Neurobiology of Disease 46 (2012) 476–485

Contents lists available at SciVerse ScienceDirect

Neurobiology of Disease

j ourna l homepage: www.e lsev ie r .com/ locate /ynbd i

Age-related functional and structural retinal modifications in theIgf1−/− null mouse☆

L. Rodriguez-de la Rosa a,1, L. Fernandez-Sanchez b,1, F. Germain c, S. Murillo-Cuesta a,I. Varela-Nieto a, P. de la Villa c,⁎, N. Cuenca b

a Instituto de Investigaciones Biomédicas ‘Alberto Sols’, Consejo Superior de Investigaciones Científicas (CSIC), Universidad Autónoma de Madrid (UAM), CIBERER Unit 761, ISCiii, IdiPAZ,Arturo Duperier 4, 28029 Madrid, Spainb Departamento de Fisiología, Genética y Microbiología, Universidad de Alicante, 03690 Alicante, Spainc Departamento de Fisiología, Universidad de Alcalá, 28871 Alcalá de Henares, Spain

Abbreviations: ABR, Auditory brainstem response; aERG mixed rod and cone response; bmixed, b-wave amand cone response; bphot, b-wave amplitude of the ERGb-wave amplitude of the ERG scotopic rod response; EInsulin-like growth factor-I; INL, Inner Nuclear LayerONL, Outer Nuclear Layer; OPL, Outer Plexiform Layer;☆ LR-dlR, SM-C and IV-N carried out the ABR studyphenotypes and the genotypes of the animal colony. FGtest. LF-S and NC carried out the immunohistochemistperformed the statistical analysis. LR-dlR, IV-N, PdlV andthe study and drafted the manuscript. All authorsmanuscript.⁎ Corresponding author at: Departamento de Fisio

Universidad de Alcalá de Henares, 28871 Alcalá de HenarE-mail address: pedro.villa@uah.es (P. de la Villa).

1 These authors contributed equally to this work.Available online on ScienceDirect (www.scienced

0969-9961/$ – see front matter © 2012 Elsevier Inc. Alldoi:10.1016/j.nbd.2012.02.013

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 7 December 2011Revised 6 February 2012Accepted 20 February 2012Available online 28 February 2012

Keywords:IGF1RetinaDegenerationDeaf-blindness

Background: Mutations in the gene encoding human insulin-like growth factor-I (IGF-I) cause syndromicneurosensorial deafness. To understand the precise role of IGF-I in retinal physiology, we have studied themorphology and electrophysiology of the retina of the Igf1−/− mice in comparison with that of the Igf1+/−

and Igf1+/+ animals during aging.Methods: Serological concentrations of IGF-I, glycemia and body weight were determined in Igf1+/+, Igf1+/−

and Igf1−/− mice at different times up to 360 days of age. We have analyzed hearing by recording the audi-tory brainstem responses (ABR), the retinal function by electroretinographic (ERG) responses and the retinalmorphology by immunohistochemical labeling on retinal preparations at different ages.Results: IGF-I levels are gradually reduced with aging in the mouse. Deaf Igf1−/− mice had an almost flat sco-topic ERG response and a photopic ERG response of very small amplitude at postnatal age 360 days (P360). At

the same age, Igf1+/− mice still showed both scotopic and photopic ERG responses, but a significant decreasein the ERG wave amplitudes was observed when compared with those of Igf1+/+ mice. Immunohistochem-ical analysis showed that P360 Igf1−/− mice suffered important structural modifications in the first synapseof the retinal pathway, that affected mainly the postsynaptic processes from horizontal and bipolar cells. Adecrease in bassoon and synaptophysin staining in both rod and cone synaptic terminals suggested a reducedphotoreceptor output to the inner retina. Retinal morphology of the P360 Igf1+/− mice showed only smallalterations in the horizontal and bipolar cell processes, when compared with Igf1+/+ mice of matched age.Conclusions: In the mouse, IGF-I deficit causes an age-related visual loss, besides a congenital deafness. Thepresent results support the use of the Igf1−/− mouse as a new model for the study of human syndromicdeaf-blindness.

© 2012 Elsevier Inc. All rights reserved.

mixed, a-wave amplitude of theplitude of the ERG mixed rodphotopic cone response; bscot,RG, Electroretinography; IGF-I,; IPL, Inner Plexiform Layer;OP, Oscillatory Potentials.and characterized the generaland PdlV carried out the ERG

ry study. LR-dlR, SM-C and FGNC designed and coordinated

read and approved the final

logía, Facultad de Medicina,es, Spain. Fax: +34 918854525.

irect.com).

rights reserved.

Introduction

Patients with homozygous mutations of the human IGF1 genecausing inactivation of the IGF-I molecule present severe prenatalgrowth retardation, postnatal growth failure, microcephaly, mentalretardation and bilateral sensorineural deafness [(Bonapace et al.,2003; Walenkamp et al., 2005; Woods et al., 1997; Woods et al.,1996) ORPHA73272, OMIM 608747]. Family members carrying aheterozygous mutation have lower weight at birth and lowerheight in adulthood, but no early hearing loss has been reported(Woods et al., 1996), although several reports point to a correla-tion between serum IGF-I levels and presbyacusis (Murillo-Cuestaet al., 2011).

Mouse models are a powerful tool for the discovery and character-ization of genes for sensorineural defects in humans. Mouse models

477L. Rodriguez-de la Rosa et al. / Neurobiology of Disease 46 (2012) 476–485

for hearing deficit have allowed the study of the genetic and molecu-lar basis of both syndromic and non-syndromic human hearing im-pairment (http://hereditaryhearingloss.org; Rodríguez-de la Rosa etal., 2011).

IGF-I has a central role in inner ear development and functionthroughout evolution (Sanchez-Calderon et al., 2007). Peak expres-sion of IGF-I in the cochlea occurs during the late embryonic and neo-natal periods, with a reduced expression in the adult. IGF-I actions areexerted through the regulation of intracellular signaling pathways,and the modulation of the transcription factors FoxP3, FoxM1, andMEF2 (Magarinos et al., 2010; Sanchez-Calderon et al., 2007). Igf1−/−

mice develop a smaller cochlea and present congenital sensorineuraldeafness and age-related metabolic cochlear alterations (Camarero etal., 2001; Camarero et al., 2002; Cediel et al., 2006; Riquelme et al.,2010; Sanchez-Calderon et al., 2010). Interestingly, the comparativestudy of gene expression in the Igf1−/− mice showed important alter-ations in the expression levels of genes associated with retinal devel-opment (mammalian achaete–scute homolog 1, fibroblast growthfactor 15 and sine oculis-related homeobox 6) and physiopathology(tubby candidate gene, retinitis pigmentosa 1, harmonin Usher syn-drome 1C and RAR-related orphan receptor beta) (Sanchez-Calderon et al., 2010).

For the study of retinal degeneration, diverse animal modelshave been used, including the rd1 mice (Bowes et al., 1990;Farber and Lolley, 1974; Strettoi and Pignatelli, 2000; Strettoi etal., 2003; Strettoi et al., 2002), the rd10 mice (Barhoum et al.,2008; Chang et al., 2002; Gargini et al., 2007), RCS rats (Cuencaet al., 2005; Herron et al., 1974; Milam et al., 1998; Villegas-Perez et al., 1998) and p23H rats (Cuenca et al., 2004). The spatialpattern of degeneration of the retina in these models quite resem-bles the human illness. Nonetheless, this does not occur with re-gard to the temporal pattern, since all these animals rapidly loseall their photoreceptor cells, contrarily to that described in the ma-jority of the human retinal dystrophy. Thus, the rd1 micecompletely lose the photoreceptor cells at the outer nuclear layer(ONL) in just three weeks (Bowes et al., 1990; Chang et al.,2002; Farber and Lolley, 1974), whereas the rd10 does by onemonth of age (Chang et al., 2002). In the RCS and P23H rats the de-generation progress is slower than in the rd1 and rd10 mice. In RCSrats the degeneration of the photoreceptors progresses during thefirst three months of life of the animal (Cuenca et al., 2005). Inthe P23H rats the degeneration begins early, so that at 20 days ofage a substantial loss of rods is already observed (Cuenca et al.,2004), whereas by 40 days of age only 1–3 rows of photoreceptorsare observed in the ONL of the retina and only a few photorecep-tors are identified by nine months of postnatal life (Cuenca et al.,2004). With regard to the functional variations detected in thesemodels, they do not appear to correspond to the human illness.In this way, in the rd1 mice the electroretinography responses me-diated by rods do not become normal at any age (Strettoi et al.,2003), whereas in the rd10 mice responses begin to attenuate by20 days of age and become undetectable for two months of age(Barhoum et al., 2008; Gargini et al., 2007).

In this context, members of the team previously assessed themorphological and immunohistological alterations of the auditoryreceptor of Igf1−/− mice (Camarero et al., 2001; Camarero et al.,2002) as well as their functional correlations (Cediel et al., 2006;Riquelme et al., 2010), but no information is currently availableon their visual function. To evaluate the role of IGF-I deficit in vi-sual function, we studied the retinal electrophysiological and im-munohistochemical features of Igf1−/− mice. Through assessmentof electroretinography (ERG) and selective labeling of retinalcells, we show here that Igf1−/− mice show age-related accelerat-ed loss in visual function, that may enlighten the physiopathologyof human IGF-I deficiency and resistance (OMIM entries 608747and 147370).

Materials and methods

Mouse handling and genotyping

Heterozygous mice with a targeted disruption of the Igf1 genewere bred andmaintained on a hybrid MF1 and 129/sv mouse geneticbackground to increase Igf1−/− mutant survival (Liu et al., 1993).Igf1−/− mice mortality before adulthood is high, although between20 and 30% survived. Igf1+/+, Igf1+/− and Igf1−/− mice were studiedat the time points indicated to follow their progression from youngadults (1 month) to aged mice (1 year). Mouse genotypes were iden-tified as described (Sanchez-Calderon et al., 2010). All procedureswere carried out in accordance with the European Council Directive(86/609/ECC) and ARVO statement for use of animals in ophthalmicand vision research with the approval of the Bioethics Committee ofthe CSIC.

Analytical procedures

The determination of serum concentration of circulating IGF-I wasmeasured using a standard OCTEIA Rat/Mouse IGF-I kit (sensitivity63 ng/ml and variability 4–8%) (IDS Ltd., Boldon, UK) according tothe manufacturer's recommendations as reported in Riquelme et al.,2010. ELISA data are expressed in ng/ml. Statistical comparisons ofIGF-I sera levels between the different age groups were performedwith a Mann–Whitney rank sum test using GraphPad InStat 3.06(Software Inc., San Diego, CA, USA). The results were considered tobe statistically significant when p≤0.05.

Animals were housed in a 12 h dark/light cycle and fed standardlaboratory diet ad libitum with free access to tap water. Mice from 1to 12-month-old were used for weight and blood glucose measures.For glycemia measurement, mice were fasted overnight and glucoselevels in blood samples collected from the tail vein were determinedusing an automatic analyzer (Glucocard Gmeter; Menarini Diagnos-tics, Badalona, Spain).

Auditory brainstem response (ABR)

Mice were anesthetized by intraperitoneal administration of keta-mine (Imalgene® 500, Merial, 100 mg/kg) and xylazine (Rompum ©2%, Bayer Labs, 10 mg/kg), and they were maintained at 37 °Cthroughout the testing period to avoid hypothermia. Both femaleand male mice were used; no sex-associated parameters were identi-fied in this study. Acoustic stimulation and auditory evoked potentialamplification and recording were performed with TDT System 3™workstation and the specific software SigGenRP™ and BioSigGenRP™(Tucker Davis Technologies, Alachua FL 32615), as described previ-ously (Cediel et al., 2006) with the modifications reported inRiquelme et al. (2010). The following parameters were analyzedfrom waves registered during the ABR tests: auditory thresholds inresponse to click and tone burst stimuli, peak and interpeak latenciesand amplitude-intensity and latency-intensity curves.

Electroretinography (ERG)

Mice were adapted to darkness overnight and subsequently, thewhole manipulation was performed in dim red light. Mice were anes-thetized with an intraperitoneal injection of a solution of ketamine(95 mg/kg) and xylazine (5 mg/kg) and maintained on a heatingpad at 37 °C. The pupils were dilated by applying a topical drop of1% tropicamide (Colircusí Tropicamida, Alcon Cusí, El Masnou, Barce-lona, Spain). A topical drop of 2% Methocel (Ciba Vision, Hetlingen,Switzerland) was instilled in each eye immediately before situatingthe corneal electrode. Flash-induced ERG was recorded from theright eye in response to light stimuli produced with a custom madeGanzfeld stimulator. The intensity of light stimuli was measured

478 L. Rodriguez-de la Rosa et al. / Neurobiology of Disease 46 (2012) 476–485

with a photometer (Mavo Monitor USB, Gossen, Nürenberg, Germa-ny) at the level of the eye. For each light intensity stimulus, 4–64 con-secutive stimuli were averaged, with an interval between light flashesin scotopic conditions of 10 s for dim flashes and up to 60 s for thehighest intensity. Under photopic conditions the interval betweenlight flashes was fixed at 1 s. The ERG signals were amplified andband filtered between 0.3 and 1000 Hz with a Grass amplifier(CP511 AC amplifier, Grass Instruments, Quincy, MA). Electrical sig-nals were digitized at 20 kHz with a Power Lab data acquisitionboard (AD Instruments, Chalgrove, UK). Bipolar recording was per-formed between an electrode fixed on a corneal lens (Burian-Allenelectrode, Hansen Ophthalmic Development Lab, Coralville, IA) anda reference electrode located in the mouth, with a ground electrodelocated in the tail. Rod mediated responses were recorded underdark adaptation to light flashes of −2 log cd·s·m−2. Mixed rod andcone mediated responses were recorded in response to light flashesof 1.5 log cd·s·m−2. Oscillatory potentials (OP) were isolated usingwhite flashes of 1.5 log cd·s·m−2 in a recording frequency range of100–10,000Hz. Cone mediated responses were recorded in responseto light flashes of 2 log cd·s·m−2 on a rod saturating background of30 cd·m−2. The amplitudes of the a-wave and b-wave and peak topeak OP were measured off-line and the results averaged (n=4–8per group). Measurements were performed by an observer blind tothe experimental condition of the animal.

Statistical analysis

Statistical analysis of ABR and ERG data was performed using SPSSv19.0 software, and each group of age was considered independent. Ageneral linear model procedure with analysis of the variance(ANOVA) was carried out. Post hoc multiple comparisons includedBonferroni test when equal variances are assumed and Tamhanetest when equal variances are not assumed. Data are expressed asmean±SD or SEM. The results were considered significant at pb0.05.

Immunohistochemistry

Animals were killed with a lethal dose of pentobarbital, and theeyes were enucleated and immersion-fixed with 4% paraformalde-hyde for two hours at 4 °C, and then washed with 0.1 M Na-phosphate buffer, pH 7.4. The eye cups were cryoprotected sequen-tially in 15, 20 and 30% sucrose. The lens were removed and eyescups were embedded in Tissue-Tek OCT (Sakura Finetek, Zoeterwou-den, Netherlands) and frozen in liquid N2. Sixteen μm-thick sectionswere obtained in a cryostat (Leica CM 1900, Leica Microsystems)mounted on Plus glass slides. (Thermo Scientific, Madrid, Spain) andair dried. For blocking non-specific staining, sections were incubatedin 10% normal donkey serum for 1 h (Jackson, West Grove, PA, USA)with 0.5% Triton X-100 and then incubated overnight at room tem-perature with combinations of different primary antibodies dilutedin PBS containing 0.5% Triton X-100. All primary antibodies used in

Table 1Primary antibodies used in this work.

Molecularmarker

Antibody Source Workingdilution

Bassoon Mouse, cloneSAP7F407

Stressgen(Ann Arbor, MI),VAM-PS003

1:1000

Calbindin D-28K Rabbitpolyclonal

SWant (Bellinzona,Switzerland), CB-38

1:500

Proteinkinase C (PKC),α isoform

Rabbitpolyclonal

Santa CruzBiotechnology,sc-10800

1:100

Synaptophysin Mouse,clone SY38

Chemicon, MAB5258 1:500

this study had been previously shown to be useful in the mouse retina(Table 1). Subsequently, the sections were washed in PBS and incu-bated in the secondary antibodies, Alexa Fluor 488-conjugated don-key anti-rabbit IgG (green) and Alexa 546-conjugated donkey antimouse IgG (red, Molecular Probes, Eugene, OR, USA) at a 1:100 dilu-tion for 1 h. The nuclear dye TO-PRO-3 iodide (Molecular Probes) wasadded at 1 μM simultaneously to secondary antibodies and slideswere incubated at room temperature. The sections were finallywashed in PBS, mounted in fluoromount Vectashield (Vector Labora-tories) and coverslipped for viewing by laser-confocal microscopy(Leica TCS SP2 Leica Microsystems). Immunohistochemical controlswere performed by omission of either the primary or secondary anti-bodies. Unless otherwise indicated, images were obtained from cen-tral retinal sections. TIFF images were enhanced using AdobePhotoshop CS3.

Results

Igf1 gene-dosage determines differences in the mouse phenotype

The body weight of Igf1+/+, Igf1+/− and Igf1−/− mice was mea-sured at 1, 3, 6, 9 and 12 months of age. All genotypes showed anage-related increase in body weight, although it was lower for theIgf1−/− mouse (Fig. 1A). The greatest body weight gain occurred dur-ing the first trimester of life in all cases, while in the following months(3–9) the body weight was maintained (Igf1−/−) or showed a smallincrease (Igf1+/+ and Igf1+/−). Finally, from 9 to 12 months of agethere was a small decrease in the weight of Igf1+/+ and Igf1−/−

mice. The body weight of Igf1+/+ and Igf1+/− mice was comparedwith that of Igf1−/− mice and the correlation was statistically signifi-cant at all ages, while the comparison between Igf1+/+ and Igf1+/−

mice was only significant between the ages of 3 and 9-month-old.The fasting blood glucose levels in each of the three genotypes weremeasured between one month and one year of age (Fig. 1B). Igf1+/+

and Igf1+/− mice showed similar values of glucose that were main-tained throughout the study. However, Igf1−/− mice showed an aver-age value of glucose level greater than those of Igf1+/+ and Igf1+/−

mice during the first month of life and then decreased graduallyfrom 6 to 12 months of age. Comparisons between different geno-types and ages did not show significant differences. Sera fromIgf1+/+, Igf1+/− and Igf1−/− mice were collected at several timesbetween 1 and 12 months of age to analyze the circulating IGF-Ilevels using a specific ELISA assay. As expected, the Igf1−/− micehad no detectable levels of IGF-I in any of the ages studied(Fig. 1C) and secondly, Igf1+/− mice presented mean values ofIGF-I levels lower than those of Igf1+/+ mice. Moreover, Igf1+/−

mice showed a statistically significant progressive age-related de-crease in serum IGF-I levels, measured between 1 and 12 monthsof age (data not shown).

To study the hearing phenotype, ABR recordings of Igf1+/+,Igf1+/− and Igf1−/− mice were obtained at 1, 3, 6, 9 and 12 monthsof age, at least 6 mice per group were tested. ABR thresholds in re-sponse to click stimulus in Igf1+/+ and Igf1+/− mice indicated nor-mal hearing from one month up to six months of age, whereasIgf1−/− mice showed a profound deafness (Figs. 1D, E). Statisticallysignificant differences were found when Igf1−/− mice were com-pared with the other genotypes at 1, 3 and 6 months of age. Ac-cordingly, the audiogram of Igf1−/− mice presented elevatedthresholds in response to pure frequencies (8–28 kHz). The com-parison of the audiograms of 1-month-old Igf1−/− mice withthose of Igf1+/+ and Igf1+/− mice showed highly statistically signif-icant differences for all the frequencies studied (Fig. 1F). However,the evolution of ABR thresholds from 6 to 12 months of age differedbetween Igf1+/+ and Igf1+/− mice. Igf1+/− mice exhibited an earli-er increase of the thresholds, especially for the high frequencies, al-though no statistical differences were found compared to Igf1+/+

Fig. 1. Temporal evolution of body weight, glycemia, IGF-I levels and hearing loss. (A) Body weight of Igf1+/+ (open circles; WT), Igf1+/− (open triangles; Hz) and Igf1−/− (closedcircles; KO) mice was measured. Weight data of WT and Hz mice were statistically significant at all ages studied compared with that of KO (***, pb0.001), while significant differ-ences between Igf1+/+ and Igf1+/− were shown only in mice from 3 to 9 month-old (§§§, pb0.001; §, p=0.014). At least 80 mice were studied per genotype. (B) Glycemia levelswere measured in Igf1+/+, Igf1+/− and Igf1−/− fasted mice from blood samples. Glucose levels in Igf1+/+ and Igf1+/− mice were similar throughout the study, but Igf1−/− nullmouse glucose values were not steady. Comparisons did not show significant differences. At least 28 mice/genotype were measured. (C) Serum IGF-I levels. Igf1−/− mice showedno detectable levels of IGF-I during the study. Taken together all the ages studied, the mean values of IGF-I levels were lower in Igf1+/− mice than those of Igf1+/+ mice (*,p=0.026). Circulating levels of IGF-I were analyzed in at least 8 mice/genotype. (D) Representative ABR recordings in response to click stimulus of Igf1+/+ (left column), Igf1+/−

(middle column) and Igf1−/− (right column) mice, at 1, 6, 9 and 12 months of age. Wild type and heterozygous mice showed a normal pattern of ABR waves up to six months,whereas null mice exhibited a congenital profound deafness. (E) ABR thresholds in response to click stimulus in Igf1+/+ (open bars), Igf1+/− (closed bars) and Igf1−/− mice (graybars) at different ages. Igf1+/+ and Igf1+/− mice showed an age-related increase in ABR thresholds, whereas the Igf1−/− mice were deaf from the youngest age studied. ***, pb0.001;###, pb0.001; †, p=0.033. (F) ABR thresholds in response to tone burst stimuli (8–28 kHz) in Igf1+/+, Igf1+/− and Igf1−/− mice. The audiogram in the Igf1−/− mice was alwayselevated. ***, pb0.001; ###, pb0.001; ##, p=0.001 (6 months, 16 and 28 kHz; 9 months, 8 and 28 kHz) or p=0.003 (6 months, 20 kHz); #, p=0.046; ††, p=0.006; †, p=0.024(9 months, 8 kHz), p=0.013 (9 months, 28 kHz) or p=0.032 (12 months, 28 kHz). (G) Peak I latency-intensity curves after click stimulation in Igf1+/+, Igf1+/− and Igf1−/− mice.The null mutant mice showed an elevated curve compared to wild type and heterozygous mice. Latency-intensity curves at 12 months of age were similar in all the genotypes.Statistical analysis was performed with ANOVA and post hoc Bonferroni test. Data are presented as the mean±SEM. * indicates comparison between Igf1−/− and the other geno-types; # between Igf1−/− and Igf1+/+ mice; † between Igf1−/− and Igf1+/− mice and §§§ between Igf1+/− and Igf1+/+ mice. Degrees of freedom were 2 and F values were >1 for allthe ANOVAs.

479L. Rodriguez-de la Rosa et al. / Neurobiology of Disease 46 (2012) 476–485

mice (Figs. 1D–F). On the other hand, Igf1−/− mice maintained highABR thresholds, with significant differences when compared to ei-ther Igf1+/+ or Igf1+/− mice at 9 months of age. Finally, micefrom the three genotypes showed similar increased ABR thresholdsin response to click and tone bursts at 12 months of age, withoutany relevant statistical differences (Figs. 1E, F). Fig. 1G shows thelatency-sound intensity curve for peak I in the three genotypes atdifferent ages, indicating that, at the same sound level, Igf1−/−

mice presented an increased peak I latency, compared to Igf1+/+

and Igf1+/− mice.

Electroretinography reveals altered retinal function in homozygousIgf1−/− mice

To determine possible effects of IGF-I deficiency on retinal func-tion, a series of ERG experiments was performed on Igf1−/− mice,and on their corresponding control littermates Igf1+/− and Igf1+/+,respectively at different times of animal development (Fig. 2).

Initially, visual function was tested in 2-month-old mice (p60).Fig. 2A shows the retinal responses to light stimuli on Igf1−/−,Igf1+/− and Igf1+/+ mice obtained under dark and light adaptationconditions. The rod-driven circuitry was tested under scotopic condi-tions. At p60, the amplitude of the scotopic b-wave (bscot), did not

show significant differences between wild type and mutant mice(Fig. 2B). Under scotopic conditions, light stimuli of high intensityevoked mixed responses. The a-wave amplitude measured in theseconditions (amixed) reflects the functionality of rod and cone photore-ceptors. Averaged data for the amixed amplitude did not show signifi-cant differences between the three genotypes by p60 (Fig. 2B). The b-wave amplitudes measured in response to high-intensity stimuli(bmixed) did not show any significant differences among Igf1−/−,Igf1+/− and Igf1+/+ mice (Fig. 2B). To determine the contributionof third order neurons to light-induced ERG, oscillatory potentials(OP) were isolated from the electrophysiological recordings. TheOP recorded in response to high-intensity light stimuli under scoto-pic conditions showed no significant differences among Igf1−/−,Igf1+/− and Igf1+/+ mice (Fig. 2B). The cone driven circuitry wastested in p60 old mice under photopic conditions by measuringthe ERG response to intense flashes of light in the presence ofrod-saturating light stimulation. Photopic b-wave responsesrecorded from p60 old mice to light stimuli of very high intensity(bphot) did not show significant modifications among Igf1−/−,Igf1+/− and Igf1+/+ mice (Fig. 2B).

Visual function was tested in 4-month-old mice (p120) (Figs. 2A,B). The amplitude of the scotopic b-wave, did not show significantdifferences between wild type Igf1+/+ and Igf1+/− mice, although a

Fig. 2. Temporal evolution of visual loss in Igf1+/+, Igf1+/− and Igf1−/− mice by electroretinography recordings. (A) Representative ERG recordings of Igf1+/+ (left column, WT),Igf1+/− (middle column, Hz) and Igf1−/− (right column, KO) mice, at 2, 4, and 12 months of age. Rod responses, mixed responses and oscillatory potentials (OP) were recordedunder dark adapted conditions; cone responses were recorded under light adapted conditions. Igf1+/+, Igf1+/− and Igf1−/− mice showed normal ERG responses up to two monthsof age. Igf1+/+ and Igf1+/− mice showed normal ERG responses by 4 months of age, whereas Igf1−/− mice exhibited a significant decrease of ERG wave amplitudes. Igf1+/+ miceshowed normal ERG responses by 12 months of age, whereas Igf1+/− mice exhibited a significant decrease of ERG wave amplitudes and Igf1−/− mice showed almost no ERG lightresponses. Vertical calibration corresponds to 100 μV for rod, mixed and OP responses, and 50 μV for cone responses. Horizontal calibration corresponds to 150 ms for rod, mixedand cone responses, and 50 ms for OP responses. (B) Histogram representation of the ERG wave amplitudes averaged from Igf1+/+ (WT, n=8), Igf1+/− (Hz, n=8) and Igf1−/− mice(KO, n=5) at different ages (2, 4, and 12 months). The ERG wave amplitudes (bscot, amixed, bmixed, OP and bphot) of the light responses were measured as shown on traces in A.Statistical analysis was performed with ANOVA and post hoc Bonferroni tests. Data are presented as the mean±SD. * indicates comparison between Igf1−/− and Igf1+/+ mice; #indicates comparison between Igf1−/− and Igf1+/− mice; † indicates comparison between Igf1+/− and Igf1+/+ mice. The ERG response amplitudes of the Igf1+/+, Igf1+/− andIgf1−/− mice showed no statistical differences by p60. The ERG response amplitudes of the Igf1−/− mice showed a significant decrease by p120 and p360 when compared withIgf1+/+ mice (***, pb0.001). The ERG response amplitudes of the Igf1+/− mice also showed a significant decrease by p360 when compared with Igf1+/+ mice (†††, pb0.001). Com-parison between of the Igf1+/− and Igf1−/− ERG response amplitudes also showed significant differences by p120 (##, p=0.001 [bscot], p=0.003 [amixed, bmixed] or p=0.008 [OPand bphot]) and p360 (###, pb0.001, #, p=0.022 [bmixed] or p=0.031 [OP]). F values were >10, degrees of freedomwere 2 and for all the ANOVAs. (C) Age dependent reduction inthe ERG mediated responses in the Igf1+/+ (WT), Igf1+/− (Hz) and Igf1−/− (KO) mice. Averaged bscot, amixed, bmixed, OP and bphot amplitudes from ERG recordings shown in B forp60, p120 and p360, and those averaged by p180 and p270 are represented as a function of postnatal age. Data are presented as the mean±SD.

480 L. Rodriguez-de la Rosa et al. / Neurobiology of Disease 46 (2012) 476–485

statistically significant reduction was observed in Igf1−/− mice whencompared with Igf1+/+ or Igf1+/− mice. The a- and b-wave ampli-tudes from mixed responses neither showed significant differencesamong Igf1+/− and Igf1+/+ mice genotypes by p120, but again, a sta-tistically significant reduction in the amixed wave was consistently ob-served in Igf1−/− mice when compared with Igf1+/+ or Igf1+/− mice.The amplitude of the bmixed wave in the in Igf1−/− mice also showedstatistically significant differences between the Igf1−/− and Igf1+/+

mice. Oscillatory showed no significant differences between Igf1+/−

and Igf1+/+ mice by p120, but a statistically significant decrease inOP amplitude was observed in p120 Igf1−/− mice when comparedwith Igf1+/+ (pb0.001) or Igf1+/− mice. Photopic b-wave responsesrecorded from p120 old mice showed no significant differences be-tween Igf1+/− and Igf1+/+ mice by p120, but again, a significant de-crease in bphot amplitude was observed in Igf1−/− mice by p120when compared with Igf1+/+ mice.

Visual function was also tested in 6 and 9-month-old mice (p180and p270) of Igf1−/−, Igf1+/− and Igf1+/+ genotypes. As a rule, a de-crease in the ERG wave amplitudes was observed for all mice geno-types, being differences statistically significant between Igf1−/− andIgf1+/+ genotypes (data not shown). Visual function was finally test-ed in 1-year-old mice (p360) of the three genotypes (Figs. 2A,B). The

amplitude of the scotopic b-wave was significant reduced in theIgf1+/− mice, when compared with Igf1+/+ mice and was almostnull in the Igf1−/− mice. Mixed responses were also significantly re-duced in the Igf1+/− mice. The a- and b-wave amplitudes of themixed response showed statistically significant differences withIgf1+/+ mice. Mixed responses were also almost null in the Igf1−/−

mice. Similarly, OP showed significant differences among Igf1−/−,Igf1+/− and Igf1+/+ mice by p360. Photopic b-wave responsesrecorded from p360 old mice also showed significant differencesamong Igf1−/−, Igf1+/− and Igf1+/+ mice. Average data on ERGwave amplitudes and statistically significant analyses are shown inFig. 2B for 2, 4 and 12-month-old mice.

Plot representation of averaged ERG wave amplitudes as a func-tion of animal ages showed the time course of retinal dysfunction inall three genotypes (Fig. 2C). As it has been previously shown, wildtype animals suffered a slight decrease in ERG wave amplitudesalong life (Li et al., 2001); for one year old animals, amplitudes ofthe ERG waves decreased one third of their initial values. Our workin the Igf1−/− and the Igf1+/− mice showed significant differenceswith the Igf1+/+ animals. From our analysis we appreciated thatIgf1−/−, Igf1+/− and Igf1+/+ mice showed equally normal functionat young ages (p60). By p120, rod and cone mediated responses

481L. Rodriguez-de la Rosa et al. / Neurobiology of Disease 46 (2012) 476–485

were decreased in the Igf1−/− mice, when compared with Igf1+/−

and Igf1+/+ mice. From this age, a linear steep reduction in the rodand cone mediated responses was clearly observed in the Igf1−/−

mice with age, showing an almost absent light response for 1-year-old animals. The Igf1+/− mice also showed a linear reduction of ERGwave amplitudes from p120, although detectable ERG light responsesmay be still observed at one year of animal age. Our electrophysiolog-ical results indicated that retinal function in the Igf1−/− mutants wasseverely altered by p360, and that the dysfunction may mainly affectthe cells postsynaptic to rod and cone photoreceptors.

Immunohistochemistry reveals altered retinal structure in homozygousIgf1−/− mice

To study changes in retinal morphology in the Igf1−/− animals weused TOPRO to stain the nucleus of all cell types (Fig. 3). Vertical sec-tion of Igf1+/+ and Igf1−/− retinas at four months of age showed nodifferences in retinal thickness of any retinal nuclear layer (Figs. 3Aand B). At 12 months of age, only the thickness of the outer plexiformlayer (OPL) in the Igf1−/− was reduced (Fig. 3D) when comparedwith Igf1+/+ animals (Fig. 3C). Since we have found a decrease inthe ERG b-wave, we decided to explore whether connectivity markersassociated with rods, cones and their postsynaptic cells may be al-tered in the transgenic Igf1−/− mice. In the mouse retina, there is asingle type of ON rod bipolar cell, which is immunoreactive for PKC(protein kinase C alpha). The rod bipolar cell bodies are mostly

Fig. 3. General view of the Igf1+/+ and Igf1−/− retinas stained with TOPRO 3 showingall retinal cells layers. No differences in layer thickness were found at 4 (A,B) and12 months (C,D) between Igf1+/+ and Igf1−/− mice. Only the thickness of the OPL inthe Igf1−/− mouse at 12 months was thinner that in wild type mouse. Scale bar repre-sents 20 μm.

aligned in the outermost part of the inner nuclear layer (INL). Eachrod bipolar cell has a single primary dendrite, and a tuft of dendriticterminals establishes connections with rod spherules through alarge dendritic arbor in the OPL. Their single axon runs perpendicular-ly through the IPL, ending in its innermost stratum as large axon ter-minal end-bulbs, together with some lateral terminal varicosities,close to the ganglion cell layer. In order to identify synaptic contactsbetween photoreceptors and ON rod bipolar cells in the OPL duringaging, antibodies against bassoon and PKC were used. Bassoon is apresynaptic protein located at the synaptic ribbon in cone and rodterminals in the OPL. In the inner plexiform layer (IPL), bassoon isconcentrated at conventional GABAergic amacrine synapses but is ab-sent from the bipolar cell ribbon synapses (Brandstatter et al., 1999).We used antibodies against PKC to identify the dendritic terminals ofrod bipolar cells that represent one of the postsynaptic elements torod photoreceptors. Double labeling with bassoon and PKC showedthe relationship between photoreceptors and dendritic terminals ofrod bipolar cells (Fig. 4). At 2 months of age, no differences in thepaired bassoon-PKC staining were found in the Igf1+/− and Igf1−/−

compared with Igf1+/+ (Figs. 4A, B and C). At 4 months of age,Igf1−/− retinas showed a retraction of rod bipolar dendrites and a de-crease in the synaptic contacts between rod spherules and ON rod bi-polar dendrites compared with Igf1+/− and Igf1+/+ retinas (Figs. 4D,E and F). The most remarkable differences in connectivity betweenphotoreceptors and ON rod bipolar cells were found in Igf1−/− retinasat 12 months of age, where a loss of ON rod bipolar cell dendrites wasevident and a few paired contacts could be observed (Fig. 4I). Also,bassoon immunostaining was mislocated, moving from the OPL tothe outer nuclear layer, around photoreceptor cell bodies (Fig. 4I). Al-though the loss of contact in the Igf1+/− (Fig. 4H) was not as dramaticas in the Igf1−/−, differences with Igf1+/+ (Fig. 4G) could be observed,with less contacts between rod terminal tips and bassoonimmunoreactivity.

We studied further cell connections between horizontal cells andaxon terminals of photoreceptors. Horizontal cells can be identifiedusing antibodies against calbindin (Fig. 5). In the mouse retina, den-dritic tips of horizontal cells make contacts with the cone pediclewhile horizontal cell axon terminal tips contact rod photoreceptorsin the OPL, at the rod spherules. A continuous plexus in the OPL andtip terminal of horizontal cells may be easily identified in the 2-month-old Igf1+/+ mouse retina (Fig. 5A). No differences werefound in the horizontal morphology and tip terminal in Igf1+/− andIgf1−/− (Figs. 5A, B and C). At 4 months of age, the regular anddense plexus of horizontal cell processes and tip terminals ofIgf1+/− and Igf1−/− mice were different from Igf1+/+, whereas aclear reduction of dendrites and axon terminal tips was observedin Igf1−/− mice (Figs. 5D, E and F). At 12 months of age, horizontalcell processes in Igf1+/− mice were diminished and a decrease ofterminal tips could be observed compared with Igf1+/+. In Igf1−/−

mice, calbindin-immunoreactive terminal tips were difficult to rec-ognize, and the horizontal cell plexus showed gaps in the OPL(Figs. 5G, H and I).

To identify if photoreceptor axon terminals were lost with aging inIgf1−/− mice, synaptophysin immunostaining was analyzed. The 2-month-old Igf1+/+ mouse showed a continuous and thick immunore-activity band of synaptophysin located on top of the horizontal cellterminals. At this age, retinas of the Igf1+/− and Igf1−/− mice didnot show significant differences in synaptophysin immunoreactivitywhen compared with Igf1+/+ (Figs. 6A, B and C). At 4 months ofage, a decrease of photoreceptor axon terminals immunolabelingwas observed in the Igf1−/− compared with Igf1+/− and Igf1+/+

(Figs. 6D, E and F). Finally, at 12 months of age (Figs. 6G, H and I),immunostaining for the synaptophysin in the Igf1−/− mouse retinawas no longer distributed as a continuous layer, and it showed label-ing gaps corresponding to the lack of horizontal cells processes in theOPL. Also a thin band of photoreceptor axon terminals was located on

Fig. 4. Synaptic contacts between rod photoreceptors and rod bipolar cell dendrites in the Igf1+/+, Igf1+/− and Igf1−/− mice. Immunostaining for rod bipolar cells (PKC, green) andsynaptic ribbon (bassoon, red). (A) The interaction of ON rod bipolar cell dendrites with paired photoreceptor synaptic ribbons could be observed in the Igf1+/+ mouse at 2 months.(B) The picture shows the synaptic contacts of ON rod bipolar cells with paired photoreceptors in the Igf1+/− mouse at 2 months. (C) At two months of age, the Igf1−/− mouseshowed no apparent changes in connectivity between ON rod bipolar cells and rod axon terminals compared with Igf1+/− and Igf1+/+ mice. Retraction of bipolar cell dendritesand loss of photoreceptor synaptic ribbon were observed in Igf1−/− mouse by 4 months of age (F), compared with Igf1+/− (E) and Igf1+/+ mice (D). At 12 months of age, loss ofsynaptic contacts between ON rod bipolar and photoreceptors (arrowheads) was evident in the Igf1−/− (I) compared with Igf1+/− (H) and Igf1+/+ mice (G). Scale bar represents20 μm.

482 L. Rodriguez-de la Rosa et al. / Neurobiology of Disease 46 (2012) 476–485

top of the horizontal cells compared with Igf1+/+, this indicating thataxon terminals were lost with aging in the Igf1−/− mouse. At this age,Igf1+/− retinas showed intermediate synaptophysin immunoreactiv-ity between Igf1+/+ and Igf1−/− mice.

Discussion

The present work on the IGF-I deficient mouse supports the use ofnewmouse models for the study of human syndromic deaf-blindness.The study on Igf1−/− mice, Igf1+/− mice and Igf1+/+ mice allowed usto compare the serological levels of IGF-I with phenotype manifesta-tions in body weight and glycemia and a good correlation was ob-served. Moreover, we showed that Igf1−/− mice suffer a profounddeafness from birth, while the Igf1+/− mice start to loose auditory ca-pacity by 6 months of age. Although auditory function is also affectedin the Igf1+/+ mice, the process seems to be slower in Igf1+/+. Allthree mice genotypes become equally deaf by 12 months of age. Con-trary to the auditory function, visual function in the Igf1−/− mice isnormal until four months of age. From then, a progressive loss in visu-al function is observed in the Igf1−/− mice becoming almost blind at

twelve months of age. The evolution of the visual function in theIgf1+/− mice is also affected, and a decrease in retinal responses is ob-served by 6 months of age, although no complete blindness is reachedby 12 months of age. Visual impairment in the IGF-I deficient mouseis parallel to a decrease in cell contacts at the first synapse of the vi-sual pathway, and no clear decrease of retinal cell numbering is ob-served in this mouse model of syndromic blindness.

Patients that suffer homozygous mutations of the human IGF1gene present severe bilateral sensorineural deafness (Walenkampand Wit, 2007; Murillo-Cuesta et al., 2011). Accordingly, the Igf1−/−

mice experience a profound deafness from birth, due to an abnormaldevelopment of the auditory nervous system, where IGF-I plays a crit-ical role (Sanchez-Calderon et al., 2010). On the contrary, develop-ment of the retina in the Igf1−/− mice seems not to becompromised by the deficiency in IGF-I, which seems to be criticalfor the normal vascularization of the mammalian retina, but not forits development (Calvaruso et al., 1996; Hellstrom et al., 2002).Therefore, normal retinal function and structure are observed in theIgf1−/− mice at birth. Human heterozygous mutation in the IGF-Igene has lower weight at birth and lower height in adulthood, but

Fig. 5. Horizontal cell modifications in the Igf1+/+, Igf1+/− and Igf1−/− mice. Confocal fluorescence micrographs of retinal cross sections showing horizontal cell using antibodiesagainst calbindin. (A) In the 2-month-old Igf1+/+ mouse, horizontal cell terminal tips were easily identified (arrows). At this age, no loss of horizontal plexus in the OPL and ter-minal tips was found in Igf1+/− compared with Igf1−/− mice retina (B,C respectively). At 4 months of age, the regular and dense plexus of horizontal cell processes and tip terminalsin the OPL in Igf1+/− (E) and Igf1−/− (F) were different from Igf1+/+ mice retinas (D), and there was a clear reduction of dendrites and axon and terminal tips (arrows) comparedwith Igf1+/+ mice. At 12 months of age, (I), the horizontal cell terminal tips showed a clear retraction where sparse calbindin immunoreactivity was shown no longer distributed asa continuous layer in the OPL, compared with Igf1+/− (H) and Igf1+/+ mice (G). Scale bar represents 10 μm.

483L. Rodriguez-de la Rosa et al. / Neurobiology of Disease 46 (2012) 476–485

no early hearing loss (Woods et al., 1996). Accordingly, these muta-tions do not present a phenotype strikingly different to that foundin the wild type mouse. In the Igf1+/− mouse, the hearing loss andthe degeneration of the visual system progress in parallel withaging. From the present experiments, no clear relationship betweenpartial IGF1 gene deficiency and visual progressive loss may beestablished.

Sensorineural visual defects have not been addressed thoroughlyin patients that suffer homozygous mutations of the human IGF1gene. To our knowledge, no visual tests have been shown in these pa-tients to date. IGF-I has been associated with the pathogenesis of dia-betic retinopathy, although its role is not fully understood(Gerhardinger et al., 2001; Hellstrom et al., 2002; Ruberte et al.,2004). The synthesis of IGF-I is decreased in both human and experi-mental diabetes. However, no changes in the activation of its receptorand downstream antiapoptotic effector, nor retinal microvascular cellapoptosis, were found (Gerhardinger et al., 2001). In our experi-ments, the Igf1−/− mouse experiences a decrease in wave amplitudeof the retinal electrical responses that become almost flat at12 months of age. Absence of ERG responses has been addressed inseveral mouse models of retinal degeneration (Chang et al., 2002).In most dystrophic mice, decrease in the number of retinal photore-ceptors parallels the decrease in ERG wave amplitudes (Barhoum etal., 2008; Cuenca et al., 2004; Cuenca et al., 2005; Gargini et al.,

2007; Strettoi and Pignatelli, 2000; Strettoi et al., 2003; Strettoi etal., 2002). However, the current Igf1−/− mouse model does not expe-rience a significant loss of retinal photoreceptors, but a significant lossof cell contacts in the OPL between photoreceptors and their postsyn-aptic cells, bipolar and horizontal cells was found. In a transgenicmice model overexpressing IGF-I no changes in the retinal layerthickness were found and interestingly the specific expression ofIGF-I was detected in the OPL and the inner segment of photore-ceptors (Ruberte et al., 2004). These transgenic mice showed al-tered retinal vascularization at 6 months of age and older(Ruberte et al., 2004).

Since IGF-I has been shown to be involved in synaptic plasticity(Torres-Aleman, 2009), the decrease in ONL synaptic contacts maybe related to the absence of IGF-I and thus may have an effect on ret-inal responses; the decrease of rod photoreceptor-ON rod bipolar cellcontacts would result in a decrease of the ERG b-wave, but should notcompletely compromise the ERG a-wave generated at the photore-ceptor cells. From the present work, we can confirm the structuraland functional alteration of retinal connectivity in the Igf1−/−

mouse that parallels deficit in visual function, but we cannotcompletely explain the absence of a-wave in the mixed ERG re-sponses. It is possible that the lack of IGF-I could affect photoreceptorresponses but not their morphology. Therefore, we cannot discard thepossibility that photoreceptor function may be also affected in this

Fig. 6. Synaptic contacts between photoreceptors and horizontal cell terminals in the Igf1+/+, Igf1+/− and Igf1−/− mice. Confocal fluorescence micrographs of retinal cross sectionsshowing the degeneration of horizontal cell terminals contacting photoreceptor axon terminals, using antibodies against synaptophysin to label photoreceptor terminals (red chan-nel) and calbindin to label horizontal cells (green). (A) The Igf1+/+ mouse retina showed a continuous and thick immunoreactivity band of synaptophysin located on top of thehorizontal cell terminals. At 2 months of age, the Igf1+/− (B) and Igf1−/− (C) mice showed no synaptophysin-IR differences with Igf1+/+ mice. At 4 months of age, a decrease ofaxon terminals immunostained with antibodies against synaptophysin was found in the Igf1−/− (F) compared with Igf1+/− (E) and Igf1+/+ mice (D). At 12 months of age, synap-tophysin immunoreactivity in the Igf1−/− mouse retina (I) was no longer distributed as a continuous layer, and a thin band of axon terminals was located on top of horizontal cellscompared with Igf1+/+. At 12 months of age, Igf1+/− mouse retinas showed intermediate synaptophysin immunoreactivity between Igf1+/+ and Igf1−/− retinas. Scale bar repre-sents 10 μm.

484 L. Rodriguez-de la Rosa et al. / Neurobiology of Disease 46 (2012) 476–485

mouse model and we suggest that the functional role of pigment ep-ithelial cells may be compromised due to the lack of IGF-I (Takagi etal., 1994). It has been previously shown that disruptions of the IGF-Igene results in hypomyelination and interneuronal loss of differentbrain regions (Beck et al., 1995), but to our knowledge, this workshows for the first time that disruptions of the Igf1 gene results in adecrease of the number of cell contacts in a specific brain regionsuch as the retina.

We conclude that the maintenance of normal levels of IGF-I is re-quired for normal visual function and its lack leads to a loss of visionover time. Finally, we propose the Igf1−/− and Igf1+/− mice as newmodels of progressive deaf-blindness of potential use to testnovel IGF-I-based therapies aimed at the protection and repair of sen-sory systems.

Acknowledgments

This research was funded by grants from the Spanish Ministry ofScience and Innovation SAF2010-21879 and RETICS RD07/0062/0008 to PdlV; BFU2009-07793/BFI, Fundaluce, ONCE and RETICSRD07/0062/0012 to NC; and SAF2008-0064, SAF2011-24391 andIntra-CIBERER programs to IV-N. We kindly acknowledge the

technical help received from Mrs. L. Ramirez and the support andcomments from Drs. R. Cediel and J.M. Zubeldia.

References

Barhoum, R., et al., 2008. Functional and structural modifications during retinal degen-eration in the rd10 mouse. Neuroscience 155, 698–713.

Beck, K.D., et al., 1995. Igf1 gene disruption results in reduced brain size, CNS hypo-myelination, and loss of hippocampal granule and striatal parvalbumin-containing neurons. Neuron 14, 717–730.

Bonapace, G., et al., 2003. A novel mutation in a patient with insulin-like growth factor1 (IGF1) deficiency. J. Med. Genet. 40, 913–917.

Bowes, C., et al., 1990. Retinal degeneration in the rd mouse is caused by a defect in thebeta subunit of rod cGMP-phosphodiesterase. Nature 347, 677–680.

Brandstatter, J.H., et al., 1999. Differential expression of the presynaptic cytomatrixprotein bassoon among ribbon synapses in the mammalian retina. Eur. J. Neurosci.11, 3683–3693.

Calvaruso, G., et al., 1996. Insulin-like growth factors in chick embryo retina during de-velopment. Regul. Pept. 61, 19–25.

Camarero, G., et al., 2001. Delayed inner ear maturation and neuronal loss in postnatalIgf-1-deficient mice. J. Neurosci. 21, 7630–7641.

Camarero, G., et al., 2002. Cochlear abnormalities in insulin-like growth factor-1 mousemutants. Hear. Res. 170, 2–11.

Cediel, R., et al., 2006. Sensorineural hearing loss in insulin-like growth factor I-nullmice: a new model of human deafness. Eur. J. Neurosci. 23, 587–590.

Chang, B., et al., 2002. Retinal degeneration mutants in the mouse. Vis. Res. 42,517–525.

485L. Rodriguez-de la Rosa et al. / Neurobiology of Disease 46 (2012) 476–485

Cuenca, N., et al., 2004. Regressive and reactive changes in the connectivity patternsof rod and cone pathways of P23H transgenic rat retina. Neuroscience 127,301–317.

Cuenca, N., et al., 2005. Early changes in synaptic connectivity following progressivephotoreceptor degeneration in RCS rats. Eur. J. Neurosci. 22, 1057–1072.

Farber, D.B., Lolley, R.N., 1974. Cyclic guanosine monophosphate: elevation in degener-ating photoreceptor cells of the C3H mouse retina. Science 186, 449–451.

Gargini, C., et al., 2007. Retinal organization in the retinal degeneration 10 (rd10) mu-tant mouse: a morphological and ERG study. J. Comp. Neurol. 500, 222–238.

Gerhardinger, C., et al., 2001. IGF-I mRNA and signaling in the diabetic retina. Diabetes50, 175–183.

Hellstrom, A., et al., 2002. IGF-I is critical for normal vascularization of the human ret-ina. J. Clin. Endocrinol. Metab. 87, 3413–3416.

Herron Jr., W.L., et al., 1974. Retinal dystrophy in the pigmented rat. Invest. Ophthal-mol. 13, 87–94.

Li, C., et al., 2001. Age-related changes in the mouse outer retina. Optom. Vis. Sci. 78,425–430.

Liu, J.P., et al., 1993. Mice carrying null mutations of the genes encoding insulin-likegrowth factor I (Igf-1) and type 1 IGF receptor (Igf1r). Cell 75, 59–72.

Magarinos, M., et al., 2010. RAF kinase activity regulates neuroepithelial cell prolifera-tion and neuronal progenitor cell differentiation during early inner ear develop-ment. PLoS One 5, e14435.

Milam, A.H., et al., 1998. Histopathology of the human retina in retinitis pigmentosa.Prog. Retin. Eye Res. 17, 175–205.

Murillo-Cuesta, S., et al., 2011. The role of insulin-like growth factor-I in the physiopa-thology of hearing. Front. Mol. Neurosci. 4 (11), 1–17.

Riquelme, R., et al., 2010. A comparative study of age-related hearing loss in wild typeand insulin-like growth factor I deficient mice. Front. Neuroanat. 4. doi:10.3389/fnana.2010.00027.

Rodríguez-de la Rosa, L., et al., 2011. Regulation of cochlear development and function:a scenario for growth factor actions. In: Ahuja, Satpal (Ed.), Usher Syndrome: Path-ogenesis, Diagnosis and Therapy. Nova Science Publishers, Hauppauge, NY (USA).

Ruberte, J., et al., 2004. Increased ocular levels of IGF-1 in transgenic mice lead todiabetes-like eye disease. J. Clin. Invest. 113, 1149–1157.

Sanchez-Calderon, H., et al., 2007. A network of growth and transcription factors con-trols neuronal differentation and survival in the developing ear. Int. J. Dev. Biol. 51,557–570.

Sanchez-Calderon, H., et al., 2010. RNA microarray analysis in prenatal mouse cochleareveals novel IGF-I target genes: implication of MEF2 and FOXM1 transcription fac-tors. PLoS One 5, e8699.

Strettoi, E., Pignatelli, V., 2000. Modifications of retinal neurons in a mouse model ofretinitis pigmentosa. Proc. Natl. Acad. Sci. U. S. A. 97, 11020–11025.

Strettoi, E., et al., 2002. Morphological and functional abnormalities in the inner retinaof the rd/rd mouse. J. Neurosci. 22, 5492–5504.

Strettoi, E., et al., 2003. Remodeling of second-order neurons in the retina of rd/rd mu-tant mice. Vis. Res. 43, 867–877.

Takagi, H., et al., 1994. Insulin-like growth factor-related genes, receptors, and bindingproteins in cultured human retinal pigment epithelial cells. Invest. Ophthalmol.Vis. Sci. 35, 916–923.

Torres-Aleman, I., 2009. Toward a comprehensive neurobiology of IGF-I. Dev. Neuro-biol. 70, 384–396.

Villegas-Perez, M.P., et al., 1998. Ganglion cell loss in RCS rat retina: a result of com-pression of axons by contracting intraretinal vessels linked to the pigment epithe-lium. J. Comp. Neurol. 392, 58–77.

Walenkamp, M.J., Wit, J.M., 2007. Genetic disorders in the GH IGF-I axis in mouse andman. Eur. J. Endocrinol. 157 (Suppl 1), S15–S26.

Walenkamp, M.J., et al., 2005. Homozygous and heterozygous expression of anovel insulin-like growth factor-I mutation. J. Clin. Endocrinol. Metab. 90,2855–2864.

Woods, K.A., et al., 1996. Intrauterine growth retardation and postnatal growth failureassociated with deletion of the insulin-like growth factor I gene. N. Engl. J. Med.335, 1363–1367.

Woods, K.A., et al., 1997. Insulin-like growth factor I gene deletion causing intrauterinegrowth retardation and severe short stature. Acta Paediatr. Suppl. 423, 39–45.