genetica s3 sindrom down 1

8/6/2019 Genetica S3 Sindrom Down 1

1/6

A YEAR OF UNPRECEDENTED PROGRESS IN

DOWN SYNDROME BASIC RESEARCH

Roger H. Reeves1,2* and Craig C. Garner

3

1Department of Physiology, Johns Hopkins University School of Medicine, Baltimore, Maryland2Institute of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland

3Department of Psychiatry and Behavioral Sciences, Nancy Pritzker Laboratory, Stanford University, Palo Alto, California

The years 2006 and 2007 saw the publication of three new anddifferent approaches to prevention or amelioration of Down syndromeeffects on the brain and cognition. We describe the animal model sys-

tems that were critical to this progress, review these independent break-through studies, and discuss the implications for therapeutic approachessuggested by each. '2007 Wiley-Liss, Inc.MRDD Research Reviews 2007;13:215220.

Key Words: mouse model; cognitive testing; GABA inhibitor; APP; sonic

hedgehog; synapse function; brain development

ANIMAL MODELS OF DOWN SYNDROME

F

eatures of Down syndrome (DS) arise due to overex-pression of genes on human chromosome 21 (Hsa21)usually through the triplication of the entire chromo-

some. Divergence from euploid of the structures and functionsthat are affected in DS may result from upregulation of Hsa21genes in adult cells and tissues and/or because of inappropriatecues during development that are perpetuated in descendentsof the cells that are affected initially [Potier et al., 2006; Roperand Reeves, 2006]. The central nervous system begins to de-velop very early in embryogenesis and substantial structuraland functional changes are still evident for several years fol-lowing birth, and to some degree, throughout life. Thus, thebrain is especially susceptible to perturbations caused by subtlechanges in gene expression. Small structural differences canresult in a substantial degree of dysfunction at the level ofindividual neurons, which will affect the complex circuitsinvolved in cognitive function.

Development cannot be studied in human beings at thelevel necessary to understand the complex processes that areperturbed by trisomy. In vitro molecular biology and cell cul-ture systems are important tools but do not replicate the de-velopmental processes that are the targets for intervention toameliorate DS features. Animal models are essential to studythe disruption of the developmental process caused by trisomy.

The laboratory mouse provides the best-characterized ex-perimental system for studies of mammalian aneuploidy. Thefirst line of evidence supporting the use of mouse models tostudy DS is genetic. Detailed comparative and physical mappingand now comparative sequence analysis [Hattori et al., 2000;Mural et al., 2002; Toyoda et al., 2002] demonstrate highly

conserved linkage with Hsa21 on mouse chromosomes 16(Mmu16), 17, and 10. Ts65Dn mice are trisomic for nearly halfof the mouse equivalents of Hsa21 genes, whereas other modelscontain subsets of these (Fig. 1). The elevated gene expressiondue to trisomy is very comparable between these mice andhuman beings. But will this comparable genetic effect have thesame outcome on the features of the mouse, i.e., will analogousphenotypes be produced in the same way?

For many features of DS, there is strong evidence to vali-date the idea that developmental processes affected by trisomy21 in humans can be studied in mice. Quantitative phenotypeassessments in Ts65Dn mice demonstrate directly comparabledysmorphologies in craniofacial structure, in which the samebones affected in DS are affected in the same way in mice[Richtsmeier et al., 2000, 2002]. Histopathological analysis ofthe cerebellum in Ts65Dn mice predicted a pattern of cell loss

that was subsequently shown to occur in humans with DS[Baxter et al., 2000]. Age-related loss of forebrain cholinergicneurons occurs in Ts65Dn mice and in DS as well as in Alzhei-mer disease. Deficits in retrograde transport of NGF are respon-sible for this loss in mice [Cooper et al., 2001] and suggest atherapeutic approach to this problem in humans. Detailed stud-ies of synaptic dysfunction in mouse models point towards spe-cific neuronal pathologies in DS [Belichenko et al., 2004].

Trisomic mouse models also provide a good representa-tion of important aspects of cognitive dysfunction in DS.Work by Nadel and coworkers has demonstrated that func-tions of the hippocampus are among those most severelyaffected in DS [Nadel, 2003; Pennington et al., 2003]. A ro-bust learning and memory deficit in tests requiring hippocam-

pal function has been demonstrated in Ts65Dn mice [Escori-huela et al., 1995; Reeves et al., 1995; Holtzman et al., 1996].Hippocampal dysfunction extends to studies of the electro-

Grant sponsors: Anna and John J. Sie Foundation; The National Heart, Lung andBlood Institute; National Institute of Child Health and Human Development;National Cancer Institute.*Correspondence to: Roger H. Reeves, Ph.D., Department of Physiology, JohnsHopkins University School of Medicine, 725 N. Wolfe St., Baltimore, MD21205. E-mail: [email protected] 10 August 2007; Accepted 13 August 2007Published online in Wiley InterScience (www.interscience.wiley.com).DOI: 10.1002/mrdd.20165

MENTAL RETARDATION AND DEVELOPMENTAL DISABILITIES

RESEARCH REVIEWS 13: 215220 (2007)

' 2007 Wiley-Liss, Inc.


2/6

physiological properties of neuronal cir-cuits in the hippocampus of trisomicmice [Siarey et al., 1997; Kleschevnikovet al., 2004].

Trisomic mouse models havebecome an essential tool in research intothe basic mechanisms of gene dosageeffects in DS. The highly parallel out-comes that result when the same evolu-tionarily conserved genetic programs

are perturbed in mice and man validatestudies that focus on phenotypes. Thethree studies discussed here rely onmouse genetic models to explore when,where, and how trisomic developmentand function deviate from euploid, andeach suggests possible therapeutic app-roaches to overcome these differences.

To illustrate the application ofgenetic models and introduce tests ofhippocampal function in trisomic mice,we review a recent study analyzing thegenetic contribution to hippocampal-based cognitive dysfunction measured in

the Morris maze deficit [Olson et al.,2007]. We then review the threebreakthrough studies. The study ofSalehi et al. (2006) uses a genetics-basedapproach to identify a candidate targetgene, while the other two studies reachtheir conclusions by focusing ondescription of pathogenesis and modera-tion of phenotypes without lookinginto the genetic basis for the origins ofthe phenotype under consideration. Allthree reach their important conclusionsby careful attention to detailed and

quantitative analysis of phenotypic out-comes in these models.

THE CONTRIBUTION OFCRITICAL REGIONS ONHsa21 TO PHENOTYPES OF DS

Efforts to identify regions ofHsa21 containing genes that contributeto specific aspects of the DS phenotypehave focused on individuals with trans-

locations resulting in trisomy for a sub-set of Hsa21 genes (translocation DS).Phenotype maps correlating dosageimbalance of specific regions with spe-cific characteristics provide useful infor-mation about segments in which tosearch for genes that perturb specificdevelopmental pathways [Delabar et al.,1993; Korenberg et al., 1994].

The resolution and interpretationof the critical regions identified byphenotype maps is severely limited byseveral factors. Notably, there is no col-lection of human beings who are triso-

mic for only the critical region, whereasassessment of variable human pheno-types requires assessment of many indi-viduals to draw meaningful conclusions.To address this issue directly, Olsonet al. [2004] used chromosome engi-neering of Mmu16 to create mice witha duplication or a deletion of the so-called DS critical region (DSCR) (Fig.1) and asked whether the 33 DSCRgenes were necessary and/or sufficientto produce DS features attributed togenes in that region. Several DSCR

features have parallels that have beendocumented in trisomic Ts65Dn miceincluding short stature; anomalies of thecraniofacial skeleton and skull that resultin flat facies, brachycephaly and pro-truding tongue (small mandible); and(with caveats) mental retardation of theDS type, specifically impaired hippo-campal function. Ts1Rhr mice, whichare trisomic only for the DSCR,

showed none of the features of the cra-niofacial skeleton attributed to theDSCR. Ms1Rhr/Ts65Dn mice are tri-somic for all the genes triplicated inTs65Dn except those in the DSCR (Fig.1). These mice had craniofacial featuressimilar to Ts65Dn. Thus, trisomy forthe DSCR is not sufficient and for themost part is not necessary to producethese features of trisomy.

The picture is more complexwhen the question of mental retardationof the DS type is addressed [Olson et al.,2007]. Disruption of hippocampal func-

tions in DS can be modeled in the per-formance of Ts65Dn mice in the Morriswater maze (MWM) paradigm. InMWM, a mouse placed in a swimmingtank is trained to navigate to a platformbased on visual cues in the room (Fig.2a), a visiospatial integration task that isdependent on hippocampal function.Ts65Dn mice are significantly impairedin the MWM test, whereas Ts1Rhrmice that are trisomic for only theDSCR genes performed the same aseuploid in both phases of the test [Olsonet al., 2007]. Thus, trisomy for the

DSCR is not sufficient to produce thekind of hippocampal disruption seen inTs65Dn mice and in DS, contrary to theprediction of the DSCR hypothesis.Next, Ts65Dn mice were crossed toMs1Rhr mice, returning the DSCRgenes to the normal two copies in amouse that was trisomic for all of theother genes triplicated in Ts65Dn.These Ms1Rhr/Ts65Dn mice per-formed like euploid (Fig. 2b), showingthat a gene (or genes) from this segmentis necessary to produce this specific hip-pocampal deficit, and acts in combina-

tion with other genes from Mmu16 (andHsa21) to do so.

GABAA ANTAGONISTS ANDLONG-TERM CORRECTION OFLEARNING AND MEMORY[FERNANDEZ ET AL., 2007]

The previous experiments de-scribe functional tests at the organismallevel that demonstrate the effects of tris-omy. The deficit in MWM performancein Ts65Dn mice is paralleled by achange in the way that neurons com-

Fig. 1. Aneuploid segments in mouse models used to study Down syndrome. Gene numbersfor each segment are based on conserved genes identified by Gardiner et al. [2003].

216 MRDD Research Reviews DOI 10.1002/mrdd

TOWARDS THERAPY FOR DOWN SYNDROME

REEVES


3/6

municate. These changes in synapticplasticity are measured by sensitive elec-trophysiological tests as changes in long-

term potentiation (LTP) and long-termdepression (LTD), changes that are ro-bust in Ts65Dn mice [Sago et al., 1998;Siarey et al., 1999, 2006; Kleschevnikovet al., 2004]. In fact, Ts65Dn micecan show LTP in the dentate gyrus butare normally inhibited from inducing itdue to excessive inhibitory input [Kle-schevnikov et al., 2004]. Studies thatexamined excitatory and inhibitoryinputs to various parts of the hippocam-pus found that inhibitory inputs weresomewhat more prevalent and moreactive in trisomic than in euploid mice

[Hanson et al., 2007]. These inhibitoryneurons are characterized by their useof a neurotransmitter called GABA.Use of GABAA receptor antagonist inin vitro studies results in enhancement ofthe development of normal LTP in thedentate gyrus of Ts65Dn brain slices.

These observations led Dr. CraigGarner and coworkers to examine thepossibility that administration of non-competitive GABAA receptor antagonistscould have a positive effect in mice [Fer-nandez et al., 2007]. We consideredthree antagonists of GABA, compounds

that counter the effects of this neuro-transmitter by blocking its receptor.The first compound is called picrotoxin(PTX) and is widely used in experi-mental paradigms to induce seizures.(Indeed, all three of these molecules cancause seizures, discussed further below.)PTX has a very small window betweenthe amount required to show a positiveeffect in LTP experiments and theamount required to induce seizures, andthus is an unlikely candidate as a drug.However, its effects and mode of action

are very well studied and thus the resultshere can be related to a wealth of addi-tional findings. The second compound is

bilobalide (BB), one of several com-pounds extracted from the leaves of theginkgo biloba tree. A variety of effects,many conflicting, have been reported forthis compound, but the most remarkableis as a noncompetitive antagonist ofGABAA receptors. The third com-pound, pentylenetetrazole (PTZ), is alsoa noncompetitive GABAA receptor an-tagonist. It was administered to patientsfor many years as a circulatory and respi-ratory stimulant [Levy, 1953]. It has alsobeen used ineffectively, at higher dosesthat cause convulsion, for the treatment

of psychiatric disorders such as schizo-phrenia and at low nonepilepitic doses asa procognitive drug for the treatment ofsenility in geriatric patients. Given thatit was ineffective for the treatment of ei-ther of these disorders and is potentiallyharmful at high doses, approval for PTZwas revoked by the FDA in 1982. Thefact remains that large numbers of peo-ple were treated with this compound for

years, and although the drug showed lit-tle efficacy, it also showed few sideeffects in most people at low doses.Given this history, of the three com-

pounds tested in this study, PTZ is themost intriguing to consider as a drug.

To test these compounds, theinvestigators turned to the Ts65Dnmouse model since hippocampal impair-ments in these mice parallel those in DS.They used two nonstressful tests of hip-pocampal function. The first is calledObject recognition. The mouse isplaced in a cage with two objects for15 min, during which time it will spendabout half the time examining eachobject. The next day, the mouse is placed

back in the cage with one of the sameobjects and a second, new object.Euploid (normal) mice will rememberthe old object and spend most of theirtime examining the new one, but triso-mic mice show no evidence that theyrecognize the first object. The secondtest is an Alternating T maze, a long

platform with two arms. In multiple tri-als, a mouse with normal hippocampalfunction comes to the T, remembersthe arm that it investigated in the pre-ceding trial, and goes down the otherarm. Mice in which the hippocampus isimpaired, including Ts65Dn mice, choosethe arms randomly, suggesting that theydo not remember the preceding trial.

In the first series of experiments,Ts65Dn mice given BB or PTX atdoses far below those used to induceseizures showed marked improvementcompared to saline-treated trisomic mice,

performing the same as their euploid lit-termates in both the object recognitionand T-maze tests. The authors observedthat if animals were injected with PTXfor 2 weeks followed by saline for2 weeks, they performed just as well asanimals that had just finished the drugtreatment.

Based on these results, a series ofexposures to PTZ was designed. Thisdrug can be given orally, so euploid andtrisomic mice were trained to drinkchocolate milk with or without thedrug daily for 3 weeks. Two to three

months after the animals stopped get-ting the drug, they were tested in theobject recognition test. Trisomic micethat got PTZ performed as well aseuploid (Fig. 3). This long-term im-provement was also evident in a sensi-tive electrophysiological test of LTP inthe dentate gyrus, a measure of synapticplasticity that is normally lacking inTs65Dn mice. Trisomic mice that hadreceived PTZ had significantly im-proved LTP relative to saline-treatedTs65Dn mice [Fernandez et al., 2007].

Fig. 2. Testing hippocampal function in mice. (a) The Morris water maze tests the ability of amouse to navigate to a platform (black rectangle) based on visual cues placed around the room.(b) Ts65Dn mice perform poorly in the MWM (final trials 7, 8, and 9), but when the so-calledDSCR is genetically subtracted in Ms1Rhr/Ts65Dn mice, they perform like euploid [adaptedfrom Olson et al., 2007].

Fig. 3. Ts65Dn mice are impaired relativeto euploid in the hippocampal-based objectrecognition task (Milk), but treatment withPTZ allows them to perform normally, evenseveral months after the end of treatment[adapted from Fernandez et al., 2007].

MRDD Research Reviews DOI 10.1002/mrdd


REEVES 217


4/6

Serious questions remain to beanswered before clinical trials can beconsidered. The first is the phenom-enon known as kindling, in whichrepeated medium doses of chemicalsincluding all three of those used inthese experiments (2030 mg/kg)make animals more susceptible to seiz-ures. It remains to be seen whether thefar lower doses used here (13 mg/kg),

which fail to elicit seizures in euploidrats and mice, will increase seizure riskin trisomic mice. A related question iswhether there is a genetic predispositionto seizures that would make the use ofthis drug more dangerous in some peo-ple than in others and/or whether

young children are more susceptiblethan young adults. A second importantgoal is the elucidation of the mecha-nisms by which short-term treatment ofnoncompetitive GABAA receptor antag-onists leads to a long-term improvementin function. This is not only important

for the direct consideration of safety ofthese drugs, but is important to choos-ing (or designing) safer and more effec-tive treatment strategies in the future.Even given these cautions, this studyrepresents an extremely exciting break-through on which to focus researchefforts in the near term.

Conclusion Study 1Decreased hippocampal perform-

ance in mouse models of DS is linkedto an imbalance in the weighting ofexcitatory and inhibitory circuits. At

present, the underlying etiology of thisimbalance is unknown. Administrationof low, nonepileptic doses of noncom-petitive GABAA antagonists can restorenormal cognitive function in Ts65Dnmice. This information suggests a newand exciting avenue for the possibletreatment of cognitive dysfunction inDown syndrome. While serious ques-tions remain, the questions are welldefined and there is every reason tothink they can be answered. WhetherPTZ ends up as a drug of choice, theseexperiments indicate that drugs directedtoward GABA pathways that presum-ably lower inhibitory inputs in the tri-somic hippocampus hold great promisefor improvement of cognition in DS.

DEGENERATION OF BASALFOREBRAIN CHOLINERGICNEURONS [SALEHI ET AL., 2006]

The second major breakthrough ofthe past year was established by Dr. Wil-liam Mobleys group at Stanford [Salehiet al., 2006]. The basis for this discovery

is described in depth elsewhere in thisissue and is reviewed briefly here to putit in context with the other two studies.Mobley and coworkers have used mousemodels extensively to understand howtrisomy causes changes in structure andfunction of neuronal synapses, includingsome key observations about inhibitoryinput to the hippocampus as discussedearlier [Kleschevnikov et al., 2004]. They

have also studied the loss of a specificpopulation of neurons that communicatesbetween the basal forebrain and the hip-pocampus. These basal forebrain cholin-ergic neurons (BFCN) utilize acetylcho-line as a neurotransmitter and are seen todegenerate in Alzheimer disease (AD), inDS, and in Ts65Dn mice. (In fact, thestudy by Salehi et al. presents evidencesuggesting that rather than degenerating,these neurons actually become quiescent,turning off expression of the markercommonly used to identify them.) ThisBFCN quiescence is believed to contrib-

ute to the age-related decline in cogni-tive function that is seen in many indi-viduals with DS.

Like many neurons, the BFCNrequire specific signals to remain activeand alive. Several years ago, Mobley andcoworkers [Cooper et al., 2001] usedTs65Dn trisomic mice to demonstratethat neurons do not properly transportan important signal, nerve growth factor(NGF), from the nerve terminals in thehippocampus back to the neuronal cellbody. It was known that this retrogradetransport of NGF was a signal that

maintains BFCN, and so the absence ofthis transport in the trisomic neuronswas considered likely to be the cause oftheir demise.

At this point, the incomplete butvery powerful set of mouse models forDS and also for AD played a prominentrole. Ts1Cje mice [Sago et al., 1998]are trisomic for about 80% of thegenes triplicated in Ts65Dn (Fig. 1).When Mobleys g roup compared BFCN-related neuropathology in euploid,Ts65Dn, and Ts1Cje mice, they foundthat retrograde transport of NGF was

much closer to that of euploid mice inthe Ts1Cje mice (Fig. 4a). Accordingly,attention was focused on the genes thatare present in three copies in Ts65Dnbut not Ts1Cje. Several of these haveknown functions in the CNS, includingGabpa, Grik1, App, and Sod1. Of these,APP plays a major role in AD, as shownby the fact that a cleavage product ofthe APP protein is a major constituentof the neuritic plaques that characterizethe histopathology of this disease. Fur-ther, mutations in APP are associated

with inherited forms of AD with veryearly age of onset.

To determine whether dosage (3copies vs. 2 copies) of App contributed

to the NGF retrograde transport deficitin Ts65Dn mice, a genetic cross wasmade between Ts65Dn, which havethree copies of App (Ts65Dn:App//), and mice with only one copy ofApp (App/2) and the offspring wereanalyzed. The severe reduction in trans-port of Ts65Dn mice was substantiallyimproved in the trisomic mice that hadtwo instead of three copies of App(Ts65Dn:App//2, Fig. 4b). Fur-thermore, mice that were euploidexcept for extra copies of the APP genewere affected. Careful experiments es-

tablished that an APP cleavage productfrom the COOH-terminal end of theprotein was responsible for the retro-grade transport deficit, possibly throughinteraction at the level of the earlyendosomes that carry NGF.

Conclusion, Study 2Restoration of NGF transport and

survival of BFCN might counteract, tosome degree, the age-related cognitivedecline seen frequently in DS. The

Fig. 4. BFCN must transport NGF pro-duced in the hippocampus back to the cell

body in the basal forebrain to remainactive. (a) Retrograde transport of NGF isgreatly impaired in Ts65Dn mice, butaffected to a lesser degree in Ts1Cje mice.(b) Genetic subtraction of App from thetrisomic gene set in Ts65Dn substantiallyimproves retrograde NGF transport(Ts65Dn:App//2 vs. Ts65Dn:App//);*indicates a significant difference [adaptedfrom Salehi et al., 2006].



REEVES


5/6

pharmaceutical industry is not likely tomake a large investment in DS becausethe number of affected individuals istoo small to return a profit in their eco-nomic model. In that light, the associa-tion of APP C-terminal fragments withthe pathology that is shared in AD andDS is fortuitous, because extensiveefforts are already focused on APPprocessing as a possible intervention for

AD. The indications in this study ofAPP interference with retrograde trans-port via early endosomes suggest thatreducing levels of APP could have apositive effect in reducing BFCNdegeneration, with beneficial conse-quences in DS and AD.

REVERSAL OF A GROWTHFACTOR RESPONSE DEFICITAFFECTING BRAINDEVELOPMENT [ROPERET AL., 2006]

The third breakthrough studycomes from Dr. Reeves laboratory[Roper et al., 2006]. It appears likely tobe farther removed from potential clini-cal application, but may have morewide-reaching effects regarding thebroad range of features that characterizeDS. This study also demonstrates theability to reverse one very specific aspectof the phenotype using a potential drug.In the course of fundamental researchon the development of the trisomicbrain, these studies identified a deficit inresponse to a growth factor that affects

many different tissues at different stagesof development. We went on to demon-strate that administration of an experi-mental compound could return a spe-cific aspect of the developmental processto normal in the cerebellum. As in thepreceding studies, these findings reliedon genetic mouse models of DS.

The experiments started several years ago with efforts to compare braindevelopment in trisomic Ts65Dn miceto that in human beings with DS. Thecerebellum is significantly reduced inDS, even relative to the overall smaller

size of the brain [Crome et al., 1966;Aylward et al., 1997]. Unlike many fea-tures of DS which may or may not bepresent in an individual with trisomy21, a small cerebellum appears to occurin everyone with this condition. Thus,it was expected that a mouse with asimilar genetic constitution should dis-play corresponding effects on develop-ment of the cerebellum.

Baxter et al. [2000] used high reso-lution 3d MRI to show that while thebrain was the same size as euploid in

Ts65Dn mice, the cerebellum was dis-proportionately small. (The fact that thebrain was not reduced overall in Ts65Dn

was not terribly surprising as much of thereduction seen in DS occurs in the cere-bral cortex, which represents a muchlarger proportion of the brain in humanbeings than in other mammals.) Giventhe reduced Ts65Dn cerebellum, histo-logical approaches were used to deter-mine where the reduction occurred.Both the granule layer and molecularlayer were affected. Unexpectedly, Baxteret al. demonstrated that not only was thevolume occupied by granule cells andPurkinje cells reduced, but in additionthere was a reduction in the cell density

of both types of neurons, further reduc-ing the total number of cerebellar neu-rons. This phenotype had never beenreported in DS. Using age-matched brainspecimens from individuals with two orthree copies of Hsa21, the same reduc-tion of granule cell density was seen inDS.

The ability of a mouse model tocorrectly predict a new phenotype of ahuman condition is a strong indicationthat the developmental processes are thesame in the two species. Accordingly,we next asked at what point the pattern

of development diverged between triso-mic and euploid mice [Roper et al.,2006]. Again using histological app-roaches, we found that every parameterthat could be measured 1 week afterbirth (postnatal day 6, P6) was reducedin Ts65Dn mice, but these differenceswere not evident at P0. Very precisestatistically validated sampling of cere-bellum revealed that while the numberof granule cell precursors (GCP) thatgive rise to granule cell neurons was thesame in Ts65Dn and euploid litter-

mates, the number of GCP undergoingcell division was significantly reduced intrisomic mice.

Collectively, these findings identi-fied a specific effect of trisomy on braingrowth and defined the developmentaltiming (P0), the affected cell type(GCP), and the disrupted process (celldivision). Studies from a number of lab-oratories had established that thegrowth factor, sonic hedgehog (SHH),is the major signal to induce cell divi-sion of GCP. Crosses between Ts65Dnand mice with a reporter of SHH path-way activity suggested that this path-way was somewhat downregulated inTs65Dn mice. Accordingly, we isolated

GCP and cultured them in the presenceof increasing amounts of duly-lipidatedSHH (Fig. 5a). This analysis showedtwo important things. First, trisomicGCP responded less at every concentra-tion of SHH, demonstrating a cell au-tonomous deficit in response to this im-portant growth factor. Second, trisomicGCP did respond, suggesting that asmall increase in SHH might elevateproliferation rates of trisomic GCP tothe same levels as euploid.

To test this, we injected a smallmolecule that activates the SHH path-

way (SHH agonist, SAG) into trisomicpups the day they were born andassessed the number of cells in thecerebellum 6 days later. The singleadministration restored the number ofGCP (Fig. 5b) and the volume of thecerebellum to euploid levels. Further,the number of dividing GCP was alsorestored to euploid level, suggestingthat the ameliorative effect of this sin-gle treatment would continue throughcerebellar development [Roper et al.,2006].

Fig. 5. Growth of the Ts65Dn brain is reduced because trisomic mice do not respond properlyto the SHH growth factor. (a) Purified GCP from both trisomic and euploid mice respond toincreasing concentrations of SHH, but trisomic GCP respond less. A representative experiment isshown. (b) A single treatment at P0 with SAG is sufficient to restore the number of GCP in thecerebellum 1 week later; *indicates a significant difference [adapted from Roper et al., 2006].

MRDD Research Reviews DOI 10.1002/mrdd


REEVES 219


6/6

Conclusion, Study 3SHH is important in brain devel-

opment from very early stages. In addi-tion, this potent morphogen and growthfactor affects the proliferation, migration,differentiation, and survival of manypopulations of cells beginning early indevelopment and continuing throughoutlife. Thus, any stimulation of this path-way must be precisely targeted, both in

which cells are exposed and when duringdevelopment this stimulation takes place.

On the other hand, only thosecells with the appropriate receptors andtransducers of the SHH signal canrespond; those without the receptor areimpervious to its affects. If all trisomiccells show the same attenuated responseto SHH, a more general application of adrug targeted to this specific responsedeficit could have wide-reaching amelio-rative effects. SHH is important in theformation of a substantial part of thebrain. In addition, it affects all structures

derived from neural crest, including thecraniofacial skeleton, enteric ganglia,and components of the outflow tract ofthe heart. This morphogen is primarilyresponsible for establishment of pattern-ing anteriorposterior polarity in thehand. Thus, an attenuated SHH responsecould contribute to multiple aspects ofthe DS phenotype. However, substantialadditional fundamental discovery will berequired to determine whether SHH isbeneficial in the longer term and whataspects of development may specificallybenefit from this type of approach.

OVERALL CONCLUSIONSThe occurrence of three break-

through studies in 1 year is a harbinger ofgreat progress to come in DS and a call tothe wider scientific community to con-sider what other areas of research mightbenefit from the perspectives provided bydefined genetic effects in trisomy. Collec-tively, these studies represent a synthesisof approaches based in development,structure, function, and age-relatedchanges that must be the basis for anyconsideration in this complex disorder.

These multi-faceted considerations canonly be addressed using mouse modelsthat reflect the developmental and physi-ological processes affected in DS, and thatprovide tools essential for understandingthe genetic basis of the divergencebetween trisomic and euploid indi-viduals. A common element of theseapproaches is that all were designed withconsideration of the possibilities for phar-maceutical (or other) amelioration. Com-bined with ongoing clinical trials, e.g., of

antioxidants and cholinesterase inhibitors,the prospects for improving cognitivefunction in DS in the foreseeable futureappear to be strongly positive.n

ACKNOWLEDGMENTThe authors thank Dr. William

Mobley for insightful comments.

REFERENCESAylward EH, Habbak R, Warren AC, et al. 1997.

Cerebellar volume in adults with Down syn-drome. Arch Neurol 54:209212.

Baxter LL, Moran TH, Richtsmeier JT, et al.2000. Discovery and genetic localization ofDown syndrome cerebellar phenotypes usingthe Ts65Dn mouse. Hum Mol Genet 9:195202.

Belichenko PV, Masliah E, Kleschevnikov AM,et al. 2004. Synaptic structural abnormalitiesin the Ts65Dn mouse model of Down syn-drome. J Comp Neurol 480:281298.

Cooper JD, Salehi A, Delcroix JD, et al. 2001.Failed retrograde transport of NGF in amouse model of Downs syndrome: reversal

of cholinergic neurodegenerative phenotypesfollowing NGF infusion. Proc Natl Acad SciUSA 98:1043910444.

Crome R, Cowie V, Slater E. 1966. A statisticalnote on cerebellar and brain-stem weight inDown syndrome. J Mental Defect Res10:6972.

Delabar JM, Theophile D, Rahmani Z, et al.1993. Molecular mapping of twenty-fourfeatures of Down syndrome on chromosome21. Eur J Hum Genet 1:114124.

Escorihuela RM, Fernandez-Teruel A, Vallina IF,et al. 1995. A behavioral assessment ofTs65Dn mice: a putative Down syndromemodel. Neurosci Lett 199:143146.

Fernandez F, Morishita W, Zuniga E, et al. 2007.Pharmacotherapy for cognitive impairmentin a mouse model of Down syndrome. NatNeurosci 10:411413.

Gardiner K, Fortha A, Bechtel L, et al. 2003.Mouse models of Down syndrome: howuseful can they be? Comparison of the genecontent of human chromosome 21 withorthologous mouse genomic regions. Gene318:137147.

Hanson JE, Blank M, Valenzuela RA, et al. 2007.The functional nature of synaptic circuitry isaltered in area CA3 of the hippocampus in amouse model of Downs syndrome. J Physiol579 (Part 1):5367.

Hattori M, Fujiyama A, Taylor TD, et al. 2000.The DNA sequence of human chromosome21. Nature 405:311319.

Holtzman DM, Santucci D, Kilbridge J, et al.1996. Developmental abnormalities and age-related neurodegeneration in a mouse modelof Down syndrome. Proc Natl Acad SciUSA 93:1333313338.

Kleschevnikov AM, Belichenko PV, Villar AJ,et al. 2004. Hippocampal long-termpotentiation suppressed by increased inhi-bition in the Ts65Dn mouse, a geneticmodel of Down syndrome. J Neurosci24:81538160.

Korenberg JR, Chen XN, Schipper R, et al.1994. Down syndrome phenotypes: the con-sequences of chromosomal imbalance. ProcNatl Acad Sci USA 91:49975001.

Levy S. 1953. Pharmacological treatment of agedpatients in a state mental hospital. J AmMed Assoc 153:12601265.

Mural RJ, Adams MD, Myers EW. 2002. A com-parison of whole-genome shotgun-derivedmouse chromosome 16 and the human ge-nome. Science 296:16611671.

Nadel L. 2003. Downs syndrome: a genetic disor-der in biobehavioral perspective. GenesBrain Behav 2:156166.

Olson LE, Richtsmeier JT, Leszl J, et al. 2004. Achromosome 21 critical region does not

cause specific down syndrome phenotypes.Science 306:687690.

Olson LE, Roper RJ, Sengstaken CL, et al. 2007.Trisomy for the Down syndrome criticalregion is necessary but not sufficient forbrain phenotypes of trisomic mice. HumMol Genet 16:774782.

Pennington BF, Moon J, Edgin J, et al. 2003. Theneuropsychology of Down syndrome: evi-dence for hippocampal dysfunction. ChildDev 74:7593.

Potier MC, Rivals I, Mercier G, et al. 2006. Tran-scriptional disruptions in Down syndrome: acase study in the Ts1Cje mouse cerebellumduring post-natal development. J Neuro-chem 97 (Suppl 1):104109.

Reeves R, Irving N, Moran T, et al. 1995. A

mouse model for Down syndrome exhibitslearning and behaviour deficits. Nat Genet11:177183.

Richtsmeier J, Baxter L, Reeves R. 2000. Parallelsof craniofacial maldevelopment in Downsyndrome and Ts65Dn mice. Dev Dyn 217:137145.

Richtsmeier JT, Zumwalt A, Carlson EJ, et al.2002. Craniofacial phenotypes in segmen-tally trisomic mouse models for Down syn-drome. Am J Med Genet 107:317324.

Roper R, Reeves R. 2006. Understanding thebasis for Down syndrome phenotypes. PLoSGenet 2:231236.

Roper RJ, Baxter LL, Saran NG, et al. 2006. De-fective cerebellar response to mitogenicHedgehog signaling in Down syndrome

mice. Proc Natl Acad Sci USA 103:14521456.

Sago H, Carlson EJ, Smith DJ, et al. 1998.Ts1Cje, a partial trisomy 16 mouse modelfor Down syndrome, exhibits learning andbehavioral abnormalities. Proc Natl Acad SciUSA 95:62566261.

Salehi A, Delcroix JD, Belichenko PV, et al. 2006.Increased App expression in a mouse modelof Downs syndrome disrupts NGF transportand causes cholinergic neuron degeneration.Neuron 51:2942.

Siarey RJ, Carlson EJ, Epstein CJ, et al. 1999.Increased synaptic depression in the Ts65Dnmouse, a model for mental retardation inDown syndrome. Neuropharmacology 38:

19171920.Siarey RJ, Kline-Burgess A, Cho M, et al. 2006.Altered signaling pathways underlyingabnormal hippocampal synaptic plasticity inthe Ts65Dn mouse model of Down syn-drome. J Neurochem 98:12661277.

Siarey RJ, Stoll J, Rapoport SI, et al. 1997.Altered long-term potentiation in the youngand old Ts65Dn mouse, a model for Downsyndrome. Neuropharmacology 36:15491554.

Toyoda A, Noguchi H, Taylor TD, et al. 2002.Comparative genomic sequence analysis ofthe human chromosome 21 Down syndromecritical region. Genome Res 12:13231332.



REEVES

genetica s3 sindrom down 1

Documents