Université Pierre et Marie Curie

ÉCOLE DOCTORALE CERVEAU-COGNITION-COMPORTEMENT

CIRB – Collège de France / Équipe Alain Prochiantz

ENGRAILED, AN ANTI-AGEING HOMEOPROTEIN

Par François-Xavier BLAUDIN de THÉ

Thèse de Doctorat de Biologie

Dirigée par Alain PROCHIANTZ

Présentée et soutenue publiquement le 28 septembre 2015 Devant le jury suivant:

Pr Alain Trembleau Président

Dr Philippe Vernier Rapporteur

Dr Vincent Colot Rapporteur

Pr Wolfgang Wurst Examinateur

Dr Julia Fuchs Examinateur

Pr Alain Prochiantz Directeur de thèse

« Pour soulever un poids si lourd,

Sisyphe, il faudrait ton courage !

Bien qu’on ait du cœur à l’ouvrage,

L’Art est long et le Temps est court. »

Baudelaire C. (1857) Les Fleurs du Mal, XI le Guignon

ACKNOWLEDGMENTS

Thank you to all the members of the jury for accepting to be part of my PhD defense comity. I especially wish to thank Vincent Colot and Philippe Vernier for accepting to read this manuscript in preview.

I would like to thank my Ph.D. director Alain for, unlike others, believing in me in the first place despite my lack of biology formation. I want to thank him for the time he spends with his students and for the profound interest he has in our daily work, which is rare for someone with so many responsibilities and so little leisure time.

I would like to thank Julia my supervisor for the kindness and patience she managed to keep while teaching me basically all there is to know about the context of this project. Thank you also for laughing at my not always funny jokes and imitations. Thank you very much Hocine for being my partner in crime for nearly everything in the lab: from experiments and analysis to dragging me to the terrace, when smoke was coming out of my ears during the writing of the thesis! Thank you to Olivia for your very valuable help and incredible knowledge and frankness. Thank you Rajiv for your encyclopedic expertise on bibliography as well as for your kindness. I also want to acknowledge Thomas who helped me in the lab, while I was writing this and what follows.

A big thank you to all the members of the lab for their advice, help and support during these three wonderful years! Thank you in particular for your jokes and for all the gastronomic experiences we shared. Thank you also to all the other people of the CIRB, particularly all the members of the Joliot-Vriz team for their contagious happiness. A big thank for the administrative staff, and particularly Francine, for relieving our shoulders from the bureaucratic burden.

I would like to thank all the persons who, through their advice or example, showed me that science and research was worth fighting for. I must acknowledge in particular the very positive influence of both my grandfathers, of my father, of Philip Gallic and Jean-Pierre Foulon at Henri IV, of Samir Zard at Polytechnique, of the members of the lab of Nobuhiko Kayagaki in San Francisco and of Cleo Kontoravdi in London.

A very big thank to my family for everything they are. Some say that the most important thing in life is to choose your parents well. I cannot imagine a better choice than my wonderful parents and my three incredible sisters!

A very warm thank you to all my friends for being there during these three wonderful but hard years. It would be useless and ink consuming to quote all their names, furthermore they know who they are. I want to thank in particular my roommates who did a wonderful support job on a day-to-day basis! SUMMARY

LIST OF ILLUSTRATIONS ...... 1

LIST OF ABBREVIATIONS ...... 3

INTRODUCTION ...... 4

I. Homeoproteins ...... 7 A. Homeoproteins as developmental factors ...... 7 B. Homeoprotein transfer ...... 8 C. Engrailed 1 and 2 in the midbrain ...... 10 II. Parkinson disease ...... 13 A. History and symptoms of the disease ...... 13 B. Origin of the disease ...... 14 C. The biochemistry of Parkinson disease ...... 15 D. Animal models of Parkinson disease ...... 18 III. LINEs: a class of retrotransposons not quite extinct ...... 20 A. LINEs mechanism of action ...... 21 B. Defenses against the LINEs ...... 22 C. Retrotransposition in brain neural progenitors ...... 25 D. The implication of LINEs in diseases of the brain and in evolution ...... 26

ARTICLE 1 - Engrailed Homeoprotein, an Anti-Ageing Factor that Protects Mesencephalic Dopaminergic Neurons from Oxidative Stress ...... 30

I. Summary ...... 33 II. Introduction ...... 34 III. Results ...... 36 A. mDA neuron gene expression in the SNpc of En1+/- and wt mice ...... 36 B. DNA strand breaks and modified chromatin marks in wt and En1+/- mDA neurons ...... 37 C. Induction of DSBs and degeneration by 6-OHDA in wt and En1+/- mDA neurons ...... 38 D. Oxidative stress induces changes in heterochromatin marks ...... 39 E. Engrailed protects mDA neurons from oxidative stress ...... 40 F. Engrailed activates anti-apoptotic pathways ...... 42 IV. Discussion ...... 43 V. Experimental procedures ...... 48 VI. Author contributions ...... 51 VII. Acknowledgments ...... 51

PART 2 - Full-length LINE-1 expression in adult dopaminergic neurons of the Substantia Nigra is enhanced by oxidative stress and repressed by Engrailed homeoprotein ...... 72

I. Introduction ...... 74 II. Results ...... 75 A. LINE-1 families are expressed in adult mesencephalic dopaminergic neurons ...... 75 B. LINE-1 expression and nuclear localization are increased upon oxidative stress ...... 76 C. Oxidative stress-induced and direct LINE-1 expression initiate DNA damage ...... 77 D. LINE-1 expression is increased in En1-het mice ...... 78 E. Engrailed protects against LINE-1- induced DNA damage and is a direct repressor of LINE- 1 expression ...... 79 III. Experimental procedures ...... 87 IV. Acknowledgments ...... 91

DISCUSSION ...... 92

I. Genomic instability and epigenetic alterations ...... 96 A. Genomic instability ...... 96 B. Epigenetic alterations ...... 98 C. LINE-1 the missing link between some hallmarks of ageing ...... 100 D. Engrailed direct repression of LINEs could mediate its anti-ageing properties ...... 104 II. Loss of proteostasis and mitochondrial dysfunction ...... 107 A. Loss of proteostasis ...... 107 B. Reactive oxygen species and oxidative stress...... 110 C. Mitochondrial dysfunctions ...... 112 III. Other hallmarks of ageing ...... 115 A. Telomere attrition and cellular senescence ...... 115 B. Deregulated nutrient sensing and altered cellular communication ...... 118 C. Stem cell exhaustion ...... 125 IV. General conclusion ...... 126

BIBLIOGRAPHY ...... 128

LIST OF ILLUSTRATIONS

Figure 1. Antennapedia mutation in Drosophila...... 7

Figure 2. Example of boundaries defined by the expression of abutting homeoproteins with self-activating and reciprocal inhibiting properties...... 8

Figure 3. Embryonic and adult expression of Engrailed in the mouse brain ...... 10

Figure 4. Progressive loss of DA neurons in the ventral midbrain of En1+/- mice...... 11

Figure 5. Engrailed protects SNpc mDA neurons against MPTP ...... 12

Figure 6. Mechanisms of neurodegeneration in PD ...... 16

Figure 7. Classes of mobile genetic elements in the human genome ...... 20

Figure 8. Mechanism of LINE-1 retrotransposition ...... 21

Figure 9. piRNA biogenesis and amplification (Ping-Pong cycle) in the Drosophila and in the mouse...... 23

Figures for Article 1 ...... 52

Figure 1. Altered expression of genes related to DNA damage, chromatin remodeling and apoptosis in En1+/- mice ...... 52

Figure 2. Increased DNA damage and chromatin alteration in mDA neurons in En1+/- SNpc ...... 54

Figure 3. Enhanced sensitivity of En1+/- mice SNpc TH+ neurons to 6-OHDA ...... 56

Figure 4. Modification of heterochromatin marks upon 6-OHDA injection ...... 58

Figure 5. Engrailed rescues TH+ cells from 6-OHDA-induced cell death ...... 60

Figure 6. RNA-seq analysis reveals Engrailed anti-apoptotic activity ...... 62

Figure S1. Analysis and validation of RNA-seq data ...... 64

Figure S2. TH cell death and chromatin changes in 6-OHDA injected mice ...... 66

Figure S3. Engrailed transcription activity involved in TH neuron protection ...... 68

Figure S4. Detection of cell cycle markers in TH+ neurons upon acute oxidative stress in vivo and Engrailed- mediated protection against H2O2-induced DSBs in vitro ...... 70

Figure S5. Differentially expressed genes in the SNpc of En2 vs sham injected mice ...... 71

Figures for Part 2 ...... 82

Figure 1. Full-length Long Interspersed Nuclear Elements (LINEs) are expressed in the adult mouse ventral midbrain...... 82

Figure 2. LINE-1 expression and nuclear localization are increased upon oxidative stress ...... 83

1 Figure 3. The piRNA binding protein Piwil1 represses the formation of DSBs and nuclear LINE-1 RNA foci under conditions of oxidative stress ...... 84

Figure 4. Neurodegeneration is accompanied by LINE-1 activation in En1+/- mice...... 85

Figure 5. Engrailed inhibits retrotransposition by directly repressing LINE-1 transcription...... 86

Figure 1. The nine hallmarks of ageing ...... 94

Figure 2. KEGG pathway analysis of differentially regulated mRNAs between 6 week-old En1+/- and wt mice ...... 95

Figure 3. MeDIP-seq partial results ...... 100

Figure 4. Model for Sirt 6 mediated age related activation of LINE-1 ...... 103

Figure 5. Some of the miRNAs bound to Engrailed ...... 106

Figure 6. Genes from the “Ubiquitin mediated proteolysis” pathway modified in the En1+/- mouse ...... 109

Figure 7. Genes from the “Protein processing in the ER” pathway modified in the En1+/- mouse ...... 110

Figure 8. Genes from the “peroxisome” pathway modified in the En1+/- mouse ...... 112

Figure 9. γ-H2AX Chromatin immunoprecipitation after 6-OHDA injection in the SNpc ...... 116

Figure 10. Genes from the “p53 signalling pathway” modified after Engrailed infusion in the SNpc ...... 117

Figure 11. Genes from the “insulin signalling pathway” modified after Engrailed infusion in the SNpc ...... 119

Figure 12. Genes from the “mTOR signalling pathway” modified after Engrailed infusion in the SNpc ...... 120

Figure 13. The LPA pathway and its potential link to the En1+/- mouse ...... 120

Figure 14. Genes from the “MAPK signalling pathway” modified after Engrailed infusion in the SNpc ...... 123

Figure 15. Genes from the “MAPK signalling pathway” modified in the En1+/- mouse ...... 123

Figure 16. The principle of single-chain antibodies ...... 124

Figure 17. Three main hallmarks of ageing strongly regulated by Engrailed ...... 126

2 LIST OF ABBREVIATIONS

(q)PCR (quantitative)polymerase chain reaction 6-OHDA 6-Hydroxydopamine A-T Ataxia telangiectasia AAV2 Adeno-associated virus 2 ChIP Chromatin immunoprecipitation CHX Cycloheximide CI Complex I CNS Central nervous system CR Calorie restriction DIG Dioxygenin DSB(s) DNA strand break(s) En(1/2) Engrailed (1/2) ER Endoplasmic reticulum FACS Fluorescence activated cell sorting FISH Fluorescent in situ hybridization H3K27(me3) Histone 3 lysine 27 (trimethylated) H3K9(me3) Histone 3 lysine 9 (trimethylated) HD Homeodomain IIS Insulin and IGF-1 signalling Levodopa L- 3,4-dihydroxyphenylalanine LINE(s) Long interspersed nuclear element(s) mDA Midbrain/mesencephalic dopaminergic MHB Midbrain/hindbrain boundary MPP+ 1-methyl-4-phenylpyridinium MPTP 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine mtDNA Mitochondrial DNA ncRNA non coding RNA ORF1/2 Open reading frame 1/2 PD Parkinson disease piRNA Piwi interacting RNA PV Parvalbumin RGC Retinal ganglion cell RNA-seq RNA-sequencing ROS Reactive oxygen species RT Reverse transcription RT plasmid Retrotransposition plasmid RTT Rett syndrome SINE(s) Small interspersed nuclear element(s) SNpc Substantia Nigra pars compacta TPRT Target primed reverse transcription VP16 Transcriptional activator domain of the Herpes virus VP64 4 times VP16 VTA Ventral tegmental area wt Wild type

3

INTRODUCTION

4

5 Homeoproteins are a major class of transcription factors, first described in the fly for their ability to induce morphogenetic changes. They have been described later in many other organisms, including all mammals, signing their great importance. Two of these homeoproteins, Engrailed1 and Engrailed2 (collectively Engrailed) are implicated in midbrain formation during embryogenesis. They are still expressed in the adult midbrain, where they maintain dopaminergic neurons. Indeed, in mice heterozygous for Engrailed1 (En1+/-) dopaminergic neurons from the Substantia Nigra pars compacta (SNpc) start degenerating at 6 weeks of age, a condition reminiscent of Parkinson disease (PD). These mice also display several motor and non-motor symptoms characteristic of this pathology. Finally Engrailed infusion is curative in several mouse models of PD, leading to the hypothesis that Engrailed could be implicated in the pathway of the disease.

PD is the second most common neurodegenerative disorder, characterized by a progressive degeneration of SNpc dopaminergic neurons. As these neurons project to the striatum, their degeneration induces a massive loss of striatal dopamine, leading to the majority of the symptoms. Most patients have idiopathic forms of the disease; nevertheless between 5 to 10% of the patients have familial forms linked to DNA mutations. The treatments are for the moment symptomatic, due to insufficient knowledge of the mechanisms of neurodegeneration. However environmental and genetic studies have shed light on several related pathways, including oxidative stress, mitochondrial dysfunction, defective autophagy and protein misfolding and aggregation.

Long interspersed nuclear elements (LINEs) are retrotransposons, which constitute approximately 20% of the human genome. Their ability transpose in the genome make them potentially harmful for the cell, as they can induce mutagenesis or modify the expression of nearby genes. This could explain why a wide variety of mechanisms exists that inhibit their activity. However, their inhibition is not systematic and LINE-1 activation has been observed in several processes and pathologies particularly in the brain, often in relationship with ageing. This challenges the classical view that the genome is uniform and stable in nerve cells and could suggest a role for LINE transposition in some disease-associated pathways.

6 I. Homeoproteins

Homeoproteins are a class of transcription factors with a highly conserved 60 amino acid-long DNA binding domain, the homeodomain (HD). This HD is coded by a 180-long nucleotide sequence (the homeobox), a signature used to identify a large number of these genes, many of them first described in the fruit fly (Drosophila) thanks to their developmental roles (Gehring 1987).

A. Homeoproteins as developmental factors

The first homeogenes were described in the 1970s in the Drosophila, where they are revealed through homeotic mutations: when a segment or part of a segment is replaced by another segment or segment part. The most famous example is the Antennapedia mutant in which the antennas of the head are replaced by legs (Gehring 1967) (Figure 1). These observations gave rise to the idea that homeoproteins control morphogenesis, meaning that they govern the pattern of a tissue, in particular through the control of cell identity, proliferation, differentiation or positioning. Beyond the fly, similar genes were found in all eukaryotic cells, plants and animals (Derelle et al. 2007), demonstrating their ancestry and importance.

Figure 1. Antennapedia mutation in Drosophila. On the left a wild type fly, on the right a mutated fly in which the antennae are replaced by legs (Photo by FR Turner, Indiana Univ.)

7 One of the roles of homeoproteins is in brain compartmentalization (Prochiantz & Di Nardo 2015). During early embryonic development, the brain is divided thanks to the positioning of boundaries between different territories of the neuroepithelium (Kiecker & Lumsden 2005). This process, controlled by different factors including homeoproteins, is of prime importance, since these compartments will give rise to different brain areas, with distinct functions (Zilles & Amunts 2010). As illustrated in Figure 2, in many cases two homeoproteins are expressed at the two sides of the boundary, with auto-activating and reciprocal inhibitory activities at a transcription level (Quiñinao et al. 2015).

Figure 2. Example of boundaries defined by the expression of abutting homeoproteins with self-activating and reciprocal inhibiting properties. Adapted from (Prochiantz & Di Nardo 2015) MHB midbrain/hindbrain boundary, DMB diencephalon/midbrain, ZLI Zona limitans intrathalamica, PSB palial/sup-palial

Two classical examples of this can be found in the MHB and in the boundary between the visual (V1) and somatosensory cortex (S1). The MHB boundary is governed by the balance between Otx2 and Gbx2 (Simeone 2000; Wurst & Bally-Cuif 2001; Joyner et al. 2000) and the position of the frontier between S1 and V1 depends on the relative expressions of Emx2 and Pax6 (O'Leary et al. 2007).

B. Homeoprotein transfer

A striking feature of several homeoproteins is their non-cell autonomous functions. The hypothesis of homeoprotein intercellular transfer was first proposed following experiments demonstrating the internalization and nuclear addressing of the Antennapedia HD. Ironically, what was initially designed as a negative control turned out to be much more positive than

8 expected and opened a whole new field in homeoprotein studies. In brief, to test the impact of homeogenes on neuronal cells, the Antennapedia HD was scraped into mammalian neurons in culture, a procedure which transiently disrupts the membrane, allowing the HD to enter the cell and, due to HD conservation, to chase the endogenous homeoproteins from their DNA targets. As a negative control, intact cells were treated with the HD. Surprisingly in this last condition, an effect similar to that observed by scrape loading (change in neuronal morphogenesis) was observed, suggesting that the HD has by itself the ability to enter the cells (Joliot, Pernelle, et al. 1991). To test this rather unorthodox hypothesis, intact cells were treated with a fluorescein isothiocyanate (FITC) tagged HD, allowing the observation of its capture and diffusion trough the cytoplasm and into the nucleus (Joliot, Triller, et al. 1991; Joliot, Pernelle, et al. 1991; Le Roux et al. 1993).

Further experiments proved that the internalization sequence corresponds to the third alpha-helix of the HD, an observation extremely important since this helix is highly conserved among homeoproteins; meaning that the 200 or so homeoproteins, encoded by the human genome, may also transfer between cells. This capacity to enter cells has already been shown for several homeoproteins in vitro (Joliot & Prochiantz 2004) and for six of them in vivo, namely Otx2, Pax6, Engrailed1 (En1), Engrailed2 (En2), Vax1 and Hoxd1 (Bardine et al. 2014; Di Lullo et al. 2011; N. Kim et al. 2014; Layalle et al. 2011; Lesaffre et al. 2007; Sugiyama et al. 2008; Wizenmann et al. 2009). More importantly, homeoprotein transfer has also been shown to have physiological roles. Indeed, Engrailed regulates the guidance of retinal ganglion cell (RGC) axons and has a role in retino-tectal patterning (I. Brunet et al. 2005; Wizenmann et al. 2009; Stettler et al. 2012). Otx2 is secreted by the choroid plexus and internalized by the Parvalbumin (PV) interneurons of the visual cortex, thereby regulating their plasticity (Spatazza et al. 2013). This is probably the tip of the iceberg and we believe that more developmental and physiological functions for homeoprotein transfer are still to be described.

Finally as shown by the last example, homeoproteins functions are not restricted to embryogenesis; indeed several homeoproteins have physiological roles in the adult. Engrailed, for example, participates in the maintenance of dopaminergic neurons throughout life.

9 C. Engrailed 1 and 2 in the midbrain

The two genes encoding Engrailed1 and Engrailed2 (En1/2 or Engrailed) start to be expressed at the stage of 1 and 5-somites respectively, in the anterior neuroepithelium, around the Mid/Hindbrain Boundary (MHB) (Davis & Joyner 1988). The two genes remain expressed throughout life in the midbrain dopaminergic (mDA) neurons, in the Substantia Nigra pars compacta (SNpc) and in the ventral tegmental area (VTA) (Prochiantz & Di Nardo 2015) (Figure 3). The two proteins play biochemically equivalent roles in the midbrain (Hanks et al. 1995).

Figure 3. Embryonic and adult expression of Engrailed in the mouse brain Adapted from (Prochiantz & Di Nardo 2015)

En1-/-, En2-/- mouse embryos do not form dopaminergic (mDA) neurons (Simon et al. 2001; Albéri et al. 2004) and En1-/- mice die at birth with many defects, including midbrain/hindbrain deletion (Wurst et al. 1994). Thus revealing an essential role of Engrailed in brain development. Further work has shown that Engrailed is also required for maintenance of dopaminergic neurons in the adult.

Indeed, Engrailed is expressed in several zones of the adult mouse brain, including the ventral midbrain and the cerebellum (Figure 3). In mice heterozygous for En1 (En1+/LacZ, En2+/+, Swiss OF1 background called En1+/- thereafter), the mDA neurons of the SNpc start degenerating progressively after 6 weeks and this until 1 year, when a plateau is observed with around 60% neurons still present (Sonnier et al. 2007) (Figure 4). The VTA neurons degenerate as well but at a slower rate. This phenotype could be reversed by infusing En2 recombinant protein in the ventral midbrain of these mice, suggesting that En2 overexpression

10 can compensate for the loss of one En1 allele. Of note, En1+/- mice also display motor and non-motor phenotypes; such as decreased performance in the forced swimming test, diminished preference for saccharine, which may reflect an anhedonic behavior and/or be the sign of olfactory decline. En1+/- mice also exhibit decreased social interactions (Sonnier et al. 2007).

Figure 4. Progressive loss of DA neurons in the ventral midbrain of En1+/- mice. A-B: Progressive mDA death in the SNpc after 6 weeks. C: Quantification for the VTA mDA neurons Adapted from (Sonnier et al. 2007)

These features, especially the faster loss of mDA neurons in the SNpc compared to that in the VTA, the motor symptoms, the anhedony and olfactory decline are reminiscent of the pathology of PD. Therefore, if an En1 partial loss of function leads to a Parkinson-like phenotype, could an Engrailed gain of function revert such a phenotype? Indeed Engrailed prevents neuronal death in several mouse models of PD, namely injection of 6- Hydroxydopamine (6-OHDA), of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) or of cell permeable mutated A30P α-Synuclein (Alvarez-Fischer et al. 2011) (Figure 5) (see part II for the description of these models).

11

Figure 5. Engrailed protects SNpc mDA neurons against MPTP Adapted from (Alvarez-Fischer et al. 2011)

Engrailed, in addition to being a transcription factor, also regulates the translation of local mRNAs, as demonstrated in axon growth cones and brain synaptoneurosomes (I. Brunet et al. 2005; Wizenmann et al. 2009; Stettler et al. 2012; Alvarez-Fischer et al. 2011). In the case of MPTP injections, Engrailed protective activity involves its ability to regulate the translation of mRNAs produced from nuclear genes encoding mitochondrial proteins. In particular, Engrailed regulates the translation of Ndufs1, a key component of the electron chain transfer Complex I (CI). Indeed, in mice first infused with Engrailed and a siRNA designed against Ndufs1 and then intoxicated with MPTP, the protective effect of the homeoprotein is abolished (Alvarez-Fischer et al. 2011). This means that, in this model, part of Engrailed-mediated neuroprotection goes through the regulation of Ndufs1 translation. In vitro experiments with specific transcription and translation inhibitors have confirmed that Engrailed regulates the translation of Ndufs1 and not its transcription (Alvarez-Fischer et al. 2011). However, because Engrailed is a transcription factor, one can anticipate that its protective activity also presents a transcriptional component.

As mentioned earlier, En1+/- mice display Parkinson-like symptoms, raising the possibility that Engrailed could be implicated in the Parkinson pathway. Although no En1 or En2 mutations have been associated with familial forms of the disease, it is noteworthy that En1/2 polymorphisms were associated with the risk to develop the disease (Fuchs et al. 2009; Rissling et al. 2009; Haubenberger et al. 2011).

12 II. Parkinson disease

PD is the second most common neurodegenerative disease after Alzheimer’s disease (Dauer et al. 2003) and both diseases are a major public health problem given the global ageing of the population.

A. History and symptoms of the disease

James Parkinson first described PD in 1817 in his monograph “Essay on the Shaking Palsy”. It is a progressive disease normally starting after the age of 55, with an incidence that increases with age. The disease was associated with the degeneration of the mDA neurons of the SNpc more than a century later; but the understanding of the disease made a huge step forward in 1958 after Arvid Carlsson‘s discovery of dopamine in the mammalian brain; this neurotransmitter is indeed implicated in the disease. Following this discovery, SNpc neurons, which synthesize dopamine, were found to project to the striatum, unraveling the organization of the nigro-striatal pathway in the mammalian brain.

This led to the discovery of two key features of the disease. First, the loss of SNpc neurons leads to a severe dopamine deficiency in the striatum, at the origin of the major symptoms of PD. Secondly, oral administration of the dopamine precursor levodopa (L- 3,4- dihydroxyphenylalanine) alleviates most of these symptoms. However, after a few years of treatment, most patients develop involuntary movements called “dyskinesia” which are hard to control and strongly reduce their quality of life. Usually, when the symptoms appear, 60% of the SNpc neurons are lost, alongside with 80% of striatal dopamine; although this may be overestimated: it is now considered that the loss could more likely be around 50-60% for axon terminals and 30% for dopamine neuron cell bodies (Cheng et al. 2010). During the disease, although SNpc mDA neurons preferentially die, cell death is present to a lesser extent in other cells: the VTA mDA neurons degenerate but less than the SNpc ones (Uhl et al. 1985), and other nuclei in the brain are also affected (Hornykiewicz & Kish 1987), as for example the dorsal vagal nucleus (Gai et al. 1992) and the hippocampus (Pereira et al. 2013).

13 The main motor symptoms of the disease are tremor at rest, rigidity and slowness or even absence of movement, postural instability and freezing (Dauer et al. 2003). However these symptoms are not restricted to PD and are found in other pathological conditions, leading to the notion of parkinsonism; nevertheless, PD accounts for 80% of the cases. The tremor usually occurs at rest and decreases during motor activity, which means that it affects daily life to a moderate level. Apart from tremor, other symptoms can be found in the disease, such as stooped posture, olfactory dysfunctions (Doty 2012) or loss of normal postural reflexes, which can lead to falls. Finally, there are also cognitive deficits such as passivity, lack of initiative, depression and, in older patients, dementia. At the cellular level, another aspect of the disease is the presence in neurons of protein inclusions called Lewy bodies. Although it is still debated whether these bodies are a cause or a consequence of the disease, these inclusions contains mostly α-Synuclein and Ubiquitin proteins (Spillantini et al. 1997).

Apart from a symptomatic treatment, there is no cure for PD and most research is aimed at finding ways to prevent the degeneration of SNpc neurons. One difficulty with this approach is the lack of understanding of the molecular mechanisms implicated in the specific degeneration of SNpc neurons.

B. Origin of the disease

The origin of PD is still largely a mystery in spite of the many studies devoted to this question. An important issue, mostly unresolved, is the reason why SNpc mDA neurons preferentially degenerate in the disease. However, a staging theory by Braak suggests that the disease originates elsewhere, namely in the olfactory bulb or in the enteric plexuses of the foregut and would propagate in the brain via different routes. A “neurotropic pathogen” would propagate from the nasal or gastric entry point to the dorsal vagal nucleus, the medulla, the pons and finally the midbrain where SNpc neurons degenerate (Hawkes et al. 2009; Dickson et al. 2010).

An important breakthrough was made in 1983 with the discovery that people intoxicated with a toxin: 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) develop a pathology similar to that of PD (Langston & Ballard 1983). This was the first example of how

14 an endogenous toxin could induce a Parkinson-like condition. Several other toxins were then found: paraquat, an herbicide with a structure similar to that of 1-methyl-4-phenylpyridinium (MPP+) the active metabolite of MPTP, or rotenone a mitochondrial poison used as an insecticide. Although human epidemiological studies correlated the frequent contact with pesticides or insecticides with an elevated risk to develop the disease (Tanner 1992) (Allen & Levy 2013), no convincing data implicate neurotoxins as the only cause for sporadic PD. On a lighter note, consumption of coffee or tobacco is negatively associated with the risk to develop the disease (Hernán et al. 2002)! This reinforces the idea that environmental factors modify PD susceptibility (Chin-Chan et al. 2015).

Nevertheless, exogenous toxins are not the only suspects. On the quest to find the cause(s) of PD, science made a big step forward when genetic mutations were associated with the disease. In 1997, mutations in the α-Synuclein gene were shown to cause an inherited form of the disease (Polymeropoulos et al. 1997; Krüger et al. 1998). After this breakthrough several other gene mutations causing PD were found: Parkin (Kitada et al. 1998), Ubiquitin C-Terminal Hydrolase-L1 (UCH-L1) (Leroy et al. 1998), DJ-1 (Bonifati et al. 2003), PTEN- induced putative kinase 1 (Pink1) (Beilina et al. 2005), Leucine-Rich Repeat Kinase 2 (Lrrk2) (Zimprich et al. 2004; Gilks et al. 2005), Vacuolar Protein Sorting 35 (VPS35) (Zimprich et al. 2011) and ATPase Type 13A2 (ATP13A2) (Ramirez et al. 2006). These genetic mutations are all associated with familial forms of the disease, which account for 5 to 10% of all cases.

These genes provided the first insight into the etiology and the pathogenesis of the disease by deciphering the different pathways in which the corresponding proteins are involved. These studies led to a new field of research on the biochemistry of Parkinson.

C. The biochemistry of Parkinson disease

The biochemistry of Parkinson still remains a bit of a mystery, as, although many features were found, it is hard to distinguish the causes from the consequences of the disease. The main mechanisms of neurodegeneration are summarized in Figure 6.

15

Figure 6. Mechanisms of neurodegeneration in PD Adapted from (Dauer et al. 2003)

As stated earlier one of the main hallmark of PD is the presence of protein inclusions in the cytoplasm of neurons; the presence of these so-called Lewy bodies shed some light on a possible pathway of the disease: protein degradation. Indeed protein degradation is perturbed in PD, which correlates well with the fact that UCH-L1 and DJ-1 are related to the Ubiquitin- proteasome pathway (K. D. Wilkinson 2000; Y. Liu et al. 2002; Takahashi et al. 2001). Along with this idea, it is worth mentioning that Lewy bodies are composed of aggregated α- Synuclein, a protein encoded by a gene mutated or overexpressed in some genetic forms of early onset PD. However, it is of note that mutated α-Synuclein has not been found in sporadic PD (P. Chan et al. 1998). The normal function of α-Synuclein is still under investigation, but it seems that the protein modulates synaptic vesicle function and dopamine content (Abeliovich et al. 2000; Bendor et al. 2013; Sidhu et al. 2004) and participates in cellular membrane dynamics. However it is believed that its tendency to misfold and form amyloid fibrils is key to its pathogenic function in Parkinson (reviewed in (Goedert 2001; W. Xu et al. 2015)) This protein has even been proposed to have a prion-like nature, where the misfolded form would induce misfolding of the correctly folded protein (J.-Y. Li et al. 2008; Herva & Spillantini 2014). This could complete the Braak staging theory, if one considers that the “neurotrophic pathogen” is misfolded α-Synuclein (Visanji et al. 2013).

16 In the same order of thoughts, Parkin is an E3 Ubiquitin ligase (Y. Zhang et al. 2000; Shimura et al. 2000) that marks unfolded proteins with the small peptide Ubiquitin, which then targets them to the proteasome to be destroyed (Sherman & Goldberg 2001). Parkin targets have been studied (Petrucelli et al. 2002) and some of them converge onto the Synuclein pathway as well as onto Cyclin E, a protein linked to neuronal apoptosis. These data are still correlative and the pathways by which Parkin induces PD are not yet fully deciphered, although mitochondria and autophagy seem also involved.

Autophagy, a process in which proteins and damaged organelles are recycled, has been linked to PD (Lynch-Day et al. 2012). In PD, but also in Alzheimer and Huntington disease, defective autophagy induces the accumulation of damaged proteins and organelles (Banerjee et al. 2010). Interestingly, several Parkinson genes have been linked to the control of autophagy: α-Synuclein and Lrrk2 (Bandyopadhyay et al. 2007; Alegre-Abarrategui & Wade- Martins 2009), but also Pink1 and Parkin for their role in mitophagy, the autophagy process that specifically degrades mitochondria (Narendra et al. 2008; Matsuda et al. 2010).

Another classical feature of PD is oxidative stress, which is associated with mitochondrial activity. Mitochondria and oxidative stress play key roles in PD, and - as mentioned above - most PD-related toxins target the mitochondria. MPTP for instance, inhibits mitochondrial CI, which correlates with the fact that CI abnormalities were found in PD (Greenamyre et al. 2001). Mitochondria are very active in mDA neurons (due to their high metabolic activity) (Tufi et al. 2014) and thus produce reactive oxygen species (ROS) that can be toxic to the cells. ROS lead to protein oxidation which can induce protein misfolding or aggregation, as it is the case for α-Synuclein (Giasson et al. 2000). They can also induce oxidation in the nucleus, in particular at the DNA level, provoking DNA damage, chromatin relaxation and ultimately apoptosis. In the latter context, it is of note that increased oxidative stress and reduced antioxidant activity were found in the SNpc of PD patients (Jackson-Lewis et al. 2000; Sian et al. 1994). Along this idea, DJ-1 has been linked to the response against oxidative stress (Lev et al. 2006). Likewise, Park1 has been associated to mitochondria, in particular their fission/fusion as well as to Complex I (Buhlman et al. 2014; Morais et al. 2009) it also interacts with Parkin and α-Synuclein (Norris et al. 2015). The complexity of this problem is well illustrated by Lrrk2, the function of which is still not fully understood.

17 Lrrk2 has been associated to α-Synuclein, tau, inflammatory responses, oxidative stress, mitochondrial dysfunction, synaptic dysfunction and autophagy (J.-Q. Li et al. 2014).

Finally, the question of whether or not the mechanisms of cell death are implicated in the disease remains debated. As in post-mortem human SNpc samples, most neurons are already dead (whatever cell death mechanisms are operating), it is difficult to address this question with this material (Dauer et al. 2003). However, it seems that apoptosis pathways are activated in Parkinson disease, with an increased presence of both Caspases 8 and 9 (Viswanath et al. 2001).

D. Animal models of Parkinson disease

PD has only been described in humans, probably because of our increased life expectancy and of the evolution of our brain. However several animal models are available; they recapitulate many symptoms of the disease, but they have the classical limitations of models, as these animals have parkinsonian symptoms, but not PD per se. Nevertheless these models are of prime importance in research, because of the limitations of post mortem tissues mentioned above and of the need to study the brain as part of the organism, a complexity which cannot be found in in vitro models.

There are several genetic models of the disease, for example expression of a human mutated gene, which mimic in mice an early onset familial form of the disease. The most widely used is the A30P mutation of α-Synuclein. This model is however debated, since these mice do not display specific neurodegeneration of the SNpc (Matsuoka et al. 2001; Giasson et al. 2002; M. K. Lee et al. 2002). Nevertheless injection of a lentiviral vector with the mutated protein in the SNpc of rats induced mDA neuron neurodegeneration (Kirik et al. 2002; Bianco et al. 2002). This discrepancy might be explained by species differences or the need to have a sudden burst of the protein to induce cell death; recent data suggest that it might be linked to α-Synuclein conformation (Peelaerts et al. 2015). On the other hand, mice heterozygous or knock-out for various transcription factors show specific degeneration of the SNpc, among which are: Lmx1b (Laguna et al. 2015), Pitx3 (Veenvliet et al. 2013), Nurr1 (Peng et al. 2011; Kadkhodaei et al. 2013) and Engrailed1 (Sgadò et al. 2006; Sonnier et al. 2007). Indeed,

18 En1+/- mice, a model that I have studied during my thesis, recapitulate many motor and non- motor symptoms of the human disease (Sonnier et al. 2007).

Toxin-based animal models are widely used. This has been the case for 6- Hydroxydopamine (6-OHDA) for over 30 years (Ungerstedt 1968). 6-OHDA is selectively taken up by dopaminergic neurons and induces their loss (in particular in the SNpc) through ROS production (Di Salvio, Di Giovannantonio, Acampora, et al. 2010). 6-OHDA can be injected either directly in the SNpc or in the striatum where it induces the retrograde degeneration of the SNpc dopaminergic neurons. However, no Lewy bodies were observed in this model. The toxins paraquat and rotenone are also used, they induce neurodegeneration and the formation of α-Synuclein-positive inclusions, however their specificity to mDA neurons is still debated (Manning-Bog et al. 2002; McCormack et al. 2002; Betarbet et al. 2000). MPTP remains the gold standard for PD studies in animals. In human and monkeys, MPTP produces a strong parkinsonian syndrome, which recapitulates many features of the disease, as well as a positive response to levodopa. This is the reason why MPTP injection in monkeys is the classical phase I test for therapeutics developed and tested in the treatment of PD. MPTP activity seems also linked to ageing, as both older mice and monkeys are more sensitive to the toxin (Rose et al. 2002; I. Irwin et al. 1993; Ovadia et al. 1995). The mechanism of action of MPTP has been well documented: upon systemic injection, it crosses the blood brain barrier and enters dopaminergic cells where it is oxidized into MPP+. MPP+ then inhibits mitochondrial CI (Nicklas et al. 1985) and creates a strong oxidative stress. It is noteworthy that MPP+ has a high affinity for the dopamine transporter DAT, which is specifically expressed by dopaminergic neurons.

Among the main pathways of Parkinson disease, oxidative stress is of prime interest to us as we have observed that Engrailed protects neurons from oxidative stress. A protective role of Engrailed-triggered translation of CI proteins had already been demonstrated (Alvarez- Fischer et al. 2011), but we now provide new data that demonstrate a role of Engrailed at the chromatin and transcriptional levels and propose a hypothesis based on the regulation of the expression of Long Interspersed Nuclear Element (LINEs) retrotransposons by Engrailed.

19 III. LINEs: a class of retrotransposons not quite extinct

Transposable elements were discovered by Barbara McClintock in the 1940s while studying chromosome breakage in maize (McCLINTOCK 1950). At that time, she proposed two main functions for these elements: mutagenesis by insertion and transcriptional control of nearby genes, which led her to call transposons controlling elements. This work earned her the Nobel Prize in 1983. Transposable elements are classified by whether they act via a DNA (DNA transposons) or a RNA (retrotransposons) intermediate (Figure 7). They are further distinguished by whether they encode proteins to mediate their own mobility (autonomous elements) or rely upon proteins encoded by other elements (nonautonomous elements) (Beck et al. 2011). Among retrotransposons, we focused our attention on a class of autonomous retrotransposons, which do not contain a long terminal repeat (LTR): long interspersed nuclear elements (LINEs).

Figure 7. Classes of mobile genetic elements in the human genome Taken from (Beck et al. 2011)

20 A. LINEs mechanism of action

LINEs are 6kb-long when full length and encode two proteins required for their transposition. They compose nearly 20% of the human genome (W. Li et al. 2013). Even if most of them are unable to retrotranspose, because they are truncated due to previous retrotransposition problems, there are around 3000 and 100 active elements in the mouse and human genomes, respectively. They produce a polycistronic transcript composed of a variable 5’UTR sequence defining the LINE-1 family (Tf, Gf or A) (Sookdeo et al. 2013), two open reading frames ORF1 and ORF2, respectively encoding a RNA chaperone protein and an endonuclease plus reverse transcriptase and finally a 3’UTR with a poly(A) tail (Martin & Bushman 2001; Mathias et al. 1991). Their mechanism of action is depicted in Figure 8.

Figure 8. Mechanism of LINE-1 retrotransposition Taken from (Ostertag & Kazazian 2001)

The two proteins (ORF1p and ORF2p) bind to their own polycistronic mRNA creating a ribonucleoprotein (RNP) (Hohjoh & Singer 1996). This RNP translocates to the nucleus, either by a transport mechanism or because of nuclear envelop breakdown (Kubo et al. 2006). The ORF2p endonuclease then creates a nick in the DNA at the consensus sequence AA/TTTT (Feng et al. 1996) where ORF2p-reverse transcriptase retrotranscribes the mRNA, a process ending with a new LINE-1 copy inserted at the site of the DNA break. LINEs also mediate the retrotransposition of several types of so-called non-autonomous elements, mainly small interspersed nuclear elements (SINEs), variable number of tandem repeats and Alu

21 sequences (SVA) (Esnault et al. 2000; Kajikawa & Okada 2002; Dewannieux et al. 2003; Hancks et al. 2012; Raiz et al. 2012), which can also have deleterious activities (Vogt et al. 2014; Kaer & Speek 2013; Ade et al. 2013).

Integration is called target primed reverse transcription (TPRT) and ORF2p provides all the necessary functions including DNA nicking, DNA priming and reverse transcription (Feng et al. 1996). However, this mechanism only applies to the small number of full length LINEs. Most LINEs are 5’ truncated, because of the poor processing of retrotranscription (Lander et al. 2001). These truncated copies are therefore inactive for integration but still have deleterious activities. First, ORF2 can create DNA strand breaks and mutagenize active genes through full or partial insertion (S. Kim et al. 2015; H. Xu et al. 2014; Kines et al. 2014; Morisada et al. 2010; Gasior et al. 2006) or unfaithful DNA repair (K. Han et al. 2008). Secondly, LINEs have several promoters, including antisense ones (J. Li et al. 2014; Wolff et al. 2010; Speek 2001). This implies that truncated LINEs can modify the expression of nearby genes by adding a new sense promoter or creating an antisense long non coding RNA (Kazazian et al. 1988; Skowronski & Singer 1985; H. Kim et al. 2015). Third, partial LINE-1 transcripts can have regulatory functions (Feschotte 2008; Rebollo et al. 2012; Faulkner et al. 2009). Several strategies exist to counteract the harmful effects of LINE-1 expression.

B. Defenses against the LINEs

A first defense strategy is to repress their transcription through epigenetic mechanisms. LINEs contain CpG-rich sequences in their 5’UTR (Yu et al. 2001; Coufal et al. 2009), the methylation of which is associated to LINE-1 repression. Moreover, loss of the DNA-binding protein MeCP2 leads to increased LINE-1 transcription and transposition, thus further suggesting a causal link between DNA methylation and LINE-1 repression (Muotri et al. 2010). In addition to hypermethylation, LINEs are often associated with the condensed state of chromatin called heterochromatin (Elsässer et al. 2015; Chow et al. 2010; Garcia- Perez et al. 2010). This modification can also repress LINE-1 transcription, as heterochromatin regions are, in general, poorly accessible to transcription complexes (Saksouk et al. 2015). There are two main type of heterochromatin associated with multiple methylation of either Histone 3 lysine 9 (H3K9) for chromocenters (Muramatsu et al. 2013; F. L. Chan & Wong 2012) or of H3K27 for peri-nucleolar/peri-nuclear heterochromatin

22 (Bártová et al. 2008). Repetitive elements are often found in H3K27me3 heterochromatin (Pauler et al. 2009; Bulut-Karslioglu et al. 2012). However, these mechanisms are not sufficient to prevent the deleterious activity of LINEs, thus justifying the existence of other defense mechanisms.

A second line of defense is the Piwi/piRNA pathway. Piwi (P-element induced wimpy testis in Drosophila) is a member of the Argonaut family of proteins and piRNAs are 24-32 small RNAs, bound by Piwi. First In the Drosophila, where they were first discovered, piRNAs originate from long genomic piRNA clusters, mostly composed of remains of all types of transposons (Vagin et al. 2006; Brennecke et al. 2007; Houwing et al. 2007). Most of these piRNAs are antisense to transposon transcripts and likely act by enabling Piwi proteins to target those transcripts for degradation (Saito & Siomi 2010). The importance of this pathway was demonstrated in the Drosophila male germline, where Piwi loss of function leads to a derepression of the transposons and an insertion of their copies all over the genome, leading to infertility (Kalmykova et al. 2005). By the same token, Piwi inactivation can also serve the creation of adaptive mutations (Piacentini et al. 2014). The Piwi pathway is illustrated in Figure 9.

Figure 9. piRNA biogenesis and amplification (Ping-Pong cycle) in the Drosophila and in the mouse. Taken from (Ishizu et al. 2012)

23 In Drosophila and mammals, piRNAs are transcribed, trimmed and 2’-O methylated at their 3’ end by the Pimet protein associated with Piwi (Horwich et al. 2007; Saito et al. 2007). Hen1, the equivalent of Pimet, was first described in Arabidopsis, where unlike in animals, it additionally methylates miRNAs (X. Chen 2005; Jay et al. 2011). The piRNA defense can be amplified by a mechanism called the Ping-Pong cycle (illustrated in Figure 9) (Brennecke et al. 2007; Gunawardane et al. 2007; Houwing et al. 2008). After being loaded on Piwi protein, the piRNA binds in an antisense orientation to a LINE-1 transcript bound to Ago3. This induces the cleavage of the target transcript, creating a new antisense piRNA (secondary piRNA), which, in turn, induces another Ping-Pong cycle with the Aubergine (Aub) protein (Saito et al. 2006; Gunawardane et al. 2007; Nishida et al. 2007). In mice, the mechanism is almost identical, but involves three different Piwi proteins: MIWI, MILI and MIWI2. The two former associate with primary piRNAs and the latter is the equivalent of Ago3 and mostly binds to secondary piRNA (Siomi et al. 2011).

The Piwi/piRNA pathway primarily works by RNA interference but there is also an epigenetic component to it. In mice, the piRNA/MIWI2 complex, bound to a nascent transcript, induces de novo DNA methylation through the recruitment of a DNA-methyl- transferase (DNMT3a). In Drosophila, the piRNA/Piwi complex induces de novo H3K9 methylation recruiting the Heterochromatin Protein 1 (HP-1) (Weick & Miska 2014; Kuramochi-Miyagawa et al. 2008; Ross et al. 2014; Huang et al. 2013). HP-1 is an evolutionary conserved protein which binds to H3K9me3 and induces the formation of higher-order chromatin structure (Nishibuchi & Nakayama 2014). Therefore, guided by a piRNA, Piwi is also able to repress retrotransposon expression through the nucleation of repressive heterochromatin.

These mechanisms are considered to be efficient enough to fully block LINE-1 retrotransposition in the whole organism, with the exception of the germ cells where retrotransposition was observed, creating genetic diversity in response to stress in flies (Ishizu et al. 2012). Importantly in the present context, the presence of piRNAs was also demonstrated in the central nervous system with several criteria; their size (24 to 27nt), binding to MIWI and 2’-O methylation on their 3’ end (E. J. Lee et al. 2011; Rajasethupathy et al. 2012). Functionally, brain piRNAs have been implicated in the control of memory-

24 related plasticity (Rajasethupathy et al. 2012). The presence of piRNAs in the brain could be a sign that LINEs may have neurophysiological functions.

C. Retrotransposition in brain neural progenitors

The use of a LINE-1 expression cassette allowing one to follow retrotransposition events through GFP expression (Freeman et al. 1994; Moran et al. 1996), has indicated that LINE-1 activity can be induced in the brain, primarily in neurons as glial cells show little retrotransposition (Muotri et al. 2005). In particular, LINEs are highly expressed in neural progenitor cells (NPCs) compared to neural stem cells, both in vitro and in vivo, in the mouse, the rat and the human (Coufal et al. 2009; Muotri et al. 2005). This suggests that neural stem cell differentiation into neurons is accompanied by a strong increase in retrotransposition. Furthermore LINE activity perturbs the expression of close-by genes through promoter addition and/or epigenetic remodeling (Muotri et al. 2005).

Another study, comparing LINE-1 copy numbers in different post-mortem tissues, showed a significantly higher LINE-1 content in the brain than in the heart or the liver. Within the nervous system, the dentate gyrus (in the hippocampus), the frontal lobes and the spinal cord contain the highest numbers of LINE-1 copies (Coufal et al. 2009). Still in the hippocampus, the subgranular zone, an important neurogenic niche (P. S. Eriksson et al. 1998) shows high levels of retrotransposition. The amount and role of retrotransposition in the brain are poorly understood (Thomas et al. 2012). It could be partially explained by the presence of YY1 and Sox2 transcription factor binding sites in the promoter sequence of the transposon (Athanikar et al. 2004; Tchénio et al. 2000). Sox2 was proposed to inhibit LINE-1 expression in neural stem cells, through its participation in a complex with histone deacetylase 1 (HDAC1) and methylated H3K9 (Muotri et al. 2005). In this scheme, inhibition would be released upon cell commitment to a neuronal lineage when Sox2 is replaced by NeuroD, thus allowing neuronal diversity through LINE-1 insertion (Kuwabara et al. 2009).

Environment also plays a role in retrotransposition. Mice in an enriched environment, for example having access to a running wheel, display threefold more retrotransposition than mice in a sedentary environment (Muotri et al. 2009). Similarly, stress (Ponomarev et al.

25 2010; Hunter et al. 2012) and alcohol (Ponomarev et al. 2012) can alter LINE-1 expression in the brain. Retrotransposition in the brain, which contains as much as 86 billion neurons (Azevedo et al. 2009), indicates how specific sub-populations of neurons could form mosaics at genetic and physiological levels. Recent findings suggest that LINE-1 expression might also be associated with diseases either neurological or affecting other systems such as hemophilia (Hancks & Kazazian 2012).

D. The implication of LINEs in diseases of the brain and in evolution

A pathophysiological implication of LINE-1 elements has been studied in two diseases affecting the central nervous system (CNS): Rett syndrome and Ataxia Telangiectasia.

Rett syndrome (RTT) is a neurodevelopmental syndrome caused by mutations in the gene MeCP2, which is located on chromosome X. Boys do not survive and the disease only affects heterozygous girls (Amir et al. 1999), who develop normally until the age of 6 to 18 months when their development starts to stall before regressing (Chahrour & Zoghbi 2007). The classical symptoms include autism, loss of speech, anxiety and disappearance of coordinated hand movement (replaced by stereotyped repetitive movements) that eventually lead to motor deterioration. The link between MeCP2 mutation and RTT symptoms is not yet perfectly understood. However, LINE-1 involvement is plausible, as MeCP2 inhibits retrotransposition, likely by binding CpG-rich regions in their promoters (Coufal et al. 2009; Muotri et al. 2010). This hypothesis is further supported by the fact that MeCP2 deficient mice display an increased retrotransposition not only in neural progenitor cells (Muotri et al. 2010), but also in mature neurons (Skene et al. 2010). A similar increase has been observed in human induced pluripotent stem cells (iPSCs) from RTT patients that display a 2.5 fold increase in retrotransposition compared to age-matched controls (Muotri et al. 2010). These data, although very suggestive, are correlative and there is yet no demonstrated causal link between retrotransposition and the symptoms of the disease.

Ataxia telangiectasia (A-T) is a hereditary disease caused by a mutation in the gene Ataxia Telangiectasia Mutated (ATM). The symptoms include loss of motor function as well as many other illnesses, including diabetes, lymphomas, immunodeficiency and paralysis, the

26 latter owing to cerebellar degeneration. Patients most often die between their teens and twenties. ATM is a serine/threonine protein kinase implicated in the DNA damage response (Savitsky et al. 1995). It is one of the first factors which bind to DNA double strand breaks and initiates a signalling cascade through its kinase activity. This cascade eventually leads to DNA repair or to p53 induction, which activates the DNA damage checkpoint and possibly cell cycle arrest (in dividing cells) and apoptosis (Bar-Shira et al. 2002; Madabhushi et al. 2014; Mossalam et al. 2012). In ATM-deficient cells, DNA repair is impaired, leading to increased genomic instability. Of note, for post-mitotic cells (neurons), homologous recombination - the most efficient DNA repair mechanism - is not possible because the cells do not undergo divisions, making them more sensitive to DNA damage (Iyama & Wilson 2013). Interestingly, increased retrotransposition was found in NPCs from ATM-deficient mice and in post mortem brain tissues from patients (Coufal et al. 2011). The authors discovered that, in mice and in patients, ATM deficiency also leads to longer LINE-1 insertions, suggesting that ATM may impact retrotransposition by an unknown mechanism (Coufal et al. 2011). The exact mechanism of neuron degeneration is not really known, but the hypothesis is that, due to ATM loss, DNA double strand breaks are not detected in NPCs and that damaged NPCs, instead of going into apoptosis, differentiate into neurons, more sensitive to oxidative stress and therefore sensitive to degeneration, leading to the A-T phenotype.

High LINE-1 retrotransposition was also found in schizophrenia, where an increased number of insertions was found in neurons from the prefrontal cortex and in patient-derived iPSCs (Bundo et al. 2014). When the authors studied the insertion loci, they observed that they were enriched in genes encoding factors involved in synaptic functions as well as genes associated with schizophrenia. Although not directly related to retrotransposition, the tau protein, implicated in Alzheimer disease, has been shown to induce genome wide chromatin relaxation (Frost et al. 2014), and it is tempting to speculate that this could induce LINE-1 activation. Similarly LINE-1 retrotransposition increases during ageing (W. Li et al. 2013). Finally LINE-1 hypomethylation, which could induce their activation, has been observed in cancer (Kitkumthorn & Mutirangura 2011).

Another aspect of the study of LINEs is their possible link with hominine evolution. A striking correlation was seen between the evolution of LINE-1 families and that of hominines, it seems that the arrival of new species correlated with the emergence of a new family of

27 active LINEs (J. Lee et al. 2007; Smit et al. 1995). Could LINEs be drivers of evolution towards humanity? Albeit novel, this correlation could prove to be fundamental for the study of evolution. Conversely, the Apobec3 family is a group of cytidine deaminases, which inhibit retrotransposons and retroviruses by replacing a cytidine by a uridine in RNA and DNA (Koito & Ikeda 2013). This family of proteins links LINEs to evolution as well, since their number increases from one in the mice to seven in primates and humans (Koito & Ikeda 2013). This could correlate with the hypothesis that evolution goes along with reduced genetic variability, and could perhaps be driven by a decrease in genetic instability (Erwin et al. 2014). This may seem to be a little paradoxical, but one could imagine that the appearance of a new LINE-1 family is associated with arrival of new species and that it is then stabilized by increased LINE-1 inhibition.

Although correlative, these data challenge the initial view that LINEs are useless selfish genes constituting junk DNA. In marked contrast, a growing body of evidence suggests that LINEs may play a role in the physiology and the pathology of several organs and particularly of the brain. This new avenue promises to be very stimulating, provided that the role of retrotransposition in post-mitotic neurons is understood. Whatever time and energy this will take, one can already see an end to the myth of neural genome uniformity and stability.

The aim of the work presented here was to link these three elements: Engrailed, Parkinson disease and LINEs and to study Engrailed as a possible therapeutic protein for Parkinson disease. The first approach was to study En1+/- mice to analyze the mechanisms of Engrailed-mediated neuroprotection, particularly at the transcriptional and chromatin levels. The second part was to pursue the hypothesis of LINE-1 activation in En1+/- mice and in models of oxidative stress and, if so, to verify its impact on neurodegeneration. Finally the project aimed at deciphering the interaction between Engrailed and retrotransposition and to ask if Engrailed neuroprotection could not be mediated, at least partially, by LINE-1 repression. As the reader will, hopefully, be convinced, these three lines of research strongly suggest that Engrailed is a therapeutic agent worthy to explore in the treatment of PD.

28

29

ARTICLE 1 - Engrailed Homeoprotein, an Anti- Ageing Factor that Protects Mesencephalic Dopaminergic Neurons from Oxidative Stress

30

31 ENGRAILED HOMEOPROTEIN, AN ANTI-AGEING FACTOR THAT PROTECTS MESENCEPHALIC DOPAMINERGIC NEURONS FROM OXIDATIVE STRESS

Hocine Rekaik1,2, François-Xavier Blaudin de Thé1,2, Julia Fuchs1, Olivia Massiani-

Beaudoin1, Alain Prochiantz1,* and Rajiv L. Joshi1,

1Centre for Interdisciplinary Research in Biology (CIRB) Labex Memolife, CNRS UMR 7241 / INSERM U1050, Collège de France, 11 place Marcelin Berthelot, 75231 Paris Cedex 05, France

2Co-first author

*Correspondence: [email protected] (A.P.)

Cell Reports (in press)

32 I. Summary

Engrailed homeoproteins (En1/En2) are expressed in adult dopaminergic neurons of the substantia nigra. In En1 heterozygous mice, these neurons start dying at 6 weeks, correlatively with the deregulation of genes controlling chromatin structure, the DNA damage response and apoptosis. Analysis of the nuclear phenotype reveals that they develop signs of progressive oxidative stress. Accordingly, strong oxidative stress inflected to wild-type dopaminergic neurons induces similar changes, including DNA strand breaks, modification of several heterochromatin marks and nucleolar stress indicators, as well as expression of apoptosis modulators. En1/En2 are biochemically equivalent transducing proteins previously used to antagonize neuronal death in mouse models of PD. We now show that En2 treatment, after an acute oxidative stress reverses the nuclear/nucleolar phenotypes, decreases the number of phosphorylated H2AX histone foci, represses cell cycle and pro-apoptotic gene expression and protects against oxidative stress-induced apoptosis.

33 II. Introduction

The two homeoprotein transcription factors En1 and En2 (collectively Engrailed or En1/2) play equivalent roles in the development of the mesencephalic dopaminergic (mDA) neurons (Joyner 1996; Simon et al. 2001). In the adult, Engrailed is required for the survival of these neurons in the substantia nigra pars compacta (SNpc) and the ventral tegmental area (VTA) (Albéri et al. 2004; Sgadò et al. 2006). In the En1-/+; En2+/+ (En1+/- thereafter) mouse (Swiss OF1 background), mDA neurons degenerate progressively starting at 6 weeks of age (Sonnier et al. 2007). Neuronal death is higher in the SNpc than in the VTA, a selective vulnerability as described in Parkinson disease (PD). Furthermore, En1+/- mice display motor and non-motor phenotypes reminiscent of symptoms observed in PD (Sonnier et al. 2007).

These observations suggest that Engrailed might be in the PD pathway, in line with a possible association between EN1 polymorphisms and the risk to develop PD (Fuchs et al. 2009; Haubenberger et al. 2011; Rissling et al. 2009) even though this was not confirmed in genome-wide association studies. Taking advantage of the internalization properties of Engrailed (Joliot & Prochiantz 2004), it was shown that En1 and En2 proteins save mDA neurons in the En1+/- mouse (Sonnier et al. 2007) and in wild-type (wt) mice exposed to striatal 6-hydroxy-dopamine (6-OHDA), systemic 1-methyl-4-phenyl-1,2,3,6- tetrahydropyridin (MPTP) or the toxic A30P variant of α-Synuclein (Alvarez-Fischer et al. 2011). In all cases, En1 and En2 have an equivalent capacity to protect mDA neurons from cell death.

Engrailed has multiple functions acting as a transcription factor and also regulating the translation of capped mRNAs (I. Brunet et al. 2005). Among its translation targets are nuclear-encoded mitochondrial mRNAs that surround the mitochondria. Engrailed transduction up-regulates the translation of mitochondrial CI subunits, Ndufs1 and Ndufs3, and increases ATP synthesis (Alvarez-Fischer et al. 2011; Stettler et al. 2012). Blocking Engrailed-induced Ndufs1 translation in vivo abolishes the protective effects of Engrailed against MPTP, a complex-I-specific neurotoxin (Alvarez-Fischer et al. 2011).

34 Neuroprotection by Engrailed in a variety of models raises the possibly that it acts not only locally (e.g. CI activity), but also more globally and that its decrease may render the cells more sensitive to other mutations or to oxidative lesions, either acute or resulting from normal ageing. Indeed ageing is a major risk factor for PD, even in its familial forms. To better identify the pathways involved, we used deep sequencing and found that losing one En1 allele leads to the up-regulation in the SNpc of many genes associated with the DNA damage response and chromatin structure.

This led us to study the protective activity of Engrailed against oxidative stress at the nuclear level. We find that mDA neurons from En1+/- mice present signs of premature ageing and are more sensitive to an acute oxidative stress (intranigral 6-OHDA injection). Moreover, Engrailed gain of function protects against oxidative stress-induced apoptosis, restores a healthy chromatin structure and leads to a reduction in the number of phosphorylated histone H2AX (γ-H2AX,) foci, suggesting a repair of stress-induced DNA strand breaks (DSBs) (Löbrich et al. 2010; Morrison & Shen 2005). It is thus proposed that Engrailed is an anti- ageing protein protecting against acute or age-dependent oxidative stress.

35 III. Results

A. mDA neuron gene expression in the SNpc of En1+/- and wt mice

To investigate the molecular mechanisms that lead to mDA neuron degeneration in En1+/- mice, we performed a RNA-sequencing (RNA-seq) analysis. Total RNA from SNpc was obtained by laser capture micro-dissection following tyrosine hydroxylase (TH) immunostaining of midbrain sections (Figure 1A). Comparable numbers of reads were obtained by RNA–seq in samples from both wt and En1+/- animals and we identified 989 differentially expressed genes (p values < 0.05; Table S1 lists 1837 differentially expressed genes with p < 0.1). This analysis was performed on 6 week-old animals because, even though cell death initiates at 6 weeks in the in En1+/- mice, all neurons are still present at that age (Sonnier et al. 2007). RNA-seq data analyzed using gene set enrichment analysis (Pathway Studio Ontology, Pathway Studio software) indicated that the three most represented entities within significantly differential enriched ontologies where DNA repair (p = 0.002), chromatin remodeling (p = 0.004) and transcription factors (p = 0.007) (Figure 1B). Cell signalling pathway analysis also revealed apoptosis regulation as differentially enriched (p = 0.01). Genes within these ontologies and pathways found significantly different between wt and En1+/- mice were ranked by increasing p values and Figure 1C highlights the genes with highly significant differences in read numbers.

As shown in Figure 1B, transcription factors genes represented the most abundant category followed by genes involved in DNA damage response and chromatin modification. Figure S1A ranks (by p values) the "transcription factor" genes with a modified expression in the En1+/- mice. However, in an Engrailed gain of function experiment (Engrailed infusion in the SNpc of wt mice), transcription factor genes were also regulated whereas the two other families were absent (Figure S1B). DNA damage response and chromatin modifying genes are thus directly connected to the pathological phenotype resulting from the loss of one En1 allele. Quantitative RT-PCR using total RNA from the SNpc (Figure S1C) confirmed that 6 week-old En1+/- mice display significantly altered expression of several genes related to DNA damage, chromatin remodeling and apoptosis (Asagoshi et al. 2010; Burikhanov et al. 2014; Choi & Bakkenist 2013; Hong et al. 2008; Ma & D'Mello 2011; Melis et al. 2013; Y. Yang et al. 2013; G. Zhang et al. 2013)

36 B. DNA strand breaks and modified chromatin marks in wt and En1+/- mDA neurons

In view of the RNA-seq data, we examined mDA neurons in the SNpc of En1+/- mice (between 6-8 weeks of age) for signs of increased DNA damage by immunostaining for the DSBs marker γ-H2AX (Löbrich et al. 2010; Morrison & Shen 2005). This revealed the presence of multiple γ-H2AX foci in about 16% of TH+ neurons in the SNpc of these mice (Figures 2A,B). Although 16% may seem a small percentage, it must be recalled that cell death is progressive in the mutant and that only 40% die within 48 weeks (Sonnier et al. 2007). Furthermore, the presence of DSBs suggests a DNA repair problem (Suberbielle et al. 2013), but not necessarily rapid death. Indeed, in the En1+/- mutant, almost no death occurs between 24 and 48 weeks (Sonnier et al. 2007) even though 16% of the cells have multiple γ- H2AX foci at 24 weeks (Figure S1D). In wt mice, about 98% of mDA neurons displayed a single ring of γ-H2AX seemingly surrounding the nucleolus (Figures 2A). This γ-H2AX ring in healthy neurons was seen throughout the brain, thus not specific of mDA neurons from the SNpc and its presence was not correlated with the loss of one En1 allele. We also observed that TH+ neurons in the SNpc appear to be more vulnerable to DNA damage than TH+ neurons of the VTA, where γ-H2AX staining was similar between wt and En1+/- mice (Figure 2B). This difference is consistent with the fact that mDA neurons from the VTA are less sensitive to the loss of one En1 allele (Sonnier et al. 2007).

In accordance with the RNA-seq data suggesting remodeling of chromatin in the En1+/- mice, 8 week-old mutants also exhibit changes in chromatin marks. Figure 2C-E illustrates that the pattern of H3K27me3 (K27 trimethylated histone H3) perinucleolar and perinuclear staining is changed in a significant fraction of En1+/- TH-positive neurons. Similarly, the size distribution of the DAPI-dense regions (chromocenters) corresponding to specific spots of heterochromatin (Guenatri et al. 2004) shows a reduction in the percentage of large spots in the En1+/- mouse (Figure 2F). Changes in heterochromatin structure are often associated with the expression of partially silenced genes, among which Long Interspersed Nuclear Elements (LINEs) (Beisel & Paro 2011; Guetg & Santoro 2012; Padeken & Heun 2014; Wutz 2011). Figure 2G shows that LINEs are expressed in the SNpc, an expression significantly up-regulated in the mutant. Finally, in agreement with the

37 induction of apoptosis genes (Figure 1C), some TH+ cells in the SNpc of En1+/- mice were co-stained by activated caspase-3, an observation never made in wt littermates (Figure 2H).

C. Induction of DSBs and degeneration by 6-OHDA in wt and En1+/- mDA neurons

To investigate the link between degeneration and DNA damage, we applied an acute oxidative stress to mDA neurons by injecting 6-OHDA (a superoxide producing drug specifically captured by mDA neurons) directly into the SNpc. This induced the formation of multiple γ-H2AX foci (Figure 3A). Quantification revealed that abnormal foci appear in about 26% of TH+ neurons in the SNpc of the ipsilateral (injected) side within 6h, sham- injected mice only showing the single perinucleolar ring of γ-H2AX (Figure 3B). TH+ neurons in the contralateral SNpc and in the VTA were spared, the latter presumably because they do not express the dopamine (DA) transporter necessary for 6-OHDA uptake (Di Salvio, Di Giovannantonio, Acampora, et al. 2010).

Counting TH+ neurons indicated that about 35 % of TH+ neurons were lost in the ipsilateral SNpc at 6h after 6-OHDA injection (Figure 3C). This rapid loss was also reflected by a reduction in TH protein and mRNA (En1 mRNA as well), 6h post 6-OHDA injection (Figures S2A,B). Finally, 6-OHDA-induced TH cell loss possibly involved caspase-3, since activated caspase-3-positive TH neurons were detected in the ipsilateral SNpc 6h after 6- OHDA injection, and never on the contralateral side (Figure S2C).

We next injected 6-OHDA in En1+/- mice and analyzed the induction of γ-H2AX foci in mDA neurons. The percentage of TH+ neurons with γ-H2AX foci in the 6-OHDA-injected side was significantly higher (51%) than in wt mice (Figure 3A, B). This higher level of 6- OHDA-induced DSBs in En1+/- mice correlated with an increased in the loss of TH+ neurons (60%) on the SNpc ipsilateral side (Figure 3C). This indicates that endogenous En1 has a protective effect against an acute oxidative stress. Because En1 is primarily a repressor gene (Tolkunova et al. 1998), we speculated that the infusion in the SNpc of an activator form of Engrailed, composed of the En1/2 HD (for target recognition) and of four copies of the VP16 transcriptional activator domain of the Herpes Virus (EnHD-VP64), may have an anti-En1

38 activity. EnHD-VP64 infusion indeed leads to mDA cell death (Figure S3A,B) and to an increase in the number of γ-H2AX foci (Figures S3C). This experiment supports the view that Engrailed is a survival factor and strongly suggests that part of its activity takes place at a transcription level.

D. Oxidative stress induces changes in heterochromatin marks

The enhanced sensitivity of SNpc mDA neurons in En1+/- mice to oxidative stress led us to look, for comparison, at various heterochromatin marks that may be modified upon 6- OHDA injection in the SNpc of wt mice. H3K27me3 showed dense perinucleolar and perinuclear staining in TH+ neurons of sham-injected mice whereas SNpc TH+ neurons of 6- OHDA-injected mice displayed a diffuse nucleoplasmic staining (Figure 4A). The relative fluorescence intensity of perinucleolar H3K27me3 staining was measured along an axis running through the nucleolus (dotted line in Figure 4A) showing that the enhanced perinucleolar ring present in TH+ neurons of sham-injected mice was lost upon 6-OHDA injection (Figure 4B, left). The same observation was made for perinuclear H3K27me3 staining. Indeed, the ratio of H3K27me3 fluorescence intensities between the peripheral nuclear lamina and the nuclear stroma of TH+ neurons showed a drop from 1.4 to 1 six hours following 6-OHDA (Figure 4B, right).

The loss of peripheral H3K27me3-enriched heterochromatin in TH+ neurons upon 6- OHDA injection was accompanied by a disruption of H3K9me3 (K9 trimethylated histone H3) and MeCP2 staining (Figure S2D,E), the loss of the neat and defined Lamin B2 staining (Figure S2F) and a change in the size distribution of DAPI-dense spots (Figure S2G). MeCP2 that binds methylated CpGs is also found at the level of the chromocenters in the brain (Deng et al. 2010; Garg et al. 2013) and changes in its staining might reflect guanine oxidation in CpGs (Skene et al. 2010). It is of note that the loss of MeCP2 in SNpc mDA neurons compromises the nigrostriatal dopaminergic pathway (Gantz et al. 2011).

Perinucleolar γ-H2AX loss upon 6-OHDA injection, possibly reflected a nucleolar stress. We thus examined whether oxidative stress induced by 6-OHDA is associated with nucleolar disruption. Midbrain sections from sham-injected mice showed a dense nucleolar-specific

39 nucleolin staining in 70% of TH+ neurons (Figure 4C). This number dropped to 30% in the ipsilateral SNpc of 6-OHDA-injected mice, with various levels of nucleolin in the nucleoplasm outside of the nucleolus (Figure 4C,D) and a strong up-regulation of ribosomal pre-45S RNA (Figure 4E) which can be attributed to nucleolar damage and heterochromatin loss (Guetg et al. 2010; Guetg & Santoro 2012; Larson et al. 2012).

The similar increase in the number of γ-H2AX foci in En1+/- mice and 6-OHDA- injected wt mice led us to study the nucleolar structure of TH+ cells in En1+/- mice (in absence of 6-OHDA). Nucleolin staining showed signs of nucleolar disruption after one year but not at 6 weeks (Figure 4F). However, the expression of genes involved in the organization of the nucleolus in the initial RNA-seq (Figure 1) and the qRT-PCR analysis of pre-45S rRNA suggest a change in nucleolar physiology in 6-8 week-old En1+/- mice (Figure S3D,E).

Nucleolar disruption potentially links to the p53 pathway for senescence and apoptosis (Krenning et al. 2014; Rieker et al. 2011; Sloan et al. 2013; Teng et al. 2013). qRT-PCR on total RNA from SNpc of sham- and 6-OHDA-injected mice showed that the level of p53 transcripts was increased by 50% upon 6-OHDA injection in parallel with that of p21, a p53 transcriptional target (Sperka et al. 2012) (Figure 4G). Finally and in line with data showing that retrotransposition in neurons is regulated by MeCP2 (Muotri et al. 2010), LINE-1 mRNA levels were increased upon 6-OHDA injection (Figure 4H).

E. Engrailed protects mDA neurons from oxidative stress

The above results suggest that the loss of one En1 allele and the acute oxidative stress imposed to wt mDA neurons in vivo lead to very similar changes in nuclear phenotypes. This, added to the phenotype of EnHD-V64 infused mice and the enhanced sensitivity of En1+/- mice to oxidative stress, points toward Engrailed as an anti-ageing agent able to protect mDA neurons against naturally occurring (in the course of ageing) or acute oxidative stress. To evaluate this possibility, wt mice were unilaterally injected with 6-OHDA, re-injected 30 min later with vehicle (sham) or En2, and analyzed 6h, 24h or 7 days later. Analysis at 24h showed a dramatic loss of TH+ neurons in the SNpc of 6-OHDA/sham injected (ipsilateral

40 side) and a marked protection against this loss in the SNpc of 6-OHDA/En2 injected mice, in particular in the its most dorso-lateral part (Figure 5A). Engrailed injection 24h (instead of 30 min) after 6-OHDA treatment and analyzed 6h later showed no recovery, demonstrating that the enhanced number of TH+ cells reflects actual survival and not TH re-expression (Figure S3E).

En2-mediated neuroprotection was quantified at various times post injections (Figure 5B). Protection was clear at 24h and still visible at 7 days post injections with 40% and 20% of surviving neurons in En2 and sham injected mice, respectively. As observed in other models of apoptosis (Casafont et al. 2011; Krenning et al. 2014; W. Li et al. 2013), 6-OHDA- induced apoptosis is paralleled by an abnormal expression of several cell cycle markers (Figure S4A-D). We thus examined the expression of Cyclin A in the En2-mediated protection experiment and found that Cyclin A expression still detected in En2-treated TH+ neurons 6h post injections had almost completely disappeared from the surviving neurons at 24h (Figure 5C).

The partial rescue of TH+ neurons by En2 in the 6-OHDA model is associated with the reappearance of normal staining patterns for H3K27me3, nucleolin and γ-H2AX (Figure 5D). The percentage of TH+ neurons with "wild-type" perinucleolar H3K27me3 staining increased from 20% to 37% at 6h and reached 60% and 80% at 24h and 7 days, respectively. Wild-type nucleolin pattern took longer to reappear as recovery was observed only at 24h with little changes between 24h and one week. Finally the decrease in the number of γ-H2AX foci was slower and an abnormal percentage of 20% of neurons with more than two spots showed full recovery only at 7 days.

The rescue of TH+ neurons also correlated with a significant increase in TH mRNA expression and a decrease in the level of pre-45S rRNA already at 6h after En2 injection (Figures 5E). Finally, the expression of selected genes related to cell cycle and apoptosis (Albéri et al. 2004; Lim & Kaldis 2013; Smith et al. 2003; Tapias et al. 2014; Tilstra et al. 2012; W. Wang et al. 2009) up-regulated either in the En1+/- mouse model or upon 6-OHDA injection were significantly repressed in the SNpc of 6-OHDA/Engrailed, compared to 6- OHDA/sham mice (Figure 5E). These analyses were done at 6h because, at this time point,

41 sham and En2-injected animals still have the same number of mDA neurons (Figure 5B), thus allowing a strict comparison. But given the fact that, even with En2, at least 50% of the cells are irreversibly engaged in an apoptotic pathway, the numbers in Figure 5E probably underestimate the En2-controlled repression of deleterious genes in the long-term mDA neuron survivors.

As mentioned, VTA neurons are less sensitive to the loss of one En1 allele. This might be due to the expression of Otx2 specifically in the VTA (Di Salvio, Di Giovannantonio, Omodei, et al. 2010) and to its enhanced expression in the En1 mutant (Figure S1A). We thus verified if Otx2 injected in the SNpc also protected against 6-OHDA. Figure 5F demonstrates that this is the case raising the issue of whether Engrailed could act as a more general survival factor for midbrain neurons exposed to an oxidative stress. To address this point, E14.5 midbrain neurons that express Engrailed but of which only 1-2% are dopaminergic were exposed to H202 with or without En2. Figure S4E-F illustrates that En2 decreases the number of DSBs induced by H2O2 in parallel with a decrease in the formation of comet tails that are signs of DNA damage.

F. Engrailed activates anti-apoptotic pathways

The results described above suggested that, even injected 30 min after 6-OHDA, En2 modifies the expression of "survival" genes. To identify some of them, the SNpc RNA from 6-OHDA/sham or 6-OHDA/En2 mice at 6h was sequenced. We used Pathway Studio annotations to select the genes in the chromatin remodeling, DNA damage, apoptosis and cell cycle pathways. Figure 6A,B ranks by order of significance the genes showing the highest differences in the number of reads. Some genes appearing in a given category could also be placed in other ones. For example Pathway Studio places Lamin B1 in the "apoptosis" category whereas it could as well be in "chromatin remodeling". Similarly, Gadd45b appears in "cell cycle" but it also acts in the epigenetic and apoptosis pathways (De Smaele et al. 2001; Wu & Y. E. Sun 2009). Figure 6C illustrates qRT-PCR confirmation of enhanced expression for genes taken as top representative of the 4 categories mentioned above. Figure 6D where cycloheximide was added to En2 in the rescue experiments demonstrates that, as opposed to Pml, Gadd45b is a direct target of En2 (does not require the translation of an

42 intervening protein, left panel), a result that can be placed in parallel with the decrease in Gadd45b expression in 8 week-old En1+/- mice (Figure 6D, right panel).

Interestingly, the number of genes annotated in the apoptosis pathway is much higer that in the En1+/- mouse (Figure 1) and this was not due to the En2 injection as illustrated per se (Figure S5), thus suggesting that the rapid (6 hours) and strong up-regulation of anti- apoptotic pathways by En2 takes place in the context of an acute oxidative stress. Among the anti-apoptotic pathways, the high induction of Gadd45b/g and NF-κB pathway mediators suggested pathways involving JNK (c-Jun N-terminal kinase) signaling already implied in the context of several neurodegenerative diseases, including PD (Coffey 2014). Figure 6E,F confirms that 6-OHDA injection rapidly increases the percentage of neurons stained for p- JNK and that En2 strongly and rapidly reduces this number (Figure 6F).

IV. Discussion

This study has demonstrated that an acute oxidative stress applied specifically to mDA neurons is more deleterious in an En1+/- context than in a wt one. Conversely, it is shown that wt mDA neurons exposed to acute stress can be rescued by an Engrailed gain of function. Given that, in the En1+/- mouse, mDA neurons show signs of oxidative insults and die progressively, it can be proposed that mDA neurons age more rapidly in the mutant mice. In other words, Engrailed, through its protective effect against oxidative stress, may act as an anti-ageing factor for mDA neurons. This hypothesis fits with the fact that even when of genetic etiology, PD develops late in life, an observation that applies to other neurodegenerative diseases, including Alzheimer and Huntington diseases.

In former studies (Alvarez-Fischer et al. 2011; Sonnier et al. 2007) we showed that Engrailed protects mDA neurons in the En1+/- mouse and in three PD mouse models. The hypothesis, explored here, that part of Engrailed action is at an chromatin level, takes its origin in an RNA-seq analysis comparing 6 week-old wt and En1+/- mice and placing in the limelight many genes associated with the DNA damage response and chromatin remodeling.

43 These pathways are interconnected since chromatin-associated changes can result from DNA damage and are necessary to give access to the DNA repair machinery (Madabhushi et al. 2014; Rulten & Caldecott 2013; Soria et al. 2012). Indeed the same pathways are activated during ageing (Orozco-Solis & Sassone-Corsi 2014; Vijg & Suh 2013)

A link between ageing, DNA damage repair and chromatin changes can be found in oxidative stress. Abnormal levels of reactive oxygen species (ROS) are toxic, in particular at the DNA level where proteins and DNA bases are subjected to oxidation (Vijg & Suh 2013). For example guanine oxidation in CpGs can block the access of MeCP2 and thus participate in chromatin modifications (Skene et al. 2010). ROS can induce DNA breaks and elicit DNA damage responses (Marteijn et al. 2014; O'Sullivan & Karlseder 2012). A striking example is given by the observation that neural activity - thus ATP and ROS synthesis - is physiologically accompanied by the formation of DSBs rapidly repaired in wt mice but not in a mouse model of Alzheimer disease (Suberbielle et al. 2013). More generally, neurons with high metabolic activity, because they consume more ATP and produce more ROS, are at a higher risk for degeneration.

This has led us to establish a model of acute oxidative stress selectively exerted on mDA neurons of the SNpc. This model consists in intranigral 6-OHDA injection, a drug only taken-up by mDA neurons expressing the DA transporter, thus primarily those of the SNpc(Di Salvio, Di Giovannantonio, Acampora, et al. 2010) The injection at the level of the SNpc, compared to the classical injection in the striatum, has the advantage of rapidity, not allowing secondary events to take place (e.g. changes in the striato-nigral physiology).

After 6-OHDA injection, several new γ-H2AX foci appear throughout the nucleus suggesting the formation of DSBs. This is a phenocopy of the En1+/- phenotype and the synergy between En1+/- and 6-OHDA for cell death suggests that Engrailed protects against oxidative stress and its deleterious consequences. This protection takes place in the experimental 6-OHDA paradigm as demonstrated by the latter "synergy experiment" but also in "real life" since many changes observed in the acute 6-OHDA stress also appear at specific times and with various intensities, in the En1+/- mutant. This includes DSB formation, H3K27me3 distribution, size of DAPI-dense regions, nucleolar phenotype, activated caspase-3 staining

44 and abnormal gene expression. The protective role of Engrailed is further confirmed by the experiment where an En2 injection, 30 minutes after that of 6-OHDA, partially protects against cell death evaluated 6h, 24h and 7 days later. It must be recalled that En1 and En2 are biochemically equivalent in the midbrain (Alvarez-Fischer et al. 2011; Hanks et al. 1995; Sonnier et al. 2007) justifying the use of En2 - much easier to produce than En1 - as a protective protein.

The low expression of the DA transporter explains the poor sensitivity of VTA mDA neurons to 6-OHDA but not why VTA neurons are relatively spared in the En1+/- mutant even though they express En1. A possibility is that Otx2, a homeoprotein expressed in the VTA and not in the SNpc, protects VTA neurons from death (Di Salvio, Di Giovannantonio, Acampora, et al. 2010; Di Salvio, Di Giovannantonio, Omodei, et al. 2010), thus dampening the loss of one En1 allele and resulting in the slower degeneration of mDA neurons in the VTA of En1+/- mice (Sonnier et al. 2007). Indeed Otx2 prevents the degeneration of Retinal Ganglion Cells (RGCs) in a mouse model of Glaucoma (Torero Ibad et al. 2011) and confers a protection similar to that of En2 in the SNpc (this study).

When testing a battery of nuclear/nucleolar antigens in the 6-OHDA-injected mice we were struck by the diffusion of nucleolin from the nucleolus to the entire nucleoplasm, a classical sign of nucleolar stress (Boulon et al. 2010). It is of note in this context that nucleolar stress was demonstrated in mDA neuron from PD patients (Rieker et al. 2011). We thus looked for morphological signs of nucleolar stress in En1+/- mice at 6 weeks. Although we did not find signs of nucleolar disruption, the transcription of the pre-45S rRNA that takes place in the nucleolus (Németh & Längst 2011) was dramatically up-regulated in the mutant mouse suggesting nucleolar stress (Cong et al. 2012; Larson et al. 2012; Németh & Längst 2011) not yet translating into nucleolus disruption. However, 40% of one year-old En1-/+ mDA neurons displayed nucleolar disruption. A percentage to compare to the 70% observed in 6-OHDA injected mice.

There exists a strong correlation between DSB formation and changes in the heterochromatin (Seeber et al. 2013). Using markers for H3K27me3, H3K9me3, MeCP2 and Lamin B2, we found major and similar reorganizations in En1+/- mice and after an acute

45 oxidative stress. Following 6-OHDA injection, H3K27me3 which decorates perinucleolar and perinuclear heterochromatin looses this specific localization, suggesting a change in heterochromatin structure. In parallel, a delocalization of Lamin B2 perinuclear staining was noted. Dramatic changes in H3K9me3 and MeCP2 staining patterns were also observed with a dilution throughout the nucleus as if the proteins had been released from the chromocenters decorated with DAPI. Interestingly, although the DAPI spots were still present after 6-OHDA injection, their size distribution demonstrated a loss of the largest ones. These traits suggest that 6-OHDA may lead to the expression of genes or repeats normally silent or expressed at low levels and possibly toxic to mDA cells when expressed at abnormally high levels (Gantz et al. 2011; Matsumoto et al. 2010). Among over-expressed repeats are those encoding LINE- 1 retrotransposons as shown by their enhanced expression following 6-OHDA injection.

Giving support to the idea that EN1/2 is in the PD pathway, possibly through the ability of Engrailed to antagonize ROS negative effects, a decrease in the number of mDA neurons with perinucleolar H3K27m3 staining and an increase in LINE-1 transcripts were also observed in 6-8 week-old En1+/- mice. All deregulated genes common to the 6-OHDA and En1+/- models were not searched for, but we speculate that the increased expression of several genes encoding cell-cycle regulators (e.g. Cyclin A, PCNA and pH3) and anti- oncogenes (e.g. p53, p21) in post-mitotic neurons may be explained by the DNA damage response and heterochromatin modifications. The link between their expression and cell death remains to be investigated, in particular that of LINEs, thus of ORF2 that encodes a reverse transcriptase and an endonuclease that can participate in DSB formation (Gasior et al. 2006).

A physiological protection against oxidative stress by Engrailed is supported by the enhanced toxicity of 6-OHDA in the En1+/- mice and the numerous parallels between the En1+/- phenotype and the changes provoked by an acute oxidative stress. Protection is confirmed by the "therapeutic" effects of En2 following 6-OHDA injection. Of note, En2 injection follows that of 6-OHDA by 30 min, indicating that Engrailed does not simply neutralize the 6- OHDA-induced oxidative stress. This En2-mediated protection was paralleled by a return to normal H3K27me3 and nucleolin localization. In addition, the percentage of mDA neurons with abnormal γ-H2AX spots dropped from 20 to less than 5% and the expression of Cyclin A and several cell cycle and pro-apoptotic genes was strongly repressed.

46

These results mirror the enhanced sensitivity of En1+/- mice to oxidative stress and confirm that Engrailed protection against oxidative stress includes an activity at the chromatin level, probably involving the regulation of gene expression by this transcription factor. The RNA-seq experiment comparing 6-OHDA/Sham and 6-OHDA/En2 mice, 6h after En2 injection confirms changes in the transcription of many genes in the chromatin remodeling, cell cycle and apoptosis pathways and less so in the DNA-damage pathway. Interestingly, the apoptosis pathway involving GADD45b/g, NF-kappa B and JNK seems to be predominant and confirms that Engrailed acts as a strong suppressor of apoptosis (Albéri et al. 2004; Beltran et al. 2014). The absence of a large induction of genes in the DNA damage pathway is likely due to the fact that the RNA-seq analysis was done at 6h in order to compare the same number of mDA neurons in both conditions. Indeed the decrease in the number of γ-H2AX foci is a late event in mDA neuron recovery. It is thus possible that DNA repair is posterior to the activation of anti-apoptotic pathways and chromatin remodeling. It remains to be understood how these pathways are interconnected and which genes are direct targets of the transcription factor. Indeed, Engrailed activity at the transcription level (also supported by the deleterious activity of EnHD-VP64 and the direct regulation Gadd45b transcription by En2) may associate with translation regulation (Alvarez-Fischer et al. 2011) and a chromatin scaffolding activity.

In conclusion, it is proposed that Engrailed protects against acute oxidative stress but also against oxidative lesions that accumulate with age. This is of importance as many neurodegenerative diseases are ageing diseases, including their genetic forms as demonstrated by the late onset of the familial forms of Alzheimer disease, PD and Huntington disease. The use of Engrailed as a therapeutic protein in 3 PD models (Alvarez-Fischer et al. 2011) and in the En1+/- mutant (Sonnier et al. 2007), together with the present data gives credit to the idea of using Engrailed as a therapeutic protein. Indeed it is tempting to speculate that the observed restructuration of the chromatin after a single injection of En2 may allow a long-lasting effect of a single treatment, possibly allowing one to avoid a gene therapy approach (not to be excluded anyhow). Because homeoproteins are expressed in the entire brain and have the peptidic signals allowing their intercellular transport (Joliot & Prochiantz 2004), it will be of interest to verify if what was shown here for Engrailed and SNpc mDA neurons can be generalized to other homeoproteins and neuronal populations as suggested by Otx2 rescuing

47 activity. On a more speculative mode, it is intriguing that this class of developmental transcription factors that intervene in early and late development through cell autonomous and non-cell autonomous pathways (Prochiantz & Di Nardo 2015) might antagonize the ageing process once the adult stage has been reached.

V. Experimental procedures

Animals

Mice were treated as per the guidelines for the care and use of laboratory animals (US National Institute of Health) and the European Directive 86/609 (EEC Council for Animal Protection in Experimental Research and Other Scientific Utilization). Swiss OF1 wt (Janvier) and En1+/- mice (Hanks et al., 1995) were maintained in conventional animal facility. Experimental groups consisted of 6 to 9 week-old mice.

RNA-seq analysis

mDA neurons in the SNpc of wt and En1+/- mice labeled using the quick TH-staining protocol (Chung et al., 2005) were isolated by Laser Capture Micro-dissection (LMD7000, Leica). Samples from 4 animals per group were pooled and total RNA was extracted using the AllPrep DNA/RNA Micro Kit (Qiagen) followed by DNase I using the RNeasy MinElute Cleanup protocol for on-column DNase I treatment. Construction of cDNA libraries (Ovation RNA-seq System V2) and Illumina RNA-seq were performed by the Ecole Normale Supérieure Genomic platform (Paris). p values of differentially expressed genes between wt and En1+/- samples were computed using DESeq package (Anders & Huber 2010). Normalized read counts were injected into the Gene Set Enrichment Analysis algorithm (Pathway Studio, Elsevier) to test for statistical enrichment. Pathway Studio Ontology and Cell Process Pathways collections were selected as Gene Set Categories for the analysis. RNA-seq analysis was also performed using RNA extracted from Laser Capture micro- dissected mDA neurons after Engrailed infusion (Alvarez-Fischer et al. 2011) in the SNpc of wt mice and from SNpc tissue punches collected 6h after 6-OHDA/sham and 6-OHDA/En2 injection.

48 In vivo treatments

For 6-OHDA injections, mice were placed in a stereotaxic instrument and a burr hole drilled into the skull 3.3 mm caudal and 1 mm lateral to the bregma. The needle was lowered 4 mm from the surface of the skull and 6-OHDA (2 µl; 0.5 µ g/µl Sigma) or sham (NaCl 0.9%) injections were performed over 4 min. For Engrailed rescue experiments, a solution (2 µl) of bacterial recombinant En2 (300 ng; 4 µM) and colominic acid (3 µg) (Sonnier et al., 2007) or vehicle (NaCl 0.9%) was injected 30 min after 6-OHDA injection using the same coordinates. When indicated, cycloheximide (0.1 µg/µl, sigma) was added. For Otx2 protein injection, a 2 µl solution containing 300 ng of the protein was used. Mice were killed at indicated times for analysis. SNpc tissues for qRT-PCR and Western blot analysis were obtained by performing 1 mm punches from 2 mm thick frozen coronal slices. For EnHD- VP64, mice were infused for 7 days with an osmotic mini pump (Alzet 1002, Charles River Laboratories) connected to a 4 mm-long cannula placed at the same stereotaxic coordinates as above. The pump was filled with 100 µl containing EnHD-VP64 (400 nM, 0.9% NaCl or the equivalent volume of an empty-plasmid-containing bacterial extract) and colominic acid (1.5 µg/µl).

qRT-PCR

Total RNA from SNpc tissues was extracted using the RNeasy Lipid Tissue kit (Qiagen) followed by DNase I (Thermo) digestion. RNA (200 ng) was transcribed using the QuantiTect Reverse Transcription kit (Qiagen). qRT-PCR was performed using SYBR-Green (Roche Applied Science) and values normalized to Gapdh and/or Hprt. Data were analyzed using the ddCt method. In some experiments, RNA was isolated specifically from SNpc nuclei (Subcellular Protein Fractionation Kit, Thermo). Primers are listed in Table S2.

Immunostaining

Mice were perfused with 4% paraformaldehyde in PBS; brains were postfixed for 1h and cryoprotected in 15% sucrose. Tissues embedded in Tissue-Tek O.C.T. (Sakura Finetek) were frozen isopentane prior to storage at −80°C. Brains were cut at the level of the SNpc into 18 µm-thick sections. For immunofluorescence, slides were permeabilized in 1% Triton X- 100 in PBS for 20 min and incubated at 100°C for 20 min in citrate buffer (10 mM citric acid, 0.05% Tween 20, pH 6.0). After blocking for 1h (10% normal goat serum, 0.05% % Triton X-

49 100 in PBS), tissues were incubated overnight at 4°C with primary antibodies diluted in the blocking solution (mouse anti-γ-H2AX, 1:200, Millipore; chicken anti-TH, 1:500, Abcam; rabbit anti-activated-caspase3, 1:200, Abcam; rabbit anti-nucleolin, 1:200, Sigma; rabbit anti- fibrillarin, 1:200, Cell Signalling; rabbit anti-H3k27me3, 1:200, Millipore; rabbit anti- H3k9me3, 1:200, Abcam; a generous gift from Edith Heard; rabbit anti-Lamin B2, 1:100, Santa Cruz; rabbit anti-Mecp2, 1:300, Abcam; rabbit anti-PCNA, 1:200, Cell Signalling; rabbit anti-Cyclin A, 1:200, Santa Cruz; mouse anti-pH3, 1:100, Cell Signalling; mouse anti- p-JNK, 1:200, Santa Cruz). Sections were incubated with appropriate secondary antibodies (488 anti-chicken, 647 anti-chicken, 488 anti-mouse, 546 anti-mouse and 546 anti-rabbit Alexa Fluor, Life Technologies) for 1h at room temperature. Labeled sections were imaged by confocal microscopy (SP5, Leica). For TH immunohistochemistry, sections were permeabilized in 1% Triton X-100 and incubated overnight at 4°C with 10% normal goat serum in PBS containing a rabbit polyclonal antibody against TH (1:1000; Pel-Freez Biologicals). Sections were treated with biotinylated secondary antibody and incubated with avidin-biotinylated horseradish peroxidase complex (ABC system, Vectastain). Peroxidase was revealed using DAB peroxidase (HRP) substrate kit (Vectastain) and imaged with a Nikon Eclipse i90 microscope.

Image quantification

Images were analyzed with ImageJ. For immunofluorescence, all quantifications were performed using 60X magnification and 0.7 µm-thick successive focal planes. For H3K27me3 nucleolar pattern analysis, a graph of the intensities of pixels along a line positioned through the nucleolus was created. Perinuclear/nuclear ratio of H3K27me3 fluorescence intensity was determined by measuring pixel density at the periphery of the nucleus and in the nucleoplasm. DAPI-dense regions in sham and 6-OHDA injected mice were quantified by measuring individual DAPI-surface areas in each TH+ cell and plotting them as relative frequency distribution histograms.

50 VI. Author contributions

HR and FXBT contributed with OBM to the conception and realization of most experiments. JF and RJ were involved in the experimental work, the day to day supervision of the younger investigators and the conception of the study. AP coordinated the work, participated in the conception of the experiments and wrote the manuscript with the help of RJ, primarily.

VII. Acknowledgments

We thank Pr. Edith Heard and Dr. Ariel Di Nardo for their reading of the manuscript and helpful discussions. We also acknowledge Dr. Alain Joliot for his generous gift of EnHD- VP64 and the CIRB imaging facilities for their help. The study was supported by Région Ile de France, Fondation Bettencourt Schueller, GRL program N°2009-00424 and ERC Advanced Grant HOMEOSIGN n°339379.

51 Figures for Article 1

Figure 1. Altered expression of genes related to DNA damage, chromatin remodeling and apoptosis in En1+/- mice (A) SNpc from wt and En1+/- mice was laser microdissected. The total number of reads and differentially expressed genes (p < 0.05) are indicated. Decreased expression of En1 in RNA samples from En1+/- SNpc was confirmed by RNA-seq.

(B) RNA-seq data analysis (Pathway Studio) reveals a highly significant enrichment (p < 0.05) in Ontology gene subgroups involved in DNA damage and chromatin remodeling between wt and En1+/- littermates. Analysis using Cell Process Pathways shows differential expression of apoptosis-related genes.

(C) The differentially expressed genes in En1+/- SNpc TH+ neurons belonging to DNA damage, chromatin remodeling and apoptosis groups and pathways are ranked by p values. Also see Figure S1 and Table S1.

52

53 Figure 2. Increased DNA damage and chromatin alteration in mDA neurons in En1+/- SNpc

(A) TH+ neurons in the SNpc of wt mice display a single ring of γ-H2AX staining whereas En1+/- TH+ neurons show additional γ-H2AX foci scattered in the nucleus. Scale bar, 10 µm.

(B) The percentage of TH+ neurons with >2 γ-H2AX foci in the SNpc increases from 2% in wt to 16% in 8 week-old En1+/- mice (n = 3; **p < 0.01; Student’s t test). Error bars represent SEM. γ-H2AX staining in TH+ neurons in the VTA of En1+/- and wt mice is similar. Between 110 and 162 neurons were counted in each condition. Also see Figure S1D.

(C) Perinucleolar and perinuclear H3K27me3 staining in TH+ neurons is decreased in the SNpc of En1+/- mice. Scale bar, 10 µm.

(D) H3K27me3 perinucleolar staining was quantified by measuring fluorescence intensity (left) along the dotted lines (Figure 2C). The percentage of cells with dense staining drops (right) in En1+/- mice (n = 3, *p < 0.05; Student’s t test; 129 and 159 neurons counted in wt and En1+/- mice, respectively). Error bars represent SEM.

(E) Perinuclear H3K27me3 staining in En1+/- TH+ neurons is reduced as shown by the decreased ratio of nuclear lamina/nucleoplasm fluorescence intensities (n = 3, **p < 0.01; Student’s t test; 30 and 30 neurons counted in sham and 6-OHDA conditions, respectively). Error bars represent SEM.

(F) Surface quantification of DAPI-dense regions. The distribution of the relative frequencies indicates a shift towards smaller DAPI-dense areas in En1+/- mice (n = 162-211; **p < 0.01; Kolmogorov-Smirnov test; 3 wt and En1+/- mice analyzed).

(G) LINE-1 transcripts in the SNpc of En1+/- mice analyzed by qRT-PCR increase, compared to wt (n = 5, *p < 0.05; Student’s t test). Error bars represent SEM.

(H) Activated caspase-3 positive TH+ neurons are detected in the SNpc of En1+/- mice (8 week-old). Scale bar, 50 µm.

54

55

Figure 3. Enhanced sensitivity of En1+/- mice SNpc TH+ neurons to 6-OHDA (A) En1+/- mice are more sensitive to DNA damage induced by 6-OHDA

(B) Quantification of (A). 6-OHDA injected in the SNpc of wt mice leads 6h later to the appearance of γ-H2AX foci in about 25% of TH+ neurons (n = 3-6, **p < 0.01; Student’s t test). En1+/- mice are more sensitive to 6-OHDA with 50% of neurons showing multiple γ-H2AX foci (n = 3-6, **p < 0.01; Student’s t test; 130, 210 and 146 neurons counted for each condition, respectively). Error bars represent SEM. Scale bar, 10 µm.

(C) 6-OHDA injection provokes (6h later) the loss of about 30% and 60% of TH+ neurons in the ipsilateral SNpc of wt and En1+/- mice, respectively (n = 3, **p < 0.01; Student’s t test). The contralateral non-injected side is taken as a reference. In each condition between 1534 and 2034 neurons were counted. Error bars represent SEM. Also see Figure S2A-C and Figure S3A-C.

56

57 Figure 4. Modification of heterochromatin marks upon 6-OHDA injection

(A) Midbrain sections stained for γ-H2AX, H3k27me3 and TH and analyzed by confocal microscopy. Perinucleolar and perinuclear H3K27me3 staining is decreased upon 6- OHDA injection. Mice were analyzed 6h post injection. Scale bar, 10 µm.

(B) The percentage of TH+ neurons displaying dense H3K27me3 perinucleolar and perinuclear staining quantified as in Figures 2D,E drops dramatically (left) in 6- OHDA-injected mice (n = 3, ***p < 0.001; Student’s t test; 148 and 91 neurons counted in sham and 6-OHDA conditions, respectively). Error bars represent SEM. Perinuclear H3K27me3 staining in TH+ neurons also decreases upon 6-OHDA injection (right) (n = 3, *p < 0.05; Student’s t test; 40 and 47 neurons counted in sham and 6-OHDA conditions, respectively). Error bars represent SEM. Also see Figure S2D-G.

(C) Midbrain sections stained for γ-H2AX, nucleolin and TH and analyzed by confocal microscopy. Nucleolin nucleolar localization in sham-injected mice is lost upon 6- OHDA-injection (analyzed 6h post injection). NCL, nucleolin. Scale bar, 10 µm.

(D) The percentage of TH+ neurons with nucleolar nucleolin is significantly decreased upon 6-OHDA injection (n = 3, *p < 0.05; Student’s t test; 161 and 97 neurons were counted in sham and 6-OHDA conditions, respectively). Error bars represent SEM. NCL, nucleolin.

(E) The level of pre-45S rRNA analyzed by qRT-PCR is up-regulated following 6- OHDA injection in SNpc purified nuclei (n = 3, *p < 0.05 ; **p < 0.01; Student’s t test). Error bars represent SEM.

(F) Midbrain sections from (1 year-old animals) stained for nucleolin and TH were analyzed by confocal microscopy. Nucleolin presents a nucleolar localization in wt TH+ neurons whereas 40% of TH+ neurons in En1+/- mice present a diffuse staining pattern (arrows). Scale bar, 50 µm. Higher magnification images of dotted square areas are shown in the right most panels. Scale bar, 10 µm. The percentage of TH+ neurons with nucleolar nucleolin is significantly decreased in En1+/- mice (n = 3, **p < 0.01; Student’s t test; 99 and 87 neurons counted in wt and En1+/- mice, respectively). Error bars represent SEM. NCL, nucleolin. Also see Figure S3D,E.

(G) p53 and p21 mRNAs were quantified by qRT-PCR using total RNA from the SNpc of 6-OHDA or sham injected mice (n = 6, *p < 0.05; Student’s t test). Error bars represent SEM.

(H) Higher expression of repeat elements in 6-OHDA injected SNpc. The levels of LINE- 1 transcripts analyzed by qRT-PCR using total SNpc RNA increase in 6-OHDA- injected mice. The level of LINE-1 RNA was also analyzed using RNAs from SNpc purified nuclei (n = 3-5, *p < 0.05 ; **p < 0.01; Student’s t test). Error bars represent SEM.

58

59 Figure 5. Engrailed rescues TH+ cells from 6-OHDA-induced cell death (A) Mice injected with 6-OHDA in the SNpc were re-injected 30 min later with either sham or Engrailed and analyzed at 24h. Compared to sham, Engrailed injection prevents the 6-OHDA-induced TH cell loss in the ipsilateral SNpc. Scale bar, 500 µm.

(B) Protective effect of Engrailed assessed by comparing the ratio of TH+ cell counts in ipsilateral versus contralateral SNpc 6h, 24h and 7 days post injections (n = 3-5, *p < 0.05; **p < 0.01; Student’s t test). The number of neurons counted ranged between 1670 and 2275 per condition. Also see Figure S3F.

(C) The rescue of TH+ neurons by Engrailed in 6-OHDA-injected mice is paralleled by a disappearance of cyclin A staining at 24h. Scale bar, 100 µm. Also see Figure S4A-D.

(D) Immunostaining for H3K27me3 indicates a progressive recovery following En2 injection. Recovery of nucleolar nucleolin is almost complete at 24h. The percentage of TH+ neurons with γ-H2AX foci returns to normal between 24h and 7 days after Engrailed injection. For each analysis, n = 3. **p < 0.01; ***p < 0.001; ****p < 0.0001; one way ANOVA followed by Dunett’s test (vs sham). The number of neurons counted ranged between 102 and 150 for each condition. Error bars represent SEM.

(E) The expression of selected genes related to apoptosis and cell cycle in the SNpc of 6- OHDA/sham and 6-OHDA/Engrailed mice analyzed 6h post injections (n = 5, *p < 0.05, **p < 0.01; Student’s t test). Error bars represent SEM.

(F) Otx2 protects TH+ neurons against 6-OHDA-induced cell death. Otx2 was injected 30 min after 6-OHDA injection and mice were analyzed 24h later. Protective effect was assessed as in the case of En2 injections (n = 3; *p < 0.05; Student’s t test). Number of neurons counted ranged between 1268 and 1495 for each condition. Error bars represent SEM. Also see Figure S4E,F.

60

61 Figure 6. RNA-seq analysis reveals Engrailed anti-apoptotic activity (A) Differentially expressed genes in the SNpc of 6-OHDA versus 6-OHDA/En related to DNA damage, chromatin remodeling and apoptosis and cell cycle are ranked by p values.

(B) Differentially expressed genes in the SNpc of 6-OHDA versus 6-OHDA/En related to cell cycle are ranked by p values.

(C) The expression of selected genes in the SNpc of 6-OHDA and 6-OHDA/En is confirmed by qRT-PCR (n = 5, *p < 0.05, **p < 0.01; Students t test). Error bars represent SEM.

(D) Left panel: Gadd45b and Pml transcripts were measured by qRT-PCR in the SNpc of 6-OHDA and 6-OHDA/En at 6h (injection with cycloheximide, CHX; n = 5, **p < 0.01; Student’s t test). Right panel: Gadd45b transcripts in the SNpc of 8 week-old wt and En1+/- mice (n = 4, ***p < 0.001; Student’s t test). Error bars represent SEM.

(E) Midbrain sections from sham, 6-OHDA/sham and 6-OHDA/En stained for p-JNK and TH and analyzed by confocal microscopy. Scale bar, 50 µm. Higher magnification images of dotted squares are shown in the right most panels. Scale bar, 10 µm.

(F) Engrailed significantly decreased the percentage of SNpc TH+ neurons with p-JNK staining (n = 3, *p < 0.05, ***p < 0.001; one way ANOVA followed by Tukey’s multiple comparisons test). Number of neurons analyzed ranged from 227 and 351 for each condition). Error bars represent SEM.

62

63 Figure S1. Analysis and validation of RNA-seq data (A) Differentially expressed transcription factors in RNA-seq analysis by Pathway Studio of wt and En1+/- SNpc are ranked by p value.

(B) Analysis of RNA-seq data from Engrailed gain of function experiments by Pathway Studio. The gene subgroups enriched (p < 0.01) are indicated (n = 5).

(C) Altered expression of a selection of genes related to DNA damage, chromatin remodeling and apoptosis is confirmed by qRT-PCR using SNpc RNAs from 6 week- old wt and En1+/- mice (n = 3-6; *p < 0.05; **p < 0.01; Student’s t test). Error bars represent SEM.

(D) The percentage of TH+ neurons with >2 γ-H2AX foci in the SNpc increases from 5% in wt to 17% in 24 week-old En1+/- mice (n = 3; *p < 0.05; Student’s t test; 107 and 94 neurons were counted in wt and En1+/- conditions, respectively). Error bars represent SEM.

64

65 Figure S2. TH cell death and chromatin changes in 6-OHDA injected mice (A) Western blot analysis shows a 50% decrease in TH protein levels at 6h in the 6- OHDA injected ipsilateral SNpc (n = 3, *** p<0.001).

(B) The levels of mRNAs for TH and En1 are decreased (at 6h) by 70 and 50%, respectively in 6-OHDA injected SNpc (n = 6, *p < 0.001; Student’s t test). Error bars represent SEM.

(C) Activated caspase-3 is detected at 6h in TH+ neurons in the 6-OHDA-injected side but not in the contralateral side. Scale bar, 100 µm.

(D) Triple immunostaining of midbrain sections for γ-H2AX, H3K9me3 and TH shows that H3K9me3 colocalization with DAPI in sham condition is lost in 6-OHDA- injected mice. Similarly, MeCP2 colocalization with DAPI in TH+ neurons of sham- injected mice is lost in 6-OHDA-injected mice. Scale bar, 10 µm.

(E) The percentage of TH+ neurons with dense H3K9me3 or MeCP2 marks is significantly decreased in 6-OHDA-injected mice (n = 3, ***p < 0.001; Student’s t test). The numbers of neurons counted were 161 (sham) and 97 (6-OHDA) for H3K9me3 and 150 (sham) and 128 (6-OHDA) for MecP2. Error bars represent SEM.

(F) Loss of perinuclear lamin B2 staining in TH+ neurons in the SNpc of 6-OHDA- injected mice. Scale bar, 20 µm.

(G) Surface quantification of DAPI-dense regions in TH+ neurons in sham or 6-OHDA- injected mice. Compared to sham-injected controls, the distribution of the relative frequencies indicates a shift towards smaller DAPI-dense areas in TH+ neurons of 6- OHDA-injected mice (n = 137-177; ****p < 0.0001; Kolmogorov-Smirnov test; 3 mice per condition).

66

67 Figure S3. Engrailed transcription activity involved in TH neuron protection (A) Mice were unilaterally infused with sham or EnHD-VP64 in the SNpc and analyzed at 7 days. Immunostaining of midbrain sections for TH shows that EnHD-VP64 infusion (but not sham) provokes TH cell loss in the ipsilateral SNpc. Scale bar, 500 µm.

(B) EnHD-VP64 induced TH cell loss was assessed by comparing the ratio of TH+ cell counts in ipsilateral versus contralateral SNpc at 7 days post infusion (n = 6, **p < 0.01; Student’s t test). The number of neurons counted ranged between 1902 and 2118 per condition

(C) Quantification of γ-H2AX staining shows that the percentage of TH+ neurons with >2 γ-H2AX foci in the SNpc increases upon EnHD-VP64 infusion (n = 3; *p < 0.05; Student’s t test; 138 and 129 neurons were counted in sham and EnHD-VP64 conditions, respectively). Error bars represent SEM.

(D) Analysis of RNA-seq data also reveals differentially expressed genes in Pathway Studio Ontology subgroups related to nucleolar organization and biogenesis in the SNpc of En1+/- mice (p < 0.1).

(E) The level of pre-45S rRNA in the SNpc of wt and En1+/- mice (8 week-old animals) was analyzed by qRT-PCR (n = 4, **p < 0.01; Student’s t test). Error bars represent SEM.

(F) Engrailed promotes TH+ neurons survival in the 6-OHDA model. The survival effect of Engrailed was not observed if Engrailed was injected 24h after 6-OHDA injection (when the majority of TH+ neurons are lost) and recovery assessed 6h later (n = 3).

68

69 Figure S4. Detection of cell cycle markers in TH+ neurons upon acute oxidative stress in vivo and Engrailed-mediated protection against H2O2-induced DSBs in vitro (A) Detection of PCNA and pH3 cell cycle markers in TH+ neurons in the 6-OHDA- injected side. Scale bar, 20 µm.

(B) Immunostaining of midbrain sections reveals TH+ neurons expressing cyclin A only in the ipsilateral side of 6-OHDA-injected mice. Mice were analyzed 6h post injection. Scale bar, 100 µm.

(C) Detection of activated caspase-3 in TH+ neurons staining positive for pH3. Scale bar, 50 µm.

(D) The percentage of TH+ neurons expressing cyclin A upon 6-OHDA injection was quantified (left, n = 3, ***p < 0.001; Student’s t test) and confirmed by Western blot analysis (right, n = 4, *p < 0.05; Student’s t test). Error bars represent SEM.

(E) Engrailed treatment protects midbrain neurons against H2O2-induced DSBs. Primary mesencephalic neuron cultures of E14.5 mouse embryos were treated with H2O2 at 5 µM after a 24h incubation with NaCl or En2 (15 nM). Cells were fixed 1h after H2O2 treatment, stained for γ-H2AX and the number of foci was quantified (n = 68-137, ***p < 0.001, ****p < 0.0001; one way ANOVA followed by Tukey’s multiple comparisons test). Error bars represent SEM, scale bar, 10 µm.

(F) Engrailed-mediated protection against H2O2-induced DSBs using midbrain neuron cultures was assessed by the comet assay performed according to Trevigen CometAssay kit instructions and analyzed with OpenComet plugin, scale bar, 100 µm. The number of neurons with a comet tail was quantified (n = 39-46, ***p < 0.001; one way ANOVA followed by Tukey’s multiple comparisons test). Error bars represent SEM.

70 ! !

Figure! S5! (related! to! Figure! 6).! Differentially! expressed! genes! in! the! SNpc! of! En2! Figure S5. Differentially expressed genes in the SNpc of En2 vs sham injected mice versus!sham!injected!mice! GenesGenes! relatedrelated to!to!DNA!damage,!chromatin!remodeling,!apoptosis!are!ranked!by!p DNA damage, chromatin remodeling, apoptosis are ranked by p0values. value !

! ! ! !

71

PART 2 - Full-length LINE-1 expression in adult dopaminergic neurons of the Substantia Nigra is enhanced by oxidative stress and repressed by Engrailed homeoprotein

72

73

I. Introduction

In the first article we have demonstrated that Engrailed protects mDA neurons from oxidative stress induced by 6-OHDA. In the same study we showed that Engrailed not only prevents neuronal death, but also restores DNA integrity and a healthy chromatin state. We thus needed to provide a better insight into the link between protection and heterochromatin marks restoration. This led us to explore the “LINE-1 hypothesis”. Transposons were until recently thought to be specifically expressed in germ cells, in conditions when the repressive activities of the Piwi/piRNA pathway are alleviated. These conditions correspond to an endangering stress and the resulting LINE-1-induced mutations in germ cells have been described as the last line of defense of organisms in highly unfavorable environmental conditions (Siomi et al. 2011; Piacentini et al. 2014). However, this view has changed recently with the discovery that LINEs are also active in somatic tissues. Most of the latter observations demonstrate transposition in dividing cells, including neural stem cells (Muotri et al. 2005), but there also exists a few reports supporting LINE-1 expression in post-mitotic cells (Kubo et al. 2006), including neurons, primarily upon stress and illness conditions or in ageing individuals (W. Li et al. 2013; Erwin et al. 2014).

As stated previously, LINEs can induce DNA strand breaks, via the endonuclease domain of ORF2p; moreover they can be activated in case of heterochromatin opening or oxidative stress. Thus, they appear as interesting candidates in the search for the missing cog between DNA strand breaks, heterochromatin relaxation and oxidative stress. Given the number of new systems in which LINEs are being implicated, we undertook to explore a possible relationship between LINE-1 expression, oxidative stress and Engrailed protective activity.

74 II. Results

A. LINE-1 families are expressed in adult mesencephalic dopaminergic neurons

In a next-generation RNA sequencing (RNA-seq) experiment on RNA extracted from laser captured SNpc, we found that the three main active LINE-1 families (A, Tf and Gf) are expressed in this structure (Figure 1A). The Tf and A show a similar high number of reads (1864 - 1049 respectively) to be compared to the hypoxanthine-guanine phosphoribosyltransferase (Hprt) mRNA (343 reads), a well-transcribed gene, usually used for normalization (Figure 1A). This information was confirmed by qRT-PCR with the same LINE-1/Hprt RNA ratio on SNpc punches-extracted mRNA (Figure 1A). Because LINE-1 RNAs are not necessarily full-length, mRNA was purified from mouse ventral midbrain tissue on oligo-dT columns to ensure the presence of the 3’UTR, digested with DNase I and reverse transcribed with oligo-dT primers. PCR was achieved with the forward primer in the 5’UTR and the reverse in the 3’ region of ORF2 (Figure 1B – top panel). All three LINE-1 families are detectable at the expected sizes and no amplicon was visible in the absence of the reverse transcriptase in the RT reaction (-RT) establishing that the bands (+RT) are not due to a contamination with genomic DNA (Figure 1B – bottom panel). We verified the amplicons by enzymatic digestion (BamHI, NcoI, PstI) and Sanger-sequencing, the amplicons yielding the corresponding LINE-1 sequences (data not shown). The full–length mRNAs are translated into protein, as shown by the Western Blot of Figure 1C, where proteins from ventral midbrain were extracted, separated by gel electrophoresis and identified with a specific anti- ORF1p antibody. Finally ORF1p expression in post-mitotic neurons of the ventral midbrain, including mDA neurons, was verified. Figure 1D illustrates by in situ hybridization and immunohistochemistry the co-localization of TH and ORF1p and Figure 1E shows by double immunohistochemistry that ORF1p is present in less than 50% of all post-mitotic neurons, but in all TH-positive mDA neurons in the SNpc.

All in all, this series of experiments demonstrate that LINEs from at least 2 families are strongly expressed in post-mitotic neurons from the SNpc, including all mDA neurons. We then wondered whether, as we hypothesized, LINE-1 expression was modified after oxidative stress.

75 B. LINE-1 expression and nuclear localization are increased upon oxidative stress

Midbrain dopaminergic neurons are particularly sensitive to oxidative stress due to intrinsic sustained activity and dopaminergic metabolism, itself a generator of oxidative stress and a risk factor for neurodegeneration (C.-M. Chen et al. 2009). Several reports have highlighted an induction of LINE-1 elements upon oxidative stress in different systems (Giorgi et al. 2011; Rockwood et al. 2004). We thus tested whether oxidative stress regulates LINE-1 expression in mDA neurons. To that end, NaCl (sham) or 6-OHDA, were injected in the SNpc of wt mice and RNA from SNpc punches was quantified by qRT-PCR. An increase in LINE-1 Tf/Gf transcription was measured (Figure 2A); going along with the idea that oxidative stress can activate retrotransposition.

This 6-OHDA-mediated increase was verified in another experiment where LINE-1 transcripts were quantified by qRT-PCR separately in the nucleus and in the cytoplasm (Figure 2B). Clearly, the global increase was reproduced, but there was a higher increase of LINE-1 RNA in the nuclear fraction. In the course of LINE-1 replication, LINE-1 poly- cistronic mRNA associates with ORF1p and ORF2p and is translocated into the nucleus (Kubo et al. 2006). We thus verified if the abundance of LINE-1 transcripts in the nucleus could be in part explained by an oxidative stress-induced transport of the ribonucleoprotein particle into the nucleus. As illustrated and quantified in Figure 2C, 6-OHDA-induced stress results in a dramatic import of ORF1p into the nucleus of TH positive cells, a feature which was not observed for non TH positive cells. It is unlikely that this import reflects a damage of the nuclear membrane, since the protein tyrosine hydroxylase (TH) remains in the cytoplasm. Another explanation for this RNA nuclear increase could be that LINE-1 stress-induced transcription is a dynamic process. As we have seen, LINEs are massively present already in the steady-state level, and therefore the increase in LINE-1 RNA is diluted by the amount of mRNA already present in the entire cell. On the other hand, the nuclear fraction is probably very enriched in newly synthesized LINE-1 RNA, and therefore the stress-related LINE-1 activation is more visible in this compartment. We do not have enough data to support one hypothesis against the other, and the reality is probably a combination of both. However these results strongly support the hypothesis that 6-OHDA nigral injection is accompanied by LINE-1 activation.

76 C. Oxidative stress-induced and direct LINE-1 expression initiate DNA damage

Upon induction, full-length LINE-1 elements enter the nucleus in form of a ribonucleoprotein complex encompassing the LINE-1 RNA, the ORF1 protein (the RNA binding unit) and the ORF2 protein (encoding endonuclease and reverse transcriptase activities). The endonuclease activity generates a nick in the DNA and LINE-1 RNA retrotranscription can be initiated using the exposed 3'OH as a primer (Feng et al. 1996). ORF1p translocation into the nucleus after 6-OHDA injection coupled with increased LINE-1 RNA was already a strong clue that retrotransposition might be induced by 6-OHDA- mediated oxidative stress; however, we wanted to verify if LINE-1 activation could create DNA damage. To do so, we used an in vitro model of oxidative stress. E13.5 midbrain primary neurons were lipofected with a plasmid encoding Piwil1 or with an empty vector and treated 48 hours later with 100µM H202 for 1h. The void condition validated that oxidative stress induced DNA strand breaks (γ-H2AX immuno-cytochemistry), whose number were strongly reduced in the Piwil1 transfected neurons (Figure 3A - left panel). To confirm that it was Piwil1 which protected these neurons, DNA strand breaks were quantified in non- transfected neurons of the same culture and in those neurons, there was no difference between the two conditions (Figure 3A – right panel). All neurons were co-transfected with GFP encoding plasmids, which allowed us to discriminate the transfected neurons (the co- transfection rate is usually around 95% when two plasmids are used).

To validate these results, similar in vitro assays were made using adeno-associated virus 2 (AAV2) coding for Piwil1 or for GFP. Here again, overexpression of Piwil1, 7 days before an acute oxidative stress (100µM H202 for 1h), decreased the number of γ-H2AX foci in the neurons compared to the GFP controls (Figure 3B – left panel). Using the same assay, we measured LINE-1 transcription by fluorescent in situ hybridization (FISH), using probes spanning both ORF1 and ORF2 mRNA and counting the number of transcription spots in the infected neuronal nuclei (Figure 3B – right panel). Similarly, Piwil1 overexpression repressed LINE-1 transcriptional hotspots, which were increased by oxidative stress in the control condition. As a negative control, scrambled probes were used and in the most unfavorable condition (AAV2 GFP/H2O2), negligible numbers of foci were counted in the nuclei (20 times less than with the LINE-1 probes, data not shown).

77 Since Piwil1 is part of a major pathway for transposon silencing in germ cells, it is very tempting to speculate that oxidative-stress induced DNA damage was dependent on LINE-1 activation. This was shown by lipofection of either wild type LINE-1 or an ORF2 double mutant, with no endonuclease and reverse transcriptase activity. As shown in Figure 3C, in contrast with its inactive mutant version, wt LINE-1 induced a large number of DNA strand breaks. Moreover, the fact that DNA damage almost goes back to its initial value in presence of Piwil1 (Figure 3A-B) shows that LINE-1-induced DNA strand breaks are not a secondary mechanism but an important pathway for oxidative stress-induced DNA damage. The mechanism of LINE-1 transcriptional repression mediated by Piwil1 remains to be elucidated; indeed, Piwi mostly represses LINEs by RNA interference – the Ping-Pong cycle – which should not reduce a priori LINE-1 transcription. But Piwi proteins have also been shown to be epigenetic regulators (Huang et al. 2013). Thus, a possible explanation is that addition of Piwil1 induces the formation of heterochromatin around LINE-1 sequences, preventing their activation by oxidative stress.

D. LINE-1 expression is increased in En1-het mice

Midbrain DA neurons from En1+/- mice die progressively, starting from 6 weeks of age. Importantly, at the time when neuronal death is initiated, all neurons are still present but several nuclear phenotypes become visible, including DNA damage and the loss of heterochromatin, coinciding with an aberrant induction of LINE-1 expression in En1-het mice (Article 1). Furthermore, Engrailed addition protects against DSBs in vitro and in vivo (Article 1). These observations led us to ask whether Engrailed could protect against DNA damage via a repression of LINE-1 expression and we thus followed LINE-1 expression in En1+/- mice. RNA from laser captured SNpc (Article 1) was sequenced and mapped onto a consensus LINE-1 Tf sequence (AF016099.1). Figure 4A demonstrates an increase in the number of reads in 6 week-old En1+/- mice compared to wild type siblings. This increase at 6 weeks was confirmed by qRT-PCR for LINE-1 Tf while LINE-1 A RNA was not increased (Figure 4B). ORF1 protein was also increased as verified by immunohistochemistry (Figure 4C).

Because the progressive death of mDA neurons is initially accelerated in En1+/- mice (Sonnier et al. 2007) and in view of a possible link between LINE-1 expression and ageing

78 (W. Li et al. 2013), we compared LINE-1 expression in wt and En1+/- mice at one year. As opposed to 6 weeks, qRT-PCR at one year failed to show a difference between heterozygous and wild type mice (Figure 4B). To verify that this was not due to the 40% mDA neuronal loss at this age (Sonnier et al. 2007), LINE-1 Tf/Gf RNA level was normalized to TH (specifically expressed in mDA neurons), which also failed to show a significant difference (data not shown). In fact, expression levels of LINE-1 Tf/Gf in En1+/- mice at 6 weeks are not very different from expression levels of LINE-1 Tf/Gf in wild type littermates at 1 year. A hypothesis is that, as much as LINE-1 Tf/Gf is concerned, 6 week-old mutant animals have the same “age” as 1 year old wild-type mice, leading to the concept of premature ageing in the En1+/- mouse. This fits with the observation that little difference in cell death takes place between wt and En1+/- animals after 6 months (Sonnier et al. 2007).

E. Engrailed protects against LINE-1- induced DNA damage and is a direct repressor of LINE-1 expression

From the above-described results it can be suggested that Engrailed represses LINE-1 expression and that part of the DNA damage observed in En1+/- mice is directly due to LINE- 1 up-regulation and ORF2p endonuclease activity. As a first step into testing this hypothesis, we used a retrotransposition reporter assay (RT plasmid) using HEK cells bearing a tetracyclin-inducible En2 coding sequence (graciously given to us by the lab of Alain Joliot at the Collège de France). Engrailed2 was induced with doxycycline and one day later the RT plasmids (graciously given to us by Wenfeng An – South Dakota State University) were transfected. The plasmid (WA125), used previously to overexpress LINE-1 (Figure 3C), contains a codon-optimized mouse LINE-1 cassette with its endogenous 5’UTR region and in its 3’UTR a retrotransposition reporter cassette consisting of a CMV promoter followed by the EGFP gene interrupted by a sense orientation γ-globin intron. Only if transcription, splicing and retrotransposition occur, the intact GFP protein can be expressed allowing one to follow and quantify retrotransposition. The control plasmid has a double-mutation in ORF2p leading to a loss of endonuclease and reverse transcriptase activity (WA126, D212G; D709Y). After selection of transfected cells with puromycine, cells were sorted by fluorescence activated cell sorting (FACS). This experiment confirmed that Engrailed is, similarly to Piwil1, a repressor of retrotransposition (Figure 5A).

79 This, and the fact that the LINE-1 5‘UTR contains a canonical in silico Engrailed binding site, made us wonder whether Engrailed might not directly repress LINEs. To evaluate this hypothesis, we performed chromatin immunoprecipitation with an anti Engrailed antibody. To that end, nuclei purified from midbrain primary neuron were incubated with En2, without addition or with that of polydI:dC (to decrease unspecific binding) or of NP6 (a competing multimerized Engrailed binding site (Mainguy et al. 2000)). After fixation, the nuclei were sonicated and the obtained chromatin was immunoprecipitated with an anti- Engrailed antibody (Sonnier et al. 2007). Figure 5B shows that following En2 incubation, the anti-En1/2 antibody specifically pulls down DNA fragments of the 5'UTR of both LINE-1 Tf/Gf and A families around the putative Engrailed binding site. Because Engrailed is primarily a repressor, we constructed an activator Engrailed (thus an anti-Engrailed) by fusing the HD, which confers binding specificity to the genome, to a tetramerized Herpes virus activator domain (VP16x4= VP64; EnHD-VP64) (Figure 5A - bottom panel). To ensure that this construct would activate only direct targets of Engrailed, we simultaneously added the translation inhibitor cycloheximide (CHX). This treatment significantly increased LINE-1 Tf/Gf and A expression after 10 min in vitro or 1 hour in vivo, validating LINEs as a direct Engrailed target (Figure 5C – top panel). In these experiments, treatment with the flow- through of a void plasmid transformed into bacteria and purified in the same conditions as EnHD-VP64 (Control) proved that LINE-1 activation is specifically due to EnHD-VP64 activity and not to a contaminant in the recombinant protein production procedure.

Engrailed infusion into the midbrain is followed by its internalization into mDA cells. This internalization saves mDA neurons in several models of PD (Alvarez-Fischer et al. 2011), or following a direct oxidative stress induced by 6-OHDA injection in the SNpc (Article 1). EnHD-VP64 infused for one week in the SNpc of wt mice activates the formation of DSBs and induces mDA neuron death (Article 1). Given that, as shown above, EnHD- VP64 directly activates LINE-1 transcription, we hypothesized that EnHD-VP64-mediated cell death and DNA damage are induced - at least partially - by LINE-1 activation.

Conversely, we showed in Figure 5D that LINE-1 expression, which is increased by 6- OHDA injection in the SNpc (Figure 2A), is downregulated upon En2 injection. The fact that CHX was co-injected with Engrailed shows that, here again, this regulation is direct. The timing used in this experiment – injection of 6-OHDA followed by En2 + CHX 30 min later

80 and analysis 6 hours post-injection – is the same as in the Engrailed protection protocol described in the first article. This demonstrates that Engrailed is a direct repressor of LINEs and strongly suggests that its protective activity against oxidative stress-induced DSBs involves direct LINE-1 repression.

These experiments show that oxidative stress and neurodegeneration can be accompanied by LINE-1 activation, which can induce DNA damage and potentially cell death. Furthermore one can infer that Engrailed-mediated protection against DNA damage in the H2O2 in vitro and 6-OHDA in vivo oxidative stress models is, at least partially, due to its ability to directly repress LINE-1 transcription.

81 Figures for Part 2

Figure 1. Full-length Long Interspersed Nuclear Elements (LINEs) are expressed in the adult mouse ventral midbrain. (A) LINE-1 RNA from the three main families (A, Tf and Gf) is expressed in the Substantia Nigra pars compacta (SNpc) (RNA-seq) and in the ventral midbrain (qRT- PCR). RPM: reads per million, Ct: qPCR cycle threshold.

(B) Poly(A)+ RNA was purified from the ventral midbrain, digested with DNase I and reverse transcribed with oligodT primers. A 3598bp sequence was amplified by PCR between the 5’UTR and ORF2. Importantly, there was no amplification in the absence of RT, thus no genomic DNA contamination (RT-). This validates the presence of full length LINE-1 RNA.

(C) ORF1p presence in the ventral midbrain was confirmed by Western blot.

(D) In situ hybridization with ORF1 RNA antisense probes (blue) shows that LINEs are transcribed in all midbrain dopaminergic neurons (mDA) of the SNpc marked by IHC/IF of tyrosine hydroxylase (TH). Scrambled probes were used as a negative control (right panel).

(E) ORF1p staining in TH-positive neurons in the SNpc confirms that ORF1p is expressed in all mDA neurons.

82

Figure 2. LINE-1 expression and nuclear localization are increased upon oxidative stress (A) Increased LINE-1 Tf/Gf transcription in the SNpc, 6 hours after nigral 6-OHDA injection. LINE-1 Tf/Gf RNA extracted from SNpc punches was quantified by qRT- PCR (***p<0.001, Student’s t test).

(B) Cell fractionation experiment showing a significant increase in nuclear, but not cytoplasmic LINE-1 Tf/Gf RNA from SNpc punches of 6-OHDA injected animals. LINE-1 A failed to yield a significant result. (*p<0.05; n=4 One-way ANOVA, Bonferroni's Multiple Comparison Test).

(C) Immunofluorescence staining of ORF1p and TH in the SNpc shows that ORF1p translocates in the nucleus after 6-OHDA nigral injection. Staining intensity in the nucleus of the injected and non-injected side of the same animals (n= 2) was quantified using Image J.

83

Figure 3. The piRNA binding protein Piwil1 represses the formation of DSBs and nuclear LINE-1 RNA foci under conditions of oxidative stress (A) Lipofection of a plasmid encoding Piwil1 protects midbrain primary neurons from oxidative stress. After H2O2 treatment, neurons transfected with Piwil1 show significantly less γ-H2AX positive foci than neurons transfected with a void plasmid. (***p<0.001; n=6, One-way ANOVA, Bonferroni's Multiple Comparison Test). No significant changes were observed in non-transfected (GFP-negative) neurons.

(B) Midbrain primary neurons infected with an AAV2 encoding Piwil1 show less γ- H2AX foci (left panel) and less active LINE-1 transcription sites (measured by fluorescent in situ hybridization (FISH)) (right panel) after H2O2 treatment than neurons infected with an AAV2 encoding GFP. (***p<0.001; n=6, One-way ANOVA, Bonferroni's Multiple Comparison Test).

(C) Mutated or wt plasmids overexpressing mouse LINE-1 were lipofected in midbrain primary neurons. 48 hours later, γ-H2AX immunofluorescence was performed and the number of γ-H2AX foci in the nucleus was counted. (***p<0.001; n=6, Student’s t test).

84

Figure 4. Neurodegeneration is accompanied by LINE-1 activation in En1+/- mice. (A) Increased LINE-1 Tf transcripts in the SNpc of 6 week-En1+/- mice compared to wild type littermates. RNA from laser dissected SNpc was sequenced and mapped against a consensus LINE-1 Tf sequence.

(B) This increase was verified by qRT-PCR in SNpc punches of 6 week-old heterozygous mice and their wild type littermates. However, no difference was observed between 1- year-old mice, although the abundance of LINE-1 transcripts increases with age. At 6 weeks no significant increase in transcription was measured for the LINE-1 A family.

(C) Midbrain sections of 8 week-old wild type and En1+/- mice were stained for ORF1p and analyzed by confocal microscopy. Quantification of ORF1p fluorescence intensity showed an increase in ORF1p in heterozygous TH+ neurons compared to wild-type (***p<0.001; n=30 neurons counted, Student’s t test).

85

Figure 5. Engrailed inhibits retrotransposition by directly repressing LINE-1 transcription. (A) Engrailed represses retrotransposition in HEK cells transfected with a retrotransposition reporter plasmid. (***p<0.001, n=3 more than 10000 cells counted per condition, Student’s t test)

(B) Chromatin immunoprecipitation (ChIP) experiments show that Engrailed binds to the 5’UTR of both LINE-1 Tf/Gf and A families. Before ChIP, purified nuclei, from midbrain primary neurons, were treated with Engrailed or sham, pre-incubated with polydI:dC (a polymer mimicking DNA to eliminate non specific binding) or En-NP6 (an oligonucleotide comprised of 6 copies of the Engrailed canonical binding sequence to eliminate specific binding). (***p<0.001; one-way ANOVA, Bonferroni's Multiple Comparison Test).

(C) En binding to LINEs was confirmed using a recombinant protein, EnHD-VP64, composed of the fusion of an extended Engrailed HD and an activator domain from the Herpes virus. Treatment with this protein in the presence of an inhibitor of translation, cycloheximide (CHX), both in vitro (left panel, Mbpn) and in vivo (nigral injection in the SNpc of wt mice), induced a significant increase in the transcription of both LINE-1 Tf/Gf and A. (*p<0.05, Student’s t test).

(D) Engrailed directly represses LINE-1 transcription in the SNpc after oxidative stress. Mice injected with 6-OHDA, then Engrailed and CHX display less LINE-1 transcription than 6-OHDA/sham/CHX injected ones. (**p<0.01, *p<0.05; Student’s t test).

86 III. Experimental procedures

Animals

Mice were treated in accordance with the guidelines for the care and use of laboratory animals (US National Institute of Health) and the European Directive 86/609 (EEC Council for Animal Protection in Experimental Research and Other Scientific Utilization). The En1 LacZ mouse line (Hanks et al., 1995) was bred from male heterozygous En1 mice with OF1 female mice (Janvier). Experimental groups consisted of at least 3 mice per genotype and condition aged from 6 weeks to 1 year.

RNA-seq analysis

The SNpc of wt and En1+/- mice was labeled using a quick TH-staining protocol (Chung et al., 2005) and isolated by Laser capture micro-dissection (LMD7000, Leica). Samples from 4 animals per group were pooled and total RNA was extracted using the AllPrep DNA/RNA Micro Kit (Qiagen) followed by DNase I using the RNeasy MinElute Cleanup protocol for on-column DNase I treatment. Construction of cDNA libraries (Ovation RNA-seq System V2) and Illumina RNA-seq were performed by the IBENS Genomic platform (Ecole Normale Supérieure, Paris).

In vivo treatments

For 6-OHDA injections, mice were placed in a stereotaxic instrument and a burr hole was drilled into the skull 3.3 mm caudal and 1 mm lateral to the bregma. The needle was lowered 4 mm from the surface of the skull and 6-OHDA (2 µl; 0.5 µg/µl Sigma) or sham (NaCl 0.9%) injections were performed over 4 min. For Engrailed rescue experiments, a solution (2 µl) of recombinant En2 (300 ng; 4 µM) (Sonnier et al., 2007) or vehicle (NaCl 0.9%) in the presence of colominic acid (3 µg) was injected 30 min after 6-OHDA injection. When indicated, cycloheximide (0.1 µg/µl, Sigma) was added. SNpc tissues for qRT-PCR and Western blot analysis were obtained by performing 1 mm punches from 2 mm thick frozen coronal slices. The ventral midbrain was dissected under a microscope with the help of micro-knives (Fine Science Tools).

87 Cell culture

Midbrain primary neurons were dissected from E 13.5 mice embryos and cultured in Neurobasal medium (Life Technologies) supplemented with glutamine (500 µM, Sigma), glutamic acid (3.3 mg/l Sigma), aspartic acid (3.7 mg/l, Sigma) and B27 (Gibco). Cells were treated with H202 (100µM) for 1 hour. Cells were fixed in 4% PFA/ PBS for 20 min and then treated with 100 mM glycine/PBS for 10 min. Cells were blocked in PBS containing 10% Goat serum for 1 hour, then overnight at 4°C with the primary antibodies (PBS + 3% goat serum + 0,1% triton X-100) and 1h at room temperature with the secondary antibodies. The transfection protocol was adapted from (Dalby et al. 2004). The plasmids (0,75 µg per transfection) were pre incubated with 8µl lipofectamine 2000 (Life Technologies) for 20 min at room temperature in Optimem medium (Life Technologies). Medium was added and the mix was added to the cells for 48 hours at 37°C, after which immunofluorescence was performed.

Retrotransposition assay

En inducible or normal HEK were cultured and treated with doxycycline for 1 day to induce the expression of En2. Then, cells were transfected with a retrotransposition reporter plasmid containing a GFP-based retrotransposition reporter cassette in the 3’UTR of a codon- optimized mouse LINE-1, similar to the ones described in (Xie et al. 2011). After one day, cells were split and three days later treated with puromycine (0,7 µg/µl Sigma) to select for transfected cells. Finally, after 1 week, the percentage of GFP positive cells per condition was measured using Fluorescence-activated cell sorting (FACS).

Chromatin immunoprecipitation

Nuclei from midbrain primary neurons were purified in a cytoplasmic lysis buffer

(10mM HEPES, 40mM KCl, 1mM CaCl2, 0.5% NP-40) for 10 min on ice and washed twice in the same buffer without NP-40 by centrifugation at 800g for 10 min at 4°C. Nuclei were then treated with En2 (500 ng/ml), sham (0.9% NaCl), polydI:dC (Sigma, 50 ng/µl) or NP6 (0.4 pmol/µl; an oligonucleotide composed of 6 times the Engrailed binding sequence (6x TCAATTAAATGA)) for 20 min at 37°C. Nuclei were then fixed in 1% formaldehyde (Sigma) in PBS. Chromatin was purified using the Magna ChIP kit (Millipore). Immunoprecipitations were performed with 1µg of anti Engrailed antibody (Ab) (86/8, in-

88 house rabbit polyclonal Ab against En1/2) or 1 µg of the kit rabbit IgG overnight at 4°C on a rotating wheel. The obtained chromatin was quantified by qPCR.

qRT-PCR

Total RNA from SNpc tissues was extracted using the RNeasy Lipid Tissue kit (Qiagen) followed by DNase I (Thermo) digestion. RNA (200 ng) was transcribed using the QuantiTect Reverse Transcription kit (Qiagen). qRT-PCR was performed using SYBR-Green (Roche Applied Science) on a Light Cycler 480 and values were normalized to Gapdh and/or Hprt. Data were analyzed using the ddCt method. In cell fractionation experiments, the nuclei purification part of the Magna ChIP protocol was used in the presence of RNase inhibitors (New England BioLabs, 1/500). The primers used for qPCR or qRT-PCR are the following:

Primers Forward Reverse

LINE-1 Tf/Gf CTGGGAACTGCCAAAGCAAC CCTCCGTTTACCTTTCGCCA

LINE-1 A TTCTGCCAGGAGTCTGGTTC TGAGCAGACCTGGAGGGTAG

Hprt AGCAGGTGTTCTAGTCCTGTGG ACGCAGCAACTGACATTTCTAA

Gapdh TGACGTGCCGCCTGGAGAAAC CCGGCATCGAAGGTGGAAGAG

Immunostaining

Mice were perfused with 4% paraformaldehyde in PBS, brains were post-fixed for 1h and cryoprotected in 20% sucrose overnight at 4°C. Tissues were embedded in Tissue-Tek O.C.T. (Sakura Finetek) and frozen in isopentane prior to storage at −80°C. Brains were cut into 20 µm-thick sections covering the whole SNpc. For immunofluorescence, slices were permeabilized in 1% Triton X-100 in PBS for 20 min and incubated at 100°C for 20 min in citrate buffer (10 mM citric acid, 0.05% Tween 20, pH 6.0). After blocking for 1h (10% normal goat serum, 0.05% Triton X-100 in PBS), slices were incubated overnight at 4°C in primary antibodies diluted in blocking solution (mouse anti-γ-H2AX, 1:200, Millipore; chicken anti-TH, 1:500, Abcam; ORF1p, 1:500). Sections were incubated with the appropriate secondary antibodies (488 anti-chicken, 647 anti-chicken, 488 anti-mouse, 546 anti-mouse and 546 anti-rabbit Alexa Fluor, Life Technologies) for 1h at room temperature. Labeled sections were imaged by confocal microscopy (SP5, Leica). For TH immunohistochemistry, sections were permeabilized in 1% Triton X-100 and incubated

89 overnight at 4°C with 10% normal goat serum in PBS containing a rabbit polyclonal antibody against TH (1:1000; Pel-Freez Biologicals). Sections were treated with biotinylated secondary antibody (Vector lab) and incubated with avidin-biotinylated horseradish peroxidase complex (ABC system, Vectastain). Peroxidase was revealed using the DAB peroxidase (HRP) substrate kit (Vector lab) and imaged with a Nikon Eclipse 90i microscope.

Image quantification

Images were analyzed with ImageJ. For immunofluorescence, all quantifications were performed using 63X magnification and 1 µm-thick successive focal planes. Cytoplasmic/nuclear ratio of ORF1p fluorescence intensity was determined by measuring pixel density in the nucleus and in the cytoplasm. For γ-H2AX, the number or foci was counted in the nucleus of transfected/infected cells. For each experiment 3 images in separate regions of the lamella were quantified, each point corresponds to more than 60 neurons.

In situ hybridization

20 µm mouse brain slices were fixed in 4% PFA in PBS for 10 min at RT, then permeabilized 2x10 min in RIPA buffer (150 mM NaCl, 1% NP-40, 0.5% Na deoxycholate, 0.1% SDS, 1 mM EDTA, 50 mM Tris-HCl pH8). They were fixed a second time for 5 min and demasked with TEA buffer (Triethanolamine 100 mM, 0.8% acetic acid pH8) + 0.25% acetic anhydride for 10 min. Slices were then permeabilized for 30 min in PBS + 1% triton X- 100 and blocked for 1 hour in hybridization buffer (50% formamide, 5x SSC, 5x Denhardt (1% Ficoll, 1% SSC, 1% tween 20), 500 µg/ml Salmon sperm DNA, 250 µg/ml yeast tRNA). Slices were incubated overnight with 10nM of two digoxigenin (DIG) labeled oligonucleotide probes (DIG Oligonucleotide 3'-End Labeling Kit, 2nd generation – Roche) in hybridization buffer at 37°C. Probes were designed against ORF1 (2 probes) and ORF2 (4 probes). Scrambled probes containing an equal base composition were used as a negative control. Slices were rinsed with FAM/SSC (50% formamide, 2x SSC, 0.1% tween 20) twice for 30 min at 37°C, then twice in 0.2X SCC at 42°C. Slices were blocked in B1 buffer (100 mM maleic acid pH 7.5, 150 mM NaCl) containing 10% fetal bovine serum (FBS) for 1 hour and incubated overnight at 4°C in B1 buffer +10% FBS with an antibody against DIG (Roche, 1/2000). Slices were rinsed three times in B1 buffer and once in B3 buffer (100 mM Tris-HCl

90 pH9, 50 mM MgCl2, 100 mM NaCl, 0.1% Tween 20) and revealed using a NBT/BCIP kit (Vector lab). Slices were rinsed with PBS and TH immunostaining was performed.

Oligo Séquence

ORF1-1 ACTAACAGGAACCAAGACCACTCACCATCACCAGAACCCAGCACACCCAC ORF1-1-Scr CCAAACGCACAACACGCACACCACCCTAACACGATGGCCCCATCCAAAAA ORF1-2 CTGCAGCCCAGGCTACTATACCCAGCCAAACTCTCAATTATCATAGAGGG ORF1-2-Scr ACAATAAGCAGCTGCTGAACCCTCTTAGTGTAAAAACCGCCATCTGGCCC

ORF2 -1 GCACACATTGCGCCTCACACAATAATAGTGGGAGACTTCAACACACC ORF2-1-Scr GTTTAACATTCCAACACGCCACTAAACGCAAAGACGCACGGTCCTGA ORF2-2 GCAATCATCCATCCTGACCAAGTAGGTTTTATTCCAGGGATGCAGGGATGG ORF2-2-Scr TGGCCCAATCACTTTGGAGAGTATCGTTTAGTCATGCGAATGGACGCAAGG ORF2-3 AATCAGGGACTAGACAAGGCTGCCCACTTTCTCCCTACCT ORF2-3-Scr ACTGCGCTGCACCAGTAGTTCACGGTTCCATCCCACAATA ORF2-4 GATCTCAGAAGATGGAAAGATCTCCCATGCTCATGGATTGGCAGGACC ORF2-4-Scr GTGACATGGATTTAGGCGGCGCACTATACAAATCAATGACCGGACGCT

IV. Acknowledgments

We would like to thank Dr Wenfeng An for the retrotransposition plasmids (WA125- 126). We also acknowledge the contribution of Dr Alain Joliot for the protein En2, EnHD- VP64 and for the En2-inducible HEK cells. We thank Pr Sandy Martin for the ORF1p antibody. We finally acknowledge the CIRB imaging and animal facilities for their help. The study was supported by Région Ile de France, Fondation Bettencourt Schueller, GRL program N°2009-00424 and ERC Advanced Grant HOMEOSIGN n°339379.

91

DISCUSSION

92

93 The data presented in this manuscript indicate that Engrailed counteracts several age- related processes. Ageing has been very well described in a review written by Lopez-Otin et al. in the journal Cell (López-Otín et al. 2013), that followed similar reviews of the hallmarks of cancer (Hanahan & Weinberg 2000; Hanahan & Weinberg 2011). To be considered as hallmarks of ageing the processes need to fulfill the following criteria (López-Otín et al. 2013):

• It should manifest during normal ageing • Its experimental aggravation should accelerate ageing • Its experimental amelioration should retard the normal ageing process and hence increase healthy lifespan.

This last condition, the hardest to achieve, is therefore not perfectly fulfilled by the nine chosen hallmarks (Figure 1). Finding nine different hallmarks was also difficult because of the interconnectivity between these different pathways.

The nine hallmarks of ageing are the following:

Figure 1. The nine hallmarks of ageing (López-Otín et al. 2013)

94 These pathways will be described in depth and their potential links with Engrailed will be discussed. This a posteriori approach, although repetitive, allows an unbiased assessment of the anti-ageing activity of Engrailed, by studying Engrailed under the prism of ageing and not the other way round. For clarity, the nine hallmarks of ageing will be divided into three different groups, each of them linked to an aspect of the present study; first, genomic stability and epigenetic alterations, in echo with retrotransposition, secondly, mitochondrial dysfunctions and loss of proteostasis, which could be linked to PD and finally the five other hallmarks of ageing which are telomere attrition, cellular senescence, deregulation of nutrient sensing, altered cellular communication and stem cell exhaustion, which could be linked to some aspects of Engrailed activity. Indeed, a KEGG pathway analysis performed on the differentially regulated RNAs in En1+/- compared to wt mice showed the involvement of several pathways, many of which can be associated with one or several of the nine hallmarks of ageing (Figure 2). When we independently analyzed upregulated and downregulated genes in En1+/- mice, additional gene categories appeared, in particular “protein processing in the endoplasmic reticulum” (ER), showing that Engrailed could be linked to other ageing-related pathways.

Figure 2. KEGG pathway analysis of differentially regulated mRNAs between 6 week-old En1+/- and wt mice

95 I. Genomic instability and epigenetic alterations

A. Genomic instability

Genomic instability is the first hallmark of ageing. Indeed, because of exogenous or endogenous stress, genome integrity is constantly under threat and genomic alterations accumulate with age (Moskalev et al. 2013).

Exogenous toxins are able to induce DNA damage, either directly for mutagens, or indirectly for example through oxidative stress like 6-OHDA (Bernstein et al. 2011). A very striking observation is that endogenous compounds or processes, particularly reactive oxygen species (ROS) produced by mitochondria, also challenge DNA integrity (Hoeijmakers 2009).

As previously seen, mitochondria produce oxidized components, particularly H2O2, which can oxidize DNA, creating mutations and DNA strand breaks. This illustrates the interconnections in ageing, as mitochondrial dysfunction, another hallmark, induces ROS formation, which in turn will create DNA damage (Hoeijmakers 2009). Many forms of DNA damage have been associated with ageing, indicating that this first hallmark is a key one. For example, somatic mutations accumulate in the nucleus of cells from either aged humans or model animals (Moskalev et al. 2012). Apart from this, chromosomal aneuploidies and copy number variations have also been linked to ageing (Faggioli et al. 2012; Forsberg et al. 2012). Ageing has also been associated with increased clonal mosaicism for large chromosomal abnormalities (Jacobs et al. 2012). Concomitantly, defects in DNA repair cause premature ageing in mice as well as progeroid diseases in humans: Xeroderma Pigmentosum, Werner syndrome, Cockayne syndrome, Seckel syndrome or Bloom syndrome among others (Gregg et al. 2012; Hoeijmakers 2009; Murga et al. 2009; Burtner & Kennedy 2010). Of note, mice overexpressing Bubr1, a protein involved in chromosome segregation, live longer and are more resistant to aneuploidy or cancer (Baker et al. 2013). Although particular, this example shows that increasing defenses against DNA damage may delay ageing.

Increased DNA damage has also been found in En1+/- mice, suggesting that Engrailed may be implicated in this hallmark of ageing. In these mice, gene set enrichment analyses performed on differentially regulated mRNAs at 6 weeks, corresponding to the beginning of

96 neurodegeneration, show DNA damage as one of the most altered category (Article 1 – Figure 1B). This was confirmed in the SNpc, where more DNA strand breaks (γ-H2AX staining) were observed in mDA neurons of En1+/- mice (Article 1 – Figure 2A). Very interestingly, and similarly to the situation in Parkinson disease, the VTA mDA neurons are partly spared (Article 1 – Figure 2B). An attractive explanation would be that these neurons, contrary to SNpc mDA neurons, express Otx2, another homeoprotein, which also protects mDA neurons from oxidative stress (Article 1 – Figure 5F). This raises the hypothesis that, similarly to Engrailed, other homeoproteins could be anti-ageing factors. Indeed Otx2 protects retinal ganglion cells in a model of glaucoma (Torero Ibad et al. 2011). Correlatively, overexpression of Engrailed reduces DNA damage in several oxidative stress models both in vitro and in vivo, showing the generality of Engrailed protection against oxidative-stress induced DNA damage. Indeed, in midbrain primary neurons, an overnight pretreatment with En1/2 diminishes the number of γ-H2AX positive foci after H2O2 treatment (Article 1 – Figure S4E and unpublished data). The fact that increased γ-H2AX means stronger DNA damage and not enhanced DNA repair was confirmed using a comet assay (Article 1 – Figure S4F). The same association was made in vivo in the SNpc after 6-OHDA injection, where neurons co-injected with Engrailed 30 minutes later, return to a normal γ-H2AX staining pattern after 1 week compared to sham-injected ones (Article 1 – Figure 5D). Timing is indeed an important issue, since the 6-OHDA injection experiment is the first example of protection with Engrailed added after the induction of oxidative stress.

Defects in nuclear architecture can also cause genomic instability (Dechat et al. 2008). For example mutations in nuclear lamins, the major components of the nuclear lamina, cause accelerated ageing in humans (Cabanillas et al. 2011; De Sandre-Giovannoli et al. 2003; M. Eriksson et al. 2003). Even more interesting, alterations of the lamina and synthesis of a malfunctioning form of Prelamin A, called Progerin, have been detected during normal ageing in humans (Ragnauth et al. 2010; Scaffidi & Misteli 2006). These nuclear lamina provide a scaffold for chromatin and chromatin regulating protein complexes (Gonzalez-Suarez et al. 2009; B. Liu et al. 2005). This aspect links genetic instability with another hallmark of ageing: epigenetic alterations.

97 B. Epigenetic alterations

As mentioned above, epigenetic alterations accumulating with age are another hallmark of ageing (Talens et al. 2012). These changes affect nearly all the aspects of the epigenetic network, and seem to favor a more open chromatin structure. These changes include modifications of DNA methylation patterns, and changes in post-translational modifications of histones. Indeed, increased H4K16 acetylation, H4K20 and H3K4 trimethylation or decreased H3K9 and H3K27 trimethylation have been associated with ageing (Fraga & Esteller 2007; S. Han & A. Brunet 2012). This network is regulated by a variety of proteins, including DNA and histone methylation or demethylation complexes and histone acetylases or deacetylases. In that context, puzzling observations have been made, as deficiencies in histones methyltransferase complexes increase lifespan in nematodes and flies (Greer et al. 2010; Siebold et al. 2010). Moreover inhibition of H3K27 demethylase extends lifespan in worms by targeting the insulin/IGF1-signalling (IIS) complex (Jin et al. 2011). The problem is less straightforward than expected, as the data in flies (H3K27 decreased methylation) and worms (H3K27 decreased demethylation) seem contradictory, probably due to the different downstream signalling pathways involved.

Thankfully, epigenetic changes are reversible, opening the way to new therapeutic strategies against ageing (Freije & López-Otín 2012; Rando & H. Y. Chang 2012). Indeed, over-expressing histone deacetylase in mice restores H4 acetylation and avoids age-related memory impairment (Peleg et al. 2010). Along with this idea, inhibition of histone acetyltransferase ameliorates symptoms in progeroid mice and extends their lifespan (Krishnan et al. 2011). These data indicate that the reversion of epigenetic states may have neuroprotective effects. Similar data were obtained with Engrailed, which reverses the epigenetic changes induced by 6-OHDA injection (Article 1). This lends further weight to the hypothesis of its anti-ageing activity.

The other very important category highlighted by our RNA-seq results is chromatin remodeling (Article 1 – Figure 1). The impact of Engrailed on chromatin structure was confirmed by the progressive disappearance of H3K27me3 heterochromatin and DAPI stained chromocenters in En1+/- mice (Article 1 – Figure 2C-F). The loss of H3K27 heterochromatin was reminiscent of the phenotype described previously as a marker of ageing. Similarly to

98 DNA strand breaks, Engrailed gain of function could reverse chromatin ageing phenotypes. We injected 6-OHDA in the SNpc of mice, inducing neurodegeneration accompanied by a global loss of heterochromatin marks (H3K9me3, H3K27me3, MeCP2 and Lamin B2) (Article 1 – Figure 4, S2). In this model, we injected Engrailed, 30 minutes after inducing this strong oxidative stress, and all the marks seemed to progressively go back to their “normal” values. This kinetic is very interesting, since it proves that Engrailed action could be curative, aligned with the theory of reversible epigenetic changes. An interesting experiment would be to take old wt mice, to verify that they display chromatin relaxation and to use Engrailed to try to reverse this age marker. Another example of the link between Engrailed and chromatin can be found in the terminals of Xenopus retinal ganglion cells where Engrailed can regulate the local translation of Lamin b2 (Yoon et al. 2012), a member of the nuclear lamina.

DNA methylation yields more complex results. Although ageing is generally associated with a global decrease in DNA methylation, several loci corresponding to tumor suppressor genes or Polycomb target genes are hypermethylated during ageing (Maegawa et al. 2010). Similar results were found in cells derived from patients or mice with progeroid syndromes (Osorio et al. 2010; Shumaker et al. 2006) as well as in cancer cells (Ehrlich 2002; Song & He 2012). Therefore, more work is needed to understand the causality - if any - between ageing and DNA methylation. DNA and histone-modifying proteins work in concert with key proteins like HP1α, and protein complexes like the Polycomb or NuRD complexes. The expression of these key factors is diminished during normal and pathological ageing (Pegoraro et al. 2009; Pollina & A. Brunet 2011). One consequence of this could be the loss or redistribution of heterochromatin, which is a characteristic of ageing (Oberdoerffer & Sinclair 2007; Tsurumi & W. X. Li 2012). This relaxation is associated with increased transcriptional noise and aberrant transcription and maturation of many mRNAs (Bahar et al. 2006; Harries et al. 2011; Nicholas et al. 2010).

DNA methylation is another possible route by which Engrailed could induce epigenetic changes. Indeed we have shown that 6-OHDA-mediated neurodegeneration is accompanied by a nearly complete disappearance of MeCP2, which binds to methylated cytosines (Article 1 – Figure S2D). Although we have not quantified the reappearance of MeCP2 after Engrailed injection in our 6-OHDA model, nor seen decreased MeCP2 in the

99 En1+/- mouse (only the size distribution of DAPI foci which superpose with MeCP2 expression), we have unpublished data linking Engrailed to methylated DNA. Indeed we were puzzled by the fact that Engrailed infusion increases dopaminergic activity (Alvarez-Fischer et al. 2011) and this up to at least 11 weeks after the end of the experiment (unpublished data). One hypothesis that we decided to follow was whether Engrailed could induce epigenetic changes, in particular changes in DNA methylation. To do so we performed methylated DNA immunoprecipitation after Engrailed infusion in the SNpc and deep- sequenced the precipitated DNA fragments. Much to our surprise nearly 50% of the obtained sequences derived from the LINE-1 class of retrotransposons. To check the specificity of this result we used a random dataset with an equal chromosomal distribution and size and found that this dataset did not overlap with LINE-1 annotations at all (0% overlap), validating LINEs as an epigenetic target of Engrailed (Figure 3). This finding is extremely interesting as it suggests that LINEs could link the first two hallmarks of ageing, namely epigenetic alterations and genomic instability.

Annotations Significant islands Random dataset Cpg Island 25 1% none 0% Whole gene 1545 64% none 0% Coding exons 1184 49% none 0% LINE-1 1189 49% none 0%

Figure 3. MeDIP-seq partial results

C. LINE-1 the missing link between some hallmarks of ageing

As already mentioned, LINEs can induce DNA strand breaks via ORF2p endonuclease activity. Furthermore, their expression is repressed by the Piwi/piRNA complex but also by epigenetic changes, including DNA methylation and heterochromatin formation. Similarly to the restructuration of chromatin architecture, the genome undergoes during ageing, we have observed changes in LINE-1 methylation in both directions (hypo- and hypermethylation in equivalent proportions) after Engrailed infusion. Despite this ambiguous result, LINE-1 association with ageing is a promising possibility, as their transcription increases with age in the central nervous system (W. Li et al. 2013). This may be a consequence of chromatin relaxation, as heterochromatin contains many of these repetitive sequences (Fadloun et al.

100 2013). Age-related LINEs activation in the brain could participate to genomic instability through their ability to induce DNA strand breaks or mutagenesis. This hypothesis is reinforced by the association between LINE-1 activation and schizophrenia or several age- related processes, including cancer or Alzheimer disease (associated with chromatin relaxation) (Kitkumthorn & Mutirangura 2011; Bundo et al. 2014; Frost et al. 2014), described in the introduction.

In our second study, we have further implicated LINE-1 with neurodegeneration. First, we found that LINE-1 transcription was increased in all our experimental models of neurodegeneration. In the En1 heterozygous mice, LINE-1 Tf RNA, the most active family of LINE-1, was increased at 6 weeks, a time when mDA neurons start to degenerate but not at 1 year when a plateau is reached in neurodegeneration (Part 2 – Figure 4B). Similar results were found when quantifying LINE expression as percent of TH instead of Hprt (data not shown), showing that the plateau is not due to the loss of TH+ neurons (around 40% at this age (Sonnier et al. 2007)). In the heterozygote, LINE-1 RNA levels at 6 weeks and 1 year were identical (Part 2 – Figure 4B). As retrotransposition increases with age (W. Li et al. 2013), we see this result as a sign of premature ageing in the En1+/- mice. Conversely, in the wild type, we confirmed the age-related increase in LINE-1 transcription. This was paralleled by an increase in ORF1p in 8 week-old heterozygous mice, showing that LINE-1 RNA increase is followed by translation and does not reflect pure transcriptional deregulation. Similarly, in 6- OHDA injected mice retrotransposition is increased (Part 2 – Figure 2A-B). Another striking observation is that after 6-OHDA injection, LINE-1 RNA increase was higher in the nucleus. This may reflect both transcription increase and the translocation of the RNP into the nucleus. Indeed, following 6-OHDA, ORF1p translocated from the cytoplasm to the nucleus of mDA SNpc neurons (Part 2 – Figure 2C). This could be a sign of retrotransposition, if we consider that ORF1p is present in the nucleus with ORF2p, the latter being able to induce DNA damage and mutagenesis by insertion. These data support the idea that LINE-1 activation is a consequence and/or a mediator of ageing and neurodegeneration.

To go further into this direction, we needed to prove that LINE-1 activation after oxidative stress could create DNA damage. Our experimental paradigm was to block LINE-1 activation during oxidative stress and to measure the levels of DNA damage. This causative link was first proved in vitro through the overexpression of Piwil1. Overexpression of Piwil1

101 in midbrain primary neurons strongly decreased the number of γ-H2AX foci, reinforcing the hypothesis that LINEs are active in neurons and have the ability to compromise genomic integrity (Part 2 – Figure 3A-B). Interestingly, in this experiment, Piwil1 transfection seemed to decrease DNA damage also in control condition (although it is not statistically significant), showing that LINEs may have an impact on the stress mediated by cell culturing and/or transfection. Piwil1 mainly represses LINE-1 expression by RNA interference, however viral overexpression of Piwil1 also decreased LINE-1 transcription as shown by the reduced number of transcription spots (fluorescent in situ hybridization) (Part 2 – Figure 3B right panel), suggesting that other pathways may be at stake in Piwil1-mediated repression. One hypothesis may be that Piwil1 induces methylation of the LINE-1 promoters and /or induces heterochromatinization, which represses their transcription, as described in other models (Huang et al. 2013). To obtain direct proof that LINE-1 transcription mediates DNA damage, we transfected a plasmid overexpressing a mouse LINE-1 (A generous gift from Jeff Boecke John Hopkins University School of Medicine, Baltimore, USA & Wenfeng An South Dakota University, Brookings, USA); here again increased DNA damage was measured compared to transfection with a plasmid containing a double mutated ORF2, without endonuclease or reverse transcriptase activity (Part 2 – Figure 3C). If what we show in our in vitro models could be extrapolated to PD or to other ageing-derived pathologies, new therapeutic strategies could be imagined. We intend to test this hypothesis using DNA from the SNpc of PD patients and non-parkinsonian controls and measure the number of LINE-1 insertions.

A link between LINE-1 expression and ageing can be found in a class of NAD- dependent protein deacetylases, the Sirtuins (Guarente 2011). A single member of this family, Sir2, has very promising anti-ageing properties in yeast, flies and worms (Kaeberlein et al. 1999; Rogina & Helfand 2004; Tissenbaum & Guarente 2001). However, recent studies have questioned the extension of lifespan in worms (Burnett et al. 2011). In fact drosophila Sir2 extends lifespan but in a more modest way than initially claimed (Viswanathan & Guarente 2011). Similar results were found with Sirt1, the closest equivalent in mammals. Although it does not extend lifespan, Sirt1 ameliorates several age-related processes in mice (Herranz et al. 2010). The underlying mechanisms are rather complex but involve genetic stability (Oberdoerffer et al. 2008; R.-H. Wang et al. 2008) as well as enhanced metabolic activity (Nogueiras et al. 2012), showing again the interconnections between the different hallmarks of ageing. Similarly the seven mammalian Sirtuins can ameliorate various aspects of ageing in

102 mice (Houtkooper et al. 2010; Sebastián et al. 2012). A more compelling example can be found with Sirt6, which regulates NK-κB signalling and glucose homeostasis via H3K9 deacetylation (Kanfi et al. 2010; Kawahara et al. 2009; Zhong et al. 2010). Moreover, mice mutated for Sirt6 exhibit accelerated ageing (Mostoslavsky et al. 2006), and overexpression of this protein in mice increases their lifespan, probably through IGF-1 signalling (Kanfi et al. 2012).

A recent article demonstrates a direct link between Sirt6 and LINE-1 (Van Meter et al. 2014). The authors report that Sirt6 directly binds to the LINE-1 5’UTR where it ribosylates the nuclear corepressor protein Kap6. This increases the interaction of Kap6 with the heterochromatin protein HP1α, which induces the formation of repressive heterochromatin around LINE-1 elements and inhibits their activation (Van Meter et al. 2014). Moreover, during the course of time or in response to DNA damage, Sirt6 is depleted form LINE-1 promoters, which could participate in the increased LINE-1 activity observed during ageing (Van Meter et al. 2014). The link between Sirt6, a very important anti ageing factor, and LINE-1 repression further validates LINE-1 activation as a hallmark of ageing, at the crossroad of epigenetic alterations and genomic instability (Figure 4).

Figure 4. Model for Sirt 6 mediated age related activation of LINE-1 (Van Meter et al. 2014)

We have obtained similar data with Engrailed, showing that, similarly to Sirt6, its anti- ageing capacity might include the repression of retrotransposition.

103 D. Engrailed direct repression of LINEs could mediate its anti-ageing properties

Having observed that Engrailed infusion induces methylation changes in LINE-1 5’UTR (Figure 3), we wondered whether Engrailed, similarly to Sirt6, might not directly bind the LINE-1 5’UTR. This hypothesis was made even more appealing by the presence of in silico canonical Engrailed binding sequences in the LINE-1 5’UTR. To verify if these predicted binding sites were occupied, we performed chromatin immunoprecipitation (ChIP) experiments, which showed that Engrailed binds to the 5’UTR of both LINE-1 Tf/Gf and A families (Part 2 – Figure 5B). To further validate this result, we used a recombinant activator form of Engrailed (EnHD-VP64), which increased LINE-1 Tf/Gf and A transcription, both in vitro (midbrain primary neurons) and in vivo (in the SNpc) in the presence of cycloheximide (CHX) an inhibitor of translation (Part 2 – Figure 5C). The presence of CHX reinforces the hypothesis of a direct LINE-1 regulation by Engrailed (although we cannot formally exclude a non coding RNA intermediate).

To study the functionality of this regulation, we transfected the LINE-1-expressing retrotransposition reporter plasmid, mentioned above, in HEK cells inducible for En2 expression. This plasmid allows the expression of GFP if at least one full retrotransposition event takes place. After Engrailed induction the percentage of GFP expressing cells was significantly reduced (Part 2 – Figure 5A), showing that Engrailed indeed represses retrotransposition. Similarly, in the nigral 6-OHDA injection model, we co-injected Engrailed with CHX and observed that transcription of both LINE-1 Tf/Gf and A was significantly reduced (Part 2 – Figure 5D). This confirms that Engrailed is a direct repressor of LINE-1 transcription, in particular in stress conditions, in which their transcription is activated. It is therefore very tempting (and not too absurd) to speculate that Engrailed neuroprotection in the 6-OHDA mouse model is partially through LINE-1 repression. Most of the logical steps for such an assertion have been verified in this study. Engrailed protects mDA SNpc neurons from degeneration, Engrailed represses LINE-1 and DNA strand breaks are, at least partially, induced by LINE-1 activation. However the last step would be to show that it is indeed the direct repression of LINE-1 by Engrailed that mediates neuroprotection; unfortunately it is extremely hard to prevent the interaction between LINEs and Engrailed. One possibility to disturb this interaction would be to mutate the Engrailed-binding site in the LINE-1 5’UTR. However, even if this could be done with a plasmid construct, the endogenous LINE-1 would

104 remain, and their large number (3000 “theoretically” active LINEs in the mouse) would forbid efficient mutagenesis. Another possibility would be to use an Engrailed unable to repress LINE-1. This would not be fully satisfactory, as other direct transcriptional targets of Engrailed might also be neuroprotective. Gadd45ß provides a good example, as it is a direct target of Engrailed (Article 1 – Figure 6C-D) and because its induction could inhibit apoptosis via the decrease in JNK phosphorylation (Coffey 2014) (Article 1 – Figure 6E, F). In fact, Gadd45ß is also a promising protein against ageing since it increases longevity in flies, and potentially impacts different other hallmarks of ageing (Moskalev et al. 2012). Its direct induction by Engrailed further increases the great potential of Engrailed as an anti-ageing protein. Gadd45ß and γ could also be, according to preliminary data in the lab, direct targets of Otx2, generalizing the concept that homeoproteins could be anti-ageing factors. Despite the impossibility to specifically disturb the LINE-1/Engrailed interaction in vivo, the data presented here strongly suggests that Engrailed repression of LINE-1 transcription mediates parts of its anti-ageing properties.

The fact that Engrailed represses LINE-1 via a direct transcriptional mechanism is hardly questionable, as the presence of CHX during Engrailed injection or the absence of true chromatin on the RT plasmid does not favor an indirect mechanism. But, as we have seen in our preliminary MeDIP analysis, Engrailed infusion also induces methylation changes in the promoter region of LINE-1 suggesting that there might be, in addition, a second and epigenetic level of repression. Similar parallel mechanisms were described with Piwil1- mediated methylation of LINE-1 promoters alongside the Ping-Pong cycle (Huang et al. 2013). Experiments are still ongoing, but they already underscore the potential therapeutic anti-ageing power of Engrailed, which could, shortly after it administration, directly and rapidly repress retrotransposition and then install an epigenetic long-lasting repression of LINE-1 transcription, via the methylation of their promoter. As Engrailed regulates translation (I. Brunet et al. 2005; Wizenmann et al. 2009; Alvarez-Fischer et al. 2011; Stettler et al. 2012), we performed a pull-down assay to identify potential regulatory small RNAs bound to the protein (unpublished data). Among these small RNAs are miRNAs (some of them described succinctly on Figure 5), as well as potential piRNAs, already described in the brain (E. J. Lee et al. 2011). Regarding the identified miRNAs, some of them are potential regulators of key processes and will be studied in depth (Donzelli et al. 2012; Wu et al. 2010; Chandrasekar & Dreyer 2009; Dubinsky et al. 2014)

105 miRNAs bound to Engrailed Validated targets Trp53, Cdkn1a/b, Cpeb1, Bcl2, Akt1, Mmu-mir-128-1 Dnmt3b, TNF, Pten, Dicer1 Trp53, Jarid1b (Kdm5b), Piwil2, Tnf, Trp53, Mmu-let-7d (linked to Igf2, Cdkn2a/1a/1b, Smad3/5, Dicer1, Ephb2, neurodegeneration- p53 Akt1, Egr1, NeuroD1, Bcl2, Pten, Sox2, Lrrk2) Dnmt3b; Casp3, Gadd45α, Jarid1b, Drd3

Figure 5. Some of the miRNAs bound to Engrailed

Although we need to thoroughly validate the latter observations, an exciting idea emerges from these results. These still potential piRNAs could guide Engrailed toward the active LINEs and mediate de novo DNA methylation of their promoters, similarly to what has been described for Piwi (Huang et al. 2013). We can even push the theory forward and imagine that Engrailed transfers small non coding RNAs (ncRNAs) (including pi and miRNAs) between cells and methylate active LINE-1 loci or regulate translation in a non-cell autonomous manner. This is for the moment very theoretical, but nevertheless the binding of ncRNAs to Engrailed opens the possibility to consider homeoproteins as part of the Argonaut family.

Engrailed transcriptional protective abilities have been well demonstrated, as a transcriptional “anti-Engrailed”, EnHD-VP64 (which activates Engrailed direct transcriptional targets), induces DNA strand breaks and cell death in the SNpc (Article 1 – Figure S3A-C). To understand the action of this protein, one has to remember that Engrailed is primarily a repressor, although we know that this is at least untrue for Gadd45ß, which is directly activated by Engrailed (Article 1 – Figure 6C-D). As EnHD-VP64 induces LINE-1 transcription (Part 2 – Figure 4), we can hypothesize that EnHD-VP64 mediated cell death is, at least partially, via LINE-1 activation, although other targets are probably involved. The latter targets will be identified in the future through RNA-seq following EnHD-VP64 and En1/2 injections in the presence of CHX.

Engrailed seems to be a very general repressor of LINE-1 activity and this at several levels, first as a direct transcriptional repressor, but also as an epigenetic regulator, possibly through its ability to bind small non coding RNAs. To realize the multifunction abilities of

106 this homeoprotein, one must remember that it can also control local translation in growth cones and synaptoneurosomes (I. Brunet et al. 2005; Alvarez-Fischer et al. 2011; Yoon et al. 2012). Therefore a powerful image for Engrailed could be that of an anti-ageing Swiss army knife protein, protecting neurons at several levels, transcriptional, translational and epigenetic. As far as translational control goes, Engrailed regulates the translation of nuclear-encoded CI mRNAs, thus providing an entry into the mitochondria hallmark of ageing.

II. Loss of proteostasis and mitochondrial dysfunction

Proteostasis and mitochondrial dysfunction, two hallmarks of ageing, are of immediate interest as they recapitulate two of the most important biology mechanisms implicated in Parkinson disease, where age is the principal risk factor (Rodriguez et al. 2015). As we have seen in the introduction, one of the first pathological signs of PD is the presence of protein aggregates, the Lewy bodies, in neurons (Goedert 2001; Spillantini et al. 1997). Cause or consequence of the disease, α-Synuclein inclusions indicate that proteostasis is malfunctioning in Parkinson disease. Furthermore, many Parkinson-related mutations affect proteostasis-related pathways, including the Ubiquitin pathway and autophagy (Leroy et al. 1998; Lynch-Day et al. 2012). Mitochondrial dysfunction is another hallmark of PD: as already mentioned, SNpc mDA neurons from PD patients show a higher level of oxidative stress as well as mitochondrial CI deficiencies (M. H. Irwin et al. 2013; Ciccone et al. 2013). It is also noteworthy that PD-inducing drugs, like MPTP/MPP+, target this complex, showing its importance in the pathology of the disease. Moreover, several PD genes are related to mitochondria quality control and oxidative stress. Finally, these two hallmarks are probably linked, as MPTP/MPP+ treatment in mice, which inhibits CI, induces the formation of Lewy bodies. This shows that mitochondrial dysfunction can lead to loss of proteostasis and that similarly to other hallmarks of ageing, these two pathways are interconnected.

A. Loss of proteostasis

Loss of proteostasis is a common feature of many age-related diseases (Powers et al. 2009). PD, the second most important neurodegenerative disorder, has already been

107 mentioned, but the most frequent one, Alzheimer disease, is characterized by the accumulation of Amyloid ß plaques and hyper phosphorylated tau proteins in the neurons (Small & Duff 2008). Proteostasis is composed of two different mechanisms, the stabilization of correctly folded proteins, mainly via the heat-shock family of proteins, and the degradation of unfolded or misfolded proteins, through the proteasome or the lysosomes (Hartl et al. 2011; Koga et al. 2011; Mizushima et al. 2008). These mechanisms prevent the accumulation of unfolded proteins, and are responsible for the continuous renewal of cellular proteins by degrading the useless or damaged ones. Several studies have demonstrated that proteostasis is altered with ageing (Koga et al. 2011; Liang et al. 2014; Morimoto & Cuervo 2014). As a consequence many ageing diseases, like neurodegenerative diseases but also cataracts, are associated with the accumulation of unfolded, misfolded or aggregated proteins, which participate to their development (Powers et al. 2009).

Protein degradation declines with age, in particular the two main systems, the autophagy/lysosome and the Ubiquitin/proteasome pathways (Rubinsztein et al. 2011; Tomaru et al. 2012). This supports the view that protein homeostasis defects are a key component of ageing. Overexpression of the chaperone mediated autophagy receptor LAMP2a prevents the apparition of ageing symptoms in the liver, which are associated with a decline in autophagy (C. Zhang & Cuervo 2008). Very strikingly, rapamycin, a chemical inducer of macro-autophagy (another form of autophagy), extends lifespan in middle-aged mice (Blagosklonny 2011; Harrison et al. 2009). This has attracted considerable attention to this product, which is also able to decrease other aspects of ageing in mice (J. E. Wilkinson et al. 2012). For many organisms, like worms, nematodes or flies, the effect of rapamycin is strictly dependent on the induction of autophagy (Rubinsztein et al. 2011; Bjedov et al. 2010). However, in mammals the problem is more complex and it seems that other pathways such as the inhibition of the S6 protein kinase 1 (S6K1), which regulates protein synthesis, are also implicated (Selman et al. 2009). Similarly, other compounds related to autophagy increase lifespan in model organisms: spermidine, polyamines and ω-6 polyunsaturated acids in particular (Eisenberg et al. 2009; Matsumoto et al. 2010; Soda et al. 2009; O'Rourke et al. 2013).

108 Although En1+/- mice do not have Lewy bodies, protein degradation seems modified in heterozygous mice. When we analyzed the results of the RNA-seq in these mice, using KEGG pathways, gene categories related to proteostasis appeared differentially regulated. For instance the “Ubiquitin mediated proteolysis” category is modified in the En1 heterozygotes (Figure 6). Autophagy is also modified in En1+/- mice and could even be an early feature appearing before actual degeneration, with the presence of abnormal autophagy vacuoles in degenerating nigral terminals in the striatum (Nordström et al. 2015). Although these data do not necessarily mean that Engrailed directly regulates proteolysis, they strongly suggest a link between Engrailed and the protein degradation pathway. One of the possible routes could be similar to the one between inhibition of CI by MPTP and the formation of Lewy bodies. Indeed, in MPTP injected mice, Lewy bodies are formed in SNpc mDA neurons in the mouse and in the baboon (Fornai et al. 2005; Kowall et al. 2000). In this PD drug model, Engrailed directly increases CI activity (Alvarez-Fischer et al. 2011) and could therefore reverse the formation of these protein inclusions.

Ubiquitin mediated proteolysis LOF - adjP=0.0119

12 0.06 10 8 0.04 6 Figure 6. Genes from the

4 0.02 value p 2 “Ubiquitin mediated proteolysis” 0 0.00 pathway modified in the En1+/- -2 -4 -0.02 mouse -6

Fold change Fold -8 -0.04 2 -10 adjP = adjusted p value

Log -12 -0.06

Traf6 Ube2i Cdc16 Ube2k Ube2l6 Anapc4

Another facet of this problem is the correct folding of proteins; indeed, overexpression of protein folding chaperones increases lifespan in worms and flies (Morrow et al. 2004; Walker & Lithgow 2003). Of particular interest are proteins of the heat-shock family: mice deficient in one member of this family show accelerated ageing, whereas long-lived mouse strains present an increased expression of these proteins (Min et al. 2008; Swindell et al. 2009). Moreover, activation of a master regulator of the heat-shock response, HSF-1, increases longevity in nematodes (Chiang et al. 2012; Hsu et al. 2003); equivalent results were found using amyloid binding components (Alavez et al. 2011). These processes could be linked to the previously described Sirtuin family, as deacylation of HSF-1 by Sirt1 in

109 mammalian cells potentiates the transactivation of HSP70, one of the key factors of the heat- shock response; conversely, downregulation of Sirt1 has the opposite effect (Westerheide et al. 2009). Of note are also results on treatments inducing the heat-shock response or increasing the expression of chaperone proteins, which lead to interesting ameliorations in ageing models (Gehrig et al. 2012; Calamini & Morimoto 2012). Here again, we have not observed a direct effect of Engrailed on protein folding. However in one of our RNA-seq analysis we have seen a significant modification of the “Protein processing in the ER” pathway in the En1+/- mouse (Figure 7.), which could indicate that this hallmark might be linked in some way to the Engrailed pathway, although deciphering it will require further study.

Protein processing in the ER LOF - adjP=0.008 (genes down) 12 0.06 10 8 0.04 6

4 0.02 value p Figure 7. Genes from the 2 “Protein processing in the ER” 0 0.00 +/- -2 pathway modified in the En1 -4 -0.02 -6 mouse

Fold change Fold -8 -0.04 2 -10

Log -12 -0.06 Preb Derl2 Uggt2 Fbxo6 Man1c1 Dnajb11

Loss of proteostasis, one of the key hallmarks of ageing and probably a fundamental feature of Parkinson disease, may thus be linked to the Engrailed pathway. The absence of protein inclusions in the En1 heterozygous mouse probably indicates that this link is not direct. However, protein folding and hydrolysis of unfolded proteins, particularly autophagy, seem to be modified in the heterozygote. As previously mentioned, oxidative stress could be an inducer of protein unfolding, and our data show that Engrailed protects mDA neurons from oxidative stress.

B. Reactive oxygen species and oxidative stress.

As age increases, the efficiency of the mitochondrial chain tends to decrease, inducing electron leakage and the creation of reactive oxygen species (ROS) (Green et al. 2011). This led to the mitochondrial free radicals theory: age-related progressive mitochondrial

110 dysfunction results in increased ROS production, which will in turn induce more mitochondrial damage and global cell ageing (HARMAN 1965).

This “catch 22 theory” has been revaluated in recent years, due to the unexpected discovery that ROS may prolong lifespan in both yeast and worms (Hekimi et al. 2011; Doonan et al. 2008; Mesquita et al. 2010; Van Raamsdonk & Hekimi 2009). Similarly, genetic manipulations to increase ROS production and oxidative damage in mice do not accelerate ageing (Van Remmen et al. 2003; Y. Zhang et al. 2009). Conversely, mice with increased antioxidant defense do not have extended lifetime (Pérez et al. 2009). Finally, genetic manipulations, which induce mitochondrial dysfunctions but do not increase ROS production, accelerate ageing (Edgar et al. 2009; Hiona et al. 2010; Kujoth et al. 2005; Trifunovic et al. 2004; Vermulst et al. 2008). These data show that more studies need to be done to decipher the link between ROS and ageing (Ristow & Schmeisser 2011). Another example has been given by another team from our research centre: in the zebrafish, ROS serve as signals for the regeneration of the caudal fin, and their inhibition compromises healing and regeneration (Gauron et al. 2013). Therefore the view of ROS as purely negative entities is challenged by recent data in many different organisms and a more unifying theory needs to be developed. A more precise explanation for this apparent contradiction may be that ROS could serve as stress signalling factors. A mild oxidative challenge could constitute a stress-elicited survival signal, designed to activate compensatory homeostatic responses. This apparent positive response could be overrun by a stronger stress, which might raise ROS concentration to pathological levels, as we see for example in our own H2O2 in vitro or 6- OHDA in vivo models. Going above this pathological threshold might induce DNA damage, protein inclusions or chromatin relaxation: all hallmarks of ageing.

Engrailed protective capacity in at least three oxidative stress models (H2O2 in vitro, MPTP and 6-OHDA in vivo) suggests that it might act at the source of these stresses, ie against general oxidative stress mechanisms or through direct ROS control. However, we have very little data yet supporting the latter hypothesis: the most solid ones come from the RNA-seq in the heterozygous mouse, where the gene category “peroxisome” is differentially regulated (Figure 8). If confirmed, the possible “regulation” by Engrailed of these organelles, which detoxify cells from oxidized compounds, would be very interesting. Engrailed conformation could also be regulated by the oxidative status of the cell, as shown by

111 preliminary results from the team of Alain Joliot in our center of research. There is a very evolutionary conserved cysteine near the homeodomain of Engrailed, which has attracted their attention, because of the known sensitivity of cysteines to oxidative stress. When mutated at this amino acid, Engrailed looses the ability to form dimers and to exert its paracrine activity (Rampon et al. 2015), although its transcriptional activity and binding to PBX remain untouched. It is therefore tempting to imagine that oxidative stress influences Engrailed conformation, therefore its activity, via the oxidation of this cysteine.

Peroxisome

12 LOF - adjP=0.0013 0.06 10 8 0.04 6 4 0.02 p value p 2 Figure 8. Genes from the 0 0.00 “peroxisome” pathway -2 modified in the En1+/- mouse -4 -0.02 -6

Fold change Fold -8 -0.04 2 -10

Log -12 -0.06 Xdh Idh2 Gnpat Abcd1 Pex12 Mpv17l

In this line of thought, one could imagine Engrailed as a ROS sensor, participating in maintaining ROS physiological levels, perhaps via its “regulation” of peroxisomes. This could explain its impressive ability to buffer the harmful consequences of a violent oxidative stress as demonstrated in the case of cells exposed to MPTP, 6-OHDA or H2O2.

C. Mitochondrial dysfunctions

As previously mentioned, mitochondrial dysfunctions can induce ageing phenotypes independently of oxidative stress. For example, mice deficient for mitochondrial polymerase- γ exhibit premature ageing and reduced lifespan accompanying the accumulation of random point mutations in the mitochondrial genome. These symptoms are paralleled by decreased mitochondrial functions, but unexpectedly without significant increase in ROS production (Edgar et al. 2009; Hiona et al. 2010; Park & Larsson 2011). This provides another example of the interconnections between the hallmarks of ageing; indeed mitochondrial DNA (mtDNA) is particularly sensitive to DNA damage because of the highly oxidative environment, the absence of protective histones and the rather inefficient mtDNA repair

112 mechanisms (Linnane et al. 1989). However the link between mtDNA mutations and ageing is not so direct: because of the multiplicity of mitochondria in one cell, some will carry the mutation but some will not. The mechanisms for mitochondrial-induced ageing are not yet known, but several hypotheses have been proposed. For example mitochondrial dysfunction might enhance the permeability of mitochondrial membranes in response to stress, thus creating a “vicious circle” by increasing oxidative stress in the cells (Kroemer et al. 2007). Another possibility is that mitochondrial dysfunction perturbs inter-organelles communication, particularly the crosstalk between the outer mitochondrial membrane and the endoplasmic reticulum; such dysfunctions could hamper proteostasis regulation (Raffaello & Rizzuto 2011).

These dysfunctions could result from several converging pathways, for example telomere attrition, another hallmark of ageing. In telomerase deficient mice, mitochondrial dysfunctions are observed, with a p53-induced repression of both PPARγ co-activator 1 (PGC1) α and ß (E. Sahin & DePinho 2012), which are members of a family controlling mitochondria biogenesis (Scarpulla 2011). Similar observations were made during normal ageing. Interestingly, in both conditions, this repression could be reversed by telomerase activation (E. Sahin & DePinho 2012; Bernardes de Jesus et al. 2012). Sirt1 modulates mitochondrial biogenesis via the co-activator PGC1α and the removal of malfunctioning mitochondria by autophagy (Rodgers et al. 2005; I. H. Lee et al. 2008). Another member of this definitively interesting family of Sirtuins - Sirt3 - targets many proteins implicated in mitochondrial activities: electron chain transfer and tricarboxylic acid cycle among others (Lombard et al. 2007; Giralt & Villarroya 2012). Sirt3 may also control ROS production by deacetylating manganese superoxide dismutase, one of the most important antioxidant enzyme in mitochondria (Tao et al. 2010; Qiu et al. 2010). This supports the view that mitochondrial dysfunctions are closely linked to several other hallmarks of ageing.

Another important feature of mitochondrial homeostasis is mitophagy, the mitochondria specific autophagy, which targets deficient mitochondria for degradation (K. Wang & Klionsky 2011). This process, which becomes deficient with age, is one of the features of PD (Ryan et al. 2015). Mitophagy may also be linked to the beneficial effects of fasting, which constitutes a very potent trigger of autophagy (Rubinsztein et al. 2011). Its

113 direct activation of mitochondrial CI (Alvarez-Fischer et al. 2011) makes Engrailed a promising anti-ageing protein. Indeed a lessened efficiency of the respiratory chain, particularly of CI, observed during ageing, decreases ATP synthesis and increases electron chain leakage (Green et al. 2011). Engrailed has the opposite effect as it potentiates CI activity; a slightly paradoxical phenomenon however, as activation of CI induces also increases ATP synthesis, thus ROS production and oxidative stress (Alvarez-Fischer et al. 2011). In this context it is quite interesting to note that Engrailed protects against oxidative stress to which it contributes (this study).

This takes us to the “mitohomeosis” theory, on which many ageing-related researches seem to converge (Calabrese et al. 2011). According to this theory, mild toxic treatments could induce a stress response, the beneficial effects of which surpass the negative effects of the initial stress. Therefore mild stresses could have positive effects, rendering the cells more able to fight stronger stresses, a sort of mithridatism or vaccine against stress. For example, mild respiratory chain deficiencies may increase lifespan (Haigis & Yankner 2010), probably because of mitochondrial defense mechanisms, which can also transfer to distant cells in C. elegans (Durieux et al. 2011). Similarly, mild mitochondrial poison like metformin and resveratrol retard ageing, via AMPK signalling for metformin and Sirtuin and PGC1α activation for resveratrol (Hawley et al. 2010; Onken & Driscoll 2010). PGC1α seems to be a key player in ageing as its overexpression extends lifespan in flies and improves their mitochondrial activity (Rera et al. 2011). Despite what the loi Evin says, we should advertise for French red wines, and even consider having them reimbursed by social care systems, as they contain resveratrol and may therefore have anti-ageing properties (Bastianetto et al. 2015). This is linked to the concept of the “French paradox”, as French consume a lot of saturated fats but have less coronary heart diseases (Ferrières 2004). This paradox could be explained by the consumption of resveratrol in red wine (B.-L. Liu et al. 2007). However, for wine as for stress, there is no good or bad, only dose matters (Biagi & Bertelli 2015).

114 III. Other hallmarks of ageing

The five remaining hallmarks of ageing do not seem directly linked to the object of the present study. However, a careful analysis shows that they might be relevant to the physiology of homeoproteins and of Engrailed in particular. This further underscores the interest of the “anti-ageing homeoprotein” hypothesis.

A. Telomere attrition and cellular senescence

As seen previously, DNA damage is one of the key hallmarks of ageing. It seems to affect random regions of the genome with the exception of short DNA repeats at the end of chromosomes, the telomeres. Telomeres are indeed particularly susceptible to age-related deterioration (Blackburn et al. 2006). Replicative DNA polymerases do not have the ability to replicate these regions due to their terminal position on chromosomes; this function is thus taken up by reverse transcriptases, called telomerases, which elongate telomeres. However, most mammalian cells do not express telomerases, or enough of them, therefore a continuous shrinking of the telomeres takes place, division after division. Telomere exhaustion was first described in vitro in cell cultures, where it explains limited cell proliferation: a limit called the Hayflick limit (HAYFLICK & MOORHEAD 1961; Olovnikov 1996). This theory was further validated by the fact that telomerase expression is sufficient to confer immortality to cells without causing oncogenic transformations (Bodnar et al. 1998). It is of particular interest that telomere shortening was observed during normal ageing in humans and mice (Blasco 2007).

Telomeric repeats are identical between chromosomes and homologous recombination is blocked by a protein complex called Shelterin that avoids inter- chromosomal fusion at the telomere level (Palm & de Lange 2008). Because DNA repair is difficult in these regions, damage is very persistent, inducing senescence and/or apoptosis (Fumagalli et al. 2012; Hewitt et al. 2012). We performed a γ-H2AX ChIP experiment 6 hours after 6-OHDA injection in the SNpc to see where DNA breaks appear and confirmed that telomeres are very sensitive, as we found more telomeres bound to γ-H2AX on the 6- OHDA injected side (ipsi) compared to the contralateral side (contra). This was also the case,

115 albeit slightly less, for LINE-1 sequences (Figure 9.). As a control, ChIP was also performed with an IgG antibody, to measure the background noise.

ChIP γ-H2AX - 6-OHDA 6h 20 Ipsi_γ-H2Ax Contra_γ-H2Ax Ipsi_IgG Contra_IgG 15

10 Figure 9. γ-H2AX Chromatin

% input immunoprecipitation after 6-OHDA injection

5 in the SNpc

0

Telomere LINE-1 A LINE-1 Tf/Gf LINE-1 ORF2

To further study this problem, we sequenced the DNA obtained by this method and although still preliminary, the results suggest that DNA damage preferentially takes place in active genes. Indeed, nearly all regions bound to γ-H2AX are located in coding regions and when we analyzed the list using KEGG pathway, neuronal-specific categories were found. To validate this observation, we are repeating this experiment, using either midbrain primary neurons or fibroblast. We want to see if DNA damage is still found in genes, and if so, whether enriched gene categories depend on the cell type and thus, as we hypothesize, on a particular chromatin environment of affected genes.

Telomerase deficiency is associated with premature development of several diseases in humans, for example pulmonary fibrosis or aplastic anemia, all of them related to the loss of tissue regenerative capacities (Armanios et al. 2009). Moreover, mutations in the Shelterin pathway are associated with accelerated ageing, even in the presence of normal length telomeres (Martínez & Blasco 2010). In mice, shorter or longer telomeres are associated with reduced or increased lifespan, respectively (Armanios et al. 2009; Blasco et al. 1997; Herrera et al. 1999; Rudolph et al. 1999; Tomás-Loba et al. 2008). Similarly, shorter telomeres in humans have been associated with increased mortality risks, particularly in younger individuals (Boonekamp et al. 2013). Finally recent studies suggest that overexpression of telomerase in mice decreases ageing phenotypes, in both pathological and normal ageing (Jaskelioff et al. 2011; Bernardes de Jesus et al. 2012).

116 Cellular senescence, another hallmark of ageing, is described as a stable arrest of the cell cycle, coupled with stereotyped phenotypic changes (Campisi & d'Adda di Fagagna 2007; Collado et al. 2007; Kuilman et al. 2010). Senescence was first described in fibroblast cultures (HAYFLICK & MOORHEAD 1961), and in this model, it was a consequence of telomere shortening (Bodnar et al. 1998). However, other age-related triggers exist independently of telomeres: notably non-telomeric DNA damage and derepression of the INK4/ARF locus, both triggers increase during ageing and induce senescence (Collado et al. 2007). Furthermore, the percentage of senescent cells increases with age in the liver, skin and spleen but not in the heart, muscle and kidney (C. Wang et al. 2009), showing that this process is organ specific. Several pathways have been associated with this state, notably the p16INK4a/Rb and the p19ARF/p53 pathways (Serrano et al. 1997). A recent genome-wide association study has identified the INK4/ARF locus as the major locus for age-related diseases, including cardiovascular diseases, diabetes, glaucoma - for which the homeoprotein Otx2 has been suggested as a possible treatment (Torero Ibad et al. 2011) - and Alzheimer’s disease (Jeck et al. 2012).

Telomere shortening-associated senescence cannot be applied to our model, as neurons are post-mitotic and thus arrested in G0. Nevertheless, senescent-like phenotypes were described in the brain, linked to proteins like p53, p21Cdkn1a or Sirt1 (Herskovits & Guarente 2014; Jurk et al. 2012). Interestingly, p53 and p21 are increased after 6-OHDA injection in the SNpc, meaning that mDA neurons could enter a senescent-like state (Article 1 – Figure 4G). Furthermore, p53 signalling is modified after Engrailed infusion (Figure 10), although this could be due to other functions of p53, which is a regulator of a very large number of biological processes (Gonfloni et al. 2014).

p53 signalling pathway GOF - adjP=0.0126 12 0.06 10 8 0.04 6

4 0.02 value p Figure 10. Genes from the “p53 2 signalling pathway” modified 0 0.00 -2 after Engrailed infusion in the -4 -0.02 SNpc -6

Fold change Fold -8 -0.04 2 -10

Log -12 -0.06 Tsc2 Trp73 Pmaip Casp8

117 Another protein strongly associated with senescence is Pml (Ivanschitz et al. 2013), which, similarly to p53, is a regulator of a variety of processes, including oxidative stress (U. Sahin et al. 2014) and telomere maintenance (Osterwald et al. 2015; F. T. M. Chang et al. 2013). Similarly, Gadd45ß has been implicated with senescence control (Moskalev et al. 2012). Both proteins are strongly upregulated after Engrailed injection in the SNpc/6-OHDA mouse model (Article 1 – Figure 6D), indicating that they may control a senescent-like phenotype in these neurons. An interesting hypothesis is that they might block the observed cell cycle re-entry of stressed post-mitotic mDA neurons (Article 1 – Figure S4A-D), therefore preventing apoptosis. Engrailed-mediated control of neuronal senescence or cell cycle-mediated apoptosis, via Pml and Gadd45ß, is an attractive hypothesis, not contradictory with the fact that Engrailed protects against DNA damage, a senescence inducer.

B. Deregulated nutrient sensing and altered cellular communication

In mammals, the somatotrophic axis is composed of the growth hormone (GH) from the anterior pituitary and of its secondary mediator, the insulin-like growth factor (IGF-1), which is produced by different cell types (mainly hepatocytes) after stimulation by GH. IGF-1 and insulin elicit the same cellular pathway which informs cells of high glucose presence. This pathway is therefore called the insulin and IGF-1 signalling (IIS) pathway and is the most conserved ageing-controlling pathway in evolution. It controls many proteins including AKT, the FOXO family or the mTOR complex, all involved in ageing and conserved during evolution (Barzilai et al. 2012; Fontana et al. 2010; Kenyon 2010). Mutations at many levels of this pathway (from GH to mTOR or FOXO downstream effectors) have been linked with longevity (Barzilai et al. 2012; Fontana et al. 2010; Kenyon 2010), further implicating bioenergetics and trophic pathways in longevity. Consistent with an implication of nutrient sensing in ageing, calorie restriction (CR) increases lifespan and healthy ageing in all eukaryotes tested (Barzilai et al. 2012; Fontana et al. 2010; Kenyon 2010), showing the importance of this hallmark of ageing. In flies and worms this effect is mediated, at least partially, by the IIS pathway (Barzilai et al. 2012; Fontana et al. 2010; Kenyon 2010). Although we have not yet studied it in detail, the IIS pathway could be regulated by Engrailed, as after infusion with the protein, insulin signalling is modified in SNpc mDA neurons (Figure 11.).

118

Insulin signalling pathway

GOF - adjP=0.0004 12 0.06 10 8 0.04 6 Figure 11. Genes from

4 0.02 value p the “insulin signalling 2 0 0.00 pathway” modified -2 after Engrailed infusion -4 -0.02 in the SNpc -6

Fold change Fold -8 -0.04 2 -10

Log -12 -0.06 Irs4

Akt1 Tsc2 Ikbkb Phka2 Socs2 Prkab1 Rps6ka1

The exact effect of the IIS pathway on ageing is not simple to analyze, as on one side its constitutive inhibition increases lifetime, but on the other end, it is decreased in both normal and pathological ageing (Fontana et al. 2010; Schumacher et al. 2008). A possible explanation is that of a compensatory mechanism: during ageing, the cell tries to cope with systemic damage by decreasing the IIS pathway, thus minimizing cell growth and metabolism (Garinis et al. 2008). There are at least three other nutrient sensing pathways implicated in ageing: first the mTOR pathway which senses high-energy states with amino acids, second the 5' adenosine monophosphate-activated protein kinase (AMPK) and Sirtuin pathways which sense low energy states by detecting high AMP or NAD+ respectively (Houtkooper et al. 2010). We have already discussed the interest of Sirtuins in the anti-ageing process; importantly, both Sirt2 and Sirt7 are upregulated after Engrailed infusion (Article 1 – Figure S5) suggesting that Engrailed could regulate, directly or not, their expression.

Concerning the mTOR pathway, it is composed of two complexes: mTORC1 and mTORC2, which regulate essentially all aspects of the anabolic metabolism (Laplante & Sabatini 2012b). Genetic down-regulation of mTORC1 in flies, worms and yeasts extends longevity and attenuates the longevity benefits of CR, suggesting that mTORC1 inhibition phenocopies calorie restriction (Johnson et al. 2013). Interestingly, mTOR signalling is modified after Engrailed infusion (Figure 12)

119

mTOR signalling pathway GOF - adjP=0.0056 12 0.06 10 8 0.04 6

4 0.02 value p 2 Figure 12. Genes from the 0 0.00 “mTOR signalling pathway” -2 -4 -0.02 modified after Engrailed -6 infusion in the SNpc

Fold change Fold -8 -0.04 2 -10 Log -12 -0.06 Akt1 Tsc2 Rps6 Rps6ka1

Mice with decreased mTORC1 but unmodified mTORC2 levels live longer (Harrison et al. 2009; Lamming et al. 2012), similarly to mice deficient for S6K1, a main mTORC1 downstream target (Selman et al. 2009). Moreover, mTOR activity increases with age in mouse hypothalamic neurons and is implicated in age-related obesity, which can be reversed by rapamycin, a mTORC1 inhibitor (S.-B. Yang et al. 2012). One possible initiator of mTOR is the lysophosphatidic acid (LPA) pathway (Foster 2009), which may be modified in the En1+/- mouse via the DG kinases (Dgkk and Dgkq) (Figure 13). Interestingly Dgkq has been linked with PD in China (Y. P. Chen et al. 2013).

500 * wt En1+/- 400

300 *

200

% Control / Gapdh 100

0

Dgkq Dgkk

Figure 13. The LPA pathway and its potential link to the En1+/- mouse Taken from (Foster 2009)

120 Both mTOR and IIS pathways sense high-energy states and exacerbate ageing phenotypes; moreover, inhibition of these pathways, by rapamycin in the case of mTORC1, has anti-ageing properties. This correlation is mirrored by the two other nutrient sensing pathways Sirtuins and AMPK. Both sense low-energy states and induce catabolism rather than anabolism. Similarly, their upregulation favors healthy ageing. AMPK activation has many anti-ageing properties, including inhibition of mTORC1 (Alers et al. 2012). AMPK could also mediate metformin-induced extension of longevity in worms and mice (Anisimov et al. 2011; Mair et al. 2011; Onken & Driscoll 2010). Sirt1, whose anti ageing properties have already been described, can also deacetylate and activate PGC-1α (Rodgers et al. 2005). This protein will induce a complex response, including mitochondrial biogenesis, enhanced antioxidant defenses and improved lipid oxidation (Fernández-Marcos & Auwerx 2011). Finally, Sirt1 and AMPK engage in a positive feedback loop, which mediates a unified and potent anti-stress response (Price et al. 2012).

Nutrient sensing seems to have an important part downstream of Engrailed, since 3 out of 4 pathways (IIS, mTOR and Sirtuins) are modified in either gain or loss of function of Engrailed. Furthermore it is noteworthy that metabolism is the first category, both in terms of size and in significance, in KEGG pathway analysis of Engrailed gain of function (32 genes adjP = 5.10-6) and loss of function (Discussion - Figure 2). The fact that they have a very high metabolic rate could explain the selective vulnerability of mDA SNpc neurons in PD (Surmeier & Schumacker 2013), and could also explain why Engrailed, a potent regulator of metabolism, is a key survival factor for these neurons.

Altered cellular communication is another hallmark of ageing, showing that changes induced by ageing are not purely cell autonomous. Indeed, ageing also involves changes in endocrine, neuroendocrine and neuronal intercellular communication (Laplante & Sabatini 2012a; Rando & H. Y. Chang 2012; Russell & Kahn 2007; G. Zhang et al. 2013).

In mammals, ageing is accompanied by increased inflammation, which could be elicited by many different phenomena: accumulation of pro-inflammatory tissue damage, failure of the immune system, senescence, activation of NF-κB signalling or defective autophagy (Salminen et al. 2012). Consumption of aspirin, a well-known anti-inflammatory

121 molecule, increases lifespan in mouse and healthy ageing in humans (Strong et al. 2008; Rothwell et al. 2011). We have not studied inflammation in En1+/- mice, but the category does not appear in our RNA-seq analyses using either KEGG pathways (Figure 2), or Pathway studio (Article 1 - Figure 1B). However, inflammation-related genes TNFR2 and C1qA, B and C (Griffiths et al. 2009; Fischer & Maier 2015) are strongly suspected - but not yet convicted - Engrailed direct targets. To study this hypothesis, one could imagine looking at inflammatory marks in older En1+/- mice.

On the contrary, we have strong data implicating Engrailed as a regulator of the NF- κB pathway. After 6-OHDA injection in the SNpc, this pathway is activated (RNA-seq 6- OHDA vs sham injection) and even further activated by injection of Engrailed after 6-OHDA, probably via the direct induction of Gadd45ß (Article 1 – Figure 6A-D). It could seem contradictory that a massive stress and its antidote have similar effects. However, NF-κB is a multi-function complex and can have different roles in the two conditions, depending on which downstream pathway is activated. Another possible explanation is that NF-κB activation is a compensatory mechanism against stress induced by 6-OHDA and that Engrailed exacerbates this response, fostering neuron survival. Whichever explanation is right, Engrailed is a potential regulator of this pathway.

Deciphering the role of Engrailed as an anti-ageing protein in the context of NF-κB activity is not simple. Increased NF-κB is associated with ageing, as shown by global studies of the transcriptional landscapes of aged tissues (de Magalhães et al. 2009; J. S. Lee et al. 2012). Accordingly, the expression of NF-κB inhibitors in the mouse skin reverses the ageing phenotype and induces transcription patterns typical of young cells (Adler et al. 2007). Similar results have been obtained in different mouse models of ageing (Osorio et al. 2012; Tilstra et al. 2012). Further studies will therefore be necessary to decipher the meaning of the positive regulation of NF-κB by Engrailed, but one can imagine that normal ageing and neurodegeneration are different processes and that Engrailed may not activate the NF-κB pathway in non suffering cells. This hypothesis is suggested by the differences between the genes from the DNA damage, chromatin remodeling and apoptosis categories identified after Engrailed infusion in control mice (Article 1 – Figure S5) or injection in 6-OHDA treated mice (Article 1 – Figure 6A). Indeed, Engrailed does not activate NF-κB in control mice.

122 Another pathway related to inflammation is the mitogen-activated protein kinase (MAPK) pathway (Kyriakis & Avruch 2012). This pathway seems strongly related to Engrailed, as it is modified in both gain (Figure 14) and loss of function (Figure 15) of Engrailed, both with a very low p value. In response to pro-inflammatory cytokines, MAPK signalling controls cell proliferation, differentiation, senescence and apoptosis (Y. Sun et al. 2015), many of them of great interest for our study. Furthermore, this pathway could regulate autophagy as well, in association with JNK (Zhou et al. 2015), which we have seen regulated by Engrailed (Article 1 – Figure 6E-F). Finally, spermidine anti-ageing effect could also be mediated, at least partially, by this pathway (Minois 2014).

MAPK signalling pathway GOF - adjP=9.98e-06

12 0.06 10 Figure 14. Genes 8 0.04 from the “MAPK 6

4 0.02 value p signalling pathway” 2 modified after 0 0.00 -2 Engrailed infusion in -4 -0.02 -6 the SNpc

Fold change Fold -8 -0.04 2 -10

Log -12 -0.06 Elk4 Akt1

Rac2 Ikbkb Ptpn5 Mef2c Nfatc2 Tgfbr2 Pla2g6 Cacnb2 Map2k3 Cacna1a Rps6ka1 Pla2g12a

MAPK signalling pathway LOF - adjP=0.0024 12 0.06 10 8 0.04 6 4 0.02 Figure 15. Genes p value p 2 from the “MAPK 0 0.00 signalling pathway” -2 modified in the -4 -0.02 +/- -6 En1 mouse

Fold change Fold -8 -0.04 2 -10

Log -12 -0.06

Elk4 Myc Stk4 Rac3 Traf6 Arrb2 Tgfb3 Nfatc2 Pdgfra Pla2g3

123 Cellular communication is reminiscent of the homeoprotein ability to transfer between cells. Engrailed transfer has not yet been studied in the SNpc, but we intend to do so using a technology recently developed in the laboratory: secreted single chain antibodies (Figure 16). These antibodies are composed of the variable regions of the light and heavy antibody chains, connected by a linker and with a signal peptide for secretion in their N-terminus. A stop cassette flanked by two loxP sites upstream of the sequence allows their conditional expression and secretion. We shall cross these mice with the TH-CRE-ERT2 mouse line to prevent the transfer of Engrailed in and out of TH positive cells. These experiments will allow us to decipher the non-cell autonomous role of Engrailed in the ventral midbrain, particularly its ability to transfer information between mDA neurons. We will also use this technology to immunoprecipitate extracellular Engrailed and identify putative interactors, small ncRNAs for example (see discussion I-D). It is attractive to imagine that Engrailed treatment in the SNpc might restore some elements of intercellular communication.

Figure 16. The principle of single-chain antibodies

124 C. Stem cell exhaustion

One of the most obvious characteristics of ageing is the loss of tissue regenerative potential. For example hematopoiesis decreases with age, inducing a diminished production of adaptive immune cells. This process, called immunosenescence, is responsible for the increased incidence of anemia and myeloid cancers with age (Shaw et al. 2010). A similar consequence of stem cell decay was observed in most adult stem cells, including those in the mouse forebrain (Molofsky et al. 2006), bones (Gruber et al. 2006), or muscles (Conboy & Rando 2012). Studies in old hematopoietic mouse stem cells (HSCs) showed that they undergo fewer divisions than those from young mice (Rossi et al. 2007). This decreased cell cycle activity correlated with the accumulation of DNA damage (Rossi et al. 2007) and the overexpression of cell cycle inhibitor proteins like p16INK4a (Janzen et al. 2006). Similarly INK4a-/- HSCs exhibit better engraftment and cell cycle capacities than wt HSCs (Janzen et al. 2006). Finally, telomere shortening is also an important cause of stem cell decline (Flores et al. 2005; Sharpless & DePinho 2007). Inhibition of mTOR by rapamycin, which improves healthy ageing through proteostasis and nutrient sensing, may also improve stem cell function in the epidermis, the hematopoietic system and in the intestine (Castilho et al. 2009; C. Chen et al. 2009; Yilmaz et al. 2012). Stem cell exhaustion emerges as an integrative consequence of ageing as it is a consequence of many different hallmarks. Promising studies suggested recently that stem cell rejuvenation might reverse the ageing-induced changes at the organism level (Rando & H. Y. Chang 2012).

A possible link between stem cells and homeoproteins is suggested by recent observations made in the laboratory. Injecting Cre recombinase in the lateral ventricles of Otx2flox/flox mice deletes Otx2 specifically in the choroid plexus and decreases Otx2 concentration in the cerebrospinal fluid (Spatazza et al. 2013). In these mice, Anabelle Planques and her colleagues observed a 25% decrease in newborn neurons in the olfactory bulb and a decreased doublecortin (DCX) staining in the neurogenic sub ventricular zone. Although the mechanisms involved are still unclear, this result shows that Otx2 can control neuronal stem cell production and/or differentiation. Whether Engrailed may have similar properties remains an open question.

125 IV. General conclusion

The aim of this project was to study the other neuroprotective pathways of Engrailed, apart from the translation regulation of mitochondrial CI subunits (Alvarez-Fischer et al. 2011). We found that Engrailed is linked to two additional key hallmarks of ageing, namely DNA damage and chromatin remodeling both related to the physiopathology of Parkinson disease (Figure 17).

Figure 17. Three main hallmarks of ageing strongly regulated by Engrailed

In the review on ageing by (López-Otín et al. 2013), genomic instability and epigenetic alterations are classified as primary hallmarks, meaning that they constitute the first steps of the ageing pathway. Mitochondrial dysfunction, on the other hand, is classified as an antagonistic hallmark, meaning that it has opposite effects depending on its intensity. However, this ambivalent nature fits oxidative stress better, and mitochondrial dysfunction by itself, could easily be considered as a primary hallmark. With this new categorization, Engrailed protection against ageing seem to pass mainly through the regulation of primary hallmarks. Furthermore, Engrailed regulates all of them with the exception of telomere attrition, whose role in post mitotic neurons is still debated (Eitan et al. 2014). This shows that

126 Engrailed may protect against the initial causes of ageing, and therefore could be a very potent anti-ageing treatment with an epigenetic component allowing long-lasting positive effects. To further validate this hypothesis, Engrailed, or other homeoproteins, should be tested in other age related diseases to see if they are as efficient. For example, the laboratory is starting a new project to study Engrailed potential therapeutic effects in Huntington disease. Furthermore, Otx2 protection has already been shown in glaucoma mouse models (Torero Ibad et al. 2011).

We have also addressed one of the major pending questions of the first article: the mechanism by which Engrailed protects against DNA strand breaks. We propose that LINEs, which are directly repressed by Engrailed, are a potential link between genomic instability and epigenetic alterations. Taken together, we have unraveled a “vicious circle” between chromatin relaxation, LINE-1 activation and DNA damage, which could participate in PD- neurodegeneration. Engrailed also seems a strong repressor of this “vicious circle”, and this on several levels (Figure17). To put this into perspective, one has to remember that LINEs, which were considered as junk DNA 20 years ago, are now being implicated in the origin or progression of different diseases, some related to ageing: cancer, Alzheimer’s disease, Rett syndrome, ataxia telangiectasia and schizophrenia (see introduction). Engrailed transcriptional repression of LINEs, shown in the second part of this thesis, should encourage people to test whether Engrailed may not have therapeutic effects in these diseases as well. A lot of questions remain unanswered; however the preliminary responses we obtained should definitively encourage further research in the field of homeoprotein anti-ageing functions.

127 BIBLIOGRAPHY

Abeliovich, A. et al., 2000. Mice lacking alpha-synuclein display functional deficits in the nigrostriatal dopamine system. Neuron, 25(1), pp.239–252.

Ade, C., Roy-Engel, A.M. & Deininger, P.L., 2013. Alu elements: an intrinsic source of human genome instability. Current opinion in virology, 3(6), pp.639–645.

Adler, A.S. et al., 2007. Motif module map reveals enforcement of aging by continual NF- kappaB activity. Genes & Development, 21(24), pp.3244–3257.

Alavez, S. et al., 2011. Amyloid-binding compounds maintain protein homeostasis during ageing and extend lifespan. Nature, 472(7342), pp.226–229.

Albéri, L., Sgadò, P. & Simon, H.H., 2004. Engrailed genes are cell-autonomously required to prevent apoptosis in mesencephalic dopaminergic neurons. Development (Cambridge, England), 131(13), pp.3229–3236.

Alegre-Abarrategui, J. & Wade-Martins, R., 2009. Parkinson disease, LRRK2 and the endocytic-autophagic pathway. Autophagy, 5(8), pp.1208–1210.

Alers, S. et al., 2012. Role of AMPK-mTOR-Ulk1/2 in the regulation of autophagy: cross talk, shortcuts, and feedbacks. Molecular and cellular biology, 32(1), pp.2–11.

Allen, M.T. & Levy, L.S., 2013. Parkinson's disease and pesticide exposure--a new assessment. Critical reviews in toxicology, 43(6), pp.515–534.

Alvarez-Fischer, D. et al., 2011. Engrailed protects mouse midbrain dopaminergic neurons against mitochondrial complex I insults. Nature Neuroscience, 14(10), pp.1260–1266.

Amir, R.E. et al., 1999. Rett syndrome is caused by mutations in X-linked MECP2, encoding methyl-CpG-binding protein 2. Nature Genetics, 23(2), pp.185–188.

Anders, S. & Huber, W., 2010. Differential expression analysis for sequence count data. Genome biology, 11(10), p.R106.

Anisimov, V.N. et al., 2011. If started early in life, metformin treatment increases life span and postpones tumors in female SHR mice. Aging, 3(2), pp.148–157.

Armanios, M. et al., 2009. Short telomeres are sufficient to cause the degenerative defects associated with aging. American journal of human genetics, 85(6), pp.823–832.

Asagoshi, K. et al., 2010. FEN1 functions in long patch base excision repair under conditions of oxidative stress in vertebrate cells. Molecular cancer research : MCR, 8(2), pp.204– 215.

Athanikar, J.N., Badge, R.M. & Moran, J.V., 2004. A YY1-binding site is required for accurate human LINE-1 transcription initiation. Nucleic Acids Research, 32(13), pp.3846–3855.

128 Azevedo, F.A.C. et al., 2009. Equal numbers of neuronal and nonneuronal cells make the human brain an isometrically scaled-up primate brain. Journal of Comparative Neurology, 513(5), pp.532–541.

Bahar, R. et al., 2006. Increased cell-to-cell variation in gene expression in ageing mouse heart. Nature, 441(7096), pp.1011–1014.

Baker, D.J. et al., 2013. Increased expression of BubR1 protects against aneuploidy and cancer and extends healthy lifespan. Nature cell biology, 15(1), pp.96–102.

Bandyopadhyay, U., Bandhyopadhyay, U. & Cuervo, A.M., 2007. Chaperone-mediated autophagy in aging and neurodegeneration: lessons from alpha-synuclein. Experimental gerontology, 42(1-2), pp.120–128.

Banerjee, R., Beal, M.F. & Thomas, B., 2010. Autophagy in neurodegenerative disorders: pathogenic roles and therapeutic implications. Trends in neurosciences, 33(12), pp.541– 549.

Bar-Shira, A. et al., 2002. ATM-dependent activation of the gene encoding MAP kinase phosphatase 5 by radiomimetic DNA damage. Oncogene, 21(5), pp.849–855.

Bardine, N. et al., 2014. Vertical signalling involves transmission of Hox information from gastrula mesoderm to neurectoderm. PloS one, 9(12), p.e115208.

Barzilai, N. et al., 2012. The critical role of metabolic pathways in aging. Diabetes, 61(6), pp.1315–1322.

Bastianetto, S., Ménard, C. & Quirion, R., 2015. Neuroprotective action of resveratrol. Biochimica et biophysica acta, 1852(6), pp.1195–1201.

Bártová, E. et al., 2008. Histone modifications and nuclear architecture: a review. The journal of histochemistry and cytochemistry : official journal of the Histochemistry Society, 56(8), pp.711–721.

Beck, C.R. et al., 2011. LINE-1 elements in structural variation and disease. Annual review of genomics and human genetics, 12, pp.187–215.

Beilina, A. et al., 2005. Mutations in PTEN-induced putative kinase 1 associated with recessive parkinsonism have differential effects on protein stability. Proceedings of the National Academy of Sciences of the United States of America, 102(16), pp.5703–5708.

Beisel, C. & Paro, R., 2011. Silencing chromatin: comparing modes and mechanisms. Nature Reviews Genetics, 12(2), pp.123–135.

Beltran, A.S., Graves, L.M. & Blancafort, P., 2014. Novel role of Engrailed 1 as a prosurvival transcription factor in basal-like breast cancer and engineering of interference peptides block its oncogenic function. Oncogene, 33(39), pp.4767–4777.

Bendor, J.T., Logan, T.P. & Edwards, R.H., 2013. The Function of a-Synuclein. Neuron, 79(6), pp.1044–1066.

Bernardes de Jesus, B. et al., 2012. Telomerase gene therapy in adult and old mice delays

129 aging and increases longevity without increasing cancer. EMBO molecular medicine, 4(8), pp.691–704.

Bernstein, A.I. et al., 2011. 6-OHDA generated ROS induces DNA damage and p53- and PUMA-dependent cell death. Molecular neurodegeneration, 6(1), p.2.

Betarbet, R. et al., 2000. Chronic systemic pesticide exposure reproduces features of Parkinson's disease. Nature Neuroscience, 3(12), pp.1301–1306.

Biagi, M. & Bertelli, A.A.E., 2015. Wine, alcohol and pills: What future for the French paradox? Life sciences, 131, pp.19–22.

Bianco, Lo, C. et al., 2002. alpha -Synucleinopathy and selective dopaminergic neuron loss in a rat lentiviral-based model of Parkinson's disease. Proceedings of the National Academy of Sciences of the United States of America, 99(16), pp.10813–10818.

Bjedov, I. et al., 2010. Mechanisms of life span extension by rapamycin in the fruit fly Drosophila melanogaster. Cell metabolism, 11(1), pp.35–46.

Blackburn, E.H., Greider, C.W. & Szostak, J.W., 2006. Telomeres and telomerase: the path from maize, Tetrahymena and yeast to human cancer and aging. Nature medicine, 12(10), pp.1133–1138.

Blagosklonny, M.V., 2011. Rapamycin-induced glucose intolerance: hunger or starvation diabetes. Cell cycle (Georgetown, Tex.), 10(24), pp.4217–4224.

Blasco, M.A., 2007. Telomere length, stem cells and aging. Nature chemical biology, 3(10), pp.640–649.

Blasco, M.A. et al., 1997. Telomere shortening and tumor formation by mouse cells lacking telomerase RNA. Cell, 91(1), pp.25–34.

Bodnar, A.G. et al., 1998. Extension of life-span by introduction of telomerase into normal human cells. Science (New York, N.Y.), 279(5349), pp.349–352.

Bonifati, V. et al., 2003. Mutations in the DJ-1 gene associated with autosomal recessive early-onset parkinsonism. Science (New York, N.Y.), 299(5604), pp.256–259.

Boonekamp, J.J. et al., 2013. Telomere length behaves as biomarker of somatic redundancy rather than biological age. Aging Cell, 12(2), pp.330–332.

Boulon, S. et al., 2010. The nucleolus under stress. Molecular Cell, 40(2), pp.216–227.

Brennecke, J. et al., 2007. Discrete small RNA-generating loci as master regulators of transposon activity in Drosophila. Cell, 128(6), pp.1089–1103.

Brunet, I. et al., 2005. The transcription factor Engrailed-2 guides retinal axons. Nature, 438(7064), pp.94–98.

Buhlman, L. et al., 2014. Functional interplay between Parkin and Drp1 in mitochondrial fission and clearance. Biochimica et biophysica acta, 1843(9), pp.2012–2026.

130 Bulut-Karslioglu, A. et al., 2012. A transcription factor-based mechanism for mouse heterochromatin formation. Nature structural & molecular biology, 19(10), pp.1023– 1030.

Bundo, M. et al., 2014. Increased L1 Retrotranspositionin the Neuronal Genome in Schizophrenia. Neuron, 81(2), pp.306–313.

Burikhanov, R. et al., 2014. Paracrine apoptotic effect of p53 mediated by tumor suppressor Par-4. Cell reports, 6(2), pp.271–277.

Burnett, C. et al., 2011. Absence of effects of Sir2 overexpression on lifespan in C. elegans and Drosophila. Nature, 477(7365), pp.482–485.

Burtner, C.R. & Kennedy, B.K., 2010. Progeria syndromes and ageing: what is the connection? Nature reviews. Molecular cell biology, 11(8), pp.567–578.

Cabanillas, R. et al., 2011. Néstor-Guillermo progeria syndrome: a novel premature aging condition with early onset and chronic development caused by BANF1 mutations. American journal of medical genetics. Part A, 155A(11), pp.2617–2625.

Calabrese, V. et al., 2011. Hormesis, cellular stress response and vitagenes as critical determinants in aging and longevity. Molecular aspects of medicine, 32(4-6), pp.279– 304.

Calamini, B. & Morimoto, R.I., 2012. Protein homeostasis as a therapeutic target for diseases of protein conformation. Current topics in medicinal chemistry, 12(22), pp.2623–2640.

Campisi, J. & d'Adda di Fagagna, F., 2007. Cellular senescence: when bad things happen to good cells. Nature reviews. Molecular cell biology, 8(9), pp.729–740.

Casafont, I. et al., 2011. Effect of ionizing radiation in sensory ganglion neurons: organization and dynamics of nuclear compartments of DNA damage/repair and their relationship with transcription and cell cycle. Acta Neuropathologica, 122(4), pp.481–493.

Castilho, R.M. et al., 2009. mTOR mediates Wnt-induced epidermal stem cell exhaustion and aging. Cell stem cell, 5(3), pp.279–289.

Chahrour, M. & Zoghbi, H.Y., 2007. The story of Rett syndrome: from clinic to neurobiology. Neuron, 56(3), pp.422–437.

Chan, F.L. & Wong, L.H., 2012. Transcription in the maintenance of centromere chromatin identity. Nucleic Acids Research, 40(22), pp.11178–11188.

Chan, P. et al., 1998. Failure to find the alpha-synuclein gene missense mutation (G209A) in 100 patients with younger onset Parkinson's disease. Neurology, 50(2), pp.513–514.

Chandrasekar, V. & Dreyer, J.-L., 2009. microRNAs miR-124, let-7d and miR-181a regulate cocaine-induced plasticity. Molecular and cellular neurosciences, 42(4), pp.350–362.

Chang, F.T.M. et al., 2013. PML bodies provide an important platform for the maintenance of telomeric chromatin integrity in embryonic stem cells. Nucleic Acids Research, 41(8), pp.4447–4458.

131 Chen, C. et al., 2009. mTOR regulation and therapeutic rejuvenation of aging hematopoietic stem cells. Science Signaling, 2(98), p.ra75.

Chen, C.-M. et al., 2009. Increased oxidative damage in peripheral blood correlates with severity of Parkinson's disease. Neurobiology of Disease, 33(3), pp.429–435.

Chen, X., 2005. MicroRNA biogenesis and function in plants. FEBS letters, 579(26), pp.5923–5931.

Chen, Y.P. et al., 2013. GAK rs1564282 and DGKQ rs11248060 increase the risk for Parkinson's disease in a Chinese population. Journal of clinical neuroscience : official journal of the Neurosurgical Society of Australasia, 20(6), pp.880–883.

Cheng, H.-C., Ulane, C.M. & Burke, R.E., 2010. Clinical progression in Parkinson disease and the neurobiology of axons. Annals of neurology, 67(6), pp.715–725.

Chiang, W.-C. et al., 2012. HSF-1 regulators DDL-1/2 link insulin-like signaling to heat- shock responses and modulation of longevity. Cell, 148(1-2), pp.322–334.

Chin-Chan, M., Navarro-Yepes, J. & Quintanilla-Vega, B., 2015. Environmental pollutants as risk factors for neurodegenerative disorders: Alzheimer and Parkinson diseases. Frontiers in cellular neuroscience, 9, p.124.

Choi, S. & Bakkenist, C.J., 2013. Brd4 Shields Chromatin from ATM Kinase Signaling Storms. Science Signaling, 6(293), pp.pe30–pe30.

Chow, J.C. et al., 2010. LINE-1 activity in facultative heterochromatin formation during X chromosome inactivation. Cell, 141(6), pp.956–969.

Ciccone, S. et al., 2013. Parkinson’s Disease: A Complex Interplay of Mitochondrial DNA Alterations and Oxidative Stress. International Journal of Molecular Sciences, 14(2), pp.2388–2409.

Coffey, E.T., 2014. Nuclear and cytosolic JNK signalling in neurons. Nature Reviews Neuroscience, 15(5), pp.285–299.

Collado, M., Blasco, M.A. & Serrano, M., 2007. Cellular senescence in cancer and aging. Cell, 130(2), pp.223–233.

Conboy, I.M. & Rando, T.A., 2012. Heterochronic parabiosis for the study of the effects of aging on stem cells and their niches. Cell cycle (Georgetown, Tex.), 11(12), pp.2260– 2267.

Cong, R. et al., 2012. Interaction of nucleolin with ribosomal RNA genes and its role in RNA polymerase I transcription. Nucleic Acids Research, 40(19), pp.9441–9454.

Coufal, N.G. et al., 2011. Ataxia telangiectasia mutated (ATM) modulates long interspersed element-1 (L1) retrotransposition in human neural stem cells. Proceedings of the National Academy of Sciences of the United States of America, 108(51), pp.20382–20387.

Coufal, N.G. et al., 2009. L1 retrotransposition in human neural progenitor cells. Nature, 460(7259), pp.1127–1131.

132 Dalby, B. et al., 2004. Advanced transfection with Lipofectamine 2000 reagent: primary neurons, siRNA, and high-throughput applications. Methods (San Diego, Calif.), 33(2), pp.95–103.

Dauer, W. et al., 2003. Parkinson’s Disease: Mechanisms and Models. Neuron, pp.1–21.

Davis, C.A. & Joyner, A.L., 1988. Expression patterns of the homeo box-containing genes En-1 and En-2 and the proto-oncogene int-1 diverge during mouse development. Genes & Development, 2(12B), pp.1736–1744. de Magalhães, J.P., Curado, J. & Church, G.M., 2009. Meta-analysis of age-related gene expression profiles identifies common signatures of aging. Bioinformatics (Oxford, England), 25(7), pp.875–881.

De Sandre-Giovannoli, A. et al., 2003. Lamin a truncation in Hutchinson-Gilford progeria. Science (New York, N.Y.), 300(5628), p.2055.

De Smaele, E. et al., 2001. Induction of gadd45beta by NF-kappaB downregulates pro- apoptotic JNK signalling. Nature, 414(6861), pp.308–313.

Dechat, T. et al., 2008. Nuclear lamins: major factors in the structural organization and function of the nucleus and chromatin. Genes & Development, 22(7), pp.832–853.

Deng, J.V. et al., 2010. MeCP2 in the nucleus accumbens contributes to neural and behavioral responses to psychostimulants. Nature Neuroscience, 13(9), pp.1128–1136.

Derelle, R. et al., 2007. Homeodomain proteins belong to the ancestral molecular toolkit of eukaryotes. Evolution & development, 9(3), pp.212–219.

Dewannieux, M., Esnault, C. & Heidmann, T., 2003. LINE-mediated retrotransposition of marked Alu sequences. Nature Genetics, 35(1), pp.41–48.

Di Lullo, E. et al., 2011. Paracrine Pax6 activity regulates oligodendrocyte precursor cell migration in the chick embryonic neural tube. Development (Cambridge, England), 138(22), pp.4991–5001.

Di Salvio, M., Di Giovannantonio, L.G., Acampora, D., et al., 2010. Otx2 controls neuron subtype identity in ventral tegmental area and antagonizes vulnerability to MPTP. Nature Neuroscience, 13(12), pp.1481–1488.

Di Salvio, M., Di Giovannantonio, L.G., Omodei, D., et al., 2010. Otx2 expression is restricted to dopaminergic neurons of the ventral tegmental area in the adult brain. The International journal of developmental biology, 54(5), pp.939–945.

Dickson, D.W. et al., 2010. Evidence in favor of Braak staging of Parkinson's disease. Movement disorders : official journal of the Movement Disorder Society, 25 Suppl 1, pp.S78–82.

Donzelli, S. et al., 2012. MicroRNA-128-2 targets the transcriptional repressor E2F5 enhancing mutant p53 gain of function. Cell death and differentiation, 19(6), pp.1038– 1048.

133 Doonan, R. et al., 2008. Against the oxidative damage theory of aging: superoxide dismutases protect against oxidative stress but have little or no effect on life span in Caenorhabditis elegans. Genes & Development, 22(23), pp.3236–3241.

Doty, R.L., 2012. Olfactory dysfunction in Parkinson disease. Nature Publishing Group, 8(6), pp.329–339.

Dubinsky, A.N. et al., 2014. Let-7 coordinately suppresses components of the amino acid sensing pathway to repress mTORC1 and induce autophagy. Cell metabolism, 20(4), pp.626–638.

Durieux, J., Wolff, S. & Dillin, A., 2011. The cell-non-autonomous nature of electron transport chain-mediated longevity. Cell, 144(1), pp.79–91.

Edgar, D. et al., 2009. Random point mutations with major effects on protein-coding genes are the driving force behind premature aging in mtDNA mutator mice. Cell metabolism, 10(2), pp.131–138.

Ehrlich, M., 2002. DNA methylation in cancer: too much, but also too little. Oncogene, 21(35), pp.5400–5413.

Eisenberg, T. et al., 2009. Induction of autophagy by spermidine promotes longevity. Nature cell biology, 11(11), pp.1305–1314.

Eitan, E., Hutchison, E.R. & Mattson, M.P., 2014. Telomere shortening in neurological disorders: an abundance of unanswered questions. Trends in neurosciences, 37(5), pp.256–263.

Elsässer, S.J. et al., 2015. Histone H3.3 is required for endogenous retroviral element silencing in embryonic stem cells. Nature.

Eriksson, M. et al., 2003. Recurrent de novo point mutations in lamin A cause Hutchinson- Gilford progeria syndrome. Nature, 423(6937), pp.293–298.

Eriksson, P.S. et al., 1998. Neurogenesis in the adult human hippocampus. Nature medicine, 4(11), pp.1313–1317.

Erwin, J.A., Marchetto, M.C. & Gage, F.H., 2014. Mobile DNA elements in the generation of diversity and complexity in the brain. Nature Reviews Neuroscience, 15(8), pp.497–506.

Esnault, C., Maestre, J. & Heidmann, T., 2000. Human LINE retrotransposons generate processed pseudogenes. Nature Genetics, 24(4), pp.363–367.

Fadloun, A. et al., 2013. Chromatin signatures and retrotransposon profiling in mouse embryos reveal regulation of LINE-1 by RNA. Nature structural & molecular biology, 20(3), pp.332–338.

Faggioli, F. et al., 2012. Chromosome-specific accumulation of aneuploidy in the aging mouse brain. Human molecular genetics, 21(24), pp.5246–5253.

Faulkner, G.J. et al., 2009. The regulated retrotransposon transcriptome of mammalian cells. Nature Genetics, 41(5), pp.563–571.

134 Feng, Q. et al., 1996. Human L1 retrotransposon encodes a conserved endonuclease required for retrotransposition. Cell, 87(5), pp.905–916.

Fernández-Marcos, P.J. & Auwerx, J., 2011. Regulation of PGC-1α, a nodal regulator of mitochondrial biogenesis. The American journal of clinical nutrition, 93(4), pp.884S–90.

Ferrières, J., 2004. The French paradox: lessons for other countries. Heart (British Cardiac Society), 90(1), pp.107–111.

Feschotte, C., 2008. Transposable elements and the evolution of regulatory networks. Nature Reviews Genetics, 9(5), pp.397–405.

Fischer, R. & Maier, O., 2015. Interrelation of oxidative stress and inflammation in neurodegenerative disease: role of TNF. Oxidative medicine and cellular longevity, 2015, p.610813.

Flores, I., Cayuela, M.L. & Blasco, M.A., 2005. Effects of telomerase and telomere length on epidermal stem cell behavior. Science (New York, N.Y.), 309(5738), pp.1253–1256.

Fontana, L., Partridge, L. & Longo, V.D., 2010. Extending healthy life span--from yeast to humans. Science (New York, N.Y.), 328(5976), pp.321–326.

Fornai, F. et al., 2005. Parkinson-like syndrome induced by continuous MPTP infusion: convergent roles of the ubiquitin-proteasome system and alpha-synuclein. Proceedings of the National Academy of Sciences of the United States of America, 102(9), pp.3413–3418.

Forsberg, L.A. et al., 2012. Age-related somatic structural changes in the nuclear genome of human blood cells. American journal of human genetics, 90(2), pp.217–228.

Foster, D.A., 2009. Biochimica et Biophysica Acta. BBA - Molecular and Cell Biology of Lipids, 1791(9), pp.949–955.

Fraga, M.F. & Esteller, M., 2007. Epigenetics and aging: the targets and the marks. Trends in genetics : TIG, 23(8), pp.413–418.

Freeman, J.D., Goodchild, N.L. & Mager, D.L., 1994. A modified indicator gene for selection of retrotransposition events in mammalian cells. BioTechniques, 17(1), pp.46–48–9– 52.

Freije, J.M.P. & López-Otín, C., 2012. Reprogramming aging and progeria. Current opinion in cell biology, 24(6), pp.757–764.

Frost, B. et al., 2014. Tau promotes neurodegeneration through global chromatin relaxation. Nature Neuroscience, 17(3), pp.357–366.

Fuchs, J. et al., 2009. The transcription factor PITX3 is associated with sporadic Parkinson's disease. Neurobiology of aging, 30(5), pp.731–738.

Fumagalli, M. et al., 2012. Telomeric DNA damage is irreparable and causes persistent DNA- damage-response activation. Nature cell biology, 14(4), pp.355–365.

Gai, W.P. et al., 1992. Age-related loss of dorsal vagal neurons in Parkinson's disease. Neurology, 42(11), pp.2106–2111.

135 Gantz, S.C. et al., 2011. Loss of Mecp2 in substantia nigra dopamine neurons compromises the nigrostriatal pathway. Journal of Neuroscience, 31(35), pp.12629–12637.

Garcia-Perez, J.L. et al., 2010. Epigenetic silencing of engineered L1 retrotransposition events in human embryonic carcinoma cells. Nature, 466(7307), pp.769–773.

Garg, S.K. et al., 2013. Systemic delivery of MeCP2 rescues behavioral and cellular deficits in female mouse models of Rett syndrome. Journal of Neuroscience, 33(34), pp.13612– 13620.

Garinis, G.A. et al., 2008. DNA damage and ageing: new-age ideas for an age-old problem. Nature cell biology, 10(11), pp.1241–1247.

Gasior, S.L. et al., 2006. The human LINE-1 retrotransposon creates DNA double-strand breaks. Journal of molecular biology, 357(5), pp.1383–1393.

Gauron, C. et al., 2013. Sustained production of ROS triggers compensatory proliferation and is required for regeneration to proceed. Scientific reports, 3, p.2084.

Gehrig, S.M. et al., 2012. Hsp72 preserves muscle function and slows progression of severe muscular dystrophy. Nature, 484(7394), pp.394–398.

Gehring, W., 1967. Clonal analysis of determination dynamics in cultures of imaginal disks in Drosophila melanogaster. Developmental biology, 16(5), pp.438–456.

Gehring, W.J., 1987. Homeo boxes in the study of development. Science (New York, N.Y.), 236(4806), pp.1245–1252.

Giasson, B.I. et al., 2002. Neuronal alpha-synucleinopathy with severe movement disorder in mice expressing A53T human alpha-synuclein. Neuron, 34(4), pp.521–533.

Giasson, B.I. et al., 2000. Oxidative damage linked to neurodegeneration by selective alpha- synuclein nitration in synucleinopathy lesions. Science (New York, N.Y.), 290(5493), pp.985–989.

Gilks, W.P. et al., 2005. A common LRRK2 mutation in idiopathic Parkinson's disease. Lancet, 365(9457), pp.415–416.

Giorgi, G., Marcantonio, P. & Del Re, B., 2011. LINE-1 retrotransposition in human neuroblastoma cells is affected by oxidative stress. Cell and tissue research, 346(3), pp.383–391.

Giralt, A. & Villarroya, F., 2012. SIRT3, a pivotal actor in mitochondrial functions: metabolism, cell death and aging. The Biochemical journal, 444(1), pp.1–10.

Goedert, M., 2001. Alpha-synuclein and neurodegenerative diseases. Nature Reviews Neuroscience, 2(7), pp.492–501.

Gonfloni, S. et al., 2014. P53 and Sirt1: routes of metabolism and genome stability. Biochemical pharmacology, 92(1), pp.149–156.

Gonzalez-Suarez, I. et al., 2009. Novel roles for A-type lamins in telomere biology and the

136 DNA damage response pathway. The EMBO journal, 28(16), pp.2414–2427.

Green, D.R., Galluzzi, L. & Kroemer, G., 2011. Mitochondria and the autophagy- inflammation-cell death axis in organismal aging. Science (New York, N.Y.), 333(6046), pp.1109–1112.

Greenamyre, J.T. et al., 2001. Complex I and Parkinson's disease. IUBMB life, 52(3-5), pp.135–141.

Greer, E.L. et al., 2010. Members of the H3K4 trimethylation complex regulate lifespan in a germline-dependent manner in C. elegans. Nature, 466(7304), pp.383–387.

Gregg, S.Q. et al., 2012. A mouse model of accelerated liver aging caused by a defect in DNA repair. Hepatology (Baltimore, Md.), 55(2), pp.609–621.

Griffiths, M.R., Gasque, P. & Neal, J.W., 2009. The multiple roles of the innate immune system in the regulation of apoptosis and inflammation in the brain. Journal of neuropathology and experimental neurology, 68(3), pp.217–226.

Gruber, R. et al., 2006. Fracture healing in the elderly patient. Experimental gerontology, 41(11), pp.1080–1093.

Guarente, L., 2011. Sirtuins, aging, and metabolism. Cold Spring Harbor symposia on quantitative biology, 76, pp.81–90.

Guenatri, M. et al., 2004. Mouse centric and pericentric satellite repeats form distinct functional heterochromatin. The Journal of cell biology, 166(4), pp.493–505.

Guetg, C. & Santoro, R., 2012. Formation of nuclear heterochromatin: the nucleolar point of view. Epigenetics : official journal of the DNA Methylation Society, 7(8), pp.811–814.

Guetg, C. et al., 2010. The NoRC complex mediates the heterochromatin formation and stability of silent rRNA genes and centromeric repeats. The EMBO journal, 29(13), pp.2135–2146.

Gunawardane, L.S. et al., 2007. A slicer-mediated mechanism for repeat-associated siRNA 5' end formation in Drosophila. Science (New York, N.Y.), 315(5818), pp.1587–1590.

Haigis, M.C. & Yankner, B.A., 2010. The aging stress response. Molecular Cell, 40(2), pp.333–344.

Han, K. et al., 2008. L1 recombination-associated deletions generate human genomic variation. Proceedings of the National Academy of Sciences of the United States of America, 105(49), pp.19366–19371.

Han, S. & Brunet, A., 2012. Histone methylation makes its mark on longevity. Trends in cell biology, 22(1), pp.42–49.

Hanahan, D. & Weinberg, R.A., 2011. Hallmarks of cancer: the next generation. Cell, 144(5), pp.646–674.

Hanahan, D. & Weinberg, R.A., 2000. The hallmarks of cancer. Cell, 100(1), pp.57–70.

137 Hancks, D.C. & Kazazian, H.H., 2012. Active human retrotransposons: variation and disease. Current opinion in genetics & development, 22(3), pp.191–203.

Hancks, D.C. et al., 2012. The minimal active human SVA retrotransposon requires only the 5'-hexamer and Alu-like domains. Molecular and cellular biology, 32(22), pp.4718–4726.

Hanks, M. et al., 1995. Rescue of the En-1 mutant phenotype by replacement of En-1 with En-2. Science (New York, N.Y.), 269(5224), pp.679–682.

HARMAN, D., 1965. THE FREE RADICAL THEORY OF AGING: EFFECT OF AGE ON SERUM COPPER LEVELS. Journal of gerontology, 20, pp.151–153.

Harries, L.W. et al., 2011. Human aging is characterized by focused changes in gene expression and deregulation of alternative splicing. Aging Cell, 10(5), pp.868–878.

Harrison, D.E. et al., 2009. Rapamycin fed late in life extends lifespan in genetically heterogeneous mice. Nature, 460(7253), pp.392–395.

Hartl, F.U., Bracher, A. & Hayer-Hartl, M., 2011. Molecular chaperones in protein folding and proteostasis. Nature, 475(7356), pp.324–332.

Haubenberger, D. et al., 2011. Association of transcription factor polymorphisms PITX3 and EN1 with Parkinson's disease. Neurobiology of aging, 32(2), pp.302–307.

Hawkes, C.H., Del Tredici, K. & Braak, H., 2009. Parkinson's disease: the dual hit theory revisited. Annals of the New York Academy of Sciences, 1170, pp.615–622.

Hawley, S.A. et al., 2010. Use of cells expressing gamma subunit variants to identify diverse mechanisms of AMPK activation. Cell metabolism, 11(6), pp.554–565.

HAYFLICK, L. & MOORHEAD, P.S., 1961. The serial cultivation of human diploid cell strains. Experimental cell research, 25, pp.585–621.

Hekimi, S., Lapointe, J. & Wen, Y., 2011. Taking a “good” look at free radicals in the aging process. Trends in cell biology, 21(10), pp.569–576.

Hernán, M.A. et al., 2002. A meta-analysis of coffee drinking, cigarette smoking, and the risk of Parkinson's disease. Annals of neurology, 52(3), pp.276–284.

Herranz, D. et al., 2010. Sirt1 improves healthy ageing and protects from metabolic syndrome-associated cancer. Nature Communications, 1, p.3.

Herrera, E. et al., 1999. Disease states associated with telomerase deficiency appear earlier in mice with short telomeres. The EMBO journal, 18(11), pp.2950–2960.

Herskovits, A.Z. & Guarente, L., 2014. SIRT1 in neurodevelopment and brain senescence. Neuron, 81(3), pp.471–483.

Herva, M.E. & Spillantini, M.G., 2014. Parkinson's disease as a member of Prion-like disorders. Virus research.

Hewitt, G. et al., 2012. Telomeres are favoured targets of a persistent DNA damage response

138 in ageing and stress-induced senescence. Nature Communications, 3, p.708.

Hiona, A. et al., 2010. Mitochondrial DNA mutations induce mitochondrial dysfunction, apoptosis and sarcopenia in skeletal muscle of mitochondrial DNA mutator mice. PloS one, 5(7), p.e11468.

Hoeijmakers, J.H.J., 2009. DNA damage, aging, and cancer. The New England journal of medicine, 361(15), pp.1475–1485.

Hohjoh, H. & Singer, M.F., 1996. Cytoplasmic ribonucleoprotein complexes containing human LINE-1 protein and RNA. The EMBO journal, 15(3), pp.630–639.

Hong, Z. et al., 2008. A polycomb group protein, PHF1, is involved in the response to DNA double-strand breaks in human cell. Nucleic Acids Research, 36(9), pp.2939–2947.

Hornykiewicz, O. & Kish, S.J., 1987. Biochemical pathophysiology of Parkinson's disease. Advances in neurology, 45, pp.19–34.

Horwich, M.D. et al., 2007. The Drosophila RNA methyltransferase, DmHen1, modifies germline piRNAs and single-stranded siRNAs in RISC. Current biology : CB, 17(14), pp.1265–1272.

Houtkooper, R.H., Williams, R.W. & Auwerx, J., 2010. Metabolic networks of longevity. Cell, 142(1), pp.9–14.

Houwing, S. et al., 2007. A role for Piwi and piRNAs in germ cell maintenance and transposon silencing in Zebrafish. Cell, 129(1), pp.69–82.

Houwing, S., Berezikov, E. & Ketting, R.F., 2008. Zili is required for germ cell differentiation and meiosis in zebrafish. The EMBO journal, 27(20), pp.2702–2711.

Hsu, A.-L., Murphy, C.T. & Kenyon, C., 2003. Regulation of aging and age-related disease by DAF-16 and heat-shock factor. Science (New York, N.Y.), 300(5622), pp.1142–1145.

Huang, X.A. et al., 2013. A Major Epigenetic Programming Mechanism Guided by piRNAs. Developmental cell, 24(5), pp.502–516.

Hunter, R.G. et al., 2012. Acute stress and hippocampal histone H3 lysine 9 trimethylation, a retrotransposon silencing response. Proceedings of the National Academy of Sciences of the United States of America, 109(43), pp.17657–17662.

Irwin, I., DeLanney, L.E. & Langston, J.W., 1993. MPTP and aging. Studies in the C57BL/6 mouse. Advances in neurology, 60, pp.197–206.

Irwin, M.H., Parameshwaran, K. & Pinkert, C.A., 2013. Mouse models of mitochondrial complex I dysfunction. The international journal of biochemistry & cell biology, 45(1), pp.34–40.

Ishizu, H., Siomi, H. & Siomi, M.C., 2012. Biology of PIWI-interacting RNAs: new insights into biogenesis and function inside and outside of germlines. Genes & Development, 26(21), pp.2361–2373.

139 Ivanschitz, L., De Thé, H. & Le Bras, M., 2013. PML, SUMOylation, and Senescence. Frontiers in oncology, 3, p.171.

Iyama, T. & Wilson, D.M., 2013. DNA repair mechanisms in dividing and non-dividing cells. DNA repair, 12(8), pp.620–636.

Jackson-Lewis, V. et al., 2000. Developmental cell death in dopaminergic neurons of the substantia nigra of mice. Journal of Comparative Neurology, 424(3), pp.476–488.

Jacobs, K.B. et al., 2012. Detectable clonal mosaicism and its relationship to aging and cancer. Nature Genetics, 44(6), pp.651–658.

Janzen, V. et al., 2006. Stem-cell ageing modified by the cyclin-dependent kinase inhibitor p16INK4a. Nature, 443(7110), pp.421–426.

Jaskelioff, M. et al., 2011. Telomerase reactivation reverses tissue degeneration in aged telomerase-deficient mice. Nature, 469(7328), pp.102–106.

Jay, F. et al., 2011. Misregulation of AUXIN RESPONSE FACTOR 8 underlies the developmental abnormalities caused by three distinct viral silencing suppressors in Arabidopsis. PLoS pathogens, 7(5), p.e1002035.

Jeck, W.R., Siebold, A.P. & Sharpless, N.E., 2012. Review: a meta-analysis of GWAS and age-associated diseases. Aging Cell, 11(5), pp.727–731.

Jin, G. et al., 2011. Histone deacetylase inhibitors enhance endothelial cell sprouting angiogenesis in vitro. Surgery, 150(3), pp.429–435.

Johnson, S.C., Rabinovitch, P.S. & Kaeberlein, M., 2013. mTOR is a key modulator of ageing and age-related disease. Nature, 493(7432), pp.338–345.

Joliot, A. & Prochiantz, A., 2004. Transduction peptides: from technology to physiology. Nature cell biology, 6(3), pp.189–196.

Joliot, A., Pernelle, C., et al., 1991. Antennapedia homeobox peptide regulates neural morphogenesis. Proceedings of the National Academy of Sciences of the United States of America, 88(5), pp.1864–1868.

Joliot, A.H., Triller, A., et al., 1991. alpha-2,8-Polysialic acid is the neuronal surface receptor of antennapedia homeobox peptide. The New biologist, 3(11), pp.1121–1134.

Joyner, A.L., 1996. Engrailed, Wnt and Pax genes regulate midbrain--hindbrain development. Trends in genetics : TIG, 12(1), pp.15–20.

Joyner, A.L., Liu, A. & Millet, S., 2000. Otx2, Gbx2 and Fgf8 interact to position and maintain a mid-hindbrain organizer. Current opinion in cell biology, 12(6), pp.736–741.

Jurk, D. et al., 2012. Postmitotic neurons develop a p21-dependent senescence-like phenotype driven by a DNA damage response. Aging Cell, 11(6), pp.996–1004.

Kadkhodaei, B. et al., 2013. Transcription factor Nurr1 maintains fiber integrity and nuclear- encoded mitochondrial gene expression in dopamine neurons. Proceedings of the

140 National Academy of Sciences of the United States of America, 110(6), pp.2360–2365.

Kaeberlein, M., McVey, M. & Guarente, L., 1999. The SIR2/3/4 complex and SIR2 alone promote longevity in Saccharomyces cerevisiae by two different mechanisms. Genes & Development, 13(19), pp.2570–2580.

Kaer, K. & Speek, M., 2013. Retroelements in human disease. Gene, 518(2), pp.231–241.

Kajikawa, M. & Okada, N., 2002. LINEs mobilize SINEs in the eel through a shared 3' sequence. Cell, 111(3), pp.433–444.

Kalmykova, A.I., Klenov, M.S. & Gvozdev, V.A., 2005. Argonaute protein PIWI controls mobilization of retrotransposons in the Drosophila male germline. Nucleic Acids Research, 33(6), pp.2052–2059.

Kanfi, Y. et al., 2010. SIRT6 protects against pathological damage caused by diet-induced obesity. Aging Cell, 9(2), pp.162–173.

Kanfi, Y. et al., 2012. The sirtuin SIRT6 regulates lifespan in male mice. Nature, 483(7388), pp.218–221.

Kawahara, T.L.A. et al., 2009. SIRT6 links histone H3 lysine 9 deacetylation to NF-kappaB- dependent gene expression and organismal life span. Cell, 136(1), pp.62–74.

Kazazian, H.H. et al., 1988. Haemophilia A resulting from de novo insertion of L1 sequences represents a novel mechanism for mutation in man. Nature, 332(6160), pp.164–166.

Kenyon, C.J., 2010. The genetics of ageing. Nature, 464(7288), pp.504–512.

Kiecker, C. & Lumsden, A., 2005. Compartments and their boundaries in vertebrate brain development. Nature Reviews Neuroscience, 6(7), pp.553–564.

Kim, H., Bakshi, A. & Kim, J., 2015. Retrotransposon-derived promoter of Mammalian aebp2. PloS one, 10(4), p.e0126966.

Kim, N. et al., 2014. Regulation of retinal axon growth by secreted Vax1 homeodomain protein. eLife, 3, p.e02671.

Kim, S., Park, J.Y. & Yang, T.-J., 2015. Characterization of three active transposable elements recently inserted in three independent DFR-A alleles and one high-copy DNA transposon isolated from the Pink allele of the ANS gene in onion (Allium cepa L.). Molecular genetics and genomics : MGG, 290(3), pp.1027–1037.

Kines, K.J. et al., 2014. Potential for genomic instability associated with retrotranspositionally-incompetent L1 loci. Nucleic Acids Research.

Kirik, D. et al., 2002. Parkinson-like neurodegeneration induced by targeted overexpression of alpha-synuclein in the nigrostriatal system. Journal of Neuroscience, 22(7), pp.2780– 2791.

Kitada, T. et al., 1998. Mutations in the parkin gene cause autosomal recessive juvenile parkinsonism. Nature, 392(6676), pp.605–608.

141 Kitkumthorn, N. & Mutirangura, A., 2011. Long interspersed nuclear element-1 hypomethylation in cancer: biology and clinical applications. Clinical epigenetics, 2(2), pp.315–330.

Koga, H., Kaushik, S. & Cuervo, A.M., 2011. Protein homeostasis and aging: The importance of exquisite quality control. Ageing research reviews, 10(2), pp.205–215.

Koito, A. & Ikeda, T., 2013. Intrinsic immunity against retrotransposons by APOBEC cytidine deaminases. Frontiers in microbiology, 4, p.28.

Kowall, N.W. et al., 2000. MPTP induces alpha-synuclein aggregation in the substantia nigra of baboons. Neuroreport, 11(1), pp.211–213.

Krenning, L. et al., 2014. Transient activation of p53 in G2 phase is sufficient to induce senescence. Molecular Cell, 55(1), pp.59–72.

Krishnan, V. et al., 2011. Histone H4 lysine 16 hypoacetylation is associated with defective DNA repair and premature senescence in Zmpste24-deficient mice. Proceedings of the National Academy of Sciences of the United States of America, 108(30), pp.12325–12330.

Kroemer, G., Galluzzi, L. & Brenner, C., 2007. Mitochondrial membrane permeabilization in cell death. Physiological reviews, 87(1), pp.99–163.

Krüger, R. et al., 1998. Ala30Pro mutation in the gene encoding alpha-synuclein in Parkinson's disease. Nature Genetics, 18(2), pp.106–108.

Kubo, S. et al., 2006. L1 retrotransposition in nondividing and primary human somatic cells. Proceedings of the National Academy of Sciences of the United States of America, 103(21), pp.8036–8041.

Kuilman, T. et al., 2010. The essence of senescence. Genes & Development, 24(22), pp.2463– 2479.

Kujoth, G.C. et al., 2005. Mitochondrial DNA mutations, oxidative stress, and apoptosis in mammalian aging. Science (New York, N.Y.), 309(5733), pp.481–484.

Kuramochi-Miyagawa, S.S. et al., 2008. DNA methylation of retrotransposon genes is regulated by Piwi family members MILI and MIWI2 in murine fetal testes. Genes & Development, 22(7), pp.908–917.

Kuwabara, T. et al., 2009. Wnt-mediated activation of NeuroD1 and retro-elements during adult neurogenesis. Nature Neuroscience, 12(9), pp.1097–1105.

Kyriakis, J.M. & Avruch, J., 2012. Mammalian MAPK signal transduction pathways activated by stress and inflammation: a 10-year update. Physiological reviews, 92(2), pp.689–737.

Laguna, A. et al., 2015. Dopaminergic control of autophagic-lysosomal function implicates Lmx1b in Parkinson's disease. Nature Neuroscience, 18(6), pp.826–835.

Lamming, D.W. et al., 2012. Rapamycin-induced insulin resistance is mediated by mTORC2 loss and uncoupled from longevity. Science (New York, N.Y.), 335(6076), pp.1638–1643.

142 Lander, E.S. et al., 2001. Initial sequencing and analysis of the human genome. Nature, 409(6822), pp.860–921.

Langston, J.W. & Ballard, P.A., 1983. Parkinson's disease in a chemist working with 1- methyl-4-phenyl-1,2,5,6-tetrahydropyridine. The New England journal of medicine, 309(5), p.310.

Laplante, M. & Sabatini, D.M., 2012a. mTOR signaling in growth control and disease. Cell, 149(2), pp.274–293.

Laplante, M. & Sabatini, D.M., 2012b. mTOR Signaling. Cold Spring Harbor perspectives in biology, 4(2).

Larson, K. et al., 2012. Heterochromatin Formation Promotes Longevity and Represses Ribosomal RNA Synthesis S. K. Kim, ed. PLoS Genetics, 8(1), p.e1002473.

Layalle, S. et al., 2011. Engrailed homeoprotein acts as a signaling molecule in the developing fly. Development (Cambridge, England), 138(11), pp.2315–2323.

Le Roux, I. et al., 1993. Neurotrophic activity of the Antennapedia homeodomain depends on its specific DNA-binding properties. Proceedings of the National Academy of Sciences of the United States of America, 90(19), pp.9120–9124.

Lee, E.J. et al., 2011. Identification of piRNAs in the central nervous system. RNA, 17(6), pp.1090–1099.

Lee, I.H. et al., 2008. A role for the NAD-dependent deacetylase Sirt1 in the regulation of autophagy. Proceedings of the National Academy of Sciences of the United States of America, 105(9), pp.3374–3379.

Lee, J. et al., 2007. Different evolutionary fates of recently integrated human and chimpanzee LINE-1 retrotransposons. Gene, 390(1-2), pp.18–27.

Lee, J.S. et al., 2012. Meta-analysis of gene expression in the mouse liver reveals biomarkers associated with inflammation increased early during aging. Mechanisms of ageing and development, 133(7), pp.467–478.

Lee, M.K. et al., 2002. Human alpha-synuclein-harboring familial Parkinson's disease-linked Ala-53 --> Thr mutation causes neurodegenerative disease with alpha-synuclein aggregation in transgenic mice. Proceedings of the National Academy of Sciences of the United States of America, 99(13), pp.8968–8973.

Leroy, E. et al., 1998. The ubiquitin pathway in Parkinson's disease. Nature, 395(6701), pp.451–452.

Lesaffre, B. et al., 2007. Direct non-cell autonomous Pax6 activity regulates eye development in the zebrafish. Neural Development, 2, p.2.

Lev, N. et al., 2006. Role of DJ-1 in Parkinson's disease. Journal of molecular neuroscience : MN, 29(3), pp.215–225.

Li, J. et al., 2014. An antisense promoter in mouse L1 retrotransposon open reading frame-1

143 initiates expression of diverse fusion transcripts and limits retrotransposition. Nucleic Acids Research.

Li, J.-Q., Tan, L. & Yu, J.-T., 2014. The role of the LRRK2 gene in Parkinsonism. 9(1), pp.1– 17.

Li, J.-Y. et al., 2008. Lewy bodies in grafted neurons in subjects with Parkinson's disease suggest host-to-graft disease propagation. Nature medicine, 14(5), pp.501–503.

Li, W. et al., 2013. Activation of transposable elements during aging and neuronal decline in Drosophila. Nature Neuroscience.

Liang, V. et al., 2014. Altered proteostasis in aging and heat shock response in C. elegans revealed by analysis of the global and de novo synthesized proteome. Cellular and molecular life sciences : CMLS, 71(17), pp.3339–3361.

Lim, S. & Kaldis, P., 2013. Cdks, cyclins and CKIs: roles beyond cell cycle regulation. Development (Cambridge, England), 140(15), pp.3079–3093.

Linnane, A.W. et al., 1989. Mitochondrial DNA mutations as an important contributor to ageing and degenerative diseases. Lancet, 1(8639), pp.642–645.

Liu, B. et al., 2005. Genomic instability in laminopathy-based premature aging. Nature medicine, 11(7), pp.780–785.

Liu, B.-L. et al., 2007. New enlightenment of French Paradox: resveratrol's potential for cancer chemoprevention and anti-cancer therapy. Cancer biology & therapy, 6(12), pp.1833–1836.

Liu, Y. et al., 2002. The UCH-L1 gene encodes two opposing enzymatic activities that affect alpha-synuclein degradation and Parkinson's disease susceptibility. Cell, 111(2), pp.209– 218.

Lombard, D.B. et al., 2007. Mammalian Sir2 homolog SIRT3 regulates global mitochondrial lysine acetylation. Molecular and cellular biology, 27(24), pp.8807–8814.

López-Otín, C. et al., 2013. The hallmarks of aging. Cell, 153(6), pp.1194–1217.

Löbrich, M. et al., 2010. gammaH2AX foci analysis for monitoring DNA double-strand break repair: strengths, limitations and optimization. Cell cycle (Georgetown, Tex.), 9(4), pp.662–669.

Lynch-Day, M.A. et al., 2012. The role of autophagy in Parkinson's disease. Cold Spring Harbor perspectives in medicine, 2(4), p.a009357.

Ma, C. & D'Mello, S.R., 2011. Neuroprotection by histone deacetylase-7 (HDAC7) occurs by inhibition of c-jun expression through a deacetylase-independent mechanism. Journal of Biological Chemistry, 286(6), pp.4819–4828.

Madabhushi, R., Pan, L. & Tsai, L.-H., 2014. DNA Damage and Its Links to Neurodegeneration. Neuron, 83(2), pp.266–282.

144 Maegawa, S. et al., 2010. Widespread and tissue specific age-related DNA methylation changes in mice. Genome research, 20(3), pp.332–340.

Mainguy, G. et al., 2000. An induction gene trap for identifying a homeoprotein-regulated locus. Nature biotechnology, 18(7), pp.746–749.

Mair, W. et al., 2011. Lifespan extension induced by AMPK and calcineurin is mediated by CRTC-1 and CREB. Nature, 470(7334), pp.404–408.

Manning-Bog, A.B. et al., 2002. The herbicide paraquat causes up-regulation and aggregation of alpha-synuclein in mice: paraquat and alpha-synuclein. Journal of Biological Chemistry, 277(3), pp.1641–1644.

Marteijn, J.A. et al., 2014. Understanding nucleotide excision repair and its roles in cancer and ageing. Nature reviews. Molecular cell biology, 15(7), pp.465–481.

Martin, S.L. & Bushman, F.D., 2001. Nucleic acid chaperone activity of the ORF1 protein from the mouse LINE-1 retrotransposon. Molecular and cellular biology, 21(2), pp.467– 475.

Martínez, P. & Blasco, M.A., 2010. Role of shelterin in cancer and aging. Aging Cell, 9(5), pp.653–666.

Mathias, S.L. et al., 1991. Reverse transcriptase encoded by a human transposable element. Science (New York, N.Y.), 254(5039), pp.1808–1810.

Matsuda, N. et al., 2010. PINK1 stabilized by mitochondrial depolarization recruits Parkin to damaged mitochondria and activates latent Parkin for mitophagy. The Journal of cell biology, 189(2), pp.211–221.

Matsumoto, L. et al., 2010. CpG demethylation enhances alpha-synuclein expression and affects the pathogenesis of Parkinson's disease. PloS one, 5(11), p.e15522.

Matsuoka, Y. et al., 2001. Lack of nigral pathology in transgenic mice expressing human alpha-synuclein driven by the tyrosine hydroxylase promoter. Neurobiology of Disease, 8(3), pp.535–539.

McCLINTOCK, B., 1950. The origin and behavior of mutable loci in maize. Proceedings of the National Academy of Sciences of the United States of America, 36(6), pp.344–355.

McCormack, A.L. et al., 2002. Environmental risk factors and Parkinson's disease: selective degeneration of nigral dopaminergic neurons caused by the herbicide paraquat. Neurobiology of Disease, 10(2), pp.119–127.

Melis, J.P.M. et al., 2013. Slow accumulation of mutations in Xpc-/- mice upon induction of oxidative stress. DNA repair, 12(12), pp.1081–1086.

Mesquita, A. et al., 2010. Caloric restriction or catalase inactivation extends yeast chronological lifespan by inducing H2O2 and superoxide dismutase activity. Proceedings of the National Academy of Sciences of the United States of America, 107(34), pp.15123– 15128.

145 Min, J.-N. et al., 2008. CHIP deficiency decreases longevity, with accelerated aging phenotypes accompanied by altered protein quality control. Molecular and cellular biology, 28(12), pp.4018–4025.

Minois, N., 2014. Molecular basis of the “anti-aging” effect of spermidine and other natural polyamines - a mini-review. Gerontology, 60(4), pp.319–326.

Mizushima, N. et al., 2008. Autophagy fights disease through cellular self-digestion. Nature, 451(7182), pp.1069–1075.

Molofsky, A.V. et al., 2006. Increasing p16INK4a expression decreases forebrain progenitors and neurogenesis during ageing. Nature, 443(7110), pp.448–452.

Morais, V.A. et al., 2009. Parkinson's disease mutations in PINK1 result in decreased Complex I activity and deficient synaptic function. EMBO molecular medicine, 1(2), pp.99–111.

Moran, J.V. et al., 1996. High frequency retrotransposition in cultured mammalian cells. Cell, 87(5), pp.917–927.

Morimoto, R.I. & Cuervo, A.M., 2014. Proteostasis and the aging proteome in health and disease. The journals of gerontology. Series A, Biological sciences and medical sciences, 69 Suppl 1, pp.S33–8.

Morisada, N. et al., 2010. Branchio-oto-renal syndrome caused by partial EYA1 deletion due to LINE-1 insertion. Pediatric nephrology (Berlin, Germany), 25(7), pp.1343–1348.

Morrison, A.J. & Shen, X., 2005. DNA repair in the context of chromatin. Cell cycle (Georgetown, Tex.), 4(4), pp.568–571.

Morrow, G. et al., 2004. Overexpression of the small mitochondrial Hsp22 extends Drosophila life span and increases resistance to oxidative stress. FASEB journal : official publication of the Federation of American Societies for Experimental Biology, 18(3), pp.598–599.

Moskalev, A.A. et al., 2012. Gadd45 proteins: relevance to aging, longevity and age-related pathologies. Ageing research reviews, 11(1), pp.51–66.

Moskalev, A.A. et al., 2013. The role of DNA damage and repair in aging through the prism of Koch-like criteria. Ageing research reviews, 12(2), pp.661–684.

Mossalam, M. et al., 2012. Direct induction of apoptosis using an optimal mitochondrially targeted p53. Molecular pharmaceutics, 9(5), pp.1449–1458.

Mostoslavsky, R. et al., 2006. Genomic instability and aging-like phenotype in the absence of mammalian SIRT6. Cell, 124(2), pp.315–329.

Muotri, A.R. et al., 2009. Environmental influence on L1 retrotransposons in the adult hippocampus. Hippocampus, 19(10), pp.1002–1007.

Muotri, A.R. et al., 2010. L1 retrotransposition in neurons is modulated by MeCP2. Nature, 468(7322), pp.443–446.

146 Muotri, A.R. et al., 2005. Somatic mosaicism in neuronal precursor cells mediated by L1 retrotransposition. Nature, 435(7044), pp.903–910.

Muramatsu, D. et al., 2013. Pericentric heterochromatin generated by HP1 protein interaction- defective histone methyltransferase Suv39h1. Journal of Biological Chemistry, 288(35), pp.25285–25296.

Murga, M. et al., 2009. A mouse model of ATR-Seckel shows embryonic replicative stress and accelerated aging. Nature Genetics, 41(8), pp.891–898.

Narendra, D. et al., 2008. Parkin is recruited selectively to impaired mitochondria and promotes their autophagy. The Journal of cell biology, 183(5), pp.795–803.

Németh, A. & Längst, G., 2011. Genome organization in and around the nucleolus. Trends in genetics : TIG, 27(4), pp.149–156.

Nicholas, A. et al., 2010. Age-related gene-specific changes of A-to-I mRNA editing in the human brain. Mechanisms of ageing and development, 131(6), pp.445–447.

Nicklas, W.J., Vyas, I. & Heikkila, R.E., 1985. Inhibition of NADH-linked oxidation in brain mitochondria by 1-methyl-4-phenyl-pyridine, a metabolite of the neurotoxin, 1-methyl-4- phenyl-1,2,5,6-tetrahydropyridine. Life sciences, 36(26), pp.2503–2508.

Nishibuchi, G. & Nakayama, J.-I., 2014. Biochemical and structural properties of heterochromatin protein 1: understanding its role in chromatin assembly. Journal of biochemistry, 156(1), pp.11–20.

Nishida, K.M. et al., 2007. Gene silencing mechanisms mediated by Aubergine piRNA complexes in Drosophila male gonad. RNA, 13(11), pp.1911–1922.

Nogueiras, R. et al., 2012. Sirtuin 1 and sirtuin 3: physiological modulators of metabolism. Physiological reviews, 92(3), pp.1479–1514.

Nordström, U. et al., 2015. Progressive nigrostriatal terminal dysfunction and degeneration in the engrailed1 heterozygous mouse model of Parkinson's disease. Neurobiology of Disease, 73, pp.70–82.

Norris, K.L. et al., 2015. Convergence of parkin, PINK1 and α-synuclein on stress-induced mitochondrial morphological remodelling. Journal of Biological Chemistry.

O'Leary, D.D.M., Chou, S.-J. & Sahara, S., 2007. Area Patterning of the Mammalian Cortex. Neuron, 56(2), pp.252–269.

O'Rourke, E.J. et al., 2013. ω-6 Polyunsaturated fatty acids extend life span through the activation of autophagy. Genes & Development, 27(4), pp.429–440.

O'Sullivan, R.J. & Karlseder, J., 2012. The great unravelling: chromatin as a modulator of the aging process. Trends in Biochemical Sciences, 37(11), pp.466–476.

Oberdoerffer, P. & Sinclair, D.A., 2007. The role of nuclear architecture in genomic instability and ageing. Nature reviews. Molecular cell biology, 8(9), pp.692–702.

147 Oberdoerffer, P. et al., 2008. SIRT1 redistribution on chromatin promotes genomic stability but alters gene expression during aging. Cell, 135(5), pp.907–918.

Olovnikov, A.M., 1996. Telomeres, telomerase, and aging: origin of the theory. Experimental gerontology, 31(4), pp.443–448.

Onken, B. & Driscoll, M., 2010. Metformin induces a dietary restriction-like state and the oxidative stress response to extend C. elegans Healthspan via AMPK, LKB1, and SKN-1. PloS one, 5(1), p.e8758.

Orozco-Solis, R. & Sassone-Corsi, P., 2014. Epigenetic control and the circadian clock: linking metabolism to neuronal responses. Neuroscience, 264, pp.76–87.

Osorio, F.G. et al., 2010. Nuclear envelope alterations generate an aging-like epigenetic pattern in mice deficient in Zmpste24 metalloprotease. Aging Cell, 9(6), pp.947–957.

Osorio, F.G. et al., 2012. Nuclear lamina defects cause ATM-dependent NF-κB activation and link accelerated aging to a systemic inflammatory response. Genes & Development, 26(20), pp.2311–2324.

Ostertag, E.M. & Kazazian, H.H., 2001. Biology of mammalian L1 retrotransposons. Annual review of genetics, 35, pp.501–538.

Osterwald, S. et al., 2015. PML induces compaction, TRF2 depletion and DNA damage signaling at telomeres and promotes their alternative lengthening. Journal of cell science, 128(10), pp.1887–1900.

Ovadia, A., Zhang, Z. & Gash, D.M., 1995. Increased susceptibility to MPTP toxicity in middle-aged rhesus monkeys. Neurobiology of aging, 16(6), pp.931–937.

Padeken, J. & Heun, P., 2014. Nucleolus and nuclear periphery: velcro for heterochromatin. Current opinion in cell biology, 28, pp.54–60.

Palm, W. & de Lange, T., 2008. How shelterin protects mammalian telomeres. Annual review of genetics, 42, pp.301–334.

Park, C.B. & Larsson, N.-G., 2011. Mitochondrial DNA mutations in disease and aging. The Journal of cell biology, 193(5), pp.809–818.

Pauler, F.M. et al., 2009. H3K27me3 forms BLOCs over silent genes and intergenic regions and specifies a histone banding pattern on a mouse autosomal chromosome. Genome research, 19(2), pp.221–233.

Peelaerts, W. et al., 2015. α-Synuclein strains cause distinct synucleinopathies after local and systemic administration. Nature, 522(7556), pp.340–344.

Pegoraro, G. et al., 2009. Ageing-related chromatin defects through loss of the NURD complex. Nature cell biology, 11(10), pp.1261–1267.

Peleg, S. et al., 2010. Altered histone acetylation is associated with age-dependent memory impairment in mice. Science (New York, N.Y.), 328(5979), pp.753–756.

148 Peng, C. et al., 2011. Pitx3 is a critical mediator of GDNF-induced BDNF expression in nigrostriatal dopaminergic neurons. Journal of Neuroscience, 31(36), pp.12802–12815.

Pereira, J.B. et al., 2013. Regional vulnerability of hippocampal subfields and memory deficits in Parkinson's disease. Hippocampus, 23(8), pp.720–728.

Petrucelli, L. et al., 2002. Parkin protects against the toxicity associated with mutant alpha- synuclein: proteasome dysfunction selectively affects catecholaminergic neurons. Neuron, 36(6), pp.1007–1019.

Pérez, V.I. et al., 2009. The overexpression of major antioxidant enzymes does not extend the lifespan of mice. Aging Cell, 8(1), pp.73–75.

Piacentini, L. et al., 2014. Transposons, environmental changes, and heritable induced phenotypic variability. Chromosoma, 123(4), pp.345–354.

Pollina, E.A. & Brunet, A., 2011. Epigenetic regulation of aging stem cells. Oncogene, 30(28), pp.3105–3126.

Polymeropoulos, M.H. et al., 1997. Mutation in the alpha-synuclein gene identified in families with Parkinson's disease. Science (New York, N.Y.), 276(5321), pp.2045–2047.

Ponomarev, I. et al., 2010. Amygdala transcriptome and cellular mechanisms underlying stress-enhanced fear learning in a rat model of posttraumatic stress disorder. Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology, 35(6), pp.1402–1411.

Ponomarev, I. et al., 2012. Gene coexpression networks in human brain identify epigenetic modifications in alcohol dependence. Journal of Neuroscience, 32(5), pp.1884–1897.

Powers, E.T. et al., 2009. Biological and chemical approaches to diseases of proteostasis deficiency. Annual review of biochemistry, 78, pp.959–991.

Price, N.L. et al., 2012. SIRT1 is required for AMPK activation and the beneficial effects of resveratrol on mitochondrial function. Cell metabolism, 15(5), pp.675–690.

Prochiantz, A. & Di Nardo, A.A., 2015. Homeoprotein Signaling in the Developing and Adult Nervous System. Neuron, 85(5), pp.911–925.

Qiu, X. et al., 2010. Calorie restriction reduces oxidative stress by SIRT3-mediated SOD2 activation. Cell metabolism, 12(6), pp.662–667.

Quiñinao, C., Prochiantz, A. & Touboul, J., 2015. Local homeoprotein diffusion can stabilize boundaries generated by graded positional cues. Development (Cambridge, England), 142(10), pp.1860–1868.

Raffaello, A. & Rizzuto, R., 2011. Mitochondrial longevity pathways. Biochimica et biophysica acta, 1813(1), pp.260–268.

Ragnauth, C.D. et al., 2010. Prelamin A acts to accelerate smooth muscle cell senescence and is a novel biomarker of human vascular aging. Circulation, 121(20), pp.2200–2210.

149 Raiz, J. et al., 2012. The non-autonomous retrotransposon SVA is trans-mobilized by the human LINE-1 protein machinery. Nucleic Acids Research, 40(4), pp.1666–1683.

Rajasethupathy, P. et al., 2012. A Role for Neuronal piRNAsin the Epigenetic Controlof Memory-Related Synaptic Plasticity. Cell, 149(3), pp.693–707.

Ramirez, A. et al., 2006. Hereditary parkinsonism with dementia is caused by mutations in ATP13A2, encoding a lysosomal type 5 P-type ATPase. Nature Genetics, 38(10), pp.1184–1191.

Rampon, C. et al., 2015. Control of brain patterning by Engrailed paracrine transfer: a new function of the Pbx interaction domain. Development (Cambridge, England), 142(10), pp.1840–1849.

Rando, T.A. & Chang, H.Y., 2012. Aging, rejuvenation, and epigenetic reprogramming: resetting the aging clock. Cell, 148(1-2), pp.46–57.

Rebollo, R., Romanish, M.T. & Mager, D.L., 2012. Transposable elements: an abundant and natural source of regulatory sequences for host genes. Annual review of genetics, 46, pp.21–42.

Rera, M. et al., 2011. Modulation of longevity and tissue homeostasis by the Drosophila PGC-1 homolog. Cell metabolism, 14(5), pp.623–634.

Rieker, C. et al., 2011. Nucleolar disruption in dopaminergic neurons leads to oxidative damage and parkinsonism through repression of mammalian target of rapamycin signaling. Journal of Neuroscience, 31(2), pp.453–460.

Rissling, I. et al., 2009. Haplotype analysis of the engrailed-2 gene in young-onset Parkinson's disease. Neuro-degenerative diseases, 6(3), pp.102–105.

Ristow, M. & Schmeisser, S., 2011. Extending life span by increasing oxidative stress. Free radical biology & medicine, 51(2), pp.327–336.

Rockwood, L.D., Felix, K. & Janz, S., 2004. Elevated presence of retrotransposons at sites of DNA double strand break repair in mouse models of metabolic oxidative stress and MYC-induced lymphoma. Mutation research, 548(1-2), pp.117–125.

Rodgers, J.T. et al., 2005. Nutrient control of glucose homeostasis through a complex of PGC-1alpha and SIRT1. Nature, 434(7029), pp.113–118.

Rodriguez, M. et al., 2015. Parkinson's disease as a result of aging. Aging Cell, 14(3), pp.293– 308.

Rogina, B. & Helfand, S.L., 2004. Sir2 mediates longevity in the fly through a pathway related to calorie restriction. Proceedings of the National Academy of Sciences of the United States of America, 101(45), pp.15998–16003.

Rose, S. et al., 2002. Age-related effects of 1-methyl-4-phenyl-l,2,3,6-tetrahydropyridine treatment of common marmosets. European Journal of Pharmacology, pp.1–9.

Ross, R.J., Weiner, M.M. & Lin, H., 2014. PIWI proteins and PIWI-interacting RNAs in the

150 soma. Nature, 505(7483), pp.353–359.

Rossi, D.J. et al., 2007. Deficiencies in DNA damage repair limit the function of haematopoietic stem cells with age. Nature, 447(7145), pp.725–729.

Rothwell, P.M. et al., 2011. Effect of daily aspirin on long-term risk of death due to cancer: analysis of individual patient data from randomised trials. Lancet, 377(9759), pp.31–41.

Rubinsztein, D.C., Mariño, G. & Kroemer, G., 2011. Autophagy and aging. Cell, 146(5), pp.682–695.

Rudolph, K.L. et al., 1999. Longevity, stress response, and cancer in aging telomerase- deficient mice. Cell, 96(5), pp.701–712.

Rulten, S.L. & Caldecott, K.W., 2013. DNA strand break repair and neurodegeneration. DNA repair, 12(8), pp.558–567.

Russell, S.J. & Kahn, C.R., 2007. Endocrine regulation of ageing. Nature reviews. Molecular cell biology, 8(9), pp.681–691.

Ryan, B.J. et al., 2015. Mitochondrial dysfunction and mitophagy in Parkinson's: from familial to sporadic disease. Trends in Biochemical Sciences, 40(4), pp.200–210.

Sahin, E. & DePinho, R.A., 2012. Axis of ageing: telomeres, p53 and mitochondria. Nature reviews. Molecular cell biology, 13(6), pp.397–404.

Sahin, U., Lallemand-Breitenbach, V. & De Thé, H., 2014. PML nuclear bodies: regulation, function and therapeutic perspectives. The Journal of pathology, 234(3), pp.289–291.

Saito, K. & Siomi, M.C., 2010. Small RNA-mediated quiescence of transposable elements in animals. Developmental cell, 19(5), pp.687–697.

Saito, K. et al., 2007. Pimet, the Drosophila homolog of HEN1, mediates 2“-O-methylation of Piwi- interacting RNAs at their 3” ends. Genes & Development, 21(13), pp.1603–1608.

Saito, K. et al., 2006. Specific association of Piwi with rasiRNAs derived from retrotransposon and heterochromatic regions in the Drosophila genome. Genes & Development, 20(16), pp.2214–2222.

Saksouk, N., Simboeck, E. & Déjardin, J., 2015. Constitutive heterochromatin formation and transcription in mammals. Epigenetics & chromatin, 8, p.3.

Salminen, A., Kaarniranta, K. & Kauppinen, A., 2012. Inflammaging: disturbed interplay between autophagy and inflammasomes. Aging, 4(3), pp.166–175.

Savitsky, K. et al., 1995. A single ataxia telangiectasia gene with a product similar to PI-3 kinase. Science (New York, N.Y.), 268(5218), pp.1749–1753.

Scaffidi, P. & Misteli, T., 2006. Lamin A-dependent nuclear defects in human aging. Science (New York, N.Y.), 312(5776), pp.1059–1063.

Scarpulla, R.C., 2011. Metabolic control of mitochondrial biogenesis through the PGC-1

151 family regulatory network. Biochimica et biophysica acta, 1813(7), pp.1269–1278.

Schumacher, B. et al., 2008. Delayed and accelerated aging share common longevity assurance mechanisms. PLoS Genetics, 4(8), p.e1000161.

Sebastián, C. et al., 2012. From sirtuin biology to human diseases: an update. Journal of Biological Chemistry, 287(51), pp.42444–42452.

Seeber, A., Hauer, M. & Gasser, S.M., 2013. Nucleosome remodelers in double-strand break repair. Current opinion in genetics & development, 23(2), pp.174–184.

Selman, C. et al., 2009. Ribosomal protein S6 kinase 1 signaling regulates mammalian life span. Science (New York, N.Y.), 326(5949), pp.140–144.

Serrano, M. et al., 1997. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell, 88(5), pp.593–602.

Sgadò, P. et al., 2006. Slow progressive degeneration of nigral dopaminergic neurons in postnatal Engrailed mutant mice. Proceedings of the National Academy of Sciences of the United States of America, 103(41), pp.15242–15247.

Sharpless, N.E. & DePinho, R.A., 2007. How stem cells age and why this makes us grow old. Nature reviews. Molecular cell biology, 8(9), pp.703–713.

Shaw, A.C. et al., 2010. Aging of the innate immune system. Current opinion in immunology, 22(4), pp.507–513.

Sherman, M.Y. & Goldberg, A.L., 2001. Cellular defenses against unfolded proteins: a cell biologist thinks about neurodegenerative diseases. Neuron, 29(1), pp.15–32.

Shimura, H. et al., 2000. Familial Parkinson disease gene product, parkin, is a ubiquitin- protein ligase. Nature Genetics, 25(3), pp.302–305.

Shumaker, D.K. et al., 2006. Mutant nuclear lamin A leads to progressive alterations of epigenetic control in premature aging. Proceedings of the National Academy of Sciences of the United States of America, 103(23), pp.8703–8708.

Sian, J. et al., 1994. Glutathione-related enzymes in brain in Parkinson's disease. Annals of neurology, 36(3), pp.356–361.

Sidhu, A. et al., 2004. The role of alpha-synuclein in both neuroprotection and neurodegeneration. Annals of the New York Academy of Sciences, 1035, pp.250–270.

Siebold, A.P. et al., 2010. Polycomb Repressive Complex 2 and Trithorax modulate Drosophila longevity and stress resistance. Proceedings of the National Academy of Sciences of the United States of America, 107(1), pp.169–174.

Simeone, A., 2000. Positioning the isthmic organizer where Otx2 and Gbx2meet. Trends in genetics : TIG, 16(6), pp.237–240.

Simon, H.H. et al., 2001. Fate of midbrain dopaminergic neurons controlled by the engrailed genes. Journal of Neuroscience, 21(9), pp.3126–3134.

152 Siomi, M.C. et al., 2011. PIWI-interacting small RNAs: the vanguard of genome defence. Nature reviews. Molecular cell biology, 12(4), pp.246–258.

Skene, P.J. et al., 2010. Neuronal MeCP2 is expressed at near histone-octamer levels and globally alters the chromatin state. Molecular Cell, 37(4), pp.457–468.

Skowronski, J. & Singer, M.F., 1985. Expression of a cytoplasmic LINE-1 transcript is regulated in a human teratocarcinoma cell line. Proceedings of the National Academy of Sciences of the United States of America, 82(18), pp.6050–6054.

Sloan, K.E., Bohnsack, M.T. & Watkins, N.J., 2013. The 5S RNP couples p53 homeostasis to ribosome biogenesis and nucleolar stress. Cell reports, 5(1), pp.237–247.

Small, S.A. & Duff, K., 2008. Linking Abeta and tau in late-onset Alzheimer's disease: a dual pathway hypothesis. Neuron, 60(4), pp.534–542.

Smit, A.F. et al., 1995. Ancestral, mammalian-wide subfamilies of LINE-1 repetitive sequences. Journal of molecular biology, 246(3), pp.401–417.

Smith, P.D. et al., 2003. Cyclin-dependent kinase 5 is a mediator of dopaminergic neuron loss in a mouse model of Parkinson's disease. Proceedings of the National Academy of Sciences of the United States of America, 100(23), pp.13650–13655.

Soda, K. et al., 2009. Polyamine-rich food decreases age-associated pathology and mortality in aged mice. Experimental gerontology, 44(11), pp.727–732.

Song, C.-X. & He, C., 2012. Balance of DNA methylation and demethylation in cancer development. Genome biology, 13(10), p.173.

Sonnier, L. et al., 2007. Progressive Loss of Dopaminergic Neurons in the Ventral Midbrain of Adult Mice Heterozygote for Engrailed1. Journal of Neuroscience, 27(5), pp.1063– 1071.

Sookdeo, A. et al., 2013. Revisiting the evolution of mouse LINE-1 in the genomic era. Mobile DNA, 4(1), pp.1–1.

Soria, G., Polo, S.E. & Almouzni, G., 2012. Prime, repair, restore: the active role of chromatin in the DNA damage response. Molecular Cell, 46(6), pp.722–734.

Spatazza, J. et al., 2013. Choroid-plexus-derived Otx2 homeoprotein constrains adult cortical plasticity. Cell reports, 3(6), pp.1815–1823.

Speek, M., 2001. Antisense promoter of human L1 retrotransposon drives transcription of adjacent cellular genes. Molecular and cellular biology, 21(6), pp.1973–1985.

Sperka, T., Wang, J. & Rudolph, K.L., 2012. DNA damage checkpoints in stem cells, ageing and cancer. Nature reviews. Molecular cell biology, 13(9), pp.579–590.

Spillantini, M.G. et al., 1997. Alpha-synuclein in Lewy bodies. Nature, 388(6645), pp.839– 840.

Stettler, O. et al., 2012. Engrailed homeoprotein recruits the adenosine A1 receptor to

153 potentiate ephrin A5 function in retinal growth cones. Development (Cambridge, England), 139(1), pp.215–224.

Strong, R. et al., 2008. Nordihydroguaiaretic acid and aspirin increase lifespan of genetically heterogeneous male mice. Aging Cell, 7(5), pp.641–650.

Suberbielle, E. et al., 2013. Physiologic brain activity causes DNA double-strand breaks in neurons, with exacerbation by amyloid-β. Nature Neuroscience, 16(5), pp.613–621.

Sugiyama, S. et al., 2008. Experience-Dependent Transfer of Otx2 Homeoprotein into the Visual Cortex Activates Postnatal Plasticity. Cell, 134(3), pp.508–520.

Sun, Y. et al., 2015. Signaling pathway of MAPK/ERK in cell proliferation, differentiation, migration, senescence and apoptosis. Journal of receptor and signal transduction research, pp.1–5.

Surmeier, D.J. & Schumacker, P.T., 2013. Calcium, bioenergetics, and neuronal vulnerability in Parkinson's disease. Journal of Biological Chemistry, 288(15), pp.10736–10741.

Swindell, W.R. et al., 2009. Endocrine regulation of heat shock protein mRNA levels in long- lived dwarf mice. Mechanisms of ageing and development, 130(6), pp.393–400.

Takahashi, K. et al., 2001. DJ-1 positively regulates the androgen receptor by impairing the binding of PIASx alpha to the receptor. Journal of Biological Chemistry, 276(40), pp.37556–37563.

Talens, R.P. et al., 2012. Epigenetic variation during the adult lifespan: cross-sectional and longitudinal data on monozygotic twin pairs. Aging Cell, 11(4), pp.694–703.

Tanner, C.M., 1992. Epidemiology of Parkinson's disease. Neurologic clinics, 10(2), pp.317– 329.

Tao, R. et al., 2010. Sirt3-mediated deacetylation of evolutionarily conserved lysine 122 regulates MnSOD activity in response to stress. Molecular Cell, 40(6), pp.893–904.

Tapias, A. et al., 2014. Trrap-dependent histone acetylation specifically regulates cell-cycle gene transcription to control neural progenitor fate decisions. Cell stem cell, 14(5), pp.632–643.

Tchénio, T., Casella, J.F. & Heidmann, T., 2000. Members of the SRY family regulate the human LINE retrotransposons. Nucleic Acids Research, 28(2), pp.411–415.

Teng, T., Thomas, G. & Mercer, C.A., 2013. Growth control and ribosomopathies. Current opinion in genetics & development, 23(1), pp.63–71.

Thomas, C.A., Paquola, A.C.M. & Muotri, A.R., 2012. LINE-1 retrotransposition in the nervous system. Annual review of cell and developmental biology, 28, pp.555–573.

Tilstra, J.S. et al., 2012. NF-κB inhibition delays DNA damage-induced senescence and aging in mice. The Journal of clinical investigation, 122(7), pp.2601–2612.

Tissenbaum, H.A. & Guarente, L., 2001. Increased dosage of a sir-2 gene extends lifespan in

154 Caenorhabditis elegans. Nature, 410(6825), pp.227–230.

Tolkunova, E.N. et al., 1998. Two distinct types of repression domain in engrailed: one interacts with the groucho corepressor and is preferentially active on integrated target genes. Molecular and cellular biology, 18(5), pp.2804–2814.

Tomaru, U. et al., 2012. Decreased proteasomal activity causes age-related phenotypes and promotes the development of metabolic abnormalities. The American journal of pathology, 180(3), pp.963–972.

Tomás-Loba, A. et al., 2008. Telomerase reverse transcriptase delays aging in cancer-resistant mice. Cell, 135(4), pp.609–622.

Torero Ibad, R. et al., 2011. Otx2 promotes the survival of damaged adult retinal ganglion cells and protects against excitotoxic loss of visual acuity in vivo. Journal of Neuroscience, 31(14), pp.5495–5503.

Trifunovic, A. et al., 2004. Premature ageing in mice expressing defective mitochondrial DNA polymerase. Nature, 429(6990), pp.417–423.

Tsurumi, A. & Li, W.X., 2012. Global heterochromatin loss: a unifying theory of aging? Epigenetics : official journal of the DNA Methylation Society, 7(7), pp.680–688.

Tufi, R. et al., 2014. Enhancing nucleotide metabolism protects against mitochondrial dysfunction and neurodegeneration in a PINK1 model of Parkinson's disease. Nature cell biology, 16(2), pp.157–166.

Uhl, G.R., Hedreen, J.C. & Price, D.L., 1985. Parkinson's disease: loss of neurons from the ventral tegmental area contralateral to therapeutic surgical lesions. Neurology, 35(8), pp.1215–1218.

Ungerstedt, U., 1968. 6-Hydroxy-dopamine induced degeneration of central monoamine neurons. European Journal of Pharmacology, 5(1), pp.107–110.

Vagin, V.V. et al., 2006. A distinct small RNA pathway silences selfish genetic elements in the germline. Science (New York, N.Y.), 313(5785), pp.320–324.

Van Meter, M. et al., 2014. SIRT6 represses LINE1 retrotransposons by ribosylating KAP1 but this repression fails with stress and age. Nature Communications, 5, pp.1–10.

Van Raamsdonk, J.M. & Hekimi, S., 2009. Deletion of the mitochondrial superoxide dismutase sod-2 extends lifespan in Caenorhabditis elegans. PLoS Genetics, 5(2), p.e1000361.

Van Remmen, H. et al., 2003. Life-long reduction in MnSOD activity results in increased DNA damage and higher incidence of cancer but does not accelerate aging. Physiological genomics, 16(1), pp.29–37.

Veenvliet, J.V. et al., 2013. Specification of dopaminergic subsets involves interplay of En1 and Pitx3. Development (Cambridge, England), 140(16), pp.3373–3384.

Vermulst, M. et al., 2008. DNA deletions and clonal mutations drive premature aging in

155 mitochondrial mutator mice. Nature Genetics, 40(4), pp.392–394.

Vijg, J. & Suh, Y., 2013. Genome instability and aging. Annual review of physiology, 75, pp.645–668.

Visanji, N.P. et al., 2013. The prion hypothesis in Parkinson's disease: Braak to the future. Acta neuropathologica communications, 1, p.2.

Viswanath, V. et al., 2001. Caspase-9 activation results in downstream caspase-8 activation and bid cleavage in 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-induced Parkinson's disease. Journal of Neuroscience, 21(24), pp.9519–9528.

Viswanathan, M. & Guarente, L., 2011. Regulation of Caenorhabditis elegans lifespan by sir- 2.1 transgenes. Nature, 477(7365), pp.E1–2.

Vogt, J. et al., 2014. SVA retrotransposon insertion-associated deletion represents a novel mutational mechanism underlying large genomic copy number changes with non- recurrent breakpoints. Genome biology, 15(6), p.R80.

Walker, G.A. & Lithgow, G.J., 2003. Lifespan extension in C. elegans by a molecular chaperone dependent upon insulin-like signals. Aging Cell, 2(2), pp.131–139.

Wang, C. et al., 2009. DNA damage response and cellular senescence in tissues of aging mice. Aging Cell, 8(3), pp.311–323.

Wang, K. & Klionsky, D.J., 2011. Mitochondria removal by autophagy. Autophagy, 7(3), pp.297–300.

Wang, R.-H. et al., 2008. Impaired DNA damage response, genome instability, and tumorigenesis in SIRT1 mutant mice. Cancer cell, 14(4), pp.312–323.

Wang, W. et al., 2009. Neural cell cycle dysregulation and central nervous system diseases. Progress in neurobiology, 89(1), pp.1–17.

Weick, E.-M. & Miska, E.A., 2014. piRNAs: from biogenesis to function. Development (Cambridge, England), 141(18), pp.3458–3471.

Westerheide, S.D. et al., 2009. Stress-inducible regulation of heat shock factor 1 by the deacetylase SIRT1. Science (New York, N.Y.), 323(5917), pp.1063–1066.

Wilkinson, J.E. et al., 2012. Rapamycin slows aging in mice. Aging Cell, 11(4), pp.675–682.

Wilkinson, K.D., 2000. Ubiquitination and deubiquitination: targeting of proteins for degradation by the proteasome. Seminars in cell & developmental biology, 11(3), pp.141– 148.

Wizenmann, A. et al., 2009. Extracellular Engrailed participates in the topographic guidance of retinal axons in vivo. Neuron, 64(3), pp.355–366.

Wolff, E.M. et al., 2010. Hypomethylation of a LINE-1 promoter activates an alternate transcript of the MET oncogene in bladders with cancer. PLoS Genetics, 6(4), p.e1000917.

156 Wu, H. & Sun, Y.E., 2009. Reversing DNA methylation: new insights from neuronal activity- induced Gadd45b in adult neurogenesis. Science Signaling, 2(64), p.pe17.

Wu, H. et al., 2010. Genome-wide analysis reveals methyl-CpG-binding protein 2-dependent regulation of microRNAs in a mouse model of Rett syndrome. Proceedings of the National Academy of Sciences of the United States of America, 107(42), pp.18161–18166.

Wurst, W. & Bally-Cuif, L., 2001. Neural plate patterning: upstream and downstream of the isthmic organizer. Nature Reviews Neuroscience, 2(2), pp.99–108.

Wurst, W., Auerbach, A.B. & Joyner, A.L., 1994. Multiple developmental defects in Engrailed-1 mutant mice: an early mid-hindbrain deletion and patterning defects in forelimbs and sternum. Development (Cambridge, England), 120(7), pp.2065–2075.

Wutz, A., 2011. Gene silencing in X-chromosome inactivation: advances in understanding facultative heterochromatin formation. Nature Reviews Genetics, 12(8), pp.542–553.

Xie, Y. et al., 2011. Characterization of L1 retrotransposition with high-throughput dual- luciferase assays. Nucleic Acids Research, 39(3), pp.e16–e16.

Xu, H. et al., 2014. Cohesin Rad21 Mediates Loss of Heterozygosity and Is Upregulated via Wnt Promoting Transcriptional Dysregulation in Gastrointestinal Tumors. Cell reports, 9(5), pp.1781–1797.

Xu, W., Tan, L. & Yu, J.-T., 2015. The link between the SNCA gene and parkinsonism. Neurobiology of aging, 36(3), pp.1505–1518.

Yang, S.-B. et al., 2012. Rapamycin ameliorates age-dependent obesity associated with increased mTOR signaling in hypothalamic POMC neurons. Neuron, 75(3), pp.425–436.

Yang, Y. et al., 2013. Polycomb group protein PHF1 regulates p53-dependent cell growth arrest and apoptosis. Journal of Biological Chemistry, 288(1), pp.529–539.

Yilmaz, Ö.H. et al., 2012. mTORC1 in the Paneth cell niche couples intestinal stem-cell function to calorie intake. Nature, 486(7404), pp.490–495.

Yoon, B.C. et al., 2012. Local translation of extranuclear lamin B promotes axon maintenance. Cell, 148(4), pp.752–764.

Yu, F. et al., 2001. Methyl-CpG-binding protein 2 represses LINE-1 expression and retrotransposition but not Alu transcription. Nucleic Acids Research, 29(21), pp.4493– 4501.

Zhang, C. & Cuervo, A.M., 2008. Restoration of chaperone-mediated autophagy in aging liver improves cellular maintenance and hepatic function. Nature medicine, 14(9), pp.959–965.

Zhang, G. et al., 2013. Hypothalamic programming of systemic ageing involving IKK-β, NF- κB and GnRH. Nature, 497(7448), pp.211–216.

Zhang, Y. et al., 2009. Mice deficient in both Mn superoxide dismutase and glutathione peroxidase-1 have increased oxidative damage and a greater incidence of pathology but

157 no reduction in longevity. The journals of gerontology. Series A, Biological sciences and medical sciences, 64(12), pp.1212–1220.

Zhang, Y. et al., 2000. Parkin functions as an E2-dependent ubiquitin- protein ligase and promotes the degradation of the synaptic vesicle-associated protein, CDCrel-1. Proceedings of the National Academy of Sciences of the United States of America, 97(24), pp.13354–13359.

Zhong, L. et al., 2010. The histone deacetylase Sirt6 regulates glucose homeostasis via Hif1alpha. Cell, 140(2), pp.280–293.

Zhou, Y.-Y. et al., 2015. MAPK/JNK signalling: a potential autophagy regulation pathway. Bioscience reports, 35(3).

Zilles, K. & Amunts, K., 2010. Centenary of Brodmann's map--conception and fate,

Zimprich, A. et al., 2011. A mutation in VPS35, encoding a subunit of the retromer complex, causes late-onset Parkinson disease. American journal of human genetics, 89(1), pp.168– 175.

Zimprich, A. et al., 2004. Mutations in LRRK2 cause autosomal-dominant parkinsonism with pleomorphic pathology. Neuron, 44(4), pp.601–607.

158 Homeoproteins are a major class of transcription factors, which in addition to their established roles as products of morphogenetic genes have cell autonomous and non-cell autonomous functions in the adult brain. For example, En1 and En2 (collectively Engrailed) are responsible for midbrain dopaminergic (mDA) neuron maintenance during adulthood. Indeed, En1+/- mice display a selective and progressive degeneration of the mDA neurons of the Substantia Nigra pars compacta (SNpc) reminiscent of Parkinson disease. Furthermore, Engrailed infusion in the SNpc saves neurons in several models of the disease, namely mutated A30P α-Synuclein, MPTP/MPP+ and 6-hydroxydopamine (6-OHDA). In the case of MPTP, Engrailed-mediated neuroprotection partially necessitates its ability to up-regulate the translation of Ndufs1, a subunit of the mitochondrial Complex I. The main aim of my thesis was to study additional neuroprotective pathways mediated by Engrailed.

We first studied En1+/- mice and showed that neurodegeneration was accompanied by increased DNA damage and specific changes in chromatin marks. Similar but accelerated symptoms were seen in an acute oxidative stress model, consisting in the injection of 6- OHDA directly in the SNpc of wild type mice. In this model, Engrailed injection, 30 minutes after that of 6-OHDA, saves mDA neurons and progressively restores genomic stability (the number of DNA breaks) and a healthy chromatin state (monitored by the intensity and distribution of chromatin marks). The Engrailed-mediated positive impact on these key hallmarks of ageing lends weight to the idea that this protein is a promising anti-ageing factor, possibly useful in the treatment of Parkinson disease.

Long interspersed nuclear elements (LINE-1) are a class of retrotransposons which have the ability to multiply in the genome, via a “copy and paste” mechanism. They are normally repressed, leading to the assumption that they are silent. However, we have shown that they are expressed in mouse mDA neurons and that their expression is increased by an acute oxidative stress, leading to DNA damage and degeneration. Their direct repression by Engrailed allows us to propose that this transcriptional and/or epigenetic silencing explains part of Engrailed protective activity in mouse models of acute and progressive oxidative stress. Whether the interest of controlling the expression of mobile elements can be extended to other mouse models of human pathologies and to other members of the homeoprotein family remains an open question.