Cerebral cortex development is a complex process that integrates both extrinsic and intrinsic mechanisms. The surrounding cellular environment triggers receptor activation, which in turn initiates components of different signalling cascades and subsequently gene transcription, influencing cell survival, proliferation, and differentiation. Genetic mutations causing a loss-of-function or gain-of-function of signalling pathways elements can lead to cortical abnormalities and result in cognitive dysfunctions. In this thesis, I examined the receptor tyrosine kinase (RTK) TrkB and TrkC, the small GTPase Ras, and the C/EBP family of transcription factors, investigating their roles during cerebral cortex development. First, I looked at the role of C/EBPs during cortical cell fate determination. I determined that inhibition of C/EBPs decrease neurogenesis, keeping precursors in an undifferentiated state and later promoting their differentiation into astrocytes, while expression of an activated form of C/EBP promoted neurogenesis and reduced astrogenesis. Moreover, the inhibition of MEK, a mediator of C/EBPβ phosphorylation, also caused a decrease in neurogenesis. Thus, activation of the MEK-C/EBP pathway biases precursor cells to become neurons rather than astrocytes, thereby acting as a differentiation switch. Second, I examined the involvement of Trk signalling during cortical development. I showed that genetic knockdown using shRNA, or inhibition using dominant negative of TrkB and TrkC lead to a decrease in proliferation and later to postnatal precursor cells depletion. Moreover, it caused a reduction in number of neurons combined with mislocalization of the generated neurons to the different cortical layers. Thus, Trk signalling plays an essential role in the regulation of cortical precursor cell proliferation and differentiation during embryonic development. Third, I elucidated the effect of Costello syndrome H-Ras mutations during cerebral cortex formation. I determined that these mutations promoted cell proliferation and astrogenesis, while reducing neurogenesis. Together, these data support a model where proper Trks/Ras/MEK/C/EBP signalling is essential for normal genesis of neurons and astrocytes and show that cortical development perturbations can ultimately lead to cognitive dysfunction as seen in Costello syndrome patients.

Cerebral cortex development is a complex process that integrates both extrinsic and intrinsic mechanisms. The surrounding cellular environment triggers receptor activation, which in turn initiates components of different signalling cascades and subsequently gene transcription, influencing cell survival, proliferation, and differentiation. Genetic mutations causing a loss-of-function or gain-of-function of signalling pathways elements can lead to cortical abnormalities and result in cognitive dysfunctions. In this thesis, I examined the receptor tyrosine kinase (RTK) TrkB and TrkC, the small GTPase Ras, and the C/EBP family of transcription factors, investigating their roles during cerebral cortex development. First, I looked at the role of C/EBPs during cortical cell fate determination. I determined that inhibition of C/EBPs decrease neurogenesis, keeping precursors in an undifferentiated state and later promoting their differentiation into astrocytes, while expression of an activated form of C/EBP promoted neurogenesis and reduced astrogenesis. Moreover, the inhibition of MEK, a mediator of C/EBPβ phosphorylation, also caused a decrease in neurogenesis. Thus, activation of the MEK-C/EBP pathway biases precursor cells to become neurons rather than astrocytes, thereby acting as a differentiation switch. Second, I examined the involvement of Trk signalling during cortical development. I showed that genetic knockdown using shRNA, or inhibition using dominant negative of TrkB and TrkC lead to a decrease in proliferation and later to postnatal precursor cells depletion. Moreover, it caused a reduction in number of neurons combined with mislocalization of the generated neurons to the different cortical layers. Thus, Trk signalling plays an essential role in the regulation of cortical precursor cell proliferation and differentiation during embryonic development. Third, I elucidated the effect of Costello syndrome H-Ras mutations during cerebral cortex formation. I determined that these mutations promoted cell proliferation and astrogenesis, while reducing neurogenesis. Together, these data support a model where proper Trks/Ras/MEK/C/EBP signalling is essential for normal genesis of neurons and astrocytes and show that cortical development perturbations can ultimately lead to cognitive dysfunction as seen in Costello syndrome patients.

Forebrain development is controlled by a set of transcription factors which are expressed in dynamic spatiotemporal patterns in the embryonic forebrain and are known to regulate complex gene networks. Pax6 is a transcription factor that regulates corticogenesis and mutations affecting Pax6 protein levels cause neurodevelopmental defects in the eyes and forebrain in both humans and mice. In previous studies, it was shown that the graded expression pattern of Pax6 protein, which is high rostro-laterally to low caudo-medially in the cerebral cortex, is critical for its control of cell cycle progression and proliferation of cortical progenitors. However the underlying mechanisms are still unclear. Based on a microarray analysis carried out in our laboratory, a number of cell cycle-related candidate genes that may be affected by Pax6 have been identified. One such gene, Cell division cycle associated 7 (Cdca7) is expressed in a counter-gradient against that of Pax6. In my current study, I found that Cdca7 mRNA expression in the telencephalon is upregulated in Pax6 null (Small eye) mutants and downregulated in mice that overexpress PAX6 (PAX77) across developing time points from E12.5 to E15.5. There are several potential Pax6 binding motifs located in the genomic locus upstream of Cdca7. However, by chromatin immunoprecipitation, it is showed that none of the predicted binding sites are physically bound by Pax6. Promoter luciferase assays using fragments combining five suspected binding motifs show that Pax6 is functionally critical. Cdca7 is also identified as a Myc and E2F1 direct target and is upregulated in some tumours but its biological role is not fully understood. Current work using in utero electroporation to overexpress Cdca7 around the lateral telencephalon, where Cdca7 expression levels are normally low, tested the effects on the proliferation and differentiation of cortical progenitor cells in this region. In E12.5 mice embryos, overexpression of Cdca7 protein causes fewer intermediate progenitor cells and post-mitotic neurons to be produced but these effects were not found in E14.5 embryos. This result implies that Cdca7 may affect cell fate decision during cortical development.

Pocket proteins (pRb, p107 and p130) are well studied in the role of regulating cell proliferation by controlling progression through the G1/S phase of the cell cycle. Increasing genetic and anatomical evidence suggests that these proteins also control early differentiation and even later stages of cell maturation including neural migration. However, the multifaceted functions of pocket proteins in the regulation of cell proliferation and cell death has complicated our interpretation of their role during development. As a result, the mechanisms through which pocket proteins regulate neuronal migration and neural maturation remain unknown. Using a pRb and p107 double knock out model, we show that a population of upper layer cortical neurons fails to pass through the intermediate zone into the cortical plate. Importantly, these neurons are born at the appropriate time and have exited the cell cycle. In addition, the role of pocket proteins in radial migration is independent cell death, since this migration defect cannot be rescued by eliminating ectopic cell death through Bax deletion. We also show a novel role of pRb and p107 in development of the dorsal midline and guidance of callosal axons. In the absence of pRb and p107, the structures of the commissural plate are highly disorganized and the callosal axons fail to cross the midline. We identify primary defects in axon extension and expression of multiple guidance cues, which can be observed prior to the disorganization of the midline axon guidance structures. Through the use of in vitro cortical explants and in utero electroporation, we identify defects in the rate of axon extension and directional guidance independent from the midline. In addition, protein levels of Netrin and Neuropilin-1 are decreased in the absence of pRb and p107, which could mediate the function of pocket proteins in guiding callosal axons. Indeed, we identify a previously undescribed population of Netrin expressing cells in the cingulate cortex of control embryos which is lost in the pRb/p107 deficient littermates. We propose that these cells play a significant role in callosal axon guidance during normal development. The results presented in this dissertation define multiple novel roles of pRb and p107 in the regulation of radial migration and axon guidance, independent from the role of these pocket proteins in cell death and proliferation.

Einführung: Eine entscheidende Ursache für das Aufkommen der den modernen Menschen charakterisierenden kognitiven Funktionen ist in der beachtlichen Vergrößerung des menschlichen Neocortex innerhalb der letzten 5-7 Millionen Jahre zu finden. Die Identifizierung der dieser Entwicklung zu Grunde liegenden genomischen Veränderungen ist letztlich nicht nur bedeutsam für die Beantwortung der Frage, welche evolutionären Anpassungen den Menschen kennzeichnen, sondern auch für ein besseres Verständnis einer möglicherweise besonderen Anfälligkeit gegenüber neurologischen und psychiatrischen Erkrankungen. Kürzlich konnten 15 menschenspezifische Gene, deren Expression sich vorzugsweise in neuronalen Vorläuferzellen (NPCs) des menschlichen fetalen Neokortexes nachweisen lässt, identifiziert werden (Florio et al., 2018). Drei davon (FAM72B, C und D) sind vor 3,4 – 1 Millionen Jahren im menschlichen Genom durch Genduplikationen entstanden und gehören zur Family of sequence similarity 72 (FAM72). Zielsetzung und Ansätze: Konkret wurde betrachtet, ob FAM72D durch die spezifischen Substitutionen von Aminosäuren eine sich von der Funktion des anzestralen Gens FAM72A unterscheidende Rolle in der neokortikalen Entwicklung einnimmt. Untersucht wurden deshalb die Effekte von FAM72A und D auf die Proliferationskapazität und Genexpression von NPCs nach der ektopen Expression von FAM72A oder D während der embryonalen Entwicklung des Neocortex der Maus. Methoden: Die in utero Elektroporation (IUE) embryonaler Mäusegehirne erfolgte zur Expression eines rot oder grün fluoreszierenden Proteins (RFP oder GFP) entweder gemeinsam mit einem leeren DNA pCAGGS Vektor als Kontrollbedingung oder aber einem pCAGGS-FAM72A oder pCAGGS-FAM72D Plasmid. Die in der zweiten Ergebnissektion (Results II) präsentierten IUE wurden dabei im dorsolateralen Neokortex zum Höhepunkt der Neurogenese am 14. Entwicklungstag (E 14.5) durchgeführt, im Unterschied zu den Experimenten in der dritten Sektion (Results III), die im medialen Neokortex am 18. Entwicklungstag (E 18.5) während der Spätphase der embryonalen Neurogenese realisiert wurden. Die Proliferation der NPCs wurde durch Immunfluoreszenzanalysen zweier Marker (Ki67 und phosphoryliertes Histon 3) bestimmt. Zudem wurde die Häufigkeit wichtiger Subtypen von NPCs ebenfalls durch Immunfluoreszenzanalysen eines Markers für basale intermediäre Vorläuferzellen (bIPs → Tbr2) sowie für basale und apikale radiale Gliazellen (aRGs, bRGs → Sox2) ermittelt. Die Gliogenese wurde durch Olig2 Immunfluoreszenz quantifiziert. Weitere Experimente wurden durchgeführt, um die Fähigkeit der NPCs, den Zellzyklus nach der IUE von FAM72D erneut einzuleiten, zu untersuchen. Zu diesem Zweck wurde schwangeren Mäusen 24 h nach der IUE das Thymidin-Analogon 5-Ethinyl-2'-desoxyuridin (EdU) intraperitoneal injiziert. Damit wurden alle Zellen markiert, die sich zu diesem Zeitpunkt in der S-Phase des Zellzyklus befanden und damit den Zellzyklus nach der IUE fortsetzten. Nach weiteren 24 h (48 h post-IUE) erfolgte die Auswertung: alle Ki67- und EdU-doppelt positiven Zellen wurden als solche betrachtet, die den Zellzyklus nach IUE fortführten (EdU+) und nach weiteren 24 h noch immer proliferierten (Ki67+). Zur Durchführung der Transkriptomanalyse wurden Mäuse am 13. Entwicklungstag mit pCAGGS-GFP und entweder dem leeren DNA-Vektor (pCAGGS, Kontrolle) oder einem die Expression von FAM72A (pCAGGS-FAM72A) oder FAM72D (pCAGGS-FAM72D) ermöglichenden Vektor elektroporiert. Anschließend wurden die elektroporierten dorsolateralen neokortikalen Bereiche am 14. Entwicklungstag mikroskopisch seziert und in einzelne Zellen dissoziiert. Die Isolation der elektroporierten (GFP+) Zellen erfolgte aus den Einzelzellsuspensionen durch Fluoreszenz-aktivierte Zellsortierung (FACS). Im Anschluss wurden die isolierten Zellen für die RNA-Sequenzierung vorbereitet. Die primäre Datenanalyse der Ergebnisse der RNA-Sequenzierung wurde entsprechend etablierter Protokolle durchgeführt (Florio et al., 2015). Ergebnisse: Die Analyse der Immunfluoreszenzquanitfizierungen (Results II und III) ergab keine signifikanten Veränderungen der proliferativen Parameter oder der Häufigkeit der NPCs in der ventrikulären Zone (VZ) oder subventrikulären Zone (SVZ) des sich entwickelnden Mausneokortex nach der ektopen Expression von FAM72A oder FAM72D im Vergleich zur Kontrollbedingung. Die Transkriptomanalyse (Results IV) zeigte jedoch 88 signifikant hoch- und 52 herunterregulierte Gene in Folge der FAM72A sowie 91 signifikant hoch- und 67 herunterregulierte Gene nach der FAM72D Expression im Vergleich zur Kontrolle. Es wurde festgestellt, dass nur zwei dieser differentiell exprimierten Gene in Folge der ektopen Expression sowohl von FAM72A als auch FAM72D hochreguliert wurden und ein Expressionslevel > 1 fpkm aufwiesen: Syde1 und Shisa5. Darüber hinaus wurden sechs Gene mit > 1 fpkm identifiziert, die spezifisch nach der Expression von FAM72D hochreguliert waren: Tapbp, Mtfp1, Slitrk5, Parp9, Cnp, Rbm43. Darüber hinaus zeigte die Genontologie-Analyse (Gen Ontology) eine signifikante Anreicherung von Angiogenese-assoziierten Genen (z. B. Vegfc) im Datensatz der artifiziell FAM72A exprimierenden Zellen. Interessanterweise konnte beobachtet werden, dass unter den im Vergleich zur Kontrolle differentiell exprimierten Genen mehr Gene mit typischer Expression in NPCs in Folge von FAM72D als FAM72A Expression hochreguliert und mehr NPC typische Gene nach FAM72A Expression herunterreguliert wurden. Im Falle der Gene, deren Expression eher in Neuronen zu finden ist, zeigte sich ein entgegengesetztes Bild (Results IV). Diese Befunde lassen den vorsichtigen Schluss zu, dass FAM72D stärker als FAM72A die Aufrechterhaltung von NPC-Eigenschaften positiv beeinflussen kann. Schlussfolgerungen: In einer früheren Studie erhöhte der Knockdown von Fam72a in NPCs erwachsener Mäuse die Neurogenese (Benayoun et al., 2014). Dies legt in Verbindung mit den vorliegenden Ergebnissen nahe, dass FAM72A und FAM72D nicht hinreichend, möglicherweise jedoch notwendig sind, um die Aufrechterhaltung des Vorläuferzellcharakters von NPCs zu fördern (Results II, III). Aus diesem Grund sollte das in dieser Studie verfolgte Gain of Function Design durch einen Loss of Function Ansatz ergänzt werden. Als Modellsystem bieten sich hierfür insbesondere Hirnorganoide aus Stammzellen des Schimpansen oder Menschen an. Da alle der kürzlich identifizierten menschenspezifischen Gene in den gleichen NPCs exprimiert werden, sollte auch die potenzielle synergistische Wirkung auf die NPC-Proliferation der FAM72 und der zwölf anderen humanspezifischen Gene wie etwa ARHGAP11B analysiert werden. Neben anderen möglichen Mechanismen, die auf Grundlage der Genexpressionsanalyse im Diskussionsteil dieser Arbeit (Results IV und Discussion) erörtert wurden, könnte die Hochregulierung von Slitrk5 in Folge der ektopen Expression des humanspezifischen FAM72D besonders relevant sein. Es ist bekannt, dass Slitrk5 am Recycling des TrKB-Rezeptors beteiligt ist (Song et al., 2015), der wiederum grundlegende Aspekte der Gehirnentwicklung beeinflusst. Ebenfalls konnte bereits gezeigt werden, dass FAM72A die Funktion des TrKB Rezeptors hemmt (Nehar et al., 2009). Somit ist denkbar, dass FAM72D im menschlichen Neokortex die Wiederherstellung der TrKB-Rezeptorfunktion indirekt über Slitrk5 verbessert und dadurch wesentliche Parameter wie das Überleben von Vorläuferzellen und die Neurogenese beim Menschen verlängern oder verstärken könnte. Diese Studie stellt damit die erste funktionelle Charakterisierung der evolutionär hochinteressanten, die FAM72 Gene beinhaltende Region des menschlichen Genoms während der Entwicklung in utero dar. Daraus ergeben sich zahlreiche Ansatzpunkte für zukünftige Untersuchungen, die in ihrer Gesamtheit ein umfassendes Verständnis der Evolution des menschlichen Gehirns ermöglichen werden.:1 INTRODUCTION 11 1.1 WHAT MADE US HUMAN? 11 1.2 THE NEOCORTEX 12 1.2.1 Origin and structure 12 1.2.2 Neurogenesis in the developing neocortex 14 1.2.3 How to increase the neuronal output 18 1.3 EVOLUTION AND GENE DUPLICATION 19 1.3.1 Gene duplication and evolutionary novelty 19 1.3.2 Mechanisms of replication 21 1.3.3 Fates of duplicated genes 22 1.3.4 Which genes tend to duplicate? 24 1.3.5 Human adaptation and gene duplication 24 1.4 HUMAN-SPECIFIC SIGNATURES OF NEOCORTICAL EXPANSION 25 1.5 IDENTIFICATION OF HUMAN-SPECIFIC GENES EXPRESSED IN THE DEVELOPING NEOCORTEX.. 25 1.6 FAMILY WITH SEQUENCE SIMILARITY 72 (FAM72) 26 1.6.1 Evolutionary origin 26 1.6.2 Subcellular localization 27 1.6.3 Cell cycle regulation 28 1.6.4 NPC maintenance 28 2 AIMS & APPROACHES 30 3 RESULTS I 31 3.1 FROM GENES TO PROTEINS: 1 FAMILY – 4 PARALOGUES 31 3.2 FAM72 MRNA EXPRESSION LEVELS IN THE DEVELOPING MOUSE AND HUMAN NEOCORTEX 32 3.3 COMPUTATIONAL ANALYSES 34 3.3.1 Proportion of cysteines 34 3.3.2 Transmembrane domain 34 3.4 AMPLIFICATION, SUBCLONING AND MUTAGENESIS 36 3.4.1 Amplification from human cDNA 36 3.4.2 Verification of the pCAGGs vectors 36 4 RESULTS II 38 4.1 ECTOPIC EXPRESSION OF FAM72A AND FAM72D IN THE MOUSE DORSOLATERAL NEOCORTEX AT MID-NEUROGENESIS 38 4.2 NPC PROLIFERATION 39 4.2.1 Assessment of NPC proliferation using Ki67 immunofluorescence 39 4.2.2 Cell cycle reentry 41 4.2.3 Assessment of mitosis using PH3 immunofluorescence 43 4.2.4 Conclusion 45 4.3 NPC ABUNDANCE 46 4.3.1 Assessment of NPC abundance using Tbr2 and Sox2 immunofluorescence 46 4.3.2 Conclusion 49 5 RESULTS III 50 5.1 ECTOPIC EXPRESSION OF FAM72A and FAM72D IN THE MOUSE MEDIAL CORTEX AT LATE-NEUROGENESIS 50 5.2 NPC PROLIFERATION 51 5.2.1 Assessment of the NPC proliferation using Ki67 immunofluorescence 51 5.3 NPC ABUNDANCE 52 5.3.1 Assessment of NPC abundance using Tbr2 and Sox2 immunofluorescence 52 5.4 GLIOGENESIS 53 5.4.1 Assessment of gliogenesis using Olig2 immunofluorescence 53 5.5 CONCLUSION 54 6 RESULTS IV 55 6.1 DIFFERENCES IN GENE EXPRESSION UPON ANCESTRAL FAM72A AND HUMAN-SPECIFIC FAM72D EXPRESSION AT MID-NEUROGENESIS 55 6.1.1 Rationale and experimental setup 55 6.2 DIFFERENTIALLY EXPRESSED GENES UPON ECTOPIC FAM72A AND FAM72D EXPRESSION IN THE DEVELOPING MOUSE DORSOLATERAL NEOCORTEX 56 6.3 UPREGULATED GENES UPON THE ECTOPIC FAM72A OR FAM72D EXPRESSION 57 6.3.1 Upregulated genes upon the ectopic FAM72A and FAM72D expression 57 6.3.2 Upregulated genes upon the ectopic FAM72A or D expression – cut off: fpkm >1 58 6.3.3 Upregulated genes upon the ectopic FAM72A and D expression – cut off: fpkm >1 59 6.4 UPREGULATED GENES SPECIFICALLY UPON THE ECTOPIC FAM72D EXPRESSION – CUT OFF: FPKM >1 60 6.4.1 Tapbp (TAP binding protein, Tapasin) 60 6.4.2 Mtfp1 (mitochondrial fission protein 1, Mtp18) 61 6.4.3 Slitrk5 (Slit and Ntrk-like protein 5) 61 6.4.4 Parp9 (Poly(ADP-ribose) polymerase 9) 63 6.4.5 Cnp (2',3'-Cyclic-nucleotide 3'-phosphodiesterase) 63 6.4.6 Rbm43 (RNA binding motif protein 43) 65 6.5 DOWNREGULATED GENES UPON THE ECTOPIC FAM72A OR FAM72D EXPRESSION 66 6.5.1 Downregulated genes upon the ectopic FAM72A or D expression – cut off: fpkm >1 66 6.5.2 Downregulated genes upon ectopic FAM72A expression – cut off: fpkm >1 66 6.5.3 Downregulated genes upon ectopic FAM72D expression – cut off: fpkm >1 67 6.6 GENES PREVIOUSLY SHOWN TO BE DIFFERENTIALLY EXPRESSED UPON FORCED FAM72A EXPRESSION 69 6.6.1 Cell cycle regulators 69 6.6.2 Tumor suppressor genes 69 6.6.3 PROTEINS PREVIOUSLY OBSERVED TO INTERACT WITH FAM72A 70 6.7 EFFECT OF ECTOPIC FAM72A AND FAM72D EXPRESSION ON GENES IMPLICATED IN NEURAL LINEAGE FATE DECISION 70 6.7.1 Upregulated and NPC-enriched genes 71 6.7.2 Downregulated and NPC-enriched genes 73 6.7.3 Upregulated and neuron-enriched genes 74 6.7.4 Downregulated and neuron-enriched genes 75 6.8 GO ENRICHMENT ANALYSIS 75 6.9 CONCLUSION 75 7 DISCUSSION 78 7.1 WHAT MAKES US HUMAN? 78 7.2 IN UTERO ELECTROPORATION OF A HUMAN-SPECIFIC GENE IN THE DEVELOPING MOUSE NEOCORTEX 81 7.2.1 Opportunities and limitations of the approach 81 7.3 THE FAMILY OF SEQUENCE SIMILARITY 72 AND HUMAN UNIQUENESS 83 7.3.1 Cell cycle regulation and NPC maintenance 83 7.3.2 Cell death 84 7.3.3 Neurogenic period 85 7.3.4 TrkB signaling 85 7.3.5 Mitochondria 86 7.3.6 Angiogenesis 88 7.3.7 An evolutionary immunological adaptation in the brain? 89 7.3.8 FAM72 and SRGAP2 90 7.3.9 FAM72, Neanderthals, and lncRNAs 91 7.4 FUTURE DIRECTIONS 92 7.4.1 Loss of function 92 7.4.2 Gain of function 92 8 SUMMARY / ZUSAMMENFASSUNG 95 8.1 SUMMARY 95 8.2 ZUSAMMENFASSUNG 98 9 MATERIALS AND METHODS 101 9.1 CHART OF ALL EXPERIMENTS 101 9.2 COMPUTATIONAL ANALYSIS 101 9.2.1 Reference sequences and multiple sequence alignments 101 9.2.2 Transmembrane domain prediction 102 9.3 AMPLIFICATION, SUBCLONING, MUTAGENESIS 102 9.3.1 Amplification from human brain cDNA 102 9.3.2 Subcloning 103 9.2.3 Mutagenesis 103 9.4 PLASMID VERIFICATION 104 9.4.1 Transfection of Cos7 cells 104 9.4.2 Immunoblots 104 9.4.3 In situ hybridization (ISH) 105 9.5 MICE 105 9.6 IN UTERO ELECTROPORATION 105 9.7 FIXATION AND CRYOSECTIONS 106 9.8 IMMUNOFLUORESCENCE AND ANTIBODIES 106 9.9 EDU DETECTION 107 9.10 IMAGE ACQUISITION 108 9.11 STATISTICS 108 9.12 MICRODISSECTION AND SINGLE CELL SUSPENSION 108 9.13 FACS 109 9.14 RNA SEQUENCING 109 9.15 TRANSCRIPTOME ANALYSIS 110 10 REFERENCES 111 11 APPENDIX 145 11.1 CONFERENCE PRESENTATION 145 V. ACKNOWLEDGMENTS 146 / Introduction: The higher cognitive functions that characterize modern humans can be attributed to the cerebral neocortex and its remarkable expansion in size during the last 5 – 7 million years of human evolution. The identification of the underlying genomic changes will be not only of importance to better understand the unique complexity of the human brain, but also its susceptibility to neurological and psychiatric diseases. Recently, 15 human-specific genes preferentially expressed in neural progenitor cells (NPCs) of the human fetal neocortex were identified (Florio et al., 2018). Three of them, FAM72B, C and D belong to the Family of sequence similarity 72 (FAM72) and occurred in the human genome by gene duplication 3.4 – 1 mya. Aims & Approaches: Specifically, it was asked whether FAM72D plays a diverse role compared to the ancestral FAM72A (Results II, III, IV) due to the specific sets of amino acid substitutions it acquired (Results I). Effects of FAM72A and FAM72D on the proliferative capacity and gene expressions of embryonic mouse NPCs were analyzed upon ectopic expression either of FAM72A or FAM72D during embryonic mouse neocortical development. Methods: In utero electroporation (IUE) of embryonic mouse brains was performed to drive the expression of a red or green fluorescent protein (RFP or GFP) either plus empty DNA vector (pCAGGS; control), pCAGGS-FAM72A or pCAGGS-FAM72D plasmids in the dorsolateral neocortex at mid-neurogenesis (embryonic day 13.5, E13.5; Results II) or in the medial neocortex at late-neurogenesis (E15.5; Results III). NPC proliferation was evaluated by immunofluorescence of Ki67 (immunohistochemistry, IHC), a cell proliferation marker, and phosphorylated Histone H3 (PH3), a marker of cell mitosis. Moreover, the abundance of NPCs using immunofluorescence of basal intermediate progenitor (Tbr2) and apical and basal radial glia (Sox2) markers, and the gliogenesis by Olig2 immunofluorescence was analyzed. Additional experiments were carried out to study the capacity of NPCs to reenter the cell cycle upon IUE of FAM72D. To this end, pregnant mice were intraperitoneally injected with the thymidine analog 5-Ethynyl-2´-deoxyuridine (EdU) 24 h post-IUE, to label all cells undergoing S-phase of the cell cycle (i.e., all cells that reentered the cell cycle after IUE) in the developing mouse brains. Embryonic brains were collected 24 h after EdU injection and co-stained with Ki67. Ki67 and EdU double positive cells were considered as cells that reentered the cell cycle. To execute the transcriptome analysis E13.5 mice were electroporated with pCAGGS-GFP either plus an empty DNA vector (pCAGGS, control), a vector driving expression of FAM72A (pCAGGS-FAM72A) or FAM72D (pCAGGS-FAM72D). Subsequently, the electroporated dorsolateral neocortical areas were microdissected at E14.5 and dissociated into single cells. The electroporated (GFP+) cells were isolated from the single cell suspensions by the fluorescence-activated cell sorting (FACS). The isolated cells were processed for RNA sequencing. Data analysis was performed as previously reported (Florio et al., 2015). Results: By immunohistochemistry, no significant changes in any of the proliferative parameters or in the abundance of progenitors in the ventricular zone (VZ) and subventricular zone (SVZ) of the developing mouse neocortex upon ectopic expression of FAM72D compared to FAM72A and control samples were detected (Results II, III). However, the transcriptome analysis (Results IV) showed 88 significantly up- and 52 down-regulated genes upon FAM72A and 91 significantly up- and 67 downregulated genes upon FAM72D expression compared to the control. Only two of these differentially expressed genes were found to be upregulated upon FAM72A and FAM72D with an expression >1 fpkm: Syde1 and Shisa5. Besides, six genes specifically upregulated upon ectopic expression of FAM72D exhibiting fpkm > 1 were identified and characterized using the existing literature: Tapbp, Mtfp1, Slitrk5, Parp9, Cnp, Rbm43. Beyond that, gene ontology analysis showed significant enrichment of angiogenesis-related genes (e.g., Vegfc) upon FAM72A expression. Interestingly, there were more genes found to be enriched in NPCs that were upregulated compared to control upon FAM72D than FAM72A expression, but more NPC enriched genes downregulated upon FAM72A compared to FAM72D expression. In the case of differentially expressed neuron-enriched genes, the data was were inverse, which slightly supports the idea that FAM72D rather than FAM72A could positively affect the maintenance of NPC characteristics. Conclusions: In a previous study knockdown of Fam72a in adult mouse NPCs increased neurogenesis (Benayoun et al., 2014). This suggests, in conjunction with the present results, that FAM72A and FAM72D are not sufficient, but may be required, to promote NPC maintenance (Results II, III). This is why the gain of function experiments conducted in this study should be complemented by a loss of function approach in the developing mouse neocortex, in chimpanzee or human-derived brain organoids. Because of their expression in the NPCs of the developing human neocortex, it might be productive to analyze the potential synergistic effect on NPC proliferation of the FAM72s and the 12 other human-specific genes such as ARHGAP11B. Among other mechanisms discussed based on the gene expression analysis in this thesis (Results IV and Discussion), the upregulation of Slitrk5 upon ectopic expression of the human-specific FAM72D could be particularly remarkable. Slitrk5 is known to be involved in the recycling of the TrKB receptor (Song et al., 2015), which affects fundamental aspects of brain development. While FAM72A was found to inhibit the TrKB receptor (Nehar et al., 2009), the occurrence of FAM72D could indirectly rescue the TrKB receptor function via Slitrk5 and thereby prolonging or enhancing essential features such as precursor cell survival and neurogenesis in humans. Therefore, this study provides the first functional characterization of the evolutionary highly interesting region in the human genome comprising the FAM72 genes during embryonic neocortical development in vivo and offers numerous starting points for further investigations, that will collectively facilitate a comprehensive understanding of the genomic adaptations underlying the astonishing evolution of the human brain.:1 INTRODUCTION 11 1.1 WHAT MADE US HUMAN? 11 1.2 THE NEOCORTEX 12 1.2.1 Origin and structure 12 1.2.2 Neurogenesis in the developing neocortex 14 1.2.3 How to increase the neuronal output 18 1.3 EVOLUTION AND GENE DUPLICATION 19 1.3.1 Gene duplication and evolutionary novelty 19 1.3.2 Mechanisms of replication 21 1.3.3 Fates of duplicated genes 22 1.3.4 Which genes tend to duplicate? 24 1.3.5 Human adaptation and gene duplication 24 1.4 HUMAN-SPECIFIC SIGNATURES OF NEOCORTICAL EXPANSION 25 1.5 IDENTIFICATION OF HUMAN-SPECIFIC GENES EXPRESSED IN THE DEVELOPING NEOCORTEX.. 25 1.6 FAMILY WITH SEQUENCE SIMILARITY 72 (FAM72) 26 1.6.1 Evolutionary origin 26 1.6.2 Subcellular localization 27 1.6.3 Cell cycle regulation 28 1.6.4 NPC maintenance 28 2 AIMS & APPROACHES 30 3 RESULTS I 31 3.1 FROM GENES TO PROTEINS: 1 FAMILY – 4 PARALOGUES 31 3.2 FAM72 MRNA EXPRESSION LEVELS IN THE DEVELOPING MOUSE AND HUMAN NEOCORTEX 32 3.3 COMPUTATIONAL ANALYSES 34 3.3.1 Proportion of cysteines 34 3.3.2 Transmembrane domain 34 3.4 AMPLIFICATION, SUBCLONING AND MUTAGENESIS 36 3.4.1 Amplification from human cDNA 36 3.4.2 Verification of the pCAGGs vectors 36 4 RESULTS II 38 4.1 ECTOPIC EXPRESSION OF FAM72A AND FAM72D IN THE MOUSE DORSOLATERAL NEOCORTEX AT MID-NEUROGENESIS 38 4.2 NPC PROLIFERATION 39 4.2.1 Assessment of NPC proliferation using Ki67 immunofluorescence 39 4.2.2 Cell cycle reentry 41 4.2.3 Assessment of mitosis using PH3 immunofluorescence 43 4.2.4 Conclusion 45 4.3 NPC ABUNDANCE 46 4.3.1 Assessment of NPC abundance using Tbr2 and Sox2 immunofluorescence 46 4.3.2 Conclusion 49 5 RESULTS III 50 5.1 ECTOPIC EXPRESSION OF FAM72A and FAM72D IN THE MOUSE MEDIAL CORTEX AT LATE-NEUROGENESIS 50 5.2 NPC PROLIFERATION 51 5.2.1 Assessment of the NPC proliferation using Ki67 immunofluorescence 51 5.3 NPC ABUNDANCE 52 5.3.1 Assessment of NPC abundance using Tbr2 and Sox2 immunofluorescence 52 5.4 GLIOGENESIS 53 5.4.1 Assessment of gliogenesis using Olig2 immunofluorescence 53 5.5 CONCLUSION 54 6 RESULTS IV 55 6.1 DIFFERENCES IN GENE EXPRESSION UPON ANCESTRAL FAM72A AND HUMAN-SPECIFIC FAM72D EXPRESSION AT MID-NEUROGENESIS 55 6.1.1 Rationale and experimental setup 55 6.2 DIFFERENTIALLY EXPRESSED GENES UPON ECTOPIC FAM72A AND FAM72D EXPRESSION IN THE DEVELOPING MOUSE DORSOLATERAL NEOCORTEX 56 6.3 UPREGULATED GENES UPON THE ECTOPIC FAM72A OR FAM72D EXPRESSION 57 6.3.1 Upregulated genes upon the ectopic FAM72A and FAM72D expression 57 6.3.2 Upregulated genes upon the ectopic FAM72A or D expression – cut off: fpkm >1 58 6.3.3 Upregulated genes upon the ectopic FAM72A and D expression – cut off: fpkm >1 59 6.4 UPREGULATED GENES SPECIFICALLY UPON THE ECTOPIC FAM72D EXPRESSION – CUT OFF: FPKM >1 60 6.4.1 Tapbp (TAP binding protein, Tapasin) 60 6.4.2 Mtfp1 (mitochondrial fission protein 1, Mtp18) 61 6.4.3 Slitrk5 (Slit and Ntrk-like protein 5) 61 6.4.4 Parp9 (Poly(ADP-ribose) polymerase 9) 63 6.4.5 Cnp (2',3'-Cyclic-nucleotide 3'-phosphodiesterase) 63 6.4.6 Rbm43 (RNA binding motif protein 43) 65 6.5 DOWNREGULATED GENES UPON THE ECTOPIC FAM72A OR FAM72D EXPRESSION 66 6.5.1 Downregulated genes upon the ectopic FAM72A or D expression – cut off: fpkm >1 66 6.5.2 Downregulated genes upon ectopic FAM72A expression – cut off: fpkm >1 66 6.5.3 Downregulated genes upon ectopic FAM72D expression – cut off: fpkm >1 67 6.6 GENES PREVIOUSLY SHOWN TO BE DIFFERENTIALLY EXPRESSED UPON FORCED FAM72A EXPRESSION 69 6.6.1 Cell cycle regulators 69 6.6.2 Tumor suppressor genes 69 6.6.3 PROTEINS PREVIOUSLY OBSERVED TO INTERACT WITH FAM72A 70 6.7 EFFECT OF ECTOPIC FAM72A AND FAM72D EXPRESSION ON GENES IMPLICATED IN NEURAL LINEAGE FATE DECISION 70 6.7.1 Upregulated and NPC-enriched genes 71 6.7.2 Downregulated and NPC-enriched genes 73 6.7.3 Upregulated and neuron-enriched genes 74 6.7.4 Downregulated and neuron-enriched genes 75 6.8 GO ENRICHMENT ANALYSIS 75 6.9 CONCLUSION 75 7 DISCUSSION 78 7.1 WHAT MAKES US HUMAN? 78 7.2 IN UTERO ELECTROPORATION OF A HUMAN-SPECIFIC GENE IN THE DEVELOPING MOUSE NEOCORTEX 81 7.2.1 Opportunities and limitations of the approach 81 7.3 THE FAMILY OF SEQUENCE SIMILARITY 72 AND HUMAN UNIQUENESS 83 7.3.1 Cell cycle regulation and NPC maintenance 83 7.3.2 Cell death 84 7.3.3 Neurogenic period 85 7.3.4 TrkB signaling 85 7.3.5 Mitochondria 86 7.3.6 Angiogenesis 88 7.3.7 An evolutionary immunological adaptation in the brain? 89 7.3.8 FAM72 and SRGAP2 90 7.3.9 FAM72, Neanderthals, and lncRNAs 91 7.4 FUTURE DIRECTIONS 92 7.4.1 Loss of function 92 7.4.2 Gain of function 92 8 SUMMARY / ZUSAMMENFASSUNG 95 8.1 SUMMARY 95 8.2 ZUSAMMENFASSUNG 98 9 MATERIALS AND METHODS 101 9.1 CHART OF ALL EXPERIMENTS 101 9.2 COMPUTATIONAL ANALYSIS 101 9.2.1 Reference sequences and multiple sequence alignments 101 9.2.2 Transmembrane domain prediction 102 9.3 AMPLIFICATION, SUBCLONING, MUTAGENESIS 102 9.3.1 Amplification from human brain cDNA 102 9.3.2 Subcloning 103 9.2.3 Mutagenesis 103 9.4 PLASMID VERIFICATION 104 9.4.1 Transfection of Cos7 cells 104 9.4.2 Immunoblots 104 9.4.3 In situ hybridization (ISH) 105 9.5 MICE 105 9.6 IN UTERO ELECTROPORATION 105 9.7 FIXATION AND CRYOSECTIONS 106 9.8 IMMUNOFLUORESCENCE AND ANTIBODIES 106 9.9 EDU DETECTION 107 9.10 IMAGE ACQUISITION 108 9.11 STATISTICS 108 9.12 MICRODISSECTION AND SINGLE CELL SUSPENSION 108 9.13 FACS 109 9.14 RNA SEQUENCING 109 9.15 TRANSCRIPTOME ANALYSIS 110 10 REFERENCES 111 11 APPENDIX 145 11.1 CONFERENCE PRESENTATION 145 V. ACKNOWLEDGMENTS 146

Des mutations hétérozygotes faux-sens dans le gène de la tubuline gamma TUBG1, ont été identifiées dans le contexte des malformations du développement cortical, associées à une déficience intellectuelle et à l'épilepsie. Ici, nous avons étudié par la technique d’électroporation in-utero et par des études in vivo, l’effet de quatre de ces variantes sur le développement cortical. Nous montrons que les mutations dans TUBG1 affectent le positionnement neuronal dans la plaque corticale, en perturbant la locomotion des neurones nouvellement nés, mais sans affecter la neurogenèse. Nous proposons que la γ-tubuline mutante affecte le fonctionnement global de ses complexes, et en particulier leur rôle dans la régulation de la dynamique des microtubules. De plus, nous avons développé un modèle de souris knock-in Tubg1Y92C/+ et évalué les conséquences de la mutation sur le développement cortical, les caractéristiques neuroanatomiques et le comportement. Les souris mutantes présentent une microcéphalie globale, des anomalies du néocortex et de l'hippocampe, des altérations du comportement et une susceptibilité épileptique. Ainsi, nous montrons que les souris Tubg1Y92C/+ miment au moins partiellement le phénotype humain et représentent donc un modèle pertinent pour d'autres investigations de la physiopathologie des malformations du développement cortical. / Missense heterozygous variants in the gamma tubulin gene TUBG1 have been linked to malformations of cortical development, associated with intellectual disability and epilepsy. Here, we investigated through in-utero electroporation and in-vivo studies, how four of these variants affect cortical development. We show that TUBG1 mutants affect neuronal positioning within the cortical wall, by a disrupting the locomotion of newly born neurons but without affecting neurogenesis. We propose that mutant γ-tubulin affects overall functioning of γ-tubulin complexes, and in particular their role in the regulation of microtubule dynamics. Additionally, we developed a knock-in Tubg1Y92C/+ model and assessed consequences of the mutation on cortical development, neuroanatomical features and behaviour. Mutant mice present with global microcephaly, neocortical and hippocampal abnormalities, behavioural alterations and epileptic susceptibility. Thus, we show that Tubg1Y92C/+ mice partially mimic the human phenotype and therefore represent a relevant model for further investigations of the physiopathology of malformations of cortical development.