記住我
Abstract
Background: Uterine carcinosarcomas (UCS) constitute 3–4% of all uterine malignancies and 16% of deaths caused due to uterine neoplasms. Aim: In this study, we aimed to perform DNA-based mutation analysis in 12 genes (KRAS, NRAS, EGFR, C-KIT, BRAF, PDGFRA, ALK, ERBB2, ERBB3, ESR1, RAF1, PIK3CA) to determine the molecular subtypes of UCS using next-generation sequencing (NGS) in patients with aggressive UCS and poor prognosis. We aimed to compare the results of our analysis with clinicopathological data to contribute to the development of targeted therapy approaches related to the molecular changes of UCS. Materials and Methods: In this study, we included 12 cases diagnosed with uterine carcinosarcomas and examined the changes in oncogenes that play a role in UCS pathogenesis. For the analysis of mutation, the clinicopathological data were compared with the variations in the DNA-based gene panel consisting of 12 genes and 1237 variants in the UCS using the NGS method. Results: EGFR mutation was found in 91.7% of the cases, mutation in 41.7%, PDGFRA mutation in 25%, KRAS and PIK3CA mutation in 16.7%, and C-KIT mutation in 8.3% of the cases. Although no statistical significance was found between the detected mutation and clinicopathological data, it was concluded that PDGFRA mutation might be associated with advanced-stage disease development. Conclusion: This study's findings regarding different molecular types of UCS and information on oncogenesis of UCS can provide inferences for targeted therapies in the future by identifying targetable mutations representing early oncogenic events and thereby contribute toward further studies on this subject.
Keywords: Molecular analysis, molecular variations, mutations, next-generation sequencing, uterine carcinosarcoma
How to cite this article: Introduction 
Uterine carcinosarcomas (UCS) constitutes approximately 3–4% of all uterine malignancies.[1],[2] UCS are clinically aggressive malignancies that contain high-grade carcinomatous (epithelial) and sarcomatous (mesenchymal) components in a mixed manner.[3] Although they are encountered rarely, they constitute 16% of deaths due to uterine malignancies.[4],[5] In the epithelial component, endometrioid (most common) or non-endometrioid (serous, clear cell, undifferentiated, or mixed) histological subtypes can be observed, whereas sarcomatous component may show homologous or heterologous elements.[6],[7] Morphologically, intratumoral heterogeneity of these tumors is also dependent on the combination of mesenchymal components comprising heterologous tissues, such as skeletal muscle, cartilage or bone tissue, and carcinomatous elements.[8] Today, the consensus relies that sarcomas develop from carcinoma as a result of epithelial-mesenchymal transition during tumor development, and the majority of carcinomas (CS) are metaplastic carcinomas of monoclonal origin.[1],[8],[9],[10] Molecular heterogeneity associated with biphasic tumor progression has unknown therapeutic effects and may be associated with a more aggressive clinical course of UCS compared to high-grade endometrial carcinomas.[11] Therefore, understanding the mechanisms that trigger CS progression before histological transformation is crucial in developing rational therapeutic strategies that can equally target both tumor components.[12]
The 5-year progression-free survival is 40–75% when the disease is limited to the uterus at the time of diagnosis, whereas this rate drops to 20-35% in the presence of extrauterine spread.[6],[13] Due to the difficulty of treatment and poor prognosis of CS, new strategies have been developed in the last few decades to investigate therapeutic agents targeting dysfunctional molecular pathways. Some of the molecular studies, conducted in the recent years, point out that approximately 50% of the UCS carry mutations in the PI3K/AKT and RAS/RAF pathways, and the PIK3CA mutation most frequently affected by these pathways is seen in approximately 20% of the tumors.[14],[15] On a review published in 2017, metaanalysis of 23 studies were examined. It was emphasized that HER 2 paves the way for new possibilities in the field of targeted therapy of CS.[16] Studies conducted in the last few years reveal the mutational status of UCS, whose primary treatment is surgery. The efficacy of adjuvant therapies on UCS is quite poor and these studies provide an opportunity to define pathways as potential new targets for the treatment of this tumor with poor prognosis.
In many tumors (especially breast, lung, and colon), targeted treatment approaches after mutation analysis and genetic profiling have gained importance in recent years. Therefore, there is a need for comprehensive molecular studies highlighting oncogenes that are targets or modulators of response to specific treatments in cases of recurrent and aggressive UCS refractory to chemotherapy and radiotherapy.
[TAG:2]Materials and Methods[/TAG:2]
Case selection
This study was approved by the Scientific Research Ethics Committee of Trakya University Faculty of Medicine (Approval Number: TÜTF-BAEK 2018/249). Patients who were diagnosed with UCS and underwent total abdominal hysterectomy at Trakya University Faculty of Medicine, Department of Pathology between January 1, 1983 and June 1, 2018 were included in the study. A total of 12 patients who met the criteria were included in the study.
Evaluation criteria
We evaluated the histopathological features of the tumor and determined the carcinomatous (epithelial) and sarcomatous (mesenchymal) components on hematoxylin and eosin (H&E) stained slides. Histological subtyping was performed in both malignant epithelial and mesenchymal components by evaluating histopathological and previously applied immunohistochemical staining findings (epithelial component histological subtyping; endometrioid carcinoma/serous carcinoma/endometrioid-serous-other mixed carcinoma) according to the current World Health Organization (WHO) Endometrial Tumor Classification (5th edition). Histopathological evaluation was performed to assess whether the mesenchymal component contained heterologous elements (presence/absence of heterologous elements) [Figure 1].[1] In addition, the depth of myometrial invasion (exceeding/not exceeding more than ½ of myometrial thickness), lymphatic invasion, blood vessel invasion, and spread to other tissues or organs were evaluated. Age information of the cases was recorded (≤60 years old vs. >60 years old).
Figure 1: Histological features of selected carcinosarcoma cases; chondroid areas as heterologous elements in the sarcoma component (H&E ×40) (a), endometrioid carcinoma in the epithelial component and mesenchymal areas containing heterologous elements side by side (H&E ×100) (b), serous carcinoma foci as epithelial component (H&E ×40) (c), endometrioid carcinoma foci as epithelial component (H&E ×40) (d), rhabdoid differentiation (H&E ×100) (e), Immunohistochemical Desmin positivity of rhabdoid differentiation areas (attached thumbnail, e), Chondrosarcoma area (H&E ×200) (f), Immunohistochemical S100 positivity of chondroid areas (attached thumbnail, f)Molecular examination
Tumor areas containing both carcinomatous and sarcomatous areas adjacent to each other were selected from the H and E stained slides and marked with a pen. From formalin-fixed paraffin embedded blocks, 2–3 sections of 5-micron thickness containing the marked tumor area were taken, and DNA isolation was performed. Mutation analysis was performed on the isolated DNA using QIAGEN (Hilden, Germany, Serial No: G1709015) Gene Reader Next Generation Sequencing device, Gene Read DNA Sequencing Panel PCR V2 kit (Ref. No: 181940) and Gene Read QIAact Panel, Actionable Insights Tumor Panel kit (Ref. No: 181910) in the Trakya University Faculty of Medicine Molecular Pathology Laboratory. Actionable Insights Tumor Panel comprising 12 genes and 1237 variants (KRAS, NRAS, EGFR, C-KIT, BRAF, PDGFRA, ALK, ERBB2, ERBB3, ESR1, RAF1, PIK3CA) were studied.
Molecular evaluation
Molecular analysis was conducted at the data analysis/bioinformatics stage, which is performed automatically by the QIAGEN Clinical Insight (QCI) platform (QCI Analyze and QCI Interpret).[17]
QCI software platform provided information about the mutation status, including the report of variant, alteration, and function of any existing mutation. Furthermore, mutations were classified as 1) pathogenic, 2) likely pathogenic, 3) uncertain significance, 4) likely benign or 5) benign according to the TIER[18] classification, which is a methodology used to provide standardization in data analysis centers.
Statistical analysis
Descriptive statistics (mean ± standard deviation, minimum, maximum, number, and percentage (%)) were calculated for the data. With respect to the presence of mutation, Pearson or Fisher Chi-square tests were used to compare the parameters of age, histological subtype of the epithelial component, presence or absence of a heterologous element in the mesenchymal component, lymphatic vessel invasion, blood vessel invasion, spread to other tissues and organs, and whether the depth of tumor invasion exceeded half of the myometrial thickness. Odds ratio and 95% confidence interval values were calculated in the evaluation of possible risk factors.
A value of P < 0.05 was accepted as statistically significant in all analyses. Statistical analyses were carried out using SPSS 20.0 package program in T.U. Faculty of Medicine, Department of Biostatistics and Medical Informatics (License No: 10240642).
Results Histopathological and clinical findings
The mean age of the patients was 68.25 ± 10.788 years, and the median age was 69.5.
The clinicopathological parameters of the patients (age, histological subtype of epithelial component, presence of heterologous element in mesenchymal component, depth of myometrial invasion, lymphatic and blood vessel invasion, spread to other tissues and organs) are shown in [Table 1].
Molecular evaluation
On mutation analysis of the selected cases, variations were detected in 6 genes. According to the TIER classification, variations including pathogenic, likely pathogenic, and uncertain significance were taken into consideration. Mutations on the 6 genes (ERBB2, EGFR, C-KIT, PIK3CA, PDGFRA, and KRAS) included 11 different variants. At least 1 gene variation was observed in each case. No variation was found in ALK, BRAF, ERBB3, NRAS, RAF 1, and ESR1 genes. Mutations regarded as likely benign or benign according to TIER classification were not included in the statistical analysis.
On performing molecular analysis, gene variation was detected in all cases (100%). Variation in a single gene was observed in 3 (25%) of the cases, whereas synchronous variations including 2 and 3 genes were observed in 6 (50%) and 3 (25%) cases, respectively. Pathogenic PIK3CA mutation was detected in 2 (16.7%) cases, pathogenic mutation and mutation of uncertain significance on KRAS gene was detected in 2 (16.7%), pathogenic CKIT mutation in 1 (8.3%), mutation of uncertain significance on EGFR gene in 11 (91.7%), mutation of uncertain significance on ERBB2 gene in 5 (58.3%), and mutation of uncertain significance on PDGFRA gene was detected in 3 (25.0%) cases [Figure 2].
The distribution of cases according to the detected 6 gene variations, type of detected variants, and distribution of these variants is shown in [Table 2].
Table 2: Variants of the mutations detected and distribution of variants according to the casesComparison of detected mutations and clinicopathological findings in patients with carcinosarcoma
KRAS variation was detected in 2 (16.7%) of 12 cases. One of these mutations was pathogenic (KRAS p.G12D), whereas other mutation (KRAS c. 2505) was in a non-coding region, therefore no amino acid change was observed and the mutation was classified as uncertain significance (VUS). When KRAS mutation status was compared with clinicopathological parameters, no statistical significance was found.
Two different variations were found in the EGFR gene in 11 of the 12 cases (91.7%) on the QCI platform. These were classified as uncertain significance (VUS) according to the TIER classification. The detected variants were coded as follows: p.Q787Q in exon 20 and p.R521K in exon 13. Of the 11 cases with EGFR variation, both variants were present in 5 (45.5%) cases. Although 5 (45.5%) of 11 cases carried only the p.Q787q variant, only 1 (9.0%) case carried the other variant (p.R521K). When the variations in the EGFR gene were compared with clinicopathological parameters, no statistical significance was found.
C-KIT gene variation was detected in only 1 (8.3%) of 12 cases. Two different variants were present in this case. One of the variants was the p.I798I variant in exon 17, whereas the other was the p.M541L variant in exon 10. The p.I798I variant was classified as pathogenic in the TIER 3 classification for CS on the QCI platform, whereas the p.M541L variant was included in the uncertain significance class. When the C-KIT mutation status was compared with clinicopathological parameters, no statistical significance was found.
PDGFRA variation was detected in 3 (25%) of 12 CS cases. The variant was p.G313G in exon 7. In the QCIs platform, this mutation was included in the uncertain significance class. The presence of heterologous elements, lymphatic invasion, and depth of myometrial invasion were compared with the PDGFRA gene variation status. Although the result was not statistically significant, it was found that the presence of heterologous elements in the PDGFRA variant could increase the risk of lymphatic invasion 2-fold (P = 0.182 OR = 2.0 (95% confidence interval: 0.899–4.452), and the risk of myometrial invasion exceeding more than half of the myometrial thickness by 1.5 times [P = 0.509 OR = 1.5 (95% confidence interval: 0.945–2.381)] [Figure 3].
Figure 3: PDGFRA mutation status according to the presence of heterologous element, lymphatic invasion, whether the depth of tumor invasion exceeds half of the myometrial thickness or notERBB2 gene variation was detected in 5 (58.3%) of 12 CS cases. This mutation was detected in exon 17 with the variant code p.I655V. In the QCI platform, this variant was included in the uncertain significance class. No statistical significance was found when the ERBB2 mutation was compared with clinicopathological parameters.
Patogenic PIK3CA mutation on p.G106V and p.M1043V was detected in 2 of 12 cases (16.7%). No statistical significance was found when PIK3CA gene variation was compared with clinicopathological data.
Discussion Uterine carcinosarcomas (UCS) are rare but aggressive malignancies that consist of metaplastic transformation of epithelial elements and comprise a small part of endometrial cancers.[19] Considering the development of therapeutic strategies, it is important to understand the gene mutations underlying the biological behavior of these tumors, which are quite challenging to treat, and the mechanisms that trigger biphasic transformations. Recently, numerous extensive sequencing studies have demonstrated mutations or abnormal activation of multiple genes/pathways in CS, including genes involved in cell cycle regulation such as HER2, PI3K/AKT/mTOR, EGFR, MAPK, and reported the genetic makeup of CS identifying potential therapeutic targets.[20] Some of these studies have emphasized on the overexpression of HER2 in UCS.[21],[22] Genetic mutations and abnormal activation of PI3K/AKT/mTOR pathways have been reported by several studies, including sequencing studies.[9],[15],[19],[23] There are also studies reporting recurrent KRAS mutations and suggesting possible targeted therapies for the MAPK pathway for these mutations.[9],[19],[23]
In this study, 12 different genes (EGFR, ERBB2, KRAS, PIK3CA, C-KIT, PDGFRA, NRAS, RAF1, ESR1, ERBB3, BRAF, and ALK) were evaluated in UCS using the next-generation sequencing (NGS) method. Variations were detected in 6 of these 12 evaluated genes (EGFR, ERBB2, KRAS, PIK3CA, C-KIT, and PDGFRA). They were compared with clinicopathological parameters of the patients such as age, histological subtype of the epithelial component, heterologous element in the mesenchymal component, lymphatic and blood vessel invasion, and spread to other tissues/organs. Molecular analysis revealed gene mutations in all cases. Mutations were most frequent in the EGFR gene (91.7%), followed by the ERBB2 (HER 2) gene (41.7%). Mutations detected in both genes were included in the uncertain significance class according to the TIER classification. No definite pathogenic variant was defined for the mutations seen in ERBB2 and EGFR genes. Although the affected gene localizations are similar to those seen in previous studies, the possible reasons for not identifying a pathogenic variant may be the small size of our case series and the limited number of studies in the literature. Therefore, further studies involving large case series should be conducted so that the importance of variations in these genes can be better demonstrated with accumulation of data on this subject.
There are studies predicting that structural and/or numerical changes of chromosomes may be partially involved in evaluating the overexpression of tyrosine kinase receptors including EGFR and/or HER-2.[21],[22] Kanthan et al.[12] (2000–2011) conducted a review of studies on CS and emphasized that the overexpression of tyrosine kinase receptors, such as HER-2 and EGFR, was prominent among the genetic changes seen in UCS, and these tyrosine kinase receptors can reveal potential targets for therapeutic applications in UCS.[21],[22],[24],[25] Vitale et al.[16] reviewed the studies on CS between 2001 and 2016, and concluded that HER-2 gene was one of the most prominent targets in CS and created new possibilities in the field of targeted therapy. In this study, 4 out of 5 cases with ERBB2 variation also had EGFR variation. The frequent coexistence of EGFR and ERBB2 gene mutations in this study was similar to the literature and was considered as one of the features that should be considered when planning targeted treatment in these tumors.
In the study of Biscuola et al.,[14] widespread EGFR expression was found in most cases of UCS, and exon 15 mutation was detected in 2 cases with NGS method. In this study, EGFR variations were detected at exon 13 and exon 20. No information was found in the literature regarding CS cases with variations in these exons.
Amplification of HER2 in UCS has been reported at rates ranging from 25 to 56%.[11],[21],[22],[24] No clinical trials have yet been reported in patients with CS, but preclinical data and case report in a uterine serous carcinoma patient with HER2 overexpression and resistance to chemotherapy support the use of trastuzumab in cases of HER2-overexpressing CS.[20],[26] There are other studies evaluating the treatment efficiency of tyrosine kinase inhibitors in CS.[27],[28] Based on the results of these studies, it was concluded that HER-2 inhibitors constitute a good potential for treatment in cases of HER-2 overexpressing/amplified/mutant CS. The presence of variation in the ERBB2 exon 17 in 5 cases (41.7%) we examined also support that ERBB2 mutations may be involved in the carcinogenesis of CS. Molecular analysis studies including larger case series are needed to determine whether this variation is pathogenic for CS and whether tyrosine kinase inhibitors can be a treatment option in cases carrying this mutation.
Phosphatidylinositol 3-kinase (PI3K) pathway is also an important regulator of cellular activity, whose inhibitors represent a promising new therapeutic strategy in uterine cancers; furthermore, it has also been shown that PI3K/mTOR inhibitors may be effective against uterine sarcomas.[29]
Biscuola et al.[14] investigated the oncogene changes in CS, and the PI3K/AKT pathway was detected in 32% of tumors (6 mutations were seen in PIK3CA and 1 mutation in AKT1), and was found to be the pathway that was most frequently subjected to changes.[14] The PIK3CA mutation detected in 17% of the cases was shown in exon 9, exon 20, and exon 1. There are studies reporting different rates of PIK3CA mutations. In the genotyping study conducted by Growdon et al.[15] PIK3CA mutation was reported at a rate of 32%, whereas Murray et al.[30] reported PIK3CA mutations (exon 20) at a rate of 11% in their mutational analysis. Bashir et al.[31] reported PIK3CA mutation in exon 20 and exon 6 at a rate of 15%, whereas McConechy et al.[32] reported PIK3CA mutations (40%) in exon 9 and exon 20 in CS cases. Zhao et al.[19] conducted a mutational analysis study, and emphasized that the PIK3CA gene mutation detected in CS increases the possibility of treatments targeting this signaling pathway. In two recent studies, PIK3CA mutation was detected in 31% and 34% of the UCS cases.[16],[32],[33],[34] PIK3CA variations were detected at a rate of 16.7%, and mutations were present in exon 2 in one case and in exon 21 in one case. Both variants were classified as pathogenic for CS. The possibility that PI3KCA may be one of the oncogenes involved in the tumorigenesis of UCS as shown in endometrial carcinomas and one of the oncogenic loci that can be targeted with PIK3CA/mTOR inhibitors should be supported by studies involving larger case series.
There are limited studies investigating C-KIT expression in CS. In studies where C-KIT expression was evaluated immunohistochemically in UCS, conflicting and different rates of expression were reported.[22],[24],[35],[36],[37] The C-KIT mutation was reported for the first time by Biscuola et al.[14] and was detected only in one case. In this study, the C-KIT variation was detected as 2 variants in only one patient (8.3%). One of these variants (exon 17) was classified as pathogenic. Further studies on C-KIT expression/mutation in larger CS series are needed to resolve controversial issues, such as C-KIT expression frequency and prognostic and clinical value in UCS.
Wada et al.[38] conducted one of the earliest studies contributing to the elucidation of UCS pathogenesis, and found K-RAS gene mutation at a rate of 24% using PCR analysis, suggesting that the mutation of K-RAS gene probably emerged as an early event before biphasic differentiation of tumors. In the genotyping study conducted by Growdon et al.,[15] all of the KRAS mutations were detected in UCS cases, which constituted 50% of all cases, and in another study in which the NGS method was used, KRAS mutation was observed in 3 cases.[32] In two recent studies, KRAS mutations were found in 6% and 16% of cases.[33],[34] In this study, KRAS gene mutation was detected in 2 cases (16.7%). Although one case was classified as pathogenic with exon 2 mutation, the other case was classified as uncertain significance with the C.2505 variant (non-coding region). These rates obtained were in accordance with the mutation rates found in the studies mentioned above. When KRAS mutation was compared with clinicopathological parameters, all of the cases carrying this mutation were found to have the mixed carcinoma subtype (including serous/endometrioid/other carcinomas) of the malignant epithelial component. Although statistical significance was not detected between KRAS mutation and subtype of epithelial component (P = 0.091), it was thought that more significant results could be obtained with studies including large case series.
In the study conducted by Adams et al.,[39] strong immunoexpression for PDGFRA was detected in 66% of CS cases. Biscuola et al.[14] detected PDGFRA gene mutation (exon 18) in three CS cases, and it was reported that PDGFRA mutation was detected for the first time in the literature in this study. In this study, PDGFRA gene mutation was detected in 3 out of 12 cases (25%), as a single variant in exon 7. This mutation was included in the uncertain significance category according to the TIER classification. When PDGFRA gene mutation was compared with clinicopathological data, PDGFRA variation was thought to be associated with the presence of heterologous elements and lymphatic invasion, although the results were not statistically significant. It was concluded that further mutational analysis studies with larger case series are needed to determine the pathogenicity of exon 7 variation detected in 3 cases in this study, the possibility of this mutation being a driver mutation, and its potential to be a target gene mutation in targeted therapies.
One of the limitations of our study was the inclusion of a low number of cases owing to the research fund provided to us.
In conclusion, this study can contribute to a better understanding of the oncogenesis of UCS, identification of targetable mutations representing early oncogenic events, revealing different molecular findings in UCS in parallel with molecular subtypes of endometrial carcinoma, as well as the development of targeted treatment options. However, the results should be supported by larger case series and multiple gene panels.
Financial support and sponsorship
This work was supported by Research Fund of the Trakya University. Project Number: 2019/04.
Conflicts of interest
There are no conflicts of interest.
References 
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/ijpm.ijpm_777_21
留言 (0)