RESEARCH ARTICLE

A novel mouse model of tuberous sclerosis complex (TSC):eye-specific Tsc1-ablation disrupts visual-pathway developmentIwan Jones, Anna-Carin Hagglund, Gunilla Tornqvist, Christoffer Nord, Ulf Ahlgren and Leif Carlsson*

ABSTRACTTuberous sclerosis complex (TSC) is an autosomal dominantsyndrome that is best characterised by neurodevelopmental deficitsand the presence of benign tumours (called hamartomas) in affectedorgans. This multi-organ disorder results from inactivating pointmutations in either the TSC1 or the TSC2 genes and consequentactivation of the canonical mammalian target of rapamycin complex 1signalling (mTORC1) pathway. Because lesions to the eye are centralto TSC diagnosis, we report here the generation and characterisationof the first eye-specific TSC mouse model. We demonstrate thatconditional ablation of Tsc1 in eye-committed progenitor cells leadsto the accelerated differentiation and subsequent ectopic radialmigration of retinal ganglion cells. This results in an increase in retinalganglion cell apoptosis and consequent regionalised axonal losswithin the optic nerve and topographical changes to the contra- andipsilateral input within the dorsal lateral geniculate nucleus. Eyes fromadult mice exhibit aberrant retinal architecture and display all theclassic neuropathological hallmarks of TSC, including an increase inorgan and cell size, ring heterotopias, hamartomas with retinaldetachment, and lamination defects. Our results provide the firstmajor insight into the molecular etiology of TSC within the developingeye and demonstrate a pivotal role for Tsc1 in regulating variousaspects of visual-pathway development. Our novel mouse modeltherefore provides a valuable resource for future studies concerningthe molecular mechanisms underlying TSC and also as a platformto evaluate new therapeutic approaches for the treatment of thismulti-organ disorder.

KEY WORDS: TSC, Retina, Hamartoma

INTRODUCTIONGenetic disruption during the formation of the central nervous system(CNS) is one of the underlying causes of neurodevelopmental deficits;the incidence rate of such disruptions is high within multi-organsyndromes such as tuberous sclerosis complex (TSC) (OMIM191100)(Han and Sahin, 2011; Smalley, 1998; Thiele, 2004). TSC is anautosomal dominant disorder that is caused by inactivating pointmutations in either theTSC1 (9q34) or the TSC2 (16p13.3) genes. Theprotein products of TSC1 and TSC2 (hamartin and tuberin,respectively) form a heterodimeric complex that is stabilised by athird protein partner (TBC17D). This complex negatively regulatescell growth and proliferation through a canonical signalling pathway

involving Ras homologue enriched in brain (Rheb) and themammalian target of rapamycin complex 1 (mTORC1). TSC is bestcharacterised by the presence of benign tumours (called hamartomas)in affected organs due to uncontrolled cell growth driven bymTORC1 hyperactivity. Hamartomas commonly present as cardiacrhabdomyomas, renal angiomyolipomas and facial angiofibroma. Atthe neuropathological level, hamartomas take the form of white matterradial migration lines (RMLs), subependymal nodules (SENs),subependymal giant cell astrocytes (SEGAs) and cortical tubers(Capo-Chichi et al., 2013; Cheadle et al., 2000; Dibble et al., 2012;DiMario, 2004; Garami et al., 2003; Han and Sahin, 2011; Joneset al., 1999; Kwiatkowski and Manning, 2005; Samueli et al., 2015).Individuals with TSC also present with a myriad of complexneurological deficits, with autism and epilepsy being prevalentamongst affected individuals. These observations clearlydemonstrate that TSC is a multifaceted syndrome in which multipleCNS regions contribute to both the neurological and behaviouralcomponents (Costa-Mattioli and Monteggia, 2013; Han and Sahin,2011; Jeste et al., 2008; Smalley, 1998).

The generation of rodent models has proved to be a robustapproach for establishing the molecular etiology underlying TSC.Germline deletion of either Tsc1 or Tsc2 is embryonic lethal owingto organ dysgenesis, whereas heterozygous animals develop aspectrum of phenotypes, with hepatic hemangiomas, renalcarcinoma and renal cysts being prevalent (Kobayashi et al.,2001; Kwiatkowski et al., 2002; Onda et al., 1999). ConditionalTsc1- and Tsc2-knockout animal models have also replicated someof the neuropathological features of TSC. Ablation of hamartin ortuberin in cerebellar Purkinje cells leads to an increase in soma sizeand dendritic spine density (Reith et al., 2013; Tsai et al., 2012).Moreover, loss of hamartin in the cerebral cortex and hippocampusleads to dysplastic neurons, ectopic neurons and reducedmyelination, whereas astrocyte-specific deletion of Tsc1 initiatesastrogliosis and the aberrant migration of hippocampal pyramidalneurons (Meikle et al., 2007; Uhlmann et al., 2002). Such changesto CNS architecture subsequently lead to functional and autistic-likebehavioural deficits (McMahon et al., 2014; Meikle et al., 2007;Reith et al., 2013; Tavazoie et al., 2005; Tsai et al., 2012; Uhlmannet al., 2002). However, although these previous conditional ablationstudies have generated substantial insight into the neurological andbehavioural aspects of TSC, it is still imperative to generateinnovative models that specifically address the roles of hamartin andtuberin in other TSC-affected organs. This is especially true ifanimal models are to be used as platforms to preclinically evaluatenovel therapeutic approaches for the treatment of this multi-organdisorder (Bissler et al., 2013; Franz et al., 2013; Napolioni et al.,2009; Samueli et al., 2015).

An animal model that addresses the involvement of the eye andvisual system in TSC is currently overlooked. This is especiallysurprising because: (i) clinical examination of the eye is one ofthe original diagnostic procedures used to demonstrate CNSReceived 15 June 2015; Accepted 30 September 2015

Umeå Center for Molecular Medicine (UCMM), Umeå University, Umeå 901 87,Sweden.

*Author for correspondence (leif.carlsson@umu.se)

This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.

1517

© 2015. Published by The Company of Biologists Ltd | Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms

involvement in TSC, (ii) three distinct morphological groups ofretinal hamartomas are routinely observed in individuals with TSC,and (iii) approximately 50% of all TSC-affected individuals presentwith eye involvement (Crino, 2013; Gomez, 1991; Mennel et al.,2007; Samueli et al., 2015; Sepp et al., 1996; Shields et al., 2004).We report here the generation and characterisation of an eye-specificTSC mouse model that recapitulates the classic neuropathologicalhallmarks of this syndrome, and also demonstrate a pivotal role forTsc1 in regulating various aspects of visual-pathway development.Our results provide the first major insight into the molecularetiology of TSC within the developing eye.

RESULTSGeneration of an eye-specific Tsc1 conditional-knockoutmouse modelTo generate an eye-specific TSC animal model, we used a loss-of-function approach by breeding mice carrying a conditionalTsc1 allele (Tsc1tm1Djk; referred to as Tsc1+/f or Tsc1f/f ) with miceharbouring an Lhx2-promoter-driven Cre-recombinase transgene[Tg(Lhx2-Cre)1Lcar; referred to as Lhx2-Cre] (Hägglund et al.,2011; Uhlmann et al., 2002). This transgene promotesrecombination starting at embryonic day 8.25 (E8.25) and is

expressed in eye-committed neural progenitor cells that give rise tothe retinal pigment epithelium (RPE), neural retina (NR) and opticstalk (Fig. S1A) (Hägglund et al., 2011). To confirm deletion ofTsc1, we performed immunoblot analysis on eye homogenates(Fig. 1A). A significant reduction in the amount of hamartin wasdetected in both Lhx2-Cre:Tsc1+/f and Lhx2-Cre:Tsc1f/f micecompared with controls (n=3), suggesting that Tsc1 was ablated inan eye-specific manner (Fig. 1B). A major cellular consequence ofTsc1 loss is the activation of mTORC1 kinase complex andsubsequent increase in the levels of phospho-S6 ribosomal protein(pS6) (Kwiatkowski et al., 2002; Meikle et al., 2007). We observeda significant increase in pS6 (S235/236) and pS6 (S240/244) levelsin Lhx2-Cre:Tsc1f/f mice when compared to controls (Fig. 1A,B).Thus, conditional deletion of Tsc1 during eye development results

Fig. 1. Conditional deletion of Tsc1 during eye development leads to anactivation of the mTORC1 pathway. (A) Immunoblot analysis of hamartin,total S6, pS6 (S235/236), pS6 (S240/244) andGAPDH from control, Lhx2-Cre:Tsc1+/f and Lhx2-Cre:Tsc1f/f eye extracts. (B) Densitometry analysisdemonstrates that conditional deletion of Tsc1 results in a reduction ofhamartin level and a parallel upregulation in pS6 (S235/236) and pS6 (240/244) levels. GAPDHwas used for normalization of hamartin densitometry data.Total S6 was used for normalization of pS6 (S235/235) and pS6 (S240/244)densitometry data. All data represent the mean±s.e.m. of three mice from eachgenotype. Abbreviations: GAPDH, glyceraldehyde 3-phosphatedehydrogenase; kDa, kilodalton. P-values are denoted as follows: ns, notsignificant, *P≤0.05, ***P≤0.001, ****P≤0.0001.

TRANSLATIONAL IMPACT

Clinical issueTuberous sclerosis complex (TSC) is a rare, inherited syndrome thatis characterised by neurodevelopmental deficits and the presence ofbenign tumours, known as hamartomas, in affected organs. Thedisease is caused by mutations in either of two genes, TSC1 or TSC2,which are involved in cell proliferation and differentiation. Mutations inthese genes induce aberrant activation of the mammalian target ofrapamycin complex 1 (mTORC1) pathway, leading to uncontrolled cellgrowth and tumour development. In addition to hamartomascommonly affecting the brain, kidney, skin and heart, lesions to theeye are central to TSC diagnosis. However, the consequences of TSCfor the visual system are not known. Moreover, an animal model thatcan be used to specifically assess the involvement of the visualsystem in TSC is currently lacking.

ResultsHere, the authors use a loss-of-function approach to generate a noveleye-specific TSCmouse model by conditional ablation of the Tsc1 gene.Levels of hamartin, the protein encoded by Tsc1, were reducedspecifically in the eyes upon ablation of the gene. The mutant micerecapitulate many of the neuropathological hallmarks of TSC, such ashamartoma-like lesions with retinal detachment, eye enlargement andloss of retinal architecture. The authors performed a detailedcharacterisation of the model at the cellular level, which includedanalysis by optical projection tomography, immunohistochemistry andscanning electronmicroscopy. These analyses revealedmultiple defectsin visual-pathway development, including aberrant neuronaldifferentiation, migration and circuit formation.

Implications and future directionsThis work provides the first retina-specific model of TSC. The modelstrikingly mimics the neuropathological characteristics of the disease,validating its use in studies examining the pathogenic effects of Tsc1 lossin the eye. Moreover, the authors provide the first major insight into themolecular etiology of TSC within the visual system, paving the way for abetter understanding of the underlying retinal pathology. This novelmodel establishes a foundation for future studies that aim to predictclinical complications of TSC in the visual system. In addition, the modelprovides a platform for evaluating new therapeutic approaches for thetreatment of this multi-organ disorder.

1518

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms

in a reduction of hamartin levels and subsequent hyperactivation ofthe mTORC1 pathway.

Conditional deletion of Tsc1 leads to enlarged eyes,hamartomas and loss of retinal architectureWe first examined the morphology of the eye in Lhx2-Cre:Tsc1f/f

mice, because hamartoma formation is one of the pathological

hallmarks of TSC (Samueli et al., 2015). Optical projectiontomography (OPT) analysis demonstrated that the eye of postnatalLhx2-Cre:Tsc1f/f mice was enlarged when compared to controlanimals (Fig. 2A,B). Most striking, however, was the presence ofsevere retinal folding (Fig. 2B, white arrowhead) and the loss of oraserrata (ORS) integrity (Fig. 2B, white arrow). These observationsindicated that Tsc1 ablation had a profound impact upon retinal

Fig. 2. Conditional deletion of Tsc1 leads to enlarged eyes, hamartomas and loss of retinal architecture. (A,B) OPT 3D volume rendering from control (A)and Lhx2-Cre:Tsc1f/f (B) mice, showing an enlargement of the eye, retinal folding (arrowhead) and loss of ora serrata (ORS) integrity (arrow) in Lhx2-Cre:Tsc1f/f

mice. (C-F) Histological analysis of coronal eye sections demonstrating that the retinal folds observed during OPT analysis of Lhx2-Cre:Tsc1f/f mice arehamartomas that were organised into ring heterotopias. Moreover, retinal detachment was a common occurrence in areas of hamartoma formation (C,E,asterisks). (G-J) Histological analysis of coronal eye sections demonstrating that hamartomas first become evident during late embryogenesis (G, arrowhead) inLhx2-Cre:Tsc1f/f mice. These lesions then become more pronounced during postnatal development (H-J, arrowheads). (K-P) Immunostaining and in situhybridisation analyses demonstrating that hamartomas in Lhx2-Cre:Tsc1f/f mice are enriched in pS6 (S235/236) protein (L) and consist of all the cellular classesthat populate the retina: Muller glia (K, GFAP+, p27Kip1+), RGCs (L, Brn3+), bipolar cells (M, PKCα+), horizontal cells (N, calbindin-D+), amacrine cells(O, calretinin+) and photoreceptors (P, Crx+). (Q-T) Histological analysis of coronal eye sections from control (Q,S) and Lhx2-Cre:Tsc1f/f (R,T) mice demonstratethe irregular ordering of the INL, an IPL populated with ectopic cells (R,T, arrowheads) and a widening of the GCL and NFL (S,T, brackets) in Lhx2-Cre:Tsc1f/f

mice. Scale bars: (A,B,G-J) 500 µm; (C-F,Q-T) 50 µm; (K) 100 µm; (L-P) 200 µm. Abbreviations: Brn, brain-specific homeobox; Crx, cone-rod homeobox;D, dorsal; GCL, ganglion cell layer; GFAP, glial fibrillary acidic protein; IPL, inner plexiform layer; IS, inner segments; INL, inner nuclear layer; NFL, nerve fiberlayer; NR, neural retina; ORS, ora serrata; OS, outer segments; PKC, protein kinase C; RPE, retinal pigment epithelium; V, ventral.

1519

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms

morphology and that Lhx2-Cre:Tsc1f/f mice were therefore aplausible model to assess eye involvement in TSC.The adult retina contains one glial cell type (Müller glia) and six

classes of neurons (rod and cone photoreceptors, and horizontal,bipolar, amacrine and ganglion cells) whose cell bodies residewithin three distinct layers: photoreceptor cell bodies reside in theouter nuclear layer (ONL); the inner nuclear layer (INL) containshorizontal, bipolar, amacrine and Müller glia cell bodies; and theganglion cell layer (GCL) contains ganglion cells, whose axonsform the nerve fiber layer (NFL), and displaced amacrine cells.Synapses between photoreceptors and bipolar and horizontal cellsoccur in the outer plexiform layer (OPL), whereas connectionsbetween bipolar, amacrine and ganglion cells occur in the inner

plexiform layer (IPL) (Fig. 2Q) (Bassett andWallace, 2012; Liveseyand Cepko, 2001).

Histological staining of coronal eye sections demonstrated thatthe retinal folding observed in the OPT-derived three-dimensional(3D) volumes were hamartomas (Fig. 2C-F), which were commonlyorganised into ring heterotopias (Fig. 2C-E), with retinaldetachment also being occasionally observed in areas ofhamartoma formation (Fig. 2C,E, asterisks). The hamartomaswere subsequently determined to arise during late embryogenesisand were first consistently detectable at E17.5 (Fig. 2G). Moreover,these lesions were unlikely to arise as a secondary consequence toaberrant proliferation at the early embryonic ages we analysedbecause the percentage of BrdU+ cells in the neural retina of controland Lhx2-Cre:Tsc1f/f mice did not differ (Fig. S2).

The hamartomas then became more pronounced during postnataldevelopment (Fig. 2H-J) and consisted of all the cellular classes thatare contained within the retina. In addition, the hamartomas werealso enriched in pS6 (S235/236) and contained elevated levels ofglial fibrillary acidic protein (GFAP), which is indicative of reactivegliosis (Fig. 2K-P). Finally, although laminar organisation of theretina in Lhx2-Cre:Tsc1f/fmice was grossly maintained (Fig. 2Q-T),several other abnormalities were readily apparent, including anirregular ordering of the INL and an IPL populated with ectopiccells (Fig. 2R,T, arrowheads). We also observed a widening of theGCL and NFL (Fig. 2S,T, black brackets). This was corroboratedby neurofilament (NF) immunostaining, which demonstrateddisorganised nuclei and aberrant neurite trajectories in the GCLand NFL of Lhx2-Cre:Tsc1f/f mice (Fig. S3).

Conditional deletion of Tsc1 leads to aberrant neuritestratification of retinal interneuronsConditional deletion of Tsc1 in the eye led to several morphologicalchanges within the retina that recapitulated many of the hallmarks ofTSC. We therefore took advantage of this eye-specific model todeterminewhether the laminar disorganisation observed in Lhx2-Cre:Tsc1f/f mice influenced neuronal circuitry. We first chose to focus onamacrine and bipolar neurons because we observed a disorganisedINL and IPL that contained cells with elevated levels of pS6 (S235/236) protein in Lhx2-Cre:Tsc1f/f mice (Fig. 3A,B). The IPL issubdivided into two parallel circuits that respond to either a decrease(OFF pathway) or an increase (ON pathway) in light intensity. Thesecircuits are organised into distinct sublaminae (S) that allow for theprocessing of visual information (OFF sublaminae: S1 and S2; ONsublaminae: S3, S4 and S5) (Sanes and Zipursky, 2010).

Fig. 3. Conditional deletion of Tsc1 leads to aberrant neurite stratificationof retinal interneurons. (A,B) Coronal eye sections from control (A) and Lhx2-Cre:Tsc1f/f (B) mice showing global upregulation of pS6 (S235/236) in the INLand GCL of Lhx2-Cre:Tsc1f/f mice. (C-J) Coronal retinal sections from control(C,E,G,I) and Lhx2-Cre:Tsc1f/f (D,F,H,J) mice. Amacrine cells (D, Pax6+) haveirregular cell-body positions in the INL of Lhx2-Cre:Tsc1f/fmice. Also evident isthe presence of ectopic amacrine cells or RGCs (Pax6+) in the IPL (D,arrowheads). Stratification deficits were observed in the Lhx2-Cre:Tsc1f/f micethat ranged from ectopic neurites extending into other sublaminae (F, ChAT+,arrowheads) to a loss of specific sublamina stratification (H, calretinin+,arrowhead). Moreover, rod bipolar cells in Lhx2-Cre:Tsc1f/fmice have distortedaxonal projections that extend through the INL (J, PKCα+, arrowheads) andterminate with enlarged axon pedicles in the IPL (J, PKCα+, arrow).(K,L) Scanning electron microscopic (SEM) analysis of the IPL in control(K) and Lhx2-Cre:Tsc1f/f (L) mice. The density of the IPL in Lhx2-Cre:Tsc1f/f

mice was increased when compared to control mice. Scale bars: (A-J) 25 µm;(K-L) 1 µm. Abbreviations: ChAT, choline acetyltransferase; GCL, ganglion celllayer; IPL, inner plexiform layer; INL, inner nuclear layer; Pax, paired boxprotein; PKC, protein kinase C; S1-S5, sublaminae 1-5.

1520

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms

To identify all amacrine cells, we analysed for the presence ofPax6 protein (Ma et al., 2007). The amacrine cell layer (Pax6+)was severely disrupted within the INL of Lhx2-Cre:Tsc1f/f mice(Fig. 3C,D). A high number of Pax6+ cells were also observedwithinthe IPL, suggesting that the majority of ectopic cells observed in ourmutant mice are either amacrine or retinal ganglion cells (RGCs)because this neuron class is also recognised by the anti-Pax6antibody (Fig. 3D, yellow arrowheads) (Ma et al., 2007). Toinvestigate whether the disorganised cell-body layers influencedneurite positioning within the IPL, we analysed the stratificationpatterns of cholinergic (ChAT+, S2 and S4) and calretinin+ (S1/S2,S2/S3 and S3/S4) amacrine cell subtypes (Fig. 3E-H). EctopicChAT+ neurites could be clearly seen extending into othersublaminae (Fig. 3F, arrowheads), whereas a loss of specificsublamina stratification (e.g. S2/S3) was observed for the calretinin+

cells (Fig. 3H, arrowhead). Moreover, rod bipolar cells (PKCα+)elaborated distorted axonal projections through the INL (Fig. 3J,arrowheads) and terminated with hypertrophic axon pedicles withinenlarged S4 and S5ON sublaminae (Fig. 3J, arrow). The ectopic-celland dendritic-stratification phenotype had full penetrance: weobserved at least one region in each eye analysed (n=18) thatcontained ectopic cells, ectopic neurites or loss of stratification. Thefact that we observed ectopic Pax6+ cells, aberrant neuritestratification and changes to sublaminae formation suggestsprofound changes to the structure of the IPL in Lhx2-Cre:Tsc1f/f

mice. We therefore performed scanning electron microscopy (SEM)and observed that the IPL in Lhx2-Cre:Tsc1f/f mice was overgrownand densely packed when compared to control animals (Fig. 3K,L).In conclusion, our observations demonstrate that Tsc1 is involved inregulating cell-body positionwithin the INL and stratificationwithinthe IPL. Moreover, the overgrown and densely packed IPL mightform a physical barrier that could partly explain the presence ofectopic cells within this region.

Conditional deletion of Tsc1 leads to accelerated RGCdifferentiation and radial-migration defectsOur immunostaining analyses demonstrated that the disorganisedcells in the GCL of Lhx2-Cre:Tsc1f/f mice were enriched in pS6(S235/236) protein (Fig. 3B).Moreover, aberrant and defasciculatedaxonal projections were also evident in the NFL (Fig. S3B). Thesecombined observations suggest that Tsc1 is involved in regulatingseveral aspects of RGC biology. We therefore chose to examineRGC differentiation in our TSC mouse because this cell type is thefirst neuronal class to differentiate in the retina and they appearfollowing Tsc1 ablation. RGCs are also an ideal model in which tostudy visual-pathway development from the perspectives of bothneuronal circuitry and axonal guidance (Erskine and Herrera, 2007).Firstly, we examined RGC differentiation during embryonic andpostnatal ages and chose time points that coincided with theimportant milestones of RGC development: neurogenesis (E12.5),migration (E17.5) and target innervation [postnatal day 1 (P1)through P7] (Huberman et al., 2008; Reese, 2011).To follow RGC differentiation, we analysed for the presence of

Brn3 protein (Pan et al., 2005). Initially, we found a few scatteredBrn3+ RGCs at E11.5 in both control and Lhx2-Cre:Tsc1f/f mice(Fig. 4A,B, arrowheads). The initial wave of RGCs born in thedorsocentral retina was detected at E12.5 in control mice (Fig. 4C,arrowhead). However, we found an increase in the number ofBrn3+ cells in the Lhx2-Cre:Tsc1f/f mice that were dispersedboth dorsally and ventrally throughout the NR (Fig. 4D,arrowheads). To confirm this accelerated differentiation, weanalysed the expression patterns of Math5, Dlx1 and Dlx2,

because these genes are essential for RGC development (de Meloet al., 2003; Yang et al., 2003). Whereas the expression of Math5,Dlx1 andDlx2 in control micewas observed only in the dorsocentralretina (Fig. 4E,G,I, arrowheads), the expression domains of thesegenes in Lhx2-Cre:Tsc1f/f mice were expanded both dorsally andventrally into the peripheral NR, as was observed for Brn3 (Fig. 4F,H,J, arrowheads). Moreover, the maintenance of dorsoventralidentity based on the expression domains of Tbx5 and Vax2demonstrates that the accelerated differentiation of RGCs in theLhx2-Cre:Tsc1f/f mice was not due to changes in dorsoventralpatterning (Fig. S4) (Behesti et al., 2006).

RGC differentiation in control mice proceeded in a wave fromcentral to peripheral retina and, by E17.5, all the Brn3+ cells hadmigrated into their final laminar position within the future GCL(Fig. 4K). The GCL then became more defined during the firstpostnatal week (P1, Fig. 4M and P7, Fig. 4O) and, by thecompletion of neurogenesis at P14, it contained an ordered radialdistribution of Brn3+ RGCs (Fig. 4Q) (Young, 1985). In contrast,although the majority of Brn3+ cells in the Lhx2-Cre:Tsc1f/f micewere observed to migrate into their final laminar position, ectopiccells were detected beginning at E17.5 (Fig. 4L, arrowheads), andthese cells became readily apparent in all mice analysed at postnatalages (P1, Fig. 4N; P7, Fig. 4P; P14, Fig. 4R; red arrowheads). Thus,conditional deletion of Tsc1 induces accelerated differentiation andaberrant radial migration of RGCs.

Conditional deletion of Tsc1 leads to a reduction in RGCnumber with the remaining cells exhibiting enlarged nucleiand an asymmetric mosaic distributionThe proportion of Brn3+ RGCs seemed to be reduced in postnatalLhx2-Cre:Tsc1f/f mice. This was confirmed by cell count analysis(Fig. 4S). Although the initial numbers of Brn3+ RGCs wasincreased in Lhx2-Cre:Tsc1f/fmice due to accelerated differentiation(E12.5), there was a subsequent reduction in the population ofBrn3+ cells through embryonic (E17.5) and postnatal (P3)development compared to control animals. We thereforeperformed flat-mount analyses in order to further quantify thereduction in RGC numbers following the completion of retinalneurogenesis (Fig. 5A,B) (Young, 1985). In control retinas, Brn3+

nuclei were dispersed in regularly patterned mosaics. By contrast,loss of RGC nuclei in Lhx2-Cre:Tsc1f/fmice resulted in asymmetricmosaic distribution. To confirm this irregularity, we examined thespatial properties using Voronoi analysis, which computes theterritory surrounding each cell relative to its neighbours (Cantrupet al., 2012; Galli-Resta et al., 1997; Kay et al., 2012). The Voronoidomain diagrams revealed an increase in cell territory occupied inLhx2-Cre:Tsc1f/f mice, whereas the Voronoi regularity indexdemonstrated that these enlarged territories were organised asirregular mosaics (Fig. 5C-E). To confirm RGC loss, we performedcell-count analysis and observed a significant reduction in thenumber of Brn3+ cells in Lhx2-Cre:Tsc1f/f mice compared tocontrols (Fig. 5F). We reasoned that an increased rate of apoptosiscould account for this observation and therefore quantified thenumber of activated-caspase3+ cells on serial sections taken fromcomparable regions of control and Lhx2-Cre:Tsc1f/f retinas. At allages analysed, we observed a significant increase in the number ofapoptotic cells in Lhx2-Cre:Tsc1f/fmice (Fig. S5). Our observationstherefore demonstrate that an increased level of programmedcell death is one mechanism that contributes to the decreasednumbers of RGCs in Lhx2-Cre:Tsc1f/f mice. Finally, the remainingRGC nuclei seemed to be enlarged in Lhx2-Cre:Tsc1f/f mice. Toquantify this enlargement, we measured the cross-sectional area

1521

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms

of Brn3+ cells in the same retinal flat mounts. We observed asignificant increase in the size of the remaining RGC nuclei withinLhx2-Cre:Tsc1f/f mice compared to controls (Fig. 5G). In summary,

conditional deletion of Tsc1 leads to a loss of RGCs at postnatalages and this reduction in cell number leads to asymmetric mosaicsin Lhx2-Cre:Tsc1f/f mice. Of the RGCs that survive, these cells

Fig. 4. Conditional deletion of Tsc1 leads to accelerated RGC differentiation and radial-migration defects. (A-D) Immunostaining analysis of Brn3+ RGCs(arrowheads) in control and Lhx2-Cre:Tsc1f/f mice at E11.5 (A,B) and E12.5 (C,D) demonstrates the accelerated appearance of RGCs in Lhx2-Cre:Tsc1f/f mice.(E-J) In situ hybridisation analysis at E12.5 demonstrates expanded expression domains for Math5 (E,F), Dlx1 (G,H) and Dlx2 (I,J) in Lhx2-Cre:Tsc1f/f mice(arrowheads). (K-R) Spatiotemporal analysis of RGC position during migration (E17.5, K,L), target innervation (P1, M,N and P7, O,P) and refinement (P14, Q,R)demonstrates that RGCs radially migrate to reach their final laminar position within the GCL by early postnatal ages in control mice (K,M,O,Q). In contrast, ectopicRGCs begin to appear at E17.5 (L, arrowheads) and become more pronounced after birth in Lhx2-Cre:Tsc1f/f mice (N,P,R, arrowheads). (S) Quantification ofBrn3+ RGC number in control and Lhx2-Cre:Tsc1f/f mice during neurogenesis (E12.5), migration (E17.5), and target innervation and synapse formation (P3).Although the initial numbers of Brn3+ RGCs was increased in themutant mice (E12.5), there is a subsequent reduction in the number of Brn3+ cells that continuesthrough late embryonic (E17.5) and early postnatal (P3) development. The data represent the mean±s.e.m. of at least threemice from each genotype. Scale bars:(A-D;E-J) 50 µm; (K-P;Q,R) 25 µm. Abbreviations: Brn, brain-specific homeobox; D, dorsal; Dlx, distal-less homeobox; GCL, ganglion cell layer; INBL, innerneuroblastic layer;Math, mouse atonal; NR, neural retina; ONBL, outer neuroblastic layer; RGC, retinal ganglion cell; V, ventral. P-values are denoted as follows:ns, not significant, *P≤0.05, ****P≤0.0001 compared with controls.

1522

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms

exhibit increased nuclei size. Tsc1 is therefore critical for regulatingcell survival and size.

Conditional deletion of Tsc1 leads to a loss ofretinogeniculate topographyRGCs project their axons into the NFL, and these axonssubsequently elaborate in a radial fashion to exit the eye at theoptic disc. From here, they enter the optic nerves (ONs) and project

through the optic tracts (OTs) to target the visual centres of the brain,including the dorsal lateral geniculate nucleus (dLGN) of thethalamus (Erskine and Herrera, 2007; Godement et al., 1984).Analysis of RGC axon trajectory during embryonic developmentdemonstrated a comparable optic disc exit pattern in control andLhx2-Cre:Tsc1f/fmice (Fig. S6).We therefore reasoned that changesin ON composition would be observed during postnataldevelopment in Lhx2-Cre:Tsc1f/f mice because this time framecoincided with the greatest loss of RGC nuclei within the retina(Figs 4S, 5F and Fig. S5I). Transverse sections through the ON ofcontrol mice demonstrated a clearly defined network of GFAP+

astrocytes that enmeshed the neurofilament-positive (NF+) RGCaxons within the ON at both P7 and P18 (Fig. 6A,C,E,G). Incontrast, regionalised areas of reactive gliosis were clearly visible inseveral domains of the Lhx2-Cre:Tsc1f/f ON at P7 (Fig. 6B,D,arrowheads). Reactive gliosis in addition to regionalised axonal lossbecame more apparent at P18 in the Lhx2-Cre:Tsc1f/f mice: weobserved cross-sectional areas that were significantly enriched inGFAP+ astrocytes and completely devoid of NF+ RGC axons(Fig. 6F,H, arrowheads). That these changes were observed aspatches and not randomly distributed across the ON clearlycorrelates with the selective loss of RGC nuclei in discretedomains across the retina observed in our flat-mount analysis(Fig. 5B). These changes observed in the ON of Lhx2-Cre:Tsc1f/f

mice were reminiscent to those documented in a DBA/2J mousemodel of glaucoma (Howell et al., 2007). We therefore measuredthe intra ocular pressure (IOP) in control and Lhx2-Cre:Tsc1f/f miceand found them to be comparable (mean±s.e.m. of 18±0.82 and21±1.9 mmHg for control and Lhx2-Cre:Tsc1f/fmice, respectively).

We next proceeded to investigate whether the remaining axons inLhx2-Cre:Tsc1f/f mice would target the dLGN. RGC projectionswere labelled with an anterograde tracer to visualize the binocularretinogeniculate projections within the dLGN (Fig. S7A). Theinputs from each eye in control mice were segregated with a singleipsilateral patch in the dorsomedial quadrant of the dLGN beingsurrounded by the contralateral projection (Fig. 7A,B). As expected,a significant reduction in the area occupied by both contralateral andipsilateral projections was observed in Lhx2-Cre:Tsc1f/f mice(Fig. 7C-E). This result clearly correlates with the dramaticreduction in total RGC number and regionalised axonal losswithin the ON. We also observed a significant decrease in thepercentage of the dLGN occupied by ipsilateral projections in Lhx2-Cre:Tsc1f/f mice (Fig. 7F-J). Also of note was the observation thatthe topography and position of the ipsilateral patch within the dLGN

Fig. 5. Conditional deletion of Tsc1 leads to a reduction in RGC numberand asymmetric mosaic distribution. (A,B) Retinal flat-mount analysis ofBrn3+ RGCs in the tangential plane of control (A) and Lhx2-Cre:Tsc1f/f (B) miceshows a reduction in the number of RGCs in Lhx2-Cre:Tsc1f/f mice.(C,D) Voronoi domain analysis of the immunostaining images demonstrates anincrease in cell territories occupied in Lhx2-Cre:Tsc1f/f mice. (E) Voronoidomain regularity indices for control and Lhx2-Cre:Tsc1f/f mice. A reduction inmosaic regularity was observed for Lhx2-Cre:Tsc1f/f mice. The data representthe mean±s.e.m. of three mice from each genotype. (F) Quantification of RGCnumber in the tangential plane of control and Lhx2-Cre:Tsc1f/f mice. Adecrease in the number of Brn3+ cells was observed in Lhx2-Cre:Tsc1f/f mice.The data represent the mean±s.e.m. of three mice from each genotype.(G) Quantification of the cross-sectional RGC nucleus area in the tangentialplane of control and Lhx2-Cre:Tsc1f/f mice. An increase in the cross-sectionalarea of Brn3+ cells was observed in Lhx2-Cre:Tsc1f/f mice. The data representthe mean±s.e.m. of three mice from each genotype. Scale bar: (A,B) 50 µm.Abbreviations: Brn, brain-specific homeobox; RGC, retinal ganglion cell; VDRI,Voronoi domain regularity index. P-values are denoted as follows: **P≤0.01,****P≤0.0001 compared with controls.

1523

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms

of Lhx2-Cre:Tsc1f/f mice was altered in all animals along thedorsoventral and lateromedial axis (Fig. 7H,I) and, in someinstances, we observed more than one ipsilateral patch (Fig. S7B-E, yellow arrows). Our data therefore demonstrates that conditionaldeletion of Tsc1 has a profound influence on the development of thevisual pathway as demonstrated by a decrease in occupied contra-and ipsilateral territories within the dLGN.

DISCUSSIONThis study describes the generation of a novel eye-specific model ofTSC that recapitulates many of the neuropathological hallmarks ofthis multi-organ disorder: increase in organ and cell size, ringheterotopias, hamartomas with retinal detachment, and laminationdefects (Crino, 2013; Shields et al., 2004). Moreover, conditionalablation of Tsc1 led to severe disruptions in visual-pathway

development as evidenced by: (i) accelerated differentiation ofRGCs, (ii) ectopic radial migration, (iii) neurite stratificationdeficits, (iv) an increase in apoptosis, (v) ON degeneration and(vi) aberrant retinogeniculate topography (Fig. S8).

The mutant mice described in this report differ substantially fromthe majority of published TSC models because we ablated Tsc1during early embryonic development in a progenitor cell populationthat generates both neural and glial lineages within the retina. This isin contrast to previous studies where the Tsc1 or Tsc2 genes havebeen ablated during later embryonic development in either neuronalor glial lineages independently (Ehninger et al., 2008; Felicianoet al., 2012; Meikle et al., 2007; Uhlmann et al., 2002; Way et al.,2009; Zeng et al., 2011; Zhou et al., 2011). Although these previousmodels documented some of the neuropathological changesassociated with TSC, none of them recorded the presence ofhamartoma-like lesions. This is presumably due to the fact that Cre-mediated recombination was initiated after tissue architecture wasalready established.We therefore propose that hamartoma formationoccurs owing to the fact that the Tsc1tm1Djk allele is recombined in aprogenitor cell population prior to the formation of any rudimentaryeye structure (Hägglund et al., 2011) and that this might also bea potential mechanism that drives hamartoma formation inindividuals with TSC. Although, it must be noted that our fate-mapping experiments demonstrated that the Lhx2-Cre transgene didnot recombine the ROSA26 allele in all parts of the retina inpostnatal mice (Fig. S1B). We therefore concede that some of thehamartomas formed in Lhx2-Cre:Tsc1f/f mice could be mosaic innature and composed of both Tsc1-null and normal cells. Non-cell-autonomous mechanisms could therefore also potentially contributeto hamartoma formation in our TSC model, as was previouslydemonstrated in zebrafish (Kim et al., 2011). We also cannotexclude the possibility that reactive Müller glia cells are alsoinvolved in hamartoma formation in Lhx2-Cre:Tsc1f/f mice, as wasrecently demonstrated for other rodent models of retinaldegeneration (Hippert et al., 2015).

The eyes of Lhx2-Cre:Tsc1f/fmice were grossly indistinguishablefrom their littermates at birth. The developmental processesinvolved in the formation of the rudimentary eye structure, i.e.diencephalon evagination, optic vesicle formation, optic cuptransformation and optic fissure closing, are therefore presumablyindependent of TSC-mTORC1 function (Adler and Canto-Soler,2007). OPT analysis demonstrated that the initial striking feature ofLhx2-Cre:Tsc1f/f mice was that their eyes were enlarged comparedto control littermates. Our observations corroborate the originalDrosophila melanogaster studies in which an enlarged-eyephenotype was generated upon Tsc1 or Tsc2 ablation, and furtherdemonstrate that mTORC1 signalling controls organ size within thenervous system (Anderl et al., 2011; Ehninger et al., 2008; Gotoet al., 2011; Ito and Rubin, 1999; Magri et al., 2011; Potter et al.,2001; Tapon et al., 2001; Uhlmann et al., 2002; Way et al., 2009;Zeng et al., 2011; Zhou et al., 2011). Histological analysisdemonstrated the presence of ectopic cells in the IPL and addsfurther support to the notion that TSC should be regarded as aneuronal migration syndrome (Feliciano et al., 2012, 2011; Gomez,1991; Magri et al., 2011; Way et al., 2009). Ectopic cells within theIPL have also been observed upon deletion of other tumoursuppressors such as Zac1 and PTEN (Cantrup et al., 2012; Ma et al.,2007; Sakagami et al., 2012). The authors postulated that both Zac1and PTEN directly regulated cell migration by controlling theexpression of cell adhesion genes (Cantrup et al., 2012; Ma et al.,2007; Varrault et al., 2006). That the loss of hamartin has also beendocumented to disrupt cell adhesion therefore provides a potential

Fig. 6. Conditional deletion of Tsc1 leads to reactive gliosis and changesin optic nerve morphology. (A-D) Transverse sections of the ON at P7demonstrate a defined network of GFAP+ astrocytes that enmesh the NF+

RGC axons in control mice (A,C). In contrast, regionalised areas of reactivegliosis are clearly visible in the ON of Lhx2-Cre:Tsc1f/fmice (B,D, arrowheads).(E-H) Transverse sections of the ON at P18 demonstrate a defined network ofGFAP+ astrocytes that enmesh the NF+ RGC axons in control mice (E,G). Incontrast, axon loss is clearly visible in regionalised areas of the ON from Lhx2-Cre:Tsc1f/f mice. Also evident are cross-sectional areas that are significantlyenriched in GFAP+ astrocytes and completely devoid of NF+ RGC axons (F,H,arrowheads). Scale bars: (A,B;E,F) 50 µm; (C,D;G,H) 25 µm. Abbreviations:ON, optic nerve; GFAP, glial fibrillary acidic protein; NF, neurofilament; RGC,retinal ganglion cell.

1524

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms

mechanism for the aberrant radial migration and the appearance ofectopic cells observed in this study (Lamb et al., 2000).Neurites are segregated into distinct synaptic sublaminae within

the IPL of the vertebrate retina, and this organisation establishes thebasis for connectivity and function (Sanes and Zipursky, 2010).However, SEM analysis demonstrated that the IPL of Lhx2-Cre:Tsc1f/f mice was overgrown and densely packed. This phenotypebears gross morphological similarity to previous studies in whichTsc1 ablation resulted in increased dendritic spine density inPurkinje cells and an increase in spine length and width inhippocampal pyramidal neurons (Tavazoie et al., 2005; Tsai et al.,2012). We therefore believe that conditional deletion of Tsc1 in theeye leads to an increase in dendritic elaboration during developmentand that this aberrant neurite outgrowth is responsible for thedensely packed nature of the IPL seen in Lhx2-Cre:Tsc1f/f mice.That rod bipolar cells were also observed to possess enlarged axonpedicles supports our hypothesis. Moreover, this overgrowth couldalso function as a physical barrier and contribute to the ectopicradial-migration phenotype discussed above. Furthermore, thestratification deficits seen for the amacrine cell subtypes mightdevelop as a direct consequence of this dendritic overgrowthbecause sublaminae organisation could be compromised in Lhx2-

Cre:Tsc1f/f mice. But what is also intriguing is that the aberrant IPLneurite targeting seen in our study is also reminiscent of reports inwhich transmembrane semaphorin or plexin signalling is perturbedduring retinal stratification (Matsuoka et al., 2011a,b; Sun et al.,2013). Semaphorins mediate their effects through mTORC1signalling (Campbell and Holt, 2001) and therefore changes tosemaphorin-plexin signalling as a direct result of Tsc1 ablationcould also be a contributing mechanism for the sublaminae deficitsreported here.

We observed accelerated RGC differentiation that is reminiscentto a previous Drosophila study (Bateman and McNeill, 2004). Theauthors demonstrated a loss of temporal neurogenic control uponTsc1 ablation that induced accelerated photoreceptor differentiation.Most importantly, no changes to cell specification were observed.The authors therefore concluded that loss of Tsc1 does not alter cellfates but has an impact upon the temporal control of when cell fatedecisions are made. It is therefore plausible that the accelerateddifferentiation of RGCs observed in our study arises from a similarmechanism. That we observed the expanded expression domains forseveral genes involved in RGC development (Math5, Dlx1 andDlx2) adds support to this hypothesis. Further studies into themechanistic regulation of accelerated RGC neurogenesis in Lhx2-

Fig. 7. Conditional deletion of Tsc1leads to aberrant retinogeniculatetopography. (A-D) Medial coronalsections showing retinogeniculateprojections into the dLGN at P14 incontrol (A,B) and Lhx2-Cre:Tsc1f/f

(C,D) mice. Dashed lines representthe borders of the dLGN. A reductionin the area occupied by bothcontralateral and ipsilateralprojections is evident in Lhx2-Cre:Tsc1f/f mice. (E) Quantification of thetotal dLGN area (µm2) in control andLhx2-Cre:Tsc1f/f mice at P14. Thedata represent the mean±s.e.m. ofthree mice from each genotype. Asignificant reduction in the total areaoccupied by both contralateral andipsilateral projections is evident inLhx2-Cre:Tsc1f/f mice. (F-I) Medialcoronal sections showing theipsilateral retinogeniculate projectionsinto the dLGN at P14 in control (F,G)and Lhx2-Cre:Tsc1f/f (H,I) mice.Dashed lines represent the borders ofthe dLGN. The topographicalappearance and position of theipsilateral patch within the dLGN ofLhx2-Cre:Tsc1f/f mice was alteredalong the dorsoventral andlateromedial axis. (J) Quantification ofthe percentage ipsilateral projectionarea in control and Lhx2-Cre:Tsc1f/f

mice at P14. The data represent themean±s.e.m. of three mice from eachgenotype. A significant reduction inthe percentage area occupied by theipsilateral projection is evident inLhx2-Cre:Tsc1f/f mice. Scale bars:(A-D,F-I) 250 µm. Abbreviations:D, dorsal; dLGN, dorsal lateralgeniculate nucleus; L, lateral.P-values are denoted as follows:****P≤0.0001 comparedwith controls.

1525

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms

Cre:Tsc1f/f mice are required. But the recent demonstration thatUnkempt and Headcase act in the same pathway to negativelyregulate the temporal control of Drosophila photoreceptordifferentiation downstream of Tsc1-mTOR provides an intriguingavenue for future investigation, particularly because both theunkempt and unkempt-like genes are expressed in the mouse retina(Avet-Rochex et al., 2014; Blackshaw et al., 2004).RGCs project to the dLGN in a stereotypical manner, with

retinogeniculate topography being established by a combination ofEphA/ephrin-A signalling, patterned retinal-activity-dependentaxonal refinement and competition (Feldheim et al., 1998; Pennet al., 1998; Torborg and Feller, 2005). Notable differences wereobserved in retinogeniculate topography of the Lhx2-Cre:Tsc1f/f

mice. Firstly, an expected reduction in the total (contra- andipsilateral) and ipsilateral-only territories was evident. This waspresumably due to RGC apoptosis and the consequent disruptionto mosaic regularity we observed in Lhx2-Cre:Tsc1f/f mice.Secondly, the topography and position of the ipsilateral patchwithin the dLGN of Lhx2-Cre:Tsc1f/f mice was altered and insome instances we observed more than one ipsilateral patch. Thisphenotype is reminiscent to the dLGN ipsilateral projectionphenotype reported for nicotinic acetylcholine receptor β2(nAChRβ2), ephrin-A2/A3/A5 and Phr1 mouse models (Culicanet al., 2009; Pfeiffenberger et al., 2005; Rossi et al., 2001). Threepossible mechanisms could account for this observed ipsilateralphenotype: (i) firstly, Lhx2-Cre:Tsc1f/f mice could have defects inaxonal refinement. Previous studies have demonstrated thatnAChRβ2-driven retinal waves during the first postnatal week ofdevelopment are crucial for dLGN axonal refinement (Cang et al.,2005; Grubb et al., 2003; Rossi et al., 2001). nAChRβ2-null micehave defects in axonal refinement that are reminiscent to thoseobserved in our study and therefore we cannot rule out thepossibility that Lhx2-Cre:Tsc1f/f mice display a similar loss ofretinal activity within the first postnatal week that contributes tothe topographic phenotypes reported here. (ii) A second possiblemechanism is that Lhx2-Cre:Tsc1f/f mice could have defects inEphA/ephrin-A-dependent axon-guidance mechanisms similar tothat recently documented for Tsc2 heterozygous mice. The authorsdemonstrated that RGCs in Tsc2+/− mice had aberrant ipsilateralprojections and provided mechanistic data suggesting that TSC-mTORC1 signalling cooperates with the ephrin–Eph-receptorsystem to control projection topography within the dLGN (Nieet al., 2010). Our observations expand on this previous report andclearly demonstrate that the aberrant ipsilateral topography occursthrough an RGC-autonomous mechanism because our fate-mapping experiments verified that the Lhx2-Cre transgene wasnot expressed in the dLGN (Fig. S7F). (iii) A third potentialmechanism for the dLGN topographic phenotype observed inLhx2-Cre:Tsc1f/f mice is one that is independent of retinal activityor EphA/ephrin-A signalling. A recent study reported thatconditional deletion of a PHR protein family member (Phr1) ledto eye-specific domains within the dLGN being severely disturbedin a manner similar to that observed in Lhx2-Cre:Tsc1f/f mice(Culican et al., 2009). The authors did not propose a specificmechanism for the observed retinotopic changes upon Phr1deletion, but this third mechanism is intriguing because PHRproteins are enriched in the CNS and have been demonstrated tointeract and regulate activity of the hamartin-tuberin complex(Murthy et al., 2004).In conclusion, we have generated a novel eye-specific model of

TSC that recapitulates many of the neuropathological hallmarks ofthis syndrome: an increase in organ and cell size, ring heterotopias,

hamartomas with retinal detachment, and lamination defects.Moreover, our results demonstrate a pivotal role for Tsc1 inregulating various aspects of visual-pathway development anddemonstrate that the TSC-mTORC1 pathway is critically involvedin neuronal differentiation, migration and circuit formation. Ourmouse model will provide a valuable resource for future studiesconcerning the molecular mechanisms underlying TSC and also asa platform to evaluate new therapeutic approaches for the treatmentof this multi-organ disorder.

MATERIALS AND METHODSAnimalsAll animal experiments were approved by the Animal Review Board at theCourt of Appeal of Northern Norrland in Umeå. The generation andgenotyping of Tg(Lhx2-Cre)1Lcar transgenic mice (referred to as Lhx2-Cre), Tsc1tm1Djk mice (referred to as Tsc1+/f or Tsc1f/f ) and Gt(ROSA)26Sortm1sor reporter mice (referred to as ROSA26R) have been describedpreviously (Hägglund et al., 2011; Soriano, 1999; Uhlmann et al.,2002). The genotype of all animals was determined by PCR analysisof genomic DNA extracted from tail biopsies. Primers used to identifyLhx2-Cre and ROSA26R mice have been described previously(Hägglund et al., 2011). Primers used to identify Tsc1tm1Djk mice wereIMR4008 5′-GTCACGACCGTAGGAGAAGC-3′ and IMR4009 5′-GAATCAACCCCACAGAGCAT-3′. Breeding Lhx2-Cre:Tsc1+/f andTsc1f/f mice generated experimental animals. Breeding Lhx2-Cre andROSA26R mice generated fate-mapping animals. The morning of thevaginal plug was considered as E0.5. Littermates lacking the Lhx2-Cretransgene that were either heterozygous or homozygous for the Tsc1tm1Djk

allele were used as controls in all experiments. Lhx2-Cre:Tsc1f/f micewere born in Mendelian ratios (19 of 74 pups were Lhx2-Cre:Tsc1f/f in oneset of genotyping experiments) and no gross eye defects were evidentat birth. Lhx2-Cre:Tsc1f/f mice were indistinguishable from controlsuntil P5. Thereafter, they failed to thrive and no Lhx2-Cre:Tsc1f/f micesurvived longer than P21 (Fig. S9). The observed mortality presumablyoccurs owing to the activity of the Lhx2-Cre transgene and consequentablation of the Tsc1tm1Djk allele in other domains outside the retina(Hägglund et al., 2011).

ImmunoblottingThe cornea, lens, ON and blood vessels were removed from enucleated eyesand the retina/RPE was snap-frozen in liquid nitrogen. Tissues werehomogenized in SDS lysis buffer [100 mM Tris, pH 6.8; 2% (w/v) SDS],containing protease and phosphatase inhibitor cocktails (Complete Mini &PhosSTOP, Roche), using a TissueLyser (Qiagen Inc.). The protein extractswere centrifuged at 14,000 g for 5 min at 4°C and the concentration of thesoluble fraction was measured using the BCA Protein Assay Kit (FisherThermo Scientific). Western blotting was performed on soluble proteinextracts (10 µg) using Criterion TGX gels (Bio-Rad) and nitrocellulosemembranes (Bio-Rad) as previously described (Mahmood and Yang,2012). Reactive proteins were visualised using SuperSignal West DuraExtended Duration Substrate (Fisher Thermo Scientific). Imaging andquantification was performed using a ChemiDoc MP Imaging System (Bio-Rad). The following antibodies and dilutions were used: Hamartin (CellSignaling Technology, 1:1000), S6 (Cell Signaling Technology, 1:1000),pS6 (S235/236) (Cell Signaling Technology, 1:1000), pS6 (S240/244) (CellSignaling Technology, 1:1000) and GAPDH (Cell Signaling Technology,1:30,000).

Optical projection tomography (OPT)Enucleated eyes were pierced through the cornea with a 27-gauge needle toallow for fixative infusion and prepared for OPT scanning as describedpreviously (Alanentalo et al., 2007). Specimens were scanned intransmission OPT mode on a Bioptonics 3001 OPT scanner (Bioptonics)and the acquired data was processed as described previously (Cheddad et al.,2012). 3D volumes based on the tomographic reconstructions were renderedin Drishti software (Version 2.2).

1526

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms

HistologyHeads or enucleated eyes were fixed in 2% (w/v) glutaraldehyde in PBSovernight at 4°C. Eyes were immersed in 70% (v/v) ethanol and paraffin-embedded. Paraffin sections (10 µm) were cleared by 2×5 min incubation inxylenes before rehydration through a graded series of ethanol [99.5%, 95%,90% and 80% (v/v) in PBS]. Haematoxylin and eosin staining wasperformed as previously described (Hägglund et al., 2011).

ImmunohistochemistryFor cryosection analysis, heads or enucleated eyes were fixed in 4% (w/v)PFA in PBS for up to 2 h on ice, equilibrated overnight at 4°C in 30% (w/v)sucrose in PBS and embedded in OCT compound (Sakura Finetek). For flat-mount analysis, retinae were dissected out from enucleated eyes andindividually placed in separate wells of a 24-well plate and fixed in 4% (w/v)PFA in PBS for 2 h on ice. Immunohistochemistry was performed oncryosections (10-20 µm) or flat-mount retinae as previously described(Cantrup et al., 2012; Ma et al., 2007). The following antibodies anddilutions were used: pS6 (S235/236) (Cell Signaling Technology, 1:100),Brn3 (Santa Cruz Biotechnology, 1:50), activated caspase3 (Abcam,1:1000), NF (SMI31) (Covance, 1:1000), GFAP (Chemicon, 1:400), Pax6(DSHB, 1:100), ChAT (Millipore, 1:100), Calretinin (Swant, 1:1000),PKCα (Santa Cruz Biotechnology, 1:100), Calbindin-D (Sigma-Aldrich,1:100) and p27Kip1 (BD Biosciences, 1:750).

Proliferation analysesPregnant mice were injected intraperitoneally with 5-bromo-2′-deoxyuridine (BrdU) at 50 µg/g body weight for 5 min before sacrificeand embryo collection. Embryonic heads were fixed in 4% (w/v) PFA inPBS for up to 2 h on ice, equilibrated overnight at 4°C in 30% (w/v) sucrosein PBS and embedded in OCT compound (Sakura Finetek). Cryosections(10 µm) were washed in PBS and then denatured in 2 M HCl for 30 min at37°C. The slides were subsequently neutralised with 0.1 M borate buffer(pH 8.5) for 10 min at room temperature (RT) followed byimmunohistochemistry as described above using an anti-BrdU antibody(BD Pharmingen, 1:20).

In situ hybridisationHeads or enucleated eyes were fixed in 4% (w/v) PFA in PBS for up to 2 hon ice, equilibrated in 30% (w/v) sucrose in PBS overnight at 4°C andembedded in OCT compound (Sakura Finetek). In situ hybridization oncryosections (10 µm) was performed as previously described (Schaeren-Wiemers and Gerfin-Moser, 1993). The following IMAGE clones (SourceBioscience) were purchased to generate in situ probes: Crx (Clone#4527863), Dlx1 (Clone #30360273), Dlx2 (Clone #5718422), Math5(Clone #6824509), Tbx5 (Clone #30548234) and Vax2 (Clone #40101825).

Intra ocular pressure (IOP) measurementIOP measurements were performed using a Tonolab instrument (IcareFinland Oy) according to the manufacturer’s instructions.

Lhx2-Cre:ROSA26R fate-mappingEmbryonic heads or enucleated eyes were harvested from Lhx2-Cre:ROSA26 mice and fixed in 4% (w/v) PFA in PBS for up to 2 h on ice. Thetissues were subsequently equilibrated in 30% (w/v) sucrose in PBSovernight at 4°C, embedded in OCT compound (Sakura Finetek) andcoronal cryosections (10 µm) were prepared. Adult Lhx2-Cre:ROSA26mice (>6 weeks old) were transcardially perfused with 30 ml of ice-cold 4%(w/v) PFA in PBS. The brain was subsequently removed and embedded in4% (w/v) agarose in PBS. Coronal vibratome sections (100 µm) wereprepared. β-galactosidase staining was performed as described previously(Hägglund et al., 2011).

Retinogeniculate projection analysisHeads were fixed in 4% (w/v) PFA in PBS for 24 h at 4°C. The corneaand lens were removed from each eye and a 1-mm2 piece of NeuroVueRed Plus or NeuroVue Maroon (Polysciences Inc.) was inserted into theoptic disc (Fig. S7A). The heads were incubated in 4% (w/v) PFA in PBS

for 8 weeks at 37°C. The brain was subsequently removed and embeddedin 4% (w/v) agarose in PBS. Coronal vibratome sections (100 µm)were counterstained with DAPI and mounted with Aqua Polymount(Polysciences Inc.). Retinogeniculate analyses were performed asdescribed previously (Leamey et al., 2007) and were calculated acrossthe entire rostrocaudal extent of the dLGN. The total area (µm2) of thedLGN (as determined by DAPI staining) was first calculated using thefreehand selection tool in Fiji (Schindelin et al., 2012). Background wasthen subtracted using the rolling ball function and the area occupied byboth the contralateral and ipsilateral retinogeniculate projections wascalculated using the same approach as above for the total area analysis.The territory occupied by the ipsilateral input was subsequently calculatedas a percentage of the total dLGN area.

Scanning electron microscopy (SEM)Enucleated eyes were pierced through the cornea with a 27-gauge needle toallow for fixative infusion and fixed in 2% (w/v) glutaraldehyde in 0.1 Mcacodylate buffer (pH 7.2) overnight at 4°C. Eyes were subsequentlyimmersed in 70% (v/v) ethanol and paraffin-embedded. Paraffin sections(10 µm) were cleared by 2×5 min incubation in xylenes. Slides weresubsequently washed 2×5 min in 99.5% (v/v) ethanol and allowed to air dry.Dehydrated samples were attached onto aluminium mounts using carbonadhesive tape and coated with 5 nm gold/palladium. Section morphologywas examined using a Zeiss Merlin field emission scanning electronmicroscope using a secondary electron detector at a beam acceleratingvoltage of 4 kV and probe current of 150 pA.

Image analysesImmunohistochemistry and NeuroVue images were captured using a ZeissLSM 710 confocal microscope. Histology, in situ hybridisation and fate-mapping images were captured using a Nikon Eclipse E800 microscopefitted with a Nikon DS-Ri1 digital colour camera. All images were compiledand analysed using Fiji (Schindelin et al., 2012), CellProfiler (Carpenteret al., 2006; Kamentsky et al., 2011), Adobe Photoshop and/or AdobeIllustrator.

Statistical analysesAll statistical analyses were performed using Prism6 (GraphPad Software).Unpaired two-tailed Student’s t-tests were used to determine statisticalsignificances. All statistical analyses were performed on data derived from atleast three mice of each genotype. Error bars in all figures represent thestandard error of the mean (s.e.m.). P-values are indicated as follows: ns, notsignificant; *P≤0.05, **P≤0.01, ***P≤0.001, ****P≤0.0001.

AcknowledgementsThe authors would like to thank SaraWilson, Camilla Holmlund and CarolinaWahlbyfor critical reading of the manuscript and technical assistance. The authorsacknowledge the facilities and technical assistance of the Umeå Core FacilityElectron Microscopy (UCEM) at the Chemical Biological Center (KBC), UmeåUniversity. The Pax6 monoclonal antibody, developed by Atsushi Kawakami, wasobtained from the Developmental Studies Hybridoma Bank, created by the NICHDof the NIH and maintained at The University of Iowa, Department of Biology, IowaCity, IA 52242.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsI.J., A.-C.H., G.T., C.N., U.A. and L.C. designed and conceived the experiments. I.J.,A.-C.H., G.T. and C.N. performed the experiments. I.J., A.-C.H., G.T., C.N., U.A. andL.C. analysed the data. I.J. and L.C. wrote the paper.

FundingThis work was supported by the Swedish Research Council, the Swedish CancerSociety and the Medical Faculty at Umeå University.

Supplementary informationSupplementary information available online athttp://dmm.biologists.org/lookup/suppl/doi:10.1242/dmm.021972/-/DC1

1527

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms

ReferencesAdler, R. and Canto-Soler, M. V. (2007). Molecular mechanisms of optic vesicledevelopment: complexities, ambiguities and controversies. Dev. Biol. 305, 1-13.

Alanentalo, T., Asayesh, A., Morrison, H., Loren, C. E., Holmberg, D., Sharpe, J.and Ahlgren, U. (2007). Tomographic molecular imaging and 3D quantificationwithin adult mouse organs. Nat. Methods 4, 31-33.

Anderl, S., Freeland, M., Kwiatkowski, D. J. and Goto, J. (2011). Therapeuticvalue of prenatal rapamycin treatment in a mouse brain model of tuberoussclerosis complex. Hum. Mol. Genet. 20, 4597-4604.

Avet-Rochex, A., Carvajal, N., Christoforou, C. P., Yeung, K., Maierbrugger,K. T., Hobbs, C., Lalli, G., Cagin, U., Plachot, C., McNeill, H. et al. (2014).Unkempt is negatively regulated by mTOR and uncouples neuronal differentiationfrom growth control. PLoS Genet. 10, e1004624.

Bassett, E. A. and Wallace, V. A. (2012). Cell fate determination in the vertebrateretina. Trends Neurosci. 35, 565-573.

Bateman, J. M. and McNeill, H. (2004). Temporal control of differentiation by theinsulin receptor/tor pathway in Drosophila. Cell 119, 87-96.

Behesti, H., Holt, J. K. L. and Sowden, J. C. (2006). The level of BMP4 signaling iscritical for the regulation of distinct T-box gene expression domains and growthalong the dorso-ventral axis of the optic cup. BMC Dev. Biol. 6, 62.

Bissler, J. J., Kingswood, J. C., Radzikowska, E., Zonnenberg, B. A., Frost, M.,Belousova, E., Sauter, M., Nonomura, N., Brakemeier, S., de Vries, P. J. et al.(2013). Everolimus for angiomyolipoma associated with tuberous sclerosiscomplex or sporadic lymphangioleiomyomatosis (EXIST-2): a multicentre,randomised, double-blind, placebo-controlled trial. Lancet 381, 817-824.

Blackshaw, S., Harpavat, S., Trimarchi, J., Cai, L., Huang, H., Kuo,W. P., Weber,G., Lee, K., Fraioli, R. E., Cho, S.-H. et al. (2004). Genomic analysis of mouseretinal development. PLoS Biol. 2, e247.

Campbell, D. S. and Holt, C. E. (2001). Chemotropic responses of retinal growthcones mediated by rapid local protein synthesis and degradation. Neuron 32,1013-1026.

Cang, J., Renterıa, R. C., Kaneko, M., Liu, X., Copenhagen, D. R. and Stryker,M. P. (2005). Development of precise maps in visual cortex requires patternedspontaneous activity in the retina. Neuron 48, 797-809.

Cantrup, R., Dixit, R., Palmesino, E., Bonfield, S., Shaker, T., Tachibana, N.,Zinyk, D., Dalesman, S., Yamakawa, K., Stell, W. K. et al. (2012). Cell-typespecific roles for PTEN in establishing a functional retinal architecture. PLoS ONE7, e32795.

Capo-Chichi, J.-M., Tcherkezian, J., Hamdan, F. F., Decarie, J. C.,Dobrzeniecka, S., Patry, L., Nadon, M.-A., Mucha, B. E., Major, P., Shevell,M. et al. (2013). Disruption of TBC1D7, a subunit of the TSC1-TSC2 proteincomplex, in intellectual disability and megalencephaly. J. Med. Genet. 50,740-744.

Carpenter, A. E., Jones, T. R., Lamprecht, M. R., Clarke, C., Kang, I. H., Friman,O., Guertin, D. A., Chang, J. H., Lindquist, R. A., Moffat, J. et al. (2006).CellProfiler: image analysis software for identifying and quantifying cellphenotypes. Genome Biol. 7, R100.

Cheadle, J. P., Reeve, M. P., Sampson, J. R. and Kwiatkowski, D. J. (2000).Molecular genetic advances in tuberous sclerosis. Hum. Genet. 107, 97-114.

Cheddad, A., Svensson, C., Sharpe, J., Georgsson, F. and Ahlgren, U. (2012).Image processing assisted algorithms for optical projection tomography. IEEETrans. Med. Imaging 31, 1-15.

Costa-Mattioli, M. and Monteggia, L. M. (2013). mTOR complexes inneurodevelopmental and neuropsychiatric disorders. Nat. Neurosci. 16,1537-1543.

Crino, P. B. (2013). Evolving neurobiology of tuberous sclerosis complex. ActaNeuropathol. 125, 317-332.

Culican, S. M., Bloom, A. J., Weiner, J. A. and DiAntonio, A. (2009). Phr1regulates retinogeniculate targeting independent of activity and ephrin-Asignalling. Mol. Cell. Neurosci. 41, 304-312.

de Melo, J., Qiu, X., Du, G., Cristante, L. and Eisenstat, D. D. (2003). Dlx1, Dlx2,Pax6, Brn3b, and Chx10 homeobox gene expression defines the retinal ganglionand inner nuclear layers of the developing and adult mouse retina. J. Comp.Neurol. 461, 187-204.

Dibble, C. C., Elis, W., Menon, S., Qin, W., Klekota, J., Asara, J. M., Finan, P. M.,Kwiatkowski, D. J., Murphy, L. O. andManning, B. D. (2012). TBC1D7 is a thirdsubunit of the TSC1-TSC2 complex upstream of mTORC1.Mol. Cell 47, 535-546.

DiMario, F. J., Jr. (2004). Brain abnormalities in tuberous sclerosis complex. J. ChildNeurol. 19, 650-657.

Ehninger, D., Han, S., Shilyansky, C., Zhou, Y., Li, W., Kwiatkowski, D. J.,Ramesh, V. and Silva, A. J. (2008). Reversal of learning deficits in a Tsc2+/−mouse model of tuberous sclerosis. Nat. Med. 14, 843-848.

Erskine, L. and Herrera, E. (2007). The retinal ganglion cell axon’s journey: insightsinto molecular mechanisms of axon guidance. Dev. Biol. 308, 1-14.

Feldheim, D. A., Vanderhaeghen, P., Hansen, M. J., Frisen, J., Lu, Q., Barbacid,M. and Flanagan, J. G. (1998). Topographic guidance labels in a sensoryprojection to the forebrain. Neuron 21, 1303-1313.

Feliciano, D. M., Su, T., Lopez, J., Platel, J.-C. and Bordey, A. (2011). Single-cellTsc1 knockout during corticogenesis generates tuber-like lesions and reducesseizure threshold in mice. J. Clin. Invest. 121, 1596-1607.

Feliciano, D. M., Quon, J. L., Su, T., Taylor, M. M. and Bordey, A. (2012).Postnatal neurogenesis generates heterotopias, olfactory micronodules andcortical infiltration following single-cell Tsc1 deletion. Hum. Mol. Genet. 21,799-810.

Franz, D. N., Belousova, E., Sparagana, S., Bebin, E. M., Frost, M., Kuperman,R., Witt, O., Kohrman, M. H., Flamini, J. R., Wu, J. Y. et al. (2013). Efficacy andsafety of everolimus for subependymal giant cell astrocytomas associated withtuberous sclerosis complex (EXIST-1): a multicentre, randomised, placebo-controlled phase 3 trial. Lancet 381, 125-132.

Galli-Resta, L., Resta, G., Tan, S. S. and Reese, B. E. (1997). Mosaics of islet-1-expressing amacrine cells assembled by short-range cellular interactions.J. Neurosci. 17, 7831-7838.

Garami, A., Zwartkruis, F. J. T., Nobukuni, T., Joaquin, M., Roccio, M., Stocker,H., Kozma, S. C., Hafen, E., Bos, J. L. and Thomas, G. (2003). Insulin activationof Rheb, a mediator of mTOR/S6K/4E-BP signaling, is inhibited by TSC1 and 2.Mol. Cell 11, 1457-1466.

Godement, P., Salaun, J. and Imbert, M. (1984). Prenatal and postnataldevelopment of retinogeniculate and retinocollicular projections in the mouse.J. Comp. Neurol. 230, 552-575.

Gomez, M. R. (1991). Phenotypes of the tuberous sclerosis complex with a revisionof diagnostic criteria. Ann. N. Y. Acad. Sci. 615, 1-7.

Goto, J., Talos, D. M., Klein, P., Qin, W., Chekaluk, Y. I., Anderl, S., Malinowska,I. A., Di Nardo, A., Bronson, R. T., Chan, J. A. et al. (2011). Regulable neuralprogenitor-specific Tsc1 loss yields giant cells with organellar dysfunction in amodel of tuberous sclerosis complex. Proc. Natl. Acad. Sci. USA 108,E1070-E1079.

Grubb, M. S., Rossi, F. M., Changeux, J.-P. and Thompson, I. D. (2003).Abnormal functional organization in the dorsal lateral geniculate nucleus of micelacking the beta 2 subunit of the nicotinic acetylcholine receptor. Neuron 40,1161-1172.

Hagglund, A.-C., Dahl, L. and Carlsson, L. (2011). Lhx2 is required for patterningand expansion of a distinct progenitor cell population committed to eyedevelopment. PLoS ONE 6, e23387.

Han, J. M. and Sahin, M. (2011). TSC1/TSC2 signaling in the CNS. FEBS Lett. 585,973-980.

Hippert, C., Graca, A. B., Barber, A. C., West, E. L., Smith, A. J., Ali, R. R. andPearson, R. A. (2015). Muller glia activation in response to inherited retinaldegeneration is highly varied and disease-specific. PLoS ONE 10, e0120415.

Howell, G. R., Libby, R. T., Jakobs, T. C., Smith, R. S., Phalan, F. C., Barter, J. W.,Barbay, J. M., Marchant, J. K., Mahesh, N., Porciatti, V. et al. (2007). Axons ofretinal ganglion cells are insulted in the optic nerve early in DBA/2J glaucoma.J. Cell Biol. 179, 1523-1537.

Huberman, A. D., Feller, M. B. and Chapman, B. (2008). Mechanisms underlyingdevelopment of visual maps and receptive fields. Annu. Rev. Neurosci. 31,479-509.

Ito, N. and Rubin, G. M. (1999). gigas, a Drosophila homolog of tuberous sclerosisgene product-2, regulates the cell cycle. Cell 96, 529-539.

Jeste, S. S., Sahin, M., Bolton, P., Ploubidis, G. B. and Humphrey, A. (2008).Characterization of autism in young children with tuberous sclerosis complex.J. Child Neurol. 23, 520-525.

Jones, A. C., Shyamsundar, M. M., Thomas, M.W., Maynard, J., Idziaszczyk, S.,Tomkins, S., Sampson, J. R. and Cheadle, J. P. (1999). Comprehensivemutation analysis of TSC1 and TSC2-and phenotypic correlations in 150 familieswith tuberous sclerosis. Am. J. Hum. Genet. 64, 1305-1315.

Kamentsky, L., Jones, T. R., Fraser, A., Bray, M.-A., Logan, D. J., Madden, K. L.,Ljosa, V., Rueden, C., Eliceiri, K. W. and Carpenter, A. E. (2011). Improvedstructure, function and compatibility for CellProfiler: modular high-throughputimage analysis software. Bioinformatics 27, 1179-1180.

Kay, J. N., Chu, M. W. and Sanes, J. R. (2012). MEGF10 and MEGF11 mediatehomotypic interactions required for mosaic spacing of retinal neurons.Nature 483,465-469.

Kim, S.-H., Speirs, C. K., Solnica-Krezel, L. and Ess, K. C. (2011). Zebrafishmodel of tuberous sclerosis complex reveals cell-autonomous and non-cell-autonomous functions of mutant tuberin. Dis. Model. Mech. 4, 255-267.

Kobayashi, T., Minowa, O., Sugitani, Y., Takai, S., Mitani, H., Kobayashi, E.,Noda, T. and Hino, O. (2001). A germ-line Tsc1 mutation causes tumordevelopment and embryonic lethality that are similar, but not identical to, thosecaused by Tsc2 mutation in mice. Proc. Natl. Acad. Sci. USA 98, 8762-8767.

Kwiatkowski, D. J. and Manning, B. D. (2005). Tuberous sclerosis: a GAP at thecrossroads of multiple signaling pathways. Hum. Mol. Genet. 14 Suppl. 2,R251-R258.

Kwiatkowski, D. J., Zhang, H., Bandura, J. L., Heiberger, K. M., Glogauer, M., el-Hashemite, N. and Onda, H. (2002). A mouse model of TSC1 reveals sex-dependent lethality from liver hemangiomas, and up-regulation of p70S6 kinaseactivity in Tsc1 null cells. Hum. Mol. Genet. 11, 525-534.

Lamb, R. F., Roy, C., Diefenbach, T. J., Vinters, H. V., Johnson, M. W., Jay, D. G.and Hall, A. (2000). The TSC1 tumour suppressor hamartin regulates celladhesion through ERM proteins and the GTPase Rho. Nat. Cell Biol. 2, 281-287.

Leamey, C. A., Merlin, S., Lattouf, P., Sawatari, A., Zhou, X., Demel, N.,Glendining, K. A., Oohashi, T., Sur, M. and Fassler, R. (2007). Ten_m3

1528

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms

regulates eye-specific patterning in themammalian visual pathway and is requiredfor binocular vision. PLoS Biol. 5, e241.

Livesey, F. J. and Cepko, C. L. (2001). Vertebrate neural cell-fate determination:lessons from the retina. Nat. Rev. Neurosci. 2, 109-118.

Ma, L., Cantrup, R., Varrault, A., Colak, D., Klenin, N., Gotz, M., McFarlane, S.,Journot, L. and Schuurmans, C. (2007). Zac1 functions through TGFbetaII tonegatively regulate cell number in the developing retina. Neural Dev. 2, 11.

Magri, L., Cambiaghi, M., Cominelli, M., Alfaro-Cervello, C., Cursi, M., Pala, M.,Bulfone, A., Garcìa-Verdugo, J. M., Leocani, L., Minicucci, F. et al. (2011).Sustained activation of mTOR pathway in embryonic neural stem cells leads todevelopment of tuberous sclerosis complex-associated lesions. Cell Stem Cell 9,447-462.

Mahmood, T. and Yang, P.-C. (2012). Western blot: technique, theory, and troubleshooting. N. Am. J. Med. Sci. 4, 429-434.

Matsuoka, R. L., Chivatakarn, O., Badea, T. C., Samuels, I. S., Cahill, H.,Katayama, K.-I., Kumar, S. R., Suto, F., Chedotal, A., Peachey, N. S. et al.(2011a). Class 5 transmembrane semaphorins control selective Mammalianretinal lamination and function. Neuron 71, 460-473.

Matsuoka, R. L., Nguyen-Ba-Charvet, K. T., Parray, A., Badea, T. C., Chedotal,A. and Kolodkin, A. L. (2011b). Transmembrane semaphorin signalling controlslaminar stratification in the mammalian retina. Nature 470, 259-263.

McMahon, J. J., Yu,W., Yang, J., Feng, H., Helm,M., McMahon, E., Zhu, X., Shin,D. and Huang, Y. (2014). Seizure-dependent mTOR activation in 5-HT neuronspromotes autism-like behaviors in mice. Neurobiol. Dis. 73C, 296-306.

Meikle, L., Talos, D. M., Onda, H., Pollizzi, K., Rotenberg, A., Sahin, M., Jensen,F. E. and Kwiatkowski, D. J. (2007). A mouse model of tuberous sclerosis:neuronal loss of Tsc1 causes dysplastic and ectopic neurons, reducedmyelination, seizure activity, and limited survival. J. Neurosci. 27, 5546-5558.

Mennel, S., Meyer, C. H., Peter, S., Schmidt, J. C. and Kroll, P. (2007). Currenttreatment modalities for exudative retinal hamartomas secondary to tuberoussclerosis: review of the literature. Acta Ophthalmol. Scand. 85, 127-132.

Murthy, V., Han, S., Beauchamp, R. L., Smith, N., Haddad, L. A., Ito, N. andRamesh, V. (2004). Pam and its ortholog highwire interact with and maynegatively regulate the TSC1·TSC2 complex. J. Biol. Chem. 279, 1351-1358.

Napolioni, V., Moavero, R. and Curatolo, P. (2009). Recent advances inneurobiology of Tuberous Sclerosis Complex. Brain Dev. 31, 104-113.

Nie, D., Di Nardo, A., Han, J. M., Baharanyi, H., Kramvis, I., Huynh, T., Dabora,S., Codeluppi, S., Pandolfi, P. P., Pasquale, E. B. et al. (2010). Tsc2-Rhebsignaling regulates EphA-mediated axon guidance. Nat. Neurosci. 13, 163-172.

Onda, H., Lueck, A., Marks, P. W., Warren, H. B. and Kwiatkowski, D. J. (1999).Tsc2(+/−) mice develop tumors in multiple sites that express gelsolin and areinfluenced by genetic background. J. Clin. Invest. 104, 687-695.

Pan, L., Yang, Z., Feng, L. and Gan, L. (2005). Functional equivalence of Brn3POU-domain transcription factors in mouse retinal neurogenesis. Development132, 703-712.

Penn, A. A., Riquelme, P. A., Feller, M. B. and Shatz, C. J. (1998). Competition inretinogeniculate patterning driven by spontaneous activity. Science 279,2108-2112.

Pfeiffenberger, C., Cutforth, T., Woods, G., Yamada, J., Renterıa, R. C.,Copenhagen, D. R., Flanagan, J. G. and Feldheim, D. A. (2005). Ephrin-As andneural activity are required for eye-specific patterning during retinogeniculatemapping. Nat. Neurosci. 8, 1022-1027.

Potter, C. J., Huang, H. and Xu, T. (2001). Drosophila Tsc1 functions with Tsc2 toantagonize insulin signaling in regulating cell growth, cell proliferation, and organsize. Cell 105, 357-368.

Reese, B. E. (2011). Development of the retina and optic pathway. Vision Res. 51,613-632.

Reith, R. M., McKenna, J., Wu, H., Hashmi, S. S., Cho, S.-H., Dash, P. K. andGambello, M. J. (2013). Loss of Tsc2 in Purkinje cells is associated with autistic-like behavior in a mouse model of tuberous sclerosis complex.Neurobiol. Dis. 51,93-103.

Rossi, F. M., Pizzorusso, T., Porciatti, V., Marubio, L. M., Maffei, L. andChangeux, J.-P. (2001). Requirement of the nicotinic acetylcholine receptor beta2 subunit for the anatomical and functional development of the visual system.Proc. Natl. Acad. Sci. USA 98, 6453-6458.

Sakagami, K., Chen, B., Nusinowitz, S., Wu, H. and Yang, X.-J. (2012). PTENregulates retinal interneuron morphogenesis and synaptic layer formation. Mol.Cell. Neurosci. 49, 171-183.

Samueli, S., Abraham, K., Dressler, A., Groeppel, G., Jonak, C., Muehlebner, A.,Prayer, D., Reitner, A., Feucht, M. and Padiatrisches, T. S. C.-Z. W. (2015).Tuberous Sclerosis Complex: new criteria for diagnostic work-up andmanagement. Wien. Klin. Wochenschr. 127, 619-630.

Sanes, J. R. and Zipursky, S. L. (2010). Design principles of insect and vertebratevisual systems. Neuron 66, 15-36.

Schaeren-Wiemers, N. and Gerfin-Moser, A. (1993). A single protocol to detecttranscripts of various types and expression levels in neural tissue and culturedcells: in situ hybridization using digoxigenin-labelled cRNA probes.Histochemistry 100, 431-440.

Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch,T., Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B. et al. (2012). Fiji: anopen-source platform for biological-image analysis. Nat. Methods 9, 676-682.

Sepp, T., Yates, J. R. and Green, A. J. (1996). Loss of heterozygosity in tuberoussclerosis hamartomas. J. Med. Genet. 33, 962-964.

Shields, J. A., Eagle, R. C., Jr., Shields, C. L. and Marr, B. P. (2004). Aggressiveretinal astrocytomas in four patients with tuberous sclerosis complex. Trans. Am.Ophthalmol. Soc. 102, 139-147.

Smalley, S. L. (1998). Autism and tuberous sclerosis. J. Autism Dev. Disord. 28,407-414.

Soriano, P. (1999). Generalized lacZ expression with the ROSA26 Cre reporterstrain. Nat. Genet. 21, 70-71.

Sun, L. O., Jiang, Z., Rivlin-Etzion, M., Hand, R., Brady, C. M., Matsuoka, R. L.,Yau, K.-W., Feller, M. B. and Kolodkin, A. L. (2013). On and off retinal circuitassembly by divergent molecular mechanisms. Science 342, 1241974.

Tapon, N., Ito, N., Dickson, B. J., Treisman, J. E. and Hariharan, I. K. (2001). TheDrosophila tuberous sclerosis complex gene homologs restrict cell growth and cellproliferation. Cell 105, 345-355.

Tavazoie, S. F., Alvarez, V. A., Ridenour, D. A., Kwiatkowski, D. J. and Sabatini,B. L. (2005). Regulation of neuronal morphology and function by the tumorsuppressors Tsc1 and Tsc2. Nat. Neurosci. 8, 1727-1734.

Thiele, E. A. (2004). Managing epilepsy in tuberous sclerosis complex. J. ChildNeurol. 19, 680-686.

Torborg, C. L. and Feller, M. B. (2005). Spontaneous patterned retinal activity andthe refinement of retinal projections. Prog. Neurobiol. 76, 213-235.

Tsai, P. T., Hull, C., Chu, Y., Greene-Colozzi, E., Sadowski, A. R., Leech, J. M.,Steinberg, J., Crawley, J. N., Regehr, W. G. and Sahin, M. (2012). Autistic-likebehaviour and cerebellar dysfunction in Purkinje cell Tsc1 mutant mice. Nature488, 647-651.

Uhlmann, E. J., Wong, M., Baldwin, R. L., Bajenaru, M. L., Onda, H.,Kwiatkowski, D. J., Yamada, K. and Gutmann, D. H. (2002). Astrocyte-specific TSC1 conditional knockout mice exhibit abnormal neuronal organizationand seizures. Ann. Neurol. 52, 285-296.

Varrault, A., Gueydan, C., Delalbre, A., Bellmann, A., Houssami, S., Aknin, C.,Severac, D., Chotard, L., Kahli, M., Le Digarcher, A. et al. (2006). Zac1regulates an imprinted gene network critically involved in the control of embryonicgrowth. Dev. Cell 11, 711-722.

Way, S. W., McKenna, J., III, Mietzsch, U., Reith, R. M., Wu, H. C.-J. andGambello, M. J. (2009). Loss of Tsc2 in radial glia models the brain pathology oftuberous sclerosis complex in the mouse. Hum. Mol. Genet. 18, 1252-1265.

Yang, Z., Ding, K., Pan, L., Deng, M. and Gan, L. (2003). Math5 determines thecompetence state of retinal ganglion cell progenitors. Dev. Biol. 264, 240-254.

Young, R. W. (1985). Cell proliferation during postnatal development of the retina inthe mouse. Brain Res. 21, 229-239.

Zeng, L.-H., Rensing, N. R., Zhang, B., Gutmann, D. H., Gambello, M. J. andWong, M. (2011). Tsc2 gene inactivation causes a more severe epilepsyphenotype than Tsc1 inactivation in a mouse model of tuberous sclerosiscomplex. Hum. Mol. Genet. 20, 445-454.

Zhou, J., Shrikhande, G., Xu, J., McKay, R. M., Burns, D. K., Johnson, J. E. andParada, L. F. (2011). Tsc1 mutant neural stem/progenitor cells exhibit migrationdeficits and give rise to subependymal lesions in the lateral ventricle. Genes Dev.25, 1595-1600.

1529

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1517-1529 doi:10.1242/dmm.021972

Disea

seModels&Mechan

isms