Foxg1-Cre Mediated Lrp2 Inactivation in the Developing Mouse Neural Retina, Ciliary and Retinal Pigment Epithelia Models Congenital High Myopia

Myopia is a common ocular disorder generally due to increased axial length of the eye-globe. Its extreme form high myopia (HM) is a multifactorial disease leading to retinal and scleral damage, visual impairment or loss and is an important health issue. Mutations in the endocytic receptor LRP2 gene result in Donnai-Barrow (DBS) and Stickler syndromes, both characterized by HM. To clearly establish the link between Lrp2 and congenital HM we inactivated Lrp2 in the mouse forebrain including the neural retina and the retinal and ciliary pigment epithelia. High resolution in vivo MRI imaging and ophthalmological analyses showed that the adult Lrp2-deficient eyes were 40% longer than the control ones mainly due to an excessive elongation of the vitreal chamber. They had an apparently normal intraocular pressure and developed chorioretinal atrophy and posterior scleral staphyloma features reminiscent of human myopic retinopathy. Immunomorphological and ultrastructural analyses showed that increased eye lengthening was first observed by post-natal day 5 (P5) and that it was accompanied by a rapid decrease of the bipolar, photoreceptor and retinal ganglion cells, and eventually the optic nerve axons. It was followed by scleral thinning and collagen fiber disorganization, essentially in the posterior pole. We conclude that the function of LRP2 in the ocular tissues is necessary for normal eye growth and that the Lrp2-deficient eyes provide a unique tool to further study human HM.

found at the outer, pigmented, layer of the ciliary body epithelium and the retinal pigment epithelium ( Fig 1B). Other Lrp2 expressing tissues during eye development include the cephalic neural crest cells [23] and the optic nerve astrocytes [24]. Conditional inactivation of Lrp2 in the neural crest, via Wnt1-Cre, or in the astrocytes, via GFAP-Cre, did not alter retinal differentiation as shown in S1 Fig, or eye formation indicating that Lrp2 expressed in these tissues was not critical for eye morphogenesis.
To inactivate Lrp2 in the forebrain we used the FoxG1.Cre deleterious mouse strain active from E9.0 onward [20]. Complete Lrp2 ablation within the neural retina and the developing ciliary body was efficient by E12.5 (Fig 1C and 1D). Homozygous Lrp2 FoxG1.cre-KO mutant mice were viable and showed a striking bi-lateral eye enlargement (Fig 1E and 1F). High-resolution small animal magnetic resonance imaging (MRI) revealed that the excessive eye enlargement was associated with a shorter inter-ocular distance and that the retrobulbar space; i.e. the space between the eyeball and the orbit, was reduced in the mutant eyes (Fig 1G-1J). In addition to the ocular phenotype the Lrp2 FoxG1.cre-KO mutants lacked a corpus callosum (arrow and asterisk in Fig 1I and 1J respectively), a feature reminiscent of the human DBS phenotype [11].
To analyze the growth rates of AL and VCD biometric parameters, we distinguished three different phases; phase 1 (P15-P21), phase 2 (P21-P90) and phase 3 (P90-P330) (Fig 2O and  2P). Phase 1 was characterized by a very rapid enlargement of the eyeball, in both control and mutant mice; the AL growth rate was however higher in the mutant (64 μm/d) than in the control eyes (42.4 μm/d). During phase 2 the mean growth rate dropped to 13.4 μm/d in the mutants and 6.5 μm/d in the controls and further slowed to 0.3 μm/d and 0.02 μm/d respectively in phase 3. The very rapid increase in AL during phase 1 was associated with a higher VCD growth rate in the mutants (39.4 μm/d) compared with the controls (8.1 μm/d). In phase 2 the VCD mean growth rate decelerated in both control and Lrp2 FoxG1.cre-KO eyes but remained two-fold higher in the mutants; in phase 3 a very slow increase in VCD (0.07 μm/d) was still recorded exclusively in the mutants (two-way ANOVA, AL: genotype effect: F (1, 18) = 96.98, P < 0.0001; period effect, F (2, 18) = 1151, P < 0.0001; interaction: F (2, 18) = 36.54, P < 0.0001. VCD: genotype effect, F (1,18) = 1631, P < 0.0001; period effect, F (2, 18) = 2371, P < 0.0001; interaction: F (2, 18) = 943.4, P < 0.0001, n = 4 animals per period and genotype, Fig  2O and 2P). In agreement with early onset and continuous extension of the ocular axis posterior staphyloma could systematically be evidenced from P21 onward on MRI sections of the mutant eyes (Fig 2E-2H) as well as on histological sections (Fig 1F).
Further ophthalmological examination confirmed the MRI data and showed that the anterior segment was generally normal in the mutants (Fig 3A and 3B') although occasionally pupillary ectopia could be observed (Fig 3C and 3C'). Topical endoscopy fundus imaging at P60 (Fig 3D and 3E) and P180 (Fig 3F and 3G), when the Lrp2 FoxG1.cre-KO mutant eyes were respectively 30% and 40% longer than the control ones revealed that a widespread pigment dispersion was affecting most of the retina. In some cases the visibility of the underlying sclera areas was suggesting a profound chorioretinal atrophy. A peripapillary atrophy of the pigment epithelium consistent with a peripapillary staphyloma, 3-4 disc diameter wide, trait of the socalled myopic retinopathy was also observed [26,27]. Tonometry evaluation showed that the intraocular pressure (IOP) increased rapidly during postnatal development of both control and mutant eyes, and that it was different between the two genotypes (main effect "genotype": F (1, 126) = 39.47, P < 0.0001, interaction: F (6, 126) = 11.66, P < 0.0001, Fig 3H). Apparently normal values were recorded in the mutants; however the, IOP was significantly lower in Lrp2-deficient eyes at P 330 and P390 ( ÃÃÃ P < 0.001, two-way ANOVA, n = 10 animals per age and genotype, Fig 3H).
It is of interest that normal IOP values can be recorded in highly myopic, including DBS, patients [28]. Together the results show that the lengthening of the ocular axis in Lrp2 FoxG1.cre-KO mutants is primarily due to continuous vitreal chamber enlargement. It is accompanied by myopic chorioretinopathy and normal IOP values Early post-natal onset axial elongation in Lrp2 FoxG1.cre-KO mutant eyes In highly myopic patients the excessive axial elongation of the eye may be observed early in life and is associated with retinal thinning at both the periphery and posterior pole [8,27,29]. In agreement with this retinal thinning could be seen on the histological and MRI sections of the As shown in Table A Comparative analysis of control and mutant eyes at early post-natal stages did not reveal any differences prior to P3. Between P3 and P15 the AL of the control eyes increased slightly; in Lrp2-deficient eyes the AL values were higher and increased significantly between P5 and P15 (two-way ANOVA, main effect "genotype": F (1, 90) = 5019, P < 0.0001, interaction: F (4, 90) = 746.3, P < 0.0001, n = 10 animals per age and genotype, Fig 4A). A significantly higher AL value was first evident in the mutants at P5 (P < 0.001, Fig 4A). A histological section of a P5 enlarged mutant eye is shown in Table D (Fig 4B) (two-way ANOVA, main effect "genotype": F (1, 108) = 12372, P < 0.0001, interaction: F (5, 108) = 486.8, P < 0.0001, differences between genotypes at each age studied P < 0.001, n = 10 animals per age and genotype) ( Fig 4C).
To investigate whether retinal thinning was associated with differentiation and/or lamination defects we analyzed the expression of various retinal markers. The differentiated retina cell markers Brn3a, PKCalpha, PNA-Lectin and Aquaporine 4 were similarly distributed in the retinal ganglion, bipolar, photoreceptor and Müller cells (Fig 4G and 4H). It is interesting to note that despite an overall similar distribution of the PKCalpha staining, the mutant bipolar cells had atrophic dendrites and their connections with the retinal ganglion cells appeared thicker ( Fig 4J). The above results indicate that despite early onset and excessive thinning the differentiation and lamination of the mutant retina were globally preserved.
Increased cell-death contributes to excessive retinal thinning in Lrp2 FoxG1.cre-KO mutant mice To investigate whether decreased retinal cell density was contributing to retinal thinning in the mutants we compared the number of cells in control and mutant ganglion (GCL), ONL and INL cell layers using bins of 40,000 square millimeters. From P5 onward progressively decreasing cell density was observed in the mutant ONL and INL whereas in the mutant GCL cell density first decreased at P15 (two-way ANOVA; main effect "age": F (5, 72) = 3429, P < 0.0001; main effect "genotype": F (2, 72) = 569.8, P < 0.0001, interaction: F (10, 72) = 78.2, P < 0.0001, n = 5 animals per age and genotype, Fig 5A). Decreased cell density is generally due to insufficient cell proliferation and/or increased cell death and may also reflect the continuous stretching of the mutant retina. We therefore analyzed the expression of cell proliferation and death markers at various developmental and post-natal stages. The M-phase cell cycle marker phospho-histone H3 (PH3) was normally distributed in the mutant retinal progenitor cells between E13.5 and P1 (Table A in S4 Fig), as well as at P3 (Fig 5B and 5C). The pattern of the proliferation index was globally similar in both genotypes between E13.5 and P1 (two-way ANOVA, main effect "genotype": F (1, 48) = 5.667, P = 0.0213, main effect "age": F (5, 48) = 1763, P < 0.0001, interaction: F (5, 48) = 6.385, P = 0.0001, n = 5 animals per age and genotype, Fig  5D). A significant reduction of the proliferation index was seen at P3 and P5 in the Lrp2-deficient eyes (P3, P < 0.001, and P5, P < 0.01, Fig 5D).
We used TUNEL-labeling to identify cell death of retinal neurons before birth and between P5 and P21 when programmed cell death peaks in mouse retina [30]. In control retinas TUNEL positive cells were observed between E13.5 and P1 ( Table B in Fig 5E). Accordingly, the quantitative analysis clearly showed a significantly higher number of TUNEL-positive cells in the mutant retinas between P5 and P21 (two-way ANOVA followed by post hoc Tukey test, main effect "genotype": F (1, 40) = 3681, P < 0.0001, interaction: F (4, 40) = 57.17, P < 0.0001, each age P < 0.001, n = 5 animals per age and genotype, Fig 5F). Activated caspase 3 is a pro-apoptotic mediator, expressed during retinal development. In control retinas the activation of caspase 3 peaked in the INL at P7 whereas at P21 only a few caspase 3-profiles were observed throughout the retina (Figs 5G and 4H). In agreement with increased TUNEL staining, the mutant retinas also showed a significantly higher number of caspase 3-profiles between P5 and P21 (two-way ANOVA, main effect: F (1, 40) = 1708, P < 0.0001, interaction: F (4, 40) = 18.86, P < 0.0001, each age P<0.001, n = 5 animals per age and genotype, Fig 5G and 5H) We then compared the expression of stress-related markers in control and mutant eyes. At P21 the expression of the heat shock protein Hsp70 [31] was increased in the mutant INL ( Fig  5I). Because stress-related proteins are also actors of the chaperone-mediated autophagy we hypothesized that this pathway might contribute to cell death in the mutants. This hypothesis was supported by the high expression of markers widely used to monitor autophagic activity including Hsp25/27, the autophagosome marker microtubule-associated protein light chain LC3B-II and the autophagy-related gene Atg12 exclusively in the mutant retina ( Fig 5J). Finally GFAP, another stress-related marker that labels astrocytes was strongly expressed in the mutant ganglion cell layer (Table A and B in S5 Fig) further suggesting that retinal thinning and retinal cell death, at least partly due to increased stress were associated events in the Lrp2 FoxG1.cre-KO eye.
Cell death was only occasionally observed in control retinas after P21. In the Lrp2 FoxG1.cre-KO mutant retinas however, cell death appeared to increase especially in the mutant ganglion cell layer and consequently led to a reduction of the number of axons in the mutant optic nerve observed at P150 (two-tailed, unpaired t test p ÃÃÃ = 1.07E-6<0.001, n = 4 animals per age and genotype, Table C    Scleral modifications in Lrp2 FoxG1.cre-KO mutant mice The sclera, the outer coat of the eye, is a collagen and proteoglycan containing connective tissue with flattened fibroblasts embedded in it. Thinning of the sclera, in particular at the posterior pole of the eye and the ensuing posterior staphyloma, an outward protrusion of all posterior layers, are typical features of myopic maculopathy [9,32]. At the posterior pole of control P90 eyes the mean scleral thickness was around 35μm ( Fig  6A). At the same age the scleral thickness of the Lrp2 FoxG1.cre-KO mutant eyes appeared reduced ( Fig 6B). Comparison of the mean scleral thickness at P90 and P180 showed that at these ages the mutant sclera was thinner than the control one by 33% and 50% respectively (two-tailed, unpaired t test. P90, p ÃÃÃ = 0.000255<0.001; P180, p ÃÃÃ = 3.99E-4<0.001, n = 3 animals per age and genotype, Table A in S6 Fig).
Ultrastructural analysis revealed that while in the control sclera collagen fibrils were organized into well-defined lamellae ( Fig 6C) the mutant collagen fibrils formed fewer lamellae at P90 and P180 (two-tailed, unpaired t test. P90, p ÃÃ = 0.0018<0.01; P180, p ÃÃÃ = 0.00027<0.001, n = 3 animals per age and genotype, Table B in S6 Fig) which appeared disorganized ( Fig 6D). Further analysis of P90 control and mutant eyes showed that within the mutant lamellae the organization of the collagen fibrils was also perturbed; interweaving fibrils, readily observed in the controls, were only occasionally seen throughout the mutant sclera (Table C (Table I and J in S6 Fig). Quantification of the scleral fibrils per square millimeter showed that the fibril density was significantly lower in the mutants (two-tailed, unpaired t test, inner: ÃÃÃ P<0.001, t = 20.63, df = 22; median: ÃÃÃ P<0.001, t = 24.97, df = 22; outer: ÃÃÃ P<0.001, t = 41.9, df = 22, n = 3 animals per genotype, Fig 6E) and that, in contrast with the controls, it was lower in the outer sclera (Fig 6E).
The morphology of the mutant fibrils was very heterogeneous even within a single region of the sclera and their contour was irregular with rectangular-rather than oval-shaped fibrils readily found throughout the mutant sclera (Fig 6F-6I). In control eyes the majority of collagen fibrils from the outer, middle and inner layers of the posterior sclera had a mean cross-sectional diameter of~115 nm (Fig 6J). Compared with the controls the mean cross-sectional diameter of the mutant fibrils was drastically modified in the three scleral layers (Fig 6K). It was of~135 nm in the inner and middle scleral layers and~120 nm in the outer layer. The frequency however of fibers with both smaller (<60 nm) and wider (>180 nm) mean cross-sectional diameters was increased in the mutants (Fig 6K). The above results indicate that the posterior mutant sclera undergoes important structural modifications reminiscent of human HM [33] and consistent with the development of posterior scleral staphyloma. the mutant INL (I). Western-blot analysis of the indicated autophagic markers; GAPDH is used as an internal loading control (J). Two-way ANOVA post hoc Tukey test was used, **P<0.01, ***P<0.001, ns: not statistically significant, values are mean ± SEM of 5 animals per age and genotype; ***p<0.01. Scale bars: 50 μm in B, C, I; 30 μm in E, G. doi:10.1371/journal.pone.0129518.g005

Discussion
In the present study we show that Lrp2 expressed in the anterior neuroepithelium and its derivatives the neural retina, ciliary and retinal pigment epithelium is required for normal eye growth. Lrp2 deficient eyes are abnormally enlarged essentially along the anterior-posterior axis, a feature typical of high myopia. The continuous eye elongation is accompanied by posterior segment anomalies including chorio-retinal atrophy and posterior staphyloma, features of the myopic retinopathy or maculopathy [1,9,27].
LRP2 is a large plasma membrane protein generally expressed at the apical pole of absorptive epithelia. Prominent sites of expression include the renal proximal tubules, the ependymal cells of the brain and the developing neuroepithelium [34,35]. In the kidney LRP2 acts as a clearance receptor for filtered plasma proteins including steroid hormones and carrier proteins for vitamin D metabolites and retinoids, preventing uncontrolled loss of these metabolites and regulating systemic vitamin homeostasis. Mice carrying a targeted disruption of the Lrp2 gene suffer from forebrain malformations, lack of corpus callosum and impaired eye development [16,22]. Most of them die within minutes after birth with only a small percentage surviving to adulthood [36]. Conditional inactivation of Lrp2 using a floxed Lrp2 allele [19] and the MORE-Cre active in the epiblast but not the extra-embryonic tissues results in an identical phenotype confirming the significance of embryonic LRP2 in brain and eye development [16]. In the Lrp2-deficient adult mice the brain and eye defects are associated with low molecular weight proteinuria and vitamin deficiency [17,36]. Remarkably, the nonsense mutations in the zebrafish lrp2 bugeye mutant cause eye enlargement and overt renal reabsorption deficits [13,37] phenotypes shared by patients suffering from DBS [11]. These studies clearly suggest that LRP2 functions are evolutionary conserved but do not provide any information on the tissue autonomy of the ocular LRP2 function.
The LRP2 expressing tissues involved in ocular structure formation are the neuroepithelium that forms the retina, the RPE and the ciliary body epithelium and the neural crest that gives rise to the central part of cornea [38]. LRP2 is also strongly expressed in the optic nerve astrocytes [24]. To preserve the renal function of LRP2 and identify the LRP2 expressing tissue required for normal eye growth we selectively inactivated Lrp2 in the anterior neuroepithelium, the neural crest or the astrocytes. Whereas Lrp2 inactivation in the neural crest, via Wnt1-Cre, or the astrocytes via GFAP-Cre did not alter eye morphogenesis, FoxG1-mediated ablation of Lrp2 dramatically modified post-natal eye development identifying Lrp2 expressed in the anterior neuroepithelium and its derivatives as an essential component of eye morphogenesis.
FoxG1-mediated gene ablation is first efficient around E9.0 [20], stage at which the single eye field is already separated into two, forming the optic vesicle and later, the optic cup. The formation of the optic cup as well as further differentiation of the neuroretina and RPE depends among others on BMP and SHH signaling, pathways known to be modulated by LRP2 [16,22]. Optic cup formation, subsequent retinal differentiation followed by the expression of OTX2, PAX6 and TUJ1, as well as rates of retinal cell proliferation and death appear normal between E13.5 and P5 suggesting that these signaling pathways are unlikely to be perturbed in the Lrp2-deficient eyes. Further retinal differentiation is overall preserved as suggested by the normal distribution of Brn3a, PKCalpha, PNA-Lectin and Aquaporin 4.
The excessive lengthening of the Lrp2 FoxG1.cre-KO mutant ocular axis begins around P5 and is almost exclusively due to the elongation of the vitreal chamber. Indeed the overall growth and morphology of the anterior segment are only marginally modified in the mutants. With the exception of the slight, albeit significant reduction of the lens thickness observed around P330, the CRC and ACD values as well the size of the pupil are not modified in the Lrp2-deficient eyes. Furthermore we do not find any sign of cataract and the occasionally observed pupillary ectopia may be secondary to a detachment of the ciliary body due to the excessive elongation-associated traction.
In addition to the enlarged vitreal chamber the modifications of the Lrp2-deficient posterior segment include retinal thinning and chorioretinal atrophy. Retinal thinning is not simply due to the continuous eye growth but also to a slightly reduced cell proliferation and increased retinal cell death. Between P5 and P21 cell death is particularly important in the photoreceptor and bipolar cell layers. After this age increased cell death in the retinal ganglion cell layer eventually leads to the decrease of the RGC axon number. Remarkably the expression of several autophagic markers is abnormally strong suggesting that the auto regulatory capacities of the Lrp2-deficient retinas may be affected. It is possible that the retinal modifications provide signals contributing to the continuous eye growth but it is not clear how these signals may be triggered from P5 onward. Because the eye-lids are still closed at this age and the retina does not process visual information, it is unlikely that the onset of abnormal eye-growth is visually driven in the mutants. It is rather due to signals produced by the Lrp2 deficient tissues. The ciliary epithelium and RPE are the only sites that express LRP2 throughout life. In view of the implication of LRP2 in protein endocytosis in vivo [17] it is possible that LRP2 mediates the selective uptake of macromolecules and delivery of vitamins or other nutrients to the ocular structures. Whether the impairment of this function provides signals for abnormal eye growth remains to be established.
Scleral modifications including thinning and formation of posterior staphyloma are first evident around P21 suggesting that they are the consequence rather than the cause of the excessive axial elongation in the Lrp2 FoxG1.cre-KO mutants. Although the study of the biomechanical properties of the mutant sclera is beyond the scope of the present work one would anticipate that scleral thinning, impaired collagen fibril number and morphology may increase scleral extensibility and thus favor the Lrp2 FoxG1.cre-KO mutant eye enlargement and formation of peripapillary staphyloma. Increased scleral extensibility may also at least partly explain the rather normal intraocular pressure observed in the Lrp2 FoxG1.cre-KO mutant eyes as well as in highly myopic patients [28].
In sum the present work shows that LRP2 expressed in the developing ocular tissues is required for normal eye growth and that the Lrp2 FoxG1.cre-KO mutants share many characteristics of congenital HM. Because the pathological significance of HM is not due to ametropia but rather the development and extent of the associated degenerative changes we believe that the Lrp2 FoxG1.cre-KO mutants provide a unique tool to study the onset and age-related evolution of high myopia.

Ethics statement
Animal procedures were conducted in strict compliance with approved institutional protocols (INSERM and comité d'éthique en experimentation animale Charles Darwin N°5, permit number 01519.01) and in accordance with the provisions for animal care and use described in the European Communities council directive of 22 September 2010 (2010/63/EU). Deep anesthesia for terminal procedures (perfusion) was provided with a ketamine/xylazine cocktail (80mg/ 10mg/kg).

Immunohistochemistry-Immunofluorescence
Whole embryos were immersed in a solution of 4% paraformaldehyde in PBS for 1 hour at 4°C on a rocking platform. Embryos were frozen and cut into sections 10 μm thick. Pups and adults were anesthetized and perfused transcardially with 4% PFA in 0.12 m phosphate buffer, pH 7.4. After perfusion, eyes were removed from the skull and postfixed overnight in fresh fixative. Serial frozen sections were processed for immunocytochemistry using the antibodies listed in S1 Table Lectin PNA conjugates (1/50; L-21409, Molecular Probes, OR 97402) were also used. Alexa 488-or 594-conjugated antibodies (1:200, Invitrogen) were used for secondary detection. Nuclear staining was achieved in Hoechst 33342. Fluorescent images were obtained using an Olympus confocal microscope (FV-1200-IX83).

Morphometric analysis
For all quantification, slides were coded and counts were performed with the examiner blind to the age and genotype. Coronal sections of wild-type, and Lrp2 FoxG1.cre-KO mice were Nissl stained and analyzed using a 40× objective and a millimetric eyepiece. For each specimen, the retinal thickness and the number of cells were estimated on the two eyes in ten sections spaced by 40 μm. The number of Nissl profiles was counted in an area of 40,000 μm 2 in the GCL, INL, and ONL. The count was performed in a region excluding the ciliary margin and the optic nerve head. For optic nerve studies sections of P45 and P150 wild-type and mutant mice were processed for immunohistochemistry. Fluorescent area and optic nerve diameter were measured by Fiji-ImageJ software.

Terminal deoxynucleotidyl transferase-mediated biotinylated UTP nick end labeling staining
Pups and adults were anesthetized, and their eyes were immediately removed and frozen in isopentane. To visualize nuclei with DNA cleavage, serial sections (20 μm) of the eye were cut on a cryostat, and residues of fluorescein-labeled nucleotides were catalytically added to DNA fragments by terminal deoxy-nucleotidyl-transferase (TdT). Briefly, sections were fixed in fresh 4% PFA/PBS at room temperature for 15 min, washed in PBS three times for 5 min, equilibrated at room temperature for 10 min, and incubated with nucleotide mix and TdT (ApoAlert DNA fragmentation kit; Clontech, Mountain View, CA) at 37°C for 1 h. Tailing reaction was stopped by incubating sections in 2× SSC at room temperature for 15 min. The number of apoptotic profiles was counted in an area of 40,000 μm 2 in the retina. The count was performed in a region excluding the ciliary margin and the optic nerve head.

Topical endoscopy fundus imaging
Topical endoscopy fundus imaging (TEFI) was performed [41] with an endoscope with a 5-cm long otoscope with a 3-mm outer diameter (1218AA; Karl Storz, Tuttlingen, Germany) with step index lenses and an angle of view of 0°, a field of view of 80°in air, and a crescent-shaped illuminating tip. A reflex digital camera with a 6.1-million pixel charge-coupled device (CCD) image sensor (D50 with Nikkor AF 85 /F1.8 D objective; Nikon, Tokyo, Japan) was connected to the endoscope through an adapter containing a +5 lens (approximate value; the distance from the tip to the cornea appeared to have more influence on focus than did the lens power). The preferred settings of the camera are as follows: image format, raw; focus, manual; operating mode, A (priority to opening); diaphragm, 1/1.8; white balance, automatic. The light source was a xenon lamp (reference 201315-20; Karl Storz) connected through a flexible optic fiber to the endoscope.

Anterior Segment and Tonometry Evaluation
Anterior chamber phenotypes were assessed with a slit lamp (SL-D7, Topcon, Oakland, NJ) and photo-documented with a digital video camera (DV-3, Topcon). All images were taken using identical camera settings and prepared by processing with identical image software. All ocular examinations were performed on conscious mice.

Intraocular Pressure (IOP) Measurement
IOP was measured using the TonoLab rebound tonometer for rodents (Tiolat i-care, Finland) according to the manufacturer's recommendations. All IOP measurements were performed between 10 AM and noon in conscious condition. Mice were gently restrained first by hand and placed on a soft towel bed on the desk and usually appeared calm and comfortable. These data were confirmed to be reproducible by three additional different independent studies (n = 20).

Transmission Electron Microscopy
Adult mice were anesthetized and perfused transcardially with 2% paraformaldehyde 2.5%glutaraldehyde in 0.1M PBS (pH 7.4). Retinas were sliced in 200-μm-thick sections, postfixed 2 h in 1% osmium tetroxide, dehydrated in alcohol, cleared in acetone, and embedded in Epon. For light microscopy, transverse serial sections (1 μm) were cut, heat dried, and stained with toluidine blue. Ultrathin sections were cut and stained with lead citrate and examined with a Philips (Aachen, Germany) CM100 electron microscope.

Fibril diameter analyses
For each genotype, three different animals were analyzed. Only the posterior sclera was analyzed in detail. For the posterior sclera, micrographs (12 per group) from nonoverlapping regions of the central portion of sclera wall were taken at X31,680. The diameter of 85 to 300 fibrils was measured from a single region of a photographic negative. For fibrils that contained uneven contours, the minimum diameter was included in the analysis. Micrographs were randomly chosen in a masked manner from the different groups and digitized, and diameters were measured in an image analysis system (FIJI).

Data analyses and statistics
All data are expressed as the means ± s.e.m. To determine significance in our comparisons of biometric and histologic measurements, we used a commercial software (GraphPad Prism 6, San Diego, CA) for two-way ANOVA, with one factor, to determine the main effects (genotype and postnatal day) and their interaction. Differences, when significant (p<0.05) between genotypes at specific postnatal days were analyzed by the Tukey post hoc test and added in the graph. When only two groups were compared (Fig 6E, S5C and S5D, S6A and S6B Figs), twotailed unpaired Student's t test was used.  P90 (B). The signal is increased in the astrocytes (arrows) of the mutants. Neurofilament Heavy Chain (NFH) staining is used to count the optic nerve axons at the ages indicated (C). Despite the significant reduction of NFH+ fibers (55%) observed at P150 in the mutant optic nerve, the diameter of the mutant optic nerve is slightly higher than that of the control (D). Comparisons are made between the two groups, age matched controls and Lrp2 FoxG1