Scientific Areas Addressed:

1. Germ Cell and Stem Cells: insights from Developmental Biology
2. Current treatment of GCTs in children and adults
3. Pathway discovery in germ cell tumorigenesis via whole-genome surveys
4. The molecular epidemiology of normal and abnormal germ cell development
5. Epigenetic regulation of stem cells and germ cells
6. New animal models of germ cell tumors
7. Stem cell pathways in germ cell tumors: wnt, BMP, microRNAs and others
8. New paradigms to leverage stem cell pathways for the treatment of disease

Keynote speaker Haifan Lin (Yale University) started the meeting with an in-depth look at emerging roles for small RNAs in gene regulation and germ cell self-renewal. Thousands of piRNAs are generated from a relatively small number of mainly heterochromatic sites in the genome. The piwi protein binds chromosomal sites through association with piRNAs, serving to recruit heterochromatic protein HP1A upstream of H3K9 methylation, a finding with important implications for the mechanisms of differential methylation in the germline. The observation that piwi overexpression leads to increased ovarian stem cells in Drosophila led to an exploration of possible roles for piwi in germ cell tumors. Piwi is highly expressed in seminomas, stem cell-like GCTs, and piRNAs thus may represent a new avenue for GCT therapy. Carlos Rodriguez-Galindo (Dana-Farber Cancer Institute, Boston) discussed retinoblastoma as a paradigm of a rare childhood tumor. Though the RB1 mutation was actually the first tumor-suppressor identified, this has not translated into improved therapies. The relatively small number of patients makes it imperative to build collaborative networks in order to test new therapies. He described such an integrated clinical-translational effort using improved animal xenograft models, resulting in new indications for topotecan in Rb therapy. Dr. Rodriguez-Galindo further detailed the discovery of MDMX amplification in retinoblastoma, leading to the insight that MDMX inhibitors could restore P53 function and impair the viability of Rb cells.

Julie Ross (University of Minnesota) discussed the epidemiology of GCTs. There is a bimodal age distribution of these tumors, with one peak of incidence occurring in young children aged 0-4 and a second peak beginning in adolescence. She reviewed evidence, based on analysis of imprinted genes and differing cytogenetic abnormalities, that childhood and adolescent/adult GCTs originate from germ cells at different stages of embryonic development. The incidence of GCTs is increasing worldwide for unclear reasons, especially in adolescents and adults. Dr. Ross discussed emerging data from epidemiologic studies on risk factors for testicular GCT, including low birth weight, dichotomous twining and hernia. In response to this problem, Dr. Ross and co-workers have developed molecular epidemiologic approaches examining neonatal blood spots, maternal blood and cord blood for biomarkers of exposure and effect, and changes in DNA methylation and gene expression, with plans to model potential etiologies in genetically-engineered mouse models.

Jenny Poynter (University of Minnesota) The relative rarity of childhood GCT has made investigation of epidemiologic factors for these tumors difficult. Dr. Poynter along with Dr. Ross has initiated a study through the Children’s Cancer Research Network to identify nearly all cases of childhood GCT in the US. The study will evaluate genetic variation in key pathways involved in germ cell development using a case-parent triad design. A second aim is to explore heterogeneity in DNA methylation by tumor histology. Preliminary data using SNP analysis in a case-control showed several genes associated with GCT incidence, including BAK1, an intriguing finding given the potential role of apoptotic programs in quality control of germ cells during early development. She also discussed genome-wide methylation data indicating differential methylation in the tumors of genes highly associated with human Embryonic Stem Cell pluripotency, including Wnt and TGF-b pathway genes.

John Abrams (UT Southwestern Medical Center) initiated a session on genetic models of germ cell development and tumorigenesis, describing recent work from his lab on the role of TP53 in germ cell and stem cell development. Multiple lines of evidence indicate that the evolutionary appearance of p53 preceded tumor suppression, suggesting that there may be unappreciated functions for this protein. Working in Drosophila, Dr. Abrams found that the first catalytic step in meiotic recombination instigates programmed activation of p53 during germline development. The initiation of meiotic recombination events as an intrinsic stimulus for the p53 regulatory network also occurred in mice, suggesting that tumor suppressive functions may have been co-opted from primordial activities during meiotic recombination. Using this same Drosophila reporter system, he also found that stress-induced damage selectively activates p53 in germ line stem cells. Furthermore, hyperactivation of p53 was observed in several germ line tumor models where inappropriate proliferation occurs beyond the stem cell compartment. These discoveries suggest that ancient, currently unappreciated pathways may link oncogenic stress to p53 activation.

Hao Zhu (Howard Hughes Medical Institute, Children’s Hospital Boston) discussed his work in Dr. George Daley’s laboratory. Lin28a is essential for proper PGC development both in vitro and in vivo, and recent studies have shown that LIN28A is a highly sensitive and specific marker for GCTs, but functional or mechanistic roles have not been defined in these tumors. Dr. Zhu reported unexpected roles for Lin28a/b and let-7 in regulating mammalian glucose metabolism. When overexpressed in transgenic mice, both Lin28a and LIN28B promote a metabolic state with enhanced glucose uptake and insulin sensitivity that resists obesity-induced hyperglycemia, by activating the PI3K-mTOR pathway. Overexpression of let-7, on the other hand, leads to impaired glucose uptake, due to let-7-mediated repression of the PI3K-mTOR pathway. In vitro and in vivo, the Lin28a mediated glucose uptake and insulin-sensitive phenotype is abrogated by the mTORC1 inhibitor rapamycin. These data establish the Lin28/let-7 pathway as a regulator of mammalian glucose metabolism, with implications for cancer metabolism and the use of Mtor inhibitors in GCTs.

Michael Buszczak (UT Southwestern Medical Center) followed with a discussion of how specialized niche environments control the development and differentiation of germ cells in Drosophila. The mechanisms that restrict niche cell formation during development and limit niche signaling output remain poorly understood. Dr. Buszczak presented data showing that the histone demethylase Lsd1 limits the size of the germline stem cell (GSC) niche within Drosophila ovaries. Lsd1 mutants display an expanded a GSC-like population, marked by ectopic BMP signal responsiveness throughout much of the germline. He showed that Lsd1 functions within the escort cells (ECs) that reside immediately adjacent to cap cells and prevents them from ectopically producing niche specific signals. The mammalian homolog of Lsd1 contributes to cancer and metastasis, highlighting the importance of niche specific signals for cancer biology.

Mark Krailo (University of Southern California and Children’s Oncology Group) detailed the special biostatistical approaches necessary for conducting clinical research on rare tumor types, such as GCTs. Working with the Children’s Oncology Group, he has identified new methods of risk stratification that are more sensitive and specific for childhood GCTs.

Juliet Hale (University of Newcastle upon Tyne, UK) presented data from UK and European GCT clinical trials, investigating the efficacy of different treatment strategies and the effects of dose modifications on outcomes. She also outlined initial results from a US-UK collaborative effort to merge clinical data on a very large group of patients in order to elucidate novel clinical factors associated with disease progression and outcome. Lindsay Frazier (Dana-Farber Cancer Institute and Harvard School of Public Health) discussed persistent clinical conundrums in GCTs. What are the molecular mechanisms that determine the wide array of histologic subtypes? Why does histology track with age and gender? What makes one histology aspect more chemosensitive (yolk sac) than another (embryonal carcinoma)? Why does the same “histology” –e.g. immature teratoma—have different biologic behavior in an infant compared to a 25 year old woman? She also described the troubling significant late effects beginning to emerge in long term survivors (e.g. twice the risk of heart disease and second malignancies). Working with colleagues in the UK, Dr. Frazier has started to lay the groundwork for a more sophisticated risk stratification system that incorporates biologic features, to allow therapy to be directed to those who need it and identify patients who could be safely observed and could be treated with reduced doses of drugs. These insights could help reshape clinical treatment, outcome and long term health of survivors.

James Nicholson (Cambridge University Hospitals NHS Trust, UK) turned to a discussion of the unique issues surrounding treatment of CNS germ cell tumors. CNS GCTs account for around 3% malignant tumours in the West, with peak incidence in teenagers. Treatment of CNS GCTs creates special challenges, owing to the particular problems of delivering treatment to the brain, and potential biological differences with other GCTs remain unclear. An emerging consensus supports the use of both platinum-based chemotherapy and involved-field radiotherapy for CNS GCTs. Clinical challenges in CNS GCT include reduction in late squeal of treatment for germinoma and improving survival in poor risk NGGCT. Dr. Nicholson also discussed ways in which the identification of biological patterns in CNS GCT subtypes may inform future treatment strategies and improve outcomes in terms of survival and quality of survival.

Blanche Capel (Duke University) described a key mouse model of GCT. Although testicular teratomas are rare in mice, mutation of mouse Dnd1 (Dnd1Ter/Ter), an RNA-binding protein, leads to a high incidence of testicular teratomas on the 129/SvJ genetic background, but not in other strains. In 129/SvJ male mice, Dnd1Ter/Ter germ cells fail to undergo mitotic arrest at fetal stages. Dr. Capel has shown that DND1 binds mRNAs for a large number of genes including key negative regulators of the cell cycle such as P21 and P27. In the 129/SvJ genetic background, mutant germ cells retain expression of pluripotency markers, remain proliferative, and form teratomas. This suggests that cell cycle arrest is critical to regulate the underlying pluripotent program in male germ cells, and genetic background differences that influence this process may explain the varying susceptibility to testicular cancer in the human population. David Zarkower (University of Minnesota) reviewed efforts in his group using conditional gene targeting, combined with mRNA expression profiling and genome-wide ChIP methods, to dissect the functions of DMRT1 in the mouse testis. This analysis revealed that DMRT1 is a key regulator of several distinct processes central to testis development and function. In fetal germ cells DMRT1 controls pluripotency, and mutants develop testicular teratomas at very high frequency, but only on the 129Sv strain background. Postnatally DMRT1 is required in spermatogonia to control the mitosis/meiosis decision, and mutant spermatogonia prematurely initiate meiosis. Recently, Dr. Zarkower discovered that DMRT1 also is required in the postnatal testis to maintain male fate. ChIP and expression analysis suggest that DMRT1 directly controls most of the major regulators of fetal sex determination in the postnatal testis. Based on these and other data he proposed a model for the role of DMRT1 regulation of sex maintenance. In humans DMRT1 is implicated in sex determination, DSD, and several gonadal cancers. Thus this important mouse model may be relevant to the etiology of DMRT1-linked human disorders, and the target genes regulated by DMRT1 are strong candidates for further investigation in human gonadal cancer and DSD.

Chris Wylie (Children’s Hospital, Cincinnati) presented recent results investigating how fetal germ cell development and primordial germ cell (PGC) migration impacts gonadogenesis and the development of GCTs. Elegant approaches using OCT4-GFP mice to track developing PGCs in vivo. This work emphasized the importance of signaling through the kit receptor tyrosine kinase in PGC migration and survival. Mis-migration of PGCs leads to activation of an apoptotic response mediated by P53 target genes. These findings have important implications for the development of extragonadal GCTs.

Leendert Looijenga (Erasmus Medical Center and Daniel den Hoed Cancer Center, Rotterdam) discussed the multiple GCT risk factors impinging on early gonadal development. The precursor lesion of type II GCTs can be carcinoma in situ (CIS) of the testis or gonadoblastoma (GB) of the dysgenetic gonad. The pre-malignant germ cells (as well as seminoma cells) are positive for the embryonic germ cell markers c-KIT, PLAP, OCT3/4, NANOG, and LIN28. In addition, they are hypomethylated. SOX17 is found in CIS/GB and seminoma, while embryonal carcinoma is positive for SOX2. Immunohistochemical detection of stem cell factor (KITL) is informative to identify the earliest malignant germ cells, which finding is of specific interest because of the results of the genome wide linkage studies implying relevance for the c-KIT-KITL pathway in the development of TGCTs. Dr. Looijenga also discussed recent data linking microsatellite instability and treatment resistance in testicular GCTs.

Kate Nathanson (University of Pennsylvania) reviewed the recent GWAS studies conducted by her group. Evidence from many studies has indicated a substantial genetic component to testicular germ cell tumour susceptibility, including the significantly increased familial risk and differential risk among races. However, despite extensive linkage searches on available families, no high penetrance genes have been identified. The GWAS analysis for the first time linked several genes to the development of GCTs, including KITL, SPRY4 and DMRT1. These results were discussed in the larger context of ongoing epidemiologic and biologic studies aimed at identifying the molecular origins of GCTs.

Dinesh Rakheja (UT Southwestern Medical Center) discussed recent results examining the activation status of mTORC1 pathway in a cohort of 14 pediatric yolk sac tumors and 9 pediatric germinomas with the help of immunohistochemical staining for phosphorylated mTOR (p-mTOR), phosphorylated S6 ribosomal protein (p-S6RP)and other pathway components including Cyclin D1, HIF1-alpha, GLUT1 and REDD1. While germinomas showed weak to absent staining, there was robust staining of yolk sac tumors for p-mTOR, p-S6RP, Cyclin D1, and HIF1A, indicating a remarkable dichotomy in the mTORC1 activation status between pediatric yolk sac tumors and germinomas. There was also modest staining of yolk sac tumors for p-ERK1/2, suggesting that activation of the mitogen-activated protein kinase (MAPK) pathway may at least be partly responsible for the active mTORC1 signaling in yolk sac tumors. Germinomas exhibited strong expression of REDD1, consistent with its role in suppressing mTORC1 activity in these tumors. The active mTORC1 pathway in yolk sac tumors suggests that rapamycin and its analogs should be evaluated for the treatment of this pediatric cancer.

Stefan Schönberger (University Children’s Hospital Bonn, Germany) detailed his recent work aimed at interrogating the human kinome of yolk sac tumors, the most common malignant childhood GCT. He has taken the approach of constructing a custom SequenceCapture microarray (NimbleGen) based on all exons of all the 518 members of the kinase family . Genomic DNA from the initial tumor sample and germline material of five patients with childhood YST were enriched on these SequenceCapture microarrays and sequenced on the Roche GS FLX Titanium platform. These data were coupled with kinome-wide expression profiling on TaqMan low density arrays. Initial results show that >95% of the sequence data mapped to the human reference genome. This work represents a powerful workflow for high throughput analysis of the whole kinome in YST samples.

Jim Amatruda (UT Southwestern Medical Center) presented recent data from his lab on the role of embryonic signaling pathways in childhood GCTs. Through a forward genetic screen in zebrafish, the Amatruda lab identified the Type I BMP Receptor Alk6b (Activin Receptor-like Kinase 6b) as a critical regulator of germ cell differentiation. Zebrafish with mutations in Alk6b develop GCTs resembling human seminomas. BMP pathway gene expression profiling and phospho-SMAD immunohistochemistry on a series of clinically annotated human GCTs showed that, in agreement with the zebrafish model, undifferentiated GCTs such as seminomas lack BMP signaling activity, whereas signaling is maintained in differentiated GCTs such as yolk sac tumors. Dr. Amatruda also discussed ongoing studies in conjunction with the Children’s Oncology Group to conduct genome wide interrogation of childhood germ cell tumors, including copy number analysis and deep sequencing with a goal of discovering other aberrant pathways that contribute to human germ cell tumorigenesis.

Matthew Murray (University of Cambridge, UK) discussed new findings on the GCT transcriptome and the importance of microRNA pathways in GCT development. Pathway analysis of differentially-expressed genes reveals that germinomas are particularly enriched for immune-mediated processes, which may reflect their characteristic lymphocytic infiltrate and excellent prognosis, whereas the TGF-ß/BMP pathway, amongst others, is enriched in YSTs. Importantly, a 10-gene mRNA signature is predictive of overall survival in adult males with non-seminomatous GCTs (NSGCTs). MicroRNAs (miRNAs) are short, non-protein coding RNAs that regulate gene expression via translational repression and/or mRNA degradation. Malignant GCTs universally over-express the miR-371~373 and miR-302 clusters regardless of patient age, tumour histological subtype or anatomical site. The diagnostic potential of this finding has recently been demonstrated with the detection of elevated serum levels of these miRNAs at GCT diagnosis, with miR-372 levels falling during treatment. Furthermore, global miRNA profiles segregate malignant subtypes from each other and from non-malignant controls. Integration of mRNA and miRNA datasets will assist our understanding of the intricate and complex pathways/networks involved in GCT pathogenesis.

The meeting concluded with two roundtable discussions, Identifying the important clinical/translational questions in GCT trials (Hale, Nicholson, Frazier, Rodriguez-Galindo), and Prospects for targeted therapy of Germ Cell tumors (Amatruda, Schönberger, Murray, Poynter).

New collaborations emerging from this meeting included:

1. Incorporation of GCT biology studies into MaGIC (Malignant Germ Cell Tumor International Collaboration), a US-UK joint effort (Drs. Frazier, Krailo, Hale, Nicholson, Amatruda, Murray)
2. An R01 application (Fall 2011) entitled “Biological Mechanisms of Chemosensitivity in Germ Cell Tumors” (Drs. Amatruda, Frazier, Krailo, Zarkower, Rakheja).
3. A book to be published with Springer in 2012, Pediatric Germ Cell Tumors, edited by Drs. Amatruda and Frazier, with contributions from Drs. Krailo, Nicholson, Murray, Schönberger, Poynter, Rakheja and Hale.
4. A collaboration to map genomic features of GCT associated with disease progression and treatment resistance, including Drs. Amatruda, Frazier, Looijenga, Schönberger, Krailo and Rakheja.

‍