Download Acute Inflammation and Loss of Retinal Architecture and Function

Current Eye Research, 31:955–965, 2006
c Informa Healthcare
Copyright ISSN: 0271-3683 print / 1460-2202 online
DOI: 10.1080/02713680600976925
Curr Eye Res Downloaded from informahealthcare.com by Library Schusterman Center on 11/04/11
For personal use only.
Acute Inflammation and Loss of Retinal
Architecture and Function During
Experimental Bacillus Endophthalmitis
Raniyah T. Ramadan
Oklahoma Center for
Neuroscience, University of
Oklahoma Health Sciences
Center, Oklahoma City,
Oklahoma, USA
Raul Ramirez
and Billy D. Novosad
Department of Ophthalmology,
University of Oklahoma Health
Sciences Center, Oklahoma
City, Oklahoma, USA
Michelle C. Callegan
Department of Microbiology
and Immunology, University of
Oklahoma Health Sciences
Center, Oklahoma City,
Oklahoma and Molecular
Pathogenesis of Eye Infections
Research Center, Dean A.
McGee Eye Institute, Oklahoma
City, Oklahoma, USA
ABSTRACT Rapid vision loss and explosive inflammation are devastating consequences of Bacillus endophthalmitis that have not been well defined. We
therefore analyzed the evolution of intraocular inflammation and loss of retinal
architecture and function during experimental Bacillus endophthalmitis. Mice
were intravitreally injected with 100 CFU of B. cereus, and eyes were analyzed
for bacterial growth, retinal function, architectural changes and retinal cellular
stress, inflammatory cytokines, and infiltrating cells. Retinal electrophysiologic
and structural changes began as early as 4 to 6 hr postinfection. Significant declines in retinal function paralleled the loss of retinal architecture. Glial fibrillary
acidic protein (GFAP) was detected in retina, indicating potential stress. Polymorphonuclear leukocyte (PMN) infiltration into the vitreous began as early as
4 hr postinfection, coinciding with a significant increase in TNF-α in the eye.
These results indicated that acute inflammation and detrimental architectural
and electrophysiologic changes in the retina began earlier than once thought,
suggesting that therapeutic intervention should be given at the earliest possible
time to avoid vision loss during Bacillus endophthalmitis.
KEYWORDS Bacillus; endophthalmitis; inflammation; mouse; retina
INTRODUCTION
Received 7 December 2005
Accepted 21 August 2006
Correspondence: Michelle C. Callegan,
Department of Ophthalmology DMEI
419, 608 S.L. Young Blvd., Oklahoma
City, OK 73104, USA. E-mail:
[email protected]
Endophthalmitis is a potentially blinding ocular infection resulting from the
introduction of microorganisms in the posterior segment of the eye. Infectious
agents generally gain access to the interior of the eye as a consequence of intraocular surgery (postoperative), a penetrating injury to the globe (post-traumatic), or
from metastatic spread of bacteria into the eye from a distant anatomical site (endogenous). The severity of bacterial endophthalmitis can range from relatively
mild posterior segment inflammation caused by normal flora to a refractory,
sight-threatening infection caused by virulent pathogens such as Bacillus cereus.
Aggressive and prompt therapy is especially critical for B. cereus and other
forms of severe endophthalmitis because the progression from the initial seeding of the pathogen to a severe, deep-seated infection and loss of vision
can occur in a relatively short period of time. Effective therapy for bacterial
955
Curr Eye Res Downloaded from informahealthcare.com by Library Schusterman Center on 11/04/11
For personal use only.
endophthalmitis should be immediate and precise to
kill organisms and limit widespread ocular injury. Studies using experimental endophthalmitis models have
shown that antibiotics effectively kill intraocular organisms, and anti-inflammatory agents are occasionally successful in suppressing the intraocular inflammatory response.1–6 In similar experimental models,
some of the important toxins involved in the virulence of endophthalmitis pathogens have also been
identified.7–19 For Bacillus specifically, toxins and bacterial migration contribute to intraocular virulence during
endophthalmitis.14 However, current therapeutic regimens are ineffective against bacterial toxins and are not
designed to arrest bacterial migration in the eye. Consequently, no universal regimen exists that has proved
useful in preserving vision or the eye itself during the
most severe forms of bacterial endophthalmitis, including that caused by B. cereus.
Limited information exists regarding the mechanisms involved in retinal function loss or the significance of the host response during bacterial endophthalmitis. Therefore, purpose of this study was to analyze retinal architectural and function loss and intraocular inflammation during experimental Bacillus endophthalmitis. A detailed understanding of the molecular changes in retinal physiology leading to function
loss and the triggering of the host inflammatory response will advance current knowledge of the interactions between pathogen and host during bacterial endophthalmitis and will provide information for the production of host-targeted therapeutics to prevent vision
loss during Bacillus endophthalmitis.
MATERIALS AND METHODS
Bacteria and Growth Conditions
B. cereus strain ATCC 14579 (American Type Culture
Collection, Manassas, VA, USA) was used in this study.
This strain has been used previously to initiate endophthalmitis in an experimental rabbit model, causing explosive infection and inflammation similar to that seen
in human cases.13 B. cereus was cultured overnight in
brain heart infusion media (BHI; Difco Laboratories,
Detroit, MI, USA) at 37◦ C with aeration. Overnight
cultures were subcultured into BHI, grown to logarithmic phase under the identical conditions, and serially
diluted in BHI to approximately 100-colony-forming
R. T. Ramadan et al.
units (CFU)/0.5 µl for intravitreal injection, as described below.
Experimental Murine Bacillus
Endophthalmitis
Male C57BL/6J mice (5–6 weeks of age; Jackson Laboratories, Bar Harbor, ME, USA) were used in these
studies. All animals were maintained according to institutional guidelines and the ARVO Statement for the
Use of Animals in Ophthalmic and Vision Research.
Mice were anesthetized with a mixture of ketamine
(85 mg/kg of body weight KetaVed [Phoenix Scientific,
Inc., St. Joseph, MO, USA]), and xylazine (14 mg/kg of
body weight, Rompun [Bayer Corp., Shawnee Mission,
KS, USA]). Topical anesthetic (0.5% proparacaine HCl
[Ophthetic Allergan, Hormigueros, Puerto Rico]) was
instilled in each eye before injection.
Intravitreal injections were performed with sterile borosilicate glass micropipettes (Kimble Glass Inc,
Vineland, NJ, USA) beveled to an approximate bore
size of 10 to 20 µm (BV-10 KT Brown Type micropipette beveller, Sutter Instrument Co., Novato, CA,
USA). Under stereomicroscopic visualization, the micropipettes were inserted just posterior to the superior
limbus, and 0.5-µl volumes were injected directly into
the midvitreous.20 Injection rates and volumes were
monitored using a programmable cell microinjector
(Microdata Instruments, Plainfield, NJ, USA). To ensure reproducibility of injections, 0.5 µl of fluorescein
(Fluress, Pilkington Barnes Hind, Ontario, Canada) was
injected into eyes to monitor leakage and ensure placement into the midvitreous. For the infection model,
0.5 µl of BHI with or without B. cereus was injected.
Contralateral eyes were injected with BHI (vehicle control), phosphate-buffered saline (PBS; surgical control),
or were left undisturbed (absolute control). Eyes were
analyzed throughout the course of infection by biomicroscopy, electroretinography, immunohistochemical
and histologic analysis, and whole eye quantification
of bacterial growth, inflammatory cell influx, and inflammatory cytokines, as described below.
Clinical Scoring
The clinical changes occurring during experimental
Bacillus endophthalmitis were scored independently by
masked observers with the aid of a slit-lamp biomicroscope (Topcon SL-D2, Paramus, NJ, USA). Clinical
observations included anterior segment inflammation,
956
TABLE 1 Clinical and Histological Scoring of Experimental Murine Bacillus Endophthalmitis
Score
0
No flare, no cells/field, no
iritis
Slight flare, 1–5 cells/field,
iris vessels slightly dilated
Obvious flare, 10–20
cells/field, mild dilation of
iris vessels
Moderate flare, 20–50
cells/field, significant
dilation of iris vessels
Intense flare, >50 cells/field,
severe iritis
1+
2+
3+
4+
Curr Eye Res Downloaded from informahealthcare.com by Library Schusterman Center on 11/04/11
For personal use only.
Anterior chamber
Red reflex
Red
Vitreous
Clinical scoring
Clear
Retinal clarity
Details distinct
Slightly diminished
Slight haze
Details slightly obscured by infiltrate
Mildly diminished
Mild haze
Details mildly obscured by infiltrate
Moderately diminished Dense haze
No reflex (i.e., white)
Details moderately obscured by infiltrate
Opaque vitreous Details completely obscured by infiltrate
Histologic scoring
0
1+
2+
3+
4+
Inflammation (posterior and anterior segments)
Retinal architecture
No infiltrating inflammatory cells
1–10 infiltrating cells per field, no fibrin
10–50 infiltrating cells per field, mild fibrin reaction
50–100 infiltrating cells per field, significant fibrin reaction
Both segments filled with fibrin and infiltrating cells
Retina completely intact
Retinal folds in <25% of the retina
Retinal folds/detachment in 25–50% of the retina
Retinal folds/detachment in 50–75% of the retina
Complete detachment, retinal layers indistinguishable
presence/absence of red reflex, vitreous inflammation,
and retinal clarity. Clinical changes were graded on a
scale from 0 to 4+ based on the criteria outlined in
Table 1. Photographs of mouse eyes were taken for visualization of progression of disease severity.
Quantification of Viable Bacteria
To quantitate B. cereus, globes were enucleated, placed
into 400 µl of sterile PBS, and homogenized using
sterile glass beads (1.0 mm; Biospec Products, Inc.,
Bartlesville, OK, USA). Bacteria in whole eyes were
quantified by track plating serial 10-fold dilutions onto
BHI agar.21
sphere that mimicked a Ganzfeld. The interval between
two flashes (10-ms duration) was 60 s to prevent light
adaptation to the flashes. As an indicator of retinal function, the A-wave and B-wave amplitudes and implicit
times were measured from the initiation of the light
flash to the trough of the A-wave, and the trough of the
A-wave to the peak of the B-wave, respectively.9−11,13,14
A total of five readings were recorded and averaged.
Percentages of retinal function retained compared with
controls were calculated as follows: 100 − {[1 − (experimental A or B-wave amplitude/control A or B-wave amplitude)] × 100}. The latency of retinal responses due
to infection corresponding with percentages of implicit
time (τ ) were calculated as follows: {[1 − (experimental
τ /controlτ )] × 100}.
Retinal Function Analysis
Retinal function was analyzed by electroretinography (ERG) using a protocol similar to that of Ranchon
et al.22 After dark-adaptation for at least 6 hr, mice were
anesthetized as described above and pupils were dilated
with topical phenylephrine (10%; Akorn, Inc., Buffalo
Grove, IL, USA). Retinal function was recorded with
gold-wire electrodes placed on each cornea and an oral
reference electrode. The white light stimulus used to
evoke the retinal response was delivered by a white
957
Thin-Section Histology
Globes harvested for histological analysis were fixed
in 10% formalin for 24 hr. The eyes were sectioned and
stained with hematoxylin and eosin or Giemsa by standard procedures.23 Histology sections were scored independently by masked observers based on the degree of
change in anterior and posterior segment inflammation
and retinal architecture (Table 1).
Experimental Bacillus Endophthalmitis
Curr Eye Res Downloaded from informahealthcare.com by Library Schusterman Center on 11/04/11
For personal use only.
Immunohistochemistry of Retinal
Whole-Mounts
Globes were harvested and fixed in 4% paraformaldehyde for 15 min. After fixation, the cornea, lens, and vitreous were removed. After overnight incubation of the
eye cup (choroid and retina) in fresh 4% paraformaldehyde, the choroid was carefully removed from the
retina, and the retina was washed with PBS/1% Triton
X-100. Retinas were stained primarily with mouse antiglial fibrillary acidic protein antibody (GFAP; 1:500,
Dako, Carpenteria, CA, USA) and with an endothelial
cell marker for the staining of retinal vasculature (Biotinylated Bandeiraera [Griffonia], Simplicifolia Lectin
I [Isolectin B4] 1:100, Vector Laboratories, Burlingame,
CA, USA). Whole-mounts were then stained with secondary goat anti-rat Alexa Fluor 594 and Streptavidin
Alexa Fluor 488 conjugate (1:200; Molecular Probes,
Eugene, OR, USA). Retinas were then mounted and
viewed by fluorescent microscopy using 594-nm and
488-nm filters for GFAP and lectin visualization, respectively.
Analysis of Intraocular Inflammation
Inflammatory cytokines in whole eyes were quantified. Globes were homogenized with glass beads in a
protease inhibitor cocktail (Triton X-100, 0.5 M EDTA,
10 mM sodium orthovanadate [Sigma, St. Louis, MO,
USA] and Protease Inhibitor [Calbiochem, La Jolla,
CA, USA] in PBS, pH 7.4). Supernatants were then analyzed for TNF-α, interleukin-6 (IL-6), and interleukin10 (IL-10, as a negative control cytokine) by ELISA
(Quantikine ELISA kits, R&D Systems Minneapolis,
MN, USA).
To detect infiltrating polymorphonuclear leukocytes
(PMN), myeloperoxidase (MPO) activites were assayed
in whole eyes. Globes were homogenized with glass
beads in a lysis buffer (200 mM NaCl, 5 mM EDTA,
10 mM Tris, 10% glycine [vol/vol], 1 mM PMSF, 1
µg/ml leupeptide, 28 µg/ml aprotinine). Supernatants
were then analyzed for MPO by ELISA (Mouse MPO
ELISA Test Kit, Cell Sciences, Canton, MA, USA).
Infiltrating inflammatory cells in whole eyes were
quantified and typed by flow cytometry. Globes were
removed, and the cornea, iris, ciliary body, and lens dissected from the posterior segment. The remaining eye
cups were incubated in collagenase type I (3.0 mg/ml;
Sigma) for 90 min at 37◦ C with trituration every 30
R. T. Ramadan et al.
min. Single cell suspensions were filtered using a 40µm filter (Falcon, BD Biosciences, San Jose, CA, USA)
and flushed with excess DMEM + 10% fetal bovine
serum (FBS). Following a wash in DMEM + 10% FBS,
the cells were counted, centrifuged, and resuspended at
a final concentration of 1 × 106 in 1% bovine serum
albumin (BSA)/PBS. Cells were then incubated with
BD mouse Fc block (BD Biosciences, San Diego, CA,
USA) for 20 min at 4◦ C. Cells were washed once and
resuspended in FITC- or PE-conjugated antibody to
identify PMN (CD18 and Gr-1), macrophages (F4/80
and CD11b), or dendritic cells (CD11c and I-Ab ). All
antibodies were purchased from BD Biosciences except F4/80 (Caltag, Burlingame, CA, USA) and were
used at concentrations suggested by the manufacturer.
After 30-min incubation at 4◦ C, cells were washed
twice and resuspended in 1% paraformaldehyde and
were immediately analyzed by flow cytometry. Surgical
and absolute controls were included as infection controls. Positive cell controls included PMN isolated from
whole blood, macrophages/monocytes isolated from
peritoneum, and dendritic cells isolated from bone
marrow. Isotype control antibodies were also included.
Cells were analyzed with an EPICS-XL flow cytometer (FACSCalibur, San Jose, CA, USA) and quantified
using Summit (Dako Cytomation, Fort Collins, CO,
USA).
Statistical Analysis
All values represent the mean ± standard error of the
mean (SEM) for n ≥ 5 eyes per time point assayed unless
otherwise specified. Wilcoxon’s rank sum test was used
for statistical comparison between infection groups. A
p value ≤ 0.05 was considered significant.
RESULTS
Experimental Murine Model
of Endophthalmitis
Three different inocula were tested initially to determine which would result in an intraocular infection
whose symptoms mimicked that of the experimental
rabbit Bacillus endophthalmitis model and the clinical
human infection. Inocula of 20 ± 12 CFU, 120 ± 10
CFU, or 1800 ± 150 CFU were injected per eye (n ≥
6 eyes per group). Figure 1A illustrates the number of
bacteria per eye after infection with different inocula.
At 12 hr postinfection, eyes injected with the lowest
958
Curr Eye Res Downloaded from informahealthcare.com by Library Schusterman Center on 11/04/11
For personal use only.
FIGURE 1 Quantitation of B. cereus growth in the mouse eye.
Inoculum optimization (A) and growth of optimum inoculum (102
CFU/0.5 µl) (B) in whole mouse eyes are shown. Eyes injected
with 102 or 103 CFU resulted in intraocular growth that reached a
maximum of 109 CFU/eye by 12 hr. Eyes injected with the lowest
inoculum were sterile at 12 hr postinfection (A). Therefore, the
optimum inoculum chosen to use for the infection course in this
study was 102 CFU (B). Values represented the mean ± SEM for
n ≥ 6 eyes per group.
inoculum were sterile and had no detectable inflammation. These eyes were similar in appearance to absolute
and surgical controls, which also had no detectable inflammation at 12 hr (i.e., a cumulative clinical score
of 0).
In contrast, eyes injected with 102 or 103 CFU resulted in rapid intraocular growth, reaching a maximum
109 CFU/eye by 12 hr. At this time, eyes injected with
102 bacilli presented with posterior and anterior segment inflammation and other signs of infection similar
to that of the experimental rabbit model at the same
time point (described below).9−14 At 12 hr postinfection, eyes injected with 103 bacilli presented with severe inflammation and signs of infection similar to that
observed in the experimental rabbit model at 18–20 hr
postinfection.9−14 Because the 103 inoculum caused a
more explosive and rapid infection than what was previously observed in the rabbit model, we chose the 102
CFU inoculum for the remainder of this study.
In eyes injected with the 102 CFU inoculum, bacilli
grew logarithmically to a maximum concentration of
approximately 109 CFU/eye by 12 hr (Fig. 1B). Intraocular bacterial numbers remained constant throughout
20 hr postinfection. In these eyes, mild (2+) vitreal inflammation was noted as early as 6 hr postinfection. Few
959
FIGURE 2 Clinical progression of experimental murine B.
cereus endophthalmitis. (A) Clinical photographs of the course of
endophthalmitis document evolving infection and inflammation.
(B) Signs of evolving infection were scored on a scale of 0 to 4+
based on the scoring system described in Table 1. Individual and
cumulative scores are presented. Values represented the mean ±
SEM for n ≥ 4 eyes per group.
inflammatory cells were noted in the anterior chamber
(2+), and red reflexes were present in most eyes (score
of 0). By 10–12 hr postinfection, moderate to severe inflammation was observed. Corneal haze was noted and
numerous inflammatory cells and fibrin were seen in
the anterior chamber (clinical scores ranging from 3+
to 4+). In most eyes, vitreous haze was present (3+),
and red reflexes were absent (4+). By 20 hr postinfection, the conjunctiva was highly inflamed and corneal
edema and corneal ring abscesses were present in most
infected eyes. Vitreous opacities were present (4+) and
red reflexes were absent (4+) in all eyes at this time.
Representative photographs of mouse eyes and individual/cumulative clinical scores of evolving experimental
Bacillus endophthalmitis are shown in Figure 2. The
rapid pathologic changes observed in the experimental mouse model were similar to that observed in other
experimental models of Bacillus endopthalmitis.9−14
Retinal Architecture
Whole eye and retinal histology of experimental
murine endophthalmitis is shown in Figure 3. At 2 and
4 hr postinfection, few inflammatory cells were seen
in the posterior chamber, the retinal layers were intact,
Experimental Bacillus Endophthalmitis
Curr Eye Res Downloaded from informahealthcare.com by Library Schusterman Center on 11/04/11
For personal use only.
FIGURE 3 Representative whole organ (A) and retinal (B) histology of eyes during experimental B. cereus endophthalmitis. All sections
were stained with hematoxylin and eosin. Severe inflammation and retinal detachments were observed by 8–10 h postinfection. By 12 h
postinfection, retinal layers were virtually indistinguishable. ILM, inner limiting membrane; GCL, ganglion cell layer; IPL, inner plexiform
layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; OLM, outer limiting membrane; PCL, photoreceptor
layer; RPE, retinal pigment epithelium. Magnifications:, 10× (A) and 200× (B).
and very little posterior segment pathology was seen
(i.e., histologic scores ranging from 0 to 1+ for both
anterior and posterior segments). These findings were
similar to that seen with the absolute and surgical controls.
After 6 hr, the majority of infiltrating cells observed
were located in close proximity to the optic nerve in
the posterior chamber. Morphologically, these cells appeared to be PMN (Fig. 4). The numbers of inflammatory cells in the posterior chamber were greater at 8
hr postinfection (histology score of 2+), and the retinal architecture was slightly disrupted (histology score
of 1+). After 8 hr, significant loss of retinal architecture
and significant numbers of inflammatory cells were seen
in the posterior segment, near the ciliary body, and in
the anterior chamber (histology scores ranging from 2+
to 3+). By 20 hr, most eyes were completely destroyed
(histology scores of 4+).
Retinal Function
Electroretinography of mouse eyes during Bacillus
endophthalmitis is summarized in Figure 5. Retinal
function did not decline throughout 12 hr in absolute or surgical controls. In infected eyes, decreases
in both A-wave and B-wave amplitudes were noted at
6 hr postinfection, but these declines were not signifiR. T. Ramadan et al.
cantly different from that of surgical or absolute controls (p ≥ 0.05). By 8 hr postinfection, A-wave amplitudes declined to 33.80 ± 19.49%, while B-wave amplitudes were reduced to 30.40 ± 6.92% compared with
that of the controls (p ≤ 0.0001). By 12 hr postinfection, both A-wave and B-wave amplitudes were reduced
to less than 2% of function compared with controls
(p ≤ 0.0001). In this model, the minimal dark adaptation time was 6 hr, and retinal function was not assessed
at 2 hr or 4 hr postinfection. Nevertheless, loss of retinal
function occurred at a rate similar to that seen in the rabbit endophthalmitis model.9−14 A slight but significant
increase in B-wave implicit time occurred at 6 hr postinfection (p = 0.034). Significant increases in A-wave implicit times did not occur until 8 hr postinfection
(p = 0.014).
Retinal Stress
Immunohistochemistry of retinal whole-mounts during experimental Bacillus endophthalmitis is illustrated
in Figure 6. Retinal whole-mounts were stained with
anti-GFAP to indicate generalized areas of retinal stress
and with lectin to stain the retinal vasculature. At
8 hr postinfection, GFAP immunostaining of infected
eyes appeared similar to that of surgical and absolute
controls. Lectin staining of retinal whole-mounts revealed cellular occlusion of blood vessels and leakage of
960
Curr Eye Res Downloaded from informahealthcare.com by Library Schusterman Center on 11/04/11
For personal use only.
FIGURE 5 Retinal function analysis during experimental B.
cereus endophthalmitis. A-wave and B-wave amplitudes (left) decreased significantly by 8 hr compared with controls. At 12 hr
postinfection, amplitudes were <2% or were undetectable. ERG
analysis of latencies in retinal function (right) showed increases in
the implicit times of the B-wave at 6 hr and the A-wave at 8 hr, compared with controls. Retinal function did not change after shaminjections, as indicated by the . Values represented the mean ±
SEM for n ≥ 5 eyes per group.
leakage of cells into the vitreous was also significant at
this time.
Intraocular Inflammation
Increase in MPO Paralleled an Increase
in TNF-α
FIGURE 4 PMNs in the posterior segment during experimental
B. cereus endophthalmitis. PMNs were identified morphologically
using the Giemsa stain. Higher magnification of PMNs (arrowheads) demonstrated PMNs in close proximity to the optic nerve,
the initial site of PMN influx during the earlier stages of infection.
RET, retina; VIT, vitreous; ON, optic nerve. Magnifications: 100×
(A), 400× (B), and 1000× (C).
cells from these vessels at this time. Slight increases in
GFAP immunostaining were observed at 8 hr. At 16 hr
postinfection, a significant increase in GFAP immunostaining was detected. Occlusion of retinal vessels and
961
Analysis of inflammatory cytokines and MPO in
whole eyes during experimental Bacillus endophthalmitis are summarized in Figure 7. TNF-α concentrations
were detected at 2 and 4 hr postinfection. Thereafter,
TNF-α increased significantly, reaching a peak of 49.05
± 4.9 pg/eye at 10 hr postinfection. IL-6 and IL-10 were
below the limit of detection in whole-eye homogenates
throughout the course of infection. The concentration
of MPO in whole eyes increased significantly after
4 hr postinfection, reaching a peak of approximately
280 ng/eye at 10 hr postinfection. For sham injections,
cytokine and MPO levels were not significantly different from that of uninjected control eyes.
PMN Infiltration
Flow cytometric analysis of infiltrating inflammatory
cells during experimental Bacillus endophthalmitis is
summarized in Figure 8. CD18+/Gr-1+ cells were detected in infected eyes as early as 4 hr postinfection. The
Experimental Bacillus Endophthalmitis
Curr Eye Res Downloaded from informahealthcare.com by Library Schusterman Center on 11/04/11
For personal use only.
FIGURE 6 GFAP and isolectin B4 immunostaining of retinal whole-mounts during experimental B. cereus endophthalmitis. Retinal
whole-mounts from mouse eyes infected with B. cereus were stained with anti-GFAP (red) as a marker of retinal stress and Isolectin B4
(green) to visualize the retinal vasculature. Slight GFAP immunostaining was detected at 8 hr postinfection, compared with the controls.
A significant increase in GFAP expression was seen at 16 hr postinfection, indicating a possible stress-induced response in the retina
during infection.
numbers of CD18+/Gr-1+ cells increased significantly
throughout the course of infection, correlating directly
with increases in MPO. Low numbers of CD18+/Gr1+ cells were detected in surgical or absolute control
eyes, but these numbers did not increase during infection. Less than 5% of monocytes and macrophages were
detected in infected eyes throughout the course of infection, but these numbers did not increase during infection. Dendritic cells were not detected at any time in
infected eyes or controls.
DISCUSSION
FIGURE 7 TNF-α and MPO kinetics in whole eyes during experimental B. cereus endophthalmitis. Whole eyes were homogenized and supernatants analyzed for TNF-α or MPO by ELISA. Absolute and surgical controls were included. TNF-α concentrations
were detected at 2 and 4 hr postinfection and increased significantly after 6 hr postinfection. MPO increased significantly after
4 hr postinfection. Neither TNF-α nor MPO were detected in shaminjected controls, as indicated. Values represent the mean ± SEM
for n ≥ 3 eyes per time point analyzed in triplicate.
R. T. Ramadan et al.
Endophthalmitis caused by infection with Bacillus or
other virulent ocular pathogens can result in explosive
inflammation and significant vision loss in just a few
days, despite prompt and aggressive therapeutic intervention. The mechanisms of retinal function and structural loss and explosive inflammation during B. cereus
and other types of severe bacterial endophthalmitis are
not well defined. To this end, we developed a murine
model of experimental B. cereus endophthalmitis that
962
Curr Eye Res Downloaded from informahealthcare.com by Library Schusterman Center on 11/04/11
For personal use only.
FIGURE 8 Flow cytometric analysis of infiltrating Gr-1+ /CD-
18+ cells during experimental B. cereus endophthalmitis. (A) Dotplot analyses (right) were generated by using a scatter-plot gate
(R2, left) based on whole blood PMN controls. Penel (B) represents
the increase in infiltrating PMNs in the eye during experimental B.
cereus endophthalmitis. Sham-injected and noninjected (control)
values for each time point were averaged. Values represent the
mean ±SEM for n ≥ 5 eyes per time point.
mimicked the human clinical infection and experimental rabbit models9−14 in both speed and severity of infection. In the murine model, the decline in retinal
function paralleled the loss of retinal architecture. At
6 hr postinfection, the minimal disruption in the retinal
layers observed coincided with the initial drop of retinal
function to slightly less than 80%. Significant disruptions in retinal cell layers continued at 8 hr postinfection, leading to retinal detachment and complete loss
of architecture. By 8 hr postinfection, retinal function
963
dropped to between 30% and 35%, and implicit times of
retinal function were significantly increased compared
with controls. These results suggest that the initial loss
of function may be due, in part, to changes in retinal
architecture, an observation that may be instrumental
in the development of therapeutics aimed at preserving
retinal structure.
The mechanisms for loss of retinal structure during
severe endophthalmitis are not yet known. However,
bacterial growth or toxin production near or within the
retina may contribute to the damage by affecting resident retinal cells in the vicinity. One of the cells that
may be initially affected is the Muller cell, whose endfeet in the inner limiting membrane lie in close proximity to the vitreous where Bacillus replicate, migrate,
and produce toxins during infection. The Muller cell
is the major glial cell of the retina, providing structural support and controlling the homeostasis of retina
through neuron protection, regulation of K+ ions, and
neurotransmitter recycling. Stressed Muller cells have
been shown to suffer from osmolarity imbalances and
upregulation of GFAP.24−26 During the later stages of
experimental B. cereus endophthalmitis, retinas stained
positive for GFAP, indicating possible retinal stress. We
are currently investigating whether GFAP is being upregulated from Muller cells or other retinal glial cells,
such as astrocytes. We recently reported that Muller cells
are highly sensitive to B. cereus and its toxins in vitro.27
During intraocular infection, B. cereus and/or its toxins
may negatively affect Muller cells, causing these cells to
become dysfunctional and unable to support the retina,
resulting in loss of structure and function.
Retinal function declined rapidly during experimental murine B. cereus endophthalmitis. Because Muller
cells contribute to the slow PIII component of the
B-wave of the ERG response,27,28 initial dysfunction
of Muller cells due to B. cereus infection or toxicity
may also contribute to retinal function loss regardless
of structural changes. Significant increases in implicit
times of B-wave responses occurring prior to changes
in A-wave responses suggests that initial electrophysiologic changes occur in cells responsible for the B-wave,
including Muller cells. Further, decreases in amplitude
and increases in implicit times of A-waves occurred
shortly after changes in B-waves during infection. The
A-wave component of the ERG is a function of firstorder neuron action potential firing, primarily by photoreceptor cells.27 These results suggested that photoreceptors may have suffered a similar fate in response
Experimental Bacillus Endophthalmitis
Curr Eye Res Downloaded from informahealthcare.com by Library Schusterman Center on 11/04/11
For personal use only.
to Bacillus infection and/or resident toxin production.
These findings may represent the beginnings of a cascade of retinal cell-specific dysfunction in response to
infection. The specific retinal cell types affected and
the underlying mechanisms of retinal cell dysfunction
during endophthalmitis are currently being analyzed.
Explosive inflammation is a hallmark of B. cereus endophthalmitis. Histologic analysis demonstrated that
between 4 and 6 hr postinfection, a small number of
inflammatory cells were seen in the posterior chamber
in close proximity to the optic nerve head.14 Histologic
analysis and MPO and flow cytometry data demonstrated that the primary infiltrating cell type was the
PMN. Before 6 hr postinfection, numbers of PMNs
were minimal. After 6 hr postinfection, increasing numbers of PMNs were observed throughout the vitreous, near the iris and ciliary body, and in the anterior
chamber. In many types of acute infections, including keratitis and endophthalmitis,7−14,20,29−33 PMNs
are one of the predominant initial infiltrating cell types
that contribute to the first line of defense in innate
immunity against invading organisms. In addition to
their function as phagocytes, PMNs can synthesize and
release reactive oxygen species, chemokines, and cytokines, including TNF-α.34 During infection, TNFα was detected just prior to and during increases in
PMNs in the eye. TNF-α is a robust cytokine, initiating and controlling local and systemic host responses
during the initial stages of infection. In the eye, TNFα production has been detected in several ocular infection and inflammation models, including experimental herpes simplex virus (HSV) keratitis,35 acute
HSV-1–induced acute retinal necrosis,36 dry eye,37
cytomegalovirus during mouse AIDS,38 experimental inflammatory uveitis,39,40 Pseudomonas aeruginosa
keratitis,41,42 and S. aureus endophthalmitis.43,44 The
mechanisms of PMN recruitment and the role of TNFα
and other proinflammatory cytokines during bacterial
endophthalmitis are currently being studied.
During endophthalmitis, complex exchanges occur
among bacteria, their products, and the host. Bacterial
growth kinetics and pathology of B. cereus and other
types of bacterial endophthalmitis have been described,
but the cellular and molecular events associated with
vision loss and inflammation from bacterial endophthalmitis have not been adequately characterized. Experimental murine models of bacterial endophthalmitis represent a technical advance over previous models
of intraocular bacterial pathogenesis, providing the opR. T. Ramadan et al.
portunity for use of a well-defined model for analyzing pathogen and host dynamics during infection in
both wild-type and transgenic backgrounds. Using this
model, we have identified important early events in infection, notably changes in the retina consistent with
altered cellular electrophysiology and an acute and robust inflammatory response. These events occurred during infection at a time earlier than once thought, suggesting that therapies, including those that target these
events, should be given as early as possible. Characterizing the mechanisms involved in interactions between
the host and bacterium will facilitate the design of novel
therapies strategically focused on retinal protection, inflammation and infection control to prevent vision loss
during Bacillus endophthalmitis.
ACKNOWLEDGMENTS
This work was supported by National Institutes of Health Grants R01EY12985 (M.C.C.), P30
RR12190 (NIH CORE Grant to Dr. Robert E.
Anderson, OUHSC), the Presbyterian Health Foundation (M.C.C.), and an unrestricted Career Development Award from Research to Prevent Blindness,
Inc. (M.C.C.) We gratefully acknowledge the technical assistance of Mark Dittmar, Martin Noriega, and
Andrea Mauer (Dean A. McGee Eye Institute Animal Facility), Paula Pierce (Excalibur Pathology, Oklahoma City, OK), and Andrea Moyer (OUHSC). We
also appreciate the helpful comments of Muayyad AlUbaidi, John Ash, Dan Carr, James Chodosh, Eric
Howard (OUHSC), Michael Gilmore (Schepens Eye
Research Institute, Boston, MA), and Michael Engelbert
(Columbia-Presbyterian Medical Center, New York,
NY).
REFERENCES
[1] Jett BD, Jensen HG, Atkuri RV, Gilmore MS. Evaluation of therapeutic measures for treating endophthalmitis caused by isogenic toxinproducing and toxin-nonproducing Enterococcus faecalis strains. Invest Ophthalmol Vis Sci. 1995;36:9–15.
[2] Pollack JS, Beecher DJ, Pulido JS, Lee Wong AC. Failure of intravitreal
dexamethasone to diminish inflammation or retinal toxicity in an
experimental model of Bacillus cereus endophthalmitis. Curr Eye
Res. 2004;29:253–259.
[3] Engelbert M, Mino de Kaspar H, Thiel M, et al. Intravitreal vancomycin and amikacin versus intravenous imipenem in the treatment
of experimental Staphylococcus aureus endophthalmitis. Graefes
Arch Clin Exp Ophthalmol. 2004;242:313–320.
[4] Ozkiris A, Evereklioglu C, Esel D, et al. The efficacy of intravitreal
piperacillin/tazobactam in rabbits with experimental Staphylococcus epidermidis endophthalmitis: a comparison with vancomycin.
Ophthalmic Res. 2005;37:168–174.
964
Curr Eye Res Downloaded from informahealthcare.com by Library Schusterman Center on 11/04/11
For personal use only.
[5] Ozkiris A, Evereklioglu C, Esel D, et al. The efficacy of
piperacillin/tazobactam in experimental Pseudomonas aeruginosa
endophthalmitis: A histopathological and microbiological evaluation. Curr Eye Res. 2005;30:13–19.
[6] Yildirim O, Oz O, Aslan G, et al. The efficacy of intravitreal levofloxacin and intravitreal dexamethasone in experimental Staphylococcus epidermidis endophthalmitis. Ophthalmic Res.
2002;34:349–356.
[7] Booth MC, Atkuri RV, Nanda SK, et al. Accessory gene regulator
controls Staphylococcus aureus virulence in endophthalmitis. Invest
Ophthalmol Vis Sci. 1995;36:1828–1836.
[8] Booth MC, Cheung AL, Hatter KL, et al. Staphylococcal accessory
regulator (sar) in conjunction with agr contributes to Staphylococcus
aureus virulence in endophthalmitis. Infect Immun. 1997;65:1550–
1556.
[9] Callegan, MC, Booth MC, Jett BD, Gilmore MS. Pathogenesis of Gram-positive bacterial endophthalmitis. Infect Immun.
1999;67:3348–3356.
[10] Callegan MC, Cochran DC, Kane ST, et al. Contribution of
membrane-damaging toxins to Bacillus Endophthalmitis pathogenesis. Infect Immun. 2002;70:5381–5389.
[11] Callegan MC, Jett BD, Hancock LE, Gilmore MS. Role of hemolysin
BL in the pathogenesis of extraintestinal Bacillus cereus infection as assessed using an endophthalmitis model. Infect Immun.
1999;67:3357–3366.
[12] Callegan MC, Kane ST, Cochran DC, Gilmore MS. Molecular mechanisms of Bacillus endophthalmitis pathogenesis. DNA Cell Biol.
2002;21:367–373.
[13] Callegan MC, Kane ST, Cochran C, et al. Contribution of plcR–
regulated toxins to Bacillus endophthalmitis pathogenesis. Infect
Immun. 2003;71:3116–3124.
[14] Callegan MC, Kane ST, Cochran DC, et al. Bacillus endophthalmitis:
roles of bacterial toxins and motility during infection. Invest Ophthalmol Vis Sci. 2005;46:3233–3238.
[15] Engelbert M, Mylonakis E, Ausubel FM, et al. Contribution of gelatinase, serine protease, and fsr to the pathogenesis of Enterococcus
faecalis endophthalmitis. Infect Immun. 2004;72:3628–3633.
[16] Jett BD, Atkuri RV, Gilmore MS. Enteroccus faecalis localization in experimental endophthalmitis: role of plasmid-encoded aggregation
substance. Infect Immun. 1998;66:843–848.
[17] Jett BD, Jensen HG, Nordquist RE, Gilmore MS. Contribution of the
pAD1-encoded cytolysin to the severity of experimental Enterococcus faecalis endophthalmitis. Infect Immun. 1992;60:2445–2452.
[18] Mylonakis E, Engelbert M, Qin X, et al. The Enterococcus faecalis
fsrB gene, a key component of the fsr quorum-sensing system, is
associated with virulence in the rabbit endophthalmitis model. Infect
Immun. 2002;70:4678–4681.
[19] Callegan, MC, Engelbert M, Parke DW, et al. Bacterial endophthalmitis: epidemiology, therapeutics, and bacterium-host interactions. Clin Microbiol Rev. 2002;15:111–124.
[20] Engelbert M, Gilmore MS. Fas ligand but not complement is critical
for control of experimental Staphlococcus aureus endophthalmitis.
Invest Ophthalmol Vis Sci. 2005;46:2479–2486.
[21] Jett BD, Hatter KL, Huycke MM, Gilmore MS. Simplified agar
plate method for quantifying viable bacteria. BioTechniques.
1997;23:648–650.
[22] Ranchon I, Gorrand JM, Cluzel J, et al. Light-induced variations of
retinal sensitivity in rats. Curr Eye Res. 1998;17:14–23.
[23] Sheehan DC, Hrapchak BB. Theory and Practice of Histotechnology.
Columbus, OH: Batelle Press; 1987.
[24] Pannicke T, Uckerman O, Iandiev I, et al. Ocular inflammation alters
swelling and membrane characteristics of Muller cells. J Neuroimmunol. 2005;161:145–154.
965
[25] Eisenfeld AJ, Bunt-Milam AH, Sarthy PV. Muller cell expression of
glial fibrillary acidic protein after genetic and experimental photoreceptor degeneration in the rat retina. Invest Ophthalmol Vis Sci.
1984;25:1321–1328.
[26] Sarthy V, Ripps H. The Retinal Muller Cell: Structure and Function,
1st ed. New York: Plenum Publishing; 2001.
[27] Ramadan RT, Moyer AL, Hibbard AH, et al. Bacillus endophthalmitis:
Retinal function loss and Muller cell dysfunction. Invest Ophthalmol
Vis Sci. 2004;45:E-abstract 4015.
[28] Peachey NS, Ball SL. Electrophysiological analysis of visual function
in mutant mice. Documenta Ophthalmol. 2003;107:13–36.
[29] Peachey NS, Fishman GA, Derlacki DJ, Brigell MG. Psychophysical
and electroretinographic findings in X-linked juvenile retinoschisis.
Arch Ophthalmol. 1987;105:513–516.
[30] Callegan MC, Engel LS, Hill JM, O’Callaghan. Corneal virulence of
Staphylococcus aureus: roles of alpha-toxin and protein A in pathogenesis. Infect Immun. 1994;62:2478–2482.
[31] Hazlett LD. Pathogenic mechanisms of P. aeruginosa keratitis: a review of the role of T cells, Langerhans cells, PMN, and cytokines.
DNA Cell Biol. 2002;21:383–390.
[32] Hazlett LD. Role of innate and adaptive immunity in the
pathogenesis of keratitis. Ocul Immunol Inflamm. 2005;13:133–
138.
[33] Ruan X, Chodosh J, Callegan MC, et al. Corneal expression of
the inflammatory mediator CAP37. Invest Ophthalmol Vis Sci.
2002;43:1414–1421.
[34] Hobden JA, Engel LS, Hill JM, et al. Prednisolone acetate or prednisolone phosphate concurrently administered with ciprofloxacin for
the therapy of experimental Pseudomonas aeruginosa keratitis. Curr
Eye Res. 1993;12:469–473.
[35] Bennouna S, Denkers EY. Microbial antigen triggers rapid mobilization of TNF-alpha to the surface of mouse neutrophils transforming
them into inducers of high-level dendritic cell TNF-alpha production.
J Immunol. 2005;174:4845–4851.
[36] Thomas J, Kanangat S, Rouse BT. Herpes simplex virus replicationinduced expression of chemokines and proinflammatory cytokines
in the eye: implications in herpetic stromal keratitis. J Interferon
Cytokine Res. 1998;18:681–690.
[37] Zheng M, Atherton S. Cytokine profiles and inflammatory cells during HSV-1-induced acute retinal necrosis. Invest Ophthalmol Vis Sci.
2005;46:1356–1363.
[38] Luo L, Li D, Doshi A, et al. Experimental dry eye stimulates production of inflammatory cytokines and MMP-9 and activates MAPK
signaling pathways on the ocular surface. Invest Ophthalmol Vis Sci.
2004;45:4293–4301.
[39] Dix RD, Cousins SW. Susceptibility to murine cytomegalovirus retinitis during progression of MAIDS: correlation with intraocular levels
of tumor necrosis factor-alpha and interferon-gamma. Curr Eye Res.
2004;29:173–180.
[40] Shen DF, Buggage RR, Eng HC, Chan CC. Cytokine expression in
different strains of mice with endotoxin-induced uveitis (EIU). Ocul
Immunol Inflamm. 2000;8:221–225.
[41] Woon MD, Kaplan HJ, Bora NS. Kinetics of cytokine production
in experimental autoimmune anterior uveitis (EAAU). Curr Eye Res.
1998;17:955–961.
[42] Cole N, Bao S, Willcox M, Husband AJ. TNF-α production in the
cornea in response to Pseudomanas aeruginosa challenge. Immunol
Cell Biol. 1999;77:164–166.
[43] Hazlett LD. Corneal response to Pseudomonas aeruginosa infection.
Prog Retin Eye Res. 2004;23;1–30.
[44] Giese MJ, Sumner HL, Berliner JA, Mondino BJ. Cytokine expression in a rat model of Staphlococcus aureus endophthalmitis. Invest
Ophthalmol Vis Sci. 1998;39:2785–2790.
Experimental Bacillus Endophthalmitis