Abstract
purpose. Immunocytochemistry showed strong aquaporin (AQP)-4 water channel
expression in Müller cells in mouse retina and fibrous astrocytes
in optic nerve. This study was designed to test the hypothesis that
AQP4 is required for vision by comparing electroretinograms and retinal
morphology in wild-type mice and transgenic knockout mice with no AQP4.
methods. Electroretinograms were recorded over a 105-fold range of
flash intensities in dark-adapted mice and analyzed for a- and b-wave
amplitude and latency, a-wave normalized slope, and oscillatory
potential amplitude and latency. AQP4 protein was localized in mouse
retina by immunocytochemistry, and retinal morphology was studied by
light and electron microscopy.
results. Significantly reduced electroretinogram b-wave potentials were recorded
in 10-month-old null mice with smaller changes in 1-month-old mice.
Immunocytochemistry showed strong AQP4 protein expression in retina of
wild-type mice. Morphologic analysis of retina by light and electron
microscopy showed no differences in retinal ultrastructure.
conclusions. Retinal function is mildly impaired in AQP4-null mice, suggesting a
role for AQP4 in Müller cell fluid balance. These results support
the paradigm that AQP4 expression in supportive cells in the nervous
system facilitates neural signal transduction in nearby electrically
excitable cells.
The aquaporins (AQPs) are a family of water-transporting
channels that are expressed widely in mammalian fluid-transporting
epithelia and endothelia. The eye expresses at least four AQPs: AQP1 in
corneal endothelium, ciliary epithelium, and lens epithelium; AQP3 in
conjunctiva; AQP4 in ciliary epithelium and retina; and AQP5 in corneal
epithelium and lacrimal gland.
1 2 3 4 5 6 Because of this
expression pattern, it has been proposed that AQPs play a role in
intraocular pressure regulation, corneal and lens transparency, and
vision.
Recently our laboratory has generated transgenic mice without each of
the four eye AQPs, individually and in pairs. The mice have been
informative in defining the role of AQPs in extraocular functions
(reviewed in Ref.
7 ). For example, the deletion of
AQP1 produced defects in urinary concentrating
ability,
8 9 10 lung water transport,
11 and
dietary fat processing.
12 Deletion of AQP3 resulted in
nephrogenic diabetes insipidus,
13 and deletion of AQP5
gave reduced saliva secretion,
14 airway submucosal gland
secretion,
15 and alveolar epithelial water
permeability.
16 However, the tissue-specific expression of
an AQP does not always indicate physiological significance. For
example, AQP5 deletion does not affect tear secretion by lacrimal
glands,
17 and AQP4 deletion does not produce demonstrable
abnormalities in skeletal muscle function
18 or gastric
acid secretion,
19 despite its expression in muscle cell
plasmalemma and gastric parietal cells.
Indirect evidence has suggested a role for AQP4 in retinal function.
AQP4 was first localized in the eye in ciliary epithelium and in glial
cells in the inner nuclear layer of the retina.
6 High-resolution morphology showed AQP4 protein expression in
Müller cells in the retina and fibrous astrocytes in the optic
nerve.
3 Studies of brain edema in AQP4-null mice implied
an important role for AQP4 in fluid balance,
20 supporting
the possibility that AQP4 may participate in the maintenance of retinal
water balance during synaptic transmission and retinal edema. During
neurophysiological activity, action potentials and osmotic gradients
are generated by ion fluxes from ion-solute pumps and exchangers. AQP4
has been shown to be the orthogonal array protein (OAP) by the absence
of OAPs in AQP4-null mice,
21 the creation of OAPs
in AQP4-transfected cells,
22 and label-fracture studies in
brain tissue.
23 Based on the colocalization of Kir4.1
potassium channels and AQP4-containing OAPs in specific membrane
domains of retinal Müller cells, it was proposed that AQP4 is
important in retinal signal transduction involving interactions between
Müller and bipolar cells.
24 25 Similar interactions
occur between AQP4-expressing glial cells in the central nervous system
and adjacent neurons,
26 27 as well as in AQP4-expressing
supportive cells (Claudius, Hensen, and inner sulcus cells) in cochlea
and adjacent sensory hair cells,
28 where they are proposed
to play an important role in acoustic signal
transduction.
29
The purpose of this study was to test the hypothesis that AQP4 plays a
role in retinal function. We compared electroretinograms (ERGs) in
wild-type and AQP4-null mice and performed morphologic analysis of
retina and optic nerve. We found significantly reduced ERG b-wave
amplitude and latency in AQP4-null mice without ultrastructural
abnormalities, providing the first direct evidence for a functional
role of an AQP in the eye.
As shown schematically in
Figure 1A , the instrument consisted of a flashlamp with focusing and filtering
optics and amplifiers (Biopac System, Inc., Goleta, CA) and recording
hardware and software. Optics were constructed to control light
intensity and deliver the light to the eye using a fiber-optic bundle
and Lucite (DuPont, Wilmington, DE) coupler that conformed to the
curvature of the globe (coupled with optical gel). As described by
Lyubarsky and Pugh,
32 this configuration provides
uniform full-field retinal illumination. The stimulus consisted of
single white-light pulses of 20-μsec duration produced by a
stroboscope (NovaStrobe; Monarch Instruments, Amherst, NH). The light
was collected, focused, and filtered onto a 7-mm diameter fiber-optic
bundle (Edmund Scientific, Barrington, NJ). The bundle was brought into
a Faraday cage and inserted into the tapered Lucite coupler, which was
held by a 4-axis micromanipulator (Narashige Instruments, Greenvale,
NY). ERGs were recorded using a 0.15-mm diameter platinum wire coil
that was secured to the rim of the rod tip, making electrical contact
with the cornea through a layer of methylcellulose solution (Gonisol;
Iolab Pharmaceuticals, Claremont, CA). Reference and ground subdermal
needle electrodes were placed on the forehead and right front leg,
respectively. Responses were differentially amplified at a gain of
10,000, bandpassed at 0.1–1000 Hz, digitized at a sampling rate of
4000 Hz, and recorded using a workstation (model MP100; Biopac, Inc.)
equipped with differential amplifier (model ERS100B) and stimulator
modules (model STM100A; both from Biopac, Inc.).
Mice were dark adapted overnight and anesthetized under dim red light,
with ketamine (80 mg/kg body weight) and xylazine (16 mg/kg body
weight). Their pupils were dilated and anesthetized with 2%
phenylephrine and 0.5% proparacaine hydrochloride. Mice were placed on
a heating pad (Braintree Scientific, Inc., Braintree, MA), and body
temperature was strictly maintained at 37.0 ± 0.5°C, as
measured by an indwelling rectal thermistor. The mouse, heating pad,
and micromanipulator controlling the Lucite coupler were enclosed in a
copper Faraday cage and a light-tight, sound-insulated black box. ERG
waveforms were generally recorded in triplicate and averaged, in order
of increasing light intensity. A filter wheel containing neutral
density filters was used to set illumination intensity. The time
between flashes was 30 seconds at low flash intensities and 60 seconds
at the highest flash intensity. Control studies were performed to
confirm that ERG waveforms were not affected by repeated flashes.
Statistical significance in data comparing wild-type with
AQP4-null mice was taken as P < 0.05, using the
Student’s t-test (InStat 2.03 software; GraphPad, San
Diego, CA). ERG waveforms were analyzed for a- and b-wave amplitude and
latency, the leading edge of a-wave, and oscillatory potential
amplitude and latency, as described in the Results section. Data are
reported as mean ± SE for measurements in 10 mice in each group.
Averaged amplitudes and latencies were compared between wild-type and
AQP4-null mice at each light intensity.