A review of the molecular profile of benign and malignant odontogenic lesions

ARTICLE IN PRESS Vol. 00 No. 00 && 2020

A review of the molecular profile of benign and malignant odontogenic lesions Filipe Fideles Duarte-Andrade, BSc, MSc,a J
essica Gardone Vit
orio, BSc, MSc,a Tha
ıs dos Santos Fontes Pereira, DDS, PhD,a Carolina Cavali
eri Gomes, DDS, PhD,b and Ricardo Santiago Gomez, DDS, PhDa Odontogenic cysts and tumors are heterogeneous lesions, originating from elements or remnants of the odontogenic apparatus. Although the majority of these lesions are benign and never undergo malignant transformation, rare malignant tumors may arise de novo or from benign precursors. The molecular basis of these lesions is still poorly understood. This article summarizes and discusses studies using small, medium, and large-scale and/or “-omic” techniques to describe the molecular characteristics of benign and malignant odontogenic lesions and briefly debates strategies to increase the use of “-omic” and multi-omic approaches or integrative analyses in the research of these lesions. A comprehensive understanding of the molecular aspects of odontogenic lesions by using large-scale approaches will enable us to refine the classification of this heterogeneous group of disorders and provide more accurate biomarkers for precise diagnosis, prognosis, and development of molecular tools in the management of patients with these conditions. (Oral Surg Oral Med Oral Pathol Oral Radiol 2020;000:1
12)

Odontogenic cysts and tumors consist of a group of heterogeneous lesions, originating from elements of the tooth-forming apparatus and their remnants.1 According to the World Health Organization Classification of Head and Neck Tumors published in 2017, odontogenic tumors are classified into epithelial, mixed, and mesenchymal tumors, whereas odontogenic cysts are classified as cysts of inflammatory or developmental origin.1 The majority of odontogenic tumors are benign, and the malignant ones may originate from a benign precursor or arise de novo.2 The molecular background underlying the pathogenesis of most odontogenic tumors and cysts remains poorly understood. However, in recent years, large-scale approaches at the global or “-omics” level have emerged as important tools to provide insights into the molecular pathogenesis of odontogenic lesions. The term -omics describe those fields of science whose purpose is to evaluate accurately and comprehensively the biomolecules of cells, tissues, biological fluids, organs, or organisms (Figure 1). Genomics focuses on identifying the genetic variants associated with a given condition on the whole-genome scale, for example, single nucleotide variants (SNVs), insertions or deletions (InDels), structural variants and others.3-5 Regarding cancer genomics, these assays aim to characterize the somatic and germline molecular defects driving or predisposing cancer development, respectively.6 Epigenomics studies the inheritable alterations of the genome that do not originate from a

Department of Oral Surgery and Pathology, School of Dentistry, Universidade Federal de Minas Gerais (UFMG), Belo Horizonte, Brazil. b Department of Pathology, Biological Sciences Institute, Universidade Federal de Minas Gerais (UFMG), Belo Horizonte, Brazil. Received for publication Aug 1, 2019; returned for revision Dec 27, 2019; accepted for publication Dec 29, 2019. Ó 2020 Elsevier Inc. All rights reserved. 2212-4403/$-see front matter https://doi.org/10.1016/j.oooo.2019.12.017

alterations in the DNA sequence, such as DNA modification (i.e., C5-methylcytosine “5mC,” and others), histone modification and variants, and nucleosome occupancy.7,8 Transcriptomics aims to evaluate, both quantitatively and qualitatively, RNA transcripts encompassing quantification of levels of expression, detection of sequence variation and splice variants, and analysis of the structural changes in RNA.9,10 Proteomics evaluates the level of production, occurrence of modifications, and interactions of proteins.4 Metabolomics evaluates the set metabolites arising from endogenous or exogenous processes.11 Finally, integrative “-omics” analyses combine information from different types of “-omics” approaches.4 Exhaustive description of “-omic” and multi-omic/integrative techniques is beyond the scope of this review, and interested readers are referred to other excellent reviews for further information.3-11 In this review, we summarize the molecular alterations identified in different odontogenic tumors and cysts by studies using large-scale and/or “-omic” techniques. For some odontogenic lesions, only results of small- to medium-scale approaches or even large-scale analysis are available, and these studies encompass techniques such as immunohistochemistry, targeted next-generation sequencing (NGS) panels, and microarrays (Table I and Supplementary Table SI; available at https://doi.org/ 10.1016/j.oooo.2019.12.017). Therefore, we have also included such techniques in the review. Moreover, we

Statement of Clinical Relevance Both “-omic” and multi-omic analyses of odontogenic tumors will enable clinicians and pathologists to refine the classification of this heterogeneous group of disorders, provide more accurate biomarkers, and allow the development of molecular tools for the management of affected individuals. 1

ARTICLE IN PRESS ORAL AND MAXILLOFACIAL PATHOLOGY 2 Duarte-Andrade et al.

OOOO && 2020

Fig. 1. Schematic illustration of the main “-omic” technologies and the flow of biological information. Each “-omic” focuses on the comprehensive analysis of a particular type of biomolecule (i.e., genomics, transcriptomics, proteomics, and metabolomics describe alterations involving DNA, RNA, proteins, and metabolites, respectively). Alternatively, multi-omic and integrative “-omic” approaches are capable of capturing the intricacy of interactions and alterations underlying the development of complex diseases, including tumors.

briefly present strategies for increasing “-omics” application in odontogenic cyst and tumor research and discuss future directions toward the use of multi-omics approaches.

LITERATURE REVIEW Benign epithelial odontogenic tumors Several small- to large-scale and “-omics” studies have elucidated many important molecular aspects of ameloblastomas.12-33 Calcifying epithelial odontogenic tumor and adenomatoid odontogenic tumor were also evaluated by “-omic” studies and are

reviewed in this section. Although squamous odontogenic tumor also belongs to this group, it has not been investigated by “-omic” techniques. Ameloblastoma. Ameloblastoma is a benign, locally destructive and infiltrative neoplasm of the jaws.1 Although intraosseous ameloblastomas are the most conventional ones, including the conventional solidmulticystic ameloblastomas, as well as the unicystic (UA) variant, in rare cases it can occur in the soft tissues, being classified as peripheral ameloblastoma.1 Considering that the conventional ameloblastoma, unicystic and peripheral cases have distinct biological behaviors, a

Odontogenic cysts and tumors

Genomic alterations

Transcriptomic alterations

Proteomic alterations Increased levels

References

2-hydroxypyridine, b-cyano-L-alanine, caprylic acid, glycerol, glycine, L-glutamic acid, Lserine, myristic acid, phosphoric acid, D-(+)-trehalose, glycolic acid

12-31, 33

Gene mutation

Chromosomal alterations

Increased expression

Decreased expression

Ameloblastoma

BRAF, FGFR2, HRAS, KRAS, NRAS, SMO

CNA losses: 21*, 22, 22q* 16q*, 19p*, 10, 10q* 10q26, 1p36, 3p24.3* CNA gains: 16p*, 1, 1q*

mRNAs: ODAM, FOS ncRNAs: LINC340, SNORD116-25, SNORA21, SNORA47, SNORA65, SNORA11, has-mir-135b*, hasmir-592, hsa-mir-31, has-mir-135b, has-mir-344

mRNAs: CTBP2, STK19

Unicystic ameloblastoma

BRAF, SMO*

CNA losses: 16q23.2*, 21q21.3*, Xq25* CNA gains: 7q11.21*, 1q32.3*, 9p21.1*

Peripheral ameloblastoma Calcifying epithelial odontogenic tumor Adenomatoid odontogenic tumor Primordial odontogenic tumor Cemento-ossifying fibroma

BRAF, NRAS*

19

CDKN2A*, PTEN* KRAS

35

Ameloblastic carcinoma

BRAF, TP53, CTNNB1*

18, 22, 34

CNA losses: 6p15*, 7p15.3*

36-38 mRNA: DMP1

mRNAs: IBSP, BGLAP

40

mRNAs: CTNNB1, TCF7, NKD1, WNT5A, HMMR ncRNAs: hsa-miR-181a-5p, hsa-miR-181c-5p, hsa-miR-149-5p, hsa-miR-138-5p, hsa-miR-199a-3p

mRNAs: CTNNBIP1, FRZB, FZD6, RHOU, SFRP4, WNT10A, WNT3A, WNT4 ncRNAs: hsa-miR-95-3p, hsa-miR-141-3p, hsa-miR205-5p, hsa-miR-223-3p, hsa-miR-31-5p, hsa-miR944, hsa-miR-200b-3p, hsamiR-135b-5p, hsa-miR-313p, hsa-miR-223-5p, hsamiR-200c-3 mRNA: EIF3S5

42-45

mRNAs: POLR2J, CDKN2C

GAPDH, HSP70, KRT19

17, 20, 49-53

(continued on next page)

REVIEW ARTICLE Duarte-Andrade et al. 3

CNA gains: 5q13, 5q, 6q

ARTICLE IN PRESS

Metabolomic alterations Increased abundance

OOOO Volume 00, Number 00

Table I. Summary of the main alterations identified by small- to large-scale and “-omic” studies

Clear cell odontogenic carcinoma

Genomic alterations

Transcriptomic alterations

Gene mutation

Chromosomal alterations

Increased expression

BRAF*

Gene rearrangements: EWSR1-ATF1; EWSR1-CREB1 CNA losses: 6*, 9* CNA gains: 19*, 20*, 14q*

mRNAs: ADAM28, FGF9, S100A7, PTCH1, MMP1, 2, 12

Ghost cell odontogenic carcinoma

CNA gains: SHH*, GLI1*, JAG1*, DTX3*, HEY1* Gene fusions: TCF4, PTPRG* CNA losses: 3q13.1, 5p14.3, 7q31.1 CNA gains: 12q13.1

Metabolomic alterations Increased abundance

References

20,55-59

60

mRNAs: Epidermal keratin types 1,13,15,16, transforming growth factor-b3 receptor, differentiationdependent A4 protein, ribosomal proteins L3, L8, L28, L29, L31, L35, S3, S5, S10, S24, ARF-activated phosphatidylcholine-specific phospholipase D1a, zinc finger protein, DNA binding protein FKHL15, PRAD1

78

Odontogenic keratocyst

PTCH1

Calcifying odontogenic cyst Glandular odontogenic cyst Radicular cyst

CTNNB1

19, 70, 71

TP53, PIK3CA

74

mRNAs: LOXL4, TCTA, LARP6, PTCH1, PTCH2, SMO, GLI1y, GLI2

mRNAs: ECM1, ITGA2B, ITGAM, LAMB2, ADAMST13z, SPP1z

AIDA

26, 61-67, 69

75

OOOO && 2020

For detailed information regarding the type of samples used as experimental control, please refer to the text. *Alterations observed in only 1 case. yConflicting results. zExclusive expression.

mRNAs: PTCH1, GLI1, RGS12

ARTICLE IN PRESS

mRNAs: Nuclear factor NF-6, epidermal keratin, type II, MEF2C transcription factor, metalloproteinase, tyrosine phosphatase CIP2, transforming growth factor-b binding protein, Mitogen inducible gene-2, oncofetal antigen 5T4

Primary intraosseous carcinoma

Decreased expression

Proteomic alterations Increased levels

ORAL AND MAXILLOFACIAL PATHOLOGY

Odontogenic cysts and tumors

4 Duarte-Andrade et al.

Table I. Continued

ARTICLE IN PRESS OOOO Volume 00, Number 00

great deal of effort has been made to fully characterize their molecular profile using “-omics” approaches. The following studies reported in this subsection are those involved with the conventional solid-multicystic type of ameloblastoma, unless stated otherwise. According to the latest classification of the World Health Organization published in 2017,1 the nomenclature of this clinical variant of ameloblastoma was simplified to ameloblastoma, this term will be used throughout this review. In 2014, Kurppa et al.12 identified activating BRAFV600E mutation as the most common alteration in ameloblastomas through Sanger sequencing and immunohistochemistry. After this initial description, other studies using small- to medium-scale and “-omics” approaches, such as Sanger sequencing, NGS hotspot panels,13,14 and whole-exome sequencing,15 further identified the prevalence of this alteration in up to ~80% of ameloblastomas. Additionally, a recent study based on whole-exome sequencing detected different allelic frequencies of BRAF-V600E in different parts of a single tumor, suggesting the existence of intratumor heterogeneity.15 However, this finding needs to be further assessed using larger cohorts of tumors and other methodologies for evaluating tumor clonality. Other mutations in genes related to the mitogen-activated protein kinase (MAPK) pathway, such as FGFR2, KRAS, NRAS, and HRAS, were reported in studies that combined genomics and other techniques. These alterations tend to be mutually exclusive with BRAF mutations.13,14,16-19 Mutations in genes not implicated in the MAPK pathway were also identified in ameloblastoma. Of these, Smoothened (SMO) gene mutations, comprising a component of the Hedgehog pathway, were reported to be predominant in the tumors of the maxilla.14,16 However, this finding was not confirmed by other authors.18,20 The occurrence of genomic imbalance in ameloblastoma was evaluated by means of comparative genomic hybridization (CGH), whole-genome microarray analysis and whole-exome sequencing.15,21-23 In a CGH analysis, copy number alterations (CNAs) were found in 2 out of 17 samples, including the loss of chromosome 21 in primary ameloblastoma and multiple chromosomal aberrations in another nonrelated recurrent case, encompassing a gain at 16p, and losses of chromosomes 22, 16q, and 19p.23 In another study that employed CGH, CNAs were identified in only 1 out of 9 ameloblastomas analyzed.21 However, a complementary analysis by fluorescence in situ hybridization detected the loss of chromosome 22 in 89% of these tumors and also less common alterations, such as losses of chromosome 10, 1p36, and 10q26 and gain of chromosome 1.21 Also, the rare CNA loss of 3p24.3 was identified in one ameloblastoma sample using a highdensity whole-genome microarray platform.22 In contrast, based on whole-exome sequencing data, a recent

REVIEW ARTICLE Duarte-Andrade et al. 5

study reported a stable copy number of genes in 10 ameloblastomas and absence of gene fusions.15 However, whole-exome sequencing is not the most appropriate method to detect CNAs and gene fusion.15 The transcriptome of ameloblastoma has been extensively studied by microarrays using different tissues for comparison.24-30 The gene expression profiling compared with normal references has indicated notable dysregulation of the expression of genes related to Wingless/integrated (Wnt) and Sonic Hedgehog (SHH) signaling pathways’ members.30 In relation to granular cell odontogenic tumor and malignant tumors, such as ameloblastic carcinoma and clear cell odontogenic tumor, ameloblastoma presents a similar expression of genes involved in many cellular processes, such as intercellular adhesion, signaling pathways, and cell cycle regulation.24 Ameloblastoma also has differential expression of several other genes, such as CTBP2 and STK19, encoding for a transcriptional corepressor protein and key regulator of NRAS signaling, respectively.24 Ameloblastoma has a distinct expression pattern of genes implicated in cell cycle and growth, metabolism, cell signaling, and signal transduction, compared with the dentigerous cyst.27 Furthermore, a genome-wide microarray comparison between ameloblastoma and odontogenic keratocyst (OKC) found that genes associated with tooth development are enriched in both lesions.26 The enrichment of toothspecific transcription factors in the epithelium of ameloblastoma indicates cell differentiation toward the enamel organ. In contrast, OKC had high expression of genes associated with squamous epithelial differentiation, corroborating with its closer molecular similarity with the oral epithelium than with ameloblastoma.26 In line with these findings, based on the expression of odontogenic tissue defining genes, ameloblastoma was found to present a variable gene expression pattern according to its histological types, with most of the follicular ameloblastomas having an expression profile closer to the one produced by presecretory ameloblasts.28 Also, a comparative cDNA microarray analysis between ameloblastoma and tooth germs revealed significant alteration in the expression of several genes. For instance, FOS, a proto-oncogene involved in the control of cellular proliferation, was identified as the most overexpressed gene in this odontogenic tumor.29 To describe the noncoding RNA signature of ameloblastoma, a high-resolution transcriptome microarray analysis was conducted, detecting the upregulation of 31 noncoding RNAs.25 The overexpression of LINC340, SNORD116-25, SNORA21, SNORA47, and SNORA65, previously associated with other neoplasms, was validated in a different cohort. SNORA11, which has not been implicated in the pathogenesis of other diseases, was also upregulated in ameloblastoma, making it an appealing

ARTICLE IN PRESS ORAL AND MAXILLOFACIAL PATHOLOGY 6 Duarte-Andrade et al.

candidate diagnostic biomarker for this tumor.25 Another study that employed TaqMan low-density arrays to analyze microRNA expression identified the dysregulation of 38 microRNAs in ameloblastoma and 6 in the UA compared with dentigerous cysts.31 Of those, has-mir-135b*, has-mir-592, hsa-mir-31, has-mir-135b, and has-mir-344 presented similar expression between ameloblastoma and UA, distinguishing them from dentigerous cysts. In contrast, the levels of mir-489 discriminate the clinical variants of ameloblastoma and may render a valuable biomarker for differential diagnosis.31 The methylation pattern of 22 apoptosis-related genes was investigated in ameloblastoma by a quantitative polymerase chain reaction methylation array. The promoters of TNFRSF25 and BCL2L11 had diminished methylation in ameloblastoma in relation to dental follicles.32 Although the levels of the transcript of BCL2L11 were increased in ameloblastoma, suggesting that the transcription of this gene is regulated by methylation of its promoter, the biological significance of these findings remains to be elucidated.32 Metabolic alterations in ameloblastoma were investigated using a metabolomic approach by gas chromatography
mass spectrometry.33 The pathways of aminoacyl-tRNA biosynthesis, cyanoamino acid metabolism, and ABC transporters were found to be possibly upregulated in this tumor. Moreover, BRAF-V600E mutation may contribute to dysregulate glycolysis in BRAF-V600E ameloblastomas.33 A small number of studies have described the molecular alterations of UA and peripheral ameloblastoma.18-20,22,34 BRAF-V600E was reported in 60% to 88% of UA samples,18,20,34 suggesting a clinical utility of this mutation in the differential diagnosis of UA from dentigerous and radicular cysts.34 Additionally, SMO (p.L412F) mutation was identified in a BRAF wild type UA case by NGS targeted sequencing.18 A cytogenetic evaluation of 2 UA by high-density wholegenome microarray analysis detected nonrecurrent rare CNAs, involving gains of 7q11.21, 1q32.3, and 9p21.1 and losses of 16q23.2, 21q21.3, and Xq25.22 Regarding peripheral ameloblastoma, BRAF-V600E was detected by NGS gene panel in 2 out of 3 cases, whereas the NRAS Q61R mutation was found in the BRAF wild type tumor.19 Calcifying epithelial odontogenic tumor. Calcifying epithelial odontogenic tumor is a rare benign neoplasm, most often characterized by an intraosseous mass, with nearly half of the cases associated with an unerupted tooth.1,2 In an analysis of 50 cancer-associated genes using a cancer hotspot panel, a frameshift deletion in CDKN2A, which potentially leads to premature termination of the translation of its protein, was noticed in 1 case, and an SNV, which is able to disrupt the

OOOO && 2020

suppressor activity of PTEN, was found in another sample.35 Because these alterations were identified in 1 case each, the molecular signature of this tumor still needs to be elucidated.35 Adenomatoid odontogenic tumor (AOT). Adenomatoid odontogenic tumor (AOT) is a benign, slow-growing, encapsulated tumor, microscopically characterized by nodules of spindle-shaped epithelial cells of various sizes and columnar epithelial cells forming rosette- or ductlike structures.1,2 Until recently, no genetic information was available about AOTs. In 2016, our research group used an NGS approach to sequence 50 oncogenes and tumor suppressor genes, commonly mutated in human tumors and reported recurrent genetic mutations in KRAS codon 12 in 7 out of 9 AOT.36 After this initial observation, we further sequenced a larger cohort of AOT samples and found that KRAS p.G12V and p.G12R mutations were present in a high proportion of AOTs (27/38, 71%).37 Another group also used the NGS approach and detected these mutations in 6 out of 8 cases, and KRAS p.G12D in a single case.38 In addition to hotspot mutations in oncogenes and tumor suppressor genes, we screened CNA and copyneutral loss of heterozygosity in 2 AOT samples by using a high-density whole-genome array platform.36 Two new CNA were detected, one at 6p15 and another deletion at 7p15.3, covering the IGF2BP3. Although this alteration encompasses an intronic region, in silico predictions suggested that this alteration may disrupt the first exon of 4 alternative transcripts. Nevertheless, further studies with a larger number of samples are needed to determine whether such alteration is important for the pathogenesis of AOT.36 Benign mixed epithelial and mesenchymal odontogenic tumors Although odontomas are the most prevalent odontogenic tumors, the molecular mechanisms implicated in their pathogenesis remain to be clarified, as well as those involved in ameloblastic fibroma. Therefore we report in this section the single “-omic” study involving a mixed odontogenic tumor, the primordial odontogenic tumor. Primordial odontogenic tumor. Primordial odontogenic tumor is a recently described odontogenic neoplasm composed of fibrous tissue with variable cellularity and areas resembling the dental papilla, surrounded by columnar to cuboidal epithelium similar to inner enamel epithelium.39 A panel of 151 cancer- and 42 odontogenesis-associated genes were investigated in 3 cases of primordial odontogenic tumor through NGS, but no somatic gene mutation was detected in these samples.40 Transcriptome analysis suggested that the absence of enamel formation found in this tumor is

ARTICLE IN PRESS OOOO Volume 00, Number 00

a consequence of the lack of dentin. The production of dentin, in turn, may be impaired because of altered mRNA expression of DMP1, IBSP, and BGLAP, inhibiting calcification and induction of dentin formation.40 Benign mesenchymal odontogenic tumors Until now, odontogenic fibroma and cementoblastoma remain poorly characterized at the molecular level. Therefore this section will focus on cemento-ossifying fibroma and odontogenic myxoma. Cemento-ossifying fibroma (COF). Cemento-ossifying fibroma (COF) is a benign neoplasm of the jaws composed of hypercellular fibroblastic tissue containing variable amounts of bone or cementum-like tissue.41 A targeted sequencing of an NGS panel comprised of 50 tumor suppressor genes and oncogenes revealed no pathogenic variant.42 Considering mRNA expression, 4 genes, including CTNNB1 and TCF7, were upregulated in COF compared with normal bone, and 8 were downregulated, suggesting Wnt/b-catenin signaling pathway deregulation in COF.42 The expressions of 754 microRNAs were also assessed in COF by qPCR using a fixed-content panel, including validated TaqMan microRNA assays in a high-throughput format and normal bone as control.43 The levels of 16 microRNAs were altered in COF. According to in silico analyses, these microRNAs possibly regulate the expression of EZH2, XIAP, MET, and TGFBR1.43 Nevertheless, further studies are required to determine the role of these alterations on COF development and establish the exact frequency of their occurrence in larger cohorts. A comparative microarray analysis between an immortalized COF cell line found overexpression of 26 genes in relation to a jaw osteoblast cell line, including HMMR.44 Increased levels of HMMR transcripts were also identified in a cohort of 12 COF tissues compared with normal bone tissue, in addition to positive immunostaining for the protein encoded by this gene, RHAMM, in the fibrous regions of the tumors.45 In the COF cell line, RHAMM is involved in the intracellular signaling of hyaluronic acid, leading to Extracellular signal-regulated kinase (ERK) phosphorylation and cell growth,45 but further investigations are essential to verify whether this process also occurs in the tumors in vivo. Odontogenic myxoma. Odontogenic myxoma is a benign tumor with destructive potential, histopathologically characterized by the presence of stellate and spindle-shaped fibroblasts immersed in a myxoid matrix.1 Because cardiac myxomas occur as a part of the Carney complex—a rare syndrome caused by mutations in the PRKAR1A gene that increases the risk of multiple neoplasias—PRKAR1A gene mutations were

REVIEW ARTICLE Duarte-Andrade et al. 7

assessed in 17 odontogenic myxomas, with 2 out of 17 presenting them.46 Besides the low proportion of odontogenic myxoma cases having PRKAR1A mutation, a targeted genetic assessment, using an NGS panel, detected only nonpathogenic variants in this tumor.47 Thus “-omic” and/or large-scale approaches, such as exome sequencing, might contribute to a better understanding of the pathogenesis of this lesion. Malignant odontogenic tumors Although the malignant odontogenic neoplasms are more clinically relevant because of higher morbidity and mortality, comprehensive molecular analysis of these tumors is hampered by their rarity. Therefore most of the reports were based either on single cases or small cohorts of patients. Ameloblastic carcinoma (AC). Ameloblastic carcinoma (AC) is a rare malignant tumor that presents histologic characteristics resembling ameloblastoma, cytologic atypia, and ability to metastasize.48 Similar to ameloblastomas, AC also harbors BRAF-V600E, but in a lower proportion of the cases, ranging between 37% to 40% of these tumors.20,49 Furthermore, in a case of ameloblastic carcinoma, a TP53 mutation in exon 5 was identified exclusively in the malignant area of a secondary type of tumor by Sanger sequencing, indicating that this alteration may be associated with the carcinomatous transformation of ameloblastoma.50 In agreement with this finding, other TP53 variant (p.V218E) was found in another AC case by NGS panel sequencing, which also harbored the p.D32Y CTNNB1 mutation.17 The cytogenetic analysis of an AC sample with spindle-cell morphology by CGH revealed an amplification of 5q13 and gains of 5q and 6q.51 Chromosomal alterations in this tumor were investigated using a highdensity whole-genome microarray cytogenetic microarray, but no rare CNAs were identified in the sample, leading to the hypothesis that other types of alterations are involved in AC pathogenesis.22 The evaluation of gene expression profile of an AC using microarray detected alterations in the expression of several genes implicated in cell cycle regulation, transcription, translation, differentiation, and apoptosis control compared with normal mucosa.52 A 2-dimensional gel electrophoresis-based proteomics analysis of an AC case with 3 ameloblastomas revealed an overall similar proteomic profile.53 Twenty proteins implicated in various biological processes were increased in AC, such as those involved in cellular structure, energetic metabolism, stress response, and signal transduction.53 Because differential expression of proteins may be involved in the malignant transformation of some ameloblastomas, in-depth proteomic analysis in association with functional studies are required to clarify this process.

ARTICLE IN PRESS ORAL AND MAXILLOFACIAL PATHOLOGY 8 Duarte-Andrade et al.

Clear cell odontogenic carcinoma. Clear cell odontogenic carcinoma (CCOC) is a rare malignant odontogenic tumor.54 Recurrent EWSR1 rearrangements have been observed in 83% to 100% of the cases.55,56 Although ATF1 is the most common fusion partner of EWSR1, CREB1 may represent an alternative partner.55-57 Other cytogenetic aberrations were found in a single case of CCOC using CGH.58 Regarding gene mutations, the BRAF-V600E mutation was identified in a sample of CCOC.20 However, further studies using larger cohorts are needed to determine the prevalence of this mutation in this tumor. Additionally, alterations in the expression of genes encoding for proteins with functions on cell cycle regulation, transcription, signal transduction, differentiation, and apoptosis were detected in a microarray analysis comparing this malignant tumor to a pool of healthy gingivae in another case of CCOC.59 A microarray analysis of an immortalized cell line of CCOC, harboring the EWSR1-ATF1 translocation using a cell line of mucoepidermoid carcinoma as control, had differential expression of 449 genes.57 Additionally, similarities in the expression of 138 genes were detected between this cell line and an odontogenic epithelial cell line originating from epithelial rests of Malassez.57 However, the results of this study should be interpreted carefully because immortalized cells in culture might not reflect all biological aspects of the tumors. Primary intraosseous carcinoma. Primary intraosseous carcinoma is a rare malignant odontogenic neoplasm.1 A microarray analysis comparing a primary intraosseous carcinoma and oral squamous cell carcinomas had very similar expression levels of carcinogenesis-related genes; only a small subset of 201 genes were found to be able to discriminate between these lesions.60 However, these results were limited to a single primary intraosseous carcinoma sample and to the evaluation of ~6,800 sequences, emphasizing the importance of novel investigations using larger sample sets and other techniques. Odontogenic cysts Odontogenic keratocyst. OKC is a benign cystic lesion, mainly occurring in the posterior mandible and presenting with a locally aggressive behavior and high recurrence rate. PTCH1 truncating mutations represent the most common alteration reported in sporadic and syndromic cases of OKC.61-64 Cytogenetic alterations in OKC were studied by means of CGH identifying deletions in the 3q13.1, 5p14.3, and 7q31.3 regions.65 Moreover, 1 amplification, encompassing several genes related to cell growth regulation and gene transcription, was detected on 12q13.2.65 The cDNA expression arrays in sporadic OKCs exhibited differences in the expression of genes related to tooth germ.65 According to hierarchical cluster

OOOO && 2020

analysis, an overall similar expression profile among OKC samples was identified; no separation trend according to the location of the lesion or type of manifestation (primary or recurrent) was detected.65 Another elegant study by the same investigators used a whole-genome microarray analysis and increased levels of PTCH1, SMO, GLI1, and GLI2 transcripts in OKC compared with ameloblastoma.26 In contrast, a comparative genome-wide expression analysis between sporadic OKCs and distinct odontogenic tissues (“dentome”) reported the existence of different molecular subtypes of this lesion, according to the expression of 60 genes involved in odontogenesis.66 Most of the samples were found to be more closely related to secretory ameloblasts. Although OKC subtypes had differences in several biological processes, including activation of AKT and MAPK pathways, common alterations compared with the reference were also found in the 2 subtypes. Notably, PTCH1 and GLI1 had diminished levels in both clusters, even though the SHH pathway was not differentially regulated between the odontogenic cyst and “dentome.”66 The conflicting results of these reports might be associated with the use of different types of samples as control, besides sample processing and data analysis. To specifically address the role of molecular alterations in stroma on the development of OKC, an RNAseq analysis was performed comparing primary stromal fibroblasts derived from 2 OKCs and 2 healthy gingivae.67 The expression of several genes was altered in OKC: LOXL4, TCTA, and LARP6 were upregulated, whereas RGS12 was downregulated. Functional analyses found that lysyl oxidase-like 4, the secreted protein encoded by LOXL4, is also increased in OKC stromal fibroblasts in relation to fibroblasts of dentigerous cysts and gingivae and might regulate the angiogenesis of OKC.67 Increased methylation of the promoters of LTBR and BCLAF1 genes was identified in OKC cases after marsupialization in an evaluation of a methylation array of 22 apoptosis-related genes. Despite this result, the levels of the transcripts of these genes were similar to the nonmarsupialized lesions. Consequently, the implication of this altered methylation remains unclear.68 A proteomic analysis based on mass spectrometry techniques was performed on OKC lesions and patientmatched normal oral mucosa and radicular cysts. The levels of several proteins were dysregulated in OKC compared with normal tissues and radicular cysts. Interestingly, AIDA, a protein known to play an important role in the regulation of JNK pathway, was overexpressed in OKC compared with radicular cysts and may provide a useful biomarker for differential diagnosis.69 Calcifying odontogenic cyst. Calcifying odontogenic cyst (COC) is an odontogenic cyst lined by

ARTICLE IN PRESS OOOO Volume 00, Number 00

ameloblastoma-like epithelium, with the presence of focal ghost cells.1 Sekine and co-workers identified activating CTNNB1 mutations as frequent genetic alterations in COCs by Sanger sequencing.70 In agreement with this observation, mutations in this gene (p.Ser33Phe, p.Ser37Phe, p.Ser33Cys, and p.Asp32Gly) were detected in nearly 66% to 91% of COC cases using targeted NGS panels.19,71 Ser33 and Ser37 are phosphorylation sites, related to b-catenin ubiquitination and subsequent proteasomal degradation, whereas Asp32 is situated in a degron. These mutations help to control the activity of WNT/b-catenin signaling pathway.19,71 Interestingly, the transfection of HEK 293 cells with plasmids encoding CTNNB1 harboring alterations in the phosphorylation sites upregulates keratin expression, suggesting that these mutations are involved in ghost cell formation, one core feature of COC.19 Glandular odontogenic cyst. Glandular odontogenic cyst is a rare developmental cyst that exhibits epithelial aspects that mimic glandular differentiation and commonly affect the mandible.1 In contrast to intraosseous mucoepidermoid carcinoma, glandular odontogenic cysts lack MAML2 rearrangements.72,73 A study using an NGS hotspot panel to assess mutations in 50 cancer genes found 3 SNVs predicted as damaging or probably damaging, including the TP53 variant (p.Leu289Phe) and variants in PIK3CA (p.Glu689Lys and p.Ala708Phe).74 However, these alterations could not be validated by orthogonal methods because of their low variant allelic fraction (<5%); their nucleotide changes are consistent with sequencing artifacts induced by formalin fixation.74 Radicular cyst. Radicular cyst is a cystic lesion of inflammatory origin and represents the most common cyst of the jaws. A cDNA microarray analysis focused on identifying differences in the expression of extracellular matrix molecules among radicular cyst, periapical granuloma, and healthy periodontal ligament revealed that ADAMTS13 and SPP1 are exclusively expressed in radicular cysts.75 Additionally, an increased expression of ECM1, ITGA2B, ITGAM, and LAMB2 was detected in this cyst in relation to periapical granulomas and healthy periodontal ligament. Because matrix metalloproteinases activity is controlled post-transcriptionally, functional analyses were performed. A higher activity of metalloproteinases MMP-9 and MMP-13 was detected in polymorphonuclear cells in periapical granulomas than in radicular cyst.75 An evaluation of methylation arrays targeting the promoters of 22 immune response-related genes revealed that FOXP3 promoter is highly methylated in the radicular cyst.76 The methylation pattern of this gene is inversely correlated to the expression of its transcript. The methylation of FOXP3 may function as

REVIEW ARTICLE Duarte-Andrade et al. 9

a switch controlling the mRNA levels and activity of T-regulatory cells, which balance lesion progression and latency.76

DISCUSSION The challenges of integrating the “-omic” approaches The development of complex human diseases, including neoplastic disease, involves a combination of environmental and genetic factors and their interplay.4,77 However, each “-omic” is limited to the detection of a particular molecular alteration associated with the disease. For instance, genomics can help in the identification of the underlying genetic alterations in a disorder but is unable to fully describe the intricacy of functional consequences and interactions originating from these alterations. Consequently, the integration of different types of “-omics” can provide an overview of the molecular mechanisms associated with the etiology of diseases (Figure 1).4,77 Because odontogenic lesions are rare, these diseases remain poorly understood and need better characterization. Establishing a parallel between the stages of dental development and the origin of the odontogenic lesion is both challenging and necessary in order to clarify the pathogenesis of these diseases. The integrative study of the molecular characteristics using multiomics approaches can help to clarify the molecular alterations in normal odontogenesis that lead to odontogenic lesions development and progression. To date, there have been few “-omics” or large-scale analyses performed on odontogenic lesions. To our knowledge, the only integrative study involving an odontogenic lesion was performed on a single recurrent case of ghost cell odontogenic carcinoma,78 a rare malignant odontogenic tumor characterized mainly by ghost-cell aberrant keratinization.1 The integration of whole-genome sequencing and transcriptomics data identified multiple alterations in SHH and NOTCH signaling pathways, including copy number gains of SHH, GLI1, JAG1, DTX3, and HEY1 concomitantly with their overexpression.78 A novel fusion between TCF4 and PTPRG was also detected in this tumor. An in silico prediction found that this fusion impairs all functional domains of the receptor-type tyrosine-protein phosphatase g, resulting in loss of tumor suppression activity of this protein. Furthermore, a frameshift mutation was identified in the APC coding gene with reduced levels of its transcript.78 The application and integration of “-omic” techniques in the study of odontogenic lesions is far below a desirable level, contrary to the research of other diseases, in which these approaches have been extensively used, providing plenty of information about various aspects of those disorders. This may, in part, be explained by the

ARTICLE IN PRESS ORAL AND MAXILLOFACIAL PATHOLOGY 10 Duarte-Andrade et al.

relatively low frequency of odontogenic cysts and lesions (except for radicular cysts) and also by the fact that most are benign lesions. However, the understanding of the pathogenesis of such lesions may help in the understanding of malignant transformation. For example, ameloblastic carcinoma may arise in preexisting benign ameloblastoma.50 It is reasonable to speculate that some benign tumors might represent one step of malignant tumors formation, possibly involving early complex molecular events. However, further studies are still needed to evaluate this hypothesis. One strategy to overcome the difficulties in working with these diseases is the use of formalin-fixed and paraffin-embedded tissues, given that these samples are widely available, because they are routinely used in the diagnosis and are often banked in hospitals and laboratory archives worldwide.79 This is especially relevant for some rare tumor types, for which frozen samples are rarely obtained. Even when working with common tumors, a more reasonable number of samples can be easily retrieved. Optimized methods for analyzing DNA, RNA, protein, and metabolites in formalin-fixed and paraffin-embedded tissues have been developed during recent decades and successfully applied in studies of human diseases using “-omic” technologies.79 Additionally, organotypic cultures have been considered as an attractive alternative strategy to investigate the molecular alterations in other tumor types because of their ability to mimic the neoplastic microenvironment in an ex vivo environment. Recently, the first organotypic cultures of odontogenic tumors were described for COF and odontogenic myxoma.80 This culture system produced, in vitro, good reproducibility of the growth pattern of these tumors.80 Organotypic cultures represent a promising tool to study the biological aspects of odontogenic tumors. However, a better understanding of the molecular background of these tumors is still necessary to verify whether these characteristics are reproduced in vitro, allowing a more rational use of the in vitro models.

CONCLUSIONS In summary, a comprehensive understanding of the molecular aspects of odontogenic lesions by using large-scale approaches will enable clinicians, pathologists, and researchers to refine the classification of this heterogeneous group of disorders and provide more accurate biomarkers for precise diagnosis, prognosis, and development of molecular tools in the management of patients with these conditions. This knowledge could even help in the identification of novel targets for personalized therapeutic approaches for patients with more aggressive and/or recurrent tumors.

OOOO && 2020

DISCLOSURE No conflict of interest is declared by the authors. FUNDING The following Brazilian agencies provided funding for this study: Coordination for the Improvement of Higher Education Personnel (CAPES), Finance code 001, National Council for Scientific and Technological Development (CNPq) and Research Support Foundation of the State of Minas Gerais (FAPEMIG). FFDA and TSFP receives CAPES scholarship and JGV receives CNPq scholarship. CCG and RSG are research fellows at CNPq. SUPPLEMENTARY MATERIALS Supplementary material associated with this article can be found in the online version at doi:10.1016/j. oooo.2019.12.017. REFERENCES 1. El-Naggar AK, Chan John KC, Grandis JR, Takata T, Slootweg PJ. World Health Organization Classification of Head and Neck Tumours. 4th ed. Lyon, France: World Health Organization; 2017. 2. Wright JM, Soluk-Tekkesin M. Odontogenic tumors: where are we in 2017? J Istanbul Univ Fac Dent. 2017;51(3 Suppl 1):S10-S30. 3. Zhang J, Chiodini R, Badr A, Zhang G. The impact of next-generation sequencing on genomics. J Genet Genomics. 2011;38:95-109. 4. Hasin Y, Seldin M, Lusis A. Multi-omics approaches to disease. Genome Biol. 2017;18:83. 5. Lappalainen T, Scott AJ, Brandt M, Hall IM. Genomic analysis in the age of human genome sequencing. Cell. 2019;177:70-84. 6. Berger MF, Mardis ER. The emerging clinical relevance of genomics in cancer medicine. Nat Rev Clin Oncol. 2018;15:353-365. 7. Stricker SH, K€oferle A, Beck S. From profiles to function in epigenomics. Nat Rev Genet. 2017;18:51-66. 8. Callinan PA, Feinberg AP. The emerging science of epigenomics. Hum Mol Genet. 2006;15(1):R95-101. Spec No. 9. Wang Z, Gerstein M, Snyder M. RNA-Seq: a revolutionary tool for transcriptomics. Nat Rev Genet. 2009;10:57-63. 10. Lowe R, Shirley N, Bleackley M, Dolan S, Shafee T. Transcriptomics technologies. PLoS Comput Biol. 2017;13:E1005457. 11. Beger RD, Dunn W, Schmidt MA, et al. Metabolomics enables precision medicine: “a white paper, community perspective.” Metabolomics. 2016;12:149. 12. Kurppa KJ, Cat
on J, Morgan PR, et al. High frequency of BRAF V600E mutations in ameloblastoma. J Pathol. 2014;232:492-498. 13. Brown NA, Rolland D, Mchugh JB, et al. Activating FGFR2 RAS - BRAF mutations in ameloblastoma. Clin Cancer Res. 2014;20:5517-5526. 14. Sweeney RT, McClary AC, et al. Identification of recurrent SMO and BRAF mutations in ameloblastomas. Nat Genet. 2014;46:722-725. 15. Guan P, Wong SF, Lim JQ, et al. Mutational signatures in mandibular ameloblastoma correlate with smoking. J Dent Res. 2019;98:652-658. 16. G€ultekin SE, Aziz R, Heydt C, et al. The landscape of genetic alterations in ameloblastomas relates to clinical features. Virchows Arch. 2018;472:807-814. 17. Bartels S, Adisa A, Aladelusi T, et al. Molecular defects in BRAF wild-type ameloblastomas and craniopharyngiomas— differences in mutation profiles in epithelial-derived oropharyngeal neoplasms. Virchows Arch. 2018;472:1055-1059.

ARTICLE IN PRESS OOOO Volume 00, Number 00 18. Heikinheimo K, Huhtala J, Thiel A, et al. The mutational profile of unicystic ameloblastoma. J Dent Res. 2019;98:54-60. 19. Yukimori A, Oikawa Y, Morita KI, et al. Genetic basis of calcifying cystic odontogenic tumors. PLoS One. 2017;12:e0180224. 20. Diniz MG, Gomes CC, Guimar~aes BVA, et al. Assessment of BRAFV600E and SMOF412 E mutations in epithelial odontogenic tumours. Tumour Biol. 2015;36:5649-5653. 21. Toida M, Bal
azs M, Tr
eszl A, et al. Analysis of ameloblastomas by comparative genomic hybridization and fluorescence in situ hybridization. Cancer Genet Cytogenet. 2005;159:99-104. 22. Diniz MG, Duarte AP, Villacis RA, et al. Rare copy number alterations and copy-neutral loss of heterozygosity revealed in ameloblastomas by high-density whole-genome microarray analysis. J Oral Pathol Med. 2017;46:371-376. 23. J€a€askel€ainen K, Leivo I, Heikinheimo K, Knuutila S, Jee KJ, Saloniemi I. Cell proliferation and chromosomal changes in human ameloblastoma. Cancer Genet Cytogenet. 2002;136:31-37. 24. Carinci F, Francioso F, Piatelli A, et al. Genetic expression profiling of six odontogenic tumors. J Dent Res. 2003;82:551-557. 25. Davanian H, Balasiddaiah A, Heymann R, et al. Ameloblastoma RNA profiling uncovers a distinct non-coding RNA signature. Oncotarget. 2017;8:4530-4542. 26. Heikinheimo K, Kurppa KJ, Laiho A, et al. Early dental epithelial transcription factors distinguish ameloblastoma from keratocystic odontogenic tumor. J Dent Res. 2015;94:101-111. 27. Lim J, Ahn H, Min S, Hong SD, Lee JI, Hong SP. Oligonucleotide microarray analysis of ameloblastoma compared with dentigerous cyst. J Oral Pathol Med. 2006;35:278-285. 28. Hu S, Parker J, Divaris K, Padilla R, Murrah V, Wright JT. Ameloblastoma phenotypes reflected in distinct transcriptome profiles. Sci Rep. 2016;5(6):30867. 29. Heikinheimo K, Jee KJ, Niini T, et al. Gene expression profiling of ameloblastoma and human tooth germ by means of a cDNA microarray. J Dent Res. 2002;81:525-530. 30. Devilliers P, Suggs C, Simmons D, Murrah V, Wright JT. Microgenomics of ameloblastoma. J Dent Res. 2011;90:463-469. 31. Seti
en-Olarra A, Marichalar-Mendia X, Bediaga NG, AguirreEchebarria P, Aguirre-Urizar JM, Mosqueda-Taylor A. MicroRNAs expression profile in solid and unicystic ameloblastomas. PLoS One. 2017;12:e0186841. 32. Costa SFS, Pereira NB, Pereira KMA, et al. DNA methylation pattern of apoptosis-related genes in ameloblastoma. Oral Dis. 2017;23:779-783. 33. Duarte-Andrade FF, Silva AMB, Vit
orio JG, et al. The importance of BRAF-V600E mutation to ameloblastoma metabolism. J Oral Pathol Med. 2019;48:307-314. 34. Pereira NB, Pereira KMA, Coura BP, et al. BRAFV600E mutation in the diagnosis of unicystic ameloblastoma. J Oral Pathol Med. 2016;45:780-785. 35. De Sousa SF, Diniz MG, Fran¸c a JA, et al. Cancer genes mutation profiling in calcifying epithelial odontogenic tumour. J Clin Pathol. 2018;71:279-283. 36. Gomes CC, de Sousa SF, de Menezes GHF, et al. Recurrent KRAS G12V pathogenic mutation in adenomatoid odontogenic tumours. Oral Oncol. 2016;56:e3-e5. 37. Coura BP, Bernardes VF, de Sousa SF, et al. KRAS mutations drive adenomatoid odontogenic tumor and are independent of clinicopathological features. Mod Pathol. 2019;32:799-806. 38. Bologna-Molina R, Ogawa I, Mosqueda-Taylor A, et al. Detection of MAPK/ERK pathway proteins and KRAS mutations in adenomatoid odontogenic tumors. Oral Dis. 2019;25:481-487. 39. Mosqueda-Taylor A, Pires FR, Aguirre-Ur
ızar JM, et al. Primordial odontogenic tumour: clinicopathological analysis of six cases of a previously undescribed entity. Histopathology. 2014;65:606-612.

REVIEW ARTICLE Duarte-Andrade et al. 11 40. Mikami T, Bologna-Molina R, Mosqueda-Taylor A, et al. Pathogenesis of primordial odontogenic tumour based on tumourigenesis and odontogenesis. Oral Dis. 2018;24:1226-1234. 41. Eversole LR, Leider AS, Nelson K. Ossifying fibroma: a clinicopathologic study of sixty-four cases. Oral Surg Oral Med Oral Pathol. 1985;60:505-511. 42. Pereira TSF, Diniz MG, Fran¸c a JA, et al. The Wnt/b-catenin pathway is deregulated in cemento-ossifying fibromas. Oral Surg Oral Med Oral Pathol Oral Radiol. 2018;125:172-178. 43. Pereira TSF, Brito JAR, Guimar~aes ALS, et al. MicroRNA profiling reveals dysregulated microRNAs and their target gene regulatory networks in cemento-ossifying fibroma. J Oral Pathol Med. 2018;47:78-85. 44. Hatano H, Shigeishi H, Kudo Y, et al. RHAMM/ERK interaction induces proliferative activities of cementifying fibroma cells through a mechanism based on the CD44-EGFR. Lab Investig. 2011;91:379-391. 45. Hatano H, Ogawa I, Shigeishi H, Kudo Y, Ohta K, Higashikawa K, et al. Expression of receptor for hyaluronan-mediated motility (RHAMM) in ossifying fibromas. Histol Histopathol. 2013; 28:473-480. 46. Perdig~ao PF, Stergiopoulos SG, De Marco L, Matyakhina L, Boikos SA, Gomez RS, et al. Molecular and immunohistochemical investigation of protein kinase a regulatory subunit type 1 A (PRKAR1A) in odontogenic myxomas. Genes, Chromosom Cancer. 2005;44:204-211. 47. Santos JN, Sousa Neto ES, Fran¸c a JA, et al. Next-generation sequencing of oncogenes and tumor suppressor genes in odontogenic myxomas. J Oral Pathol Med. 2017;46:1036-1039. 48. Loyola AM, Cardoso SV, de Faria PR, et al. Ameloblastic carcinoma: a Brazilian collaborative study of 17 cases. Histopathology. 2016;69:687-701. 49. Brunner P, Bihl M, Jundt G, Baumhoer D, Hoeller S. BRAF p. V600E mutations are not unique to ameloblastoma and are shared by other odontogenic tumors with ameloblastic morphology. Oral Oncol. 2015;51:e77-e78. 50. Nobusawa A, Sano T, Yokoo S, Oyama T. Ameloblastic carcinoma developing in preexisting ameloblastoma with a mutation of the p53 gene: a case report. Oral Surg Oral Med Oral Pathol Oral Radiol. 2014;118:e146-e150. 51. Kawauchi S, Hayatsu Y, Takahashi M, et al. Spindle-cell ameloblastic carcinoma: a case report with immunohistochemical, ultrastructural, and comparative genomic hybridization analyses. Oncol Rep. 2003;10:31-34. 52. Carinci F, Palmieri A, Delaiti G, et al. Expression profiling of ameloblastic carcinoma. J Craniofac Surg. 2004;15:264-269. 53. Garc
ıa-Mu~noz A, Bologna-Molina R, Adape-Barrios B, Lic
eaga-Escalera C, Montoya-P
erez LA, Rodr
ıguez MA. Identification of proteins with increased levels in ameloblastic carcinoma. J Oral Maxillofac Surg. 2014;72:1183-1196. 54. Guastaldi FPS, Faquin WC, Gootkind F, et al. Clear cell odontogenic carcinoma: a rare jaw tumor. a review of 107 cases. Int J Oral Maxillofac Surg. 2019;48:1405-1410. 55. Bilodeau EA, Weinreb I, Antonescu CR, et al. Clear cell odontogenic carcinomas show EWSR1 rearrangements: a novel finding and a biological link to salivary clear cell carcinomas. Am J Surg Pathol. 2013;37:1001-1005. 56. Vogels R, Baumhoer D, van Gorp J, Eijkelenboom A, et al. Clear cell odontogenic carcinoma: occurrence of EWSR1-CREB1 as alternative fusion gene to EWSR1-ATF1. Head Neck Pathol. 2019;13:225-230. 57. Kujiraoka S, Tsunematsu T, Sato Y, et al. Establishment and characterization of a clear cell odontogenic carcinoma cell line with EWSR1-ATF1 fusion gene. Oral Oncol. 2017;69:46-55.

ARTICLE IN PRESS ORAL AND MAXILLOFACIAL PATHOLOGY 12 Duarte-Andrade et al. 58. Brinck U, Gunawan B, Schulten HJ, Pinzon W, Fischer U, F€ uzesi L. Clear-cell odontogenic carcinoma with pulmonary metastases resembling pulmonary meningothelial-like nodules. Virchows Arch. 2001;438:412-417. 59. Carinci F, Volinia S, Rubini C, et al. Genetic profile of clear cell odontogenic carcinoma. J Craniofac Surg. 2003;14:356-362. 60. Alevizos I, Blaeser B, Gallagher G, Ohyama H, Wong DT, Todd R. Odontogenic carcinoma : a functional genomic comparison with oral mucosal squamous cell carcinoma. Oral Oncol. 2002;38:504-507. 61. Barreto DC, Gomez RS, Bale AE, Boson WL, De Marco L. PTCH gene mutations in odontogenic keratocysts. J Dent Res. 2000;79:1418-1422. 62. Qu J, Yu F, Hong Y, et al. Underestimated PTCH1 mutation rate in sporadic keratocystic odontogenic tumors. Oral Oncol. 2015;51:40-45. 63. Sun LS, Li XF, Li TJ. PTCH1 and SMO gene alterations in keratocystic. J Dent Res. 2008;87:575-579. 64. Gu XM, Zhao HS, Sun LS, Li TJ. PTCH mutations in sporadic and gorlin-syndrome-related odontogenic keratocysts. J Dent Res. 2006;85:859-863. 65. Heikinheimo K, Jee KJ, Morgan PR, Nagy B, Knuutila S, Leivo I. Genetic changes in sporadic keratocystic odontogenic tumors (odontogenic keratocysts). J Dent Res. 2007;86:544-549. 66. Hu S, Divaris K, Parker J, Padilla R, Murrah V, Wright JT. Transcriptome variability in keratocystic odontogenic tumor suggests distinct molecular subtypes. Sci Rep. 2016;6:24236. 67. Jiang WP, Sima ZH, Wang HC, et al. Identification of the involvement of LOXL4 in generation of keratocystic odontogenic tumors by RNA-Seq analysis. Int J Oral Sci. 2014;6:31-38. 68. Pereira KMA, Costa SFS, Pereira NB, et al. DNA methylation profiles of 22 apoptosis-related genes in odontogenic keratocysts before and after marsupialization. Oral Surg Oral Med Oral Pathol Oral Radiol. 2017;124:483-489. 69. Ivanisevi
c Malci
c A, Breen L, Josi
c D, et al. Proteomics profiling of keratocystic odontogenic tumours reveals AIDA as novel biomarker candidate. J Oral Pathol Med. 2015;44:367-377. 70. Sekine S, Sato S, Takata T, et al. Beta-catenin mutations are frequent in calcifying odontogenic cysts, but rare in ameloblastomas. Am J Pathol. 2003;163:1707-1712. 71. de Sousa SF, Moreira RG, Gomez RS, Gomes CC. Interrogation of cancer hotspot mutations in 50 tumour suppressor genes and oncogenes in calcifying cystic odontogenic tumour. Oral Oncol. 2016;57:e1-e3. 72. Bishop JA, Yonescu R, Batista D, Warnock GR, Westra WH. Glandular odontogenic cysts (GOCs) lack MAML2 rearrangements: a

OOOO && 2020

73.

74.

75.

76.

77.

78.

79.

80.

finding to discredit the putative nature of GOC as a precursor to central mucoepidermoid carcinoma. Head Neck Pathol. 2014; 8:287-290. Greer RO, Eskendri J, Freedman P, Ahmadian M, MurakamiWalter A, Varella-Garcia M. Assessment of biologically aggressive, recurrent glandular odontogenic cysts for mastermind-like 2 (MAML2) rearrangements: histopathologic and fluorescent in situ hybridization (FISH) findings in 11 cases. J Oral Pathol Med. 2018;47:192-197. Siqueira EC, de Sousa SF, Fran¸c a JA, et al. Targeted nextgeneration sequencing of glandular odontogenic cyst: a preliminary study. Oral Surg Oral Med Oral Pathol Oral Radiol. 2017;124:490-494. Paula-Silva FW, D’Silva NJ, da Silva LAB, Kapila YL. High matrix metalloproteinase activity is a hallmark of periapical granulomas. J Endod. 2009;35:1234-1242. Campos K, Franscisconi CF, Okehie V, et al. FOXP3 DNA methylation levels as a potential biomarker in the development of periapical lesions. J Endod. 2015;41:212-218. Sun YV, Hu YJ. Integrative analysis of multi-omics data for discovery and functional studies of complex human diseases. Adv Genet. 2016;93:147-190. Bose P, Pleasance ED, Jones M, et al. Integrative genomic analysis of ghost cell odontogenic carcinoma. Oral Oncol. 2015;51: e71-e75. Donczo B, Guttman A. Biomedical analysis of formalin-fixed, paraffin-embedded tissue samples: the holy grail for molecular diagnostics. J Pharm Biomed Anal. 2018;155:125-134. Bastos VC, Pereira NB, Diniz MG, et al. Bringing benign ectomesenchymal odontogenic tumours to the lab: an in vitro study using an organotypic culture model. J Oral Pathol Med. 2019;48:174-179.

Reprint requests: Ricardo Santiago Gomez, DDS, PhD Department of Oral Surgery and Pathology School of Dentistry Universidade Federal de Minas Gerais (UFMG) Av. Presidente Ant^onio Carlos, 6627 31270-901 Belo Horizonte Minas Gerais Brazil [email protected]