Reference
[12]
[13]
No. of samples
34 primary tumors
10 primary tumors
Type of samples
Whole-genome sequencing
Whole-exome sequencing
Total number of mutation/tumor
1483.1
NA
No. of mutations in exon
1017
202
No. mutations in exon/tumor
25.5
15.5
No. mutated genes
932
195
No. of mutated genes in common
19
No. of mutated genes found in more than two tumors
20
1
1.2.1 General Features of Genetic Alterations
1.2.1.1 Structural Variation (SV)
It is well known that karyotype of OS exhibits extreme aneuploidy with a large number of aberrant chromosomes [4]. The chromosomal instability of OS cells was also demonstrated by the frequent occurrence of loss of heterozygosity (LOH) [14]. In agreement with these previous data, the WGS by Chen et al. showed a high incidence of SVs in OS [12]. The study identified 10,806 SVs in 34 tumors, of which 377 produced in-frame fusion genes. RNA-sequencing data were available for 64 predicted fusion SVs and among them 15 fusion genes were expressed. Using the previously reported data of other tumors (embryonal rhabdomyosarcoma, acute T cell lymphocytic leukemia, and medulloblastoma), the basal mutation rate and the number of SNVs, SVs, and CNVs of OSs were compared with those of these tumors, and the number of SVs was significantly higher in OS compared with other tumors. LOH was also detected with high frequency; 10 out of 13 tumors showed more than 7 LOH, indicating again the high incidence of SVs in OS. Chromothripsis, literally “chromosome shattering,” is a recently discovered phenomenon by which thousands of clustered chromosomal rearrangements occur by a single mutation event in the localized and confined genomic region [15]. The WGS of OS found four cases showed chromothripsis in the region of chromosomes 6q, 13q, 14q, and 17q in each case.
1.2.1.2 Single-Nucleotide Variation (SNV)
The WGS of 34 OS samples by Chen et al. identified 1483.1 SNV/tumor, and this frequency (1.15 × 10−6) was comparable with the standard mutation rate in the human genome [12]. Among these SNVs, 25.2 mutations/tumor (ranging 5–103) resulted in either missense, nonsense, or splicing mutations, and the total number of genes showing these types of mutations were 932. Among them 20 genes were found in more than two cases, but only four genes (p53, RB1, ATRX, and DLG2) were identified by the statistical analyses as significantly mutated genes, and these genes were described as driver mutations in the next section.
Joseph et al. performed the WES using ten primary samples and three cell lines [13]. The number of mutations in the exon regions resulting in amino acid changes was 195, and surprisingly only one gene (the p53 gene) was mutated in more than two cases. Therefore the p53 gene was the only gene which showed mutations in more than two cases both in the WGS and WES studies. The difference between two studies may be caused by the difference in the sequencing method, but may reflect the heterogeneity of OSs.
1.2.2 Driver Mutations
As described in the previous section, the WGS isolated four genes as the driver mutation of OS, of which two (the RB1 and p53 gene) were previously recognized as the driver mutation because of their involvement of hereditary cases and the result of mouse models, and the remaining two genes (the ATRX and DLG2 gene) were novel candidates for the driver mutation of OS.
1.2.2.1 The p53 Gene
The involvement of p53 in OS was first demonstrated by somatic mutations [7], and LOH on the chromosome 17p [16], and further confirmed by the identification of its mutant as the causative for a familial cancer syndrome, Li-Fraumeni syndrome, which is characterized by a high risk for various cancers including breast, brain, and adrenal grand cancers and osteosarcomas [17, 18]. The p53 protein is responsible for monitoring the integrity of the genome and the control of cell cycle checkpoints after DNA damage [19]. The number of mutations in the p53 gene identified by the WGS was 28/34 (82.5 %), in which 19 were SVs and 9 were SNVs. Although some tumors were free from the p53 gene mutation, mutations of genes directly regulating the p53 such as the MDM2 gene were found in such tumors [20], and therefore it is acceptable to consider that almost all OSs have abnormalities in genes on the p53 pathway.
The driver function of mutant p53 in osteosarcomagenesis was further confirmed by animal models (Table 1.2). p53 knockout mice were fertile and developed a number of tumors including osteosarcoma [21]. Conditional knockout mice using the expression of genes in osteogenic lineages such as the Prx1, Osterix, and Col1A1 genes developed OS with a high frequency, almost 100 % in some cases [22–26]. Although the precise mechanism of how the loss of p53 can induce OS so frequently is not yet known, it might be related to the function of p53 as a guardian of genome [19], because the high incidence of SVs is the hallmark of OS.
Table 1.2
Tumor incidence of genetic engineered mice
Predicting target cells | Cre-driver gene | Target gene | Predicted genotype | Incidence of OS | Reference |
---|---|---|---|---|---|
Mesenchymal stem cells | Prx1 | p53 and RB1 | p53−/− | 61 % | [22] |
p53−/−:RB1−/− | 18 % | ||||
p53 and RB1 | p53−/− | 62 % | [23] | ||
p53−/−:RB1+/− | 92 % | ||||
p53−/−:RB1−/− | 29 % | ||||
Pre-osteoblast | Osterix | p53 and RB1 | p53−/− | 87.8 % | [24] |
RB1−/− | 0 % | ||||
p53+/−:RB1+/− | 50 % | ||||
p53+/−:RB1−/− | 90 % | ||||
p53−/−:RB1+/− | (207 days) | ||||
p53−/−:RB1−/− | (147 days) | ||||
p53 and RB1 | p53−/− | 100 % | [25] | ||
RB1−/− | 0 % | ||||
p53+/−:RB1+/− | 30.0 % | ||||
p53+/−:RB1−/− | 77.8 % | ||||
p53−/−:RB1+/− | (292 days) | ||||
p53−/−:RB1−/− | (127 days) | ||||
Col1A1 (3.6 kb) | p53 | p53−/− | 60 % | [26] | |
Osteoblast | Col1A1 (2.3 kb) | p53 | p53−/− | 85 % | [22] |
Notch | Exogenous NICD | 100 % | [63] | ||
Osteocalcin | Ptc | Ptc+/−:p53+/− | 70 % | [65] |
1.2.2.2 The RB1 Gene
Retinoblastoma (RB) is a malignant tumor that develops in the eyes of infants, and approximately 25 % of patients show bilateral and multiple tumors, which are caused by germline mutations of the RB1 gene [27]. The RB1 gene is a ubiquitously expressed gene, the encoded protein of which regulates the cell cycle through the control of cyclins [28]. The loss-of-function mutations of RB1 induce abnormal cell growth, and therefore this gene is called a tumor suppressor gene [28]. Patients with germline mutations of the RB1 gene have a high risk of developing other malignant tumors during their lifetime, with OSs most frequently encountered [29]. The mutation search in sporadic OS revealed a loss of functional RB protein in approximately 60 % of sporadic OSs, suggesting that RB1 plays a critical role in the development of not only RB but also OS [30]. Thus the RB1 gene is the first gene mutated in OS with a high frequency, although it is not yet clear why the loss of RB protein predisposes the high risk of OS.
The WGS by Chen et al. discovered ten mutations of the RB1 gene among 34 cases, of which seven were SVs and three were SNVs [12]. The frequency of mutation was lower than those of the p53 gene, but the function of RB1 was also inhibited by mutations of RB-associated genes such as the amplification of the CDK4 [31] and cyclin D [31] genes and the functional loss of the p16 gene by promoter methylation [32], and therefore the loss of RB function is also an important driver mutation in OS.
The mice model story of the RB1, however, was not as simple as in the case of the p53 (Table 1.2). Simple knockout mouse of the RB1 gene was embryonic lethal, and the heterozygous mice, which represented hereditary patients of RB, developed pinealomas but not retinoblastomas or osteosarcomas [33]. As same as the p53 gene, several lines of conditional knockout mice have been generated using the expression of genes on the osteogenic lineages. In contrast to the p53 gene, the loss of RB1 in these cells failed to produce OS in most of cases [24, 25]. The effect of loss of RB1 was only manifested when these mice were crossed with p53 knockout mice, in which loss of RB1 accelerated the tumor formation and reduced the survival time [24, 25].
It is not yet known that why the loss of RB1 preferentially induces osteosarcomas, even though the function of RB1 is important in any types of cells. One of the hypothetical explanations is that RB1 protein has some specific roles in osteogenic differentiation. However, conflicting results were reported by in vitro and in vivo studies as for the effect of loss of RB1 on osteogenic differentiation [34–36]. This issue should be further investigated to understand precise role of RB1 mutations in the development of OS.
1.2.2.3 The ATRX and DLG2 Gene
The mutations of the ATRX (ATP-dependent helicase ATRX) gene was found in 10/34 cases (five as SV and five as SNV) [12]. ATRX is involved in ALT (alternative lengthening of telomere) [37], which is the main mechanism for the maintenance of the telomere length in sarcomas [8]. The mutations of DLG2 (disc, large homolog 2) was found in 18/34 cases, all of which were SVs [12]. DLG2 is a member of the membrane-associated guanylate kinase family with multiple PDZ domains and involved in epithelial polarity during cell division [38]. In Drosophila, DLG is a tumor suppressor, but the tumor suppressor function was not yet confirmed in human cancers.
1.2.3 Genes Involved in the Hereditary Predisposition
1.2.3.1 DNA Helicase Genes
DNA helicases consist of family of enzymes catalyzing the separation of double-strand DNA in several cellular processes such as DNA replication and DNA repair [39]. There are several hereditary diseases caused by the mutation of DNA helicase genes including Bloom syndrome (caused by the mutation of the RECQL2 gene) [40], Werner syndrome (RECQL3 or WRN gene) [41], and Rothmund-Thomson syndrome (RECQL4 or RTS gene) [42]. Patients with genetic defects in these genes manifest a number of disorders and are predisposed to cancers including OS [39]. Among these helicase genes, mutations of the RECQL4 gene seem to be most closely linked with the development of OS [43]. The important differences between the RB1 or p53 genes and DNA helicase genes in terms of the involvement of OS are that the former genes were frequently mutated in sporadic cases as somatic mutations, whereas almost no somatic mutations have been reported in the latter genes [44]. Although the hereditary involvements of DNA helicase genes are clearly observed in human cases, no definite observation was found in mice models of these genes. Homozygous Recql4 mutant mice developed variable phenotype depending on the type of mutations, but the incidence of OS development was very low [45]. Therefore, although the extreme aneuploidy of OS suggested the link between the DNA repair systems and tumor development, it is not yet known how the mutations of DNA helicase genes were involved in the development of OS.
1.2.3.2 Single-Nucleotide Polymorphism Associated with OS
The genome-wide association study (GWAS) has been performed to find genetic factors contributing the development of each disease in various fields including sarcomas [46, 47]. In the case of osteosarcoma, two SNPs were found to be associated with the risk for the development of osteosarcoma [48]. One of them is in the GRM4 gene that encodes a metabotropic glutamate receptor, which involves c-AMP signaling cascade. The glutamate signaling is best characterized in the central nervous system, and its role in the bone metabolism is not known, although bone tissues expressed the GRM4 gene [49]. The effect of identified SNP for the regulation of the GRM4 gene is also not known. However, from the standpoint of recent focus in cancer research, the identification of a gene involving metabolic pathway as a risk factor is an interesting matter. The expression of the GRM4 gene is expressed in OS cells [50] and is associated with aggressive phenotypes of several cancers [51, 52]. Functional analyses of the GRM4 gene in OS cells may provide a key to answer this association.
1.2.4 Genes on the Signal Pathways Involved in OS
Studies of the molecular mechanisms of growth and progression in OS have identified more than 20 genetic alterations (Fig. 1.1). Most of them, however, were dysregulation in mRNA or protein level, and few mutations were found in genomic DNA level. The most striking feature was their redundancy in the growth signals. OSs expressed the receptors for IGF, VEGF, HER2, ErbB-4, PTHR, and HGF, and many of them are redundant [53–59]. This imposed the difference in the development of molecular target therapy for OS. One of typical examples was the recent clinical trial of mTOR inhibitor. Several signal pathways on OS are connected to the Akt kinase through the activation of PI3K. Activated Akt then activates mTOR via inhibition of TSC, which then activates S6K and eIF4E, resulting in the activation of invasion-related protein such as VEGF. Activated Akt also inhibits the function of GSK-3, resulting in the nuclear accumulation of β-catenin, which then drives target such as c-myc. These data suggested that Akt is a hub molecular for growth signaling in OS, and the inhibition of mTOR function seemed to be a promising molecular approach to the treatment of OS. The result of clinical trial using novel mTOR inhibitors, however, showed minimum responses possibly due to the activation of other signal pathways [60]. Although this illustrated the difficulty to apply the molecular target therapy for OS, several clinical trials using chemical targeting following signal pathways are currently ongoing.
Fig. 1.1
Genetic alterations in osteosarcoma. Genes with mutations (DNA and/or RNA level) are indicated with red
1.2.4.1 The Notch Pathway
The Notch pathway is one of the evolutionally conserved pathways and manifests various functions in the development and homeostasis [61]. Involvement of the Notch pathway in OS was reported in several studies, in which the Notch signal was upregulated in tumor samples and the inhibition of this pathway suppresses tumor cell activity [62]. Transgenic mice containing the activating domain of Notch (Notch intracellular domain, NICD) driven by the Col1A1 promoter developed human OS-like tumors with complete penetrance, and combination with p53 knockout mice accelerated the tumor formation [63]. Although no definite mutations were found in molecules on the Notch sigal pathway, inhibition of this pathyway will be one of therapeutic targets.
1.2.4.2 The Hedgehog Pathway
Hedgehog signal is known to be involved in the human OS, although no definite genomic alteration on this signal pathway is found [64]. Patched is the repressive receptor of Hedgehog signal, and the loss of patched resulted in the acceleration of Hedgehog signal. Homozygous knockout mice of the Patched gene was lethal, and heterozygous mice develop OS when there were crossed with p53+/- mice [65].
1.2.4.3 The Wnt Pathway
The Wnt signal is one of the critical signal pathway for the development, maintenance, and regeneration of bone tissues [66]. Nuclear accumulation of the beta-catenin protein in OS was reported [67], and the inhibition of WNT signal suppressed the aggressive phenotype of OS cell lines [68].