No description available.

Next-generation sequencing (NGS) techniques have led to major breakthroughs in the characterization of the chronic lymphocytic leukemia (CLL) genome with discovery of recurrent mutations of potential prognostic and/or predictive relevance. However, before NGS can be introduced into clinical practice, the precision of the techniques needs to be studied in better detail. Furthermore, much remains unknown about the genetic mechanisms leading to aggressive disease and resistance to treatment. Hence, in Paper I, the technical performance of a targeted deep sequencing panel including 9 genes was evaluated in 188 CLL patients. We were able to validate 143/155 (92%) selected mutations through Sanger sequencing and 77/82 mutations were concordant in a second targeted sequencing run, indicating that the technique can be introduced in clinical practice. In Paper II we screened 18 NF-κB pathway genes in 315 CLL patients through targeted deep sequencing which revealed a recurrent 4 base-pair deletion in the NFKBIE gene. Screening of NFKBIE in 377 additional cases identified the mutation in ~6% of all CLL patients. We demonstrate that the lesion lead to aberrant NF-κB signaling through impaired interaction with p65 and is associated with unfavorable clinical outcome. In Paper III we sought to delineate the genetic lesions that leads to relapse after fludarabine, cyclophosphamide, and rituximab treatment. Through whole-exome sequencing of pre-treatment and relapse samples from 41 cases we found evidence of frequent selection of subclones harboring driver mutations and subsequent clonal evolution following treatment. We also detected mutations in the ribosomal protein RPS15 in 8 cases (19.5%) and characterization of the mutations through functional assays point to impaired p53 regulation in cells with mutated RPS15. Paper IV aimed at characterizing 70 patients assigned to three major subsets (#1, #2, and #4) through whole-genome sequencing. Besides recurrent exonic driver mutations, we report non-coding regions significantly enriched for mutations in subset #1 and #2 that may facilitate future molecular studies. Collectively, this thesis supports the potential of targeted sequencing for mutational screening of CLL in clinical practice, provides novel insight into the pathobiology of aggressive CLL, and demonstrates the clinical outcome and cellular effects of NFKBIE and RPS15 mutations.

Neoadjuvant chemotherapy has become standard practice in patients with high-risk early breast cancer as it improves rates of breast conservation surgery and enables prediction of recurrence and survival by using response to treatment as a surrogate. Previous studies have focused on generating molecular datasets to develop prediction models of response, though little is known on how tumours and their microenvironments are modulated by neoadjuvant chemotherapy. The thesis aims at molecularly characterising tumour changes during neoadjuvant chemotherapy in a cohort of 168 patients. Serial tumour samples at diagnosis, and, when available, midway through chemotherapy and on completion of treatment were profiled by shallow whole genome sequencing, deep exome sequencing and transcriptome sequencing, resulting in the generation of an unprecedented genomics dataset with tumours in situ while patients received chemotherapy. Molecular predictors of response to chemotherapy were inferred from the diagnostic biopsy. Several novel observations were made, including previously undescribed associations between copy number alterations, mutational genotypes, neoantigen load, HLA genotypes and intra-tumoural heterogeneity with chemosensitivity. Possible mechanisms of chemoresistance included LOH at the MHC Class I locus, decreased expression of MHC Class I and II genes and drug influx molecules, as well as increased expression of drug efflux pumps. A complex relationship between proliferation, tumour microenvironment composition (TME) and response to treatment was explored by deconvoluting bulk RNAseq data and performing digital pathology orthogonal validation. Clonal and microenvironment dynamic changes induced by/associated with chemotherapy were then modelled. Two types of genomic responses were identified, one in which the clonal composition was stable throughout treatment and another where clonal emergence and/or extinction was evident. Validation by multi-region deep sequencing confirmed the dynamics of the clonal landscape. Clonal emergence was shown to be associated with higher proliferation and decreased immune infiltrate, with an increase in genomic instability and homologous recombination deficiency during treatment. The immune TME composition and activity mirrored response to treatment, with cytolytic activity and innate and adaptive immune infiltrates linearly correlating with the degree of residual disease remaining after chemotherapy. Finally, the circulating tumour DNA (ctDNA) genomic landscape was explored by using shallow whole genome sequencing and targeted sequencing of plasma DNA. Tumour mutations detected on exome sequencing were also detected in ctDNA in plasma, supporting the use of liquid biopsies as a biomarker for monitoring response to therapy and detection of minimal residual disease.

Antigens play a critical role in the development of chronic lymphocytic leukemia (CLL) by binding to and stimulating leukemic precursor cells at some point during CLL ontogeny. Nevertheless, much remains unknown and further studies are necessary before an accurate model of antigen-drive can be ascertained. In this context, intraclonal diversification (ID) analysis of immunoglobulin (IG) genes could shed light on whether antigen involvement is restricted to the malignant transformation phase or if the triggering antigen(s) continuously stimulates the CLL clone. Hence, in Paper I we conducted a large-scale analysis of 71 CLL cases and revealed that 28/71 cases carried intraclonally diversified IGHV-IGHD-IGHJ genes. Although most cases showed no or low levels of ID, intense ID was evident within all subset #4 (IGHV4-34/IGKV2-30) cases. Subsequent analysis, in Paper II, of the clonotypic light chains revealed that the outstanding exception again related to subset #4. In such cases, the expressed IGKV2-30 gene was affected by targeted ID, analogous to their partner IGHV4-34 gene. Whilst these results convincingly argued for the role of antigen(s) in the development and evolution of CLL subset #4, this analysis was limited to depicting what was occurring at a single time-point and could not provide insight into the temporal dynamics of the CLL clones. Thus, in Paper III we conducted a longitudinal study of 8 subset #4 cases which enabled us to establish a hierarchical pattern of subclonal evolution. The observed ‘stepwise’ accumulation of mutations strongly supports a role for antigen selection in the pathogenesis of CLL subset #4. In Paper IV we reported a subset of IgG-switched CLL patients with coexisting trisomies of 12 and 19, and propose that the emergence of trisomy 18 in such cases represents a clonal evolution event suggestive of selection due to a clonal advantage. Paper V focused on the IGHV3-21 gene, an adverse prognostic factor in CLL. Since ~60% of IGHV3-21-expressing cases carry stereotyped B cell receptors, recognition of a common antigenic epitope, perhaps of pathogenic significance, is envisaged. Therefore, we investigated IGHV3-21 gene frequency within a Swedish population-based cohort and assessed the impact of stereotypy on clinical outcome. Taken collectively, this thesis provides molecular and genetic evidence for the role of antigen in CLL pathogenesis by convincingly demonstrating that clonal evolution, at least for certain subsets of CLL, is functionally driven rather than a consequence of clonal expansion promoted by nonspecific stimuli.

京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第22118号 / 医博第4531号 / 新制||医||1039(附属図書館) / 京都大学大学院医学研究科医学専攻 / (主査)教授 松田 道行, 教授 小柳 義夫, 教授 小川 誠司 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

Le neuroblastome est le cancer solide extra-cranial le plus fréquent chez l’enfant. Il est caractérisé par une très grande hétérogénéité tant au niveau clinique que moléculaire. Alors que certains patients rentrent spontanément en rémission, on peut se demander quels facteurs permettent la réémergence du cancer chez d’autres malgré traitement. Pour répondre à cette question, il convient d’identifier chez les patients ayant rechuté, les différentes populations clonales coexistant au diagnostic et/ou à la rechute. Cela permet, entre autre, d’étudier les voies différemment altérées entre ces deux temps. Dans cette optique, nous présentons ici QuantumClone, un algorithme de reconstruction clonal à partir de données de séquençage, ainsi que son application à une cohorte de patients souffrant d’un neuroblastome. Sur ces données, l’application de notre méthode a permis d’identifier des différences dans le ratio de variants prédits fonctionnels par rapport à ceux prédits passagers entre les populations ancestrales, enrichies à la rechute ou appauvries à la rechute. / Neuroblastoma is the most frequent solid extra-cranial cancer of childhood. This cancer displays a high heterogeneity both at clinical and molecular levels. Even though in some patients spontaneous remission can be observed, some others relapse despite treatment and surgical resection. It may be wondered which are the factors that distinguish these two cases. In order to answer this question, identification of populations coexisting at diagnosis and/or relapse in the patients which have relapsed is a prerequisite. This would allow, between other things, to study the pathways differently altered in clones that are specific to each time point. With this in mind, we hereby present QuantumClone, a clonal reconstruction algorithm from sequencing data. In addition, we applied this method to a cohort of patients suffering from neuroblastoma. On these data, our method identified differences in the functional mutation rate, i.e. the number of putative functional variants by total number of variants, between the ancestral clones, clones expanding at relapse, and clones shrinking at relapse.

Evolution is an ubiquitous feature of living systems. The genetic composition of a population changes in response to the primary evolutionary forces: mutation, selection and genetic drift. Organisms undergoing rapid adaptation acquire multiple mutations that are physically linked in the genome, so their fates are mutually dependent and selection only acts on these loci in their entirety. This aspect has been largely overlooked in the study of asexual or somatic evolution and plays a major role in the evolution of bacterial and viral infections and cancer. In this thesis, we put forward a theoretical description for a minimal model of evolutionary dynamics to identify driver mutations, which carry a large positive fitness effect, among passenger mutations that hitchhike on successful genomes. We examine the effect this mode of selection has on genomic patterns of variation to infer the location of driver mutations and estimate their selection coefficient from time series of mutation frequencies. We then present a probabilistic model to reconstruct genotypically distinct lineages in mixed cell populations from DNA sequencing. This method uses Hidden Markov Models for the deconvolution of genetically diverse populations and can be applied to clonal admixtures of genomes in any asexual population, from evolving pathogens to the somatic evolution of cancer. To understand the effects of selection on rapidly adapting populations, we constructed sequence ensembles in a recombinant library of budding yeast (S. cerevisiae). Using DNA sequencing, we characterised the directed evolution of these populations under selective inhibition of rate-limiting steps of the cell cycle. We observed recurrent patterns of adaptive mutations and characterised common mutational processes, but the spectrum of mutations at the molecular level remained stochastic. Finally, we investigated the effect of genetic variation on the fate of new mutations, which gives rise to complex evolutionary dynamics. We demonstrate that the fitness variance of the population can set a selective threshold on new mutations, setting a limit to the efficiency of selection. In summary, we combined statistical analyses of genomic sequences, mathematical models of evolutionary dynamics and experiments in molecular evolution to advance our understanding of rapid adaptation. Our results open new avenues in our understanding of population dynamics that can be translated to a range of biological systems.

Acute myeloid leukemia (AML) is a clonal neoplastic disorder organized as a cellular hierarchy, with the self-renewing leukemia stem cell (LSC) at the apex. Recurrent mutations in transcription factors (TF) and epigenetic regulators suggest that AML is driven by aberrant transcriptional circuits, but these circuits have not been fully defined in an LSC model. To study transcriptional mechanisms relevant to leukemogenesis in vivo, we generated a murine serial transplantation model of MLL-AF9-driven, myelomonocytic leukemia with genetically- and phenotypically-defined LSCs. Using this model, we pursued two related lines of investigation. First, we performed an in vivo RNA interference (RNAi) screen to identify transcription factors required for LSC function. This screen highlighted the circadian rhythm TFs, Clock and Bmal1, as genes essential for the survival of murine leukemia cells, and we validated this finding with CRISPR/Cas-based genome editing and knockdown studies in AML cell lines. Utilizing luciferase reporter mice to track expression of the circadian target gene Per2, we demonstrated that both leukemic and normal hematopoietic cells have the capacity for oscillating, circadian-dependent gene expression. Importantly, using murine knockout models, we found that normal hematopoietic stem and progenitor cells (HSPC), in contrast to leukemia cells, do not depend on Bmal1. We further demonstrated that selective depletion of LSCs following circadian perturbation is mediated through enhanced myeloid differentiation. ChIP-Seq studies revealed that the circadian rhythm network is integrally connected to the LSC self-renewal circuitry and highlighted putative Clock/Bmal1 targets in leukemia, providing a mechanistic basis for our findings. Second, we performed a functional and genomic characterization of our MLL-AF9 serial transplantation model to explore mechanisms of disease evolution and clonal selection in AML. Limiting dilution studies demonstrated that serial transplantation results in a reduction in disease latency, dramatic enrichment of leukemia-initiating cells (LIC), and reconfiguration of the LSC hierarchy. While mutations in known AML-associated genes were not linked to disease progression, RNA-sequencing (RNA-Seq) demonstrated that the increase in LIC frequency in serially transplanted leukemias is driven by changes in cell cycle and differentiation. In aggregate, these studies offer insights into the biological mechanisms regulating LSC self-renewal and disease evolution in AML.

Multiple Myeloma (MM) is the second most common hematological malignancy, and although patient outcomes have significantly improved since the introduction of autologous stem cell transplantation (ASCT) and novel pharmacological agents such as immunomodulators (IMID), proteasome inhibitors (PI), and monoclonal antibodies (mAb), the disease remains incurable. The pathological complexity of MM results from accumulating mutations in clonal populations of malignant B-cells, which are cytogenetically heterogenous and selectively sensitive to different therapeutic agents. Drug regimens therefore include a diverse combination of therapeutics designed to target specific pathways to inhibit cell proliferation. Recent advances in genomic analytics and novel pharmacological agents potentially allow for more targeted treatments which improve patient outcomes and frequency of remission with minimal adverse effects. Only recently have studies began to correlate an increased understanding of the many subtypes of MM with optimal treatment regimens, and practices such as “Direct to Drug” screening can give clinicians a look at a patient’s likely response to a combination of drugs. By incorporating emerging pharmaceutical agents into studies based on patient characteristics, the management of MM is making incremental strides towards a more targeted treatment paradigm.

T-cell prolymphocytic leukemia (T-PLL) is the most common mature T-cell leukemia. It is a typically aggressively growing and chemotherapy-resistant malignancy with a poor prognosis. T-PLL cells resemble activated, post-thymic T-lymphocytes with memorytype effector functions. Constitutive transcriptional activation of genes of the T-cell leukemia 1 (TCL1) family based on genomic inversions/translocations is recognized as a key event in T-PLL’s pathogenesis. TCL1’s multiple effector pathways include the enhancement of T-cell receptor (TCR) signals. New molecular dependencies around responses to DNA damage, including repair and apoptosis regulation, as well as alterations of cytokine and non-TCR activation signaling were identified as perturbed hallmark pathways within the past years. We currently witness these vulnerabilities to be interrogated in first pre-clinical concepts and initial clinical testing in relapsed/refractory TPLL patients. We summarize here the current knowledge on the molecular understanding of T-PLL’s pathobiology and critically assess the true translational progress around this to help appraisal by caregivers and patients. Overall, the contemporary concepts on T-PLL’s pathobiology are condensed in a comprehensive mechanistic disease model and promising interventional strategies derived from it are highlighted.