Reports
IOVS, June 1999, Vol. 40, No.7
tistical significance was due to the presence of several eyes that
had high facilities, very high fixative volumes, and low B pore
densities, but that nonetheless did not qualify as outliers according to the criteria established in that study.5 We had no
such eyes in the present work and therefore observed a statistically significant correlation between B pore density and fix
volume. There are several possible explanations of these data.
It is possible that there is, in fact, a correlation between B pore
density and fixative volume that was masked in the earlier
study.5 Alternatively, B pore density may depend on other
factors that have not yet been identified. More work is required
to investigate this issue.
Some of the eyes included in this study had postmortem
times greater than 20 hours, which was previously identified as
a time at which changes in I pore density could be observed in
human eyes.5 Because of the paired nature of this study, such
postmortem effects would be a cause for concern if they
interacted with ECA effect(s) to alter facility and/or pore statistics. We found no evidence for a postmortem interaction
with ECA effects. More specifically, regression analysis of the
following ratios (defined as ECA-treated eye value/contralateral
eye value) against postmortem time found no statistically significant dependencies: posttreatment facility ratio (JP = 0.26);
grand facility ratio (P = 0.35); pore density ratio for I, B, and
total pores (P — 0.67, 0.49, 0.63, respectively); pore size ratio
for I, B, and total pores (P = 0.82, 0.36, 0.90, respectively); and
nD product ratio for I, B, and total pores (P = 0.83, 0.92, 0.80,
respectively). Because all statistical inferences were based on
ECA/control ratios, the above results strongly suggest that
postmortem time was not a confounding factor in the analysis
and that the results would not have changed if all eyes included
in the study had postmortem times of 20 hours or less.
Contact Lens-Induced
Infection—A New Model of
Candida albicans Keratitis
Denis M. O'Day, W. Steven Head,
Richard D. Robinson, Rong Yang, Debra Shetlar,
and Ming X. Wang
A model of experimental keratomycosis was established that mimics human disease in which the only
fungi present are those that are actively growing within
the cornea.
PURPOSE.
From the Department of Ophthalmology and Visual Sciences, Vanderbilt University School of Medicine, Nashville, Tennessee.
Supported in part by a National Eye Institute, Bethesda, Maryland,
Grant EY01621 and a grant from Research to Prevent Blindness, New
York, New York.
Submitted for publication August 25, 1998; revised February 2,
1999; accepted February 24, 1999.
Proprietary interest category: N.
Reprint requests: Denis M. O'Day, 1215 21st Avenue, S., 8th Floor
Medical Center East, Nashville, TN 37232-8808.
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
1607
References
1. Ethier CR, Croft MA, Coloma FM, Gangnon R, Ladd B, Kaufman PL.
ECA Effects on inner wall pores of Schlemm's canal in living
monkeys. Invest Ophthalmol Vis Sci. 1999;40:1382-1391.
2. Ethier CR, Coloma FM, deKater AW, Allingham RR. Retroperfusion
studies of the aqueous outflow system: part 11, studies in human
eyes. Invest Ophthalmol Vis Sci. 1995;36:2466-24753. Ethier CR, Ajersch P, Pirog R. An improved ocular perfusion
system. Curr Eye Res. 1993; 12:765-770.
4. Sit AJ, Coloma FM, Ethier CR, Johnson M. Factors affecting the
pores of the inner wall of Schlemm's canal. Invest Ophthalmol Vis
Sci, 1997;38:1517-1525.
5. Ethier CR, Coloma FM, Sit AJ, Johnson M. Two pore types in the
inner-wall endothelium of Schlemm's canal. Invest Ophthalmol
Vis Sci. 1998;39:2041-2048.
6. Allingham RR, deKater AW, Ethier CR, Anderson PJ, Hertzmark E,
Epstein DL. The relationship between pore density and outflow
facility in normal and glaucomatous human eyes. Invest Ophthalmol Vis Sci. 1992;33:1661-16697. Johnson M, Shapiro A, Ethier CR, Kamm RD. The modulation of
outflow resistance by the pores of the inner wall endothelium.
Invest Ophthalmol Vis Sci. 1992;33:l670-l6758. Epstein DL, Freddo TF, Bassett-Chu S, Chung M, Karageuzian L.
Influence of ethacrynic acid on outflow facility in the monkey and
calf eye. Invest Ophthalmol Vis Sci. 1987;28:2067-2075.
9. Erickson-Lamy K, Schroeder A, Epstein DL. Ethacrynic acid induces reversible shape and cytoskeletal changes in cultured cells.
Invest Ophthalmol Vis Sci. 1992;33:2631-2640.
10. O'Brien ET, Kinch M, Harding TW, Epstein DL. A mechanism for
trabecular meshwork cell retraction: ethacrynic acid initiates the
dephosphorylation of focal adhesion proteins. Exp Eye Res. 1997;
65:471-483.
11. Liang LL, Epstein DL, de Kater AW, Shahsafaei A, Erickson-Lamy
KA. Ethacrynic acid increases facility of outflow in the human eye
in vitro. Arch Ophthalmol, 1992;110:106-109.
METHODS. Dutch-belted
rabbits received a subconjunctival
injection of triamcinolone acetonide to one eye. One day
later the epithelium was removed from the central cornea
and a standardized inoculum of Candida albicans blastoconidia was placed on the corneal surface and covered
with a contact lens. The lids were closed with a lateral
tarsorrhaphy. After 24 hours, the lid sutures and contact
lens were removed. Five days later the animals were
killed, and their corneas were subjected to separate isolate
recovery and histology studies. A group of similarly infected rabbits without corticosteroid injection served as
controls.
RESULTS. Both groups developed invasive corneal disease.
Although isolate recovery was not significantly different
from corticosteroid-treated rabbits compared with controls, fungal biomass was increased. Hyphal invasion was
limited to the anterior cornea in control eyes, but penetrated deep stroma in most of the corticosteroid-treated
rabbits.
CONCLUSIONS. Invasive corneal disease can be established
with a surface inoculum. Corticosteroid administration
increased corneal penetration of hyphae. Quantitative isolate recovery is not a reliable measure of the fungal load
within the cornea. (Invest Ophthalmol Vis Sci. 1999;40:
1607-1611)
1608
Reports
R
ecent studies of experimental fungal keratitis have focused
on inoculation of blastoconidia to establish disease.1"4
However, it has become clear to us that the residual intrastromal inoculum, the portion of inoculum that is not actively
replicating but that remains viable for the course of an experiment can confound our ability to adequately evaluate experimental therapy when the measure of a drug's efficacy is based
on quantitative isolate recovery.4 This is not a problem with
fungicidal agents, which are capable of killing Candida albicans in both quiescent and metabolically active phases. Unfortunately, most new antifungal drugs are fungistatic. This led us
to develop a model in which the only fungus present in the
cornea is in the invasive phase. We demonstrate the feasibility
of this model with C. albicans and measure the effect of
corticosteroid administration on the fungal invasion of the
cornea.
IOVS, June 1999, Vol. 40, No.7
1. Measures of Fungal Recovery from Rabbits
Infected with Candida albicans
TABLE
Treatment
n
Control
1 mg TA
3 mgTA
12 mg TA
7
3
3
4
Isolate Recovery
(log2CFU)
12.32
13.90
16.24
13.52
±
±
±
±
0.75
1.14
1.14
0.99
Fungal Biomass
log 2 (area in /mm2)
12.40
13-84
13.78
16.68
±
±
±
±
0.87
1.33
1.33
1.15
Values are means ± SEM; n is number of rabbits.
MATERIALS AND METHODS
The tarsorrhaphy sutures and contact lens were removed
24 hours later. In the first experiment one group of animals
served as controls, and the other received 12 mg subconjunctival TA. The second experiment consisted of three groups of
three animals each: group one received 1 mg TA, group two
received 3 mg TA, and group three was a control group.
Inoculum
Postmortem Tissue Dissection
We used C. albicans strain VE-175, a stable switch variant of
strain VE-102, a human corneal isolate that we have studied
previously.2'5 The inoculum was prepared by inoculating two
Sabouraud's agar plates (BBL, Coxsackie, MD) from a frozen
agar plug that was preserved at [minus}80°C. One plate was
grown confluently, and the other was streaked for isolation to
confirm the purity of the culture. Cultures were incubated for
24 hours at 35°C, harvested in sterile normal saline, and pelleted by centrifugation. After the supernatant was poured off,
yeasts were suspended in the saline that remained in the tube,
yielding a paste of yeast containing 1 X 1010 colony-forming
units (CFU)/ml, which was pipettable using a wide-mouth
pipette tip.
Rabbits were killed by rapid intravenous injection of euthanasia solution. The infected corneas were removed at the limbus
and bisected longitudinally through the middle of the infected
area. One half of the cornea was cut into small pieces for
quantitative isolate recovery, and the other half was fixed in
formalin for histology.
Animals
Dutch-belted rabbits of either sex weighing 1 kg to 2 kg were
used in these experiments. All animals were treated in accordance with the ARVO Statement for the Use of Animals in
Ophthalmic and Vision Research.6
Corticosteroid Administration
On the day before inoculation, rabbits that were to receive
corticosteroid were given subconjunctival injections of triamcinolone acetonide (TA, 40 mg/ml, Kenalog; E. R. Squibb &
Sons, Princeton, NJ) in the dosage of 1 mg (25 /AI), 3 mg (75
/xl), or 12 mg (300 /zl). Control rabbits did not receive an
injection.
Model of Invasive Keratitis
Dutch-belted rabbits were anesthetized with intramuscular ketamine and xylazine. Corneal anesthesia was obtained with
topical 0.5% proparacaine hydrochloride. A 7-nim trephination
mark was made in the central cornea of one eye, and the
epithelium was completely removed from the enclosed area
with a knife edge. The nictitating membrane was removed by
sharp dissection. Twenty microliters of the fungal inoculum
was then placed directly on the debrided area of the cornea
and covered with a contact lens. A lateral tarsorrhaphy prevented extrusion of the contact lens. The animal remained
anesthetized for at least 2 hours after inoculation.
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
Quantitative Isolate Recovery of Fungi from
Infected Corneas
Corneal halves were placed individually into 3 ml sterile normal saline in a test tube and processed for quantitative isolate
recovery using our standard method of homogenizing the cornea and plating samples of serial dilutions in triplicate on
Sabouraud's agar.2
Histology
For each cornea, three 5-jam, paraffin-embedded sections were
cut at 20 intervals of 100 jam from the midcornea. The sections
were stained by the Gomorri's methenamine silver method to
selectively highlight fungal elements. We used Bioquant (Biometrics, Nashville, TN) software for image capture of the corneal section. The image analysis equipment included a microscope with a calibrated stage, a black and white camera, and a
video image capture card coupled to a computer with an Intel
Pentium® processor. This allowed us to measure total surface
area of the fungal elements contained in each corneal section
(fungal biomass). By adjusting the lighting for each section
scanned, we were able to capture an image from the microscope that was sufficient to differentiate the hyphal elements
from the background tissue, based on gray scale density. The
settings within the software were then adjusted to "tag" the
fungal elements according to their level of relative "darkness"
compared with the background. Automation of this process
was not possible because of confounding staining by inflammatory cells. However, by direct comparison of the gray scale
image to the slide on the microscope, we were able without
difficulty to delete manually any erroneously "tagged" objects.
Once all the hyphae were tagged, the computer determined
the area. One section at each 100jam interval was scanned
microscopically, and the total cross-sectional area of the hyphal
elements was measured in squared-micrometers.
Reports
K)\'S. lunc 1999. Vol. •!(). No."
1609
IK,I HI 1. (iimlnlii tilbicdiis in the
cornea of a Dutch-helled ruhhit T
days al'ler inoculation with blastoconidia. This rahhit was not j'iven
any lorticostcroid before inoculation Stained by Ciomorri's methenaniinc silver method. Magnification.
Statistical Analysis
We pooled the data from both experiments for analysis because there were no significant differences in the data from the
control groups in the two experiments. Because the effect of
treatment was modeled to be proportional to the prelreatmeiit
level of disease, a log transformation of the data was done.
Standard analysis of variance methods were used to examine
the effect of corticosteroid administration on disease."'' K A
further analysis of the biomass data was done to determine the
reliability of using only one, two, or three cortical sections to
determine the total fungal biomass compared with multiple-
sections cut through the entire diameter of the fungal lesion
The correlation between isolate recovery and fungal biomass
also was examined to determine whether these two measures
of response were different. All statistical analyses were performed using SAS for Windows (SAS Institute, Cary, NO.
RESULTS
All rabbits in both experiments became infected, with sufficient disease in control rabbits to permit histologic measure-
?
FH.UKI- 2. ('ii)uliilti iilhiain.s in the
cornea of a Dutch-belted rabbit S days
after injection with hlastoconidia This
rabhil was jjiven 12 mji corticosteroid
(triamcinolone acetonide) suheonjnneti\ally on the da\ before inoculation
Stained In (•omorri's methenamine sit\ermethotl Magnification. <2.SS.
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
1610
Reports
ment of hyphal invasion into the corneal stroma (Table 1).
Although isolate recovery was consistently liigher from rabbits
treated with corticosteroid compared to control rabbits, this
trend did not achieve statistical significance. In control corneas, hyphal invasion was limited to the anterior stroma and
generally was sparse (Fig. 1). In rabbits treated with corticosteroid, fungal biomass increased in all corneas, manifested by
an increase in the density of fungal elements and to a lesser
extent by an increased depth of penetration (Fig. 2). Fungal
biomass increased progressively with increasing dose of corticosteroid administered. There was a statistically significant
difference for fungal biomass measurements only between the
group of rabbits that received the highest dose (12 mg) of
corticosteroid per day and control groups (P = 0.01).
To examine the reliability of determining fungal biomass
by evaluating fewer corneal sections, we first considered control rabbits. In these eyes, an R value of 0.99 was obtained
when the data from one corneal section of each eye was
compared with the entire data set of 20 corneal sections. When
all groups of rabbits were considered (treated and untreated
controls) the correlation fell slightly (R = 0.92). When the data
from three sections for each eye were pooled and compared
with the entire data set, the correlation for control rabbits was
0.98 and for all rabbits groups it was 0.99.
To confirm that the fungal biomass was, in fact, a different
measure of response to therapy than isolate recovery in this
model, we examined the correlation between isolate recovery
and fungal biomass data. For the group of untreated control
rabbits, the correlation was poor (R = 0.39). If all rabbits were
considered, the value for R dropped to 0.25.
DISCUSSION
The development of a reliable animal model of fungal keratitis
has been an elusive goal for researchers in ocular fungal disease. Normal laboratory animals, even more so than humans,
appear resistant to the most common fungal pathogens so that
the reliable establishment of infection remains a fundamental
problem. In our laboratory we have developed several models
of fungal keratitis for use in the study of pharmacological
agents, but these models, while useful experimentally, appear
inappropriate for the study of certain classes of agents such as
fungistatic drugs. Moreover, they do not permit the study of
events in the initiation and development of corneal infection.
The complex life cycle of fungi is an obstacle to the
development of precise methods for quantifying the disease,
an essential element in any model used in therapeutic research.
With bacteria, it is relatively straightforward to estimate
organism content using colony-counting techniques. For fungi,
the fact that every nucleus-containing hyphal fragment has the
potential to form a colony introduces a large degree of uncertainty in organism counts. In addition, when the inoculum
(composed of blastoconidia) is deposited in the corneal
stroma, there is evidence that nonprolifera ting blastoconidia
remain viable for a long time (DM O'Day, unpublished data,
1998), although hyphal proliferation may not occur. These
surviving, but not replicating, colony-forming units contribute
to a background colony count that limits the detection of small,
but significant changes in fungal populations.
In setting out to establish a valid model of fungal keratitis,
we recognized four essential components: (1) the method of
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
IOVS, June 1999, Vol. 40, No.7
initiation of infection should approximate the way human
infection occurs; (2) immunologic manipulation of the host
animal should be unnecessary; (3) the invasive fungal elements
should be exclusively in the hyphal phase; and (4) quantification of disease should be objectively measurable. Previous
studies have shown Dutch-belted rabbits to be suitable experimental animals for these studies.9 For the inoculum, we chose
C. albicans strain VE-175 on the basis of our previous studies
with this organism. The method of inoculation was developed
on the analogy of human infection after contact lens wear.
Excision of the nictitating membrane and a temporary tarsorrhaphy were performed to enhance lodgment of the inoculum
on the bare stromal surface so that adequate time for invasion
could occur before the nonattached organisms were swept
from the corneal surface. In this way, we hoped to have only
hyphal-phase organisms in the corneal stroma.
The previous difficulty in objectively measuring this disease led us to attempt quantitative measurement of the amount
of fungus (fungal biomass) in corneal sections. These data
strongly support both the feasibility and practicality of this
approach.
This model appears to mimic human corneal infection:
blastoconidia deposited on the debrided corneal surface germinate and produce hyphae, which invade the corneal stroma.
Despite the brief 5-day period of observation, histologic sections reveal a relatively uniform stromal invasion by healthy
hyphae. From two perspectives these are important observations: they indicate that the bare nontraumatized corneal
stroma is susceptible to fungal infection exclusively in the
hyphal phase by this strain of C. albicans and that tissue
invasion can be evaluated histologically in this model without
the need to examine large numbers of serial sections.
The second part of the study demonstrated the utility of
the model by confirming previous reports of the effect of
corticosteroid on fungal invasion.10 Again, fungal growth was
exclusively in the hyphal phase. As we anticipated, colony
counts failed to correlate with the estimated fungal biomass;
however, this does highlight a problem—the lack of a gold
standard for measuring fungal growth in tissue. Clearly, although this quantitative histologic scanning technique will
provide a robust measure of fungal growth, because the baseline is zero, assessing viability remains an open question that
cannot be resolved by population estimates based on colony
counting techniques. The development of sensitive quantitative measures of fungal growth in tissue is an essential element
for future progress. Techniques that can measure events at this
molecular level may be the key and are the focus of our
ongoing research.
References
1. Singh SM, Khan R, Sharma S, Chatterjee PK. Clinical and experimental mycotic corneal ulcer caused by Aspergillus fumigatus and
the effect of oral ketoconazole in the treatment. Mycopathologia.
1989:106:133-141.
2. O'Day DM. Orally administered antifungal therapy for experimental keratomycosis. Trans Am Ophthalmol Soc. 1990;88:685-725.
3. O'Day DM, Ray WA, Head WS, Robinson RD, Williams TE. Influence of corticosteroid on experimentally induced keratomycosis.
Arch Ophthai 1991;109:l601-l604.
4. O'Day DM, Head WS, Robinson RD, Williams TE, Gedde S. The
evaluation of therapeutic responses in experimental keratomycosis. Curr Eye Res. 1992; 11:35- 44.
IOVS, June 1999, Vol. 40, No. 7
5. Ray WA, O'Day DM, Head WS, Robinson RD. Variability in isolate
recovery rates from multiple and single breeds of outbred pigmentecl rabbits in an experimental model of Candida keratitis.
Curr Eye Res. 1984;3:949-953.
6. Animals in Research Committee. The Association for Research
in Vision and Ophthalmology (ARVO) statement for the use of
animals in ophthalmic and vision research. In: Handbook for
the Use of Animals in Biomedical Research. Bethesda, MD:
Association for Research in Ophthalmology; 1993:15-16.
Reports
7. Steele R, Torrie J. Principles and Procedures of Statistics. New
York: McGraw-Hill; 1980.
8. Ray WA, O'Day D. Statistical analysis of multi-eye data in ophthalmic research. Invest Ophthalmol Vis Sci, 1985;26:1186-1188.
9. O'Day DM. Studies in experimental keratomycosis. Curr Eye Res.
1985;4:243-252.
10. O'Day DM, Ray WA, Robinson RD, Head WS. Efficacy of antifungal
agents in the cornea. II. Influence of corticosteroids. Invest Ophthalmol Vis Sci. 1984;25:331-335.
Immunopathology of Pineal
Glands from Horses with Uveitis
the uveitis of these horses is recurrent. Study of pineal
glands from horses with clinically documented uveitis
allows demonstration of subtle pineal changes associated
with natural uveitis. Similar changes would be difficult to
document in human patient populations. (Invest OphthalVis Sci. 1999;40:l6ll-l6l5)
Carolyn M. Kalsow,x Richard R. Dubielzig,2 and
Ann E. Dwyer13
Pinealitis accompanying uveitis is well established in laboratory models of experimental autoimmune
uveoretinitis. In naturally occurring uveitis, pinealitis has
been demonstrated in the pineal gland from a mare with
active uveitis and is suspected in some human uveitides.
We have evaluated pineal glands from horses with various
stages of uveitis for signs of immunopathology accompanying spontaneous uveitis.
PURPOSE.
METHODS. Pineal glands from 10 horses with uveitis and
from 13 horses without uveitis were evaluated for histochemical (H&E, collagen) and immunohistochemical
(MHC class II antigen expression, infiltration of T and B
lymphocytes, and glial fibrillary acidic protein (GFAP) and
vimentin upregulation) evidence of inflammation.
Septal areas of pineal glands from horses with
uveitis had clusters of MHC class II antigen-expressing
cells, T lymphocytes, and enhanced collagen deposition.
These changes were not as readily observed in pineal
glands from horses without uveitis. B lymphocytes were
detected only in the pineal gland from the one mare with
active uveitis in which T and B lymphocytes were organized into follicles. No differences in GFAP or vimentin
immunoreactivity were noted in pineal glands from horses
with or without uveitis.
RESULTS.
CONCLUSIONS. These pineal gland changes suggest that the
pinealitis associated with equine uveitis is transient just as
From the ' Department of Ophthalmology, University of Rochester
School of Medicine and Dentistry, Rochester, New York; 2 Department
of Pathobiological Sciences, School of Veterinary Medicine, University
of Wisconsin-Madison, Madison; and 3 Genesee Valley Equine Clinic,
Scottsville, New York.
Supported by U. S. Public Health Service, National Institutes of
Health, Bethesda, Maryland, Grant EY06866 (CMK); American Society
of Veterinary Ophthalmology, Stillwater, Oklahoma (AED); and an
unrestricted grant from Research to Prevent Blindness, Inc., New York,
New York, to the University of Rochester Department of OphthalmologySubmitted for publication September 30, 1998; revised January 7,
1999; accepted February 16, 1999Proprietary interest categoiy: N.
Reprint requests: Carolyn M. Kalsow, Department of Ophthalmology, University of Rochester, School of Medicine and Dentistry, Rochester, NY 14642.
Downloaded From: http://iovs.arvojournals.org/ on 06/18/2017
l6ll
ymphocytic infiltration of the pineal gland is regularly observed in laboratory models of experimental autoimmune
uveoretinitis (EAU). This experimental autoimmune pinealitis
(EAP) is a generalized phenomenon that has been induced in a
variety of species by various photoreception-associated proteins that are also present in pineal gland.1"''
In humans, lower serum levels of the pineal neurohormone melatonin5"7 in patients with uveitis suggests pineal
gland abnormalities in these individuals. However, there is no
direct histopathologic evidence of pineal gland changes in
human patients with uveitis. Affected pineal glands are usually
not available for direct observation, especially at the time of
active inflammation, and noninvasive techniques have yet to be
developed to detect pineal inflammation in situ.
Equine recurrent uveitis is a natural disorder that can serve
as a model of human uveitis.8 This is a spontaneous inflammation in which there is no artificial perturbation of the systemic
immune response, and tissues from horses euthanatized because of blindness or other reasons are more readily obtainable
for study than are tissues from human patients. Documented
pinealitis in a mare with active uveitis,9 prompted this investigation of immunohistopathologic changes of pineal glands
from several horses with and without uveitis.
L
MATERIALS AND METHODS
Pineal Glands
Pineal glands from 10 horses with a diagnosis of uveitis10'11
and from 13 horses without uveitis were provided by Ann
Dwyer of the Genesee Valley Equine Clinic, William Rebhun,
DVM, of Cornell University, and Richard Dubielzig, DVM, of
the University of Wisconsin. The tissue was recovered from
horses at the time of euthanasia or natural death, fixed in
formalin, 95% ethanol or Bouin's solution and embedded in
paraffin.
Clinical information was available for 6 of the 10 horses
with uveitis. All 6 had had a recurrence in one eye within one
year and had a duration of uveitis for greater than a year at the
time of euthanasia. Only one horse was considered to have
active uveitis at the time of euthanasia, i.e., recurrence within
2 weeks rather than 2 months or greater. These horses had
been treated expeditiously.11 The other 4 horses with uveitis