Proc. R. Soc. B

on May 8, 2018http://rspb.royalsocietypublishing.org/Downloaded from

* Autho† PresenStroke, N

Electron1098/rsp

doi:10.1098/rspb.2009.1744

Published online

ReceivedAccepted

Retinal development and functionin a ‘blind’ mole

F. David Carmona1,*, Martin Glosmann2, Jingxing Ou1,†,

Rafael Jimenez3 and J. Martin Collinson1

1School of Medical Sciences, Institute of Medical Sciences, University of Aberdeen, Foresterhill,

Aberdeen AB25 2ZD, UK2Veterinarmedizinische Universitat Wien, Department fur Biomedizinische Wissenschaften,

Institut fur Physiologie und Pathophysiologie, Veterinarplatz 1, 1210 Wien, Austria3Departamento de Genetica e Instituto de Biotecnologıa, Facultad de Ciencias, Universidad de Granada,

18071 Granada, Spain

Animals adapted to dark ecotopes may experience selective pressure for retinal reduction. No previous

studies have explicitly addressed the molecular basis of retinal development in any fossorial mammal.

We studied retinal development and function in the Iberian mole Talpa occidentalis, which was presumed

to be blind because of its permanently closed eyes. Prenatal retina development was relatively normal,

with specification of all cell types and evidence of dorsoventral regionalization. Severe developmental

defects occurred after birth, subsequent to lens abnormalities. ‘Blind’ Iberian moles had rods, cones

and rod nuclear ultrastructure typical of diurnal mammals. DiI staining revealed only contralateral pro-

jections through the optic chiasm. Y-maze experiments demonstrated that moles retain a photoavoidance

response. Over-representation of melanopsin-positive retinal ganglion cells that mediate photoperiodicity

was observed. Hence, molecular pathways of eye development in Iberian moles retain the adaptive func-

tion of rod/cone primary vision and photoperiodicity, with no evidence that moles are likely to completely

lose their eyes on an evolutionary time scale.

Keywords: Iberian mole; fossorial; evolution; retina development; melanopsin; gliosis

1. INTRODUCTIONAbout 300 mammalian species have adopted a subterra-

nean lifestyle (Burda et al. 1990; Peichl 2005). Most

fossorial mammals retain eyes, although the extent of ana-

tomical regression varies, resulting in a multitude of

defective eye phenotypes (Nevo 1979; Burda et al.

1990). Very little is known about the molecular and devel-

opmental mechanisms underlying the regression of the

visual system in mammals, and recent studies have been

performed mainly postnatally in rodents such as the

blind mole-rat and the African mole-rats (Nemec et al.

2007).

Mammals have a duplex retina containing two types of

photoreceptors: cones for colour vision and rods for low-

light vision. A small proportion of retinal ganglion cells

(RGCs) in the retina contain the visual pigment melanop-

sin and are involved in photic responses such as

entrainment of the circadian cycle and control of pupil

size (Nickle & Robinson 2007). Melanopsin is expressed

in approximately 1 to 2.5 per cent of the RGC population

in rodents and man (Hattar et al. 2002).

True moles (Talpidae) are insectivores of the order

Eulipotyphla that show striking adaptations (Gorman &

Stone 1990). Their eyes exhibit anatomical regression,

although the main ocular structural elements are present

r for correspondence (dcarmona@ugr.es).t address: National Institute of Neurological Disorders and

ational Institutes of Health, Bethesda, MD 20892, USA.

ic supplementary material is available at http://dx.doi.org/10.b.2009.1744 or via http://rspb.royalsocietypublishing.org.

28 September 200919 November 2009 1

(Quilliam 1966; Sato 1977; Carmona et al. 2008;

Glosmann et al. 2008). While the domains of the central ner-

vous system required for vision are diminished (Lund &

Lund 1966), the structures involved in controlling the

biological rhythms, the suprachiasmatic nucleus and

the retinohypothalamic projections, are superficially

normally developed (Kudo et al. 1991). The adult Euro-

pean mole (Talpa europaea Linnaeus, 1758) has small

eyes, about 1 mm in diameter, with eyelids that can

open and close. A nucleated and disorganized lens is pre-

sent; the retina displays a typical laminated organization,

with about 2000 RGCs (Quilliam 1966). Talpa europaea

successfully performs light/dark discrimination tasks

under experimental conditions (Lund & Lund 1965,

1966; Johannesson-Gross 1988), suggesting that photo-

pic vision is of relevance in the ecology of this

insectivore. Closely related talpids—the Mediterranean

mole (Talpa caeca Savi, 1882), the Roman mole (Talpa

romana Thomas, 1902), the Balkan mole (Talpa

stankovici Martino & Martino, 1931) and the Iberian

mole (Talpa occidentalis Cabrera, 1907)—have been con-

sidered totally blind because their eyes remain enclosed

under the skin throughout life (Dubost 1968; Nevo

1979; Krystufek 1994; Carmona et al. 2008).

No approaches have explicitly addressed the develop-

mental basis of morphological change in the retina of

any fossorial mammal, in spite of the potential to

inform the link between developmental biology, ecology

and evolution of the eye in a mammalian system. In this

study, we characterize the pattern of differentiation of

the main retinal cell types in T. occidentalis. We further

This journal is q 2009 The Royal Society

s3 s7

s9OC

INBL

ONBL

OPL

GCL

ONL

ONL

IPL

m

bb

bc

rhOLM

ONL

OLM

r

m

hh

b

ba

ar

bmm m

am

ONLINL

INBLINBL

GCL

IPL

INLOPL

ONL

ONBL

ONBL

s11

s14

mouse P2 mouse adult mole adult

s5b

GCL

IPL

INL

OPL

ONL

(i) (v)

(ii)

(vi)

(iii)

(vii)

(iv)

(a)

(b)

Figure 1. Morphological characterization of retinal cells types in T. occidentalis. (a) Haematoxylin and eosin staining of mole andmouse tissue sections. The optic cup (OC) is established at stage s3 in T. occidentalis. Inner neuroblastic layer (INBL) and outerneuroblastic layer (ONBL) are identifiable at s5b–s7. Inner nuclear layer (INL), outer plexiform layer (OPL) and outer nuclear

layer (ONL) differentiate during postnatal stages (s9–s14). No discrete ganglion cell layer (GCL) or inner plexiform layer(IPL) are formed. Mouse P2 and adult retinal sections represent normal mammalian retinal development. (b) Ultrastructuralstudy of the mole retina. (i) Toluidine-blue-stained semithin section of an adult Iberian mole retina. (ii–vii) TEM photomicro-graphs of adult (ii– iv) mole and (v–vii) mouse retinas. (ii) GCL- and IPL-like regions in T. occidentalis. (iii) INL and ONL in

T. occidentalis. (iv) Photoreceptor layer in T. occidentalis. (v) GCL and IPL in the mouse. (vi) INL in the mouse. (vii) Photo-receptor layer in the mouse. Muller cells were distinguishable by their characteristic electron-dense irregular nuclei. Large,round, electron-lucent nuclei were identified as amacrine cells. Horizontal cells showed less electron-dense and more sphericalnuclei than Muller cells. Bipolar cells were identified based on their round nuclei with prominent nucleoli and location (close tothe OPL). OLM, outer limiting membrane; a, amacrine cell; b, bipolar cell; h, horizontal cell; m, Muller cell; r, rod; c, cone.

(a) Scale bar represents 200 mm in s3–s14 images and 60 mm in mole adult and mouse P2 and adult images. (b) Scale bars:(i) 30 mm, (iv) 15 mm and (ii, iii, v–vii) 20 mm.

2 F. D. Carmona et al. The Iberian mole retina

on May 8, 2018http://rspb.royalsocietypublishing.org/Downloaded from

study the behavioural response of T. occidentalis to light

stimuli. We conclude that there remains selective pressure

for light detection, but, in contrast to nocturnal mammals,

there is no evidence that T. occidentalis has undergone any

adaptations for optimizing low-light vision. Moles’ vision

is probably restricted to the detection of sunlight and

maintenance of circadian rhythms.

2. RESULTS(a) Morphological characteristics of the Iberian

mole retina

The terminology of retinal lamination is described in

figure 1. Mole tissue sections showed normal progression

of prenatal retinal development (figure 1a). Early

Proc. R. Soc. B

differentiation of the retina commences properly, with

inner neuroblastic layer (INBL) and outer neuroblastic

layer (ONBL) defined before birth (s5b and s7 in

figure 1a). In newborn moles (s9), the ONBL segregates

into inner nuclear layer (INL) and outer nuclear layer

(ONL). Later in development, an outer plexiform layer

(OPL) becomes identifiable (s11 and s14). However,

cell nuclei from the INBL remain randomly distributed

along the inner side of the retina even in adults. As a con-

sequence, unlike the mouse, neither a discrete ganglion

cell layer (GCL) nor an acellular inner plexiform layer

(IPL) are established at any time.

The morphological ultrastructure of retinal cells was

studied (figure 1b; Carter-Dawson & LaVail 1979; Jeon

et al. 1998; Cernuda-Cernuda et al. 2002). In the Iberian

s7

mole mouse

E16

s9 P2

adult adult

BRN3a + DAPI

Figure 2. RGC differentiation in T. occidentalis. BRN3aimmunoreactivity in mole sections with relatively normal

differentiation of RGCs and some ectopic expression in theONBL and INL (arrows). In mice, BRN3a-positive cells dis-placed to the outer part of the retina (arrows) were observedonly at E16 and P2, not in the adult. Scale bar: 100 mm.

The Iberian mole retina F. D. Carmona et al. 3

on May 8, 2018http://rspb.royalsocietypublishing.org/Downloaded from

mole, ganglion cells and amacrine cells were randomly

distributed across the inner retina as seen by conventional

histology above (figure 1b(i)(ii)), unlike the mouse

(figure 1b(v)). The INL had all the typical neural cell

types in the Iberian mole (figure 1b(iii)), similar to the

mouse (figure 1b(vi)).

Mole rod nuclei did not exhibit the dense heterochro-

matic core typical of nocturnal species such as the mouse

(figure 1b(iv)(vii)), but the conventional architecture of

diurnal species (Solovei et al. 2009). The outer segments

of mole photoreceptors exhibited a whorl-like structure

with signs of anatomical regression. However, moles

have a surprisingly well-conserved retina, with a photo-

receptor ultrastructure that is unexpectedly typical of

diurnal mammals.

(b) Molecular characterization of differentiation

of the Iberian mole retina

(i) Retinal ganglion cells

The characterization of RGCs was assessed by studying the

expression of the RGC marker BRN3a (Nadal-

Nicolas et al. 2009; figure 2). BRN3a was expressed in

cells in the INBL (s7 in figure 2), although some positive

cells were found in the ONBL (arrows). As in the mouse

(figure 2b), in T. occidentalis BRN3a-positive cells were

less numerous after birth than in prenatal stages (s9 in

figure 2). In the adult mole, BRN3a-positive cells were

randomly distributed across the inner retina, with an

occasional ectopic location in the INL (figure 2, arrow in

inset). In the mouse, although some single positive cells

were detected in the outer retina during developmental

stages (figure 2b), all BRN3a-expressing cells were located

Proc. R. Soc. B

exclusively in the GCL in three adult mouse samples

analysed (figure 2).

In adult mice, about 1 per cent of the RGC population

expresses melanopsin (Hattar et al. 2002). We studied

melanopsin immunoreactivity in T. occidentalis

(figure 3a– f ). First evidence of expression was seen in

newborn moles at s9, but not before (n ¼ 3 samples per

stage). In adult moles (n ¼ 5), a very high proportion of

cell bodies in the inner retina was melanopsin-positive

(figure 3a) compared with the mouse (figure 3c). Mela-

nopsin-containing RGCs showed normal morphology

with axons projecting widely throughout the retina

(figure 3a–e), directed towards the optic disc and

contributing to the optic nerve (figure 3f ).

Histology and DiI labelling showed that axons

from RGCs contralaterally traverse the optic chiasm

(figure 3g,h). No ipsilateral projections were detected in

any of the five mole samples analysed. The study

showed that moles have at least two classes of

molecularly specified, though morphologically disorga-

nized, RGCs that make projections to the CNS visual

pathway.

(ii) Amacrine and horizontal cells

Amacrine and horizontal cells were characterized by

double sequential immunostaining of PAX6 and AP2a

proteins (figure 4, top). PAX6 is expressed in retinal pro-

genitors and differentiated RGCs, amacrine cells and

horizontal cells (Walther & Gruss 1991; Hitchcock et al.

1996; Philips et al. 2005). AP2a is a marker of amacrine

cells (Bisgrove & Godbout 1999; Bassett et al. 2007).

Hence, cells immunopositive for both anti-PAX6 and

anti-AP2a were identified as amacrine cells.

In the adult mouse, Pax6 immunofluorescence was

observed in the GCL and in the inner sublayer of the

INL (figure 4). Ap2a-positive amacrine cells occupied

the distal region of the INL (figure 4).

In the Iberian mole, the INL of retinas from infant

moles of s10 (n ¼ 3) showed a pattern of AP2a/PAX6

expression similar to that in the adult mouse, with an

inner band of AP2a-positive/PAX6-positive amacrine

cells and a outer zone of AP2a-negative/PAX6-positive

horizontal cells (top row of figure 4). Nevertheless, this

pattern was disrupted later in development, with the

adult mole showing a disorganized distribution of hori-

zontal and amacrine cells in both the GCL/IPL-like

regions and INL (top row of figure 4, centre).

(iii) Muller glia

Glial fibrillary acidic protein (GFAP) expression in the

apical endplates of retinal Muller cells is a common fea-

ture of vertebrates (Sassoe Pognetto et al. 1993). GFAP

was detected from s10 to adulthood in T. occidentalis

(figure 4). In infant moles at s10, cell bodies in the

inner periphery of the retina were GFAP-positive

(arrowhead in second row of figure 4, left). However,

in adult moles (n ¼ 5), GFAP immunosignal of the

Muller cells was more extensive throughout the whole

depth of the retina, with GFAP-positive processes reach-

ing the ONL (figure 4). In other mammals such as mice,

this extension of GFAP localization throughout the

Muller glia is a feature of reactive gliosis associated

with damage.

(a) (b) (c)

(d)

(g) (h)

(e) ( f )

OD

OC

melanopsin + DAPI

DiI

E

Figure 3. Immunodetection of melanopsin and optic nerve tracing in T. occidentalis. In the adult mole, melanopsin-positive cellswere found throughout the inner retina (a,b), and their density was much higher than in the adult mouse (c). (d,e) Melanopsin-immunoreactive axons were distributed across the plane of the retina, and incorporated into the optic disc (OD; arrowheadsin f ). (g) Haematoxylin staining of a transverse section of a head from s5c mole embryo; the optic nerves are observable(arrows). (h) DiI labelling of the optic nerve in an s5c embryo. OC, optic chiasm; E, eye position. Photomicrographs from

(d) and (e) were taken from flat-mounted retinas. Scale bars: (b,e) 50 mm, (a,c,d, f ) 100 mm, (g) 1.4 mm and (h) 1 mm.

4 F. D. Carmona et al. The Iberian mole retina

on May 8, 2018http://rspb.royalsocietypublishing.org/Downloaded from

(iv) Bipolar cells

Protein kinase C alpha (PKCa) robustly identifies rod

bipolar cells (Greferath et al. 1990; Osborne et al. 1991;

Glosmann et al. 2008). PKCa-immunopositive cells

were present at s10 in T. occidentalis (third row of

figure 4). At this developmental stage, and in subsequent

developmental stages, the PKCa-positive rod bipolar sub-

layer was located in the outer side of the INL, similar to

the mouse (figure 4, right). However, in contrast to the

mouse retina, some bipolar cell somata were displaced

to the ONL (figure 4) and their axon terminals were

distributed along the whole inner retina.

(v) Photoreceptors

The major rod opsin (rhodopsin) and two cone opsins

(a shortwave-sensitive S and a middle-to-longwave-

sensitive M opsin) were present in the photoreceptor

layer of the Iberian mole retina (bottom rows of

figure 4), showing that the normal range of rod and

cone photoreceptors was present. High levels of

rhodopsin expression were observed from stage s9 until

Proc. R. Soc. B

adulthood (figure 4). S opsin-positive cones first appeared

at s10, while M opsin was first detected one stage later, at

s11 (fourth row of figure 4, left image). Displaced rod and

cone photoreceptors were observed in the inner retina of

the Iberian mole (figure 4).

Mean photoreceptor densities, the cone/rod ratio and

the percentage of dual cones were assessed in four

flat-mounted retinas obtained from four different

adult individuals (electronic supplementary material,

table S1). Talpa occidentalis has a retina with an area of

around 0.6 mm2, which is rod-dominated but contains

cone photoreceptors expressing either S or M cone

opsin or both S and M cone opsins in a single

photoreceptor (dual cones).

Immunostaining on flat-mounted retinas revealed a

dorsoventral (DV) gradient of S cones (figure 5a,c). In

contrast, M cones had a homogeneous distribution

across the retina; S cones were more abundant in the ven-

tral pole (figure 5a,c). Cone densities were significantly

higher in the ventral retina (figure 5b), an unusual feature

that is discussed below.

mole s10

PAX

6 +

AP2

αPK

Cα

rod

opsi

n

mole s11*

M +

S o

psin

GFA

P

mole adult mouse adult

Figure 4. Molecular characterization of the retinal cell types in T. occidentalis. (PAX6/AP2a) The AP2a/PAX6 expression profileof the INL in stage s10 moles (top left) does not differ considerably to that in the adult mouse (right). In adult moles (centre),

this distribution pattern becomes disorganized (see inset in mole adult). (GFAP) Moderate levels of GFAP expression weredetected in the inner border of the retina in s10 moles (arrowheads in left image), similar to that in adult mouse (right). Inadult moles, the GFAP-immunosignal extended the whole depth of the retina (arrows centre) and GFAP-positive cellbodies were distributed across the inner retina (white arrowheads in centre). (PKCa) PKCa-positive bipolar cells were firstdetected at s10 located in the outermost side of the OPL, their axons and endfeet distributed along the inner retina. Adult

mouse retinas exhibited a similar spatial profile (right), although the endfeet were observed exclusively in the GCL. (Rodopsin) High expression levels during all postnatal developmental and adult stages were detected in T. occidentalis (left andcentre), similar to those in the mouse (right). (M/S opsin) Shortwave-sensitive (S) and a middle-to-longwave-sensitive (M)cone opsins, in the photoreceptor layer of the Iberian mole (left and centre). s11* stage is shown in this row because M

opsin is first expressed at this stage, whereas S opsin is expressed at s10. Cone opsin expression was maintained in adulthood(centre), with a pattern very similar to the mouse (right). Displaced cells immunoreactive for rod opsin, and S/M coneopsins were detected across the depth of the retina in T. occidentalis (arrows in left and centre). Yellow arrowhead highlightsautofluorescent erythrocytes. Scale bar: 50 mm.

The Iberian mole retina F. D. Carmona et al. 5

Proc. R. Soc. B

on May 8, 2018http://rspb.royalsocietypublishing.org/Downloaded from

15.2

S opsin

S opsin

M opsin

M opsin

14.2

14.0

14.1

16.1

8.1

10.1

14.3

15.9

16.4

12.2

21.421.9

21.3

mol

e s5

bm

ouse

E12

RALDH1 RALDH3

23.923.8

20.9

17.017.416.9

D

N

D

D

D

D

V

V

V

V

iiDIC

i

(a) (c)

(d)

(e)

(b)

M opsin + S opsin

i ii

Figure 5. Cone photoreceptor topography and dorso/ventral patterning of retina development in T. occidentalis. (a) Immuno-labelling for M opsin (green) and S opsin (red) in flat-mounted retina from an adult Iberian mole. (b) Cone density distributionmap of the retina in (a). (c) High magnifications of the indented areas (i and ii) in (a) showing a higher proportion of S cones inthe ventral retina and a homogeneous distribution of M cones. (d) Dual cones were frequent across the mole retina, althoughthe relative expression of S and M opsins within the same cone was not uniform (arrowheads). (e) Dorsal (D) expression of

Raldh1/RALDH1 and ventral (V) expression of Raldh3/RALDH3 in E12 mice and s5b moles. D, dorsal; N, nasal. Scalebars, (a,b) 300 mm, (c) 60 mm, (d) 30 mm and (e) 250 mm.

6 F. D. Carmona et al. The Iberian mole retina

on May 8, 2018http://rspb.royalsocietypublishing.org/Downloaded from

(c) Compartmentalized RALDH expression during

development of Iberian mole retina

Two members of the retinaldehyde dehydrogenase

(RALDH) family, RALDH1 and RALDH3, were used

as markers of DV regionalization during retinal develop-

ment (figure 5e). These enzymes in retinoic acid

metabolism are asymmetrically expressed in the develop-

ing mammalian retina. Raldh1 is expressed dorsally,

whereas Raldh3 shows exclusive restriction to the ventral

retina (Schulte & Bumsted-O’Brien 2008). In the retina

of E12 mice, as expected, we confirmed dorsal expression

of Raldh1 and ventral expression of Raldh3. The same DV

distribution of RALDH1 and RALDH3 was identified in

T. occidentalis from stages s4a to s5c (n ¼ 8, two embryos

per stage; figure 5e), suggesting that topographic cellular

patterning takes place during retinal development in this

mole species.

(d) Wavelength-dependent light-avoidance

capabilities of T. occidentalis moles

Our morphological and molecular data suggest that,

although it is dysgenic to some extent, the retina of

Proc. R. Soc. B

T. occidentalis is potentially functional. To test this hypoth-

esis, light-avoidance experiments were performed with

four adult moles, using flashing white light in a Y-maze

setting in which the animals were offered to approach or

avoid a light source during an escape response. The

results obtained in this study are summarized in table 1.

The choice to approach the dark tunnel was significant

in the four animals used. This indicates that T. occidentalis

is able to respond to flashing light stimuli, under the

experimental conditions employed. In all cases, the ani-

mals responded by avoiding the light source, suggesting

that in stressful situations they move towards darkness.

3. DISCUSSION(a) How strong is the selective pressure

on the mole retina?

In low-light environments, animals may be under selec-

tive pressure for increased photoreceptive sensitivity.

However, an alternative evolutionary strategy in con-

ditions where light levels are too low for any ocular

function to be possible (e.g. in caves) is for eye regression

Table 1. Response to 30 lux flashing white light. D, number

of choices to the dark Y-maze arm; L, number of choices tothe illuminated Y-maze arm.

individuals D L x2 values/p

1 24 6 10.08/0.001***2 17 3 9.8/0.002**3 15 5 5/0.025*4 22 8 6.53/0.011*

*p , 0.05.**p , 0.01.***p , 0.001.

The Iberian mole retina F. D. Carmona et al. 7

on May 8, 2018http://rspb.royalsocietypublishing.org/Downloaded from

with expansion of the other senses. Moles, as fossorial

animals, fall into a category of sporadic light exposure

for which either evolutionary strategy may be advan-

tageous. There are potentially competing pressures to

minimize the energetic cost of eyes in an environment

where photic sensory information may be valuable.

The data presented here show that T. occidentalis has a

surprisingly well-conserved retina. This suggests either

that the Iberian mole is on a pathway to further eye

regression which is not yet complete (evolutionary lag)

or that there is some adaptive advantage to maintain the

retina. Our data demonstrate that the eyes are functional,

but that some retained aspects of eye patterning could

potentially be lost in future.

(b) Differentiation of the Iberian mole retina

The pattern of differentiation of the retinal cell types is

highly conserved among vertebrates (Lamb et al. 2007).

Our results show that T. occidentalis has retained an

ordered timing of retinal cell differentiation very similar

to that of the mouse.

Photoreceptor distribution is patterned across the tan-

gential plane of the retina (Ahnelt & Kolb 2000; Peichl

2005). Surprisingly, both RALDH1 and RALDH3

showed the same robust compartmentalized distribution

in T. occidentalis as in the mouse (McCaffery et al. 1998;

Li et al. 2000), thus suggesting that the mechanisms

that establish DV position-dependent cues have been con-

served in the Iberian mole. This may be necessary to set

up the DV differential in cone photoreceptor frequency

discussed below, or it may be atavistic.

(c) Disrupted lens development could induce the

postnatal retinal defects observed in T. occidentalis

Our results suggest that prenatal retinal development pro-

gresses properly in T. occidentalis. However, postnatally,

amacrine and horizontal cell organization becomes

disrupted in the INL, with overt anomalies observed in

the adult.

The vertebrate lens plays a key role in the development

of the retina (Coulombre & Coulombre 1964; Breitman

et al. 1987; Landel et al. 1988; Kaur et al. 1989;

Ashery-Padan et al. 2000; Yamamoto & Jeffery 2000;

Kurita et al. 2003; Vihtelic et al. 2005), and in this respect

it is interesting that the retinal phenotypes observed in

moles occur subsequent to the development of the

major lens abnormalities we described previously, result-

ing from misregulation of the forkhead transcription

factor FOXE3 in the lens (Carmona et al. 2008). We

Proc. R. Soc. B

speculate that impaired regulation of lens FOXE3 may

secondarily initiate retinal dysgenesis.

(d) Extensive GFAP expression in retinal

Muller cells

Muller cells represent the principal glial cell type of the

retina. They extend apically to basally, and GFAP is nor-

mally expressed only in the very inner segments of Muller

glia in healthy retinas (Sassoe Pognetto et al. 1993). How-

ever, the whole glial surface becomes GFAP-positive

during reactive gliosis following disease or trauma

(Guerin et al. 1990; Lewis et al. 1995; Bringmann et al.

2006). In neonatal (s10, 3–7 dpp) Iberian moles,

GFAP showed a spatial expression pattern very similar

to that observed in adult mice. However, from stage s11

(7–12 dpp), GFAP immunoreactivity expanded across

the depth of the retina. Muller glia are, in mammals,

the major retinal stem cell class (Bernardos et al. 2007),

and it may be that in moles, they are responding to

some aspect of the defective retinal phenotype by

attempting to activate regenerative properties that would

‘repair’ the retina.

(e) Diurnal nuclear pattern of photoreceptor

nuclei in T. occidentalis

Solovei et al. (2009) showed that rod nuclei of nocturnal

species such as the mouse are small and present an

‘inverted’ pattern of chromatin organization in which a

heterochromatic core occupies most of the nuclear

volume. The inverted nuclei produce a more efficient

transmission of light. We found that T. occidentalis exhibits

the conventional rod nuclear pattern typical of diurnal

species. Analysis of the figures in Lluch et al. (2009)

suggests that shrews (Soricidae), which are closely related

to moles, have an intermediate rod nuclear pattern that

correlates with their polyphasic lifestyle (Genoud 1984).

The rod architecture of mole species could be a retained

trait of a diurnal common ancestor of moles and shrews,

and there is no evidence that moles have been under

selective pressure to maximize the light-gathering capacity

of their eyes—they do not ‘try’ to see in dim light, and are

not evolving to do so.

(f) Possible implications of the retina in the

ecology of moles

Cone proportions range from low (0.5–3%) in nocturnal

species and medium (2–10%) in crepuscular and

arrhythmic species to high (8–95%) in diurnal ones.

Rod densities generally are higher in nocturnal species,

with most mammals showing values ranging from 200

000 to 400 000 rods mm22 (Ahnelt & Kolb 2000;

Peichl 2005). The percentage of cone photoreceptors

varies between species and, in insectivores, seems to

relate to their activity pattern (Peichl et al. 2000). For

instance, the mostly nocturnal Crocidura spp. show rela-

tively low values (4–6%; Genoud & Vogel 1981; Vogel

et al. 1981) and the polyphasic Sorex araneus higher

ones (13%; Genoud 1984). Surprisingly, a previous

study performed in T. europaea (Glosmann et al. 2008)

and that reported here for T. occidentalis have shown

that talpid moles exhibit high cone proportions similar

(9–15%) to those of the closely related soricids. This

provides further evidence that the eyes of moles are not

8 F. D. Carmona et al. The Iberian mole retina

on May 8, 2018http://rspb.royalsocietypublishing.org/Downloaded from

adapted to low-light vision. Taking into account that the

strictly subterranean African mole-rats of the rodent

family Bathyergidae also have high cone proportions

(around 10%; Peichl et al. 2004), that the photoreceptor

mosaic is not adapted to low-level vision in subterranean

mammals may be a common phenomenon.

A common feature between T. occidentalis and the

other insectivores studied is the DV gradient of S cones,

with higher densities in the ventral retina. This striking

patterning, which seems to be conserved among Eulipo-

typhla (Peichl 2005), has been observed in several

mouse species (Szel et al. 1996). S cone asymmetry

increases the sensitivity of the ventral retina to short wave-

lengths in the mouse (Calderone & Jacobs 1995). In the

case of fossorial moles, there is a hypothetical advantage

in having this cone distribution. The ventral half-retina

receives light from the upper visual field, and an increased

S cone density may allow the mole to have a better per-

ception of breaches in their tunnel network. A possible

role of the visual systems in antipredatory behaviour has

been previously suggested for subterranean rodents

(Nemec et al. 2007, 2008).

The random distribution of melanopsin-containing

RGCs across the whole depth of the inner retina made

an accurate estimation of their relative number in

T. occidentalis not feasible. However, immunostaining

clearly showed that the percentage of RGCs expressing

melanopsin was far higher in the Iberian mole than in

adult mice, with melanopsin-positive axons successfully

incorporated to the optic nerve. Unusually high pro-

portions of the melanopsin-positive RGC subset (about

20% of the total RGC population) have been reported

also in the rodent Spalax (Hannibal et al. 2002), thus

suggesting that this feature may have evolved convergently

in unrelated fossorial mammals.

The maintenance of a high proportion of neonatal

melanopsin-containing RGCs throughout the whole life

of T. occidentalis could have represented an important

adaptive acquisition during talpid evolution. Moles

show a clear periodicity of activity and rest (Macdonald

et al. 1997; Borroni et al. 1999) and are strict seasonal

breeders, with a cycle that varies in length according to

latitude (Jimenez et al. 1990). Thus, information on

photoperiod changes may be an essential requirement

for these fossorial insectivores. An increase in the

melanopsin-containing cell density could represent a

very valuable adaptive trait.

4. MATERIAL AND METHODS(a) Material analysed

Ninety-seven specimens of T. occidentalis were obtained

from animals captured in the wild as described previously

(Barrionuevo et al. 2004a) under permission granted by the

Andalusian Environmental Council. Developmental stages

were determined according to Barrionuevo et al. (2004b)

and Carmona et al. (2009).

Positive controls were carried out with mouse samples of

the CBA/Ca strain.

(b) Immunohistochemistry and haematoxylin and

eosin staining

Tissues were embedded in wax and sectioned following stan-

dard procedures. Immunohistochemistry and haematoxylin

Proc. R. Soc. B

staining was described previously (Carmona et al. 2008).

Electronic supplementary material table S2 summarizes the

antibodies used in this study.

(c) Retinal flat-mount immunofluorescence

Retinas were dissected from fixed eyes into TBS, 0.05 per cent

Tween-20, 0.1 per cent Triton X-100 for 30 min at RT. After

washing with TBS, 0.05 per cent Tween-20 (TBS-T), retinas

were exposed to primary antibodies in PBS, 4 per cent BSA

overnight at 48C, washed again with TBS-T and incubated

in secondary antibodies diluted in PBS, 4 per cent BSA,

for 3 h at RT. Retinas were flat-mounted in medium

containing DAPI.

(d) Photoreceptor densities and topography

measurements

Densities and topographic distributions of rod and cone

photoreceptors were obtained by analysing four adult flat-

mounted retinas (from four different moles) after double

immunostaining for S and M opsin to label all cones.

Cones were counted in sampling windows of 97 � 97 mm

for analysis of density and in higher sampling windows of

160 � 120 mm for topographic assay. Rods were identified

as unlabelled photoreceptors using DIC optics. Shrinkage

was evaluated by planimetrically determining the retinal

area before and after mounting.

(e) DiI labelling of retinal ganglion cell axons

Five-stage s5b–c mole heads were fixed for DiI staining of

the optic nerve. After removing the cornea, the lens was

removed from the embryonic eye and the inside of the

retina was carefully dried with filter paper. A crystal of DiI

(D282; Molecular Probes, Eugene, OR, USA) was adhered

to the optic disc and the lens was replaced. Heads were incu-

bated in PBS, 0.02 per cent NaN3 at 378C for 4 days.

Subsequently, the brain was dissected and placed in PBS

for whole-mount photomicrography using a red single

bandpass fluorescence mirror unit.

(f) Ultrastructural analysis

Samples were processed for transmission electron

microscopy (TEM) by the Electron Microscopy Facility of

the Institute of Medical Sciences, University of Aberdeen

using a Philips CM10 microscope.

(g) Mole sight study

Behavioural tests were conducted in a Y-maze, made of

50 mm diameter black PVC tubing. Flashing white light

bulbs (12 V, 22 mm diameter, 4 flashes s21) were placed at

the ends of the two symmetric arms of the maze. The illumi-

nance that these lamps provided at the bifurcation of the

Y-maze was 30 lux. For each trial, a dark-adapted mole

(kept for 2 days prior to the experiment to get them used

to captivity) was released into the maze through the access

tube, the light switched on in one branch and the subject

allowed to choose the dark or the illuminated arm. The

light and any reflection were not visible from the entrance

of the Y-maze. Four adult moles (two males, two females)

were used. Sets of 10 trials were done with each animal.

Left and right Y-maze arms were randomly illuminated. To

minimize stress, animals were rested between consecutive

sets of experiments. The x2 distribution was determined to

test for statistical significance.

The authors thank D. G. Lupianez, J. E. Martın andF. M. Real (University of Granada, Spain) for assistance,and the Andalussian Consejerıa de Medio Ambiente for

The Iberian mole retina F. D. Carmona et al. 9

on May 8, 2018http://rspb.royalsocietypublishing.org/Downloaded from

permits. This work was supported by the Alfonso MartınEscudero Foundation and Junta de Andalucıa throughGroup PAI CVI-109 (BIO-109). J.M.C.’s laboratory isfunded by Biotechnology and Biological Sciences ResearchCouncil (BB/E015840/1). Animal handling followed theguidelines and approval of the University of Granada’sEthical Committee for Animal Experimentation.

REFERENCESAhnelt, P. K. & Kolb, H. 2000 The mammalian photo-

receptor mosaic-adaptive design. Progr. Ret. Eye Res. 19,711–777. (doi:10.1016/S1350-9462(00)00012-4)

Ashery-Padan, R., Marquart, T., Zhou, X. & Gruss, P. 2000Pax6 activity in the lens primordium is required for lensformation and for correct placement of a single retina inthe eye. Genes Dev. 14, 2701–2711. (doi:10.1101/gad.

184000)Barrionuevo, F. J., Zurita, F., Burgos, M. & Jimenez, R. 2004a

Testis-like development of gonads in female moles. Newinsights on mammalian gonad organogenesis. Dev. Biol.268, 39–52. (doi:10.1016/j.ydbio.2003.11.025)

Barrionuevo, F. J., Zurita, F., Burgos, M. & Jimenez, R.2004b Developmental stages and growth rate of themole Talpa occidentalis (Insectivora, Mammalia).J. Mammal. 85, 120–125. (doi:10.1644/BPR-010)

Bassett, E. A., Pontoriero, G. F., Feng, W., Marquardt, T.,

Fini, M. E., Williams, T. & West-Mays, J. A. 2007Conditional deletion of activating protein 2alpha(AP-2alpha) in the developing retina demonstrates non-cell-autonomous roles for AP-2alpha in optic cupdevelopment. Mol. Cell Biol. 27, 7497–7510. (doi:10.

1128/MCB.00687-07)Bernardos, R. L., Barthel, L. K., Meyers, J. R. & Raymond,

P. A. 2007 Late-stage neuronal progenitors in the retinaare radial Muller glia that function as retinal stem cells.

J. Neurosci. 27, 7028–7040. (doi:10.1523/JNEUROSCI.1624-07.2007)

Bisgrove, D. A. & Godbout, R. 1999 Differential expressionof AP-2alpha and AP-2beta in the developing chickretina: repression of R-FABP promoter activity by AP-2.

Dev. Dyn. 214, 195–206. (doi:10.1002/(SICI)1097-0177(199903)214:3,195::AID-AJA3.3.0.CO;2-8)

Borroni, A., Loi, A. & Bertolucci, C. 1999 Daily and circa-dian rhythms of rest and activity of Talpa romanaThomas (Mammalia, Insectivora: Talpidae). Rend. Fis.Acc. Lincei 10, 89–99.

Breitman, M. L., Clapoff, S., Rossant, J., Tsui, L. C., Glode,M., Maxwell, I. H. & Bernstein, A. 1987 Geneticablation: targeted expression of a toxic gene causes micro-phthalmia in transgenic mice. Science 238, 1563–1565.

(doi:10.1126/science.3685993)Bringmann, A., Pannicke, T., Grosche, J., Francke, M.,

Wiedemann, P., Skatchkov, S. N., Osborne, N. N. &Reichenbach, A. 2006 Muller cells in the healthy anddiseased retina. Prog. Retin. Eye Res. 25, 397–424.

(doi:10.1016/j.preteyeres.2006.05.003)Burda, H., Bruns, V. & Muller, M. 1990 Sensory adaptations

in subterranean mammals. Prog. Clin. Biol. Res. 335,269–293.

Calderone, J. B. & Jacobs, G. H. 1995 Regional variations inthe relative sensitivity to UV light in the mouse retina.Vis. Neurosci. 12, 463–468. (doi:10.1017/S0952523800008361)

Carmona, F. D., Jimenez, R. & Collinson, J. M. 2008 The

molecular basis of defective lens development in theIberian mole. BMC Biol. 6, 44. (doi:10.1186/1741-7007-6-44)

Carmona, F. D., Lupianez, D., Martın, E., Burgos, M.,Jimenez, R. & Zurita, F. 2009 The spatio-temporal

Proc. R. Soc. B

pattern of testis organogenesis in mammals. Insightsfrom the mole. Int. J. Dev. Biol. 53, 1035–1044.(doi:10.1387/ijdb.072470fc)

Carter-Dawson, L. D. & LaVail, M. M. 1979 Rods and conesin the mouse retina. I. Structural analysis using light andelectron microscopy. J. Comp. Neurol. 188, 245–262.(doi:10.1002/cne.901880204)

Cernuda-Cernuda, R., DeGrip, W. J., Cooper, H. M.,

Nevo, E. & Garcıa-Fernandez, J. M. 2002 The retinaof Spalax ehrenbergi: novel histologic features supportiveof a modified photosensory role. Invest. Ophthalmol.Vis. Sci. 43, 2374–2383.

Coulombre, A. J. & Coulombre, J. L. 1964 Lensdevelopment. I. Role of the lens in eye growth. J. Exp.Zool. 156, 39–48. (doi:10.1002/jez.1401560104)

Dubost, G. 1968 Les mammiferes souterrains. Rev. Ecol.Biol. Sol. 5, 99–197.

Genoud, M. 1984 Activity of Sorex coronatus (Insectivora,Soricidae) in the field. Zeitschrift fur Saugetierkunde 49,74–78.

Genoud, M. & Vogel, P. 1981 The activity of Crocidurarussula (Insectivora, Soricidae) in the field and in

captivity. Zeitschrift fur Saugetierkunde 46, 222–232.Glosmann, M., Steiner, M., Peichl, L. & Ahnelt, P. K. 2008

Cone photoreceptors and potential UV vision in a subter-ranean insectivore, the European mole. J. Vis. 8, 23.1–12.(doi:10.1167/8.4.23)

Gorman, M. L. & Stone, R. D. 1990 The natural history ofmoles. Ithaca, NY: Cornell University Press.

Greferath, U., Grunert, U. & Wassle, H. 1990 Rod bipolarcells in the mammalian retina show protein kinase C-like

immunoreactivity. J. Comp. Neurol. 301, 433–442.(doi:10.1002/cne.903010308)

Guerin, C. J., Anderson, D. H. & Fisher, S. K. 1990Changes in intermediate filament immunolabeling occurin response to retinal detachment and reattachment in

primates. Invest. Ophthalmol. Vis. Sci. 31, 1474–1482.Hannibal, J., Hindersson, P., Nevo, E. & Fahrenkrug, J. 2002

The circadian photopigment melanopsin is expressed inthe blind subterranean mole rat Spalax. Neuroreport 13,1411–1414. (doi:10.1097/00001756-200208070-00013)

Hattar, S., Liao, H. W., Takao, M., Berson, D. M. & Yau, K. W.2002 Melanopsin-containing retinal ganglion cells: archi-tecture, projections, and intrinsic photosensitivity. Science295, 1065–1070. (doi:10.1126/science.1069609)

Hitchcock, P. F., Macdonald, R. E., VanDeRyt, J. T. &

Wilson, S. W. 1996 Antibodies against Pax6 immunostainamacrine and ganglion cells and neuronal progenitors, butnot rod precursors, in the normal and regenerating retinaof the goldfish. J. Neurobiol. 29, 399–413. (doi:10.1002/

(SICI)1097-4695(199603)29:3,399::AID-NEU10.3.0.CO;2-4)

Jeon, C. J., Strettoi, E. & Masland, R. H. 1998 The majorcell populations of the mouse retina. J. Neurosci. 18,8936–8946.

Jimenez, R., Burgos, M., Sanchez, A. & Dıaz de la Guardia,R. 1990 The reproductive cycle of Talpa occidentalis inthe southeastern Iberian Peninsula. Acta Theriol. 35,165–169.

Johannesson-Gross, K. 1988 Brightness discrimination of

the mole (Talpa europaea L.) in learning experimentsapplying a modified tube-maze method. Zeitschrift furSaugetierkunde 53, 193–201.

Kaur, S., Key, B., Stock, J., McNeisch, J. D., Akeson, R. &Potter, S. S. 1989 Targeted ablation of alpha-crystallin-

synthesizing cells produces lens-deficient eyes in trans-genic mice. Development 105, 613–619.

Krystufek, B. 1994 The taxonomy of blind moles (Talpacaeca and T. stankovici, Insectivora, Mammalia) fromsouth-east Europe. Bonn. Zool. Beitr. Bd. 45, 1–16.

10 F. D. Carmona et al. The Iberian mole retina

on May 8, 2018http://rspb.royalsocietypublishing.org/Downloaded from

Kudo, M., Yamamoto, M. & Nakamura, Y. 1991 Suprachias-matic nucleus and retinohypothalamic projections inmoles. Brain Behav. Evol. 38, 332–338. (doi:10.1159/

000114399)Kurita, R., Sagara, H., Aoki, Y., Link, B. A., Arai, K. &

Watanabe, S. 2003 Suppression of lens growth byalphaA-crystallin promoter-driven expression of diptheriatoxin results in disruption of retinal cell organization in

zebrafish. Dev. Biol. 255, 113–127. (doi:10.1016/S0012-1606(02)00079-9)

Lamb, T. D., Collin, S. P. & Pugh Jr, E. N. 2007 Evolution ofthe vertebrate eye: opsins, photoreceptors, retina and eye

cup. Nat. Rev. Neurosci. 8, 960–976. (doi:10.1038/nrn2283)

Landel, C. P., Zhao, J., Bok, D. & Evans, G. A. 1988 Lens-specific expression of recombinant ricin induces develop-mental defects in the eyes of transgenic mice. Genes Dev.2, 1168–1178. (doi:10.1101/gad.2.9.1168)

Lewis, G. P., Matsumoto, B. & Fisher, S. K. 1995 Changesin the organization and expression of cytoskeletal proteinsduring retinal degeneration induced by retinal detach-ment. Invest. Ophthalmol. Vis. Sci. 36, 2404–2416.

Li, H., Wagner, E., McCaffery, P., Smith, D., Andreadis, A. &Drager, U. C. 2000 A retinoic acid synthesizing enzyme inventral retina and telencephalon of the embryonic mouse.Mech. Dev. 95, 283–289. (doi:10.1016/S0925-4773(00)00352-X)

Lluch, S., Lopez-Fuster, M. J. & Ventura, J. 2009 Cornea,retina, and lens morphology in five Soricidae species(Soricomorpha: Mammalia). Anat. Sci. Int. 84, 312–322. (doi:10.1007/s12565-009-0042-1)

Lund, R. D. & Lund, J. S. 1965 The visual system of themole Talpa europaea. Exp. Neurol. 13, 302–316.

Lund, R. D. & Lund, J. S. 1966 The central visual pathwaysand their functional significance in the mole (Talpaeuropaea). J. Zool. 149, 95–101. (doi:10.1111/j.1469-

7998.1966.tb02988.x)Macdonald, D. W., Atkinson, R. P. D. & Blanchard, G.

1997 Spatial and temporal patterns in the activity ofEuropean moles. Oecologia 109, 88–97. (doi:10.1007/s004420050062)

McCaffery, P., Lee, M. O., Wagner, M. A., Sladek, N. E. &Drager, U. C. 1998 Asymmetrical retinoic acid synthesisin the dorsoventral axis of the retina. Development 115,371–382.

Nadal-Nicolas, F. M., Jimenez-Lopez, M., Sobrado-Calvo,

P., Nieto-Lopez, L., Canovas-Martınez, I., Salinas-Navarro, M., Vidal-Sanz, M. & Agudo, M. 2009 Brn3aas a marker of retinal ganglion cells: qualitative and quan-titative time course studies in naive and optic nerve

injured retinas. Invest. Ophthalmol. Vis. Sci. 50,3860–3868. (doi:10.1167/iovs.08-3267)

Nemec, P., Cvekova, P., Burda, H., Benada, O. & Peichl, L.2007 Visual systems and the role of vision in subterraneanrodents: diversity of retinal properties and visual system

designs. In Subterranean rodents: news from underground(eds S. Begall, H. Burda & C. E. Schleich), pp. 129–160.Berlin, Germany: Springer.

Nemec, P., Cvekova, P., Benada, O., Wielkopolska, E.,Olkowitcz, S., Turlejski, K., Burda, H., Bennett, N. C. &

Peichl, L. 2008 The visual system in subterraneanAfrican mole-rats (Rodentia, Bathyergidae): retina, sub-cortical visual nuclei and primary visual cortex. Brain Res.Bull. 75, 356–364. (doi:10.1016/j.brainresbull.2007.10.055)

Nevo, E. 1979 Adaptive convergence and divergence of sub-

terranean mammals. Annu. Rev. Ecol. Syst. 10, 269–308.(doi:10.1146/annurev.es.10.110179.001413)

Proc. R. Soc. B

Nickle, B. & Robinson, P. R. 2007 The opsins of the ver-tebrate retina: insights from structural, biochemical, andevolutionary studies. Cell Mol. Life Sci. 64, 2917–2932.

(doi:10.1007/s00018-007-7253-1)Osborne, N. N., Broyden, N. J., Barnett, N. L. & Morris, N. J.

1991 Protein kinase C (alpha and beta) immunoreactivityin rabbit and rat retina: effect of phorbol esters and trans-mitter agonists on immunoreactivity and the translocation

of the enzyme from cytosolic to membrane compartments.J. Neurochem. 57, 594–604. (doi:10.1111/j.1471-4159.1991.tb03790.x)

Peichl, L. 2005 Diversity of mammalian photoreceptor

properties: adaptations to habitat and lifestyle? Anat.Rec. A. Discov. Mol. Cell Evol. Biol. 287, 1001–1012.(doi:10.1002/ar.a.20262)

Peichl, L., Kunzle, H. & Vogel, P. 2000 Photoreceptor typesand distributions in the retinae of insectivores. Vis. Neuro-sci. 17, 937–948. (doi:10.1017/S0952523800176138)

Peichl, L., Nemec, P. & Burda, H. 2004 Unusual cone androd properties in subterranean African molerats (Roden-tia, Bathyergidae). Eur. J. Neurosci. 19, 1545–1558.(doi:10.1111/j.1460-9568.2004.03263.x)

Philips, G. T., Stair, C. N., Young Lee, H., Wroblewski, E.,Berberoglu, M. A., Brown, N. L. & Mastick, G. S.2005 Precocious retinal neurons: Pax6 controls timing ofdifferentiation and determination of cell type. Dev. Biol.279, 308–321. (doi:10.1016/j.ydbio.2004.12.018)

Quilliam, A. T. 1966 The problem of vision in the ecology ofTalpa europaea. Exp. Eye Res. 5, 63–78. (doi:10.1016/S0014-4835(66)80022-2)

Sassoe Pognetto, M., Panzanelli, P., Artero, C., Fasolo, A. &

Cantino, D. 1993 Comparative study of glial fibril-lary acidic protein (GFAP)-like immunoreactivity in theretina of some representative vertebrates. Eur. J.Histochem. 36, 467–477.

Sato, Y. 1977 Comparative morphology of the visual system

of some Japanese species of Soricoidea (Superfamily) inrelation to leif habits. J. Hirnforsch. 18, 531–546.

Schulte, D. & Bumsted-O’Brien, K. M. 2008 Molecularmechanisms of vertebrate retina development: impli-cations for ganglion cell and photoreceptor patterning.

Brain Res. 1192, 151–164. (doi:10.1016/j.brainres.2007.04.079)

Solovei, I., Kreysing, M., Lanctot, C., Kosem, S., Peichl, L.,Cremer, T., Guck, J. & Joffe, B. 2009 Nuclear architec-ture of rod photoreceptor cells adapts to vision in

mammalian evolution. Cell 137, 356–368. (doi:10.1016/j.cell.2009.01.052)

Szel, A., Rohlich, P., Caffe, A. R. & van Veen, T. 1996 Dis-tribution of cone photoreceptors in the mammalian

retina. Microsc. Res. Tech. 35, 445–462. (doi:10.1002/(SICI)1097-0029(19961215)35:6,445::AID-JEMT4.3.0.CO;2-H)

Vihtelic, T. S., Yamamoto, Y., Springer, S. S., Jeffery, W. R. &Hyde, D. R. 2005 Lens opacity and photoreceptor

degeneration in the zebrafish lens opaque mutant. Dev.Dyn. 233, 52–65. (doi:10.1002/dvdy.20294)

Vogel, P., Genoud, M. & Frey, H. 1981 Rythme journalierd’activite chez quelques Crocidurinae africains et eur-opeens (Soricidae, Insectivora). Revue d’Ecologie (Terre etVie) 35, 97–108.

Walther, C. & Gruss, P. 1991 Pax-6, a murine paired boxgene, is expressed in the developing CNS. Development113, 1435–1449.

Yamamoto, Y. & Jeffery, W. R. 2000 Central role for the lens

in cave fish eye degeneration. Science 289, 631–633.(doi:10.1126/science.289.5479.631)