Down Syndrome (DS) is the most frequent human chromosomal disorder. Main symptoms include intellectual disability (ID), cardiovascular defects and craniofacial dysmorphisms. Despite ID being measured by a test of symbolic logic skills, it is common for children with DS to arouse a climate of affective intensity greater than the norm. In 1959, Jérôme Lejeune (1926-1994) and coll. described an additional chromosome 21 (Hsa21) in children with DS (trisomy 21), giving origin to the field of medical genetics. Remarkably, the discovery of trisomy 21 had relevant social consequences for the affected children, in that their parents were no longer suspected to be alcoholics or infected with syphilis. Although it is broadly agreed that the DS phenotype originates from the altered expression of the genes located on Hsa21, its molecular pathogenesis is still unknown. To date, no therapy is recognized and recommended by guidelines as being effective in improving the cognitive abilities of persons with DS. The aim of this article is to categorize main therapeutical approaches or pathways to new approaches reported in the biomedical literature, to extract critical methodological points from the works of Lejeune and then to propose a new research project aimed to generate and integrate clinical, biochemical, genetic and bioinformatic data in order to identify novel therapeutic targets for this form of trisomy. We show here that nearly all the current lines of research were pursued, theorized or foreseen by Lejeune, and that central points of his method remain current: positive hypothesis about the existence of a solution, envision of systematic investigation of cell machinery, anchoring of clinical and biochemical finding to the chromosome physical map, and continuing clinical observation of the affected children. We therefore propose a project aimed at producing both experimentally and by meta-analysis state-of-the-art maps and databases related to clinical/phenotype, cytogenetics, exome, transcriptome, methylome, molecular biology, metabolome and mutations data. The primary expected outcome of this research project is the identification of a restricted list of strong candidate genes and mechanisms for ID in persons with DS in order to devise new rational therapeutic approaches.
1 Department of Experimental, Diagnostic and Specialty Medicine (DIMES), Alma Mater Studiorum - University of Bologna, Bologna, Italy
2 Neonatology Unit, St. Orsola-Malpighi Polyclinic, Bologna, Italy
3 Department of Medical and Biological Sciences, University of Udine, Udine, Italy
4 Research Laboratory "Stem Cells", U.O.C. Immunohematology-Transfusion Medicine and Laboratory of Hematology, Santo Spirito's Hospital, Pescara, Italy
5 Department of Molecular and Translational Medicine, University of Brescia, Brescia, Italy
6 Department of Health Sciences, University of Milano Bicocca, Monza, Milan, Italy
7 Department of Oncology, Lady Davis Institute, McGill University, Montreal, Quebec, Canada
8 Laboratory of Molecular Medicine and Neuroscience, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD, USA
9 Medical Genetics Unit, St. Orsola-Malpighi Polyclinic, Department of Medical and Surgical Sciences (DIMEC), University of Bologna, Bologna, Italy
10 Neonatology Unit, St. Orsola-Malpighi Polyclinic, Department of Medical and Surgical Sciences (DIMEC), University of Bologna, Bologna, Italy
*These Authors Contributed equally to this work.
RecievedOct 1 2013 AcceptedDec 13 2013 PublishedDec 18 2013
CitationPierluigi Strippoli, Maria Chiara Pelleri, Maria Caracausi, Lorenza Vitale, Allison Piovesan, Chiara Locatelli, Maria Chiara Mimmi, Anna Concetta Berardi, Doris Ricotta, Annalisa Radeghieri, Donatella Barisani, Mark Basik, Maria Chiara Monaco, Alessandro Ghezzo, Marco Seri and Guido Cocchi (2013) An integrated route to identifying new pathogenesis-based therapeutic approaches for trisomy 21 (Down Syndrome) following the thought of Jérôme Lejeune. Science Postprint 1(1): e00010. doi:10.14340/spp.2013.12R0005
Copyright©2013 The Authors. Science Postprint published by General Healthcare Inc. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs 2.1 Japan (CC BY-NC-ND 2.1 JP) License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
FundingWe are deeply grateful to the Fondazione Umano Progresso, Milano, Italy, for supporting the research on trisomy 21 conducted at the DIMES Dept. MC fellowship is funded by a donation from the company Illumia, Bologna, Italy, that we thank greatly for their interest in our research.
We thank all the other people that very kindly contributed by individual donations to support part of the work that we are conducting on the subject. In particular, we are profoundly grateful to the Costa family, their friends (especially the Dal Monte and the Ghignone newlyweds), the "Gruppo Arzdore" and the community of Dozza (Bologna, Italy), as well as to the Morini family and the community of Pesaro (Italy), for their generous support to our trisomy 21 research.
Competing interestsNo relevant competing interests were disclosed.
Corresponding authorPierluigi Strippoli
AddressDepartment of Experimental, Diagnostic and Specialty Medicine (DIMES), Unit of Histology, Embriology and Applied Biology, University of Bologna, Via Belmeloro 8, 40126 Bologna (BO), Italy
E-mailIf you want to contact author,Please register as a member.
Down Syndrome (DS) is the most frequent human chromosomal disorder, with a frequency of 1 in ~400 conceptions and 1 in ~700 births (Morris et al. 1999, Parker et al. 2010). Main symptoms include intellectual disability (ID), cardiovascular defects and craniofacial dysmorphisms (Gardiner et al. 2010). More in detail, these symptoms may be observed: distinct facial and physical features among which are almond shaped eyes (due to epicanthal folds), a small, somewhat flat nose, a small mouth with a protruding tongue, a single crease across the palms, inturned fifth finger of their hands, a larger than normal space between the first and second toes; language production deficiency; cognitive impairment, which is present to some degree of severity in all affected individuals and which involves symbolic thought, whereas affectivity and social skills are conserved; cardiac defects in about 40% of cases; Hirschprung's disease; hypotonia; visual and hearing impairments; increased risk of leukemia, in particular megakaryoblastic; immune disorders, including increased susceptibility to infections and to autoimmune pathologies such as alopecia and celiac disease (CD); endocrine disorders, including hypothyroidism; and early-onset cognitive decline with neuropathological alterations similar to those observed in the brains of patients with Alzheimer’s disease (AD) (Epstein 1989, Roizen and Patterson 2003, Mégarbané et al. 2009, Gardiner et al. 2010, Letourneau and Antonarakis 2012, Hickey et al. 2012). Several studies have shown that individuals with DS have a specific cancer risk pattern, or tumor profile. For instance, their risk of developing leukemia and testicular cancer is much higher than age-matched controls, while women with DS almost never develop breast cancers (Hasle et al. 2000, Patja et al. 2006).
Despite ID being measured by a test of symbolic logic skills, it is common for children with DS to arouse a climate of affective intensity greater than the normal (Gruppo Facebook Genitori di ragazzi Down 2012). A recent study asked people with DS, ages 12 and older, about their self-perception, finding that the overwhelming majority of people with DS indicated that they were happy with their lives, liked who they are, liked how they look and expressed love for their families, indicating they live happy and fulfilling lives (Skotko et al. 2011).
In 1959, the young French doctor Jérôme Lejeune (1926-1994) published, together with Marthe Gautier and Raymond Turpin, the finding of an additional chromosome 21 in nine children with DS (Lejeune et al. 1959). This condition has been called trisomy 21, a genetic mutation leading to the presence of three copies of human chromosome 21 (Hsa21), instead of the normal two, in the cells of the affected individuals. This discovery is commonly recognized as a milestone in the history of genetics (National Human Genome Research Institute 2013), because it introduced the notion that a given clinical symptom may be connected to a specific alteration of the human genetic material for the first time, giving origin to the field of medical genetics. Remarkably, the discovery of trisomy 21 had relevant social consequence for the affected children, in that their parents were no longer suspected to be alcoholics or infected with syphilis. A man whose karyotype was one of the original nine having being diagnosed with three copies of Hsa21 35 years before, stated publicly at the funeral of Lejeune at Notre Dame in Paris: "Thank you, my professor, for what you did for my father and my mother. Because of you, I am proud of myself" (Lejeune-Gaymard 2012). The derogatory term "mongolism", derived from the original description of the syndrome in 1866 by John Langdon Down and based on the racial theories widespread at the time, was abandoned in favor of "Down Syndrome" or "trisomy 21" following a petition to the journal The Lancet in 1961 (Allen et al. 1961, Stevenson 2009).
The DS phenotype is expected to be associated with an altered expression of the genes located on Hsa21 (Sinet et al. 1975, Gardiner and Costa 2006, Roper and Reeves 2006, Pritchard et al. 2008, Korenberg 2009, Patterson 2009). Although DS was the first genetic alteration to have been described in humans and the most frequent form of ID caused by a microscopically demonstrable chromosomal aberration, its molecular pathogenesis is still unknown. Basic research on DS is now rapidly accelerating, and there is the possibility that the results will be beneficial for individuals with DS (Antonarakis and Epstein 2006), as well as for patients with other diseases whose risk is increased or decreased in DS. For example, one of the most frequent autoimmune diseases associated with DS is CD. Its pathogenesis involves an altered response of the immune system caused by the ingestion of proteins (gliadins) derived from wheat, rye and barley in genetically susceptible individuals, which induces the production of pro-inflammatory cytokines and duodenal mucosal lesions characterized by different degrees of villous atrophy, crypt hyperplasia and the presence of intraepithelial and lamina propria lymphocytic infiltration. Clinical manifestations can range from lack of symptoms to severe malabsorption, malnutrition, weight loss and severe anemia (Fasano and Catassi 2012). Prevalence rates of CD in patients with DS have been reported to range from 0% to 19% (Gale et al. 1997, Bonamico et al. 2001, Agardh et al. 2002, Pavlovic et al. 2010), but a recent population study in Sweden evaluated the relative risk for CD in DS to be six-fold higher (Mårild et al. 2013). The increased prevalence of CD in DS has also been considered in different guidelines; The American Academy of Pediatrics recommends CD screening in children with DS that show CD-related symptoms (Hill et al. 2005), whereas The Celiac Disease Guideline Committee of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition recommends testing for CD also in asymptomatic children with DS (Husby et al. 2012). The explanation of the increased prevalence of CD in DS could shed light on the pathogenesis of CD, which is common throughout the world and affects around one in 100 to one in 300 of the population (Rewers 2005, Bai et al. 2013).
Due to the fact that ID is the most constant symptom in DS and the one most affecting daily life of persons with DS, for the purpose of this article we refer to a "therapy for DS" as a cure able to improve the cognitive state of persons with DS. To date, no therapy is recognized and recommended by guidelines as being effective in improving the cognitive abilities of persons with DS. The aim of this article is to categorize main therapeutical approaches or pathways to new approaches reported in the biomedical literature, to extract critical methodological points from the works of Lejeune and then to propose a new research project aimed to generate and integrate clinical, biochemical, genetic and bioinformatic data in order to identify novel therapeutic targets for this form of trisomy. We will not treat environmental therapies here.
All throughout the text, italics indicates exact quotations from Prof. Lejeune.
The current approaches aimed to improve DS ID are summarized in Table 1. Therapies for other specific symptoms are not discussed here. We focus on cognitive aspects typical of children because it is believed that neurological Alzheimer's-like deterioration recognizes a different mechanism based on the simple observation that, while virtually all subjects with DS have cognitive defects and cognitive and behavioural symptoms suggestive of AD dementia affect most people with DS by 70 years of age, only 2–5% have dementia by age 40 (The Lancet Neurology 2013).
The aim of this section is not to provide a comprehensive review of each proposed therapeutic application in DS because there are several recent excellent reviews on the subject (in particular, Costa and Scott-McKean 2013; see also Mégarbané et al. 2009, Bartesaghi et al. 2011, Costa AC 2011, de la Torre and Dierrsen 2012, Busciglio et al. 2013), but we would rather propose a clustering of the described treatments to date or treatment proposals in a very few classes of main therapeutic approaches in order to highlight the central concept of each class. A presentation and a brief discussion for each approach is given below. Interestingly, nearly all the described current lines of research were pursued, theorized or foreseen in old papers published by Lejeune.
A breakthrough in research about Hsa21 has been the recent demonstration that this chromosome can be inactivated in vitro by transferring the XIST gene which is able to switch off one of the two X chromosomes in female cells, to one of the three Hsa21 present in trisomy 21 cells (Jiang et al. 2013) through zinc finger nucleases (ZNF). This is the most radical therapy: getting rid of the DNA element which is responsible for the disease and restoring normal cell function as these Authors have shown in an induced pluripotent stem (iPS) cell model (Jiang et al. 2013). The concept has come from the group that first described the expression of the XIST large noncoding RNA as the mechanism responsible for X eterochromatinization (lyonization) (Clemson et al. 1996).
Interestingly, Lejeune appears to have been the first to have gotten this suggestion from nature, as he writes: The most obvious therapy of a trisomic condition would be to silence the extra chromosome so that only a diploid amount of genetic information would remain active. Indeed nature is that shrewd and inactivation of supernumerary X chromosomes is a very efficient trick. Unfortunately the basic mechanism is still unknown although the recent use of anti-sense RNA gives already the possibility of silencing some specific sequences (Lejeune 1988). He has been somewhat prophetical because the "still unknown" mechanism would have been discovered 9 years later, in 1996, two years after his death. Indeed, an older trace of this idea may be found at least 10 years before, when he proposed turning off the extra chromosome by some kind of induced inactivation (like the lyonisation of supernumerary X chromosomes) (Lejeune 1977).
While the realization in vitro of this powerful modulation of the activity of an entire chromosome can be very useful in studying pathogenesis of DS by comparing cells derived from the same line and expressing (or not) Hsa21 genes following induction of XIST activity (Jiang et al 2013), it presently seems that many problems must be resolved before this discovery can even be applied. In particular, an effective method to transfer XIST in living humans must be found and there should be no risk of rearrangements or alterations to the germline; the transfer should be feasible in the appropriate tissues and at appropriate times; 5% of the Hsa21 genes remain active (Jiang et al. 2013) and they could maintain a possibly significant part of the phenotype; the procedure would be available in a very small number of superspecialized centers and would not be easily accessed by most of the centers interested. In any case, a new line of research has been opened and it could have unexpected consequences on the field of research on DS.
A different way to avoid the effects of the supernumerary copy of Hsa21 could be its physical elimination from the cells. A method able to do this in vitro has been recently described (Li et al. 2012) based on Hsa21 engineering with sequences able to negatively select it during cell proliferation under certain circumstances. To date, this method has been shown to have a low efficiency requiring complex treatments of the cells and raising issues related to the feasibility of the creation of transgenic cells in humans similar to those discussed in the point above. However, this is another path that may be followed, and it could take advantage of the basic studies that have demonstrated the possibility to regulate cell aneuploidy state in yeast (Waghmare and Bruschi 2005).
Theoretical bases for a possible intervention of selective chromosome destruction have been posed in recent articles (Barrangou 2012, Certo et al. 2012).
The possible establishment of uniparental disomy (UPD, retention of two homologous chromosomes from the same parent) for Hsa21 (UPD21, reviewed in Engel and Antonarakis 2002) by following approaches based on Hsa21 targeting should be carefully considered because, while in one case with UPD21 (paternal isodisomy) the phenotype was normal (Blouin et al. 1993), UPD21 (maternal heterodisomy) has been associated with early embryonic failure (Henderson et al. 1994) or acute lymphocytic leukemia (type L1) blast cells (maternal isodisomy, Rogan et al. 1995).
It is an established fact that a defect of cell proliferation, and in particular of neuron proliferation, is present in DS (Bartesaghi et al. 2011), so that an established approach to find a cure for DS is trying to rescue neuronal proliferation and interconnection. This approach is still not based on a formal theory of Hsa21 genes implied in the phenotype, and it appears to have stemmed from empirical observations of alterations of the brain in DS mouse models and in humans. Following the finding that antidepressants may stimulate neurogenesis in the hippocampus, a target organ in DS, the use of fluoxetine or lithium has been proposed (Table 1). The open issue in this case is the feasibility of the administration of powerful mood stabilizers, with potentially serious side effects, to a pediatric population based on studies conducted on mice.
Other observations have underlined a role of the protein Sonic Hedgehog (SHH) in the development of the cerebellum whose size is reduced by about 40% in DS (Table 1). An SHH agonist has very recently given remarkable results in rescuing a correct number of neurons in DS model mice (Das et al. 2013). Safety remains the main pitfall, because SHH in humans is a very powerful stimulator of mitosis and it is likely to increase the risk of tumorigenesis.
The possibility of an intervention aimed to stimulate neurogenesis was foreseen by Lejeune (reviewed in Le Méné 1997), who affirmed: a part of the network is lacking (...) For the geneticist, the idea to reactivate the embryonal steps that failed is still in the future of science, but it is not ecxluded that one day the regeneration process may be controlled (our translation).
Alterations of neurotransmission and of functions of selected brain regions in DS are specific of this condition to a certain extent, as also shown by neuropsychology investigation (e.g. Vicari and Carlesimo 2006, Menghini et al. 2011, Costanzo et al. 2013). However, there is not a clear pathogenetic scheme based on Hsa21 and most data come from experimentation with laboratory mice. To date, interventions with drugs active on the central nervous system (Gardiner 2010) have not been demonstrated to be effective, or have obtained limited or controversial results (Table 1). Again, the main issue is the potential side effects of the drugs and that sometimes they cannot be administered due to a low therapeutic index. We think that there is the need in this field to establish the proposal for a specific drug on a firmer basis. Lejeune was aware of a disturbance of connection among central nervous system neurons, and his hypothesis was to "detoxify" neurons where the implied pathways would be clarified, restoring normal functioning to the obstructed, backed up synapses (Lejeune 1978, Le Méné 1997). For this, he insisted on the need of a more detailed knowledge of the biochemistry of the involved processes and of their relationship to Hsa21. He thought deeply on the interconnected reactions at the basis of production of adrenergic and cholinergic mediators, representing them in accurate models (Figure 1, from Lejeune 1977) and pointing out that A simplified scheme shows that these metabolic deviations could provoke a disturbance of the collagen and of synthesis of chemical mediators (Lejeune 1979a).
Reprinted with kind permission by "Hereditas" from Figure 2 in Lejeune 1977. This illustration, exploiting the analogy between mechanical gears (compounds) connected by belts (enzymes), is about 25 years ahead of later commonly used interconnected graphs to show intricacies of cell machinery.
An additional approach directed toward neuroprotection has been described, to date, only in mouse models and as empirically based (Table 1).
|1. Hsa21 targeting|
|1.1 Chromosome inactivation||iPS cells||Jiang et al. 2013|
|1.2 Chromosome elimination||Yeast||Waghmar and Bruschi 2005|
|iPS cells||Li et al. 2012|
|1.3 Chromosome destruction||Bacteria||Barrangou 2012|
|Cell culture||Certo et al. 2012|
|2. Neurogenesis stimulation|
|2.1.1 Fluoxetine||Ts65Dn Mouse||Clark et al. 2006,
Bianchi et al. 2010b,
Heinen et al. 2012,
Guidi et al. 2013
|2.1.2 Lithium||Ts65Dn Mouse||Bianchi et al. 2010a,
Contestabile et al. 2013
|2.2 Peptide 6 (CNTF)||Ts65Dn Mouse||Blanchard et al. 2011|
|2.3 Sonic Hedgehog (SHH)||Ts65Dn Mouse||Roper et al. 2006,
Das et al. 2013
|3. Neurotransmission modulation|
|3.1 Choline supplement||Ts65Dn Mouse||Moon et al. 2010|
|3.2 Melatonin||Ts65Dn Mouse||Corrales et al. 2013|
|3.3 Acetyl L-Carnitine||Clinical Trial||Pueschel 2006|
|3.4 Adrenergic agonists|
|3.4.1 Formoterol (b2)||Ts65Dn Mouse||Dang et al. 2013|
|3.4.2 Xamoterol (b1)||Ts65Dn Mouse||Salehi et al. 2009,
Faizi et al. 2011
|3.5 Inhibitors of acetylcholinesterase|
|3.5.1 Pentylenetetrazole||Ts65Dn Mouse||Rueda et al. 2008|
|3.5.2 Donepezil||Clinical Trial (age 10-17 yrs)||Kishani et al. 2010|
|3.5.3 Rivastigmine||Clinical Trial (mean age 16 yrs)||Heller et al. 2010|
|3.5.4 Galantamine||Ts65Dn Mouse||de Souza et al. 2011|
|3.6 Glutamatergic neurotransmission|
|3.6.1 Memantine||Ts65Dn Mouse||Costa et al. 2008,
Rueda et al. 2010,
Lockrow et al. 2011,
Scott-McKean and Costa 2011
|Clinical Trial (age >40 yrs)||Hanney et al. 2012|
|Clinical Trial (age 18-32 yrs)||Boada et al. 2012|
|3.7 GABA Receptor antagonists|
|3.7.1 Picrotoxin||Ts65Dn Mouse||Kleschevnikov et al. 2004,
Costa and Grybko 2005,
Fernandez et al. 2007,
Kleschevnikov et al. 2012b
|3.7.2 α5IA||Ts65Dn Mouse||Braudeau et al. 2011a, b|
|3.7.3 CGP55845||Ts65Dn Mouse||Kleschevnikov et al. 2012a, b|
|3.7.4 Ethosuximide, Gabapentin||Ts65Dn Mouse||Vidal et al. 2012|
|3.7.5 RO4938581||Ts65Dn Mouse||Martínez-Cué et al. 2013|
|3.8 GABA derivative (Piracetam)||Ts65Dn Mouse||Moran et al. 2002|
|Clinical trial (age 6.5-13 yrs)||Lobaugh et al. 2001|
|4.1 Neuropeptides||Ts65Dn Mouse||Toso et al. 2008,
Vink et al. 2009,
Incerti et al. 2011, 2012
|5. Neurodegeneration rescue|
|5.1 Estrogens||Ts65Dn Mouse||Granholm et al. 2003|
|5.2 Nerve growth factor (NGF)||Ts65Dn Mouse||Cooper et al. 2001|
|5.3 Minocycline||Ts65Dn Mouse||Hunter et al. 2004|
|6. Vitamins and antioxidants|
|6.1 Antioxidants, Folate
|Clinical Trials||Ellis et al. 2008,
Blehaut et al. 2010,
Tiano et al. 2012
|6.2 Vitamin E||Ts65Dn Mouse||Lockrow et al. 2009,
Shichiri et al. 2011
|7. Pathogenesis investigation|
|7.1 Down Syndrome Candidate Regions (DSCR)||Cytogenetics in partial trisomy||Lyle et al. 2009,
Korbel et al. 2009, others
|7.2 Cellular biology, molecular biology and molecular genetics of trisomy 21|
|7.2.1 Epigallocatechine gallate||Ts65Dn Mouse||Xie et al. 2008|
|tgYAC152F7 Mouse||Guedj et al. 2009|
|Tg152F7, Tg189N3, Ts65Dn Mice||Noll et al. 2009|
|Ts65Dn Mouse||Mazur-Kolecka et al. 2012|
|Cell culture from DS subjects||Valenti et al. 2013|
|Clinical Trial (age 14-29 yrs)||de la Torre et al. 2013|
|7.2.2 DYRK1A inhibitors||Structure-based Screening||Wang D et al. 2012|
|7.2.3 Beta-amyloid||Ts65Dn Mouse||Netzer et al. 2010|
|7.3 Mouse models of DS||DS Mouse Models||Rueda et al. 2012, others|
|7.4 General ID models||Computational Biology||Sturgeon et al. 2012,
Wang W et al. 2012
Neurodegeneration rescue has been proposed, based on mouse models, as an alternative approach aimed at treating neuronal alterations (Table 1).
Some pathogenetic basis established by biochemical analysis in cell systems or in vivo has led to the proposal of different vitamins, antioxidants, or their combination as treatments for DS (Ellis et al. 2008). Recent reviews in the field fail to find evidence in favor of administration of such substances (Ellis et al. 2008). Clinical trials with antioxidants geared toward reducing dementia symptoms in DS persons have been disappointing to date (Lott 2012). However, there are recurrent clues toward the presence of metabolic disturbances that could actually benefit from similar therapies, so further investigation is needed to identify more aimed dosages or combinations, or explore old or new substances that could be effective, in possibly some, if not all, subgroups (Blehaut et al. 2010). Actually, Lejeune was the first to propose the use of folic acid (a form of vitamin B9) to prevent spina bifida (Lejeune-Gaymard 2012). This shows that a simple vitamin supplementation may significantly affect pathogenetic processes when given on the basis of a rationale established on biochemistry (Hibbard and Hibbard 1968).
We now briefly describe the main research lines aimed to better understand trisomy 21 mechanisms leading to the symptoms in order to propose therapies targeting specific molecular alterations demonstrated in DS.
There has been much debate about the existence of selected, critical regions or genes on Hsa21 as the main one responsible for ID or for the other symptoms (Rahmani et al. 1990, Korenberg et al. 1990, Delabar et al. 1993). Since the first reports in the '70's (e.g. Raoul et al. 1976 from the group of Lejeune), descriptions of partial (segmental) trisomies have accumulated, although at slow pace. Two landmark studies in 2009, each analyzing about 30 cases of these conditions, have concluded that it is not likely that there is a single region candidate for DS phenotype (Down Syndrome Candidate/Critical Region, DSCR) on Hsa21 (Lyle et al. 2009, Korbel et al. 2009), and globally the field remains controversial, to the point that further analysis on the subject has been considered unwarranted (Sturgeon et al. 2012). However, it remains to be explained how such a homogeneous phenotype, though with its natural individual variability, may be generated by completely different portions of the same chromosome when involvement of a single different gene is usually associated with completely different phenotypes of monofactorial origin. We believe that this is still a fundamental line of research. In fact, even if it would not be possible to affirm with certainty that a gene, if present in an extra copy, is more responsible than the other for a symptom, it can always be excluded that a gene not present in three copies in at least one subject with a typical DS phenotype is critical for ID, the nearly universal symptom in DS. On the other hand, subjects with nearly absent ID and partial trisomy 21 might offer complementary clues to critical regions. It is important to critically evaluate all described cases and to continue searching for partial trisomy proposing molecular cytogenetics (CGH-Array) investigations (Melis et al. 2011) in particular in cases with a discordant genotype-phenotype relationship. Interesting advancements could also come by comparison of trisomy and monosomy effects (Turchetti et al. 2011), as again highlighted by the group of Lejeune (Pangalos et al. 1992).
In the last years, several remarkable approaches for analysis of trisomy 21 have become available. We could cite, as the main approaches, in vitro cellular models and genomics and post-genomics studies (Créau 2012). In vitro models allow the study of biochemistry and molecular biology of trisomic cells (e.g. Li et al. 2006, Valenti et al. 2013). This approach has been made powerful by the recent possibility to obtain iPS cell lines by trisomy 21 cells, which are so far to date only fibroblasts (Park et al. 2008, Mou et al. 2012, Li et al. 2012, Chou et al. 2012, Briggs et al. 2013, Weick et al. 2013, Jiang et al. 2013) or amniotic cells (Lu et al. 2013). The genomics and post-genomics approaches allow the accurate study of structure, expression and function of Hsa21 genes (e.g. Sailani et al. 2013, Salemi et al. 2013), as well of genome, transcriptome by expression microarray (e.g., among the first reports, FitzPatrick et al. 2002, Giannone et al. 2004, Mao et al. 2003, Mao et al. 2005) or massive (deep) RNA sequencing (RNA-Seq, Costa V et al. 2011) and possibily proteoma of trisomy 21 cells, all made possible by the landmark sequencing of Hsa21 (Hattori et al. 2000) and then of the human genome (International Human Genome Sequencing Consortium 2001). These lines of research have produced and are producing a wealth of data that are invaluable starting points for the building of pathogenetic models particularly when integrated with the other approaches. However, more recent and powerful techniques such as deep sequencing have to date been used only at a very limited extent for the study of exome/genome, transcriptome and methylome in DS. When data of this type accumulates for many different subjects and is made available in the context of forthcoming formats available for easy data exchange (and integration with clinical data), a major advance is expected toward the comprehension of genotype-phenotype relationships to therefore identify therapy targets.
The establishment and use of mouse models of DS has perhaps been the most successful method of investigation among research in DS (reviewed in Gotti et al. 2011, Liu et al. 2011, Yamakawa 2012, Edgin et al. 2012, Kleschevnikov et al. 2012c, Rueda et al. 2012). It has led to the understanding of basic mechanisms of DS, to the description of alterations in specific subsets of neurons (e.g. Necchi et al. 2008, Di Filippo et al. 2010) and to concrete therapy proposals (Table 1). However, mice mimicking DS are difficult to grow and breed, and the data obtained from these animals are somewhat controversial (Table 1). In addition, evidence derived from the murine DS model may not be automatically translated to humans due to the great differences between the two species. Investigations on murine models have already led to experimental therapies with discordant outcomes in humans: memantine was reported to be effective in DS model mice (Table 1) but it was shown to be ineffective in adults with DS (Hanney et al. 2012) and to ameliorate some abilities in young adults with DS (Boada et al. 2012).
Sturgeon and coll. (2012) have performed an interesting analysis on the basis that different genetic conditions leading to ID must have some common step involved in the development of the cognitive defect that they eventually share. Finding pathways shared by different conditions is a promising line of research that may help to focus on particular genes and proteins as good candidates for ID, as in the case of the interaction recently identified between the DSCR1 (RCAN1) protein, encoded by Down syndrome critical region 1 (Regulator of calcineurin 1) Hsa21-located gene, and the Fragile X mental retardation protein (FMRP) (Wang W et al. 2012). The interaction regulates both dendritic spine morphogenesis and local protein synthesis.
An influential suggestion in this direction may be found in an article by Lejeune, who, comparing DS, thyroid deficiency and trisomy for the short arm of chromosome 12 at the phenotype level, linked some related biochemical activities to the location of their relative genes observed: It seems difficult to consider as purely fortuitous the fact that three mental deficiences, correlated with short nose, short stature and adiposity, and clinically sometimes very similar, are related to enzymatic changes so close to each other (Lejeune 1977).
On the other hand, DS shows very specific alteration features of brain functioning (Menghini et al. 2011, Costanzo et al. 2013), reinforcing the concept that any realistic pathogenic model for trisomy 21 has to take into account the primitive localization of the disturbances in the Hsa21 molecule.
Prof. Jérôme Lejeune has been undoubtely a unique figure in the research on DS (Fondation Jérôme Lejeune 2013). He was an expert pediatrician who took care of about 9,000 children with ID and extablished relationships with them defined by a colleague as “legendary” (Israël and Arnaldez 1997). He was a genetician and cytogenetist who discovered trisomy 21, as well as many other chromosomal disorders, and reviewed more than 30,000 karyotypes with his group. He was a skilled biochemist who, using wooden molecular models built by himself, suggested for the first time the use of folic acid to prevent spina bifida (Lejeune-Gaymard 2012). He was a fine thinker, writer, speaker, and University teacher who was able to synthesize and clarify the most complex biomedical concepts. Most of all, all of these characteristics were found in the same person. This exceptional circumstance offered a deepness of fact-based reasoning, an inspired continuous flow of ideas and concepts and an unusual freshness to his scientific articles that makes the reading of them worthwhile tens of years later without feeling that they are out-of-date. We propose in this section a reevaluation of some methodological points extracted from the scientific thought of Lejeune that we believe have to be addressed still today in order to construct an integrated path toward a therapy of ID in DS supported by a rational background, including the greatest possible number of factors that are in play in the DS condition.
Now in the case of Trisomy 21 I am not at all going to say the cure is just around the corner. I don't know, but we know enough to consider that on theoretical grounds the idea that nothing could be done because it had an extra chromosome is not warranted. On the contrary, because it has an excess of normal material they probably have some prediction of normal things but at too great an extent and if we could just block this prediction they would come back slowly to normal (Lejeune 1992). These and other similar statements of Lejeune have been critical for us to recover a positive attitude toward the possibility that trisomy 21 might be cured. This "positive hypothesis" cannot be taken for granted, due to the complexity of this condition and its affect on early development and poor knowledge about pathogenetic mechanisms, 54 years since it was first described. As a matter of fact, it has become common for researchers to intensify efforts toward prenatal diagnosis oriented to selective abortion rather than toward basic and applied research aimed at finding effective therapies. For example, a systematic search of the PubMed database of biomedical and life sciences literature (http://www.ncbi.nlm.nih.gov/pubmed/) showed that in the 1992-2012 period about 4,200 articles were published on the prenatal diagnosis of DS, with 1,500 on genetic mechanisms of the syndrome and 800 on DS therapy, the last ones including all symptomatic treatments for any manifestation of DS along with a fraction of articles actually exploring cures for the trisomy itself (Associazione Euresis 2012; data unshown). Along with this finding, despite the observation that DS is the most common disease of genetic origin in humans, funding for DS research has been only a minor fraction of that destined to the active research for many other genetic diseases, in particular monogenic diseases (e.g., Diament 2012). Maintaining a position of doubt about success in research leads to quitting, while continuing to be in an attitude of question leads to active research: "A real search always implies a positive answer as an ultimate hypothesis otherwise one would not search" (Giussani 1997). The exact answer can never be predicted, so one needs to remain open-minded about the final possible results (see point 2. below). Lejeune's idea that DS is an "intoxication" in which several compounds to be identified are present in a pathological excess within cells, and neurons in particular, opens the way to imagine a biochemical-pharmaceutical intervention aimed to remove or inhibit the excess compound/s, recovering cell functions slowed or blocked by the "intoxication". This solution, if found, could be very simple to implement without difficulties and risks typical of interventions at a DNA level, and due to plasticity of the brain, could theoretically ameliorate neuronal functions even if administered years after birth.
A typical problem faced by the investigation of a complex system (Conti et al. 2007), e.g. composed by a large number of interacting parts as biological objects indeed are, is the ability to identify components that are critical for a certain process executed by the system. This requires that we do not skew our interest in the system toward components that attract attention based on some type of a priori biased knowledge, e.g. the fact that a certain component is historically much more studied while others not. Ideally, to understand complex mechanisms we need to restart the observation of the object (a cell, an organ, a patient) as a whole and from the beginning, privileging all investigation methods that provide an "open" test of the system, exploring it in all directions to find critical checkpoints.
In molecular biology, the most relevant possibility in this sense came in the late '90's with the diffusion of genomics and then post-genomics tools. It is well known that having a virtually complete "list of parts" of the cells as they are encoded in the human genetic material became a reality following completion of sequencing of the euchromatic portion of the human genome, as well as the simultaneous determination of the expression level of thousands of genes of a given biological sample became feasible. These discoveries opened an unprecedented possibility to gain knowledge about a biological object at a system level ("system biology"), so allowing a non-biased, open investigation aimed at highlighting critical components among thousands of possibilities under the effective slogan used at the beginning of the use of microarray technology to study the global gene expression profile: "The hypothesis is there is no hypothesis" (Mir 2000).
For example, in classic biology it is historically known that hemoglobin chain genes are expressed at a very high level in red blood cells, following "macroscopic" observation of high hemoglobin content in these cells and detailed investigation of this gene that has been actually the first sequenced gene in humans (Marotta et al. 1977). However, is this gene really the one expressed at the highest level among the whole complement of tens of thousands transcript species, and at which relative extent? Post-genomics can confirm that hemoglobin genes are de facto the genes expressed at the highest level in blood red cells by building a quantitative model of global gene expression profile without any a priori assumption, as we have recently shown (Piovesan et al. 2013). It is less known that Prof. Lejeune was the first to, to our knowledge, envision system biology application to the understanding of trisomy 21 pathogenetic mechanisms. In 1977 he illustrated a remarkable review and perspective article on the mechanisms of mental deficiency in chromosomal diseases with a surprising mechanical machinery mimicking several tens of chemical compound interactions known at the time to be involved in adrenergic and cholinergic pathways (Lejeune 1977, Figure 2 reproduced here as Figure 1). The systematic biochemical investigations reviewed in the cited work by Lejeune led to disparate findings and he pointed out that the problem is precisely to relate them to each other to obtain a global and coherent picture of the studied system, a not simple task because they hardly fit together (Lejeune 1977). The powerful and popular image of the search for a single musician playing out of time within a great orchestra (Le Méné 1997) has been also used by Lejeune to simply express his belief that human reason has the ability to identify tiny disturbances in a complex but ordered context. Thirty-six years later, we now have the possibility to build much more complex and quantitative "omics" models than the "machine of Lejeune", but these resources have been used all over the world at a rather limited extent in comparison with the frequence and the relevance of the condition for the purpose of understanding trisomy 21 molecular pathogenesis.
A chromosome, a single DNA cell molecule, may be viewed as a vast territory filled with information, varying from point to point and expressed in a coordinate way to allow normal cell functioning. In trisomies, the association of the presence of genetic material of Hsa21 in excess with the clinical features specific of DS has received countless confirmations. This association has also been confirmed based on its functional foundation obtained through modern techniques showing specificity of gene expression alterations of genes located in the abnormal chromosomal molecule or in a stretch of it (e.g., Giannone et al. 2004). When studying a chromosomal disease, the coherence with the cytogenetical landscape of the affected cells becomes a final bench test for any pathogenetic model based on all other types of data. Again this was a central point in the influential papers where Lejeune traced the way for a global vision of the disorder. Accurate gene mapping was pursued by Lejeune's group for the Hsa21 genes encoding products that clinical and biochemical analysis showed altered. Clinical activity brought to attention children that were monosomic or trisomic for various segments of Hsa21 and by correlating the cytogenetic findings to the levels of the superoxyde dismutase (SOD1) enzyme and to the different phenotypes, the group was able to map SOD1 in the q22.1 sub-band in 1976 (Sinet et al. 1975). Indeed research on partial trisomy 21 is still today a very powerful tool to candidate, or to exclude, specific genes as responsible for phenotypic features of DS. Although two significant groups of cases were published a few years ago (Korbel et al. 2009, Lyle et al. 2009), this presentation of DS remains exceptional thus restricting to date any general conclusions about the role of one or more Hsa21 chromosomal regions or genes in the DS phenotype.
From this "geographical" point of view, accurate analysis and continuous monitoring of any report about partial trisomy 21 remains important because even single cases may allow for the exclusion of a main role for an Hsa21 gene if it is not present in three copies in subjects showing a typical DS phenotype.
Remarkably, this seems to be the case for DYRK1A (dual-specificity tyrosine-(Y)-phosphorylation regulated kinase 1A), one of the Hsa21 genes most studied as a gene critical for the phenotype (de la Torre and Dierssen 2012). In October 2012, a Turkish child was reported as having a typical DS phenotype, partial trisomy 21 and absence of one excess copy of DYRK1A and RCAN1 (Davies et al. 2007) by both FISH and CGH-Array (Cetin et al. 2012). This fact should be taken into serious account by rediscussing current models assuming as established a critical role of DYRK1A (and sometimes of RCAN1) for DS phenotype, often based on the mouse model (discussed at the next point below). This does not rule out that in patients with DS use of epigallocatechin-3-gallate (EGCG), a green tea catechin inhibiting DYRK1A kinase, may be beneficial, and actually EGCG administration for three months has been very recently reported to significantly improve some abilities measured by neuropsychological tests in adults with DS (de la Torre et al. 2013). The cognitive improvement observed requires further confirmation because this is a pilot study and in addition the improvement seems to be restricted to episodic and working memory, being not present in the other neuropsychological parameters investigated in the study (such as verbal, attention, executive functions and visual motor coordination skills) (de la Torre et al. 2013). The ameliorations observed could be related to actual overexpression of DYRK1A in subjects with three copies of this gene as it happens in most cases (having complete trisomy 21), so that it may contribute to some extent in ID phenotype in DS. However, the above mentioned report of a child with typical DS phenotype and absence of triplication of DYRK1A, to date not discussed in any report about the relationship between DS and DYRK1A, puts into question the assumption that this gene is critical for DS phenotype. Moreover, it should be noted that an effect on ID of EGCG could be justified by other actions of this interesting biological compound that has been described to also inhibit ADAMTS1 (Cudic et al. 2009), a metallopeptidase encoded by another Hsa21 gene and overexpressed in DS and in AD (Miguel et al. 2005). Considering that EGCG also modulates Hsa21-encoded APP cleavage (Rezai-Zadeh et al. 2005), it is a GABA antagonist by blocking the GABA(A) receptor (Adachi et al. 2006) and reduces oxidative stress in mitochondria (Valenti et al. 2013), its favorable action on DS might be complex. On the other hand it should be taken in account that EGCG inhibits DHFR (dihydropholate reductase) (Singh et al. 2011), encoded by a non-Hsa21 gene and involved in the metabolism of folate, possibly reducing folate availability. Although greeen tea catechins do not appear to affect plasma folates in a human pilot study (Augustin et al. 2009), DHFR inhibition causes toxicity in trisomic 21 children at half of the dose tolerated by nontrisomic leukemic children (discussed in Lejeune 1988). The use of EGCG in DS remains an interest open field of research, highlighting the possibility that even natural compounds could be suitable treatments, as observed by Lejeune (1988), when able to affect a step demonstrated to be critical in DS pathogenesis.
While this article was in preparation, a potential critical role for DS phenotype for USP16, which is again not present in three copies in the above mentioned child was reported (Adorno et al. 2013).
The careful discussion of the cytogenetic location of each encoded component in the physiopathological models proposed by Lejeune is, still today, a learning experience to be relived. Conclusions drawn from sophisticated mouse or cellular and molecular biology models not taking into account the correlation between the physical location on Hsa21 of a gene (or region) and the probability having a role in the phenotype based on current (and controversial) DSCR maps might be incomplete, if not misleading. Although eukaryote gene expression is finely and individually regulated, on a chromosomal scale it is known since 2000 that the order of the genes on the human chromosomes is not casual and that a higher order of gene regulation exists (Caron et al. 2001, Lercher et al. 2002, Yamashita et al. 2004). Discussing the possible deleterious effects of a given gene overdose, Lejeune noted in 1990 that the metabolic paths controlled by these genes, although unrelated at first glance, are in fact tightly related by their effects, just as if synteny was in some way related to biochemical cooperation or mutually controlled regulation (Lejeune 1990).
At the same time, firm conclusions about the role of a DNA element based on its chromosomal localization can be derived only if the annotated physical map of the chromosome is actually complete and accurate as much as possible. Our analysis by TRAM software (Lenzi et al. 2011) shows that there are hundreds of uncharacterized transcripts on Hsa21 (preliminary data). Although continuing the patient reconstruction of each single locus structure, expression and function appears to have fallen out of research trends in the last 10 years, without a really complete list of Hsa21 elements we could miss the identification of a critical component on an interesting Hsa21 region because it is not highlighted on the map (e.g., ENCODE Project Consortium et al. 2012). As we have shown by identifying a large novel gene extending for more than 100,000 bp on Hsa21, encoding a protein (CYYR1) conserved in vertebrates (Vitale et al. 2002, Vitale et al. 2007) and having gone unrecognized in the putatively complete gene catalogue published with the report of the whole Hsa21 sequence (Hattori et al. 2000), due to plasticity of gene expression the "manual" characterization of individual genomic regions is an ever-going process that should deserve active attention even in the post-genomic era.
In DS patients the main symptom is ID and high-level cognitive functions, a unique feature of humans, are concerned. Lejeune wrote several essays on the nature of men (Lejeune 1970) and on the unique features of the human intelligence (Lejeune 1979b), pointing out that medical science should try to restore this capability in disabled children (Lejeune-Gaymard 2012).
Clinical observation remains at the root of any research effort on a disease, showing what the real problems of the patient are, allowing casual recruitment of atypical cases (e.g., partial trisomies) that may offer the key for the understanding of an elusive mechanism (like patients with chromosomal translocations helped to identify disease genes that mutated more finely in the rest of the population bearing a normal karyotype and point mutations), providing phenotype variability that may be correlated to specific biomolecular imprints (e.g., Dogliotti et al. 2010) and even offering strong human motivation to struggle to help real sick people and not abstract nosographic entities (Jauhar 2006, Verghese 2008). Lejeune in this regard simply stated that eventualities for a cure of ID may sound quite futurist to us today, but if the children are to be helped, the task has to be undertaken (Lejeune 1977).
Even from this point of view, Lejeune's personal history as well as his writings is exemplar. It was clinical observation that convinced him in the '50's that DS was not an infectious disease as commonly believed at the time. In his subsequent articles, he clearly remarked the need for the integration among different biomedical fields, stating: It seems that only when clinical symptoms, gene mapping and biochemical disturbances are correlated, the first possibility of understanding the mechanism of mental deficiency will really be open (Lejeune 1977). When 10 years later the beginning of a more systematic characterization of the human genome appeared not to be impossible, he immediately highlighted the need that the forthcoming effort of DNA sequencing predicted as fundamental for understanding DS pathogenesis should remain tightly linked to the patient: the tedious and laborious comparison of the clinical data and of the DNA deciphering is, currently, the starting point of any pathogenic scheme (Lejeune 1988).
Keeping in mind the main methodological points that we have derived from the knowledge of the scientific thought of Lejeune, we have come to the conclusion that it could be worthwhile to launch an innovative research project on DS aimed at systematically integrating clinical, biochemical, genetic and bioinformatic data obtained in humans in order to identify reliable therapeutic targets for this form of trisomy. The primary outcome of the proposed approach is to gain new knowledge about the genotype-phenotype relationship in DS, focusing on ID, the most frequent and relevant symptom, in order to reduce the list of hundreds of genes located on Hsa21 to a very small number of elements. This would further focus the investigation as much as possible toward a specific search for a targeted substitution therapy capable of shifting the important metabolites back into normal level (Lejeune 1977).
The secondary outcomes will be contributions to a better knowledge of DS and to genotype-relationships correlation for other symptoms. These parts of the study would allow the advanced understanding of structure, expression and function of Hsa21 genes, further helping in delineating a general model of DS. Actually, individuals with DS suffer from an increased risk of several diseases, including congenital heart defects and autoimmune diseases (Roizen and Patterson 2003). The improved survival in DS seen in recent decades has in part been attributable to early detection and treatment of the associated diseases (Yang et al. 2002). To discuss one example, the increased prevalence of CD (celiac disease) in DS has not been explained yet; CD is a multifactorial disease and in the last years several genome-wide association studies have been performed on large cohorts of CD patients in order to identify genes that play a role in the predisposition to the disease (Dubois et al. 2010, Festen et al. 2011, Trynka et al. 2011). Although none of the genes identified up to now is localized on chromosome 21, still about 50% of the CD genetic background is unknown (Ricaño-Ponce and Wijmenga 2013). Moreover the presence of the aneuploidy in DS subjects suggests that the development of CD could, in theory, also be related to a gene-dosage effect in the various tissues. Currently no data are available on gene expression in the duodenal tissue of CD/DS patients, nor on the possible post-transcriptional mechanisms involved in the regulation of these genes.
The project that we introduce here, outlining its essential features, is aimed to build different maps and databases related to Hsa21, and to systematically integrate them to achieve the objective of selecting the very few best promising candidates for conceiving a targeted therapy of ID in DS (Table 2). The maps will be built either by experimental investigation or comprehensive meta-analysis of data available in the biomedical literature and public databases.
|1. Patients||Symptoms||Clinical||>50 features|
|2. DNA||Cytogenetics / CGH-Array||Meta-analysis; Experimental||Blood; Skin|
|3. DNA||Exome||Meta-analysis; Experimental||Blood|
|4. RNA||Hsa21/human genes expression in normal organs||Meta-analysis; Experimental||Brain, Heart, Blood,
|5. RNA||Hsa21/human genes overexpression in DS||Meta-analysis; Experimental||Brain, Heart, Blood, iPS, Intestine|
|6. DNA||Hsa21/human genes metilation in normal organs||Meta-analysis||Brain, Heart, Blood,
|7. DNA||Hsa21/human genes differential methylation in DS||Experimental||Blood|
|8. Protein||Neurological function - Hub function||Computational Biology||/|
|9. Substrate||Metabolomics||Experimental||Urine, Serum, Plasma|
|10. Protein||OMIM phenotypes associated to mutations||Computational Biology||/|
Our interdisciplinary team provides all the needed skills to complete the effort and it is composed of established research groups with expertise documented by publications in international biomedical journals and spanning the fields of: pediatrics, neonatology, neuropsychology, child neuropsychiatry, cytogenetics, molecular cytogenetics, genomics, computational biology and bioinformatics, biostatistics, molecular biology, post-genomics, stem cell biology, chemistry and biochemistry, medical genetics and genomic medicine. The project has been approved by the competent Ethics Committee of Sant'Orsola-Malpighi Hospital in Bologna and it is starting in Fall 2013.
We describe below the 10 Hsa21-oriented maps and databases that are expected to be completed in this 5-year "21-Maps Project" (also internally named "Apollo 21 Project", see Discussion), while the final goal will be the map superimposition and data correlation to identify the very few best molecular targets that are candidates for a rational therapy of DS. Depending on the actual results generated, strong suggestions toward a definite course of action could very well come during the duration of the project, if the scoring of candidates’ genes in the first maps to be produced would converge toward the same targets.
For this project we expect to study in depth, by all or most of the methods described below, 30 subjects with DS, of pediatric age, and 30 normal control subjects of corresponding age and sex, usually chosen between patients' brothers and sisters. Clinical analysis will be extended to the greatest possible number of the 130 children with DS referring to the Unit. All the people involved in the study, approved by the competent Ethics Committee, will have to agree in a written informed consent.
Systematic collection of clinical data will be performed at the Neonatology Unit – S. Orsola-Malpighi Hospital, University of Bologna (Cocchi et al. 2007, Cocchi et al. 2010, Leoncini et al. 2010, Ghezzo et al. 2012, Ghezzo et al. 2013).
Latent class analysis will be used as the main statistical tool for carrying out the results (Farina et al. 2002). The 50 major phenotypic characteristics described in DS (OMIM Entry #190685 at http://www.omim.org/entry/190685, Epstein 1989) will be investigated. The neuropsychological study (cognitive assessment) will be conducted with appropriate tests (Bayley, WPPSI III, WISC-IV) and in the form of interviews with parents (VABS scale, Aberrant Behaviour Checklist, CARS scale, Conners scale).
A database of phenotypes will be created. These data will be correlated with the characteristics of the genome of each individual obtained through Next Generation Sequencing (NGS) methods like exome analysis (exome sequencing), transcriptome analysis (RNA-Seq) and methylome analysis as illustrated below.
In selected cases, where there is a need to clarify an apparent discrepancy between karyotype and phenotype, at S.Orsola-Malpighi Hospital a molecular cytogenetics (CGH-Array) and/or a traditional cytogenetic analysis performed on fibroblasts from skin biopsy will be performed in order to search for possible segmental trisomies or mosaicisms explaining the discordance between karyotype and symptoms (Pagon et al. 1979).
An essential part of the project will be the constant monitoring of new reports describing partial trisomies of Hsa21, shedding light on genotype-phenotype relationships of specific Hsa21 segments. In addition, a meta-analysis by computational biology of all partial trisomy 21 cases ever described will be undertaken.
The analysis of exon nucleotide variations at a genome level (exome analysis) can shed light about the genomic determinants in the phenotypic variability of Down syndrome (Patel et al. 2011, Letourneau and Antonarakis 2012).
Total DNA, obtained from the peripheral blood of subjects with DS, will be submitted to exome (exome sequencing) analysis, using Illumina HiScanSQ system (Illumina, San Diego, CA) available at the Interdepartmental Center for Cancer Research "Giorgio Prodi" (CIRC) of the Bologna University. To avoid genetic privacy issues related to exome sequencing in normal young persons, a panel of DNA samples from anonymous donors acquired commercially (panel HRC-1, Sigma) will also available for the analysis, as well as samples from anonymized healthy volunteers blood donors.
The total DNA extracted will be subjected to enrichment for exonic sequences and the bi-directional exome sequencing (whole-exome Paired-End DNA sequencing) will be performed using an average "coverage" 50×. New single nucleotide variants may be identified through the comparison of all variants found with the most important public databases that provide data on human variability (dbSNP, 1000 Genomes and other similar tools). Galaxy software suite (Goecks et al. 2010) will also be used for data processing. Statistical analysis will be performed to identify significant association between specific variations and phenotypes. In particular, we will first search for mutations associated to DS (by Fisher test), rather than single nucleotide polymorphisms (SNPs), having been reported that trisomy 21 is associated with at least one specific somatic mutation in the GATA1 gene (Khan et al. 2011).
As it has already been underlined, Hsa21 genes that are expressed in tissues affected in DS patients (e.g. brain, heart, tyroid) are of special interest (Dierssen et al. 2001). A fundamental criterion by which filtering the identification of strong candidate genes for the ID phenotype in DS is the expression of the considered gene in the human brain (and possibly a relatively high expression in the brain in comparison with other tissues). We are using a novel computational biology approach that we have recently described for this purpose (Lenzi et al. 2011), able to perform meta-analysis of global gene expression profiles (Ramasamy et al. 2008) aimed at the generation of quantitative transcriptome maps linking point-by-point intensity of transcription to chromosomal map localization. The TRAM (Transcriptome Mapper) data processing algorithm includes an original and powerful method of assignment of each expression data point to a specific locus by integrating different probe/gene nomenclatures (including a parsed UniGene database, Lenzi et al. 2006) as well as flexible functions of data rescaling by a combination of intra- and inter-sample normalization (including scaled quantile normalization, thus allowing integration of data coming from experimental platforms analyzing a different number of genes) (Lenzi et al. 2011, Piovesan et al. 2013). TRAM is able to generate a quantitative model of the mean gene expression level for more than 39,000 human known and uncharacterized loci transcribed in the brain (preliminary data), allowing to direct the attention toward Hsa21 genes especially relevant for whole brain functioning. In addition, TRAM may generate comparative maps showing critical genomic regions and genes with statistically significant differential expressions between two biological conditions that may be used to compare brain vs. all non-brain tissues or brain vs. specific brain regions especially compromised in DS (Dierssen et al. 2009). Although most Hsa21 genes are likely to be expressed in the brain so that absence of expression would exclude a few genes as candidates, we expect that this fine analysis of the human brain normal transcriptome will lead to an accurate prioritization order of all accurately mapped Hsa21 transcripts with regard to their expression level in the brain or in specific brain regions, thus clearly strengthening or weakening the candidacy for responsibility in DS ID emerging from the other areas of this project.
Although TRAM may analyze gene expression profile obtained by RNA-Seq, we will first work with expression microarray data because there are many analyzed samples available and data have been obtained with established techniques, but we will not exclude the use of the RNA-Seq data to further validate our map, each method having its specific advantages and disadvantages (Malone and Oliver 2011). To validate the results, independent confirmation by Real Time reverse transcription polymerase chain reaction (RT-PCR) will be performed at regular intervals throughout the dynamic range of gene expression values.
Transcriptome maps showing differential gene expression between trisomy 21 and euploid samples in different tissues will be performed to gain additional knowledge about the role of Hsa21 gene overexpression in DS. Actually, gene-expression variation of Hsa21 genes in lymphoblastoid and fibroblast cell lines from individuals with DS may be useful to identify statistically significant differences between DS and normal samples for a subgroup of genes that are most likely involved in the constant DS traits (Prandini et al. 2007). In addition, differences in gene expression in blood have been recently found between patients with DS and lower and higher IQ, respectively (Mégarbané et al. 2013).
Total RNA collected from euploid and aneuploid cells, obtained from the peripheral blood, will be extracted and will undergo RNA-Seq using Illumina HiScanSQ system. The characterization of gene expression profile and the research of specific RNA up/down regulated isoforms will be carried out as described (Costa V et al. 2011).
An independent validation of supposedly differentially expressed genes will be performed using Real-time PCR. Individual interesting genes, found this way, will be further characterized through cloning, sequencing, gene expressions and bioinformatics analysis as described before (Bork and Koonin 1998, Strippoli et al. 2000, Vitale et al. 2002, Facchin et al. 2011). A fine characterization of gene expression at a map level is still lacking for trisomy 21 cells, and it will be obtained following analysis of RNA-Seq data by the TRAM (Transcriptome Mapper) approach recently described by us (Lenzi et al. 2011). TRAM software will be used to obtain quantitative transcriptome maps in order to identify significantly over- or under-expressed genomic regions. In particular, the correlation of the intensity of gene expression to its physical location along the chromosome is also a prerequisite for the identification of functional modules in which coexpressed genes are located within the same cluster on eukaryotic chromosome (Yamashita et al. 2004). A fine characterization of gene expression at the level of the map has not yet been conducted in cells with trisomy 21 and will be obtained after the analysis of the data of RNA-Seq obtained from human cells isolated from whole blood.
In selected cases, intestinal biopsies (if medically necessary) will be subject to the investigation of the expression profile of RNA. It will be possible to extract DNA, RNA and proteins from samples for molecular analysis of the expression profile of microRNA (miRNA) of individual proteins through Western blot analysis and expression of mRNA through quantitative real-time reverse transcriptase polymerase chain reaction (RT-PCR). Through NGS technologies, duodenal biopsies obtained from patients with DS and CD, CD only or controls will be subjected to the analysis of the transcriptome as well as to the miRNA profiling in the duodenal biopsies of children with both DS and CD which will be compared with CD patients without DS and controls. Should biopsies be available, we will add a second control group of DS patients who underwent upper endoscopy and in whom CD diagnosis has been excluded. The genes identified as differentially expressed will be further evaluated at the protein level, whereas in the case of miRNAs, possible targets will be identified by in-silico analysis and further analyzed by qPCR and Western blot. These will provide information on alteration of gene expression in the intestinal mucosa of individuals with CD with or without DS, identifying similarities and differences that could help in pointing out the altered pathways involved in CD pathogenesis.
We also intend to make use of an innovative cellular model to study the transcriptome of trisomy 21 neural cells in humans. We will use reprogrammed iPSCs (Yamanaka 2007) derived from CD34+ cells (Berardi et al. 1995, Ciraci et al. 2011, Forte and Berardi 2013) originated from blood and bone marrow of DS patients (n=2). To date these cell types have not yet been used to obtain iPSCs from trisomy 21 tissues. The generation of reprogrammed iPSCs from patients with defined genetic disorders promises avenues to understand the etiologies of complex diseases and the development of novel therapeutic interventions. This approch, together with defined in vitro differentiation protocols, suggests the possibility of developing reliable disease models. The iPSCs will be extensively characterized and stimulated to produce multiple differentiated cell lineages. RNA-Seq and single-gene characterization will be performed on the derived iPS cell lines as illustrated above for the blood samples.
By using established and publicly available databases and maps of DNA methylation in human tissues, we will perform a computational biology meta-analysis of epigenome in main tissues involved in DS, in particular brain tissue, in order to gain knowledge about the methylation state of Hsa21 genes (Table 2). This analysis, in relationship with transcriptome mapping analysis in the same tissues, could provide clues toward a model of the possible role of epigenetic regulation in the genotype-phenotype relationship in DS.
To account for epigenetic factors possibly related to ID (Kuromitsu et al. 1997, Kerkel et al. 2010, Loudin et al. 2011, Sanchez-Mut et al. 2012, Malinge et al. 2013), analysis of methylome (Ku et al. 2011) on DNA extracted from whole blood will be also performed using a commercial service with Agilent SureSelectXT Human Methyl-Seq or RRBS-seq methods (Istituto di genomica Applicata, Udine - Illumina certified service). In particular, epigenetic modifications of Hsa21 genes will be compared in detail between the DS and the control group, using the framework for enrichment in desired features of the TRAM software (Lenzi et al. 2011).
A systematic, accurate revision of Hsa21 gene function will be made using Gene Ontology tools (http://www.geneontology.org/) that provide a way to summarize biological information allowing the identification of biological proecesses associated to the investigated genes (Sun et al. 2006). A comparison of the data will be performed with trisomy 13 and trisomy 18 in order to search for more general relationships among gene function, excess of the relative gene product and specific symptoms.
Specific studies will be performed to study functions of proteins of interest whose genes are located on Hsa21. DS patients could develope early onset Alzheimer’s Disease (AD) pathology including endosomal abnormalities. The triplication of the gene coding for the beta-amyloid precursor protein (APP) is probably associated with increased exosomal secretion pathways. Indeed, intersectin 1 (ITSN1) gene, localized on Hsa21, regulates the formation of clathrin-coated vesicles (Extracellular Vesicles, EVs, Di Noto et al. 2013) and is involved in synaptic vesicle recycling being another interesting factor regulating the EVs export. The emerging role of EVs in intercellular communication can thus be a good tool to be studied for tissue damage in DS. The quantification of plasma EVs and the analysis of their protein and miRNA content is a quite helpful tool. Furthermore, the enzyme content of EVs has never been considered and it will be of extreme interest.
Although it is presumed, and in some cases documented, that some metabolic pathways are altered in trisomic cells following the imbalance of the enzymatic dosage related to the presence of an additional copy of Hsa21 (Patterson 2009, Gardiner et al. 2010), to date no systematic study has been conducted in this field. It is reasonable to assume that this type of investigation could deliver useful results to define the alteration of critical steps in the metabolism of trisomic individuals, providing ideas for possible therapeutic strategies. Results from biochemical analysis can now be related to the position on the genome map, and in particular on the Hsa21 map, of the enzymes catalyzing the reactions in which they are involved. Rather than individual substrates, it is now possible to study the metabolome, the complete set of small-molecule metabolites present in a biological cell, tissue, organ, or biofluid and it includes sugars, nucleosides, organic acids, ketones, aldehydes, amines, amino acids, lipids, steroids and alkaloids. Metabolomics, in its most ambitious global form, tries to comprehensively analyze all known and unknown metabolites in a given biological sample (Dunn et al. 2005). Targeted metabolomics have the more limited goal of quantitating selected metabolites, most typically dozens of known compounds.
Although gene/protein expression events provide useful clues to metabolic dysfunctions, it is evident that many factors like post-translational modifications, protein inhibitors and alternative gene functions, also produce important metabolic changes. Therefore, metabolomic investigations correlated with transcriptomic and proteomic studies are essential to complete a systematic biochemical understanding of phenotypes.
In the frame of the 21-Maps Project the metabolome analysis of urine, and possibly blood serum or plasma samples, derived from the same cohort of DS and control subjects recruited for genomic mapping, will be performed. The first goal of the project will be the untargeted analysis of urine samples by nuclear magnetic resonance (NMR) spectroscopy. The NMR technology represents the most suitable platform to simultaneously identify and quantify a large number of metabolites in biofluids without the need for sample manipulation (Dunn et al. 2005). Raw data, consisting of monodimensional proton NMR (1D-1H NMR) spectra, will be processed (van den Berg et al. 2006) and submitted to multivariate statistical analysis (starting from Principal Component Analysis, PCA) to highlight the most relevant variables characterizing the DS condition. The metabolites identification will take advantage of two-dimensional (2D) NMR techniques, such as 2D 1H TOCSY (total correlation spectroscopy) and 1H-13C HSQC (heteronuclear single quantum coherence spectroscopy) and will rely on database query (BMRB - Biological Magnetic Resonance Bank).
The second phase of the project will concern a more targeted investigation on the biochemical pathways that will eventually be signaled by the parallel genomic Hsa21 maps. In this phase Mass Spectrometry (MS) will likely be the eligible approach (Lu et al. 2008): in fact while the complementary NMR technique has a relatively low sensitivity, the MS analysis, if combined with effective sample preparation and chromatographic separation, can reach a very high sensitivity and specificity for a class of compounds, as well as good dynamic range.
The experimental procedures for biofluid (urine, serum/plasma) collection and NMR/MS analysis will be based on the standardized protocols reported in literature (Dunn et al. 2011, Beckonert et al. 2007), and drawn from our own experience (Mimmi et al. 2011, Mimmi et al. 2013).
Although the effects of an excess protein production are different from those of point mutations altering the amino acid sequence of the gene product, the analysis of the phenotypic effect of Hsa21 genes when mutated may help to identify a possible role in ID. While a similar concept has been explored focusing on pathways involved (Sturgeon et al. 2012) providing data that can be included in our model, we plan to develop an automated pipeline especially oriented to the parsing of OMIM database (http://www.ncbi.nlm.nih.gov/omim) and able to correlate and model the variables: symptom, mutated gene, gene expression level in the brain (obtained from the map #4 above) and gene ontology (from Gene Ontology function database).
While research using mouse models is not a primary objective of the present project, we would note that our 10-Maps route is easily translatable in an analogous study in DS model mice. It is possible that some of these maps, in particular those whose generation would be feasible by meta-analysis of available data, will be created in the context of this study when considered useful. Observing that an additional mouse/human gene homology map would be valuable in this case, the total of maps/databases possibly stemming by our project concept would be of 10+10+1=21, providentially reinforcing the choice of the name 21-Maps for the protocol, originally named after trisomy 21.
The primary expected outcome of this research project is the identification of a restricted list of strong candidate genes and mechanisms for ID in persons with DS with a quantitative prioritization based on a scoring system. This system will be developed summing the rank of each gene in each map when genes are listed in an order reflecting their decreasing probability to be critical for the ID phenotype in DS. This way, the expected results will allow more consistent and systematic criteria to associate single genes to the DS ID phenotype. Some of these criteria for defining contribution to the DS phenotype have already been pointed out: (1) expression in the normal brain; (2) overexpression in DS; (3) participation in neurobiological processes (Belichenko et al. 2007, reprised by de la Torre and Dierssen 2012). These criteria will be systematically taken into account according to points #4, #5 and #8 illustrated in the section above, in particular by defining a normal reference quantitative transcriptome map for the human brain at an unprecedented level by comprehensive meta-analysis of all available pertinent gene expression profiles. In addition, we also believe that the proposed addition of criteria such as cytogenetical localization (by excluding genes present in subjects with partial trisomy 21 and clinical DS phenotype and systematically searching for microdeletions in patiens with dissociation between karyotype and symptoms), DNA sequence or methylation variants association with specific clinical features investigated by neuropsychological tests, and metabolite profiling to identify specifically altered enzymatic pathways may represent a critical advantage for a more affordable definition of candidate genes.
Main secondary outcomes will be a better knowledge of the disease with the possibility to reveal facts of clinical interest, an advancement in genotype-phenotype relationship (McCabe and McCabe 2013) for symptoms other than ID, and an overall better knowledge of the structure and function of Hsa21.
A careful policy must be undertaken about the diffusion of news about "cures" for DS. It should be kept in mind that it affects an estimated 5.8 million persons in the world (Roizen and Patterson 2003) and that most claims of new therapies prospected only on the basis of in vitro results or theoretical models are de facto never translated into effective therapies (Frances 2013). We believe that, even if the present study will not end in the discovery of effective treatments, it is unlikely that the basic, systematic work proposed here will not generate useful contributions in the field. In the best case, it will be a possible start for subsequent projects focused on the feasibility, safety and effectiveness of strategies aimed to correct the alteration of the identified targets, strategies which could be very different depending on the nature of the precise molecular function needing modulation toward normality. Unexpected obstacles or fortuitous discoveries are also possible slowing or accelerating, respectively, the pace of the investigation, but, in general, the prudent pace we are faced with seems to us a realistic alternative to giving up as well as to giving false hopes.
Walking requires at least a positive hypothesis about the existence of the destination and is best done with a good guide. We have shown here how powerful the scientific thought of Prof. Lejeune has been in pursuing and suggesting multiple routes toward the final goal of the cure of a human disease. A fundamental part of our study, able to help connect its various sections, will be the reprise and the study of scientific articles by Jérôme Lejeune because, apart from their contribution to a solid methodological approach as we have pointed out above, they also enter into the merits of specific mechanism and pathways possibly involved in DS pathogenesis (for example Lejeune 1977, Lejeune 1988). Using knowledge and tools available at the time, he was able to formulate specific hypotheses that in many cases appear not to have yet been verified or falsified, such as the seven metabolic pathways he discusses in his 1988 review and may constitute valuable working hypotheses to guide the interpretation of the much more complete and sophisticated data set that we can create or we already have available today. Lejeune believed that, due to basic principles at the roots of genetic diseases, studying trisomy 21 could have general consequences: I could spend years discovering the genetic causes of many illnesses, and I could keep on studying even rarer diseases. But I am convinced that everything is interrelated. If I find out how to cure trisomy 21, than that would clear the way for curing all the other diseases that have a genetic origin. The patients are waiting for me; I have to find it (Lejeune-Gaymard 2012).
We would note that Lejeune used an interesting expression for defining the enterprise of discovering a cure for trisomy 21: an intellectual challenge (Lejeune-Gaymard 2012). It is a formidable challenge (the task is immense) (Lejeune "21-Thoughts"), because any pathogenetic theory for DS would be realistic only when it will be able to make sound connections among the different pathogenetic concept accumulated in time: for example, the altered development of the brain and often of the heart, the alteration of metabolic pathways, in particular those implied in the redox state and the folate cycle, the link with AD, the low prevalence of solid cancer, and the localization of primary genes starting the pathogenetic process on Hsa21. We have perhaps to recover the attitude to gather together findings of distant disciplines, an art in which Lejeune was a master, and look at all the available data with new eyes.
We would conclude this article returning to the fundamental concept that a positive hypothesis is what drives the researcher. A convinced comparison between going to the Moon and finding a cure for DS was recurrent in the talks of Lejeune: We will beat this disease. It's inconceivable that we won't. It will take much less intellectual effort than sending a man to the Moon (Lejeune-Gaymard 2012). Landing men on the Moon (Apollo Project, 1959-1972) remains a milestone of the twentieth century and it is still today a source of inspiration when trying something that seems impossible (Chaikin 1998). It is known that the Apollo Project involved the work of more than 400,000 persons with a practically unlimited budget, which was up to 5.3% of the USA total federal budget in 1965 and in the end summed up to a total of nearly $20 billion at that time (National Aeronautics and Space Administration 2004). It is difficult to estimate how many researchers are actively dedicated to the research aimed at finding a therapy for ID in DS, but from gross analysis of the literature and review of speakers in dedicated congresses it is unlikely that there are more than a few hundred dispersed all over the world. Although there has recently been an unexpected reprise of interest for doing and supporting DS research (see publication dates and acknowledgments for funding in references from Table 1), as shown above we are often working with limited resources in comparison with those available in other biomedical fields, so there is the urgent need to find adequate support for these studies. The present project has a roughly estimated total cost of 75,000 Euros/year for five years, plus about as much needed for supporting the dedication of young fellows for the project. While we are applying to traditional research funding agencies and institutions for grants on the subject, we are also starting systematic fund raising initiatives, like the one allowed through this journal, whose success will be critical for the feasibility of the project we have described here.
To use the words of Bruno, the trisomic man mentioned in our Introduction (and aware that Lejeune "discovered trisomy with me! He used my karyotype"): "The scientific research must be continued. I am glad to know that advances are continuing. Professor Lejeune was a friend of those who were handicapped" (Association Les Amis du Professeur Jérôme Lejeune at http://www.amislejeune.org/).
We would like to give special thanks to the members of the Research Office at DIMES, Drs. Claudio Negro, Elisa Bettini and Monica Fiori, as well as to the Director, Prof. Davide Treré, and the Chief Administrative Manager, Dr. Luisa Romagnoli, for their very kind and strong support for this research as regards to the administrative and strategic aspects. Authors from DIMES Dept. are sincerely grateful to Anna Fusina, who generously gave very useful advice and made extraordinary personal efforts related to support for our research. PS is indebted to Drs. Marcello Villanova, Giampaolo Ugolini and Paola de Sanctis for their constructive discussion and very useful advice about critical and methodological aspects of this research. A very special thanks from PS to Dr. Ombretta Salvucci (NIH, Bethesda, USA) for her incomparable and friendly support in discussing Lejeune's life and ideas and encouragement in pursuing the present project. This work is dedicated to Professor Jérôme Lejeune (1926-1994), who taught us to hate the disease and to love the patient, and to his fellow and teacher to some of us Professor Maria Zannotti, who brought back research on DS in Bologna in the late '60's and is still advising us on the subject.
All authors contributed to define the concepts that we present, each according to his/her specialty; they all contributed to draft the manuscript and approved the final version.