Download Sensory and sympathetic innervation of the mammalian

Investigative Ophthalmology & Visual Science, Vol. 30, No. 3, March 1989
Copyright © Association Tor Research in Vision and Ophthalmology
Sensory and Sympathetic Innervation of the
Mammalian Cornea
A Retrograde Tracing Study
Carl F. Marfurr,* Robert E. Kingsley,t
oncl
Stephen E. Echtenkamp^:
The method of retrograde transport of horseradish peroxidase-wheat germ agglutinin (HRP-WGA)
was used to study the locations, numbers and somata diameters of cornea! afferent and efferent
neurons in four different mammalian species. A 2% solution of HRP-WGA was applied to the central
cornea of the rat, rabbit, cat and monkey and the animals were perfusion-fixed 48-96 hr later.
HRP-WGA-labeled sensory neurons were distributed relatively uniformly throughout the ophthalmic
region of the ipsilateral trigeminal ganglion of the rat, cat and monkey. In marked contrast, labeled
cells in the rabbit trigeminal ganglion were clustered in a sharply defined longitudinal column located
in the midregion of the ophthalmic area. Occasional cells in some cats and monkeys were observed
near, and possibly within, the maxillary region of the ganglion. There was no evidence for a dorsoventral somatotopy of corneal afferent neurons in any species. The majority of the labeled afferent somata
were small or medium in size, although some larger diameter neurons were also observed. Modest
numbers of labeled neurons were observed in the ipsilateral superior cervical ganglion (SCG) of the
rabbit and cat; however, only occasional labeled neurons were observed in the SCG of the rat, and none
were seen in the monkey. The labeled SCG cells, when present, were concentrated in the rostral half of
the ganglion, although many cells in the cat SCG were also found further caudally. No labeled neurons
were found in the middle cervical or stellate ganglia of any animal. The results of this study have
revealed the existence of subtle interspecies differences in the organization of the mammalian corneal
afferent and efferent innervations. Invest Ophthalmol Vis Sci 30:461-472,1989
modify corneal sensitivity. 7 The mechanism by
which corneal nerves exert these influences is unknown, although the release of axonally transported,
biologically active substances such as peptides and
neurotransmitters may be involved.
Information about the cell bodies that give origin
to these different components of the corneal innervation is limited. The corneal sensory fibers in mice,
rats, cats and monkeys originate from neurons located, with rare exceptions, in the ophthalmic (medial) region of the ipsilateral trigeminal ganglion.8"13
However, the distribution of corneal afferent somata
in the rabbit, a popular animal model in ophthalmological research, has not been described. In addition,
there is little information concerning the relative
numbers of sensory cells that contribute to the corneal innervation in different species. While it is
widely reported that comeal afferent perikarya are
small or medium in size, the cell size spectra for these
neurons' 4 1 5 in most species remain unknown.
Knowledge of the origin(s) of the corneal sympathetic
innervation is even more fragmentary,716 and analysis of the limited available data suggests that there
may be significant interspecies differences.
The mammalian cornea is richly supplied by sensory nerve fibers from the trigeminal ganglion (TG)
and, to a lesser extent, by sympathetic fibers from the
superior cervical ganglion (SCG). Both populations
of nerve fibers play unique roles in maintaining the
overall structural and functional integrity of the cornea. Corneal afferent nerves, in addition to signalling
sensory information, exert important trophic influences on the corneal epithelium 1 and stimulate
wound healing following epithelial lesions.2 Corneal
sympathetic nerves, in turn, regulate epithelial ion
transport processes,3 exert inhibitory influences on
epithelial mitogenesis and wound healing4"6 and may
From the Departments of *Anatomy and ^Physiology, Northwest Center for Medical Education, Indiana University School of
Medicine, Gary, Indiana, and the fDepartment of Physiology,
South Bend Center for Medical Education, Indiana University
School of Medicine, University of Notre Dame, Notre Dame, Indiana.
Supported in part by PHS grants EY-05717 to CFM and
HL37223 to SFE.
Submitted for publication: June 24, 1988; accepted October 5,
1988.
Reprint requests: Dr. Carl F. Marfurt, Northwest Center for
Medical Education, 3400 Broadway, Gary, IN 46408.
461
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 05/05/2017
462
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / March 1989
In the current investigation, we have used the
method of retrograde transport of horseradish peroxidase-wheat germ agglutinin (HRP-WGA) to study
the afferent and efferent corneal innervation of four
mammalian species commonly used in ophthalmological research: the rat, rabbit, cat and monkey. The
specific goals of the current study were to determine:
(1) the distribution within the trigeminal and superior
cervical ganglia of corneal sensory and sympathetic
nerve cell bodies; (2) the relative numbers of trigeminal and SCG cells that contribute to the corneal innervation in different species; and (3) the somata diameters of the corneal afferent neurons. The data
generated from this study will increase our understanding of the anatomy of the mammalian corneal
innervation and will provide a useful database of information for future studies by ourselves and other
workers in such areas as nerve cell loss following corneal wounding, the development of experimental animal models of herpes virus infections, and the role of
the sympathetic nervous system in the regulation of
corneal physiology.
Materials and Methods
A total of ten rats, six rabbits, four cats and four
monkeys were used in this investigation. Some of the
data from the rat and monkey experiments have been
presented in preliminary fashion as parts of larger
studies on corneal sympathetic innervation in the
rat16 and brainstem termination sites of corneal afferent neurons in monkeys.13 AH experiments were performed in accordance with the guidelines outlined in
the ARVO Resolution on the Use of Animals in Research.
Each animal was anesthetized with ketamine (10
mg/kg, I.M.) or sodium pentobarbital (35-45 mg/kg,
I.P. or I.V.), and the central area of the corneal epithelium was gently scratched in a checkerboard fashion with a sterile number 11 scalpel blade. Partial
disruption of the epithelium was found in preliminary experiments to be necessary in order to permit
satisfactory penetration of the tracer substance into
the cornea and to obtain good retrograde transport
within the corneal nerve fibers. The scratched area
was in every case restricted to the central one-fourth
to one-half of the cornea and never involved the peripheral cornea or the coraeoscleral limbus. The prepared corneal areas in cats were, on the average,
smaller than those in the other species due to the
tendency for the cat eyeballs to be drawn back into
the orbit by contraction of the retractor bulbi muscles. The procedure was performed unilaterally in at
least two animals from each species, and bilaterally in
the remaining animals.
'
Vol. 30
In the rabbits, cats and monkeys, the scratched area
was encircled with a 2 mm high vaseline wall constructed by squeezing a thin strand of vaseline
through a 25-gauge syringe needle. Thin wafers of
gelfoam surgical sponge were then saturated with 5-8
n\ of 2% HRP-WGA in saline, placed on the
scratched corneal surface, and manipulated until they
completely filled the area inside the vaseline retaining
wall. The preparation was continuously monitored
through a dissecting microscope for 15-30 min, after
which the gelfoam pledgets were removed and the
corneal surface was rinsed for 2-3 min with sterile
saline. The eyes were then gently dabbed with cotton
applicator sticks and the eyelids were closed with
butterfly bandages or sutures. In the rats, the method
of tracer application was similar to that used in the
larger animals, except that a modified corneal trephine with an inside diameter slightly larger than that
of the scratched corneal surface was used to hold the
HRP-WGA solution in place. Postoperative recovery
for all of the animals was unremarkable and daily
monitoring of their progress revealed no evidence of
epiphora, photophobia or ocular irritation of any
sort.
At least one eye from one animal of each species
was used as a control to check the specificity of the
experimental design and to test for possible leakage of
tracer away from the central corneal application site.
In the latter cases, 3-5 n\ of 2% HRP-WGA was
applied in dropwise fashion to the central areas of
normal, nondamaged corneas and was permitted to
flow freely over the entire cornea and surrounding
extraocular tissues. Thirty minutes later, the tracer
was removed by rinsing the eye repeatedly with sterile
saline, dabbing it gently with cotton applicator swabs,
and closing the eyelids with butterfly bandages or sutures as described above.
Forty-eight to 96 hr later, all animals were reanesthetized and perfused through the left ventricle or
retrogradely through the descending aorta with
200-500 ml of warm, phosphate-buffered saline, pH
7.4, containing 5000 units/kg heparin and 3% procaine-HCl, followed in sequence by 1-2 1 of warm
fixative containing 1% paraformaldehyde-2% glutaraldehyde in 0.1 M phosphate buffer, and 1 1 of icecold 0.1 M phosphate buffer with 10% sucrose. The
trigeminal ganglia, superior cervical ganglia, corneas
and irides were immediately removed and placed in
fresh 0.1 M phosphate buffer. In at least two animals
of each species, the ciliary, middle cervical and stellate ganglia were also removed. Sections through each
ganglion were cut serially at 40 pm in a cryostat,
mounted directly on chrome alum-gelatin-coated
slides, air-dried and reacted for the presence of HRP
activity according to the tetramethylbenzidine
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 05/05/2017
No. 3
CORNEAL INNERVATION / Morfurr er al
(TMB) procedure of Mesulam.17 All tissues were then
carefully examined with darkfield optics for the .presence of HRP labeled neurons.
The corneas and irides from every animal were also
examined in order to observe the application sites.
Rat and monkey corneas and the irides of all four
species were reacted whole with TMB histochemistry
as free-floating preparations; the thicker cat and rabbit corneas were split along natural cleavage planes in
the corneal stroma into two or three layers prior to
TMB processing to permit better penetration of the
incubation media components. Four radial slits were
made in each tissue and they were then flat-mounted
on chrome alum-gelatin-coated slides and air-dried
under weighted coverslips.
The total number of labeled neurons in each ganglion was determined by counting labeled somata
containing visible nuclei. Other cells were also
counted if their maximum profiles were contained
entirely within the thickness of a section but were so
heavily filled with reaction product that the cell nucleus was obscured. Corrections for double-counting
of split neurons in the trigeminal ganglia were calculated by using the formula of Abercrombie,18 however, double-counting of the relatively small numbers
of SCG neurons was not a problem since it was possible to identify the same cell in adjacent serial sections.
The locations of every HRP-labeled neuron in three
randomly selected trigeminal and superior cervical
ganglia from each species were carefully plotted onto
line drawings made by using a drawing tube attached
to the microscope. Measurements were determined
for 500 HRP-labeled trigeminal ganglion neurons in
each species by using an ocular micrometer and averaging the maximum and minimum cell diameters
through the equatorial region of each cell. To insure
randomness of sample, cell diameters were obtained
from at least four different ganglia for each species,
and within each ganglion the labeled cells were measured in every third section to eliminate the possibility of measuring the same cell twice in adjacent sections.
Results
Corneal Injection Sites
Examination of the corneas and irides from the
experimental animals revealed a dense staining of the
epithelium and stroma of the central scarified cornea,
with no reaction product in the peripheral cornea,
limbus or iris (Fig. 1). The percentage of the corneal
area stained with TMB reaction product varied from
animal to animal, but in general covered the central
one-quarter to one-half of the cornea, and was, on the
average, slightly less in the cats than in the other spe-
463
Fig. 1. Brightfield light micrograph of a rabbit corneal whole
mount illustrating the central location and restricted size of a typical HRP-WGA application site (arrow). The application site in this
particular animal covers approximately 48.5% of the total corneal
area.
Trigeminal Ganglia
HRP-WGA application to the central cornea of
the experimental animals resulted in the production
of numerous well labeled somata and fiber processes
in the ipsilateral trigeminal ganglia of all animals
(Fig. 2). Labeled cells were never observed in the contralateral trigeminal ganglia. An average of 143
HRP-labeled neurons were observed in the trigeminal ganglia of the rat (n = 10), 172 in the monkey (n
= 7), 277 in the cat (n = 5), and 449 in the rabbit (n
= 6) (Fig. 3). The mean diameters of the labeled corneal afferent perikarya were 23.2 /im for the rat, 30.0
fim for the rabbit, 32.9 fim for the cat, and 30.8 jum
for the monkey (Fig. 4). The cell frequency distribution of the cat was relatively symmetrical with respect
to the mean (skewness = 0.86); however, the frequency of distributions of the rat, rabbit and monkey
were skewed to the left of the mean (skewness = 1.58,
1.36 and 1.62, respectively).
The labeled sensory neurons in all four species
were located in the ophthalmic region of the ipsilateral trigeminal ganglion (Fig. 5) and were intermingled extensively among other neurons in the ophthalmic region that did not contain reaction product. In
rats, cats and monkeys, the labeled cells were distributed relatively uniformly throughout both the mediolateral and anteroposterior dimensions of the ophthalmic region, except that cells in the ventral part of
the cat and monkey trigeminal ganglia tended to be
more numerous posteriorly. In marked contrast, labeled cells in the rabbit trigeminal ganglion were
much more restricted in their distribution and
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 05/05/2017
464
Vol. 30
INVESTIGATIVE OPHTHALMOLOGY G VISUAL SCIENCE / Morch 1989
Superior Cervical Ganglia
Following HRP-WGA application to the central
cornea, modest numbers of labeled neurons were observed in the ipsilateral SCGs of the rabbits (x = 53)
and cats (x = 48), however, only occasional labeled
neurons were observed in the rats (x = 1.1), and no
labeled SCG cells were observed in the monkeys (Fig.
7). The labeled SCG neurons, when present, were
stellate or spindle shaped and possessed variable
numbers of radiating dendrites (Fig. 8). In all three
species in which labeled cells were observed, the cells
were concentrated in the rostral half of the ganglion;
however, in cats, many cells were also observed in
more caudal regions of the ganglion (Fig. 9).
No labeled neurons were observed in the middle
cervical or stellate ganglia of any of the animals. One
faintly labeled neuron was found in the ciliary ganglion of one cat; all other ciliary ganglia were devoid
of labeled neurons.
Controls
HRP-WGA application to the corneal surfaces of
nondamaged, control eyes of each species labeled 23
and 35 neurons, respectively, in the trigeminal gan-
Fig. 2. Darkfield light micrograph illustrating several HRPWGA-labeled corneal afferent somata (large arrows) in the ophthalmic region of a rabbit trigeminal ganglion, 72 hr after HRPWGA was applied to the central cornea. The labeled cell in the
upper center of the figure gives origin to an axon that bifurcates
(open arrowhead) into a peripheral and a central process.
O
500-
C
bO
400-
formed a well denned, longitudinal column approximately 0.4 mm in width that did not, except near the
ventral surface of the ganglion, extend all the way to
the medial edge of the ganglion (Fig. 5). There was no
strong and consistant evidence for a somatotopic distribution of corneal afferent neurons within the dorsoventral axis of the trigeminal ganglion in any of the
species examined (Fig. 6).
Occasional labeled cells were observed in all four
species within the transitional zone representing the
apprpximate border between the ophthalmic and
maxillary regions of the ganglion (Fig. 5), and it was
difficult on the basis of somata location alone to classify these neurons as belonging to one region of the
ganglion of the other. However, by serial analyses of
adjacent sections, it was determined that the majority
of the "border cells" were connected to fibers that
entered the ophthalmic division of the trigeminal
nerve. In one monkey, however, six neurons gave
origin to peripheral axons that clearly entered the
medial part of the maxillary nerve.
bO
100-
RAT
CAT
MONKEY
Fig. 3. Histogram illustrating the mean number of HRP-WGAlabeled neurons ± SEM observed in the trigeminal ganglia of each
species following tracer application to the central cornea.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 05/05/2017
No.
465
CORNEAL INNERVATION / Morfurr er ol
3
RAT
RABBIT
X=30.8t9.0
X=23.3t5.7
32
3 6 4 0 4 4 4 8
52
5 6 6 0 6 4
68
1 2 1 6 2 0
24 28 32 3 6 4 0 4 4 4 9
52
5 6 6 0 6 4
0)
g
100 -, M O N K E Y
CAT
X=30.8t6.3
X=33.1±7.7
2 0 2 4 2 8 3 2 3 6 4 0 4 4
48
52
5 6 6 0 6 4 6 8
soma diameter (\im)
Fig. 4. Histogram illustrating the size distributions of 500 HRP-WGA-labeled, randomly selected, corneal afferent somata from each of the
four species.
glia of the rat, eight and six cells, respectively, in the
trigeminal ganglia of the rabbit, and one cell in the
trigeminal ganglion of one cat. Labeled cells were
absent from the trigeminal ganglion of the second
control cat, the trigeminal ganglia of the monkeys,
and the SCGs of all control animals.
Discussion
The results of the present investigation provide detailed information on the origins, numbers and afferent somata diameters of the sensory and sympathetic
neurons that innervate the central corneas in four
species of mammals: rat, rabbit, cat and monkey. The
data reported here extend the observations of'dther
workers in this area and provide several important,
new pieces of information, including: (1) the location
of the corneal afferent neurons within the trigeminal
ganglion of the rabbit; (2) somatotopic maps of the
trigeminal ganglia of the rat, cat and monkey that
provide more detail than those published pre-
viously9"13 by virtue of the fact that in the current
study the location of all labeled corneal afferent
neurons, and not just those in representative sections,
have been meticulously plotted; (3) detailed analyses
of the dorsoventral distribution of corneal afferent
neurons in the trigeminal ganglia of all four species,
(4) comprehensive cell size analyses of 500 randomly
selected corneal afferent somata from each species;
and (5) comparative data on the numbers and locations of SCG neurons that provide sympathetic innervation to the mammalian cornea.
Technical Considerations
Specificity of the experimental design: Examination of the TMB-reacted corneal whole mounts and
irides from the experimental animals (eg, Fig. 1) revealed that HRP-WGA applied to the central cornea
remained in the central one-fourth to one-half of the
cornea without spreading into surrounding tissues. If
any tracer diffused into the peripheral cornea, the
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 05/05/2017
466
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / March 1989
Vol. 30
Monkey
Cat
Rabbit
Fig. 5. Schematic illustrations showing the locations
of HRP-WGA-labeled corneal afferent neurons (solid
circles) in the trigeminal
ganglia of the monkey, cat,
rabbit and rat. The figures
are arranged in series with
the most dorsal section to
the left and progressively
more ventral sections to the
right. Each drawing represents a composite of six or
seven consecutive sections
and, taken collectively, each
series of drawings accurately
depicts the total number of
cells seen in that ganglion.
See text for details of cell
distribution. Oph, ophthalmic nerve; max, maxillary
nerve; mand, mandibular
nerve.
Rat
corneoscleral limbus or iris, it did so in quantitities so
small that we were unable to demonstrate its presence
by using the sensitive TMB technique. The absence
of labeled cell bodies from all but one of the ciliary
ganglia examined (and in the latter case, only one
faintly labeled cell was detected) further suggests that
tracer leakage from the cornea into the neighboring
tissues of the iris and ciliary body was negligible.
Thus, it may be concluded that the labeled neurons
observed in the trigeminal and superior cervical gan-
glia of the current investigation resulted predominantly if not solely from the uptake and retrograde
transport of tracer by nerve fibers that terminated in
the central cornea.
Controls: The small number of labeled neurons
observed in some of the trigeminal ganglia of the
control animals were probably labeled via HRPWGA uptake and retrograde transport in corneal afferent fibers that terminated near the epithelial surface. The control-labeled neurons were in all cases
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 05/05/2017
467
CORNEAL INNEIWATION / Morfurr er ol
No. 3
number of cells
0
10
I
1
20
I
L
30
40
i
I
10
I
0-.
20
1
30
40
1
1
J
RAT
» w* ¥
* rt
(ft
IT'
* 1
RABBIT
i
V
X
r
V
i
.•.:••-:.••
m
Y I
,^t,r,M;
05
0.5
',: iI]
J
1
bO
Fig. 6. Histogram illustrating the relatively uniform dorsoventral distribution of the corneal afferent
neurons in the trigeminal
ganglia of the species examined in the current study.
in
1.0-
,,
J
±"4:-:::i
*
J
40
J
10
1
0
CAT
20
I
30
I
40
I
MONKEY
.:.,..;.;^,j,,,
0.5
r
-i i uyi^r n^ni i rj^jfTru i iJTS
"
.i..A.i:.:..M
0.5i
—i—I
1.0
very faintly labeled and their somata diameters and
location within the trigeminal ganglia were identical
to the labeled corneal cells observed in the experimental animals.
Cell counts: The numbers of trigeminal and superior cervical ganglion neurons labeled in the current
investigation undoubtedly represent conservative estimates of the total number of cells that innervate the
mammalian cornea since, as mentioned above, the
tracer applications were restricted to the central cornea and did not include more peripheral areas of the
tissue. However, since the aim of the current investigation was to study the sensory and sympathetic in-
nervation of the cornea, we intentionally restricted
the size of the application area in order to prevent
nonspecific tracer uptake by nerve fibers in the corneoscleral limbus, sclera, ciliary body and iris. Nevertheless, since most large corneal nerve bundles converge radially on the central cornea and the central
cornea is more densely innervated than the peripheral cornea,19 it is likely that HRP-WGA as applied
in this study reached the terminal or preterminal
portions of a large percentage of the nerve fibers that
supply the cornea.
The lack of labeled SCG cells in the monkey experiments merits special comment. It is doubtful that
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 05/05/2017
468
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / March 1989
100'
u
.£
u
T3
60-
%
^3
4oH
M-H
o
a!
RAT
RABBIT
CAT
MONKEY
Fig. 7. Histogram illustrating the mean number of HRP-WGAlabeled neurons ± SEM observed in the superior cervical ganglia of
each species following tracer application to the central cornea.
Vol. 30
We have also shown that corneal afferent neurons
in all of the species examined here are concentrated
in the ophthalmic region of the ipsilateral trigeminal
ganglion.9"13 Similar medial localizations of corneal
neurons have been reported in other mammals, including, mice8 and guinea pigs (Keller et al, personal
communication, 1988). However, the results of this
and other studies provide no strong evidence for the
existence within the ophthalmic region of a refined
somatotopy according to the particular peripheral
target organ innervated. For example, corneal
neurons in cats overlap extensively with neurons that
give rise to the supraorbital10 and ethmoidal20 nerves,
and monkey corneal neurons overlap extensively
with cells that innervate the extraocular muscles.21
Even in the rabbit, in which corneal neurons are
clustered together to a larger degree than in other
species, individual labeled corneal cells are separated
from one another by unlabeled neurons of unknown
peripheral distribution.
A very small number of corneal afferent neurons in
some trigeminal ganglia of cats12 and monkeys13 give
rise to axons that enter the cornea via the maxillary
division of the trigeminal nerve. Theoretically, these
these "negative" data are due to inadequacies of the
HRP-WGA retrograde tracing technique because
HRP-WGA applied to the monkey central cornea
resulted in strong labeling of a large number of
neurons in the trigeminal ganglion (Fig. 3) and of
their central projections to the trigeminal brainstem
nuclear complex.13 In addition, similar methods of
corneal preparation and tracer application produced
labeled cells in both the trigeminal and superior cervical ganglia of all other species of animals used in
this study. Thus, the absence of labeled monkey SCG
cells in the current studies probably reflects a true
lack of sympathetic innervation in this species and
not a "false negative" due to technical insufficiencies.
Trigeminal Ganglion
Cell counts and somatotopic localization: The current study provides evidence that there is a positive
correlation between corneal size and numbers of corneal primary afferent neurons in the trigeminal ganglion. More HRP-labeled cells are found in the trigeminal ganglia of animals with large corneas (eg,
rabbits and cats) than of animals with smaller corneas
(eg, rats), suggesting that larger corneas require
greater numbers of primary afferent neurons to adequately subserve the sensory and trophic needs of the
tissue.
Fig. 8. Darkfield micrograph illustrating two HRP-WGA-labeled neurons (arrows) in the superior cervical ganglion of a rabbit.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 05/05/2017
No. 3
469
CORNEAL INNEIWATION / Morfurr er ol
fibers may provide important sources of preserved
corneal sensitivity following total or subtotal ophthalmic nerve injury, although the available evidence1213
suggests that the contribution made by these fibers to
the overall corneal innervation, when present, is very
small. In contrast to the findings of Morgan and coworkers'2 who found that 4% of all corneal-innervating cells in the cat are located in the maxillary division, we found that all labeled neurons in the trigeminal ganglia of cats were located in the ophthalmic
division. This variability may be the result of differences in the type of tracer used, methods of tracer
application and percentage of the cornea exposed to
the solution.
The results of this study have the revealed that corneal afferent neurons are distributed relatively uniformly in the dorsoventral axis of the trigeminal ganglia of all four species examined. According to one
current theory of trigeminal ganglion somatotopy,
cells that innervate oral and perioral regions of the
face are found in greater numbers in ventral regions
of the trigeminal ganglion, and cells that supply peripheral regions of the head are concentrated dorsally.10'22-23 Thus, the distribution of corneal afferent
neurons observed in this study may reflect the "central" position of this tissue in the orofacial somatotopic map. Interestingly, the uniform dorsoventral
distribution of corneal neurons that we observed in
the rat trigeminal ganglia differs from the pronounced dorsal concentration of corneal cells observed in an earlier study in this species1' and provides evidence for a previously unrecognized interanimal variability as regards the distribution of these
cells.
Corneal afferent somata diameters: Our results
have shown that corneal afferent neurons in all four
species examined constitute a population of predominantly small and medium sized neurons. The mean
diameter of the cat corneal afferent neurons (33.1
(un) is nearly identical to the figure reported by Nishimori and coworkers14 (33.3 nm), however, the
mean diameter of the rat corneal afferent neurons
(23.3 um) is approximately 3-4 fim larger than the
range reported by Sugimoto et al.15 Since the mean
diameter of normal adult rat trigeminal ganglion
neurons is 26.1 nm24 corneal afferent cells in the latter species are among the smaller neurons in the ganglion.
The small sizes of the corneal afferent somata reported here are consistant with electron microscopic
studies showing that the corneal innervation is derived almost exclusively from unmyelinated and
finely myelinated axons.25 Electrophysiological observations have shown that the majority of corneal
sensory axons conduct in the lower A delta fiber
Cat
Monkey
Fig. 9. Schematic illustrations showing the distribution of HRPWGA-labeled neurons (solid circles) in representative superior cervical ganglia from each species.
range, with some cold-sensitive units conducting in
the C fiber range.2627
In all of the animals examined in the current investigation, a small percentage of the labeled somata
were of relatively large size (greater than 40 fim in
diameter), suggesting that they may serve a mechanoreceptive function. Clinical testing suggests that some
tactile sensibility is present in the human eye, but that
it is normally masked by the dominant sensation of
corneal pain. 2829 A definite tactile component of corneal innervation has been described in patients with
trigeminal tractotomies; cutting the spinal tract of V
at the level of the obex in these individuals produces
corneal analgesia while maintaining tactile sensibility.3031 Alternatively, the large diameter corneal afferent neurons may, like the more numerous smaller
diameter neurons, signal some form of corneal pain.
In the latter case, the size of the perikarya may be
correlated with the total volume of the peripheral and
central nerve branches supported by the cell, and not
by the sensory modality subserved by that cell.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 05/05/2017
470
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / March 1989
Sympathetic Innervation of the Cornea
Our results have shown that the cornea of the rat,
rabbit, and cat is innervated by small to moderate
numbers of cells located in the ipsilateral superior
cervical ganglion. These observations confirm and
extend previous reports of corneal sympathetic nerve
fibers in these species as revealed by fluorescent
staining techniques32"35 and HRP-WGA anterograde
transport procedures16 and the resultant disappearance of these fibers following superior cervical ganglionectomy.36 The negative findings reported here for
the monkey also substantiate previous reports of a
lack of a sympathetic innervation to the primate cornea, as evidenced by the inability to label corneal
adrenergic fibers in this species via fluorescent staining 35 - 3738 or neuropeptide Y immunocytochemistry.39
We have also shown that the SCG neurons that
supply the cornea are located primarily in the rostral
half of the ganglion. The mammalian SCG is somatotopically organized; cells that project through the
internal carotid nerve are concentrated near the rostral pole of the ganglion, whereas cells that project
through the external carotid nerve aggregate further
caudally.40"42 However, we found no evidence for somatotopic localization of SCG neurons according to
the particular peripheral target tissue innervated. In
the rat SCG, for example, neurons that supply the
pineal gland,43'44 anterior eye chamber,45 Miiller's
smooth muscle46 and cornea (current investigation)
intermingle extensively with one another in the rostral region of the ganglion.
Based on estimated numbers of neurons in the rat
SCG of 29,000,47 in the rabbit, 76,000,48 and in the
cat, 66,000,49 the number of labeled neurons that
supply the cornea as indicated by our results (Fig. 7)
represent only 0.01%, 0.1% and 0.2%, respectively, of
the total SCG cell population. The data presented
here also show that neither the middle cervical nor
the stellate ganglia contribute to the corneal innervations in the species examined. The latter observation
is compatible with what is known of the organization
of the cervical sympathetic chain ganglia, ie, some
cells in the middle cervical and stellate ganglia send
their axons through the SCG and into the external
carotid nerve to reach target tissues in the head and
face,4150 but only cells in the SCG send their axons
through the internal carotid nerve to supply the orbit.
Functional considerations: The functional significance of the sympathetic innervation of the mammalian cornea is unclear, although several interesting
hypotheses have been proposed. The literature in this
area has recently been reviewed in detail16 and will
only be commented on here briefly. There is a good
Vol. 30
deal of experimental evidence to suggest that corneal
sympathetic nerve fibers in rabbits modulate ion
transport processes in the corneal epithelium,3 and,
therefore, may assist the endothelium in regulating
stromal hydration and corneal transparency. Sympathetic fibers in rats and rabbits have been shown to
exert antimitogenic effects on the corneal epithelium 4 ' 551 and inhibitory effects on corneal wound
healing.6-52 Finally, corneal sympathetic fibers may
play a role in modulating the sensitivity of the cornea
by interacting with the terminal or preterminal portions of trigeminal sensory nerve fibers.733 However,
solid experimental evidence in support of the latter
hypothesis is lacking. In humans with Homer's syndrome (in which the sympathetic outflow from the
SCG to the eye has been interupted), the corneal epithelium of the affected eye shows abnormal function:
epithelial thickness in the Homer's eye is slightly increased, and when subjected to an hypoxic "stress
test" it shows significantly more epithelial greying,
microcytic edema and a slower rate of desweUing.53
The significance of the pronounced interspecies
differences in SCG innervation of the mammalian
cornea as revealed in the current investigation is uncertain and will require additional study. Perhaps animals with very large corneas (eg, rabbits and cats)
require greater numbers of SCG neurons than animals with small corneas (eg, rats) in order to maintain
similar degrees of neural control over target tissues of
disproportionate size. However, this theory is untenable when one considers that the monkey cornea,
intermediate in size between that of the rat and the
rabbit, is totally lacking in sympathetic innervation.
Alternatively, sympathetic fibers may play more
"important" roles in regulating certain aspects of
corneal physiology and metabolism in some species
than in others. However, it is unclear how the corneas
of monkeys and humans, which contain few or no
sympathetic nerve fibers,37'54 differ anatomically,
metabolically or biochemically from the corneas of
rabbits and cats, which contain modest numbers of
sympathetic fibers.
Key words: corneal innervation, trigeminal ganglion, superior cervical ganglion, herpes simplex virus, ocular nerves
Acknowledgments
The authors would like to thank Mr. Mark A. Jones for
excellent technical assistance rendered during the course of
these experiments and Ms. Kathleen Drajus for caring for
the cats used in these experiments.
References
1. Duke-Elder S and Leigh AG: Neurotrophic keratitis. In System
of Ophthalmology, Vol. VIII, Diseases of the outer eye, Part 2,
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 05/05/2017
No. 3
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
CORNEAL INNERVATION / Marfurr er ol
Duke-Elder S and Leigh A, editors. St. Louis, The C.V. Mosby
Co., 1965, pp. 803-810.
Beuerman RW and Schimmelpfennig B: Sensory denervation
of the rabbit cornea affects epithelial properties. Exp Neurol
69:196, 1980.
Klyce SD and Crosson CE: Transport processes across the
rabbit corneal epithelium: A review. Curr Eye Res 4:323, 1985.
Friedenwald JS and Bushke W: The effects of excitement, of
epinephrine and of sympathectomy on the mitotic activity of
the corneal epithelium in rats. Am J Physiol 141:689, 1944.
Mishima S: The effects of the denervation and the stimulation
of the sympathetic and trigeminal nerve on the mitotic rate of
the corneal epithelium in the rabbit. Jpn J Ophthalmol 1:65,
1957.
Perez E, Lopez-Briones L, Gallar J, and Belmonte C: Effects of
chronic sympathetic stimulation on corneal wound healing.
Invest Ophthalmol Vis Sci 28:221, 1987.
Morgan C, de Groat WC, and Jannetta PJ: Sympathetic innervation of the cornea from the superior cervical ganglion. An
HRP study in the cat. J Auton Nerv Syst 20:179, 1987.
Arvidson B: Retrograde axonal transport of horseradish peroxidase from cornea to trigeminal ganglion. Acta Neuropathol
(Berlin) 38:49, 1977.
Morgan CW, Nadelhaft I, and de Groat WC: Anatomical localization of corneal afferent cells in the trigeminal ganglion.
Neurosurgery 2:252, 1978.
Marfurt CF: Somatotopic organization of the cat trigeminal
ganglion as determined by the horseradish peroxidase technique. Anat Rec 201:105, 1981.
Marfurt CF and Del Toro DR: Corneal sensory pathway in the
rat: A horseradish peroxidase tracing study. J Comp Neurol
261:450, 1987.
Morgan C, Jannetta PJ, and deGroat WC: Organization of
corneal afferent axons in the trigeminal nerve root entry zone
in the cat. Exp Brain Res 68:411, 1987.
Marfurt CF and Echtenkamp SF: Central projections and trigeminal ganglion somatotopy of corneal afferent neurons in
the monkey, Macaca fascicularis. J Comp Neurol 272:370,
1988.
Nishimori T, Sera M, Suemune S, Yoshida A, Tsuru K, Tsuiki
Y, Akisaka T, Okamoto T, Dateoka Y, and Shigenaga Y: The
distribution of muscle primary afferents from the masseter
nerve to the trigeminal sensory nuclei. Brain Res 372:375,
1986.
Sugimoto T, Takemura M, and Wakisaka S: Cell size analysis
of primary neurons innervating the cornea and tooth pulp of
the rat. Pain 32:375, 1988.
Marfurt CF: Sympathetic innervation of the rat cornea as demonstrated by the retrograde and anterograde transport of
horseradish peroxidase-wheat germ agglutinin. J Comp
Neurol 268:147, 1988.
Mesulam MM: Tetramethylbenzidine for horseradish peroxidase neurohistochemistry: A non-carcinogenic blue reactionproduct with superior sensitivity for visualizing neural afferents and efferents. J Histochem Cytochem 26:106, 1978.
Abercrombie M: Estimation of nuclear population from microtome sections. Anat Rec 94:239, 1946.
Rozsa AJ and Beuerman RW: Density and organization of free
nerve endings in the corneal epithelium of the rabbit. Pain
14:105, 1982.
Lucier GE and Egizii R: Central projections of the ethmoidal
nerve of the cat as determined by the horseradish peroxidase
tracer technique. J Comp Neurol 247:123, 1986.
Porter JD, Guthrie BL, and Sparks DL: Innervation of monkey extraocular muscles: localization of sensory and motor
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
471
neurons by retrograde transport of horseradish peroxidase. J
Comp Neurol 218:208, 1983.
Beaudreau DE and Jerge CR: Somatotopic representation in
the Gasserian ganglion of tactile peripheral fields in the cat.
Arch Oral Biol 131:247, 1968.
Kerr FWL and Lysak W: Somatotopic organization of trigeminal ganglion neurones. Arch Neurol 11:593, 1964.
Forbes DJ and Welt C: Neurogenesis in the trigeminal ganglion of the albino rat: A quantitative autoradiographic study. J
Comp Neurol 199:133, 1981.
Hoyes AD and Barber P: Ultrastructure of the corneal nerves
in the rat. Cell Tissue Res 172:133, 1976.
Belmonte C and Giraldez F: Responses of cat comeal sensory
receptors to mechanical and thermal stimulation. J Physiol
321:355, 1981.
Tanelian DL and Beuerman RW: Responses of rabbit comeal
nociceptors to mechanical and thermal stimulation. Exp
Neurol 84:165, 1984.
Giraldez F, Geijo E, and Belmonte C: Response characteristics
of corneal sensory fibers to mechanical and thermal stimulation. Pain 7:1, 1979.
Lele PP and Weddell G: Sensory neurons of the cornea and
cutaneous sensibility. Brain 79:119, 1959.
Grant FC, Groff RA, and Lewy FH: Section of the descending
spinal root of the fifth cranial nerve. AMA Arch Neurol Psychiatry 43:498, 1940.
Rowbotham GF: Observations on the effects of trigeminal denervation. Brain 62:364, 1939.
Ehinger B: Distribution of adrenergic nerves to orbital structures. Acta Physiol Scand 62:291, 1964.
33. Laties A and Jacobowitz D: A histochemical study of the
adrenergic and cholinergic innervatidn of the anterior segment
of the rabbit eye. Invest Ophthamol 3:592, 1964.
34. Ehinger B: Distribution of adrenergic nerve in the eye and
some related structures in the cat. Acta Physiol Scand 66:123,
1966.
35. Laties AM and Jacobowitz D: A comparative study of the
autonomic innervation of the eye in monkey, cat, and rabbit.
Anat Rec 156:383, 1966.
36. Tervo T and Palkama A: Sympathetic nerves to the rat cornea.
Acta Ophthalmol (Copenh) 54:75, 1976.
37. Ehinger B: A comparative study of the adrenergic nerves to the
anterior eye segment of some primates. Z Zellforsch 116:157,
1971.
38. Ehinger B and Sj6berg N-O: Development of the ocular
adrenergic neve supply in man and guinea-pig. Z Zellforsch
118:579, 1971.
39. Stone RA, Laties AM, and Emson PC: Neuropeptide Y and
the ocular innervation of rat, guinea pig, cat, and monkey.
Neuroscience 17:1207, 1986.
40. Bowers CW and Zigmond RE: Localization of Neurons in the
rat superior cervical ganglion that project into different postganglionic trunks. J Comp Neurol 185:381, 1979.
41. Dail WG, Khoudary S, Barraza C, Murray HM, and Bradely
C: The fate of adrenergic fibers which enter the superior cervical ganglion. In Histochemistry and Cell Biology of Autonomic Neurons, SIF Cells, and Paraneurons, Eranko O, Soinila S, and Paiva'rinta H, editors. New York, Raven Press, 1980,
pp. 287-297.
42. Jacobowitz D and Woodward JK: Adrenergic neurons in the
cat superior cervical ganglion and cervical sympathetic nerve
trunk. A histochemical study. J Pharmacol Exp Ther 162:213,
1968.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 05/05/2017
472
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / March 1989
43. Bowers CW, Dahm LM, and Zigmond RE: The number and
distribution of sympathetic neurons that innervate the rat pineal gland. Neuroscience 13:87, 1984.
44. Patrickson JW and Smith TE: Innervation of the pineal gland
in the rat: An HRP Study. Exp Neurol 95:207, 1987.
45. Arvidson B: Retrograde transport of horseradish peroxidase in
sensory and adrenergic neurons following injection into the
anterior eye chamber. J Neurocytol 8:751, 1979.
46. Smith PG, Brucker JW, and Mills E: Reinnervation of Muller's
smooth muscle by atypical sympathetic pathways following
neonatal ganglionectomy in the rat: Structural and functional
investigations of enhanced neuroplasticity. Neuroscience
23:781, 1987.
47. Smolen AJ, Wright LL, and Cunningham TJ: Neuron numbers in the superior cervical sympathetic ganglion of the rat: A
critical comparison of methods for cell counting. J Neurocytol
12:739, 1983.
48. Purves D, Rubin E, Snider WD, and Lichtman J: Relation of
animal size to convergence, divergence, and neuronal number
in peripheral sympathetic pathways. J Neurosci 6:156, 1986.
49. Wolf G: The ratio of preganglionic neurons to postganglionic
50.
51.
52.
53.
54.
Vol. 30
neurons in the visceral nervous system. J Comp Neurol
75:235, 1941.
Thomander L, Aldskogius H, and Arvidsson J: Evidence for a
sympathetic component in motor branches of the facial nerve:
A horseradish peroxidase study in the cat. Brain Res 301:380,
1984.
Butterfield LC and Neufeld AH: Cyclic nucleotides and mitosis in the rabbit cornea following superior cervical ganglionectomy. Expfcye Res 25:427, 1977.
Beuerman RW, Klyce SD, Vigo MC, Duruy BM, and Crosson
CE: Modulation of latency and velocity in cornea! epithelial
wound healing. ARVO Abstracts. Invest Ophthalmol Vis Sci
26(Suppl):91, 1985.
Sweeney DF, Vannas A, Holden BA, Tervo T, and Telaranta
T: Evidence for sympathetic neural influence on human corneal epithelial function. Acta Ophthalmol (Copenh) 63:215,
1985.
Toivanen M, Tervo T, Partanen M, Vanna A, and Hervonen
A: Histochemical demonstration of adrenergic nerves in the
stroma of human cornea. Invest Ophthalmol Vis Sci 28:398,
1987.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 05/05/2017