Analysis of cancers involving altered prenylated proteins

The cancer project focuses on the identification of genetic variants that contribute either to increase the risk of digestive cancer or to aggravate their evolution, according to 2 main axes:

  • Pharmacogenetics and therapeutic evaluation in digestive oncology
  • Hereditary predispositions to digestive cancers

A third axis "genetics of sarcomas and related tumors" was initiated this year
Joining Nicolas Lévy's team, the research projects have been re-centered on tumors impairing pathways involved in premature aging, mainly RAS and TGFB, to identify potential therapeutic targets useful in in both conditions. The pancreatic cancerogenesis of mucinous cysts has been chosen as study model.

Past results

The projects of the group focused on the study of gastrointestinal tumorigenesis mechanisms to identify genetic variations or mutations that contribute to or increase their risk, or impact their evolution. The clinical objective was to better integrate into medical practice DNA analyses, which take a special place in various aspects of oncology, prevention, early diagnosis and therapeutic evaluation. In the past, we described the genetic alterations responsible for major genetic predispositions to digestive cancer and their relationship with the main phenotypic characteristics. The molecular analyses have been progressively transferred to hospital laboratories and integrated to the medical care of patients. We coordinate the data collection of germ line mutations identified in the French labeled laboratories in national mutations databases and permanently implement these LSDBs with functional data allowing variants classification. Furthermore, the study of somatic genetic alterations in colorectal cancer, to which we contributed, identified two alternative pathways of colon cancer, linked to different prognostic risks. The allelic loss frequencies and the complexity of their associations, as well as the difficulty of finding a major event in tumor progression prompted us to undertake a comprehensive genomic characterization of colorectal cancer combining genome and transcriptome analysis. The direct responsibility of the WNT pathway via the APC gene appeared the most probable hypothesis that was specifically explored 59. To document the existence of genes predisposing to the development of metastases based on genotype-phenotype associations, we set up a national group of clinicians and researchers to collect such case-control series and were able to describe the first locus involved in the metastatic process 60, where a non-coding RNA was then identified by another group 61, 62.
An interface contract with the Institut Paoli-Calmettes allowed us to develop a regional laboratory that contributes to the diagnostic and therapeutic evaluation of cancer.


Joining Nicolas Lévy's team, the group will re-direct its research projects on tumors impairing pathways also involved in premature aging, mainly RAS and TGFB, to identify potential therapeutic targets in pathways involved in both conditions. Since 2006, targeted therapies have been developed to improve solid tumors treatment, like anti-EGFR and Tyrosine kinase inhibitors that are now used in metastatic colorectal and lung cancers. Anti-EGFRs have been first demonstrated as ineffective in case of K-RAS gene activation, and their use has been restricted to tumors without K-RAS somatic mutations. Subsequently, a systematic molecular analysis of tumor cells has been introduced in cancer patients’ management, and the EGFR pathway has been more extensively characterized to document the mechanisms of drug resistance. Similar observations have been made in cases of N-RAS, BRAF or PTEN mutations as examples. Drugs targeting other steps of the pathway have been tested and the use of prenylation inhibitors revealed promising directions to manage devastating tumors such pancreatic, lung and ovarian cancers 63, 64 for reviews. The use of xanthohumol, a prenylated chalcone, is able to induce cell cycle arrest and apoptosis by inhibiting phosphorylation of STAT3, together with expression of downstream targets cyclin D1 and survivin in cell lines and xenografts of pancreatic cancer 65. Combined geranylgeranyltransferase 1 and farnesyltransferase inhibitors liposomal delivery induces G1 cycle cell arrest by p21CIP1/WAF1 in human pancreatic and lung cancer cell lines 66. Zoledronic acid on lung cancer and melanoma cells with NRAS mutation inhibited prenylation, supressing downstream RAS and EGFR signaling. This effect has been reversed using geranylgeraniol and farnesol, confirming the specific prenylation inhibition 67, 68. In combination with paclitaxel, which showed promising results in pancreatic cancer, synergistic effects could be obtained. Blocking the mevalonate pathway with statins also shows interference with post-translational modification of KRAS inducing antiproliferative phenotype of ovarian cancer cells from xenografts 69, suppression of proliferation linked to endocrine resistance resulting from the accumulation of senescent cells after radio/chemotherapy of breast and prostate cancer 70, 71. Isoprenoid biosynthetic pathway inhibitors also showed Rap1A alteration in metastatic prostate cancer cells 72.

Pancreatic cancer is one of the most lethal human cancers, exhibiting RAS pathway activation in more than 80% of cases. The unique treatment consists in chemotherapy based on 5FU or gemcitabine solely or in combination with irinotecan and/or oxaliplatine, adjuvant to surgery in less than 20% of cases, often used as palliative. Clinical trials currently introduce other molecules without any positive result. Pancreatic cancer also takes part of the tumor spectrum of several genetic disorders highly predisposing to cancer (BRCA, Lynch, Peutz-Jeghers, Cowden, Von Hippel Lindau syndromes as the most frequent). Precancerous lesions have been described, particularly pancreatic cysts with mucinous component, known at high risk for cancer. These tumors were retained as an interesting model to extensively study the EGFR/RAS pathway, looking for targetable modifications. A collaborative consortium has been set up in Marseille, including all interventional digestive endoscopy units and the corresponding surgical departments (AP-HM hospitals Timone and Nord, St-Joseph hospital, European Hospital, Institut Paoli-Calmettes), and two expert centers in Lyon (Mermoz hospital) and Paris (Les Peupliers hospital). We plan to collect samples of cystic fluid and analyze“circulating” DNA from surrounding epithelial cells. As blood circulating tumor cells showed their ability to reveal molecular alterations as cancer diagnostics or recurrence markers, tumor cells are initially present in cystic fluids and can be collected for further molecular investigations 73, 74.

A pilot study has just been funded, which will test the feasibility of a systematic approach comparing the molecular profiles of cysts in pre-operative echoendoscopic biopsies and per-operative samples of 10 patients, thus validating the preoperative approach as applicable to a larger series, as echoendoscopic biopsies are systematically performed before surgery to confirm the clinical indication of pancreatectomy. Sample processing and NGS sequencing will be done within U910. Nucleic acids will be extracted and sequenced on a panel of 70 genes involved in pancreatic tumorigenesis and targeting the RAS signaling pathway, MAPK, AKT, JAK-STAT signaling pathway, WNT, TGFB, TP53 and repair BRCA, ATM. The selected sequencing technology (HaloPlexHS®, Agilent), through the incorporation of random nucleic acid sequences (10 nucleotides) during the libraries preparation, enables detection of alleles present in a 1% proportion. This new technical approach that combines high sensitivity and specificity is also compatible with degraded DNA and/or small amounts (less than 50 ng). It therefore appears as an appropriate technique for this type of samples with an expected important degradation and nucleic clonal heterogeneity. Bioinformatics analysis will detect somatic variations and define their allelic frequency as well as CNVs. A comparison of the molecular profiles of pre- and post-operative samples will be performed, then faced with the pathologic and biological characteristics. The identification of a molecular profile indicating a precancerous condition would be helpful for the surgical decision, which consists in a (cephalic) pancreatectomy, now performed twice more than a posteriori required. The characterization of the very early recurrent molecular alterations would also be helpful to refine targeted therapies, and to test cell models exhibiting similar pathways disorders.
At the same time, the group, in collaboration with French clinical oncogenetics units, is currently collecting families with aggregation of breast/pancreatic cancer without known germline alterations to define through exomes analyses a set of candidate genes that could be involved in such predispositions. Variations identified by this approach will be compared to those enlightened in the previous part, to document the genetic alterations responsible for the very early steps of pancreatic tumorigenesis.
Finally, as multichemotherapy is currently the single therapeutic approach in pancreatic cancer, the group drives pharmacogenetics studies to possibly add new parameters useful in dosage adjustment in collaboration with the oncology units of the university hospital. This year, clinical files of 98 patients have been registered and 87 have been genotyped on 1636 SNPs of 231 genes using the DMET chip (Affymetrix). Associations with toxicity have been identified for genes participating to phases I and II metabolism enzymes and transporters. The set of data will be extended, genotyping 279 patients included in the national clinical trial PRODIGE35 in collaboration with the Fédération Française de Cancérologie Digestive (FFCD) to validate primary results.