ABSTRACT INTRODUCTION: Cystic fibrosis (CF) is an autosomal recessive disease caused by mutations in the CFTR gene. The gene mutation profile is extremely heterogeneous and mutations show a variable distribution among population groups. In SA the 3120+1G->A splice site mutation has been found predominantly in Black and Coloured patients. It occurs in Black CF patients at an estimated frequency of 46%. The CF carrier frequency is estimated at 1/34 in Black and 1/55 in Coloured populations, and based on these rates, it is clear that a significant number of Black and Coloured patients remain undiagnosed. Point mutations account for the majority of the mutations that have been found in the CFTR gene. Copy number mutations are, however, increasingly being detected in CF patients through the use of gene dosage-dependant assays. These mutations have been found to occur in the CFTR gene in various African American families and exon rearrangements are thought to account for 1.3% of all CF chromosomes across all populations. AIMS: To use haplotypes to analyse the origin(s) of the 3120+1G->A mutation and the likely frequencies of the remaining unknown mutations. To increase mutation detection in the SA Black and Coloured groups by searching for CFTR gene exons for copy number mutations. METHODS: In patients with at least one copy of the 3120+1G>A mutation haplotype studies will be used to elucidate the origin(s) of this mutation in SA Black and Coloured CF patients, by analyzing pyrosequencing SNP genotype data. In patients with at least one unknown mutation, haplotype studies will reveal the likely relative frequencies of the unknown mutations in these populations. In Black and Coloured CF patients with at least one unknown mutation, a multiplex ligation dependant probe amplification (MLPA) CF kit will be used for the detection of exon copy number mutations. RESULTS: The results of the haplotype data show that there is a G-G-C-G-T-A haplotype, for markers MetD-KM19-J44-T854T-Tub18-J32, associated with the 3120+1G->A mutation in both Black and Coloured patients. Unknown mutation-associated haplotypes indicate that there are two relatively common unknown mutations in each of these populations. MLPA results show that one patient is a carrier of an exon 2 deletion. CONCLUSION: A single origin for the 3120+1G>A mutation in Black and Coloured CF patients is supported by the data. Exon copy number changes in the CFTR gene are not a major mutational mechanism leading to CF in SA Black and Coloured patients.

付記する学位プログラム名: 充実した健康長寿社会を築く総合医療開発リーダー育成プログラム / 京都大学 / 0048 / 新制・課程博士 / 博士(医科学) / 甲第22747号 / 医科博第116号 / 新制||医科||8(附属図書館) / 京都大学大学院医学研究科医科学専攻 / (主査)教授 長船 健二, 教授 武藤 学, 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

Indiana University-Purdue University Indianapolis (IUPUI) / This dissertation presents a body of research that attempts to tackle the ‘overfitting’ problem for gene signature and biomarker development in two different aspects (mechanistically and computationally). In achievement of a deeper understanding of cancer molecular mechanisms, this study presents new approaches to derive gene signatures for various biological phenotypes, including breast cancer, in the context of well-defined and mechanistically associated biological pathways. We identified the pattern of gene expression in the cell cycle pathway can indeed serve as a powerful biomarker for breast cancer prognosis. We further built a predictive model for prognosis based on the cell cycle gene signature, and found our model to be more accurate than the Amsterdam 70-gene signature when tested with multiple gene expression datasets generated from several patient populations. Aside from demonstrating the effectiveness of dimensionality reduction, phenotypic dissection, and prognostic or diagnostic prediction, this approach also provides an alternative to the current methodology of identifying gene expression markers that links to biological mechanism. This dissertation also presents the development of a novel feature selection algorithm called Predictive Power Estimate Analysis (PPEA) to computationally tackle on overfitting. The algorithm iteratively apply a two-way bootstrapping procedure to estimate predictive power of each individual gene, and make it possible to construct a predictive model from a much smaller set of genes with the highest predictive power. Using DrugMatrix™ rat liver data, we identified genomic biomarkers of hepatic specific injury for inflammation, cell death, and bile duct hyperplasia. We demonstrated that the signature genes were mechanistically related to the phenotype the signature intended to predict (e.g. 17 out of top 20 genes for inflammation selected by PPEA were members of NF-kB pathway, which is a key pre-inflammatory pathway for a xenobiotic response). The top 4 gene signature for BDH has been further validated by QPCR in a toxicology lab. This is important because our results suggest that the PPEA model not largely deters the over-fitting problem, but also has the capability to elucidate mechanism(s) of drug action and / or of toxicity.

Contexte. Les phénotypes ABO et Rh(D) des donneurs de sang ainsi que des patients transfusés sont analysés de façon routinière pour assurer une complète compatibilité. Ces analyses sont accomplies par agglutination suite à une réaction anticorps-antigènes. Cependant, pour des questions de coûts et de temps d’analyses faramineux, les dons de sang ne sont pas testés sur une base routinière pour les antigènes mineurs du sang. Cette lacune peut résulter à une allo-immunisation des patients receveurs contre un ou plusieurs antigènes mineurs et ainsi amener des sévères complications pour de futures transfusions. Plan d’étude et Méthodes. Pour ainsi aborder le problème, nous avons produit un panel génétique basé sur la technologie « GenomeLab _SNPstream» de Beckman Coulter, dans l’optique d’analyser simultanément 22 antigènes mineurs du sang. La source d’ADN provient des globules blancs des patients préalablement isolés sur papiers FTA. Résultats. Les résultats démontrent que le taux de discordance des génotypes, mesuré par la corrélation des résultats de génotypage venant des deux directions de l’ADN, ainsi que le taux d’échec de génotypage sont très bas (0,1%). Également, la corrélation entre les résultats de phénotypes prédit par génotypage et les phénotypes réels obtenus par sérologie des globules rouges et plaquettes sanguines, varient entre 97% et 100%. Les erreurs expérimentales ou encore de traitement des bases de données ainsi que de rares polymorphismes influençant la conformation des antigènes, pourraient expliquer les différences de résultats. Cependant, compte tenu du fait que les résultats de phénotypages obtenus par génotypes seront toujours co-vérifiés avant toute transfusion sanguine par les technologies standards approuvés par les instances gouvernementales, les taux de corrélation obtenus sont de loin supérieurs aux critères de succès attendus pour le projet. Conclusion. Le profilage génétique des antigènes mineurs du sang permettra de créer une banque informatique centralisée des phénotypes des donneurs, permettant ainsi aux banques de sang de rapidement retrouver les profiles compatibles entre les donneurs et les receveurs. / Background. ABO and Rh(D) phenotyping of both blood donors and transfused patients is routinely performed by blood banks to ensure compatibility. These analyses are done by antibody-based agglutination assays. However, blood is not routinely tested for minor blood group antigens on a regular basis because of cost and time constraints. This can result in alloimmunization of the patient against one or more minor antigens and may complicate future transfusions. Study design and Methods. To address this problem, we have generated an assay on the GenomeLab SNPstream genotyping system (Beckman Coulter, Fullerton, CA) to simultaneously test polymorphisms linked to 22 different blood antigens using donor’s DNA isolated from minute amounts of white blood cells. Results. The results showed that both the error rate of the assay, as measured by the strand concordance rate, and the no-call rate were very low (0.1%). The concordance rate with the actual red blood cell and platelet serology data varied from 97 to 100%. Experimental or database errors as well as rare polymorphisms contributing to antigen conformation could explain the observed differences. However, these rates are well above requirements since phenotyping and cross-matching will always be performed prior to transfusion. Conclusion. Molecular profiling of blood donors for minor red blood cell and platelet antigens will give blood banks instant access to many different compatible donors through the set-up of a centralized data storage system.

Contexte. Les phénotypes ABO et Rh(D) des donneurs de sang ainsi que des patients transfusés sont analysés de façon routinière pour assurer une complète compatibilité. Ces analyses sont accomplies par agglutination suite à une réaction anticorps-antigènes. Cependant, pour des questions de coûts et de temps d’analyses faramineux, les dons de sang ne sont pas testés sur une base routinière pour les antigènes mineurs du sang. Cette lacune peut résulter à une allo-immunisation des patients receveurs contre un ou plusieurs antigènes mineurs et ainsi amener des sévères complications pour de futures transfusions. Plan d’étude et Méthodes. Pour ainsi aborder le problème, nous avons produit un panel génétique basé sur la technologie « GenomeLab _SNPstream» de Beckman Coulter, dans l’optique d’analyser simultanément 22 antigènes mineurs du sang. La source d’ADN provient des globules blancs des patients préalablement isolés sur papiers FTA. Résultats. Les résultats démontrent que le taux de discordance des génotypes, mesuré par la corrélation des résultats de génotypage venant des deux directions de l’ADN, ainsi que le taux d’échec de génotypage sont très bas (0,1%). Également, la corrélation entre les résultats de phénotypes prédit par génotypage et les phénotypes réels obtenus par sérologie des globules rouges et plaquettes sanguines, varient entre 97% et 100%. Les erreurs expérimentales ou encore de traitement des bases de données ainsi que de rares polymorphismes influençant la conformation des antigènes, pourraient expliquer les différences de résultats. Cependant, compte tenu du fait que les résultats de phénotypages obtenus par génotypes seront toujours co-vérifiés avant toute transfusion sanguine par les technologies standards approuvés par les instances gouvernementales, les taux de corrélation obtenus sont de loin supérieurs aux critères de succès attendus pour le projet. Conclusion. Le profilage génétique des antigènes mineurs du sang permettra de créer une banque informatique centralisée des phénotypes des donneurs, permettant ainsi aux banques de sang de rapidement retrouver les profiles compatibles entre les donneurs et les receveurs. / Background. ABO and Rh(D) phenotyping of both blood donors and transfused patients is routinely performed by blood banks to ensure compatibility. These analyses are done by antibody-based agglutination assays. However, blood is not routinely tested for minor blood group antigens on a regular basis because of cost and time constraints. This can result in alloimmunization of the patient against one or more minor antigens and may complicate future transfusions. Study design and Methods. To address this problem, we have generated an assay on the GenomeLab SNPstream genotyping system (Beckman Coulter, Fullerton, CA) to simultaneously test polymorphisms linked to 22 different blood antigens using donor’s DNA isolated from minute amounts of white blood cells. Results. The results showed that both the error rate of the assay, as measured by the strand concordance rate, and the no-call rate were very low (0.1%). The concordance rate with the actual red blood cell and platelet serology data varied from 97 to 100%. Experimental or database errors as well as rare polymorphisms contributing to antigen conformation could explain the observed differences. However, these rates are well above requirements since phenotyping and cross-matching will always be performed prior to transfusion. Conclusion. Molecular profiling of blood donors for minor red blood cell and platelet antigens will give blood banks instant access to many different compatible donors through the set-up of a centralized data storage system.

1. Despite extensive research on oral squamous cell carcinoma (OSCC), its five-year survival rate has not improved for the last two decades. Effective treatment of OSCC requires the identification of molecular targets to design appropriate therapeutic strategies. To this end, DDRT-PCR analysis was used to identify molecular markers, which could be used as therapeutic targets. 2. DDRT-PCR in combination with reverse Northern analysis identified 25 differentially expressed genes in oral tumors. Fourteen genes did not show homology to any known gene in the database and therefore may represent non-specific genomic DNA sequences or novel genes that have not yet been identified. The remaining 11 genes showed homology to known genes such as DIAPH1, NJMU-R1, RBM28, PCNA, GLTP, MTATP6, ZKSCAN1, TNKS2, PAM, TUBB2C and C14orf154. TNKS2, PAM, TUBB2C and C14orf154 showed downregulation and the remaining seven genes were upregulated in oral tumor samples. 3. To reconfirm the results of DDRT-PCR and reverse Northern blot analyses, Northern blot analysis was carried out on matched normal and tumor samples for a few genes. As expected, PCNA, NJMU-R1 and ZKSCAN1 showed upregulation, whereas TUBB2C showed downregulation in the tumor sample. PCNA was also found to be upregulated in tumor samples at the protein level. 4. The expression of eight differentially expressed genes (viz., DIAPH1, NJMU-R1, RBM28, PCNA, GLTP, TNKS2, PAM and TUBB2C) was also validated in a panel of 16 matched normal and tumor samples. The mean mRNA expression levels of GLTP, PCNA, RBM28, NJMU-R1 and DIAPH1 were significantly greater in tumor samples than in normal samples. The mean expression levels of TNKS2, PAM and TUBB2C were significantly lower in tumor samples than in normal samples. 5. As some of the genes like NJMU-R1, RBM28, GLTP and PAM are found to differentially regulated in a majority of the tumors, they could be used as potential markers in oral cancer. 6. Tuberin and hamartin have been placed as a complex in the insulin signaling pathway and are known to negatively regulate this pathway. Since overexpression of TSC2 has been previously shown to exert antitumor effect on two oral cancer cell lines, and some components of the insulin signaling pathway have already been implicated in head and neck cancers, we reasoned that both TSC genes and other key players of this pathway might be differentially regulated in oral tumors. Northern blot analysis showed downregulation of the TSC2 gene in an oral tumor sample. In order to further validate the expression pattern of the TSC2 gene, a semiquantative RT-PCR analysis was carried out in a panel of 16 matched normal and tumor samples. The mean expression level of TSC2 was significantly lower in tumor samples than in normal tissue samples. The mean expression level of its interacting partner TSC1 was also significantly lower in tumor samples than in normal tissue samples, suggesting the involvement of these genes in the etiology of oral cancer. TSC1 and TSC2 were also downregulated in eight matched normal and tumor samples at the protein level. We wanted further to determine the expression of both TSC genes in cell lines. Interestingly, TSC2 did not show a detectable level of expression in an oral cancer cell line SCC 131, whereas it was expressed in two other oral cancer cell lines KB and SCC 104 as well as in four non-oral cell lines: A549, HEK-293T, HeLa and HepG2 at the protein level. The TSC2 expression in KB was, however, lower than in other cell lines. TSC1 was expressed in all the cell lines, albeit at different levels. The TSC1 expression was lower in SCC 131 as compared to two other cell lines KB and SCC 104. 7. Given the fact that both are tumor suppressors, it was hypothesized that LOH, inactivating somatic mutations and/or promoter methylation might be playing a role for their downregulation in oral tumors. Mutation analysis of all the coding regions of both the TSC genes failed to detect any mutation in a panel of 25 tumor samples. However, seven normal population variants were identified in different patients. Our analysis of the matched peripheral blood and tumor DNA samples from 52 patients showed LOH at both the TSC loci. At the TSC1 locus, 17/48 (35.42%) tumors showed an allelic loss for one or more markers. At the TSC2 locus, LOH was found in 18/48 (37.5%) informative cases. Nine patients (9/48, 18.75%) had LOH at both the TSC loci. Since PTEN is another tumor suppressor in the insulin signaling pathway, we then sought to determine if LOH is also present in the PTEN candidate region in a panel of 50 matched samples. Microsatellite analysis using three markers showed a low LOH rate of 13% in tumor samples. 8. As the OSCC cell line SCC 131 did not show a detectable level of TSC2 expression, we treated this cell line with methylation inhibition drug 5-azacytidine. The treatment restored the expression of TSC2 and increased the expression of TSC1, suggesting that the promoter methylation and LOH are the important mechanisms for their downregulation. In order to see if the downregulation of the TSC genes is due to their promoters being methylated in tumors from the patients, we examined the methylation status of their promoters in 16 oral tumors, three normal oral tissues, two peripheral blood DNA samples from normal individuals and two cell lines HeLa and SCC 131 by COBRA. Our repeated efforts to amplify the TSC1 promoter using different DNA polymerases failed. However, we were able to successfully amplify the 571 bp long TSC2 promoter. Our analysis showed methylation of the TSC2 promoter in all tumors and two cell lines. As expected, the TSC2 promoter was not methylated in normal oral tissues and control blood DNA samples. Our bisulfite sequencing data suggested a low level and a considerable heterogeneity of methylation. 9. Using Fisher’s exact test, no correlation was found between LOH at the TSC loci and different clinical parameters such as age, sex, T classification, stage, grade, histology, tobacco habits and lymph node metastasis. 10. Using Fisher’s exact test, no correlation was found between the TSC2 promoter methylation and its downregulation in 16 tumor samples. We believe that this could be due to small sample size. 11. Since TSC1 and TSC2 are important regulators of the insulin pathway, it was hypothesized that other key players of this pathway might also be dysregulated in oral cancer. To this end, the expression pattern of some of the major regulators of the insulin pathway (viz., PI3K, AKT, PDK1, RHEB, mTOR, S6K1, S6, eIF4E, 4E-BP1, PTEN, 14-3-3ﾟ and IRS1) was investigated using semiquantative RT-PCR in a panel of 16 matched normal and tumor samples. The mean expression levels of the following genes showed significant upregulation in tumor samples: AKT, PI3K, PDK1, RHEB, mTOR, S6K1, S6 and eIF4E. On the other hand, 4E-BP1 and PTEN showed significant downregulation in tumor tissues. No significant difference in the expression was found for 14-3-3ﾟ and IRS1 between tumor and normal tissues. The expression pattern of some of these genes was also analyzed at the protein level using Western blot analysis and eight matched normal and tumor tissues. The level of total AKT was upregulated in 2/8 tumor samples only. However, phosphorylated-AKT (Thr308) showed upregulation in 6/8 samples. p70S6K1 and phosphorylated-p70S6K1 (Thr389) were upregulated in 8/8 and 6/8 tumor samples, respectively. Increase in the phosphorylated forms of both AKT and its downstream effector p70S6K1 suggested an increase in their kinase activity, indicating a constitutive activation of this pathway in oral cancer. 12. Based on our findings of mutation analysis, LOH study, 5-azacytidine treatment of an oral cancer cell line and COBRA analysis, we suggest that LOH at the TSC gene loci and promoter methylation are important mechanisms for the downregulation of the TSC genes. Loss of function of these genes may thus contribute to the constitutive activation of the insulin signaling pathway in oral cancer, leading to overall cell growth and proliferation. Our studies have shown that several key members of this pathway show aberrant expression in a subset of cancers of the oral cavity and can provide useful therapeutic targets. Several inhibitors of the insulin signaling pathway, such as rapamycin and its derivatives which inhibit mTOR and the PI3K inhibitor wortmannin, are now being actively evaluated for clinical trials for other cancers. We suggest that these inhibitors could also be evaluated for the treatment of oral cancer in future. Our differential display analysis has served to identify several genes that may be important for the onset and progression of oral cancer. Further analysis of these genes is warranted.

Submitted by Diogo Misoguti (diogo.misoguti@gmail.com) on 2016-07-08T18:20:15Z No. of bitstreams: 1 cp124617.pdf: 1851616 bytes, checksum: 98299c340bd6b09b83e3fdf34f6adcf5 (MD5) / Approved for entry into archive by Diogo Misoguti (diogo.misoguti@gmail.com) on 2016-07-08T18:20:49Z (GMT) No. of bitstreams: 1 cp124617.pdf: 1851616 bytes, checksum: 98299c340bd6b09b83e3fdf34f6adcf5 (MD5) / Made available in DSpace on 2016-07-08T18:20:49Z (GMT). No. of bitstreams: 1 cp124617.pdf: 1851616 bytes, checksum: 98299c340bd6b09b83e3fdf34f6adcf5 (MD5) Previous issue date: 2009 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Associação Fundo de Incentivo à Psicofarmacologia (AFIP) / Introdução: Transplante alogênico de células tronco hematopoéticas (TCTH) é uma importante terapia para doenças hematológicas, mas o sucesso de uma porção de transplantes é limitado pela doença do enxerto versus hospedeiro(GVHD). Os fatores de risco conhecidos para GVHD agudo (aGVHD) não fornecem uma estimativa precisa do risco individual e não auxiliam na individualização da terapia. Até o momento, não existe método diagnóstico que permita predizer aGVHD. A identificação de pacientes que desenvolverão aGVHD poderia permitir a individualização da terapia para uns e evitaria imunossupressão intensa para outros. Objetivos: Revelar um classificador molecular preditivo de aGVHD. Descobrir genes diferencialmente expressos e analisar eventos precoces que desencadeiam aGVHD. Explorar categorias funcionais e tipos celulares relacionados ao desenvolvimento de aGVHD. Casuística e métodos: Foram isolados, amplificados, marcados e co-hidridados com lâminas de microarray contendo 22.000 sondas, amostras de RNA mensageiro de 89 pacientes submetidos a TCTH HLA-idêntico mieloablativo ou de toxicidade reduzida, obtido de células mononucleares periféricas durante a enxertia medular. Os pacientes foram divididos em grupo treino e grupo teste, um modo utilizado para construir um modelo que discrimine pacientes com e sem aGVHD, e para testar o modelo em amostras independentes. Os genes informativos foram selecionados utilizando recursive feature elimination, seguido de sete diferentes algoritmos de classificação multivariados para estabelecer o classificador molecular no grupo treino. Os genes diferencialmente expressos entre amostras de pacientes com e sem GVHD foram submetidos a análise de enriquecimento das vias funcionais e agrupados de acordo com o perfil de expressão com células e tecidos através do SymAtlas. Resultados: Encontramos um classificador molecular composto de 233 genes nas amostras do grupo treino, que foram selecionados baseados na mais precisa classificação. No grupo teste da amostra, cerca de 80% dos pacientes puderam ser classificados. Para estes pacientes, o classificador mostrou uma acurácia preditiva de 75% (sensibilidade de 71% e especificidade de 78%). Analisando a anotação funcional dos genes diferencialmente expressos, observamos que em pacientes que desenvolveram aGVHD, houve aumento de expressao de genes da resposta antimicrobiana, transporte de gases, metabolismo de hemoglobina, além das alarminas. Nestas amostras também observamos a diminuição da expressão da IL1 e outros genes envolvidos na via do NF-kB. Vários genes super-expressos no periodo precoce do aGVHD foram associados a células precursoras. Conclusões: Nossos resultados mostram que um classificador molecular é capaz de identificar pacientes sob alto risco de desenvolver aGVHD. Estabelecer métodos de diagnóstico preditivo para aGVHD é o primeiro passo para a individualização da estratégia terapêutica após TCTH. Além disso, os resultados da análise de enriquecimento funcional e a expressão de genes em diferentes populações celulares, sugerem que eventos precoces envolvendo múltiplas populações de células precursoras possam predefinir a interação futura entre o enxerto em desenvolvimento e o paciente. / Background: Allogeneic hematopoietic stem cell transplantation is an important last resort therapy for hematological diseases. Unfortunately, the success of a large proportion of these transplants is limited by graft-versus-host disease (GVHD). Currently known risk factors for acute GVHD (histoincompatibility, sex mismatch, older patients, previous pregnancies) do not provide a precise estimate of individual patient risk and do not help for individualization of the therapy. Early identification of those patients who will develop aGVHD may allow for individualized treatment, and also for the reduction of unnecessary treatment for those patients not at risk. Nowadays, however, there is no diagnostic method that allows prediction of aGVHD. Objectives: The goal of our study was to reveal a gene expression profile that would predict the occurrence of aGVHD. In addition, using enrichment of gene ontology categories, to analyze differentially expressed genes in order to better understand biology of the events preceding aGVHD. Material and methods: we collected blood samples from 89 recipients of myeloablative and reduced conditioning regimen HLA-identical sibling allogeneic hematopoietic stem cell transplants at the time of successful engraftment. We isolated total RNA from the peripheral blood mononuclear cells, amplified it, labeled, and co-hybridized to the microarray slides containing probes for 22,000 genes..The patients were divided into training and test groups, the former used to build a model discriminating patients with and without aGVHD and the latter - to test the model on independent samples. We selected the informative genes using “recursive feature elimination” method followed by seven different multivariate classification algorithms in order to establish a molecular classifier in the training set. Then we validated this new classifier in the test set of patients.. We found differentially expressed genes using T-test and accepted those with estimated false discovery rate below 10%. We have used Biobase Explain to find enrichment of functional groups among differentially expressed genes.Results: We found a molecular classifier comprised by 233 gene probes in the training set of samples which were selected based on the most accurate classification. In the test group of samples, we found that 80% of patients could be classified based on the concordance between classification methods as described above. For these patients, the classifier showed 75% of a predictive accuracy (71% of sensitivity and 78% of specificity). Analysis of functional annotations of differentially expressed genes showed that patients that developed acute GVHD have increased expression of antimicrobial genes, hemaglobin metabolism genes and alarmins. In these samples we also observed decreased expression of IL-1 and other genes involved in NF-kB activation. Several genes up-regulated before aGVHD were associated with multiple types of precursor cells. Conclusion: Our results show that molecular profiling is able to identify patients under high risk of acute GVHD at the time of engraftment. Establishing of a predictive diagnostic method for aGVHD is the first step of individualization of therapeutic strategy after hematopoetic stem cell transplantation. In addition, the results of functional enrichment analysis and expression in different cell populations suggest that early events during engrafment involving precursor cell populations might predefine results of interaction between stem cell allograft and patient body.

Le domaine de l'oncologie progresse de manière rapide, surtout depuis l'avènement des thérapies ciblées. Parmi elles, les inhibiteurs de tyrosine kinase multicible (ITK) antiangiogéniques ont fait la preuve de leur efficacité dans plusieurs types de cancers métastatiques. Le sorafenib, le sunitinib, le pazopanib, l'axtinib, et le regorafenib sont aujourd'hui utilisés en pratique courante. Ces premières thérapeutiques ont ouvert la voie au développement de nombreuses autres molécules ciblant d'autres récepteurs TK (crizotinib, céritinib). Les ITK ciblant les récepteurs de l'angiogénèse inhibent des récepteurs membranaires tels que les VEGFR, les PDGFR, les FGFR etc. Ces molécules améliorent généralement survie et/ou survie sans progression dans les essais cliniques pivots mais il existe une grande variabilité interindividuelle en termes de bénéfice clinique. Nous avons cherché à mettre en évidence des biomarqueurs moléculaires prédictifs de la réponse, afin de mieux sélectionner les patients susceptibles de bénéficier de ces ITKs. L'objectif final de ce travail est ainsi de mieux sélectionner les patients candidats à ce traitement, mais il est également médico économique. La part la plus importante de ce travail est axée sur le regorafenib, qui a fait preuve de son efficacité dans les cancers colorectaux métastatiques prétraités et les tumeurs stromales gastro-intestinalesen échec d'imatinib et de sunitinib. Cependant, aucun paramètre clinique ou histologique n'a été identifié pour sélectionner les patients potentiels pouvant bénéficier de ce traitement, ou, à l'inverse pour éviter de traiter les patients chez lesquels la balance bénéfice/risque est défavorable. Ce travail a été réalisé dans le cadre de programme ProfiLER (NCT01774409) en partenariat avec les plateformes de génomique tumorale (Centre de Recherche en Cancérologie de Lyon, Centre Léon Bérard). Cette étude avait pour objectif de tester l'hypothèse que l'ensemble des altérations des gènes codant pour les kinases cibles d'un ITK donné pourrait être associé au bénéfice clinique de ce traitement. Dans notre étude, nous avons observé que les cancers des patients présentant un bénéfice clinique accumulent des gains chromosomiques sur les gènes cibles, et à l'inverse, les cancers des patients nonrépondeurs possèdent plutôt un profil inverse. L'index génomique, un paramètre évaluant l'instabilité chromosomique ne permet pas de différencier les patients répondeurs, mais nous avons mis en évidence que l'accumulation de certains gains sur les gènes cibles est associée à une meilleure survie. Nous avons ainsi proposé un nouveau concept : celui de TTC (Tumor Target Charge), la somme des gains sur les gènes cibles ; et à l'inverse, celui de TTL (Tumor Target Loss), la somme des pertes sur les gènes cibles. En nous appuyant sur ces définitions de TTC et TTL, nous avons généré un algorithme nommé SUMSCAN traduisant donc la somme des gains et des pertes sur les gènes cibles. Le score SUMSCAN a été appliqué à une première cohorte composée essentiellement de patients ayant un cancer colorectal métastatique et traités par regorafenib, ainsi qu'à une 2ème cohorte de validation composée des patients ayant différentes pathologies néoplasiques. Chez les patients ayant un cancer colorectal « moléculairement sélectionné», la médiane de survie sans progression était de 9 mois contre 3 mois dans la cohorte de patients non sélectionnés (X. JIANG et al, Oncotarget, 2015). Nous avons pu montrer que le principe de ce score pronostique pouvait s'appliquer aux autres antiangiogéniques multi-ITKs.. Nous sommes ainsi en cours de validation de ce score pour la prédiction de la survie sur de larges populations de patients présentant divers types tumoraux : sarcome des tissus mous, carcinomes ovariens de haut grade, carcinome rénal carcinome de la thyroïde, etc.) / Small molecule antiangiogenic tyrosine kinase inhibitors (TKI), such as regorafenib, sorafenib, sunitinib, pazopanib, axitinib, and cabozantinib, are active in a variety of advanced cancers, including renal cell carcinoma (RCC), gastrointestinal stromal tumors (GIST), hepatocellular carcinoma (HCC), colorectal cancer (CRC) and thyroid cancers. Predictive criteria for response to these multiple kinase inhibitors (MTKI) are not as well determined as for tumors harboring key driver alterations, such as BCR-ABL translocations in chronic myeloid leukemia (CML), KIT-mutant GIST, BRAF-mutant melanoma, and ALK-positive non-small cell lung cancer among others. Regorafenib, for instance, has been shown to yield a progression-free survival (PFS) improvement in pretreated metastatic colorectal cancer (mCRC) and in imatinib and sunitinib refractory gastrointestinal stromal tumors (GIST). We report that the antitumor activity of MTKIs in tumors lacking a well-defined oncogenic driver is strongly correlated with copy number alterations of genes encoding the protein kinases targeted by these drugs. A concept of tumor target charge (TTC), defined as the total gains of the genes encoding for targets of MTKIs as well as tumor target loss (TTL) was developed, and correlated to response to regorafenib in 2 cohorts of patients composed of mCRC and STS patients. A predictive model, called SUMSCAN, was conceived as a binary classifier to identify patients as either good or poor candidates for use of MTKIs. Moreover, the PFS and OS of patients with a favorable SUMSCAN score were significantly improved. Importantly, SUMSCAN predicted exclusively response to regorafenib, but not the response to conventional chemotherapy in mCRC