www.fondation-medic.ch

Bioinformatics Core Facility (BCF)

Group presentation

The group of Mauro Delorenzi is specialised in the application of statistical and bioinformatics methods in biological data analysis, with a focus on new technologies and on molecular and medical translational research.
The core competence and activities reside in the interface between biomedical sciences, statistics and computation, particularly in the application of high-throughput omics technologies, such as gene-expression microarray, to problems of clinical importance, such as development of cancer biomarkers. The BCF also offers consulting, teaching and training, data analysis support and research collaborations for both academic and industrial partners.

The most important projects include the analysis of deep sequencing chip-seq data of the CycliX consortium, collaborations with the research groups of M. Hegi (glioblastoma), L. de Leval (T cell lymphomas), S. Tejpar (colorectal cancer) and T. Petrova (colon cancer and PROX1 transcription factor).
The group participates to international cancer collaborations, in particular in colon cancer (EU’s Coltheres consortium), as well as in method and infrastructure development (EU’s P-Medicine consortium).

We are also interested in testing tumor risk scores proposed by other researchers in independent data to understand which one are really useful, how they can be combined among themselves and with known prognostic factors and to work out implications for patient management and personalized treatment optimization.

Heatmap showing the patterns of differential gene expression used to define 5 subtypes (A-E) of CRC, in comparison with normal reference tissue (N). On the left short biological description of the gene functions.

Molecular heterogeneity of colorectal cancer

The recognition that colorectal cancer (CRC) is a heterogeneous disease in terms of clinical behaviour and response to therapy translates into an urgent need for robust molecular disease subclassifiers that can explain this heterogeneity beyond current parameters (MSI, KRAS , BRAF). Attempts to fill this gap are emerging.

We have contributed to clarifying the role of mutations in key genes for application as prognostic and predictive markers. In the analysis of gene expression profiles, we have discovered that a very specific pattern is associated with tumors that carry activating BRAF mutations. This pattern is shared with a set of additional tumors, which show similar characteristics, and in particular very poor survival in the metastatic setting. It thus identifies a very interesting group for whom new treatments are required.

Following up on this study, we have been developing a system for classifying colon cancer cases into separate "prototypical" groups. The approach chosen is to establish a system of several typical subtype patterns in gene expression and to then associate each new tumor with the most similar subtype. The subtypes might be due to different tumor development histories. They show different biological characteristics, some of likely clinical relevance, such as risk of relapse and sensitivity to different pathway-inhibition directed drug treatments.

We performed a prior free analysis of CRC heterogeneity on over 1'000 CRC gene expression profiles and confronted our findings to established molecular determinants and clinical, histopathological and survival data. Unsupervised clustering based on gene modules allowed us to distinguish at least five different gene expression CRC subtypes, which we call surface crypt-like, lower crypt-like, CIMP-H-like, mesenchymal and mixed.

A gene set enrichment analysis combined with literature search of gene module members identified distinct biological motifs in different subtypes. The subtypes, which were not derived based on outcome, nonetheless showed differences in prognosis. Known gene copy number variations and mutations in key cancer-associated genes differed between subtypes, but the subtypes provided a classification beyond what all other variables could provide. The objective existence of the subtypes and their clinical and molecular characteristics were validated in an independent set of 720 CRC expression profiles.

Characteristics associated with the CRC subtypes and normal reference tissue gene expression groups.

The subtypes provide a novel perspective on the heterogeneity of CRC. The proposed subtypes should be further explored retrospectively on existing clinical trial datasets for their association with treatment efficacy. If sufficiently robust, they could be prospectively assessed for clinical relevance in terms of prognosis and personalized treatment decision making.

We plan to continue this line of research in a wide international collaboration with the aim to generate a consensus subtyping system for colorectal cancer, with participation of various teams that have published similar studies.

Collaboration with other MEDIC partners

In addition to this main project, we bring bioinformatics, computing and statistic support to several other MEDIC projects.
We contribute to the research effort of Laurence de Laval and her group, to characterize molecular biomarkers relevant to the diagnosis and prognosis of nodal peripheral T-cell lymphomas. In this project, we attempt to identify subgroups in these rare malignancies, based on gene expression patterns. Understanding the molecular mechanisms underlying these subgroups might have important implications for diagnosis and response to therapy.
We also bring bioinformatics support to the group of Tatiana Petrova in the study of mechanisms involved in colon cancer metastasis formation and growth. Petrova's lab is particularly interested in the role of PROX1, a gene involved in the interaction between the cancer cells and the host stromal cells, closely associated with the signaling pathway involved in regulation of cell differentiation in intestinal mucosa (the Wnt pathway) and also in transformation of intestinal epithelial cells into cancer cells. Recent experiments suggest that PROX1 is involved in progression to later stages (e.g.metastasis) rather than in the initial stages of colorectal cancer development. We try to uncover the molecular mechanisms by which PROX1 may influence metastatic formation or growth by the bioinformatics analysis of gene expression experiments, when PROX1 is up- or down-regulated.