In Vivo Imaging of Ocular MCMV Infection

Abstract

Purpose.: To develop a technique by which murine cytomegalovirus (MCMV) infection can be confirmed and monitored in vivo in various ocular compartments and to investigate the dynamics and time course of primary ocular CMV infection.

Methods.: The ability of recombinant MCMV-expressing enhanced green fluorescent protein (eGFP) to serve as a tool to monitor the in vivo dynamics of experimental intraocular CMV infection was examined. Immunocompetent BALB/c mice were infected subretinally with eGFP-MCMV. Confocal scanning laser ophthalmoscopy (SLO) was used to visualize viral spread in vivo on sequential days after infection. Eyes were processed for histology and immunofluorescence microscopy to confirm viral infection and replication by means of GFP signal.

Results.: Retina was readily permissive to primary infection with eGFP-mCMV, and fluorescent signal was detected by SLO 24 hours after subretinal injection, with scattered foci around the posterior pole of the retina. GFP levels in the retina reached a maximum on day 6. Signal in the iris developed from day 4 and lasted until day 25. Examinations of retinal and iris tissue wholemounts by immunofluorescence revealed signal localized to the outer retina, iris stroma, and anterior lens capsule.

Conclusions.: The ability to noninvasively monitor infectious agents in the eye may improve current knowledge of the course and pathogenesis of intraocular infections and could lead to further clarification of the mechanisms by which the immune system responds to intraocular pathogens.

In vivo imaging, especially intravital microscopy, has advanced our understanding of dynamic biological processes and represents a rapidly growing field in medical science. ¹ This is particularly true for the investigation of cellular interactions in the course of immunologic events such as antigen presentation and T-cell dynamics in lymphoid organs. ² In contrast to intravital imaging of most tissue, which requires exposure or exteriorization of the target tissue, the eye presents a unique opportunity to analyze in vivo processes without previous surgical manipulation or trauma. ³ Recent reports on leukocyte trafficking in the murine cornea and iris have helped us gain a more detailed understanding of their dynamics at a cellular level. ⁴ In addition, investigation of real-time leukocyte dynamics in the murine retina have been established using scanning laser ophthalmoscopy. ^3,5 Although applications of intravital imaging have focused mainly on dynamics of immune cells, in vivo study of fluorescence-labeled infectious agents within the eye has, to our knowledge, received little attention. 

With these issues in mind we sought to monitor the dynamics of experimental intraocular infection with enhanced green fluorescent protein (eGFP)-labeled cytomegalovirus (CMV) as a model infection for human CMV (HCMV). This recombinant fluorescent virus has the ability to produce green fluorescent progeny virus and thus allows monitoring of viral replication and spread. Murine CMV (MCMV) is a ubiquitous beta herpesvirus that usually causes a mild, self-limiting primary infection in the eye after inoculation into immunocompetent mice. ^6–8 As do all herpesviruses, MCMV replicates in the nucleus of infected cells. ⁹ After primary infection, the virus persists in the host and eventually establishes lifelong latency. ^10,11 The possibility of ocular persistence, both in mice and humans, is still debated. ^12,13 Several case reports of ocular MCMV reactivation after targeted ocular immunosuppression in immunocompetent patients ^14–17 suggest the virus can persist within the eye after primary infection. In immunocompromised human hosts, HCMV has emerged as a significant opportunistic pathogen, and the eye is one of the organs commonly affected, with potentially blinding consequences. ¹⁸ However, despite well-known characteristics of ocular features of CMV infection in immunocompromised hosts, the pathogenesis of primary ocular infection is not well understood. The importance of HCMV disease in humans motivated us to investigate whether the dynamics of primary MCMV infection can be monitored in vivo in the eye. 

In the present study, we used in vivo confocal scanning laser ophthalmoscopy to visualize viral spread to the retina and combined that with ex vivo confocal microscopy of immunostained retinal, lens, and iris wholemounts ¹⁹ to document the in vivo pathophysiology of primary MCMV infection in the murine eye. This technique of noninvasive monitoring of fluorescence-tagged ocular pathogens will enhance the understanding of the pathophysiology of ophthalmic infections and may be useful in the assessment of treatment efficacy over time in experimental models. 

Materials and Methods

Mice

Inbred female BALB/c (H-2d, I-Ad) mice 8 weeks of age were obtained from the Animal Resources Centre (Murdoch, Western Australia, Australia). During experimentation, all mice were housed under specific pathogen-free conditions at the Animal Services Facility of the University of Western Australia. All animals were treated according to the Statement for the Use of Animals in Ophthalmic and Vision Research promulgated by the Association for Research in Vision and Ophthalmology, and animal experimentation was performed with the approval of the Animal Ethics and Experimentation Committee of the University of Western Australia and according to the guidelines of the National Health and Medical Research Council of Australia. 

Virus Stock

To prepare viral stocks mice were infected by intraperitoneal administration of 5 × 10³ plaque-forming units (pfu) of salivary gland-propagated MCMV diluted in PBS-0.05% fetal calf serum (FCS). Progeny virus was recovered from salivary glands 18 days after inoculation. Salivary gland viral progeny was purified by stepped sucrose gradient before pure viral stocks were prepared for intraocular injection. The viruses used were MCMV-K181-Perth (referred to as MCMV-wt) or MCMV-K181-Perth-eGFP (MCMV-eGFP), which encodes the eGFP gene in the intron of the m129–m131 open reading frame. A productive MCMV infection was confirmed by positive GFP signal. 

Virus Injections

To determine whether MCMV would infect ocular cells in immunocompetent BALB/c mice, 1.0 μL (10³ pfu/μL) MCMV-K181 or MCMV-eGFP was injected into the subretinal space through a beveled 35-gauge needle (NF35BV-2, World Precision Instruments, Sarasota, FL) mounted to a microsyringe (Nanofil;, World Precision Instruments) with the aid of an operating microscope. The needle was placed obliquely behind the limbus of the eye, as described previously. ⁶ Controls received mock injections with 1.0 μL sterile PBS. 

Tissue Collection and Processing for Immunohistochemistry

Mice were killed and perfusion fixed using 2% paraformaldehyde, as previously described. ¹⁹ The retina was dissected free and cut into quadrants, as documented. ¹⁹ Tissue pieces were washed in PBS, incubated in 20 mM EDTA tetrasodium at 37°C for 30 minutes, and incubated with a 0.2% solution of Triton-X in PBS with 2% BSA at room temperature for 30 minutes to assist antibody penetration. Further washes with PBS were followed by overnight incubation (4°C) with a range of monoclonal antibodies, including anti-MHC class II 1:200 (M5/114; BD PharMingen, San Jose, CA), anti-CD45 1:100 (Serotec, Raleigh, NC), anti-CD11b 1:100 (BD PharMingen), anti-GFAP 1:200 (glial fibrillary acidic protein) for glial tissue (Serotec), anti-NeuN 1:200 (neuron-specific protein; stains predominantly ganglion cells in the retina; Chemicon, Temecula, CA), ²⁰ and isotype controls 1:100 (IgG2a and IgG2b; BD PharMingen). Tissues were then treated with biotinylated goat anti-rat IgG 1:200 (Amersham Biosciences, Piscataway, NJ) at room temperature for 60 minutes, washed with PBS, and incubated with streptavidin-Cy3 1:250 (Jackson ImmunoResearch Laboratories, West Grove, PA) at room temperature for 60 minutes. 4′,6 -Diamidine-2′-phenylindole dihydrochloride (DAPI) 1:250 (Roche Diagnostics, Mannheim, Germany) was added at room temperature for 7 minutes as a nuclear stain. Stained wholemount tissues were then mounted onto slides (retinas were mounted with the vitreous side face up) using an aqueous mounting medium (Immunomount; Thermo Shandon, Waltham, MA) and were coverslipped. 

Microscopy

Semiquantitative assessment of stained specimens was performed by conventional epifluorescence microscopy (BX60; Olympus, Tokyo, Japan) or by confocal microscopy (TCS SP2; Leica, Wetzlar, Germany). Tissue wholemounts were analyzed by performing Z-stacks of the tissue from the internal to the external aspect at increments ranging from 0.5 to 1.5 μm. 

Retinal Imaging

We used a commercially available scanning laser ophthalmoscope, the Heidelberg Retina Angiograph 2 (HRA 2; Heidelberg Engineering, Dossenheim, Germany), for retinal imaging of the mice. The imaging system was adapted for the optics of the mouse eye with a 55° wide-angle lens. This reduces the laser beam diameter and allows more light to be coupled into the small mouse pupil. For fluorescein angiography, BALB/c mice were injected with 10% sodium fluorescein at a dose of 0.02 mL/5 g body weight. These mice were not used for tracking of MCMV-eGFP to exclude contamination with the fluorescein signal. Because our primary objective was to track GFP signal, we operated the HRA 2 in the fluorescence (FA) mode with the 488-nm argon laser providing the excitation light. The emission signal was collected by a photodetector using the 500-nm barrier filter. To localize the site of viral replication, native images of the murine fundus, obtained with the infrared (IR) laser (830 nm), were taken immediately before the FA images. Images taken in the FA mode were made semitransparent (75% transparency) using graphic design and illustration software (Corel Draw Graphics Suite X3; Corel Corporation, Ottawa, Canada) and were overlaid on the corresponding IR images. The structure of the vessels and the optic nerve were used to align the images, thus allowing for spatial localization of viral foci. At least nine mice per time point were analyzed. 

Results

In Vivo Imaging of MCMV Infection in the Posterior and Anterior Segments

To assess whether excitation of eGFP (l_max = 508 − 515 nm) with the 488-nm argon laser provides a sufficient signal for detection using the scanning laser ophthalmoscope, we analyzed adult BALB/c mice 4 days after intraocular injection of 10³ pfu MCMV-eGFP. To test for background autofluorescence, at wavelengths that overlap those of GFP, a group of matching BALB/c mice was injected with PBS or 10³ pfu MCMV-wt. Reproducible signal could only be detected from retinas of mice infected with MCMV-eGFP, but not from MCMV-wt or PBS-injected mice. Analysis of retinal vessels in the infrared mode and fluorescein angiogram showed tortuosity and marked dilatation of the retinal veins (Fig. 1, upper), whereas clinical fundus examination revealed a constriction of retinal arteries after infection with MCMV-wt. This was seen only in eyes infected with virus; vessels of control eyes (PBS only) appeared normal (Fig. 1, lower). 

Figure 1.

 

HRA: BALB/c mice were infected subretinally with 10³ pfu MCMV-K181. Upper: MCMV: infected BALB/c mouse retina at day 4 after infection. Lower: CTRL: PBS-injected BALB/c mouse retina (n > 3). No leakage was observed in fluorescein angiography after the injection of MCMV. FA, fluorescein mode; IR, infrared mode.

To assess ocular MCMV pathogenesis in vivo, eyes were monitored at regular intervals after intraocular infection with MCMV-eGFP. No signal could be detected by HRA in the first 24 hours after infection (Fig. 2, left). Fluorescent spots appeared 2 days after infection with MCMV-eGFP, presenting as evenly distributed foci on the retinas of infected mice (Fig. 2). The strength of the GFP signal reached a maximum at day 6 after infection, when a consistent accumulation of signal around the posterior pole and around the optic nerve was particularly evident (Fig. 2, day 6). Up to 60% of infected eyes developed cataract 10 days after infection. This diminished image quality in the posterior segment and made visualization of the retina unamenable to HRA in 30% of infected eyes at day 25. In contrast, monitoring of the anterior segment was not problematic. SLO revealed GFP signal from day 4 after infection within the iris stroma, and weaker signal was detectable around the anterior curvature of the lens (Fig. 3, HRA). Signal could be identified up to day 25 after infection and persisted mainly in the iris and, to a lesser extent, on the anterior curvature of the lens (Fig. 3, HRA). Comparison of SLO images with lens wholemount immunofluorescence (Fig. 3, IF) revealed localization of GFP-positive cells on the anterior lens curvature. In situ analysis with high magnification showed GFP-positive cells on the lens capsule (not shown). 

Figure 2.

 

HRA analysis of infected mouse retinas. BALB/c mice were infected subretinally with 10³ pfu MCMV-eGFP, and the time course of viral infection was monitored (>8 mice per time point). Images taken in the fluorescein mode (FA, upper) were merged with infrared images for anatomic detail (FA+IR, lower). At 12 hours after infection, only background noise was recorded in the FA mode. FA signal was detected at all other time points, as shown.

Figure 3.

 

Time-course analysis of MCMV infection of the anterior segment of the eye. BALB/c mice were infected subretinally with 10³ pfu MCMV-eGFP, and the time course of viral infection was monitored. HRA of the anterior segments of infected eyes is shown. Data are representative of results obtained from eight BALB/c mice at each time point. Immunohistochemistry (IF) of lens wholemounts stained with CD45 (red). Fluorescence was detected 3 days after infection. Representative images from two independent experiments are shown.

As shown in Figure 4A, the percentage of positive signal in the retinas of infected mice, as measured by HRA, correlated well with the percentage of positive signal observed in wholemount preparation of retinas from infected mice. 

Figure 4.

 

(A) Percentage of retinas with positive signal in HRA and in wholemount preparations (IF). Data are the mean ± SE for the following numbers of BALB/c mice in HRA (IF): 12 hours, n = 15 (8); day 2, n = 14 (8); day 4, n = 14 (12); day 6, n = 14 (12); day 8, n = 10 (8); day 12, n = 10 (8); day 25, n = 12 (8). (B, left) Flatmounts of MCMV-eGFP–infected optic nerve head and iris at day 4. BALB/c mice were infected subretinally with 10³ pfu MCMV-eGFP. Immunofluorescence wholemount of optic disc (top) stained for CD11b (×20), and iris (bottom) stained with concanavalin A and MHC class II (×40) are shown. Right: colocalization of GFP and CD45⁺ signals at day 6 after infection (×100).

Histologic Analysis of Retinal MCMV Infection

Intraocular infection with 5 × 10² MCMV-Smith in immunocompetent mice precipitates only very mild inflammatory ocular disease. ²¹ To assess whether the same is true for MCMV-eGFP, sequential hematoxylin eosin-stained sections of eyes of infected mice were analyzed. These revealed inflammatory cells in the vitreous and within the retina 2 days after infection with MCMV, but not in PBS-treated mice (data not shown). Fibrinous exudates, posterior synechiae, and distortions of the iris and ciliary body were present after day 4 (data not shown). At day 12 after infection, the number of inflammatory cells in the vitreous and anterior chamber decreased. No structural changes in the anterior or posterior segments were observed when mice were infected with 10³ pfu MCMV-eGFP. These findings were similar to the previous reports of MCMV infection in immunocompetent BALB/c mice. ²¹  

Confocal imaging of retinal wholemounts stained with anti-leukocyte antibodies revealed some GFP-positive cells were CD45⁺ (Fig. 4B, right), but were negative for CD11b and major histocompatibility complex (MHC) class II, suggesting that infected cells were not of the monocytic/macrophage lineage (Fig. 4B, left). Eyes with positive signal in HRA were positive in wholemount histology though the area of signal was usually smaller in the wholemount than in HRA. We suspect that this was because of the missing RPE in wholemount preparations. In some cases signal was seen in wholemount but not in HRA, especially at later time points, probably because of cataract formation. At later time points very little signal was seen in retinas; in some cases only two cells were infected within a retinal specimen. 

Accumulation of signal around the optic disc, as seen by SLO, could also be observed in wholemount preparations (Fig. 4B, top left). To assess whether infection was able to reach the optic nerve, we analyzed the optic nerve head for a GFP signal. One nerve head (1 of 6 mice) showed minimal signs of GFP at day 4, but no signal was found in the optic nerve head or along the optic nerve up to day 25 after infection (0 of 24 mice). Furthermore, no replicating virus was detected in the brain for all animals tested, suggesting that MCMV-eGFP does not spread via the optic nerve. 

Photoreceptors and RPE as Targets of MCMV in the Immunocompetent Murine Retina

Previous investigations have demonstrated that, in immunosuppressed BALB/c mice, cells (e.g., bipolar cells, horizontal cells) of the retinal pigment epithelium, glia, and retinal neurons are targets of MCMV infection. ^22,23 To assess whether these cells were also targeted in MCMV-eGFP–infected immunocompetent BALB/c mice and to confirm the SLO findings, we performed in situ analysis of infected retinal specimens on day 6, when the GFP signal on SLO examination was strongest (Fig. 2). Cryosections of retinas at days 4 and 6 days after infection showed GFP signal accumulating within the outer retinal layers, predominantly within the photoreceptor outer segments and the retinal pigment epithelium (Figs. 5A–D). These findings suggest that at the peak of primary subretinal MCMV infection, photoreceptor outer segments and retinal pigment epithelium are targets of MCMV-eGFP. No costaining with glial markers was noted. Immunofluorescence analysis of iris and ciliary body at days 4 and 6, showed no colocalization of GFP signal with CD11b or MHC class II (Figs. 5E, 5F). 

Figure 5.

 

Photomicrographs of MCMV-eGFP–infected retinas. (A) DAPI stain. OS, outer segment; ONL, outer nuclear layer; OPL, outer plexiform layer. (B) GFP signal. (C) NeuN stain and (D) merged images from 6 days after infection with 10³ pfu MCMV-eGFP into the subretinal space (arrow). Scale bars, ×20. Merged images: most of the fluorescent signal accumulates around the outer segments of the photoreceptors. Photomicrographs (×40) of CD45-stained retina wholemount at days 6 (E) and 25 (F) after infection.

Discussion

In this study we show that replication, spread, and pathogenesis of primary MCMV infection in the eye can be visualized with a scanning laser ophthalmoscope. This technique has several advantages; first, it is an excellent tool for in vivo monitoring of viral spread and allows multiple assessments over time. In this model, the presence of GFP signal equates to productive MCMV infection and thus can be used as a tool to monitor ocular infection with respect to replication and spread. Although conventional viral titration of tissue homogenates represents a temporal snapshot and usually an end point of an experiment, in vivo imaging allows one to monitor viral infection sequentially as a more complex and dynamic process. Second, the ability to simultaneously observe MCMV infection in individual ocular compartments provides insights into viral tropism within the eye. In addition, the capacity to repeatedly measure viral GFP expression in vivo will provide new avenues for pathogenesis studies and studies of MCMV latency and reactivation. In this context, it is still a matter of controversy whether virus detected in the eye of immunosuppressed mice results from virus reactivation in situ in the eye or whether it results from the spread of replicating virus from nonocular sites. CMV has been shown to establish latency in cells of the monocyte/macrophage lineage both in humans and mice. ^24–26 However, CMV might establish latency in specific cell subsets in the eye, and in vivo imaging may serve as tool to monitor viral reactivation. 

Our present study revealed that the immunocompetent eye is readily permissive to primary MCMV infection. The main targets of infection were outer segments of photoreceptors and RPE cells, and there was no spreading to bipolar or horizontal cells, though the latter has been reported in immunosuppressed animals. ²² Of note, although fluorescent virus was found to accumulate around the optic nerve head, no virus was isolated from either optic nerves or brain, suggesting that the optic nerve head represents a barrier to viral spread in immunocompetent hosts. This finding is substantiated by a recent report of HCMV-induced meningoencephalitis and optic neuritis after HCMV retinitis in a patient with AIDS. ²⁷ In addition, after subretinal infection, virus readily spreads from the posterior compartment to the iris, trabecular meshwork, and lens, suggesting that these tissues represent a target for MCMV infection. Whether this represents an active process of infected cells traveling to the anterior compartment, or whether MCMV gains access to the iris or trabecular meshwork in a soluble form through the aqueous, remains to be investigated. However, our data suggest that the mechanism of viral spread in primary MCMV infection involves free virus trafficking to the anterior compartment of the eye because viral fluorescence was not localized to trafficking cells, such as macrophages. 

At present there are a few limitations when monitoring MCMV-eGFP in vivo. First, it has to be noted that viral titers cannot be inferred from the signal intensity seen in SLO. In addition, although SLO allows excellent observation of spatial distribution of virus, it does not provide depth localization within tissues. It must be kept in mind that direct inoculation of MCMV into the subretinal space does not reflect the route of natural viral infection. In natural systemic infection, cells other than photoreceptors and RPE cells may be infected. 

Investigation of MCMV in the immunocompetent murine eye revealed that virus could be tracked even when viral titers could not be detected by standard plaque assay, suggesting that the threshold for detection by SLO is lower than the threshold for detection of virus by plaque assay, which normally lies around 40 pfu. Although no virus was detected by plaque assay at day 25 after infection, some signal could still be seen in SLO on the anterior curvature of the lens and within the iris. In addition, sporadic GFP signal was seen within the retina at day 25 after infection. Thus, in vivo retinal imaging using SLO is a valuable method for detecting small numbers of infected cells within the retina and, as such, provides valuable insight into the dynamics of MCMV-induced ocular infection over long periods of time. Because of the long-term sensitivity of SLO analysis, this technique may aid our understanding of the pathogenesis of CMV reactivation in the eye. Ultimately, it is possible that tracking ocular pathogens with pathogen-specific fluorescence-labeled antibodies may provide a novel tool for diagnosing infectious eye diseases in humans. 

Footnotes

 Supported by Swiss National Science Foundation Grant PBZH33-121040; Stiftung für Medizinische Forschung und Entwicklung, Switzerland; and the Lions Eye Institute of Western Australia.

Footnotes

 Disclosure: M.S. Zinkernagel, None; C. Petitjean, None; P. Fleming, None; H.R. Chinnery, None; I.J. Constable, None; P.G. McMenamin, None; M.A. Degli-Esposti, None