg × 32d, 4on3off, ip.) treatment significantly slowed tumor growth in vivo. By contrast, no efficacy was observed in mice bearing intracranial MG tumors, suggesting that GSK-LSD1 does not accumulate in the CNS to therapeutic levels.⁸⁵ In addition, GSK-LSD1 suppressed tumor growth following surgical resection or ionizing radiation. These results strongly support the notion that targeting LSD1 with small molecule inhibitors could be an effective strategy for treating patients with Gfi1-driven MB.

4.4. FACT antagonist

Facilitates chromatin transcription (FACT) is a heterodimeric protein complex consisting of two subunits: structure-specific recognition protein 1 (SSRP1) and suppressor of Ty insertion (Spt16).⁸⁶ FACT functions as a histone chaperone to promote H2A/H2B dimer dissociation from the nucleosome and allow RNA polymerase II (PolII) transcription on chromatin, which ultimately promotes transcriptional elongation, DNA replication and DNA repair.⁸⁶ Up-regulation of the FACT complex has been reported in many types of cancer and may play a crucial role in tumorigenesis.

From a bioinformatic analyses of MB datasets and functional genomic screening from datasets of primary MYC amplified Grp3-MB lines, SSRP1 was identified as the top druggable candidate.⁸⁷ Curaxin CBL0137, a known FACT inhibitor, has entered into Phase I clinical trials in adult patients with hematological malignancies and solid tumors. CBL0137 reduced Grp3-MB cellular growth (D425, D458 and HD-MB03 cells) with submicromolar IC₅₀ values.⁸⁷ CBL0137 (70 mg/kg × 16d, Q4D, iv.) completely abrogated tumor growth in vivo in HD-MB03 flank xenografts. In another model bearing HD-MB03 intracranial orthotopic xenografts, CBL0137 (70 mg/kg, Q4D, iv.) significantly prolonged median survival. The therapeutic effect of CBL0137 was further demonstrated by its inhibition on MYC and NEUROD1, via depletion of FACT complex from their promoter regions. The effectiveness of CBL0137 in treating Grp3-MB models in vitro and in vivo highlights its promise to advance to clinical trials in the near future.

5. Protein kinases

Recent efforts with the use of high throughput genomic and proteomic techniques on tumor specimens and pre-clinical in vitro and in vivo models have unraveled numerous molecular targets that might underlie the occurrence and progression of MB. As described below, many of these targets were originally thought to be related to classic signaling pathways that drive MB; while others were seemingly functioning independently from Wnt or Hh pathways. A number of small molecule modulators that attenuate these aberrantly activated proteins have been developed with promising preliminary profiles. Unfortunately, several candidates such as MK-0752 (a γ-secretase inhibitor that blocks Notch-signaling) and tipifarnib (a farnesyltransferase inhibitor that blocks Ras-signaling) failed to deliver efficacious results in MB patients in clinical trials.^88,89 We highlight the most recent protein kinase inhibitors as potential MB treatments in this section.

5.1. PI3K/AKT/mTOR antagonist

Phosphoinositide 3-kinases (PI3Ks) are a family of lipid kinases that play key roles in cellular proliferation, differentiation, motility, survival and intracellular trafficking. Many of these functions relate to the ability of class I PI3Ks to activate protein kinase B (PKB, AKT) as in the PI3K/AKT/mTOR cascade. Class I PI3Ks are mutated and aberrantly activated in many cancers, including blood and CNS tumors.⁹⁰ Buparlisib (BKM120, Novartis Inc.) is a class I PI3K inhibitor, which has been evaluated in clinical trials against multiple forms of cancer.^91,92 Buparlisib demonstrated potent cytotoxic effects against a panel of twelve MB cell lines, with IC₅₀ values ranging from 0.279 to 4.38 μM.⁹³ Oral gavage of buparlisib (30 or 60 mg/kg × 60d, QD, po.) dose-dependently suppressed tumor growth and prolonged survival in Daoy xenografts. Interestingly, after treatment was terminated (day 60), mice that were treated with buparlisib had a slower tumor growth rate compared to the vehicle treated control mice.

Many cancers are characterized by activation of the mammalian target of rapamycin (mTOR) protein, a serine threonine kinase involved in cell cycle regulation, angiogenesis, and apoptosis. Temsirolimus (TEM) was the first mTOR inhibitor approved by the FDA for the treatment of advanced renal cell carcinoma in 2007. TEM inhibited Daoy growth in vitro.⁹⁴ In Daoy xenografts, TEM (20 mg/kg × 2w, 5on2off, ip.) significantly delayed tumor growth. An additive antitumor effect was observed when TEM was given in combination with cisplatin or camptothecins.⁹⁴ TEM has finished its phase I study in pediatric patients with recurrent/refractory solid tumors, including MB. It was suggested weekly intravenous TEM was well tolerated in children, demonstrated antitumor activity, and inhibited the mTOR signaling pathway in peripheral-blood mononuclear cells.⁹⁵ Another phase I study of TEM in combination with irinotecan (IRN) and TMZ has also been carried out in patients with relapsed or refractory solid tumors, including MB. The combination of TEM (35 mg/m²/dose, weekly, iv.), IRN (90 mg/m²/dose, days 1–5, po.) and TMZ (125 mg/m²/dose, days 1–5, po.) administered every 21 days is well tolerated in children.⁹⁶

5.2. CDK antagonist

Cyclin-dependent kinases (CDKs) play important roles in regulating the progression of the cell cycle and are often upregulated in cancer cells, including MB.⁹⁷ A previous study using Purvalanol A, a pan-CDK inhibitor, was effective in targeting MYC-driven tumors in vitro but was less effective in MYC-overexpressing transgenic animals, presumably due to solubility issues.⁹⁸ Interestingly, specific CDK2 inhibition was found to be synthetically lethal to MYCN-driven neuroblastoma, suggesting a potential role for CDK2 inhibitors in MBs harboring amplifications in MYCN and MYC.⁹⁹

Milciclib is a selective inhibitor of CDK2 with an IC₅₀ value of 45 nM. In vitro, milciclib potently inhibited GTML2 cell growth, a Grp3-MB cell line expressing high levels of MYC, with an IC₅₀ value of 0.94 μM.¹⁰⁰ Combination treatment of JQ1 and milciclib synergistically reduced GTML2 cell growth. This combination also effectively reduced viability of the MYC-amplified Grp3-MB cell lines MB002, Sd425, and D283 and the MYCN-amplified neuroblastoma cell line Kelly. In GTML2 orthotopic xenografts, milciclib (10 mg/kg × 14d, QOD, ip.) moderately improved survival rates; whereas combination treatment with JQ1 (50 mg/kg × 14d, QOD, ip.) significantly prolonged survival. Bioluminescent analysis of GTML2 tumors treated with combination therapy showed an initial reduction in tumor size during the time of treatment; however, the tumor burden levels recovered to vehicle levels when drug was withdrawn. Similar results were seen for milciclib and JQ1 alone and in combination in MB002 xenografts.¹⁰⁰

A mutagenesis approach identified cyclin-dependent kinase Inhibitor 2A (Cdkn2a) as a novel drug target for MB.¹⁰¹ Cdkn2a is a tumor suppressor gene encoding the Cyclin-dependent kinase inhibitor 2A protein (p16), that functions to inhibit CDK4/6. Subsequent analysis revealed the CDK4/6/Cyclin D/Rb pathway as a druggable pathway for all non-Wnt MB.^102,103 Palbociclib (Pfizer Inc.) is a selective inhibitor of CDK4/6 that functions to prevent retinoblastoma (RB) hyperphosphorylation and arrest cells in the G₁ phase of the cell cycle.¹⁰⁴ It has demonstrated efficacy in a variety of RB-positive tumors, including CNS tumors, and has been FDA-approved as part of a combination therapy for advanced breast cancer. In a murine model of patient-derived MYC-amplified MB tumors (Med-211FH), reduction in RB phosphorylation at Ser⁷⁸⁰ was observed following treatment with palbociclib.¹⁰⁵ Palbociclib treatment (75 mg/kg × 28d, QD, po.) resulted in tumor regression with an average reduction in tumor volume of 63% for both Med-211FH (a representative for MYC-amplified Grp3-MB) and Med-1712FH (a representative for Hh-MB) subcutaneous xenografts. However, 15 out of 20 palbociclib-treated tumors recurred within 60 days of drug withdrawal in a study carried out in Med-211FH xenografts, suggesting the tumor regression was not sustained.¹⁰⁵ Another efficacy study in Med-211FH intracranial xenografts showed that palbociclib (75 mg/kg/day, QD, po.) led to significant tumor regression as well as increased survival.¹⁰⁵ Palbociclib is currently in phase II clinical trials for the treatment of recurrent or refractory CNS tumors, including MB (http://clinicaltrials.gov). Although advanced preclinical data for ribociclib, another CDK4/6 inhibitor, as an anti-MB agent is not available, it is currently in a phase I trial in combination with gemcitabine, trametinib or sonidegib for recurrent MB (http://clinicaltrials.gov).

5.3. CK2 antagonist

A proteome-wide analysis of phosphorylation events was performed in murine granule neuron precursors during Hh-driven cellular proliferation. Casein kinase 2 (CK2) was identified as a possible modulator since its activity changed significantly during proliferation versus differentiation.¹⁰⁶ Genetic and small molecule inhibitor studies in Hh-dependent NIH3T3 cells confirmed a role for CK2 in Hh signal transduction.¹⁰⁶ The CK2 inhibitor CX-4945 reduced Hh activity and induced cell death with IC₅₀ values ranging from 2.5 to 5.3 μM in murine Ptch^+/− (MB21 and MB55) and Ptch^+/−; Tpr53^−/− (MB53) MB cell lines. In Ptch^+/−; Tpr53^−/−; Smo^D477G MB allografts, all control mice succumbed to the tumors by day 17. By contrast, 43% of mice treated with CX-4945 (37.5 mg/kg × 30d, BID, po.) survived past 100 days even though drug treatment was discontinued after 30 days. Treatment of mice harboring either Ptch^+/−; Tpr53^−/− or Ptch^+/−; Tpr53^−/−; Smo^D477G MB allografts with the CK2 inhibitor tetrabromobenzotriazole (30 mg/kg, BID, ip.), significantly inhibited tumor growth inhibition in both models. These findings led to a phase I/II study investigating the use of the CK2 inhibitor CX-4945 in patients with Hh-MB and BCC (http://clinicaltrials.gov).

5.4. CHK1 kinase antagonist

Checkpoint kinase 1 (CHK1) is a serine/threonine kinase involved in activating and maintaining the S and G₂/M checkpoints. Inhibition of CHK1 in the absence of p53 leads to loss of DNA damage checkpoints and can enhance the activity of many DNA-damaging agents.^107,108 A recent study using Daoy and D283 cells suggested CHK1 expression is an adverse prognostic marker and potential therapeutic target in MYC-driven MB.¹⁰⁸ Prexasertib (LY-2606368), a potent and selective CHK1 inhibitor with IC₅₀ <1 nM, is currently under clinical evaluation for treatment of pediatric malignancies (http://clinicaltrails.gov). A PK/PD study performed in an orthotropic Grp3-MB xenograft model suggested good drug exposure to brain and tumor.¹⁰⁹ The CNS penetration of prexasertib in tumor-bearing mice was statistically greater than that observed in non-tumor bearing mice. The increases of nuclear pCHK1 S345 and pH2Ax levels in tumor tissues after a single dose of prexasertib (20 mg/kg, iv.) confirmed a treatment-induced target engagement.¹⁰⁹ In another study, Endersby et al found that prexasertib reduced the colony-forming ability of MB cells post-irradiation.¹¹⁰ In vivo, tumor-targeted radiation therapy and prexasertib each induced modest apoptosis in orthotopically-implanted MB, which was significantly increased when co-administered. This combination therapy also improved the survival of mice with MB. A phase I evaluation of prexasertib in combination with cyclophosphamide (CP) or gemcitabine was carried out for patients with refractory or recurrent Grp3-, Grp4- or Hh-MB brain tumors (http://clinicaltrails.gov).

5.5. ROCK antagonist

The Rho/Rho-associated coiled-coil containing protein kinase (ROCK) signaling pathway is involved in many biological processes, including cell adhesion and migration.¹¹¹ ROCK1/2 are dysregulated in a variety of cancers and have been implicated in tumor motility, invasion, and growth.¹¹² Preclinical studies have demonstrated a therapeutic potential for ROCK inhibition in a variety of cancers, including CNS tumors.¹¹³ RKI-1447 is an inhibitor of ROCK1 and ROCK2 with IC₅₀ values of 14.5 and 6.2 nM, respectively. RKI-1447 inhibited cell growth with a mean IC₅₀ of 7.28 μM in a panel of nine MB cell lines derived from different subgroups.¹¹⁴ Interestingly, in cells derived from Grp4-MB, RKI-1447 significantly inhibited cell growth in the relapse/metastatic cells (CHLA-01R-MED), but not the primary cells (CHLA-01-MED). In MB-D425 xenografts, RKI-1447 (80 mg/kg × 10d, QD, ip.) significantly suppressed tumor growth. Both Grp3- and Grp4-MB are highly metastatic and Grp4-MBs have high ROCK2 expression. These results supported RKI-1447 as a potential therapeutic candidate for metastatic MB patients.

5.6. WEE1 kinase antagonist

Also from an integrated genomic analysis of gene expression and a kinome-wide siRNA screen of MB cells and tissue, WEE1 kinase was identified as a potential therapeutic target for MB.¹¹⁵ Genetic and chemical inhibition of WEE1 potently suppresses cell growth, induces apoptosis and decreases tumor volume in MB.¹¹⁵ In vitro, Adavosertib potently suppresses WEE1 kinase activity (IC₅₀ = 18.7 nM) and Daoy and UW228 cell growth (IC₅₀ values of 150 and 232 nM, respectively). Oral gavage of adavosertib (30 mg/kg × 3w, TIW, po.) in Daoy xenograft mice decreased overall tumor size.¹¹⁵ Adavosertib displayed synergistic activity with cisplatin, further suggesting its promise as a clinically useful anti-MB drug.^115,116 Clinical studies for adavosertib are currently in phase I/II in combination with various chemotherapeutic agents against a number of cancer types (https://clinicaltrials.gov). Unfortunately, recent studies evaluating adavosertib in combination with radiation and/or chemotherapy in brain tumors have reported that adavosertib has limited efficacy and BBB penetration;¹¹⁶ therefore, further structural modification and/or formulation studies are needed to achieve optimal chemosensitizing effects in brain tumors such as MB.

5.7. Src kinase antagonist

The Src family kinases (SFKs) are a class of membrane-associated non-receptor tyrosine kinases including Src, Blk, Fgr, Fyn, Hck, Lck, Lyn, Yes, and Yrk, which are activated in response to cellular signals that promote cell growth, differentiation, cell shape, migration and survival. Particularly, Src plays an important role in angiogenesis and metastasis.¹¹⁷ Recently high Src activity was documented in an MB patient, emphasizing the potential for Src antagonists as anti-MB drugs.¹¹⁸ Giordano A. et al reported a new class of pyrazole-[3,4-d]-pyrimidine analogues as Src kinase inhibitors, which were derived from previously identified Src antagonists targeting different tumor species.¹¹⁸ A recently developed pyrimidine-based Src inhibitor S29 down-regulated p-SrcY416, which correlated to its anti-proliferative activity in Daoy (IC₅₀ = 1.72 μM) and D283-MED (IC₅₀ = 4.89 μM) cells. S29 (100 mg/kg × 40d, QD, po.) modestly inhibited tumor growth in Daoy xenografts in nude mice.

Dasatinib, an FDA-approved Src inhibitor, crosses the BBB and was well tolerated in pediatric patients.¹¹⁹ Dasatinib decreased proliferation, induced apoptosis, and abolished Src phosphorylation in MB cells (Daoy and D556).¹²⁰ Significant tumor regression was observed in a mouse syngeneic MB flank model after dasatinib treatment (15 mg/kg × 5d, QD, po.). A phase I clinical trial of dasatinib in combination with lenalidomide and TMZ in children with relapsed or refractory CNS tumors, including MB, has been completed. The results suggest that the combination at low metronomic doses is safe, meriting further evaluation in a disease-specific phase II setting.¹²¹

6. Other emerging targets

6.1. PDE4D antagonist

Phosphodiesterase 4D (PDE4D) belongs to the superfamily of PDE4 enzymes that degrade cyclic AMP (cAMP). The expression of PDE4 is essential for many physiological processes.¹²² Recently, genome-wide sequencing studies implicated mutations in PDE4D and PKA in adult Hh-MB, suggesting PDE4D and PKA as potential therapeutic targets.¹²³ Further mechanistic studies suggested that targeting PDE4D to inhibit the newly discovered Sema3-Nrp-PDE4D-PKA pathway powerfully inhibited the growth of Hh-MB.¹²⁴ Roflumilast, an FDA-approved PDE4D inhibitor, reduced Hh activity in NIH3T3 cells. In mouse MB allografts containing Smo^D477G, roflumilast significantly inhibited tumor growth. Hh signaling was decreased as assessed by Gli1 transcript levels in resected tumors.¹²⁴ These findings highlight the priority of repurposing PDE4D inhibitors as potential therapies for Hh-MB, especially for vismodegib-resistant tumors.

6.2. DDX3 antagonist

DDX3, a member of the RNA helicase family, is involved in many biological activities that regulate multiple steps in gene expression. Mutations in DDX3 have been reported in various cancers, including MB.²³ Approimately 8% of MB patients harbored mutations in DDX3.²³ The DDX3 inhibitor RK-33 reduced growth of MB cell lines Daoy and UW228, with IC₅₀ values of 2.5 μM and 3.5 μM, respectively.¹²⁵ Growth inhibition by RK-33 was associated with reduced transcript levels of Wnt-regulated genes, such as Axin2, CCND1, MYC, and Survivin. Single treatment of RK-33 (50 mg/kg × 14d, QAD, ip.) in Daoy-inoculated SCID mice resulted in a complete tumor growth inhibition by day 42. Moreover, a combination of RK-33 (50 mg/kg × 14d, QAD, ip.) and radiation (5 Gy) resulted in complete tumor reduction, suggesting RK-33 is able to radiosensitize MB tumors.

6.3. PARP antagonists

Poly (ADP-ribose) polymerase (PARP) is a family of nuclear enzymes involved in DNA damage that facilitates DNA repair.¹²⁶ Inhibition of PARP activity in certain cells results in DNA damage that leads to cell cycle arrest and apoptosis.¹²⁶ Veliparib is a PARP inhibitor that potentiates anticancer activity of chemotherapy drugs, including TMZ, and enhances radiation efficacy.¹²⁷ PARP is highly expressed in pediatric MB and GBM tumors, but absent in postnatal normal brains. Veliparib treatment potently inhibited PARP activity in mouse orthotopic xenografts of pediatric MB and GBM and enhanced TMZ efficacy.¹²⁸ Combination therapy of veliparib (25 mg/m², BID, po.) and TMZ (135 mg/m², QD, po.) was well tolerated in children with recurrent brain tumors, including MB, in a phase I clinical study.^128,129

6.4. Sp1 antagonists

Specificity protein 1 (Sp1) is a zinc finger transcription factor that binds to GC-rich motifs of many promoters, therefore, regulating several genes involved in cell proliferation and cell survival.¹³⁰ Tolfenamic acid (TA), a nonsteroidal anti-inflammatory drug (NSAID), inhibited neuroblastoma cell growth by targeting Sp1.¹³¹ Additional studies to test its activity in human MB cells (Daoy and D283) showed a dose-dependent inhibition in cell proliferation, with IC₅₀ values of 21.6 and 20.8 μg/mL, respectively.¹³² TA decreased expression of Sp1 in both Daoy and D283 cells, validating its cell-based Sp1 inhibition in MB cells. In a subcutaneous D283 xenograft model, TA (50 mg/kg × 4w, TIW, po.) significantly reduced tumor growth. Further studies assessing this agent with chemotherapy and radiation therapy are ongoing. Of note, despite TA lowering Sp1 expression in resected tumor tissue, determination that TA’s anti-MB efficacy was through Sp1-inhibition was inconclusive.

6.5. HMGCR antagonists

Lovastatin, a widely used β-hydroxy β-methylglutaryl-CoA reductase (HMGCR) inhibitor has demonstrated anti-cancer effects in multiple studies.¹³³ Lovastatin decreased proliferation of Daoy and D283 MB cells and induced apoptosis in vitro.¹³⁴ In a subcutaneous Daoy xenograft, either 0.2 or 1.0 mg/kg of lovastatin resulted in smaller tumours and lower C-MYC expression.¹³⁴ In a Daoy orthotopic model, lovastatin treated mice (1.0 mg/kg × 4w, TIW, ip.) demonstrated less tumor expansion in the ventricles; moreover, tumor invasion into the surrounding cerebellar tissues was completely blocked.

Simvastatin, another antihyperlipidemic drug, was also shown to decrease proliferation of MB cells isolated from Ptch^+/−. In addition, this HMGCR inhibitor induced apoptosis through activation of caspases and downregulation of anti-apoptotic proteins in Daoy, D283 and D341 MB cell lines.^135,136 An in vivo study revealed that simvastatin treatment (40 mg/kg × 20d, BID, ip.) of Ptch^+/− CB17/SCID subcutaneous allografted mice significantly reduced tumor growth, accompanied with repressed Hh-pathway activity. Combined treatment with vismodegib (5 mg/kg × 20d, BID, po.) and simvastatin (20 mg/kg × 20d, BID, ip.) resulted in a complete inhibition of Hh-MB progression.¹³⁶

6.6. ABC transporter antagonist

ATP-binding cassette (ABC) transporter family has 50 members in humans and is primarily responsible for the translocation of multiple substrates across membranes. A key mechanism when cells resist chemotherapy drugs is via ABC-mediated compound efflux.¹³⁷ Verapamil, an ABC transporter inhibitor used in cardiovascular diseases, has been recently reported to sensitize Daoy cells to radiation.¹³⁸ Verapamil showed antiproliferative effects in cultured MB cells TE671 and Daoy, limited Daoy expansion over time and caused a distinct downward shift in the size distribution of clonogenic colonies.^138,139 When pretreating Daoy cells with verapamil (50 μM) in vitro for 5 days, the flank xenografted tumor formation in vivo was significantly slowed compared to non-treated cells, suggesting it affected the tumor initiating process within cultures.

6.7. COX-2 antagonist

Many tumors, including MB, express cyclooxygenase 2 (COX-2) and inhibition of its activity induces apoptosis in MB cell lines and significantly suppresses the growth of MB in vivo.¹⁴⁰ Celecoxib, a known COX-2 inhibitor approved for inflammation, has been discovered as a potential anti-MB agent.¹⁴¹ Celecoxib reduced the clonogenic capacity and proliferation of D324-Med and D283-Med cells in vitro. In a mice harboring D283-Med xenografts, celecoxib (90 mg/kg × 12d, QD, po.) significantly reduced tumor growth in vivo.¹⁴¹ It also enhanced the efficacy of radiotherapy and synergistically improved survival of orthotopic MB-derived CD133/Nestin double-positive cells transplanted in immunocompromised mice, accompanied with downregulating p-STAT3 and STAT3-related gene C-MYC.^142,143 A phase II study for children with recurrent MB has been proposed to investigate the efficancy of five oral drugs, including celecoxib, in prolonging survival while maintaining good quality of life (http://clinicaltrials.gov).

6.8. DNA polymerases antagonist

Human cytomegalovirus (HCMV) is prevalent in the human population and encodes proteins that provide immune evasion strategies and promote oncogenic transformation and oncomodulation.¹⁴⁴ High prevalence of HCMV levels has been identified in primary MB, MB cell lines and MB xenografts.¹⁴¹ In addition, HCMV infection is known to induce COX-2 expression in MB cells.¹⁴¹ Ganciclovir, a FDA-approved antiviral drug targeting the DNA polymerase during HCMV replication, reduces MB growth in vitro and in vivo.¹⁴¹ It selectively reduced clonogenic capacity of HCMV-positive MB cells D234-Med, D283-Med and UW228–3. Oral gavage of valganciclovir (14 mg/kg × 12d, BID, po.), the prodrug of ganciclovir, rapidly released ganciclovir in vivo and reduced tumor growth in a D283-Med xenograft model in NMRI nu/nu mice.

6.9. PP2A agonist

Protein phosphatase 2A (PP2A) is a known tumor suppressor that is inactive or downregulated in many human tumors. Downregulated PP2A in MB led to upregulation of tumorigenic signaling pathways.¹⁴⁵ Fingolimod, an immunosuppressant drug used in multiple sclerosis, was reported to have therapeutic potential in MB.^145,146 Following treatment of MB cells D341, D384, and D425 with fingolimod (5 μM) for 4 hours, the activity of PP2A was significantly increased over baseline relative to control.¹⁴⁶ It has been proposed that fingolimod targets the endogenous PP2A inhibitors, I2PP2A/SET and CIP2A, as a mechanism for PP2A activation. Fingolimod decreased viability of MB cells, with IC₅₀ values of 7.5 μM in D425, 5.5 μM in D341 and 7.3 μM in D384. In D425 xenografts, fingolimod (10 mg/kg × 35d, QD, po.) significantly decreased tumor volume when compared to vehicle.

6.10. mGPD antagonist

Phenformin was prescribed for the treatment of diabetes until 1977 when it was removed due to the occurrence of lactic acidosis. Reports in recent years have documented its antitumor potential.¹⁴⁷ Its antitumor effect has been recently proposed through a redox-dependent mechanism.¹⁴⁸ Phenformin inhibited mitochondrial glycerophosphate dehydrogenase (mGPD), a component of the glycerophosphate shuttle, and caused elevations of intracellular NADH content in Hh-MB cells. Inhibition of mGPD promoted an association between corepressor C-terminal binding protein 2 and Gli1, thereby inhibiting Hh transcriptional output and tumor growth.¹⁴⁸ Phenformin inhibited Med1-MB cell growth in the low micromolar range. In a model of allograft tumors generated after injection of primary MB cells from Math1-CRE; Ptch^loxP/loxP mice, phenformin (300 mg/kg × 20d, QD, po.) significantly suppressed tumor growth. In a model of spontaneously generated intracranial Hh-dependent MB, the survival of mice treated with phenformin was increased by almost 40%, indicating the preclinical efficacy of this drug on MB.

6.11. Multi-target antagonists

AEE788, obtained by optimization of the 7H-pyrrolo[2,3-d]pyrimidine lead scaffold, is a potent dual inhibitor of both vascular endothelial growth factor receptor (VEGFR1/2) and human epidermal receptor (HER1/2).¹⁴⁹ Aberrant activation of VEGFRs and HERs are associated with advanced disease and poor patient prognosis in many tumor types, including CNS tumors.¹⁵⁰ In vitro, AEE788 inhibited MB cell proliferation (IC₅₀ values of 1.7 and 3.8 μM against D238 and Daoy, respectively) and prevented EGF- and neuregulin-induced HER1/2/3 activation.¹⁵¹ In vivo, AEE788 inhibited tumor growth in xenograft models of Daoy, cisplatin-resistant Daoy^Pt, and HER2-overexpressed Daoy^HER2. Clinical development of AEE788 was discontinued at phase I due to unacceptable toxicity and minimal activity for the treatment of recurrent GBM.¹⁵¹

Lapatinib is FDA approved for the treatment of postmenopausal women with hormone receptor positive metastatic breast cancer that overexpresses HER2. In vitro, lapatinib effectively inhibited Daoy, D283, and VC312 MB cell growth at 1 μM. Oral administration of lapatinib alone (100 mg/kg × 28d, BID, po.) did not inhibit tumor growth in Daoy xenografts;¹⁵² however, combination therapy with OSU-03012, a PDK1 inhibitor, resulted in significant tumor growth reduction. Multiple phase II trials for lapatinib in combination with other drugs against CNS tumors have been ineffective, presumably due to its inability to penetrate the BBB and accumulate at the tumor site.^153,154

A virtual screen for anti-Smo activity of existing approved or withdrawn drugs (1699 drugs in total), suggested the BCR-ABL kinase inhibitor nilotinib could inhibit Hh pathway activity through Smo antagonism.¹⁵⁵ These results were validated by the Gli-Luc reporter assay (IC₅₀ = 374 nM). Nilotinib was validated as a Smo antagonist through its ability to displace the known Smo antagonist BODIPY-Cyc in a competition binding assay. It inhibited growth of MB-PDX and Daoy cells, with IC₅₀ values of 5 μM and 6 μM, respectively. Oral gavage of nilotinib (40 mg/kg × 42d, QD, ip.) led to significant reduction in tumor volume as compared to vehicle group in MB-PDX xenografts.¹⁵⁵ Its anti-MB activity identified in this study was apparently a consequence of muti-pathway pharmacology antagonism. The ability of nilotinib to inhibit multiple targets simultaneously in Hh-dependent cancers makes it an interesting candidate for further development against Hh-dependent MBs.

The hepatocyte growth factor (HGF)/cMET pathway is essential for cell proliferation and migration during embryogenesis and critical for cerebellar development.¹⁵⁶ Previous studies have shown that cMET inhibition can effectively decrease MB cell migration and invasion.¹⁵⁷ Foretinib is an orally available multikinase inhibitor, targeting cMET with an IC₅₀ value of 0.4 nM.¹⁵⁸ Foretinib exhibited significant in vitro inhibitory activity in Daoy cells. In a Daoy subcutaneous xenograft model, therapy with foretinib significantly induced tumor regression at both 60 and 100 mg/kg (× 9d, QOD, po.). Similar results were achieved in ONS76 xenografts, with 50% tumor reduction at 100 mg/kg doses.¹⁵⁸ In orthotopic Daoy xenografts, oral gavage of foretinib (60 mg/kg × 2w, 6on1off, po.) significantly reduced tumor size and metastases. Lastly, foretinib (6 mg/kg × 28d, 2.5 μL/hour, osmotic pumps) effectively increased survival and reduced metastasis in Ptch^+/− mice bearing primary Hh-MB.

A Connectivity Map (C-MAP) analysis was carried out to discover small molecule inhibitors with high likelihood of efficacy against Grp3-MB.¹⁵⁹ Piperlongumine (PL), a natural product isolated from the fruit of the Piper longum, was the top candidate for non-Wnt tumors. PL was previously known to have anticancer, antiplatelet aggregation, anti-atherosclerotic, antidiabetic, anxiolytic, and antimicrobial activities through a variety of mechanisms.¹⁶⁰ PL reduced Grp3-MB cell proliferation (D425 and D458) at concentrations as low as 5 μM. Subcutaneous injection with PL (50 mg/kg × 14d, QD, sc.) markedly reduced tumor growth and increased survival in mice bearing D458 xenografts.¹⁵⁹ Another hit compound identified through C-MAP, alsterpaullone (ALP), is a GSK-3β, CDK5/p25, and CDK1/cyclin B inhibitor. It demonstrated similar anti-MB efficacy both in vitro (1 μM) and in vivo (30 mg/kg × 14d, QD, sc.). Both PL and ALP significantly increased survival and reduced tumor growth in nude mice with D425 cerebellar xenografts.¹⁵⁹

Pazopanib, a multi-kinase inhibitor that reduces tumor growth and inhibits angiogenesis, has been approved for the treatment of patients with renal cell carcinoma and soft tissue sarcoma. Pazopanib penetrates the BBB and was evaluated as a drug candidate for the treatment of MB.¹⁶¹ Pazopanib potently inhibited MB cell Daoy, Med-8A, D283 and D341 growth in vitro.¹⁶¹ Pazopanib also effectively induced apoptosis in MB cells by reduction of p-STAT3 levels. Oral gavage of pazopanib (60 mg/kg, QD, po.) significantly delayed tumor growth and prolonged survival from 22 days to 31 days in MEB-Med-8A orthotopic xenografts. A phase I study of pazopanib in children with soft tissue sarcoma and other refractory solid tumors, including MB, has finished. It was suggested pazopanib is well tolerated in children, with evidence of antiangiogenic effect and potential clinical benefit in pediatric sarcoma.¹⁶²

Aurora kinases (Aurks) play an important role in regulating cell mitotic division. Over-expressed Aurks have been found in many human cancers. In MB, overexpression of AurkA is an independent predictor of poor prognosis.¹⁶³ AT9283, a multi-targeted kinase inhibitor with potent activity against both AurkA/B, is currently being evaluated in clinical trials in patients with hematological malignancies. A recent study suggested AT9283 decreased Daoy and D556 cell proliferation and migration while inducing apoptosis, autophagy and cell cycle arrest.^120,164 As mentioned earlier, a synergy effect with dasatinib, a Src kinase inhibitor, was observed in vitro in MB cells.¹²⁰ Despite no reported in vivo efficacy results in MB models, AT9283 has advanced to a phase I trial in children and adolescents with solid tumors, including MB. It was suggested AT9283 was well-tolerated in patients with solid tumors with manageable hematologic toxicity.¹⁶⁵

Based on an informatics-driven drug repositioning strategy, digoxin, a cardiac glycoside steroid prescribed for chronic heart failure, was identified as an anti-Grp3/4-MB drug candidates.¹⁶⁶ Digoxin potently inhibited MB cells Med8A and D283 growth in vitro, with IC₅₀ values of 106 and 10 nM, respectively. Digoxin treatment (two cycles of 2 mg/kg × 14d, QD, ip., three weeks apart) significantly prolonged survival in vivo in two patient-derived orthotopic PDX models, ICb-2555MB and ICb-1078MB, representing Grp3- and Grp4-MB, respectively. In ICb-1078MB PDX, which harbors an N-MYC amplification, digoxin increased median survival to 113 days compared to 92 days for untreated controls. In mice implanted with ICb-2555MB, which harbors C-MYC amplification, digoxin-treated mice displayed a median survival of 180 days compared to 102 days for controls. Deep RNA-seq analysis on ICb-2555MB and ICb-1078MB tumors harvested from digoxin treatment group identified dozens of differentially expressed genes, many of which were reported to either play a role in or be a target of extracellular signal-regulated kinase (Erk) or Akt signaling, suggesting a potential model in which digoxin modulation of Akt and Erk signaling promotes MB cell death.

Disulfiram, an FDA-approved drug for the treatment of alcoholism, has been documented as a potential anticancer agent owing to its multiple pharmacological mechanisms in targeting tumor cells, including ROS generating, MAPK pathway activating, NF-κB inhibiting, and proteasome pathway suppressing.^167–169 Disulfiram was shown to induce apoptosis and decrease cell viability and colony formation ability in Daoy cells in vitro. Disulfiram (30 mg/kg × 5w, TIW, ip.) significantly inhibited tumor growth in Daoy intracranial xenografts.¹⁶⁹

Mebendazole (MBZ), an FDA-approved antiparasitic drug, has recently been characterized by its anti-tumor properties via inhibition of VEGFR2 and the Hh-pathway.^170,171 Larsen et al found that MBZ treatment prevented formation of the primary cilium and decreased expression of downstream Hh pathway effectors. MBZ inhibited Shh induced Hh activity in micromolar dose range in murine Hh-Light2 and C3H10T1/2 cells.¹⁷⁰ In Daoy cells, MBZ significantly reduced Gli1 and Ptch1 transcripts at 0.1 μM with almost complete suppression at 1 μM. MBZ inhibited Daoy cell proliferation and tumor growth in vitro and in vivo. In Daoy orthotopic mice, MBZ (50 mg/kg, QD, po.) significantly improved median survival from 75 days (control group) to 113 days (treated group), accompanied with significant tumor inhibition. A latter publication found that MBZ blocked VEGFR2-Y1175 autophosphorylation.¹⁷¹ Kinetic study showed MBZ inhibited VEGFR2 at an IC₅₀ of 4.3 μM. MBZ treatment (50 mg/kg, QD, po.) improved the survival of the mice by 150% in Ptch^+/−; p53^−/− MB allografts. MBZ was also efficacious in the Grp-3 MB cells D425, which carry amplification of C-MYC and OTX2. MBZ significantly reduced tumor growth and prolonged the median survival of D425 xenograft mice from 21 days to 48 days. It has been noted that both WT and Smo^D477G MEFs were functionally suppressed by MBZ in vitro and in vivo, suggesting the potential usage of MBZ as a novel anti-MB agent.^170,171 Several clinical trials of MBZ in patients with recurrent/progressive pediatric brain tumors, including MB, have been set up to evaluate the safety and efficacious profile (http://clinicaltrials.gov).

7. Conclusions and Perspectives

Current anticancer strategies rely to a great extent on targets identified from studying the genetics of cancers. Great knowledge has been gained in the field of MB in the past few decades as the origins of MB have allowed researchers and clinicians to classify these MBs into four subgroups. Through understanding the molecular pathways disrupted in MB, a better insight into molecular targets for therapy were identified and thus more specific drug candidates are being developed. Cytotoxic agents like VCR, CIS, CP, carboplatin and lomustine are currently the first-in-class chemotherapy protocols for both standard- and high-risk MBs in children under the guidance of the International Society for Pediatric Neurosurgery. As an adjuvant therapy, chemotherapy reduces the risk of tumor cells spreading through the spinal fluid and the risk of the tumor returning after surgery and/or radiation. While no targeted drugs for MB have been approved, the development of improved cytotoxic agents would definitely be a future direction as the current chemotherapy in the clinic often results in neurocognitive defects and other side effects.

Genomic and proteomic analysis in samples from MB tumors have led to the identification and characterization of new targets. Several of the most recently identified oncogenes such as MYC, MYCN, TP53, and SNCAIPR may serve better as molecular biomarkers for specific subtype, rather than as drug targets. Inhibitors selectively targeting these genes/proteins are either unavailable for clinical study or have been proven unsuccessful due to their comprehensive biological effects that cause toxicity. Mutations in epigenetic regulators have also contributed to MB tumorigenesis. Because epigenetic regulation is reversible, it has been explored for possible therapeutic targeting; however, due to the low efficiency for epigenetic drugs treating solid tumors, combination therapy with conventional anti-cancer agents may be a more effective treatment in the future.

To the best of our knowledge, we have highlighted every preclinical model used in profiling drug efficacy in this review. Multiple preclinical models have been developed and are currently being utilized to meet the requirements for anti-Hh-MB drug development, both in vitro and in vivo. By contrast, there is a desperate need for preclinical models to precisely define Wnt-, Grp3- and Grp4-MB subgroups. Another dilemma is that current preclinical models only cover a narrow spectrum of Grp3- and Grp4- heterogeneity. As our understanding of Grp3- and Grp4-MB subgroups evolves, these have been further divided into eight subtypes.¹ The emergence of these intermediate subtypes requires a more sophisticated molecular stratification to identify targets and develop drugs for these two classes of MB.

We have also discussed the current progression in anti-MB drug development at preclinical and/or clinical stages. Drug repurposing or repositioning has gained considerable popularity in cancer therapy, including MB. This is based on using previously approved drugs with known PK/PD characteristics and safety data for indications other than their traditional ones. The majority of potential anti-MB repurposed compounds target protein kinases. It is important to note that the development of kinase-targeted therapies for CNS diseases remains a challenge, as none of the approved protein kinase inhibitor drugs are for CNS indications.

Highlights.

• Medulloblastoma (MB) is clinically classified into four molecular subgroups.

• This review highlights the characteristics of each MB subgroup and identifies potential drug targets within each of these subgroups.

• The identification and current state of development for promising small molecule anti-MB agents is also described herein.