PCSK9 est un régulateur clé du métabolisme du cholestérol par le foie à travers la dégradation lysosomiale du récepteur aux LDL (low-density lipoprotein). Alors que les mutations gain de fonction (GOF) de PCSK9 induisent une hypercholestérolémie autosomique dominante, les mutations pertes de fonctions (LOF) entraînent un taux spontanément bas de LDL-cholestérol, ainsi qu’un protection cardiovasculaire. Du fait des limitations inhérentes aux modèles d’études, tels que les lignées cellulaires transfectées ou des animaux transgéniques, les fonctions de PCSK9 restent encore mal connues. Ainsi, nous avons utiliser des cellules souches pluripotentes induites (hiPSC) spécifiques de patients pour les différencier en hépatocytes et modéliser la physiopathologie liée aux mutations de PCSK9 GOF-S127R et LOF-R104C/V114A. Nous avons démontré que les hépatocytes obtenus récapitulaient la physiopathologie liés aux mutations de PCSK9. De plus, les cellules portant la mutation S127R ont montré une importante réponse au traitement par les statines, qui est corrélée à la réponse clinique des patients portant cette même mutation. Enfin, notre étude nous a permis de mettre à jour une fonction inattendue de PCSK9 dans les hiPSC et pendant leur différenciation. Elle montre que PCSK9 affecterait la prolifération des hiPSC ainsi qu’une voie de signalisation clé du développement régulée par NODAL. Cette régulation se ferait à travers une interaction directe entre PCSK9 et DACT2, un régulateur intracellulaire de la voie de signalisation de NODAL. En conclusion, les hiPSC s’avèrent être un modèle cellulaire translationnel pertinent pour mettre à jour de nouvelles fonctions hépatiques de PCSK9. / PCSK9 has been identified as a key regulator of cholesterol metabolism by the liver through inducing lysosomal degradation of the low-density lipoprotein receptor (LDLR). While PCSK9 gain-of-function (GOF) mutations induced autosomal dominant hypercholesterolemia and increased cardiovascular risk, loss-of-function (LOF) mutations are associated with low LDL-cholesterol levels and cardiovascular protection. Due to limitations inherent to current models including animal and human cells lines transfected with DNA constructs or transgenic animal models, PCSK9 functions are not fully understood. Therefore, we took advantage of patient related somatic cells reprogramming intoinduced pluripotent stem cells (hiPSC) to generate hepatocyte-like cells (HLC) and model the pathophysiology of PCSK9 mutations in dyslipidemia through focusing on two intracellular mutation forms; GOF (S127R) and LOF (R104C/V114A). We showed that HLC could recapitulate the key pathophysiological features of PCSK9 mutations. Moreover, HLC with the S127R mutation displayed an increased uptake of LDL upon statin treatment, which was correlated with the original patient clinical response. In parallel, this model enabled us to unravel a new unexpected role of PCSK9 in hiPSC and during differentiation. PCSK9 was found to affect the proliferation of hiPSC and regulate a key developmental signaling pathway mediated by NODAL. This regulation might occur by a direct interaction between PCSK9 and DACT2, an intracellular attenuator of NODAL signaling pathway. In conclusion, hiPSC provide a pertinent translational model to decipher PCSK9 hepatic functions and a novel cellular environment to highlight new functions.

京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第20605号 / 医博第4254号 / 新制||医||1023(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 川口 義弥, 教授 浅野 雅秀, 教授 中川 一路 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

The liver is a crucial human organ with a complex architecture. Although the liver has great regeneration potential, deadly liver diseases are associated with irreversible hepatocytes damage. Currently, a liver transplant is the only treatment for liver failure. A shortage of donors forced extensive research for alternative treatments. The most promising hepatocyte source could be obtained from the differentiation of induced pluripotent stem cells (iPSCs). This technology can give us great amounts of pluripotent cells without ethical restrictions, which could be available in a variety of haplotypes to minimize the possibility of rejection. There are many reprogramming protocols available. However, there is still no standardised method to obtain clinical grade iPSCs. From those stem cells, it is possible to obtain hepatic-like cells (HLCs) by direct differentiation in vitro. HLCs express multiple hepatocyte-specific features, but their names signal that they still show fetal liver identity. A variety of hepatic differentiation protocols were described, although the process of hepatic differentiation must be improved in order to be translated into the clinic. Along with genes, microRNA (miRNA) is the well-known controller of cell fate. MiRNA is a type of non-coding RNA (ncRNA) which can influence gene transcription by inhibiting gene expression. In contrast to genes, many of the miRNA can affect up to thousands of genes simultaneously. Another group of ncRNA, which is a subject of potential differences are small nucleolar RNA (snoRNA). SnoRNA are involved in RNA chemical modifications by acting as a guide, mostly for ribosomal RNA (rRNA), but some of them have additional functions. In this study, a new iPSCs line was generated from skin fibroblasts using lipotransfection of episomal vectors. This method is free from exogene integration and shows low cytotoxicity. A pluripotency of generated cells was confirmed by morphological assessment, immunocytochemical staining, and spontaneous differentiation assay. To be sure that the genome of the cells was not changed, karyotype analysis was performed. Next, HLCs were derived from those iPSCs using a four-stage hepatic differentiation protocol. The obtained HLCs were characterised using, among others, a hepatic gene expression analysis. Cells after differentiation express mature and fetal hepatic markers, which is consistent with previous results. The attempt to improve differentiation using transient overexpression of master hepatic transcription factor – HNF4α, was not sufficient, as shown by gene expression analysis and whole slide scanning. Previous studies failed to point out the genetic inhibitors of hepatic maturation and non-coding RNA (ncRNA) profiles of iPSCs – derived HLCs were not investigated. In this study, the sequencing of ncRNA was performed in order to compare the expression profiles of HLCs on two stages of differentiation (Day 20 and 24) with mature hepatocytes. The obtained results indicate that HLCs express miRNA, which control hepatic differentiation and maintain their fetal liver character. In comparison to mature hepatocytes, differentially expressed miRNAs in HLCs control the pathways of fatty acid metabolism and synthesis, proteoglycan in cancer, the Hippo signaling pathway, ECM-receptor interaction and adherens junction. Some of those highly expressed miRNAs can potentially block maturation by inhibiting epithelial-mesenchymal transition (EMT) which has an impact that is essential during hepatic differentiation. However, this should be resolved in future research. In this work, differentially expressed snoRNA were also identified. A total of 68% of differentially expressed snoRNAs was C/D box class. This is interesting because this snoRNa class was previously indicated as capable to be processed by an miRNA processing pathway. Many of the differentially expressed snoRNAs belong to the imprinted loci, in which a different expression in a human were analysed before. In obteined dataset, copies of SNORD115 were upregulated in a liver, but not in HLCs, which is consistent with an earlier comparison of a liver and other endoderm organs. Additionally, an analysis of obtained sequencing data allowed for a discovery of 19 novel snoRNA genes. In summary, this work shows a new approach to the reprogramming of a fibroblast and investigates the involvement of miRNAs and snoRNAs in the dynamics of hepatic differentiation. This study has shed a light on the molecular and regulatory mechanisms that underlie the complex process of liver differentiation and will hopefully allow existing problems with the use of in vitro derived hepatocytes to be overcome. A dataset generated here can be the foundation for a hepatic-specialised rybosomes theory and enabled to discover novel snoRNA genes.:1. INTRODUCTION 11 1.1. PLURIPOTENT STEM CELLS 11 1.1.1. Pluripotency 11 1.1.2. IPSCs 13 1.1.3. Reprogramming methods 14 1.1.4. IPSCs as an alternative cell source for disease modelling and regenerative medicine 16 1.2. LIVER 18 1.2.1. Liver anatomy and function 18 1.2.2. Liver embryonal development 20 1.3. HEPATIC DIFFERENTIATION OF IPSCS IN VITRO 22 1.3.1. HLCs 22 1.3.2. Differentiation protocols into hepatocytes 24 1.4. NCRNA 25 1.4.1. MiRNA 26 1.4.2. SnoRNA 28 2. AIMS 31 3. MATERIALS 32 3.1. EQUIPMENT 32 3.2. SOFTWARE 32 3.3. ENZYMES, KITS AND TRANSFECTION REAGENTS 33 3.4. SOLUTIONS AND REAGENTS 33 3.5. CELL LINES 34 3.6. CELL CULTURE MEDIA AND CYTOKINES 34 3.7. PLASMIDS 35 3.8. PCR REAGENTS AND PRIMERS 35 3.8.1. PCR reagents 35 3.8.2. PCR primers 35 3.9. ANTIBODIES 36 4. METHODS 37 4.1. CELL BIOLOGY 37 4.1.1. Derivation and culture of primary human foreskin fibroblasts 37 4.1.2. Counting cells 37 4.1.3. Cryo-preservation of cells 37 4.1.4. Thawing of cryo-preserved cells 38 4.1.5. Cell reprogramming 38 4.1.6. Cultivation and expansion of iPSCs 39 4.2. IMMUNOCYTOCHEMISTRY 39 4.3. IN VITRO SPONTANEOUS DIFFERENTIATION 39 4.4. KARYOTYPE ANALYSIS 40 4.5. RNA ISOLATION 40 4.6. QUANTITATIVE PCR 40 4.7. PERIODIC ACID-SCHIFF (PAS) STAINING 41 4.8. INDOCYANINE GREEN UPTAKE AND RELEASE 41 4.9. PLASMID TRANSFECTION 42 4.10. HEPATIC DIFFERENTIATION 42 4.11. WHEAT GERM AGGLUTININ STAINING 42 4.12. VALIDATION OF HEPATIC DIFFERENTIATION EFFICIENCY 43 4.13. RNA ISOLATION AND SEQUENCING 43 4.14. BIOINFORMATIC ANALYSIS 44 4.14.1. Sequencing quality and mapping 44 4.14.2. Analysis of differential expressed ncRNAs 44 4.14.3. Target pathways prediction of differentially expressed miRNAs 44 4.14.4. Identification of novel ncRNAs candidates 45 5. RESULTS 46 5.1. GENERATION OF IPSCS USING EPISOMAL VECTORS 46 5.1.1. Cell transfection 46 5.1.2. Establishment of iPSCs line 48 5.2. PLURIPOTENCY CHARACTERISATION OF THE IPSCS 49 5.2.1. Pluripotency markers 49 5.2.2. Spontaneous differentiation assay 50 5.2.3. Karyotype 52 5.3. HEPATIC DIFFERENTIATION OF IPSCS AND HLCS CHARACTERISATION 53 5.3.1. iPSCs hepatic differentiation 53 5.3.2. Expression of hepatic markers 54 5.3.3. Hepatic gene expression in HLCs 56 5.3.4. Hepatic functions in HLCs 58 5.4. HNF4A OVEREXPRESSION DURING DIFFERENTIATION 59 5.4.1. Cell transfection during differentiation 59 5.4.2. Comparison of hepatic differentiation efficiency 60 5.4.3. Whole slide scanning 62 5.5. NON-CODING RNA ANALYSIS 64 5.5.1. Non-coding RNA sequencing quality 64 5.5.2. MicroRNA analysis 68 5.5.3. SnoRNA analysis 79 5.5.4. Short reads snoRNA analysis 84 5.5.5. New gene candidates 85 6. DISCUSSION 88 6.1. METHODICAL STRATEGY 88 6.2. CHARACTERISATION OF GENERATED IPSCS 89 6.3. HEPATIC DIFFERENTIATION OF IPSCS 89 6.3.1. Characterisation of HLCs 89 6.3.2. Protocol with HNF4a overexpression 90 6.3.3. Differentially expressed miRNA 90 6.3.4. Differentially expressed snoRNA 93 6.4. NOVEL SNORNA GENES 95 7. SUMMARY 96 8. REFERENCES 99 9. APPENDIX 118 ERKLÄRUNG ÜBER DIE EIGENSTÄNDIGE ABFASSUNG DER ARBEIT 122. ACKNOWLEDGEMENTS 123

Researches from different fields around the world are searching for both new sources of biomaterials and potential hepatocytes in order to supply drug tests, cell therapies, and cell transplantation as alternative therapeutic support to liver diseases and injuries. Placenta may be eligible as a new model in tissue engineering due to its rich extracellular matrix (ECM) and availability after birth. Placental scaffolds were produced by decellularization with 0.01, 0.1 and 1% SDS, and 1% Triton X-100 which were valued by means of structure and composition. Afterwards, placental scaffolds were co-cultured with mouse embryonic fibroblasts in a tridimensional (3D) rotating system. Placental scaffolds presented a well-preserved acellular ECM containing 9.42 ± 5.2 ng dsDNA per mg of ECM. Weak collagen I of the natives clearly appears in decellularized ECM while the collagen III, once well observed in native placenta, it was absent on scaffolds. This interesting observation may have been due to the solubilization SDS-induced of the collagen III fibrils during decellularization. Fibronectin was well-observed in placental scaffolds whereas laminin and collagen IV were strongly stained. Recellularized with fibroblasts by a 3D culture system, placental scaffolds showed potential for repopulation, with cells adhered throughout its acellular ECM. Placental scaffolds were then newly recellularized, aiming now for differentiation of mouse embryonic stem cells into hepatic cells. In a protocol of 23 days, it was simulated major events of liver embryonic development by adding growth factors. As result, a high index of cells adhered, proliferated and migrated throughout outer and inner scaffolds ECM surface. Absence of Oct4 and Nanog showed that Activin A and Wnt3a (d0-6) induced primitive endoderm fate, and negative label for Foxa2 and Sox17 representing BMP4 and FGF2 (d6-10) differentiation-induced generating definitive endoderm cells. Also, FGF1, FGF4 and FG8b (d10-14) induced hepatoblast phenotype cells, that were observed positive for AFP and CK7 markers. Finally, HGF and FS-288 (d14-23) induced to hepatocyte-like cells, positive for CK18 and Alb markers. The hepatocyte-like cells functional aspects were observed by glycogen storage. Though a heterogeneous cell hepatic lineage was confirmed, mouse placental scaffolds shown a useful model to support recellularization with simultaneous differentiation into hepatic fate simulating phases of embryonic development. / Pesquisas de diferentes campos ao redor do Mundo estão em busca de novas fontes tanto de biomateriais, quanto de potenciais hepatócitos, a fim de suprir testes de drogas, terapias celulares e transplante de células, como suporte terapêutico alternativo para doenças e lesões hepáticas. Placentas podem ser elegíveis como um novo modelo em Engenharia Tecidual em decorrência de sua rica matriz extracelular (ECM), e disponibilidade após o nascimento. Os scaffolds placentários foram produzidos por decelularização com SDS 0,01, 0,1 e 1% e Triton X-100 1%, os quais foram avaliados por meio da estrutura e composição. Posteriormente, os scaffolds placentários foram co-cultivados com fibroblastos embrionários de camundongos em um sistema rotativo tridimensional (3D). Os scaffolds placentários apresentaram uma MEC acelular bem conservada, contendo 9,42 ± 5,2 ng/dsDNA/mg/MEC. O fraco colágeno I nos nativos aparece claramente na MEC descelularizada, enquanto o colágeno III bem visível na placenta nativa estava ausente nos scaffolds. Esta observação interessante pode decorrido da solubilização das fibrilas de colágeno III, induzida pelo SDS durante a decelularização. A fibronectina foi bem observada nos scaffolds placentários, enquanto a laminina e o colágeno IV estiveram fortemente marcados. Recelularizados com fibroblastos por um sistema de cultura 3D, os scaffolds placentários mostraram potencial para repovoamento, com células aderidas ao longo de sua MEC acelular. Os scaffolds placentários foram então novamente recelularizados, visando agora a diferenciação de células tronco-embrionárias de camundongos em células hepáticas. Em um protocolo de 23 dias, foram simulados os grandes eventos do desenvolvimento embrionário do fígado, pela adição de fatores de crescimento. Como resultado, um alto índice de células aderiu, proliferou e migrou através das superfícies externa e interna dos scaffolds. A ausência de Oct4 e Nanog demostraram que o Activin A e o Wnt3a (d0-6) induziram o destino endoderma primitivo, e a marcação negativa para Foxa2 e Sox17 representaram a geração de células endodermais definitivas pela diferenciação induzida por BMP4 e FGF2 (d6-10). Ainda, FGF1, FGF4 e FG8b (d10-14) induziram células do fenótipo hepatoblasto, que foram observadas positivas para os marcadores AFP e CK7. Finalmente, HGF e FS-288 (d14-23) induziram as células hepatocyte-like, positivas para os marcadores CK18 e Alb. The hepatocyte-like cells functional aspects were observed by glycogen storage. Though a heterogeneous cell hepatic lineage was confirmed, mouse placental scaffolds shown a useful model to support recellularization with simultaneous differentiation into hepatic fate simulating phases of embryonic development. Os aspectos funcionais das células hepatocyte-like foi observada pelo armazenamento de glicogênio. Embora uma linhagem hepática formada por células heterogêneas tenha sido confirmada, os scaffolds placentários de camundongos se mostraram um modelo útil para sustentar a recelularização com simultânea diferenciação em destino hepático, simulando fases do desenvolvimento embrionário.

Les formes sévères de l’infection par les virus de la fièvre jaune (YFV) et de la dengue (DENV) sont caractérisées par une atteinte du foie, plus sévère lors d’une infection YFV. L’objectif de cette thèse est de comparer les infections de YFV et DENV dans un modèle d’hépatocytes humains dérivés de cellules souches (iHeps) afin d’identifier des facteurs à l’origine de cette différence de pathogenèse. Dans un premier temps, nous avons comparé le tropisme de YFV aux 4 sérotypes de DENV dans notre modèle hépatique établi en monocouche cellulaire. Nous avons observé une faible propagation de DENV dans les iHeps par rapport YFV. Les mêmes observations ont été faites dans des hépatocytes primaires. L’utilisation de souches chimériques 17D/DENV a permis de mettre en évidence que cette faible propagation serait liée à une faible efficacité d’entrée de DENV dans les hépatocytes. Nous avons également étudié l’infection dans des sphéroïdes iHeps, métaboliquement plus actifs que les iHeps 2D. Une infection productive a été observée uniquement avec YFV. Ce résultat pourrait s’expliquer par la faible accessibilité des cellules à l’intérieur des sphéroïdes. Dans un 2ème temps, nous avons étudié les réponses cellulaires induites dans les iHeps 2D après infection par les différents virus en utilisant une approche RNAseq. Les résultats préliminaires suggèrent un lien entre le taux de réplication et le nombre de gènes activés. La réponse interféron est plus précocement détectée dans le cas de YFV, mais l’infection par DENV induit un plus grand nombre de gènes. De plus, DENV-1 et DENV-4 induisent une augmentation d’expression de certains gènes impliqués dans la présentation d’antigène comme HLA-E et TAP-2, alors que YFV diminue l’expression de certains gènes de chimiokines et molécules d’adhésion. L’analyse préliminaire des voies liées au métabolisme hépatique révèle une inhibition de la voie de la coagulation dans le cas de l’infection par YFV, qui n’est pas observée lors de l’infection par DENV. Des observations similaires ont été décrites in vivo, au niveau protéique, confirmant la pertinence du modèle iHeps / Severe forms of infection with yellow fever virus (YFV) and dengue virus (DENV) are characterized by liver damage, with more severe symptoms observed during YFV infection. The aim of this thesis is to compare YFV and DENV infections in a model of human hepatocytes derived from stem cells (iHeps) in order to identify factors that could explain their difference in pathogenesis.First, we compared YFV tropism to the four DENV serotypes in 2D iHeps. We observed a low spread of DENV compared to YFV in both iHeps and primary hepatocytes. By using chimeric 17D/DENV strains, we demonstrate that this low propagation is linked to a low DENV entry efficiency in hepatocytes. We also studied infection in iHeps spheroids, metabolically closer to primary cells than 2D iHeps. A productive infection was observed with YFV only. The low accessibility of cells inside the spheroids could explained this result. Second, we studied cellular responses induced following infection by different viruses in 2D iHeps using an RNAseq approach. Preliminary results suggest a link between replication rate and the number of activated genes. The interferon response is detected earlier following YFV infection, but DENV induces a greater number of genes implicated in this pathway. Moreover, DENV-1 and DENV-4 up-regulate some genes involved in antigen presentation such as HLA-E and TAP-2, while YFV down-regulates genes encoding chemokines and adhesion molecules. Preliminary analysis of hepatic metabolism pathways reveals inhibition of the coagulation pathway induced by YFV infection, which is not observed during DENV infection. Similar observations have been described in vivo, at the protein level, confirming the relevance of the iHeps model