The Molecular Pathology of Sarcomas

The Journal of Pathology 2011 Virtual Issue Number 3, August

Compiled and annotated by Fouad Al Dayel (1) and Peter A Hall (1,2) Department of Pathology & Laboratory Medicine, (1) and Department of Molecular Oncology (2), King Faisal Specialist Hospital & Research Centre, PO Box 3354, Riyadh 11211, Saudi Arabia

In the recent past there has been considerable progress in our understanding of the molecular events underpinning sarcomas as new technologies and insight are brought to bear on these tumours. While relatively rare for the most part, sarcomas remain a very significant burden on healthcare systems and of course the affected parents and families. Consequently progress in understanding their biology and the molecular events that underpin them will hopefully translate into enhanced diagnostics and new therapeutic approaches. This may be true not only for sarcomas but lessons here may have relevance elsewhere. A number of important papers in The Journal of Pathology underscore this view and here we highlight some of those important contributions.

Lessons and new insights from gastrointestinal stromal tumours

Gastrointestinal stromal tumours (GISTs), the most common sarcoma of the gastrointestinal tract can provide both diagnostic and clinical problems and in her incisive review Cristina Antonescu provides a comprehensive overview of the molecular pathogenesis of these tumours [1]. She also provides a clear perspective on the mode of action and problems inherent in the use of kinase inhibiting drugs in GIST and in other kinase-driven cancers. While the mutation status of receptor tyrosine kinases such as KIT and PDGFRA are central to the behaviour of GISTs it is of course important to remember that other molecular events are present in these tumours. In another recent paper [2] Florian Haller and colleagues have demonstrated differential expression of miRNAs in GISTs and emphasized that some of these clustered in a genetically imprinted region of 14q32.31. While their data need to be validated in larger series the data suggest that perturbation of miRNA networks associate with tumour progression. In principle this may have therapeutic opportunities.

1. The GIST paradigm: lessons for other kinase-driven cancers Cristina R Antonescu The Journal of Pathology 2011; 223: 252-262. (Invited review)

2. Localization- and mutation-dependent microRNA (miRNA) expressions signatures in gastrointestinal stromal tumours (GISTs), with a cluster of co-expressed miRNAs located at 14q32.32 Florian Haller, Anja von Heydebreck, Jitao David Zhang, Bastian Gunawan, Claus Langer, Giuliano Ramadori, Stefan Wiemann and Özgür Sahin The Journal of Pathology 2010: 220; 71–86. (Original paper)

Molecular events in chordoma: prospects for new therapies

Another tumour in which recent progress is opening the opportunity for therapeutic intervention is chordoma. The pathology of these tumours is well known with the pathognomonic physaliferous cells and can exist in classical and chondroid forms. There is evidence that these arise from notochordal remnants and management can be notoriously difficult because of their axial location, often at the base of the skull. Two important papers [3,4] from multicentre and multinational groups led by Adrienne Flanagan describe perturbations of the EGF receptor and the transcription factor T (also known as brachyury) in chordoma. EGFR over expression is seen in 69% of chordomas and in nearly 40% of informative cases there is high level amplification of this locus. Furthermore the functional significance of this is evident from studies of chordoma cell growth in vitro in the presence of EGFR inhibitors [3].

In the second paper Flanagan and colleagues investigated the possible role of the transcription factor T (brachyury) in chordoma biology. They provide clear evidence that T (brachyury) is over expressed in chordomas, often as a consequence of gene amplification, and that this abnormality is critical for proliferation of chordoma cells in vitro [4]. In a linked Commentary by Angelo Dei Tos [5] the relevance of these observations in regard of EGFR and T (brachyury) are placed in a broader context that highlights the importance of the community having a coherent and planned approach to the study of rare cancers. Quoting from this, Dei Tos argues that “the success of innovative clinical trials strongly depends upon the availability of preclinical evidence of efficacy. (These papers) . . . represent good examples of what is actually needed to set up solid prior probability”.

3. The role of epidermal growth factor receptor in chordoma pathogenesis: a potential therapeutic target Asem Shalaby, Nadège Presneau, Hongtao Ye, Dina Halai, Fitim Berisha, Bernadine Idowu, Andreas Leithner, Bernadette Liegl, Timothy RW Briggs, Krisztian Bacsi, Lars-Gunnar Kindblom, Nicholas Athanasou, Maria Fernanda Amary, Pancras CW Hogendoorn, Roberto Tirabosco and Adrienne M Flanagan The Journal of Pathology 2011; 223: 336–346. (Original paper)

4. Role of the transcription factor T (brachyury) in the pathogenesis of sporadic chordoma: a genetic and functional-based study Nadège Presneau, Asem Shalaby, Hongtao Ye, Nischalan Pillay, Dina Halai, Bernadine Idowu, Roberto Tirabosco, Duncan Whitwell, Thomas S Jacques, Lars-Gunnar Kindblom, Silke Brüderlein, Peter Möller, Andreas Leithner, Bernadette Liegl, Fernanda M Amary, Nicholas N Athanasou, Pancras CW Hogendoorn, Fredrik Mertens, Karoly Szuhai and Adrienne M Flanagan The Journal of Pathology 2011; 223: 327–335. (Original paper)

5. Unveiling the molecular pathogenesis of chordoma: a new paradigm for molecular targeting of rare cancers Angelo Paolo Dei Tos The Journal of Pathology 2011; 223: 565-566. (Invited commentary)

Chondroid tumours and IDH1 and IDH2

In another contribution from Flanagan’s group [6] comes important new insight into the biology of cartilaginous tumours. Mutations have been reported in isocitrate dehydrogenase 1 and 2 in patients with multiple enchondromas so it was hypothesized that in a wide range of cartilage tumours IDH1 and IDH2 mutations might be seen. Using high throughput approaches including mass spectrometry and sequencing more than 2000 samples including a wide range of cartilaginous and non cartilaginous tumours were examined. Somatic mutations in IDH1 and IDH2 (IDH1 10 x > IDH2) were only found in central and periosteal cartilaginous tumours and not peripheral cartilaginous or other tumours. These mutations would lead to the accumulation of an oncometabolite (2-hydroxyglutarate). A Commentary by David Thomas highlights the strategies employed and the biological and clinical importance of these important findings [7]. The observations also underscore the increasing importance of defects in metabolism in human cancer, an idea first proposed by Otto Warburg 70 years ago!

6. IDH1 and IDH2 mutations are frequent events in central chondrosarcoma and central and periosteal chondromas but not in other mesenchymal tumours M Fernanda Amary, Krisztian Bacsi, Francesca Maggiani, Stephen Damato, Dina Halai, Fitim Berisha, Robin Pollock, Paul O'Donnell, Anita Grigoriadis, Tim Diss, Malihe Eskandarpour, Nadège Presneau, Pancras CW Hogendoorn, Andrew Futreal, Roberto Tirabosco, Adrienne M Flanagan The Journal of Pathology 2011; 224: 334-343. (Original paper)

7. Lessons from the deep study of rare tumours David M Thomas The Journal of Pathology 2011; 224: 306-308. (Invited commentary)

Molecular events in osteosarcoma development

In the context of osteosarcomas there has also been much progress in our understanding of pathogenesis and molecular alterations. Mohseny et al. [8] have demonstrated that in mice mesenchymal stem cells can be the origin of these aggressive tumours and that this is associated with aneuploidisation and the genomic loss of the locus encoding the cyclin dependent kinase Cdkn2 (human homologue is CDKN2). Importantly these authors from Leiden linked their murine data to human samples by demonstrating CDKN2/p16 protein expression in a series of 88 human tumours.

A number of signalling pathways have been implicated in osteosarcoma. For example Wnt signalling is generally viewed as being active in cancer [9] but Cai et al. have provocative data suggesting the opposite in high grade osteosarcoma [10]. This is a slightly contentious observation but may prove to be very important [9]. Another pathway recently observed to be relevant in osteosarcoma is the Hedgehog system [10]. A Japanese group led by Hiroko Nagao have shown how the Gli2 transcription factor, a core mediator of the Hedgehog pathway, is frequently over-expressed in osteosarcoma and contributes to the behaviour of osteosarcoma cells in vitro and in xenograft models [11]. Step by step a picture of the molecular pathogenesis of osteosarcoma is emerging.

8. Osteosarcoma originates from mesenchymal stem cells in consequence of aneuploidization and genomic loss of Cdkn2 Alexander B Mohseny, Karoly Szuhai, Salvatore Romeo, Emilie P Buddingh, Inge Briaire-de Bruijn, Daniëlle de Jong, Melissa van Pel, Anne-Marie Cleton-Jansen and Pancras CW Hogendoorn Journal of Pathology 2009; 219: 294–305. (Original paper)

9. Wnts, bone and cancer David M Thomas The Journal of Pathology 2010; 220: 1–4. (Invited commentary)

10. Inactive Wnt/β-catenin pathway in conventional high-grade osteosarcoma Yongping Cai, Alexander B Mohseny, Marcel Karperien, Pancras CW Hogendoorn, Gengyin Zhou and Anne-Marie Cleton-Jansen The Journal of Pathology 2010; 220: 24–33. (Original paper)

11. Role of GLI2 in the growth of human osteosarcoma Hiroko Nagao, Kosei Ijiri, Masataka Hirotsu, Yasuhiro Ishidou, Takuya Yamamoto, Satoshi Nagano, Takumi Takizawa, Kinichi Nakashima, Setsuro Komiya and Takao Setoguchi The Journal of Pathology 2010; 224: 169-179. (Original paper)

The Power of in vivo model systems: of fish and mice

Murine systems have been of huge importance in our understanding of cancer biology and Owen Sansom and colleagues [12] from the Beatson Institute has shown how loss of p53 is not the same as p53 mutation! For some time it has simply assumed that loss of function mutations of p53 would be the same as (phenocopy) loss of p53 alleles. Not so! Mutation (as opposed to loss) of p53 was much more associated with tumour progression and metastasis in a murine model. Not only does this highlight the importance of understanding the details of molecular events in cancer but also the need to use in vivo models as argued in the accompanying Commentary by Post and Lozano [13].

New model systems are becoming available. The benefits of zebrafish (Danio rerio) as a model were reviewed in the Journal by Chico et al. [14] as part of the 2010 Annual Review issue. Their small size and relative cheapness as well as reasonably fast generation times are a benefit. In addition the ability to do both forward (classical) and reverse genetics as well as their tractable nature in genetic and other terms offers huge advantages. Hogendoorn’s group in Leiden have championed the use of zebrafish in studying connective tissue biology [15, 16] and the system has huge potential for studying sarcoma biology.

12. p53 mutation and loss have different effects on tumourigenesis in a novel mouse model of pleomorphic rhabdomyosarcoma Brendan Doyle, Jennifer P Morton, David W Delaney, Rachel A Ridgway, Julie A Wilkins and Owen J Sansom The Journal of Pathology 2010; 222: 129-137. (Original paper)

13. You can win by losing: p53 mutations in rhabdomyosarcomas Sean M Post and Guillermina Lozano. The Journal of Pathology 2010; 222: 124–128. (Invited commentary)

14. The genetics of cardiovascular disease: new insights from emerging approaches Timothy JA Chico, Marta Milo and David C Crossman The Journal of Pathology 2010; 220: 186–197. (Invited review)

15. Cartilage ultrastructure in proteoglycan-deficient zebrafish mutants brings to light new candidate genes for human skeletal disorders Malgorzata I Wiweger, Cristina M Avramut, Carlos E de Andrea, Frans A Prins, Abraham J Koster, Raimond BG Ravelli and Pancras CW Hogendoorn The Journal of Pathology 2011; 223: 531–542. (Original paper)

16. Growth plate regulation and osteochondroma formation: insights from tracing proteoglycans in zebrafish models and human cartilage Carlos E de Andrea, Frans A Prins, Malgorzata I Wiweger and Pancras CW Hogendoorn The Journal of Pathology 2011; 224: 160–168. (Original paper)

In vitro systems are still important and informative

While animal model systems are of course important much can be achieved in in vitro systems, especially when data are used to test observations made in clinical samples. Indeed nearly all of the papers cited so far in this Virtual Issue employ in vitro systems to some extent. Two other papers that substantiate that view and that are of relevance in the field of connective tissue biology and sarcomas investigate the role of RANKL in osteoclast biology [17, 18]. RANKL (or receptor activator of nuclear factor kappa-B ligand - also now known as CD254) is crucial in bone biology and has relevance to the behaviour of Ewing’s sarcoma [17] and giant cell tumours of bone [18].

17 Ewing sarcoma cells express RANKL and support osteoclastogenesis Richard Taylor, Helen J Knowles and Nicholas A Athanasou The Journal of Pathology 2011; Early View: DOI: 10.1002/path.2869. (Original paper)

18 Interleukin-34 is expressed by giant cell tumours of bone and plays a key role in RANKL-induced osteoclastogenesis Marc Baud'Huin, Romain Renault, Céline Charrier, Anne Riet, Anne Moreau, Régis Brion, François Gouin, Laurence Duplomb and Dominique Heymann The Journal of Pathology 2010; 221: 77-86. (Original paper)

New perspectives: proteomics and genomics

In concluding we hope that readers will discern from this Virtual Issue that there has been much progress and development in the area of sarcoma biology and that the Journal of Pathology continues to publish important papers in this area. Several of the papers cited contain genomic and mutation data and the developments in high throughput sequencing will have major impact in pathology in general. Certainly a recent paper from Gibault et al. [19] highlights the genomic complexity of sarcomas. Some have relatively simple genomics (Sarcomas with Simple Genomics), while others have many complex genetic changes and are thus (Sarcomas with Complex Genomics). Understanding the pathogenesis of these molecular events and their diagnostic and clinical implications will, we are sure, advance in the coming years. In addition other approaches are, and will, come to be applied. For example, Willems et al. [20] elegantly demonstrated the use of imaging mass spectrometry to define the biochemical complexity of sarcomas. Much remains to be done and the tools are now at hand!

19 New insights in sarcoma oncogenesis: a comprehensive analysis of a large series of 160 soft tissue sarcomas with complex genomics Laure Gibault, Gaëlle Pérot, Frédéric Chibon, Sarah Bonnin, Pauline Lagarde, Philippe Terrier, Jean-Michel Coindre and Alain Aurias The Journal of Pathology 2011; 223: 64–71. (Original Paper)

20 Imaging mass spectrometry of myxoid sarcomas identifies proteins and lipids specific to tumour type and grade, and reveals biochemical intratumour heterogeneity Stefan M Willems, Alexandra van Remoortere, René van Zeijl, André M Deelder, Liam A McDonnell and Pancras CW Hogendoorn The Journal of Pathology 2010; 222: 400–409. (Original paper)

Questions

The following questions can be answered by reading and reflecting upon the above annotation and the papers that are cited within it. Within the Royal College of Pathologists Continuing Professional Development (CPD) scheme, CPD points may be earned by writing reflective notes on the papers in this Virtual Issue and the questions are designed to act as a focus for this activity. To do this, you may wish to use the Royal College of Pathologists' reflective notes form.

Question 1 How do the products of the KIT and PDGFRA genes contribute to gastrointestinal stromal tumours and how do they provide therapeutic opportunities?

Question 2 What are the typical locations of chordomas and does this relate to their histogenesis? How does T (brachyury) and the EGF receptor contribute to chordoma biology?

Question 3 Otto Warburg proposed that metabolic defects were key to cancer biology in the 1930s. How does new knowledge about the molecular basis of cartilaginous tumours provide support for that view?

Question 4 Outline the Wnt and the Hedgehog signalling pathways. How are these deranged in neoplasia?

Question 5 What are the advantages of zebrafish (Danio rerio) as models of human disease?