Les petits ARN nucléolaires (snoARN) sont d'abondants petits ARN non codants impliqués dans la modification post-transcriptionnelle des ARN ribosomiques. Plus récemment, ils ont été associés à de nouvelles fonctions et des dérégulations dans les cancers. La leucémie lymphoïde chronique (LLC) est l'hémopathie maligne la plus courante dans les pays occidentaux. Cette pathologie, bien qu'indolente, est toujours incurable et est très hétérogène en termes d'évolution et de réponse au traitement. Il est ainsi nécessaire de découvrir de nouveaux marqueurs permettant de stratifier le risque d'évolution de la LLC afin d'améliorer la prise en charge thérapeutique des patients. Le but de mon projet a été d'étudier les profils d'expression des snoARN dans la LLC et de les corréler aux données cliniques et biologiques. Par des expériences de PCR quantitative à grande échelle (Fluidigm), j'ai mis en évidence la dérégulation des snoARN dans la LLC. De plus, j'ai pu montrer qu'une signature spécifique était capable de définir un nouveau sous-groupe de mauvais pronostic au sein des patients IGHV-mutés, initialement classés dans un groupe de bon pronostic. La surexpression de la signature est corrélée à un temps de survie sans traitement plus court et semble être principalement activée par les signaux de prolifération. Ainsi, cette étude démontre l'intérêt d'étudier la valeur pronostique des snoARN dans la LLC et plus largement dans les hémopathies malignes. / Small nucleolar RNAs (snoRNAs) are an abundant class of small non-coding RNAs responsible for the post-transcriptional modifications of ribosomal RNAs. They have been recently associated with new functions and described as deregulated in many cancers. Chronic lymphocytic leukemia (CLL) is the most frequent leukemia in the western world. This disease has a slow progression rate but is still incurable and is also very heterogeneous in terms of clinical course and response to therapy. Thus, it is essential to find new molecular markers allowing improvement of patient therapeutic care. This study aimed at establishing the expression profiles of snoRNAs in a CLL cohort and to correlate them to the clinico-biological parameters. By means of high-throughput quantitative PCR, I showed that snoRNAs were deregulated in CLL. Moreover, a specific signature was able to define a new adverse prognostic subgroup among IGHV-mutated patients, initially classified as good prognosis cases. The overexpression of the signature is correlated to a shorter treatment-free survival and seems to be mainly activated by proliferation signals. All in all, this study demonstrates the prognostic value of snoRNAs in CLL and prompts us to further explore their deregulation in hematological malignancies.

B-cell chronic lymphocytic leukemia (B-CLL) is the most common leukemia among the adult population in western countries and accounts for 30-40% of all leukemias. With survival time ranging from months to decades, the clinical course of individual CLL patients is highly variable. This heterogeneity and in the end the need for means to identify the patients with less favorable disease has encouraged the search for biomarkers that can predict the prognosis. Telomeres are repetitive structures protecting the chromosomal endings and shorten at each cell division. Telomere length (TL) has been indicated as a prognostic factor both in hematological malignancies and solid tumors. In B-CLL, TL is associated with mutation status of the immunoglobulin heavy chain variable (IGHV) gene and with clinical course. In the present thesis the main aim was to evaluate TL as a biomarker in B-CLL using a quantitative PCR-based method for TL determination. In paper I, TL was shown to be a prognostic factor for stage A and stage B/C patients, whereas IGHV mutation status predicted outcome only in stage A patients. Moreover, IGHV mutated CLL cases were subdivided by TL into two groups with different prognosis, a subdivision not seen for unmutated cases. Interestingly, the IGHV-mutated group with short telomeres had en overall survival close to that of the unmutated cases. Thus, a combination of IGHV mutation status and telomere length gave an improved subclassification of CLL identifying previously unrecognized patient groups with different outcomes. TL correlates with cellular origin of B-cell malignancies in relation to the germinal center (GC). In paper II different B-cell lymphoma/leukemia subtypes were analyzed. Shortest telomeres were found in IGHV unmutated CLLs, differing significantly from IGHV mutated cases. Contrary to this, mantle cell lymphomas (MCL) demonstrated similar TL regardless of IGHV mutation status. TL differed significantly between GC-like and non-GC-like diffuse large B-cell lymphomas (DLBCL) and follicular lymphomas (FL) had shorter telomeres than GC-like DLBCL. Hairy cell leukemias, which display Ig gene intraclonal heterogeneity, had longer telomeres than FLs and non-GC-DLBCL, but shorter than GC-DLBCL. In conclusion, TL seemed not to simply correlate with GC origin. Paper III presents a B-CLL cohort assessed for TL, genomic aberrations, IGHV mutation status, CD38 and ZAP-70 expression. An inverse correlation existed between TL and IGHV homology, CD38 and ZAP-70 expression. The presence of genomic aberrations was similar among patients regardless of TL. In contrast, 13q deletion, a favorable biomarker, was more frequent in patients with long telomeres, while 11q and 17p deletions (markers of less favorable outcome) were more frequent in the subgroup with short telomeres. In paper IV a large group of mainly indolent CLL cases from a population based cohort was studied again showing an association between TL and prognosis, especially in “good” prognosis cases as defined by other biomarkers. Multivariate analysis indicated a strong connection between IGHV mutation status, lipoprotein lipase (LPL) expression and TL. A comparison of TL in diagnostic and follow up samples demonstrated a significant correlation, and also in the follow samples TL constituted a significant biomarker for survival.