Selective Photoreceptor Degeneration by Intravitreal Injection of N-Methyl-N-Nitrosourea

Abstract

Purpose.: To characterize the effects of intravitreal injections of N-methyl-N-nitrosourea (MNU) in comparison to its systemic application as a measure of inducing unilateral photoreceptor degeneration.

Methods.: Eight-week-old male C57BL/6J mice received either intraperitoneal injections (three animals) or intravitreal injections (24 animals) of MNU in different concentrations and were observed over a period of 2 weeks using full-field electroretinography (ERG), spectral-domain optical coherence tomography (SD-OCT), and immunohistochemistry.

Results.: The intraperitoneal application of MNU showed moderate systemic toxic effects, indicated by a loss of body weight of 12% within the first 2 days. In both eyes the ERG became extinguished, and SD-OCT scans showed a thinning of the retina, predominantly in the outer nuclear layer (ONL). Immunohistochemistry demonstrated the selective loss of rods and cones. Mice that received intravitreal MNU injections displayed nearly no weight loss, and no degeneration of their general welfare was observed. After 2 weeks, ERG, SD-OCT, and immunohistochemistry revealed changes identical to those seen after systemic application in the injected eye, but not in the control eye.

Conclusions.: The intraperitoneal application of MNU led to moderate systemic side effects in mice and to selective photoreceptor degeneration. Intravitreal injections of MNU also induced photoreceptor degeneration; however, no systemic side effects were observed. This tool may be helpful in larger species, where genetic models of receptor degenerations are not applicable but where the size of the eye is more suitable to study surgical or other approaches to treat blindness caused by receptor degeneration.

Introduction

Inherited retinal dystrophies, such as retinitis pigmentosa (RP), are important and frequent causes of blindness and are observed in approximately 1 of 4000 persons. ^1,2 Research efforts are focused on the mechanisms of degeneration of these diseases and the genetic background, as well as on treatment strategies. Genetic models in rodents expressing retinal degeneration, for example, the retinal degeneration (rd) 1 mouse, the rd10 mouse, ^3–13 and the RCS (Royal College of Surgeons) rat, ^14–17 are indispensable tools with which to explore experimental therapies for currently untreatable conditions. In rd1 and rd10 mice, mutations in the gene encoding the β-subunit of rod cGMP phosphodiesterase result in photoreceptor degeneration with a central-to-peripheral progression, which is in contrast to human RP. ^3–13 Because of the milder retinal degeneration and later onset compared to the rd1 mouse, the rd10 mouse provides a better genetic mouse model for human RP. ¹² In the RCS rat, a deletion in the gene encoding the receptor tyrosine kinase (Mertk) expressed in retinal pigment epithelium (RPE) cells leads to a disturbed phagocytosis of photoreceptor outer segments and consequently to photoreceptor degeneration by apoptosis. ¹⁷ Regardless of the cause of outer retinal degeneration in the rd1 mouse, the rd10 mouse, and the RCS rat, a progressive loss of photoreceptors with age, comparable to that in RP, has been observed, which is due to apoptosis. ¹⁸ The thickness of the inner nuclear layer and the ganglion cell layer remained unaffected. ^12–17 However, cellular as well as organizational and functional changes in the inner retina were observed. ^7,19 Beside stem cell and gene therapy, retinal implants have been developed for treatment of RP and are currently tested in humans. Further development and improvement of such implants require experimental surgical approaches that can be more easily performed in larger animal species than in mice or rats—for example, in pigs, cats, or rabbits exhibiting a specific loss of photoreceptors. However, genetic models in larger animals are difficult to utilize due to both the cost and duration of disease manifestation. ²⁰  

Photoreceptor degeneration may also be induced by extrinsic physical measures, such as light exposure or ionizing radiation on the developing retina. ^21,22 Retinal degenerations are also observed after the systemic application of pharmaceuticals such as iodoacetic acid (IAA) ^19,23,24 or N-methyl-N-nitrosourea (MNU). ^2,18,25–29 After systemic treatment, MNU leads to extensive oxidative stress in a dose-dependent manner, resulting in retinal photoreceptor cell death, due to apoptosis, as confirmed by TUNEL staining. ^18,25–27 Photoreceptors can be rescued by several inhibitors of apoptosis. ^26,27 The toxic effects of MNU on photoreceptors in a 7-day period after systemic application have been described in different animal species. ^25–29 In mice and rats, a concentration of 60 mg/kg body weight (BW) MNU was effective when applied intraperitoneally, while in rabbits the intravenous application of 40 mg/kg BW effectively caused photoreceptor degeneration. ^25–29 Similar to what is seen in the genetic models in mice, the loss of photoreceptors proceeded with a central-to-peripheral gradient. ¹⁸ However, while in rd1 and rd10 mice, rods degenerate first, followed by a secondary degeneration of cones, rods and cones were lost simultaneously and to the same extent in MNU-treated eyes. ^25–29 The number of cells within the inner nuclear layer and within the ganglion cell layer did not decrease; but the inner retinal neurons showed features of morphologic remodeling comparable to observations in the genetic models. ^25–29 A side effect of systemic MNU application is the induction of neoplasms. ^30–32 After intravenous injection in rabbits, 68.8% of the animals developed brain tumors. ³⁰ Furthermore, tumors frequently occurred as small intestine carcinomas and as multiple vessel wall sarcomas in different organs. ³⁰ In rats, mammary tumors or prostate tumors can be induced by a single intraperitoneal or intravenous injection of 50 mg/kg BW MNU. ^31,32 Tumors were macroscopically detectable 12 to 16 weeks after injection. ^31,32  

We were motivated to identify a safe and reproducible procedure for inducing unilateral outer retinal degeneration by intravitreal application of pharmaceuticals, thereby avoiding systemic side effects and keeping the contralateral eye as an intraindividual nontreated control eye. Therefore, we investigated the effects of intravitreal injections of MNU on the mouse retina. 

Materials and Methods

All experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, in accordance with the German Law for the Protection of Animals, and after approval had been obtained by the regulatory authorities. 

Animals

Adult pigmented male wild-type mice (C57BL/6J, 8 weeks) were housed in standard conditions under 12/12-hour light/dark cycle with water and food available ad libitum. Because of the possible environmental hazard due to the carcinogenic and teratogenic potential of MNU, the animals were kept in special one-way cages (PET; Tecniplast Deutschland GmbH, Hohenpreißenberg, Germany) after each MNU application under fume hoods, with the same light and food conditions as before. The mice received sterile MNU (Sigma-Aldrich, Steinheim, Germany) at different concentrations, either intraperitoneally injected (three animals, 60 mg/kg BW) or intravitreally injected into the left eye (24 animals, 0.15–3 mg/kg BW). Because of its poor stability in aqueous solutions and the sustained time needed for intravitreal injections, MNU was dissolved in dimethylsulfoxide (DMSO; Serva, Heidelberg, Germany) immediately before use and further diluted with sterile phosphate-buffered saline (PBS). Animals were examined 5 days before as well as 7 and 14 days after the injection of MNU to monitor the effects of the injection on retinal morphology and photoreceptor function. Body weight was recorded on a daily basis for the entire duration of the experiment. Two weeks after treatment, the animals were euthanized for histologic examinations by an overdose of isoflurane (Forene 100% [vol/vol]; Abbott GmbH, Wiesbaden, Germany) and decapitated. 

Mice were anesthetized with ketamine (70 mg/kg Ketamin 10%; CEVA, Düsseldorf, Germany) and xylazine (10 mg/kg Xylazin 2% Bernburg; Medistar, Ascheberg, Germany) in 0.1 mL saline for in vivo measurements, examinations, and intravitreal injections. For the systemic treatment with MNU, no anesthesia was given. 

Systemic MNU Application

N-methyl-N-nitrosourea solution was slowly administered intraperitoneally in three mice at a concentration of 60 mg/kg BW. This concentration has been previously reported to cause photoreceptor degeneration in mice. ²⁷  

Intravitreal MNU Injection

Intravitreal injections were performed using a 27-gauge steel needle (BD Microlance 3; BD Biosciences, Heidelberg, Germany) and a nanoliter injection system with a three-dimensional (3D) micromanipulator (Nanoliter 2000; World Precision Instruments, Inc., Sarasota, FL). After puncture in the dorsonasal area of the limbus, 2 μL was injected with a glass capillary at an angle of 30°. Preliminary control experiments with PBS showed that the injection technique was safe and that the solution (defined volume of 2 μL) could be placed in the vitreous cavity without touching the lens or even the retina. Different concentrations of MNU were injected intravitreally into the left eye of 24 mice (0.15, 0.2, 0.25, 0.3, 0.5, 1.0, 2.0, 3.0 mg/kg BW, diluted with DMSO and PBS; n = 3 for each concentration). 

All intravitreally injected mice were treated with antibiotic eye ointment (Polyspectran; Alcon Pharma GmbH, Freiburg, Germany) up to 3 days after injection to avoid inflammatory reactions. 

Control Interventions

Two animals were kept under the same husbandry conditions to evaluate systemic effects of changing the cage conditions from normal cages to the one-way cage system. Further control animals received intraperitoneal sham injections (PBS, n = 3, or DMSO+PBS, n = 3). 

Intravitreal control injections of PBS (2 μL), DMSO (maximally 1 μL diluted to 2 μL with PBS), and fluorescein (different concentrations dissolved in 2 μL PBS) were performed to confirm the safety of the injection technique. Absence of leakage from the injection area directly after injection of fluorescein was confirmed. To this end, images were taken by a confocal laser scanning microscope (Heidelberg Retina Angiograph-1; Heidelberg Engineering, Heidelberg, Germany) in the fluorescein angiography mode. 

All control animals (PBS, DMSO+PBS) were examined using the same methods and the same time schedule as for the experimental animals. 

Electroretinogram Recordings

After a dark adaptation of 1 hour, pupils were dilated with one drop of Tropicamide 2.5% eye drops (Pharmacy of the University Hospital Aachen, Aachen, Germany). Mice were placed in front of a full-field dome stimulator. The eyes were locally anesthetized with proxymetacainhydrochloride 0.5% eye drops (Proparakain-POS; Ursapharm, Saarbrücken, Germany), and a gold ring electrode (animal electrode 0.5 ø 3 mm; Roland Consult, Brandenburg, Germany) was placed on the corneal surface of each eye. Nonirritating artificial tears were used for corneal clarity and for better ionic conduction (methylcellulose, Methocel 2%; Omni Vision, Puchheim, Germany). A gold ring electrode in the mouth (contact with the mouth mucosa) served as a reference electrode, and a subcutaneous needle electrode in the lumbar region served as a ground electrode. Electroretinography was performed using the Reti System designed for rodents (Roland Consult). ³³ Mice were examined by full-field ERGs, according to the International Society for Clinical Electrophysiology of Vision (ISCEV) standard protocol with five light stimuli per recording. ³⁴ For each recording, a- and b-wave amplitudes, as well as implicit times of rod and cone responses, were determined by averaging the responses of the five light stimuli. 

Spectral Domain Optical Coherence Tomography

After anesthesia, mice were positioned in front of the Spectralis OCT system (Heidelberg Engineering) for spectral-domain optical coherence tomography (SD-OCT). The scanning system was modified by adding a +25-diopter (D) lens (Heidelberg Engineering) to correct for rodent optics. ³³ The optic disc was located as the main landmark, and retinal thickness measurements were conducted using the horizontal cross-sectional image centered on the optic disc. The total retinal thickness, as well as the thickness of the different layers, was measured at six positions across the retina. Thickness values of each eye were averaged. For documentation purposes, confocal images of the fundus of the evaluated eyes were recorded (inner part of the retina [IR] reflection image, wavelength 715 nm). To further document possible macroscopic changes of the treated or the noninjected contralateral eyes, pictures were taken by a Nikon D80 camera (AF Micro Nikkor 105 mm, 2896 pixel × 1944 pixel × 24 bit; Nikon, Düsseldorf, Germany). 

Histology³³

Eyes were processed for immunohistochemistry as well as for hematoxylin and eosin staining. Immunohistochemistry was performed as described earlier by Mataruga et al. ³⁵ In brief, eyes were enucleated and opened by an encircling cut at the limbus. The retina in the eye cup was immersion fixed for 30 minutes in 4% paraformaldehyde (PA) in 0.1 M phosphate buffer (PB) at room temperature and washed in PB several times. Tissue was incubated in 10% sucrose/PB for 1 hour, followed by 30% sucrose/PB overnight. The retina was flat embedded and frozen in optimal cutting temperature compound (NEG-50, Richard-Allan Scientific; Thermo Fisher Scientific, Germany). Vertical sections (20-μm thickness) were cut on a cryostat (HM 560 CryoStar; Microm, Walldorf, Germany) and collected on Superfrost Plus slides (Menzel, Braunschweig, Germany). Sections were pretreated with blocking solution (5% Chemiblocker [Chemicon, Hofheim, Germany], 0.5% Triton X-100 in PB, and 0.05% NaN₃) for 1 hour, followed by incubation with primary antibodies overnight, and diluted in the same solution. Primary antibodies directed against the following antigens were used: antiglial fibrillary acidic protein (GFAP, raised in chicken, 1:2000; Novus Biologicals, Cambridge, UK), glutamine synthetase (raised in mouse, 1:4000; BD Biosciences), anti-calbindin 28K (CabP; raised in mouse, 1:1000; Sigma-Aldrich), anti-protein kinase C α (PKCa; raised in rabbit, 1:4000; Santa Cruz, Heidelberg, Germany), calretinin (AB1550, raised in goat, 1:3000; Millipore, Schwalbach, Germany), HCN1 (RTQ-7C3, raised in rat, 1:10; F. Müller, Forschungszentrum Jülich GmbH, Jülich, Germany), recoverin (AB5585, raised in rabbit, 1:2000; Millipore), and rhodopsin (1D4, 1:500; R.S. Molday, British Columbia, Canada). Sections were washed in PB and incubated in secondary antibodies diluted in 5% Chemiblocker (Chemicon), 0.5% Triton X-100 in PB for 1 hour, washed in PB, and coverslipped with Aqua Polymount (Polysciences, Eppelheim, Germany). Secondary antibodies included donkey anti-chicken Cy2 (1:400; Dianova, Hamburg, Germany), donkey anti-mouse Cy3 (1:100; Dianova), donkey anti-rabbit Cy2 (1:400; Dianova), donkey anti-goat Alexa 647 (1:200; Invitrogen, Germany), donkey anti-rat Cy3 (1:500; Dianova), and donkey anti-mouse Dy649 (1:500; Dianova). The lectin peanut agglutinin (PNA, biotinylated, 1:1600; Sigma-Aldrich) was visualized using streptavidin Alexa 647 (1:100; Invitrogen, Darmstadt, Germany). Sections were examined with a Leica TCS confocal laser scanning microscope (Leica Microsystems, Heidelberg, Germany) with ×63/1.4 oil immersion lenses. Images were processed and printed with Adobe Photoshop (Adobe Systems, München, Germany). For triple labeling, primary antibodies were mixed and applied simultaneously. All secondary antibodies were highly cross-absorbed and were carefully tested to exclude reactions with the wrong primary antibody. Concentration of the antibodies, laser intensity, and filter settings were carefully controlled; and the sequential scanning mode was employed to completely rule out cross-talk between the fluorescence detection channels. Band pass filters of 500 to 530 nm for green fluorescence (Cy2), 580 to 650 nm for red fluorescence (Cy3), and 680 to 750 nm for infrared fluorescence (Alexa 647, Dy649) were used. 

For hematoxylin and eosin staining, eyes were enucleated, punctured three times at the limbus, and fixed by immersion (described above). Eyes of the animals were dehydrated in a tissue dehydration automat (MTM; SLEE, Mainz, Germany) by incubation in a series of increasing ethanol concentrations (2× 70%, 2× 96%, 3× 100% for 1 hour), followed by xylene (3× for 1 hour) and paraffin (4× for 1 hour), and embedded in paraffin. Sections (5 μm thick) were cut with a microtome (R. Jung, Heidelberg, Germany), collected on slides, deparaffinized, rehydrated, and stained with hematoxylin and eosin. Pictures were taken using a Leica DMRX microscope (Leica Microsystems). 

The thickness of the entire retina and of the individual retinal layers was determined in immunohistochemical staining at six distinct positions close to the papilla. For each retina, thickness values obtained from these positions were averaged. 

Urine Test

Before and also after 2 and 3 hours and 1 and 2 days after systemic and intravitreal application of MNU, urine was collected and analyzed by ultra-performance liquid chromatography (UPLC, Aquity; Waters, Eschborn, Germany) to evaluate the excretion of MNU for safety reasons. 

Results

Systemic Application of MNU

Weight Loss.

Intraperitoneal application of MNU resulted in systemic side effects with a weight loss of approximately 12% within the first 2 days. Figure 1 demonstrates weight loss after different forms of MNU application in order to reflect the animals' well-being after MNU treatment and under changed husbandry conditions in one-way cages. 

Figure 1

 

Side effects of MNU application. Loss of body weight as indicator of general well-being after application of MNU given intraperitoneally (i.p.) and intravitreally (IVI) in comparison to untreated animals and control animals in the same changed husbandry conditions. **,***P < 0.003 versus i.p. MNU application, one-way ANOVA with Bonferroni's post hoc test. N-methyl-N-nitrosourea i.p.: n = 3; control DMSO+PBS i.p.: n = 3; untreated: n = 2; intravitreal (IVI) each concentration: n = 3.

Electroretinography.

Systemically administered MNU led to the extinction of the scotopic ERG responses (Figs. 2C–E) and the photopic (data not shown) ERG responses after 1 and 2 weeks. 

Figure 2

 

Scotopic ERG after systemic MNU application. (A–C) Examples of scotopic full-field ERG recordings from mice before (black trace) and 1 week after (gray) systemic intraperitoneal application of PBS (A), DMSO+PBS (B), and 60 mg/kg BW MNU (C) under the same husbandry conditions. Scotopic 3 cds/m², 0.067 Hz. Only with administration of MNU (C) was extinction of the ERG curve 1 week after application observed. (D, E) Amplitudes of a-wave (D) and b-wave (E) before and 2 weeks after i.p. MNU application (n = 3) at a light intensity of 3 cds/m² in the scotopic ERG. Mean ± SD. DMSO+PBS (n = 3) and PBS (n = 3) injections served as controls (both eyes of each animal were evaluated).

Spectral-Domain Optical Coherence Tomography.

One week after intraperitoneal MNU application, a reduction in total retinal thickness was observed, while the space between the retina and the RPE seemed to be filled with cell debris (Fig. 3C). After 2 weeks, the debris and the somata of the photoreceptors (outer nuclear layer, ONL) had disappeared completely, and the reduction of retinal thickness was more pronounced, as demonstrated in Figure 3C. Two weeks after injection, the inner nuclear layer became adjacent to the RPE (Fig. 3C). 

Figure 3

 

Thickness evaluation in OCT. Retinal images and thickness obtained by SD-OCT measurements. All images were taken at the level of the optic disc and are displayed with pigment epithelium and photoreceptors to the top. Examples of SD-OCT recordings before (left), 1 week after (middle), and 2 weeks after (right) intraperitoneal PBS (A), DMSO+PBS (B), or MNU (C) application. Red arrows (B) indicate examples for the thickness measurement in one retina. (C) N-methyl-N-nitrosourea 60 mg/kg BW i.p. The red bars mark the outer nuclear layer, which is significantly thinner 1 and 2 weeks after injection; enlargement of the area next to the optic disc shows the destruction of the ONL and the debris in the area of the ONL. (D) Retinal thickness measurements based on SD-OCT data before, 1 week, and 2 weeks after i.p. injection of PBS, DMSO+PBS, and MNU (n = 3). Values represent mean ± SD (for 60 mg/kg BW 1 and 2 weeks after application: ****P < 0.0001 versus the controls, one-way ANOVA with Bonferroni's post hoc test).

Microscopy.

Histologic analysis confirmed the reduction of retinal thickness as seen in SD-OCT recordings in vivo, the pronounced loss of photoreceptors, and the thinning of the ONL after MNU treatment. Representative sections are shown in Figure 4. Under normal conditions, GFAP was expressed only in astrocytes. However, after MNU treatment, Müller cells were labeled with anti-GFAP (Fig. 4D), indicating that they had become reactive, as also seen with other causes of retinal degeneration. ³⁶ In animals that had received vehicle injection only (Fig. 4B), the retina remained unaffected. In animals treated intraperitoneally with MNU, the ONL was reduced entirely or to one row of photoreceptor somata (Fig. 4C), while the thickness of the inner retina was not statistically different from the thickness in control eyes. The thinning of the outer retina was reminiscent of the photoreceptor loss observed in RP or in late stages of rodent models of RP such as in the rd10 mouse (Fig. 4E). ^6–13  

Figure 4

 

Confocal images of immunolabeled vertical sections obtained from retinae prepared 2 weeks after systemic application. The white bars in (A) and (B) indicate the thickness of the ONL. Red: glutamine synthetase, Müller cells spanning nearly the entire thickness of the retina. Blue: PNA, cone outer segments and endfeet. Green: GFAP. (A) Wild type, (B) DMSO+PBS i.p., and (C) 60 mg/kg BW MNU i.p. Note that the ONL has nearly vanished. (D) Glial reaction is indicated by GFAP expression in Müller cell processes (same section as in [C]). (E) Rd10 mouse retina at the age of 50 days (P50) displays total loss of photoreceptors. OR, outer part of the retina; OS, outer segments; IS, inner segments; OPL, outer plexiform layer; INL, inner nuclear layer; GCL, ganglion cell layer; NFL, nerve fiber layer.

Intravitreal Injection of MNU

Weight Loss.

Animals that received PBS, DMSO+PBS, or MNU as intravitreal injections showed no remarkable weight loss or other systemic side effects (see Fig. 1). 

Electroretinography.

The control interventions, that is, injection of PBS or DMSO+PBS, as well as the intravitreal injection of 0.15 up to 0.5 mg/kg BW MNU, had no effect on the ERG (Figs. 5A–C, 5G–I). 

Figure 5

 

Scotopic and photopic ERGs after intravitreal MNU application. Scotopic ERGs: (A–F) examples of scotopic full-field ERG recordings from mice before (black trace) and 1 week after (gray trace) intravitreal application of PBS (A), DMSO+PBS (B), 0.5 mg/kg BW MNU (C), 1 mg/kg BW MNU (D), 2 mg/kg BW MNU (E), and 3 mg/kg BW MNU (F). (D–F) A reduction of the amplitude of the ERG or a complete extinction is seen after intravitreal injection of 1, 2, and 3 mg/kg BW MNU. In (F), the ERG of the untreated right eye, 1 week after intravitreal injection of the left eye with 3 mg/kg BW, is illustrated (magenta). Treatment of the left eye had no influence on photoreceptor function in the right eye. The amplitudes of a-waves (G) and b-waves (H) in the scotopic ERG are illustrated before and 2 weeks after MNU application in different concentrations at a light intensity of 3 cds/m² in the scotopic ERG (n = 3 for each concentration). In each diagram, the bar designated CL3 illustrates the values obtained from the nontreated right eye of animals treated intravitreally with 3 mg/kg BW in the left eye. Values represent mean ± SD. Photopic ERGs: (I–L) examples of photopic full-field ERG recordings from mice before (black trace) and 1 week after (gray trace) IVI treatment with PBS (I), 1 mg/kg BW MNU (J), 2 mg/kg BW MNU (K), and 3 mg/kg BW MNU (L). The b-wave in the photopic ERG (3 cds/m²; 0.625 Hz) was diminished 1 week after intravitreal treatment with MNU (J–L). The contralateral eye (CL3) of the animal treated with 3 mg/kg BW revealed a normal photopic ERG (magenta trace in [L]).

In animals treated with 1 and 2 mg/kg BW, scotopic (Figs. 5D, 5E) and photopic (Figs. 5J, 5K) ERGs were affected. For both concentrations, in two of three animals the amplitudes of the a-wave and the b-wave in the scotopic and photopic ERG were diminished, and in one animal the ERG was completely abolished. In all animals injected intravitreally with 3 mg/kg BW, the scotopic (Fig. 5F) and photopic (Fig. 5L, gray trace) ERG was completely extinguished. 

Spectral-Domain Optical Coherence Tomography.

Spectral-domain OCT measurements 1 week after the injection of 1, 2, and 3 mg/kg BW revealed cell debris between the inner retina and the pigment epithelium, similar to the findings observed after intraperitoneal application (compare Figs. 6D–F to Fig. 3C). Upon injection of 1 or 2 mg/kg BW MNU, cell death in the ONL was not homogeneous. In the superior peripheral retina, that is, close to the injection site, most or all of the cell layers in the ONL had vanished, while in the inferior retina the thickness of the ONL was reduced by less than 50%. In animals that had received 3 mg/kg BW MNU, the ONL had disappeared throughout the entire retina. 

Figure 6

 

Retinal thickness evaluation using OCT upon intravitreal injection of MNU. Retinal images and thickness obtained by SD-OCT measurements. All images were taken at the height of the papilla and are displayed with pigment epithelium and photoreceptors to the top. Examples of SD-OCT recordings before (left), 1 week after (middle), and 2 weeks after (right) injection of PBS (A), DMSO+PBS (B), 0.5 mg/kg BW MNU (C), 1 mg/kg BW MNU (D), 2 mg/kg BW MNU (E), and 3 mg/kg BW MNU (F). (D–F) The red bars mark the outer nuclear layer, which is significantly thinner 1 week after injection. The destruction of the ONL 1 week after application is reminiscent of the effects observed after intraperitoneal application. (G) Retinal thickness measurements at the height of the papilla based on SD-OCT data before and 2 weeks after IVI injection of PBS, DMSO+PBS, and MNU. CL3 illustrates the retinal thickness of the contralateral, untreated eyes of animals injected intravitreally in the left eye with 3 mg/kg BW MNU. Values represent mean ± SD (for 3 mg/kg BW MNU 2 weeks after application: ****P < 0.0001 versus the controls and the MNU-treated animals before injection, one-way ANOVA with Bonferroni's post hoc test). In (H), the untreated right eye of one animal intravitreally injected in the left eye with 3 mg/kg MNU (F) 2 weeks after injection is illustrated. No thickness differences or morphologic changes compared to the eyes of control animals were observed (H).

Microscopy.

The reduction of retinal thickness upon intravitreal injection of 1, 2, and 3 mg/kg BW MNU observed in the OCT was verified by histology (Fig. 7). The MNU treatment led to a thinning of the ONL, while the thickness of the inner retina was not affected in a statistically significant way (Fig. 7G). 

Figure 7

 

Confocal images of immunohistochemically stained vertical sections through the retina 2 weeks after injection of MNU into the vitreous. The thickness of the ONL is indicated by white bars. Staining and abbreviations as in Figure 4. (A) Control injection of PBS, (B) control injection of DMSO+PBS, (C) 1 mg/kg BW MNU, (D) 2 mg/kg BW MNU, and (E) 3 mg/kg BW MNU. (C–E) Thickness of the ONL is progressively reduced. No photoreceptors are left after injection of 3 mg/kg BW MNU (E). The bar diagrams in (F) and (G) show the retinal thickness after different forms of treatment and in comparison to the genetic model rd10 at the age of 50 days (P50, n = 3). While in (F) the entire retinal thickness was evaluated, in (G) only the thickness of the inner retinal layers is demonstrated. (F) Retinal thickness decreases with rising intravitreal concentrations of MNU. After intravitreal injection of 3 mg/kg BW MNU the retinal thickness is comparable to the thickness after intraperitoneal (i.p.) treatment with 60 mg/kg BW MNU or to the thickness observed in the genetic model rd10. Retinal thickness in the contralateral untreated eye (CL3) is not affected. In (G), no influence of the intravitreal injection on the thickness of the inner retinal layers is observed.

In animals that had received 1 or 2 mg/kg BW MNU, immunohistochemical data (Figs. 7, 8) confirmed that the degree of ONL thinning depended on the distance to the injection site. Figure 7 shows images that were taken close to the papilla, whereas in Figure 8, sections through the superior parts (close to injection site) and inferior parts of the retina are shown, additionally stained for rods (Figs. 8B, 8D, 8F, 8H) and cones (Figs. 8C, 8E, 8G, 8I). Thickness loss in the ONL after treatment with 1 and 2 mg/kg BW MNU was more pronounced close to the injection site (Figs. 8B, 8D; note that cones are positive for red/green opsin, typical for the superior retina). In contrast, in the inferior retina, more layers of photoreceptor somata remained (cone outer segments were positive for blue opsin, typical for the inferior retina). 

Figure 8

 

Confocal images of immunohistochemically stained vertical sections through the retina 2 weeks after injection of MNU into the vitreous. (A) Illustration depicts the injection site in the superior part of the eye. (B, D, F, H) Staining against recoverin (green, entire photoreceptor), rhodopsin (blue, rod outer segment), and ion channel HCN1 (red, photoreceptor somata and inner segments, as well as processes in the inner retina). (C, E, G, I) Staining against red/green opsin (green, cone outer segments in the superior retina) (C, E), blue opsin (red, cone outer segments in the inferior retina) (G, I), PEA (cones, blue). OR, outer half of the retina; IR, inner half of the retina. Photoreceptor degeneration is more pronounced at the injection site. Both rods and cones show features of degeneration. (B, D, F, H) Staining of HCN1 channels (red) and recoverin in type 2 bipolar cells (green) does not reveal changes in the inner retinal layers of animals treated with 1 and 2 mg/kg BW MNU.

After treatment with 1 and 2 mg/kg BW, surviving rods and cones revealed signs of degeneration (Fig. 8). In Figures 8B, 8F, 8D, and 8H, antibodies against rhodopsin and recoverin were used. Rhodopsin present in rod outer segments was visualized in blue, while recoverin, which is found throughout the entirety of photoreceptors, is shown in green. Therefore, rod outer segments appear in cyan. They are shorter than normal in all sections shown, but are longest in regions most remote from the injection site (Fig. 8F, 1 mg/kg BW MNU, inferior part of the retina). Cones also showed signs of degeneration. In Figure 8G (1 mg/kg BW MNU, inferior retina), cones survived best, showing long outer segments stained in red for blue opsin. Closer to the injection site (Fig. 8C, 1 mg/kg BW MNU, superior retina), two abnormal features were observed. First, red/green opsin staining (green) was found throughout all cone compartments, while normally only small amounts of opsin were observed outside the outer segment. Secondly, the outer segments were short and round. Close to the injection site after 2 mg/kg BW MNU (Fig. 8E), cones were completely gone (green staining represents background in inner retinal cells); and in the inferior part of the retina (Fig. 8I), outer segments (red) were extremely shortened. 

In contrast to these findings, intravitreal injection of 3 mg/kg BW MNU resulted in a complete loss of the ONL throughout the entire retina (Fig. 7E). 

To determine, in detail, the effects of intravitreally injected MNU on inner retinal cells in contrast to wild type and after systemic application of MNU, additional staining was performed (Fig. 9; first row: CabP, horizontal cells; second row: PKCα, rod bipolar cells; third row: calretinin, amacrine cells). In wild type (Fig. 9, first column), CabP was strongly expressed in horizontal cell somata and their dendrites (upper cells, arrow) and weakly in certain amacrine cells (lower part of the section). In wild-type rod bipolar cells (Fig. 9, second row), somata were usually elongated, axons were clearly visible, and axon terminals were compact. Certain amacrine cell types and three typical bands formed by their dendrites in the inner plexiform layer (IPL) were stained with antibodies against calretinin (Fig. 9, third row). 

Figure 9

 

Confocal images of immunohistochemically stained vertical sections through the retina 2 weeks after injection of MNU. First column: wild-type control. Second column: 60 mg/kg BW MNU intraperitoneal (i.p.). Third column: 1 mg/kg BW MNU intravitreal (IVI). Fourth column: 2 mg/kg BW MNU IVI. Fifth column: 3 mg/kg BW MNU IVI. Antibodies against calbindin 28K (CabP) yield strong label in horizontal cells (arrow) and weaker staining in amacrine cells. Protein kinase C alpha is found in rod bipolar cells. Calretinin (Cal) is found in certain populations of amacrine cells.

Upon intraperitoneal application of MNU (Fig. 9, second column), fine horizontal cell dendrites were lost as the presynaptic photoreceptors had disappeared, but horizontal cell somata were mostly unchanged (Fig. 9, first row). Some rod bipolar somata (Fig. 9, second row, arrow) were more roundish than usual, but most rod bipolar cells were unaffected. Amacrine cells appeared unchanged (Fig. 9, third row). 

Upon intravitreal injection of 1 mg/kg BW (Fig. 9, third column), 2 mg/kg BW (Fig. 9, fourth column), or 3 mg/kg BW MNU (Fig. 9, fifth column), the following results were obtained. In those regions of the retina where photoreceptors survived (1 and 2 mg/kg BW MNU), horizontal cells (Fig. 9, first row), rod bipolar cells (Fig. 9, second row), and amacrine cells (Fig. 9, third row) were basically unaffected. The situation was different in those retinal regions where all photoreceptors had disappeared, that is, throughout the entire retina upon injection of 3 mg/kg BW MNU (Fig. 9, fifth column) and close to the injection site upon injection of 1 or 2 mg/kg BW MNU (not shown). Here, horizontal cells were missing (Fig. 9, first row); rod bipolar cell somata became roundish (Fig. 9, second row, third column, arrow); their axons and axon terminal systems were disorganized; and more amacrine cells expressed PKC (Fig. 9, second row, third column, arrowhead). The calretinin-positive bands in the IPL were slightly broader (Fig. 9, third row, third column). 

In all cases, the intravitreal injection did not affect the noninjected eye of the animal. N-methyl-N-nitrosourea was not found in any of the urine test samples (data not shown). 

Discussion

Genetic animal models of retinal degenerations revealing selective photoreceptor degeneration, such as RP, are important to study. In this way possible treatment approaches can be established and underlying disease mechanisms can be detected. Several rodent models are already well established. ^3–17 In larger animals, which would be more suitable to work on with surgical methods, genetic models are intrinsically expensive and disease progression is described as slow. ²⁰ Photoreceptor degeneration can also be induced pharmacologically and therefore in any animal species desired. ^23–29 One possible pharmacologic agent, MNU, was reported to lead to blindness, caused by selective photoreceptor apoptosis via oxidative stress, when administered intraperitoneally or intravenously. ^25–29 The resulting histologic picture was analyzed in detail by different researchers ^25–29 and is very similar to that observed in human RP. ²⁵ However, systemic application of MNU leads not only to bilateral retinal degeneration, but also to a reduction of the general health status of the experimental animals. Besides short-term effects caused by the toxicity of the substance, the induction of tumors has been described as a long-term effect in rabbits and rats after systemic treatment with MNU, due to its DNA alkylating mode of action. ^30–32 In this study we showed that it is possible to induce photoreceptor degeneration in only one eye of mice by local administration of MNU, while at the same time systemic side effects were avoided. 

Systemic Application of MNU

N-methyl-N-nitrosourea given as an intraperitoneal injection served as our positive control. In these animals, well-being was reduced for some hours after administration. They showed a weight reduction, the highest loss 2 days after injection; however, all animals survived. Regarding the outcome of systemic MNU application, a rapid and nearly complete denudation of the ONL was observed. Cells of the inner retina like horizontal, bipolar, or amacrine cells were only slightly affected. As the presynaptic photoreceptors were lost, horizontal cells and rod bipolar cells lost most of their dendrites; this result is similar to findings described earlier. ²⁹  

Intravitreal Injections of MNU

Due to the small vitreous space of the mouse eye and the proportionally large lens, intravitreal injections in mice are difficult. To optimize the injection technique, we used a 3D micromanipulator and nanoliter injection system. Injections of PBS, fluorescein, and DMSO+PBS mixture were used as negative controls, proving that the injection procedure was safe. Examination of control animals revealed neither a change in functionality in vivo nor histologic abnormalities. In contrast to the systemic application, in animals receiving MNU as intravitreal injection, well-being was only minimally affected, which was comparable to welfare after an anesthetic intervention. 

Effect on Photoreceptors

One week after intravitreal injection of 1, 2, or 3 mg/kg BW MNU, the OCT showed changes similar to those observed 1 week after intraperitoneal application. However, 2 weeks after 1 and 2 mg/kg BW MNU, not all somata of photoreceptors in the ONL had disappeared. The degree of photoreceptor degeneration depended on the distance to the injection site. While at the point of injection in the superior retina all photoreceptors had disappeared, 5 to 10 rows of photoreceptor somata were detectable in the inferior retina (Fig. 8). Detailed analysis using photoreceptor markers (Fig. 8) indicated that the surviving photoreceptor population comprised both rods and cones, but that outer segments were shorter than usual. Standard full-field scotopic and photopic ERG waves in animals injected with 1 and 2 mg/kg BW MNU were either diminished (two animals in each group) or even extinguished (one animal in each group, n = 3; Figs. 5D, 5E, 5J, 5K). 

Upon injection of 3 mg/kg BW MNU, the photoreceptors disappeared throughout the entire retina. Immunohistochemistry performed 2 weeks after injection revealed no surviving photoreceptor population. In agreement with this finding, no responses were observed in the scotopic and photopic ERG 2 weeks after injection of MNU (Figs. 5, 7). 

Effect on Inner Retinal Layers

We observed no statistically significant differences between the thickness of inner retinal layers of intravitreally injected eyes (1, 2, or 3 mg/kg BW) and uninjected eyes. The histologic picture was reminiscent of the results obtained after systemic treatment with MNU and results obtained with the genetic model (rd10) on postnatal day 50 (Fig. 7). 

In those regions of intravitreally injected eyes where photoreceptors had vanished entirely, the horizontal and rod bipolar cells also seemed to be affected. Two weeks after injection, we noticed a complete loss of horizontal cell somata and processes in the retinal regions where all photoreceptors had disappeared. While after systemic MNU treatment a loss of horizontal cell processes in a study over 3 months has been described, ²⁹ complete cell loss has not been observed. Recently, Tsuruma et al. ²⁷ reported a decrease in the thickness of the inner nuclear layer after intraperitoneal treatment with 75 mg/kg BW MNU instead of the 60 mg/kg BW used in our study. However, the authors did not investigate whether the thickness reduction was due to horizontal cell loss. In comparison, in the rd10 mouse at the age of 9 months, approximately 29% of horizontal cell somata were lost. ⁷  

The rod bipolar cells seemed to be disorganized after intravitreal injection of 3 mg/kg BW MNU. For rd10 mice up to 45 days old, no changes in the morphology, size, and complexity of the axonal terminal systems of rod bipolar cells were found; but for rd10 mice between 1.5 and 3.5 months of age, the number of rod bipolar cells decreased by 20%. ⁷  

There are two ways to explain the effect of MNU on horizontal and rod bipolar cells. First, for higher concentrations of MNU, as they are reached at the site of injections with use of 1 or 2 mg/kg BW MNU or throughout the eye with use of 3 mg/kg BW, MNU might become toxic to horizontal cells and rod bipolar cells. A detailed investigation still needs to be carried out to determine whether this toxicity is due to oxidative stress, as suggested for photoreceptors, or to DNA alkylation. Secondly, the rapid loss of all photoreceptors at a given retinal location within 2 weeks might have a more pronounced effect on the survival of postsynaptic cells than the photoreceptor degeneration occurring at a slower pace in the genetic model. 

Comparison to Previous Studies

Other pharmacologic agents have also been reported to induce photoreceptor degeneration, among them IAA and adenosine triphosphate (ATP). ^{19,23,24,37,38} However, photoreceptor degeneration was observed only if IAA was applied systemically, not after intravitreal administration (Rösch S, Johnen S, Mazinani B, Müller F, Pfarrer C, Walter P, manuscript submitted 2013). Puthussery and Fletcher ³⁷ and Notomi et al. ³⁸ demonstrated that intravitreally injected ATP induces photoreceptor degeneration by apoptosis. However, ATP also initiates apoptosis in the ganglion cell layer, as revealed by TUNEL staining. ³⁷ One possible reason is excessive calcium influx in ganglion cells upon stimulation of the P2X7 receptor by ATP. ³⁸ Moreover, after intravitreal injection of ATP, degeneration occurred to a variable degree, ranging from the destruction of photoreceptor outer segments to a complete loss of the RPE. In regions of rapid and strong retinal degeneration, highly disturbed synapses in the outer plexiform layer were reported. ³⁷ In contrast, Tsuruma et al. ²⁷ reported that MNU induced radical production only in cell cultures of murine 661W photoreceptor-derived cells, but not in RGC-5, a mouse ganglion cell line, indicating a higher specificity or efficacy of MNU for photoreceptor cells. It still remains unclear why MNU toxicity is specific for photoreceptors and also may be for horizontal cells. N-methyl-N-nitrosourea administered intravitreally needs to penetrate or needs to be transported through the entire retina in order to affect the photoreceptors. It is likely that photoreceptors are particularly susceptible due to their extremely high metabolic turnover. ³⁹ However, further experiments are needed to clarify the mode of MNU action. 

Conclusions

In summary, our results clearly indicate that MNU given intravitreally is a working model for the induction of photoreceptor degeneration to simulate disease processes as seen in RP or similar conditions. The mouse model has the shortcomings of a small eye. It was difficult to achieve a homogeneous degeneration of photoreceptors throughout the retina with the intravitreal injection of 1 or 2 mg/kg BW. Photoreceptor loss was more pronounced close to the injection site. This problem may not be encountered in larger eyes, such as in the rabbit, as injections of MNU at multiple positions in the large vitreal volume might result in a more homogeneous elimination of photoreceptors. Based on these results, we would suggest the use of MNU injections into the vitreous of mice or larger animals as a pharmaceutical model to specifically induce degeneration of photoreceptors while maintaining an intraindividual control eye. 

Acknowledgments

The authors thank Rene H. Tolba and the Institute for Laboratory Animal Science of the University Hospital RWTH Aachen for the opportunity to access one of their laboratories and to perform our experiments under special safety precautions. We further thank Christoph Aretzweiler (Institute of Complex Systems, Cellular Biophysics, ICS-4, Forschungszentrum Jülich GmbH, Jülich, Germany) for the excellent support in preparation and staining of the eyes. 

This manuscript is part of a doctoral thesis at the University of Veterinary Medicine Hannover, Hannover, Germany (January 10, 2013). 

Supported by Deutsche Forschungsgemeinschaft (DFG) Grants WA 1472/6-1 (PW) and MU 3036/3-1 (FM). 

Disclosure: S. Rösch, None; S. Johnen, None; A. Mataruga, None; F. Müller, None; C. Pfarrer, None; P. Walter, None