Cell line authentication: a necessity for reproducible biomedical research.

Immortalized or continuous cell lines are invaluable tools in basic and preclinical research. However, the widespread use of misidentified cell lines is a serious threat to scientific reproducibility. Based on the experiences of mandatory cell line authentication at the International Journal of Cancer (IJC), we provide an overview of the issues pertinent to misidentified cell lines and discuss available solutions. We also summarize the lessons learned, revealing that at least 5% of the human cell lines used in manuscripts considered for peer review are misidentified. About 4% of the considered manuscripts are rejected for severe cell line problems, and most are subsequently published in other journals. In order to diminish such malpractice and its consequences for the scientific record, we postulate that strict multi-layered quality control is essential. Besides journals and publishers, we encourage scientists, research institutions, and funders to take action on the matter and revise their respective policies. Hence, we provide concrete recommendations on introducing regular authentication schemes and staff training, and discuss future steps for enhancing good cell culture practices.

Establishment of primary prostate epithelial and tumorigenic cell lines using a non-viral immortalization approach.

BACKGROUND: Research on prostate cancer is mostly performed using cell lines derived from metastatic disease, not reflecting stages of tumor initiation or early progression. Establishment of cancer cell lines derived from the primary tumor site has not been described so far. By definition, cancer cells are able to be cultured indefinitely, whereas normal epithelial cells undergo senescence in vitro. Epithelial cells can be immortalized, accomplished by using viral integration of immortalization factors. Viral approaches, however, might be impaired by regulatory and safety issues as well as random integration into regulatory genetic elements, modifying precise gene expression. We intend to use surgical specimen of prostate cancer patients to (i) prove for establishment of cancer cell lines, and (ii) perform non-viral, Sleeping Beauty (SB) transposase-based immortalization of prostate epithelial cells. METHODS: Radical prostatectomy samples of prostate cancer patients (n = 4) were dissociated and cultured in vitro. Cells were cultivated either without or after non-viral, Sleeping-Beauty transposase-based stable transfection with immortalization factors SV40LT and hTERT. Established cell lines were analyzed in vitro and in vivo for characteristics of prostate (cancer) cells. RESULTS: Initial cell cultures without genetic manipulation underwent senescence within ≤ 15 passages, demonstrating inability to successfully derive primary prostate cancer cell lines. By using SB transposase-based integration of immortalization factors, we were able to establish primary prostate cell lines. Three out of four cell lines displayed epithelial characteristics, however without expression of prostate (cancer) characteristics, e.g., androgen receptor. In vivo, one cell line exhibited tumorigenic potential, yet characteristics of prostate adenocarcinoma were absent. CONCLUSION: Whereas no primary prostate cancer cell line could be established, we provide for the first-time immortalization of primary prostate cells using the SB transposase system, thereby preventing regulatory and molecular issues based on viral immortalization approaches. Although, none of the newly derived cell lines demonstrated prostate cancer characteristics, tumor formation was observed in one cell line. Given the non-prostate adenocarcinoma properties of the tumor, cells have presumably undergone oncogenic transformation rather than prostate cancer differentiation. Still, these cell lines might be used as a tool for research on prostate cancer initiation and early cancer progression.

Comparison of Tumour-Specific Phenotypes in Human Primary and Expandable Pancreatic Cancer Cell Lines.

There is an ongoing need for patient-specific chemotherapy for pancreatic cancer. Tumour cells isolated from human tissues can be used to predict patients' response to chemotherapy. However, the isolation and maintenance of pancreatic cancer cells is challenging because these cells become highly vulnerable after losing the tumour microenvironment. Therefore, we investigated whether the cells retained their original characteristics after lentiviral transfection and expansion. Three human primary pancreatic cancer cell lines were lentivirally transduced to create expandable (Ex) cells which were then compared with primary (Pri) cells. No obvious differences in the morphology or epithelial-mesenchymal transition (EMT) were observed between the primary and expandable cell lines. The two expandable cell lines showed higher proliferation rates in the 2D and 3D models. All three expandable cell lines showed attenuated migratory ability. Differences in gene expression between primary and expandable cell lines were then compared using RNA-Seq data. Potential target drugs were predicted by differentially expressed genes (DEGs), and differentially expressed pathways (DEPs) related to tumour-specific characteristics such as proliferation, migration, EMT, drug resistance, and reactive oxygen species (ROS) were investigated using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. We found that the two expandable cell lines expressed similar chemosensitivity and redox-regulatory capability to gemcitabine and oxaliplatin in the 2D model as compared to their counterparts. In conclusion, we successfully generated expandable primary pancreatic cancer cell lines using lentiviral transduction. These expandable cells not only retain some tumour-specific biological traits of primary cells but also show an ongoing proliferative capacity, thereby yielding sufficient material for drug response assays, which may provide a patient-specific platform for chemotherapy drug screening.

Differential Requirements for the RAD51 Paralogs in Genome Repair and Maintenance in Human Cells.

Deficiency in several of the classical human RAD51 paralogs [RAD51B, RAD51C, RAD51D, XRCC2 and XRCC3] is associated with cancer predisposition and Fanconi anemia. To investigate their functions, isogenic disruption mutants for each were generated in non-transformed MCF10A mammary epithelial cells and in transformed U2OS and HEK293 cells. In U2OS and HEK293 cells, viable ablated clones were readily isolated for each RAD51 paralog; in contrast, with the exception of RAD51B, RAD51 paralogs are cell-essential in MCF10A cells. Underlining their importance for genomic stability, mutant cell lines display variable growth defects, impaired sister chromatid recombination, reduced levels of stable RAD51 nuclear foci, and hyper-sensitivity to mitomycin C and olaparib, with the weakest phenotypes observed in RAD51B-deficient cells. Altogether these observations underscore the contributions of RAD51 paralogs in diverse DNA repair processes, and demonstrate essential differences in different cell types. Finally, this study will provide useful reagents to analyze patient-derived mutations and to investigate mechanisms of chemotherapeutic resistance deployed by cancers.

Ethical Challenges Using Human Tumor Cell Lines in Cancer Research.

Human and animal cell cultures are indispensable model systems for the biomedical research and pharmaceutical industry and already represent one of the most important alternatives to animal experiments. The development of mammalian cell culture started in the first half of the last century when fundamental questions of genetics were unresolved and the pioneers of cell culture did not care about individual personality rights of donors of biomaterials. However, cultivation of primary and continuous cell cultures was and still is usually associated with the use of FBS, which-almost universally applicable-is questionable in terms of extraction and quality variations measurably affecting reproducibility of results. The history of the cell line HeLa is a prime example for the development of biomedical research with its great successes in the fight against cancer and development of Polio Virus vaccinia, but also for limitations in the public and scientific applications of cell lines in the age of digitization and bioinformatics. HeLa leads from the establishment of the first human continuous cell line to initial cancer research using tumor cells, from disastrous cross-contaminations by HeLa cell cultures to legal and ethical controversy by reading out the individual genome and the commercial use that continues to this day.

Small molecular modulators of JMJD1C preferentially inhibit growth of leukemia cells.

Histone demethylases are promising therapeutic targets as they play fundamental roles for survival of Mixed lineage leukemia rearranged acute leukemia (MLLr AL). Here we focused on the catalytic Jumonji domain of histone H3 lysine 9 (H3K9) demethylase JMJD1C to screen for potential small molecular modulators from 149,519 natural products and 33,765 Chinese medicine components via virtual screening. JMJD1C Jumonji domain inhibitor 4 (JDI-4) and JDI-12 that share a common structural backbone were detected within the top 15 compounds. Surface plasmon resonance analysis showed that JDI-4 and JDI-12 bind to JMJD1C and its family homolog KDM3B with modest affinity. In vitro demethylation assays showed that JDI-4 can reverse the H3K9 demethylation conferred by KDM3B. In vivo demethylation assays indicated that JDI-4 and JDI-12 could induce the global increase of H3K9 methylation. Cell proliferation and colony formation assays documented that JDI-4 and JDI-12 kill MLLr AL and other malignant hematopoietic cells, but not leukemia cells resistant to JMJD1C depletion or cord blood cells. Furthermore, JDI-16, among multiple compounds structurally akin to JDI-4/JDI-12, exhibits superior killing activities against malignant hematopoietic cells compared to JDI-4/JDI-12. Mechanistically, JDI-16 not only induces apoptosis but also differentiation of MLLr AL cells. RNA sequencing and quantitative PCR showed that JDI-16 induced gene expression associated with cell metabolism; targeted metabolomics revealed that JDI-16 downregulates lactic acids, NADP+ and other metabolites. Moreover, JDI-16 collaborates with all-trans retinoic acid to repress MLLr AML cells. In summary, we identified bona fide JMJD1C inhibitors that induce preferential death of MLLr AL cells.

Cell Lines as Biological Models: Practical Steps for More Reliable Research.

Research in toxicology relies on in vitro models such as cell lines. These living models are prone to change and may be described in publications with insufficient information or quality control testing. This article sets out recommendations to improve the reliability of cell-based research.

Authentication of M14 melanoma cell line proves misidentification of MDA-MB-435 breast cancer cell line.

A variety of analytical approaches have indicated that melanoma cell line UCLA-SO-M14 (M14) and breast carcinoma cell line MDA-MB-435 originate from a common donor. This indicates that at some point in the past, one of these cell lines became misidentified, meaning that it ceased to correspond to the reported donor and instead became falsely identified (through cross-contamination or other means) as a cell line from a different donor. Initial studies concluded that MDA-MB-435 was the misidentified cell line and M14 was the authentic cell line, although contradictory evidence has been published, resulting in further confusion. To address this question, we obtained early samples of the melanoma cell line (M14), a lymphoblastoid cell line from the same donor (ML14), and donor serum preserved at the originator's institution. M14 samples were cryopreserved in December 1975, before MDA-MB-435 cells were established in culture. Through a series of molecular characterizations, including short tandem repeat (STR) profiling and cytogenetic analysis, we demonstrated that later samples of M14 and MDA-MB-435 correspond to samples of M14 frozen in 1975, to the lymphoblastoid cell line ML14, and to the melanoma donor's STR profile, sex and blood type. This work demonstrates conclusively that M14 is the authentic cell line and MDA-MB-435 is misidentified. With clear provenance information and authentication testing of early samples, it is possible to resolve debates regarding the origins of problematic cell lines that are widely used in cancer research.

False and mycoplasma-contaminated leukemia-lymphoma cell lines: time for a reappraisal.

Leukemia-lymphoma cell lines are important research tools in a variety of fields. To represent adequate model systems it is of utmost importance that cell lines faithfully model the primary tumor material and are not cross-contaminated with unrelated cell material (or contaminated with mycoplasma). As it has been previously reported that cross-contaminated cell lines represent a significant problem, it is of interest to know whether any improvement in the prevalence of such "false cell lines" had occurred since we called the alert in 1999. A retrospective review of our data archives covered 848 cell lines received from 1990 to 2014 from 290 laboratories in 23 countries spanning the spectrum of leukemia-lymphoma entities. Two variables were considered: authenticity and freedom from mycoplasma infection. Regarding provenance, we separately considered primary sources (original investigators having established the cell lines or reference repositories) and secondary sources. The percentages of mycoplasma-contaminated cell lines decreased significantly over the 25-year timespan. Among primary sourced material: mycoplasma-contamination fell from 23% to 0%; among secondary sourced: from 48% to 21%. The corresponding figures for cross-contamination declined from 15% to 6%, while among material obtained from secondary sources prevalence remained remarkably high, throughout the time periods at 14-18%. Taken together, our data indicate that using non-authenticated cell lines from secondary sources carries a risk of about 1:6 for obtaining a false cell line. The use of authentic leukemia-lymphoma cell lines holds important translational value for their model character and the reproducibility of the laboratory data in the clinical arena.

BCL6--regulated by AhR/ARNT and wild-type MEF2B--drives expression of germinal center markers MYBL1 and LMO2.

Genetic heterogeneity is widespread in tumors, but poorly documented in cell lines. According to immunoglobulin hypermutation analysis, the diffuse large B-cell lymphoma cell line U-2932 comprises two subpopulations faithfully representing original tumor subclones. We set out to identify molecular causes underlying subclone-specific expression affecting 221 genes including surface markers and the germinal center oncogenes BCL6 and MYC. Genomic copy number variations explained 58/221 genes differentially expressed in the two U-2932 clones. Subclone-specific expression of the aryl-hydrocarbon receptor (AhR) and the resulting activity of the AhR/ARNT complex underlaid differential regulation of 11 genes including MEF2B. Knock-down and inhibitor experiments confirmed that AhR/ARNT regulates MEF2B, a key transcription factor for BCL6. AhR, MEF2B and BCL6 levels correlated not only in the U-2932 subclones but in the majority of 23 cell lines tested, indicting overexpression of AhR as a novel mechanism behind BCL6 diffuse large B-cell lymphoma. Enforced modulation of BCL6 affected 48/221 signature genes. Although BCL6 is known as a transcriptional repressor, 28 genes were up-regulated, including LMO2 and MYBL1 which, like BCL6, signify germinal center diffuse large B-cell lymphoma. Supporting the notion that BCL6 can induce gene expression, BCL6 and the majority of potential targets were co-regulated in a series of B-cell lines. In conclusion, genomic copy number aberrations, activation of AhR/ARNT, and overexpression of BCL6 are collectively responsible for differential expression of more than 100 genes in subclones of the U-2932 cell line. It is particularly interesting that BCL6 - regulated by AhR/ARNT and wild-type MEF2B - may drive expression of germinal center markers in diffuse large B-cell lymphoma.

Localization of MLH3 at the centrosomes.

Mutations in human DNA mismatch repair (MMR) genes are commonly associated with hereditary nonpolyposis colorectal cancer (HNPCC). MLH1 protein heterodimerizes with PMS2, PMS1, and MLH3 to form MutLα, MutLß, and MutLγ, respectively. We reported recently stable expression of GFP-linked MLH3 in human cell lines. Monitoring these cell lines during the cell cycle using live cell imaging combined with confocal microscopy, we detected accumulation of MLH3 at the centrosomes. Fluorescence recovery after photobleaching (FRAP) revealed high mobility and fast exchange rates at the centrosomes as it has been reported for other DNA repair proteins. MLH3 may have a role in combination with other repair proteins in the control of centrosome numbers.

Stable expression of MutLγ in human cells reveals no specific response to mismatched DNA, but distinct recruitment to damage sites.

The human DNA mismatch repair (MMR) gene family comprises four MutL paralogues capable of forming heterodimeric MutLα (MLH1-PMS2), MutLß (MLH1-PMS1), and MutLγ (MLH1-MLH3) protein complexes. Human MutL subunits PMS2 and MLH3 contain an evolutionarily conserved amino acid motif DQHA(X)2E(X)4E identified as an endonucleolytic domain capable of incising a defective DNA strand. PMS2 of MutLα is generally accepted to be the sole executor of endonucleolytic activity, but since MLH3 was shown to be able to perform DNA repair at low levels in vitro, our aim was to investigate whether or not MLH3 is activated as a backup under MutLα-deficient conditions. Here, we report stable expression of GFP-tagged MLH3 in the isogenic cell lines 293 and 293T which are functional or defective for MLH1 expression, respectively. As expected, MLH3 formed dimeric complexes with endogenous and recombinant MLH1. MutLγ dimers were recruited to sites of DNA damage induced by UVA micro-irradiation as shown for MutLα. Surprisingly, splicing variant MLH3Δ7 lacking the endonucleolytic motif displayed congruent foci formation, implying that recruitment is not necessarily representing active DNA repair. As an alternative test for repair enzyme activity, we combined alkylation-directed DNA damage with comet formation assays. While recombinant MutLα led to full recovery of DNA damage response in MMR deficient cells, expression of MutLγ or single MLH3 failed to do so. These experiments show recruitment and persistence of MutLγ-heterodimers at UVA-induced DNA lesions. However, we demonstrate that in a MutLα-deficient background no DNA repair-specific function carried out by MutLγ can be detected in living cells.

Kaposi's sarcoma-derived cell line SLK is not of endothelial origin, but is a contaminant from a known renal carcinoma cell line.

Kaposi's sarcoma (KS) is an endothelial cell-derived tumor. Investigations of the molecular mechanisms of KS pathogenesis and the identification of drugs for treatment of KS depend critically on valid cell-culture models. Two major immortalized cell lines are available for KS research. Recently, the KS cell line KS Y-1 has been shown to be cross-contaminated with the T24 urinary bladder cancer cell line (ATCC HTB-4). Here, we show by short tandem repeat profiling that the second KS cell line, SLK, is indistinguishable from the clear-cell renal-cell carcinoma cell line Caki-1. Immunocytochemical detection of cytokeratin expression confirmed the epithelial-cell origin of SLK cells. Our findings indicate that SLK cells are not of endothelial origin and should not be used in future studies as a model for KS-derived endothelial tumor cells. We suggest that in the future, more attention needs to be paid to the authenticity of cells in lines derived from human tissues.

High-throughput SNP-based authentication of human cell lines.

Use of false cell lines remains a major problem in biological research. Short tandem repeat (STR) profiling represents the gold standard technique for cell line authentication. However, mismatch repair (MMR)-deficient cell lines are characterized by microsatellite instability, which could force allelic drifts in combination with a selective outgrowth of otherwise persisting side lines, and, thus, are likely to be misclassified by STR profiling. On the basis of the high-throughput Luminex platform, we developed a 24-plex single nucleotide polymorphism profiling assay, called multiplex cell authentication (MCA), for determining authentication of human cell lines. MCA was evaluated by analyzing a collection of 436 human cell lines from the German Collection of Microorganisms and Cell Cultures, previously characterized by eight-loci STR profiling. Both assays showed a very high degree of concordance and similar average matching probabilities (~1 × 10(-8) for STR profiling and ~1 × 10(-9) for MCA). MCA enabled the detection of less than 3% of contaminating human cells. By analyzing MMR-deficient cell lines, evidence was obtained for a higher robustness of the MCA compared to STR profiling. In conclusion, MCA could complement routine cell line authentication and replace the standard authentication STR technique in case of MSI cell lines.

Match criteria for human cell line authentication: where do we draw the line?

Continuous human cell lines have been used extensively as models for biomedical research. In working with these cell lines, researchers are often unaware of the risk of cross-contamination and other causes of misidentification. To reduce this risk, there is a pressing need to authenticate cell lines, comparing the sample handled in the laboratory to a previously tested sample. The American Type Culture Collection Standards Development Organization Workgroup ASN-0002 has developed a Standard for human cell line authentication, recommending short tandem repeat (STR) profiling for authentication of human cell lines. However, there are known limitations to the technique when applied to cultured samples, including possible genetic drift with passage. In our study, a dataset of 2,279 STR profiles from four cell banks was used to assess the effectiveness of the match criteria recommended within the Standard. Of these 2,279 STR profiles, 1,157 were grouped into sets of related cell lines-duplicate holdings, legitimately related samples or misidentified cell lines. Eight core STR loci plus amelogenin were used to unequivocally authenticate 98% of these related sets. Two simple match algorithms each clearly discriminated between related and unrelated samples, with separation between related samples at ≥80% match and unrelated samples at <50% match. A small degree of overlap was noted at 50-79% match, mostly from cell lines known to display variable STR profiles. These match criteria are recommended as a simple and effective way to interpret results from STR profiling of human cell lines.

DSMZCellDive: Diving into high-throughput cell line data.

Human and animal cell lines serve as model systems in a wide range of life sciences such as cancer and infection research or drug screening. Reproducible data are highly dependent on authenticated, contaminant-free cell lines, no better delivered than by the official and certified biorepositories. Offering a web portal to high-throughput information on these model systems will facilitate working with and comparing to these references by data otherwise dispersed at different sources. We here provide DSMZCellDive to access a comprehensive data source on human and animal cell lines, freely available at celldive.dsmz.de. A wide variety of data sources are generated such as RNA-seq transcriptome data and STR (short tandem repeats) profiles. Several starting points ease entering the database via browsing, searching or visualising. This web tool is designed for further expansion on meta and high-throughput data to be generated in future. Explicated examples for the power of this novel tool include analysis of B-cell differentiation markers, homeo-oncogene expression, and measurement of genomic loss of heterozygosities by an enlarged STR panel of 17 loci. Sharing the data on cell lines by the biorepository itself will be of benefit to the scientific community  since it (1) supports the selection of appropriate model cell lines, (2) ensures reliability, (3) avoids misleading data, (4) saves on additional experimentals, and (5) serves as reference for genomic and gene expression data.

Generation of Human Nociceptor-Enriched Sensory Neurons for the Study of Pain-Related Dysfunctions.

In vitro models of the peripheral nervous system would benefit from further refinements to better support studies on neuropathies. In particular, the assessment of pain-related signals is still difficult in human cell cultures. Here, we harnessed induced pluripotent stem cells (iPSCs) to generate peripheral sensory neurons enriched in nociceptors. The objective was to generate a culture system with signaling endpoints suitable for pharmacological and toxicological studies. Neurons generated by conventional differentiation protocols expressed moderate levels of P2X3 purinergic receptors and only low levels of TRPV1 capsaicin receptors, when maturation time was kept to the upper practically useful limit of 6 weeks. As alternative approach, we generated cells with an inducible NGN1 transgene. Ectopic expression of this transcription factor during a defined time window of differentiation resulted in highly enriched nociceptor cultures, as determined by functional (P2X3 and TRPV1 receptors) and immunocytochemical phenotyping, complemented by extensive transcriptome profiling. Single cell recordings of Ca2+-indicator fluorescence from >9000 cells were used to establish the "fraction of reactive cells" in a stimulated population as experimental endpoint, that appeared robust, transparent and quantifiable. To provide an example of application to biomedical studies, functional consequences of prolonged exposure to the chemotherapeutic drug oxaliplatin were examined at non-cytotoxic concentrations. We found (i) neuronal (allodynia-like) hypersensitivity to otherwise non-activating mechanical stimulation that could be blocked by modulators of voltage-gated sodium channels; (ii) hyper-responsiveness to TRPV1 receptor stimulation. These findings and several other measured functional alterations indicate that the model is suitable for pharmacological and toxicological studies related to peripheral neuropathies.

53BP1 and MDC1 foci formation in HT-1080 cells for low- and high-LET microbeam irradiations.

An improved assessment of the biological effects and related risks of low doses of ionizing radiation is currently an important issue in radiation biology. Irradiations using microbeams are particularly well suited for precise and localized dose depositions, whereas recombinant cell lines with fluorescent proteins allow the live observation of radiation-induced foci. Living cells of the fibrosarcoma cell line HT-1080 stably expressing 53BP1 or full-length reconstituted MDC1 fused to Green Fluorescent Protein (GFP) were irradiated with protons and α-particles of linear energy transfers (LETs) of 15 and 75 keV/µm, respectively. Using a microbeam, the irradiations were carried out in line patterns, which facilitated the discrimination between undefined background and radiation-induced foci. As expected, foci formation and respective kinetics from α-particle irradiations with a high LET of 75 keV/µm could be detected in a reliable manner by both fusion proteins, as reported previously. Colocalization of γ-H2AX foci confirmed the DSB nature of the detected foci. As a novel result, the application of protons with low LET of 15 keV/µm generated 53BP1- and MDC1-mediated foci of almost equal size and slightly different kinetics. This new data expands the capability of 53BP1 and wild-type MDC1 on visible foci formation in living cells after irradiation with low-LET particles. Furthermore, the kinetics in HT-1080 cells for α-particle irradiation show a delay of about 20 s for 53BP1 foci detection compared to wild-type MDC1, confirming the hierarchical assembly of both proteins. Preliminary data for proton irradiations are shown and also these indicate a delay for 53BP1 versus MDC1.

First report on establishment and characterization of a carcinosarcoma tumour cell line model of the bladder.

Carcinosarcoma of the urinary bladder is a very rare and aggressive subtype of bladder cancer with poor prognosis. Characteristically carcinosarcomas exhibit biphasic nature with both epithelial and mesenchymal differentiation. Limited information is available regarding its clinical features and appropriate treatments due to its rarity. Development of tumour models can further our understanding of bladder carcinosarcoma. We report establishment and characterization of the first-ever bladder carcinosarcoma cell line MaS-3. It is established by the outgrow method from 86 year-old caucasian male who underwent a radical pelvic resection after neoadjuvant radiotherapy. MaS-3 showed carcinosarcoma profile with high conformity with to the original tumour in terms of immunocytochemistry. Proteome analysis also aligned the MaS-3 cell line with the carcinosarcoma specimen rather than corresponding non-malignant tissue. Chemotherapy sensitivity testing revealed a great sensitivity of MaS-3 growth to 5-Fluorouracil, Gemcitabine and Cisplatin, with almost no impact of Irinotecan. Additionally, the suitability of MaS-3 for 3D in vitro experiments was also demonstrated. The newly established cell line MaS-3 shows typical characteristics of the tumour and may thus be a useful in vitro model system for studying the tumour biology and developing future of treatments of this rare but very aggressive entity.