Evaluation of the Rhodopsin Knockout Mouse as a Model of Pure Cone Function

Abstract

purpose. To determine a time window in the rhodopsin knockout (Rho^−/−) mouse during which retinal function is already sufficiently developed but cone degeneration is not yet substantial, thus representing an all-cone retina.

methods. Electroretinograms (ERGs) were obtained from 14 homozygous Rho^−/− mice and eight C57Bl/6 control mice. The same individuals were tested every 7 days, beginning as early as postnatal day (P)14. The ERG protocols included flash and flicker stimuli, both under photopic and scotopic conditions. Retinal and choroidal morphology was observed in animals of comparable age.

results. Functionally, the developmental phase lasted until postnatal week (PW)3 in both the Rho^−/− mice and the control animals. During PW4 to 6, the Rho^−/− mice showed a plateau in ERG parameters with normal or even supernormal cone responses and complete absence of rod contributions. At PW7, there was a marked onset of degeneration, which progressed so that no ERG signals were left at PW13, when the control eyes still had normal ERG responses. Microscopically, cone degeneration paralleled the functional changes, beginning at approximately PW6 and almost complete at PW13, whereas retinal pigment epithelium (RPE) and choroid did not show any abnormalities.

conclusions. From PW4 to 6, Rho^−/− mice appear to have normal cone and no rod function. Despite the missing rod outer segment (OS), the structure of retina, RPE, and choroid remained unchanged. Therefore, the Rho^−/− mice can serve during this age period as a model for pure cone function. Such a model is particularly useful to evaluate rod–cone interaction and to dissect rod- from cone-mediated signaling pathways in vivo.

The disease process in human hereditary retinal degenerations is often difficult to analyze, because the time course is usually slow, and material for morphologic studies is rare. The availability of transgenic animal models has helped much to obtain deeper insights in the pathophysiology of these disorders. Humphries et al. ¹ introduced one of these models, the rhodopsin knockout mouse, in 1997. The Rho^−/− mouse was primarily used as a model for retinitis pigmentosa (RP), a heterogeneous group of hereditary disorders of the rod system that cause progressive retinal degeneration. It is one of the most frequently seen retinal dystrophies, with an estimated incidence of 1 in 3500 to 1 in 4500 human births. All modes of Mendelian inheritance—autosomal dominant (ad), autosomal recessive (ar) and x-linked—have been described. At least two mutations, Glu249Ter ² and Glu150Lys, ³ have been reported to cause arRP due to absence of rhodopsin. RP is characterized by night blindness, progressive concentric visual field defects leading to tunnel vision, and a reduced to nondetectable electroretinogram (ERG). ^{4 5 6} Ocular findings include pallor of the optic disc, retinal vessel attenuation, and intraretinal pigment deposits (bone spicules). Rhodopsin, the visual pigment of the rods, is necessary to start the phototransduction cascade and also serves as a structure protein of the discs in the rod outer segments (OSs). Because homozygous (Rho^−/−) mice carry a replacement mutation in exon 2 of the rhodopsin gene, they show a complete absence of rhodopsin and do not build rod outer segments. Similar findings were reported for knockout mice carrying a different rhodopsin null mutation. ⁷ Eventually, this condition leads to progressive cone degeneration, ⁸ a common feature known from human RP. It is not clear whether this degeneration is mostly due to a direct effect caused by the absence of regular rods, some factors released by them, ⁹ or an indirect effect through damage or altered development of the RPE and/or the choroid. 

Besides the importance as a model for a specific disease, transgenic animals can foster a better understanding of the normal retinal physiology and pathologic processes in general. For example, the CNG3 knockout mouse, which we have recently characterized as a model for achromatopsia, ¹⁰ can also be used to dissect rod from cone-mediated signaling pathways in vivo, because it features pure rod function. In this study, we have examined whether there is a period in the young Rho^−/− mice when the ERG is already sufficiently developed but there is no substantial cone degeneration yet. Given that no rod function is possible in the Rho^−/− mice, they can be used to study pure cone function during this interval. 

Methods

Animals

The Rho^−/− mice ¹ were kindly provided by Peter Humphries (Department of Genetics, Trinity College, The University of Dublin, Ireland). To investigate the time course of development and degeneration of the Rho^−/− retina, retinal function was tested by Ganzfeld electroretinography. Assessment of the morphology was done in vivo by funduscopy and scanning laser ophthalmoscopy and, in animals of comparable age, by histology. The electrophysiological study included 14 Rho^−/− mice on a C57Bl/6 background and 8 C57Bl/6 control mice, which carried the wild-type rhodopsin gene. One of the Rho^−/− mice died during the course of the study and was excluded from the results. ERGs were recorded from P14 to 15 onward, with the same individual mouse tested every 7 days. The recordings were grouped by age to yield one group per postnatal week (PW)—that is, the PW4 group contained all records obtained from mice between P21 and P25. As an exception, PW3 was split into two groups, P14 to 15 and P16 to 18, because of developmental changes of the ERG response during this early period. The final group PW13 contained data from P91 to 95. The research was performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Electroretinograms

ERGs were obtained according to previously described procedures. ⁹ In summary, mice were dark adapted overnight (at least 6 hours) before the experiments and their pupils dilated with tropicamide (Mydriaticum Stulln; Pharma Stulln, Nabburg, Germany) and phenylephrine (Neosynephrin-POS 5%; Ursapharm, Saarbruecken, Germany) eye drops. Anesthesia was induced by subcutaneous injection of ketamine (66.7 mg/kg body weight; Ketanest; Parke-Davis, Berlin, Germany), xylazine (11.7 mg/kg body weight; Rompun; Bayer Vital, Leverkusen, Germany), and atropine (1 mg/kg body weight; Atropin; Braun, Melsungen, Germany). Silver needle electrodes served as reference (forehead) and ground (tail) and gold wire ring electrodes as active electrodes. Methylcellulose (Methocel; Ciba Vision, Wessling, Germany) was applied to ensure good electrical contact and to keep the eye hydrated during the entire procedure. The recording setup featured a Ganzfeld bowl, a DC amplifier, and a computer-based control and recording unit (Toennies Multiliner Vision; Jaeger/Toennies, Höchberg, Germany). ERGs were recorded from both eyes simultaneously after the mice were placed in the Ganzfeld bowl. Band-pass filter width was 1 to 300 Hz for single-flash and flicker-stimuli recordings. 

Single-flash and flicker recordings were obtained both under dark-adapted (scotopic) and light-adapted (photopic) conditions. Single-flash stimuli were presented with increasing intensities, reaching from 10⁻⁴ candelas · second per m² (cds/m²) to 25 cds/m², divided into ten steps of 0.5 and 1 log cds/m². Five to ten responses were averaged with an inter-stimulus interval (ISI) of 5 seconds (for 10⁻⁴, 10⁻³, 10⁻², 3 × 10⁻², 10⁻¹, 3 × 10⁻¹ cds/m²) or 17 seconds (for 1, 3, 10, and 25 cds/m²). Flicker stimuli had an intensity of 3 cds/m² with frequencies of 0.5, 2, 5, 10, 15, and 30 Hz. Light-adaptation was performed with a background illumination of 30 cds/m² presented for 10 minutes to reach a stable level of the photopic responses. ¹¹  

Fundus Photography

In vivo black-and-white images of the fundus were obtained in Rho^−/− and control mice at 3 months of age. The pictures were taken with a scanning laser ophthalmoscope (SLO; model 101; Rodenstock, Munich, Germany) together with a 20-D lens (Volk, Mentor, OH). Varia-tion of the focus of the laser beam allowed visualization of different layers of the fundus (i.e., retina versus pigment epithelium). 

Histology

For immunohistochemistry, light microscopy was obtained from Rho^−/− and control mice aged P14 up to P70. Cones were identified by biotin-conjugated peanut agglutinin (Lectin PNA), ¹² using 12-μm sections. 

For histologic studies of the retinal vasculature, animals were deeply anesthetized with ether and intraperitoneally applied pentobarbital sodium (50 mg/kg body weight). The chest was then opened, and the vessels transcardially perfused with heparinized plasma substitute (Expafusin; Pfrimmer, Erlangen, Germany). For vascular corrosion casts, three animals at P42 were transcardially perfused with liquid plastic, containing Araldite CY 223 (45%), hardener HY 2267 (25%), and acetone (30%), by thumb pressure as described previously in mice. ^{13 14} After hardening of the plastic for 24 hours the eyes were enucleated and macerated in concentrated KOH, with the solution being changed several times. The eyes were then fixed on holders. The right eyes were oriented for investigating the choriocapillaris, whereas the left eyes were oriented for investigating the choroidal vasculature from the external side. The specimens were sputtered with gold and studied with a scanning electron microscope (Stereoscan 90; Leica Cambridge, Cambridge, UK). 

For light- and electron microscopy investigation, two animals at P42 were transcardially perfused with 4% paraformaldehyde. The eyes were enucleated and postfixed in a solution containing glutaraldehyde (2.5%), paraformaldehyde (2.5%), and picric acid (0.05%) in cacodylate buffer (pH 7.3) for at least 24 hours. After they were rinsed in cacodylate buffer, the eyes were dissected sagittally and the lenses removed carefully. Small specimens were postfixed in 1% OsO₄, dehydrated in ascending series of alcohols, and embedded in Epon according to standard methods. Semithin sections were cut using a microtome (Ultracut OmU3; Reichert, Vienna, Austria) and stained with toluidine blue. Ultrathin sections were stained with uranyl acetate and lead citrate and viewed in an electron microscope (EM 902; Zeiss, Oberkochen, Germany). 

Results

Electroretinography

A set of representative ERG records of a C57Bl/6 control mouse at P37 is shown in the left column of Figure 1 . The other three columns to the right depict the results obtained under the same conditions in Rho^−/− mice at ages P14, P35, and P63. The first three rows were obtained under scotopic (dark adapted) conditions, whereas the final two rows depict the photopic results after light adaptation with a rod-desensitizing background of 30 cds/m². 

Control Mice

In the scotopic single-flash ERG series (Fig. 1 , first row), the b-wave emerged at a flash intensity of approximately 10⁻⁴ cds/m². With increasing stimulus intensity, its amplitude became larger and its implicit time shortened. At flash intensities of approximately 0.03 to 0.1 cds/m², a clear a-wave developed. The scotopic flicker ERG (Fig. 1 , second row) at a frequency of 0.5 Hz reflected practically a single-flash waveform. With increasing frequency, the amplitude diminished and the shape of the signal became more and more similar to the photopic flicker response. 

Under photopic conditions with a background of 30 cds/m² (Fig. 1 , third row), a distinct single-flash ERG response was usually not found at flash intensities lower than approximately 0.03 to 0.1 cds/m². The shape of the b-wave was more spread out than the dark-adapted counterparts, and no clear a-wave was visible, even at a stimulus intensity of 10 cds/m². The photopic flicker ERG (Fig. 1 , final row) did not show a transition of waveform shape with frequency comparable to the scotopic flicker. 

Rho^−/− Mice

The rightmost three columns in Figure 1 show exemplary recordings of a Rho^−/− mouse performed at P14, P35, and P63 to demonstrate variations of the ERG response with age. A general finding was an increase in amplitude with a shortening of implicit time from P14 to P35 and a decrease in amplitude together with a prolongation of implicit time between P35 and P63. 

Variation from Normal

The scotopic flash ERG (Fig. 1 , first row) of the P35 Rho^−/− mouse became detectable at intensities much higher than that of control eyes (above 10⁻² cds/m²), and grew further in amplitude with increased flash intensity. Also, the b-wave implicit time was considerably slower than that of the control eyes, and even at the highest flash intensity used, the response of the knockout mouse showed no distinct a-wave. The scotopic flicker ERG (Fig. 1 , second row) at a frequency of 0.5 Hz reflected, similar to that in control eyes, practically the single-flash waveform. With increasing frequency, the amplitude diminished less than in control eyes, and the shape of the signal changed only slightly, in that it was already similar to the photopic flicker response. 

Under photopic conditions with a 30 cds/m² background (Fig. 1 , third and fourth rows), there were no marked differences in the shape of the waveform. However, amplitudes in the Rho^−/− mice were in general larger than in the control eyes (“supernormal” responses). 

Time Course of Changes

Figure 2 shows amplitude and implicit time of the b-wave responses of the Rho^−/− and the control group to stimuli of 10 cds/m², determined once a week between PW3 and PW13. The left column refers to scotopic and the right to photopic conditions. In both the Rho^−/− (solid line, gray boxes) and the control mice (dashed line, white boxes), there was a manifest developmental phase during which the scotopic b-wave amplitude initially increased, together with a reduction in implicit time. A functional plateau was reached at around PW4 in the knockout mice’s eyes and PW5 to 6 in the control eyes. Between PW4 and PW5, cone responses of the Rho^−/− mice were even supernormal in terms of amplitude size and implicit time, most obvious under photopic conditions that enable a direct comparison to normal. Whereas the b-wave remained stable thereafter in the control eyes, there was a marked onset of functional loss in the Rho^−/− mice at PW7. After a steady loss of amplitude and an increase of implicit time, no clear ERG signals were detectable after PW13. In contrast, the ERGs of the control mice remained stable, at least until PW13. The time window of stable ERGs in the Rho^−/− mice between PW4 and PW7 is marked in gray. 

V log I Function

Because there was no rod function, big differences between the scotopic ERGs of the Rho^−/− mice and the C57BL/6 control mice can be expected in a- and b-wave amplitude and sensitivity. These differences can be seen in the so-called V log I function, a plot of the a- and b-wave amplitude versus the logarithm of stimulus intensity (Fig. 3) . The results for the a-wave are shown in Figure 3A , and those for the b-wave in Figure 3B . Both graphs make use of box-and-whisker-plots showing 5% and 95% quantiles (whiskers), 25% and 75% quartiles (box), and the median (marked by a dot). Data recorded under scotopic conditions are shown in black, and those recorded under photopic conditions in gray. The Rho^−/− mice data are shown by dashed lines and open boxes and the data of control mice by solid lines and striped boxes. Both the knockout and the control mice were 5 weeks old, which is within the previously established time window of stable ERG recordings of the Rho^−/− mice. 

Amplitude and sensitivity of the scotopic b-wave of the knockout mice were greatly reduced in comparison to normal (Fig. 3B , black lines). There were already responses in the control eyes at the minimum intensity used (10⁻⁴ cds/m²), whereas in the knockout mice, there was no response discernible for flash intensities of less than 3 × 10⁻² cds/m². Amplitudes of the knockout mice were almost always lower than those of control eyes, although the difference between control eyes and knockout mice decreased with stimulus intensity. Under photopic conditions, the curves for control eyes and Rho^−/− mice were very similar (Fig. 3B , gray lines); first responses were visible at approximately 10⁻² cds/m². Again, the responses of the knockout mice were supernormal. 

The findings were very different for the a-wave amplitude (Fig. 3A) . A clear a-wave was present only in the control eyes under scotopic conditions, emerging at a flash intensity of 3 × 10⁻² cds/m². In all other cases, the a-wave was found to be very small to nondetectable, although it appeared to be again supernormal in Rho^−/− mice under photopic conditions. 

Flicker ERG

The results of flicker ERGs from both knockout and control eyes in response to stimuli of varying frequency (0.5–5 Hz) but constant intensity (3 cds/m²) are shown in Figure 4 (top, scotopic conditions; bottom, photopic conditions). The scotopic amplitude of the control eyes (top; white boxes and dashed line) at a frequency of 0.5 Hz was much higher than that of the Rho^−/− mice (gray boxes, solid line). However, at frequencies of 2 Hz and more, this relationship reversed, and the median of the C57BL/6 mice was lower than that of the knockout mice. In contrast, under photopic conditions the flicker amplitudes of the Rho^−/− mice were supernormal at all frequencies (bottom; gray boxes, solid line). 

Macroscopic Fundus Appearance

Fundus imaging with a scanning laser ophthalmoscope (SLO) revealed that there was no visible degeneration in the Rho^−/− mice between postnatal weeks PW4 and 6, the period of stable ERG recordings. Figure 5 shows the regular fundus appearance in a Rho^−/− mouse at 3 months of age (Fig. 5B) and a control mouse at the same age (Fig. 5A) . 

Microscopic Appearance of the Retina and Choroid

Microscopically, cone degeneration begins at about PW6 and leads to an almost complete loss of outer segments until PW13. Figure 6 illustrates the sequence of changes that can be observed histologically, beginning with normal-looking cones at P14 in a PNA staining. At P58, the degeneration of cones is clearly visible, and in advanced stages, shown here for P77, no clear cone outer segments are detectable. 

Under light microscope, the retinas of Rho^−/− mice at PW12, the period of almost complete loss of cone function, showed degeneration of the outer nuclear layer (ONL). Only one to two layers of nuclei were seen in the central and peripheral regions (Fig. 7) . OSs were not visible in the entire retina. The outer limiting membrane appeared complete. In some areas, remnants of inner segments were present between the outer limiting membrane and the RPE. The remaining retinal layers appeared normal. The RPE cells were present and of normal shape in all areas. 

Electron microscopy of the central retina showed RPE cells with elongated apical microvilli and basolateral infoldings. The pigmentation of the RPE cells and the size and number of phagolysosomes appeared normal (Fig. 8) . The adjacent cones showed signs of degeneration in the cytoplasm of their cell bodies. Only remnants of their inner segments with swollen mitochondria and only remnants or no connecting cilia were in contact with the microvilli of the RPE. Outer segments were completely absent (Fig. 8) . The outer limiting membrane formed by Müller cells was well developed. 

Bruch’s membrane was normal and showed no disruptions or thickening. The choriocapillaris was highly fenestrated and appeared the same as in age-matched control eyes (Fig. 8) . 

Scanning electron microscopy of corrosion cast preparations of Rho^−/− mice at PW12 showed the typical branching pattern of the choroidal vessels. The nasal and temporal long ciliary arteries were most prominent and branched directly from the ophthalmic artery. One larger and two smaller vessels supported the inferior and superior quadrant respectively. This vessel pattern was the same as in control C57Bl/6 animals of the same age described previously. ¹⁴ The choriocapillaris formed a dense layer (Fig. 9A ). In the vortex region, the larger venules were located within the capillary layer. There was also no difference from control eyes. 

In the area of the optic nerve head, the central retinal artery branched into six to eight main branches and formed the typical three-layered vascular bed seen also in control mice (Fig. 9B) . The arteries could be identified by their clear impression of endothelial cell bodies into the inner surface, whereas the surface of the venules was smooth. 

Discussion

This work was triggered by a search for a mouse model that could serve electrophysiologically as a model for an “all-cone retina” as a counterpart to the CNG3 knockout mice ⁹ featuring pure rod function. Indeed, because the Rho^−/− mouse has no functional rhodopsin gene, there is no doubt that rod function is completely absent, ^{1 6} both because of the impaired phototransduction and the failing morphogenesis of rod OSs. ¹⁵ However, because many other models never have normal cone ERGs, it was not clear how well cone function would be preserved and how it would change over time. The purpose of this study was thus to determine a time window during which retinal function in the young Rho^−/− mouse would be sufficiently developed, with cone degeneration not yet substantial. 

To achieve that goal, we studied a cohort of newborn Rho^−/− mice and control mice for 6 months to avoid potential bias associated with cross-sectional approaches. Using special care, it was possible to obtain complete ERG records in these mice from P14 forward, which enabled us to determine the maturation as well as the decline of the ERG responses in the same individual animals. The results demonstrate that there are important changes of retinal function with time and that cohort studies using repetitive tests in weekly intervals are well suited to record these changes. 

Because the Rho^−/− mice, having no rod visual pigment rhodopsin, do not have any rod function, their photopic and scotopic ERG signals are entirely generated by cones. Thus, principal ERG differences between the C57BL/6 control mice and the Rho^−/− mice concern particularly the scotopic waveforms, as these are the ones most influenced by rods (Fig. 1) . The results indicate that cones contributed only slightly to the generation of the a-wave at the stimulus intensities used. This is in accordance with other findings that demonstrate that the a-wave is not reduced in mice that are completely without cone function. ⁹ The scotopic ERG of the Rho^−/− mice and the photopic waveforms of both the control eyes and the Rho^−/− mice looked very similar, because they represented the pure cone ERG in two different states of adaptation. In analogy, the higher scotopic flicker amplitude at a frequency of 0.5 Hz in the control eyes (Fig. 4) was due to a mixed response from rods and cones. At higher frequencies and under photopic conditions, only cones contributed to the ERG, leading to slightly higher amplitudes in the knockout mice due to supernormal responses. Again, this is consistent with our findings in CNG3^−/− mice that rods can follow flicker stimuli of that intensity (3 cds/m²) only up to a frequency of 2 to 3 Hz. ⁹  

The developmental phase (PW3–4) of both the Rho^−/− eyes and the control eyes was characterized by a steady increase in amplitude, paired with a drastic decrease of implicit time (Figs. 1 2) . This maturing process of the retina was also found in several studies on ERG development in other species. ^{16 17 18 19} However, because the amplitude in the knockout mice did not increase slightly as in the control eyes (Fig. 2) , it was hard to tell whether the retina had completed its maturation or the additional increase due to development and the decrease due to degeneration (which would then have to be very mild at this point) balanced each other. Between PW4 and 6, Rho^−/− mice appeared even to have supernormal cone function, followed by normal ERG levels between PW 6 and 7 (Fig. 2) . Supernormal cone function has been clearly demonstrated in a transgenic pig model of retinal degeneration due to a P347L mutation in the rhodopsin gene ²⁰ and has been attributed to an arrest in retinal development. Similar results appear also to be present in a knockout rat model (Artur V. Cideciyan, personal communication). In an earlier study on Rho^−/− mice, ⁷ the responses of the knockout mice were found to be somewhat larger than in wild-type mice, but this difference was not significant. Because no age information was given in this part of the study, this result was probably due to a pooled population between PW4 to 5 and PW7, which may have obscured the initial supernormal responses. However, the marked onset of ERG amplitude loss from PW8 onward was also found in the above study. We could additionally demonstrate that both the maturation and the degenerative phases were accompanied by an increased implicit time, the latter of which had also a marked onset between PW7 and PW8. This onset was somewhat less distinct in the photopic ERG because of the short implicit times associated with the initially supernormal responses. The functional deterioration progressed, until at PW13 no perceivable ERG response was detectable in most cases, whereas the responses of the C57BL/6 control mice remained stable. 

Both the loss of amplitude and the increase in implicit time are typical markers for the process of cone degeneration in RP and allied diseases in humans. ²¹ Because these changes have been shown to be restricted to the actual defective area, ^{22 23} the marked onset of amplitude loss and implicit time delay in Ganzfeld ERG indicates a relatively rapid impairment of a substantial proportion of the retina, possibly by apoptosis. The mechanism of how recessively and dominantly inherited rhodopsin mutations eventually cause cone disease, a process found both in animal models ^{8 24} and patients with RP, ^{25 26} is still unclear. It appears that cone dysfunction develops if more than 75% of rod OSs are lost. ²⁷ There are indications for the existence of a rod-released diffusible factor that may be necessary for the survival of cones. ⁹  

Morphologically, the ONL thickness in Rho^−/− mice is reduced until P45 to 40% to 50%, and at P80, one to two rows of photoreceptor nuclei remain. The time course of cone-specific changes is shown in Figure 6 . In addition to the previously described histologic sequence of photoreceptor degeneration, ^{1 7 8} we have also studied a potential involvement of the RPE and the choroid. It turned out that despite the absence of normal rod OSs at any phase during development, all other retinal layers, the RPE, and the choroid in the Rho^−/− mice were regularly shaped. Even after an almost complete degeneration of the cone OSs at PW12 and only one row of nuclei in the ONL left, the other retinal layers, the RPE and the choriocapillaris remained normal (Figs. 7 8 9) . Likewise, funduscopy in Rho^−/− mice at PW15, displaying a complete loss of photoreceptors and therefore a nondetectable ERG, yielded no apparent macroscopic fundus changes. 

In summary, the data show that between PW4 and PW7 there is a normal to slightly supernormal ERG in the absence of rod responses in the rhodopsin knockout mouse. The microscopic and macroscopic findings indicate that during the period of stable ERG recordings and beyond, the Rho^−/− mice exhibit well-preserved retinal layers and choroidal vasculature. Thus, the young rhodopsin knockout mouse can be used as a functional model for an all-cone retina. Possible applications of this model range from studies on the normal physiology of cone responses and rod–cone interaction in the visual pathway to the creation of double knockout mice to assess how certain genes affect cone function. 

 Part of this work has been presented at the XXXVII ISCEV (Society for Clinical Electrophysiology of Vision), Eilat, Israel, April 1999.

 Supported by Deutsche Forschungsgemeinschaft Grant SFB 430 C2, IKFZ Tuebingen, SFB 539 (Glaukome), 1ZKF Erlangen, and fortuene Grant 517.

 Submitted for publication August 24, 2000; accepted October 26, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Sequence of ERG waveform changes in a Rho^−/− mouse. Exemplary ERG records from a Rho^−/− mouse at postnatal day 14 (P14), P35, and P63. For comparison, records from a C57Bl/6 control obtained at P37 are shown.

Figure 2.

 

Time course of ERG changes in the Rho^−/− mouse. Top row: b-Wave amplitudes are shown in the Rho^−/− mice and the control group in response to a single flash of 10 cds/m² as a function of age. Bottom row: Corresponding implicit times. Left: The time course of the scotopic recordings; right: the photopic recordings. Gray-shaded area: Time window of stable ERGs in the Rho^−/− mice.

Figure 3.

 

Amplitude versus intensity (V log I) function for a- and b-wave. (A) a-Wave and (B) b-wave amplitude versus stimulus intensity for the Rho^−/− mice and the C57Bl/6 control eyes, both under scotopic and photopic conditions.

Figure 4.

 

Frequency dependence of flicker ERG responses. Top: Amplitude of responses to a scotopic flicker stimulus (3 cds/m²) obtained from the Rho^−/− mice (gray boxes, solid lines), and the control group (white boxes, dashed lines) plotted versus flicker frequency. Bottom: Same experiment under photopic conditions.

Figure 5.

 

Macroscopic view of the retina. Fundus appearance using indirect ophthalmoscopy. Arrow: Optic nerve head. (A) Rho^−/− mouse aged 3 months. (B) C57Bl/6 control at the same age.

Figure 6.

 

Microscopic view of the retina. Light microscopy using cone-specific PNA staining. Cross sections of the Rho^−/− retina at three postnatal time points. Arrow: Cone OSs (P14) or their remnants (P58) stained with PNA. (☆), RPE; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer.

Figure 7.

 

Light micrograph of a 1-μm thick cross section of the retina and choroid of a Rho^−/− mouse at PW12 (Richardson stain, magnification, ×400). The thickness of the ONL is highly decreased. Outer segments were completely absent. The inner retinal layers, the RPE, and the choroid appeared normal.

Figure 8.

 

Electron micrograph of the outer retina, RPE, and choroid in the central part of the globe of a Rho^−/− mouse at PW12. Between outer limiting membrane (arrowhead) and retinal pigmented epithelial cells (RPE) remnants of inner segments are visible. OSs are completely missing. The RPE and Bruch’s membrane appeared normal with apical microvilli (arrow) and basolateral infoldings. The choriocapillaris (CC) is highly fenestrated. Magnification, ×4500.

Figure 9.

 

Scanning electron micrographs of corrosion casts of Rho^−/− mice at PW12. (A) Choriocapillaris); (B) branches of the central retinal vessels). Both, the distribution pattern of the central retinal vessels and the choriocapillaris appear normal. Magnification, (A) ×1100; (B) ×180.

The authors thank Karin Mai for excellent technical assistance with the ERG recordings and Gudrun Haerer for help with histology.