Sensory and sympathetic innervation of the …iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/...Sensory and Sympathetic Innervation of the ... completely filled the area

Investigative Ophthalmology & Visual Science, Vol. 30, No. 3, March 1989Copyright © Association Tor Research in Vision and Ophthalmology

Sensory and Sympathetic Innervation of theMammalian Cornea

A Retrograde Tracing Study

Carl F. Marfurr,* Robert E. Kingsley,t o n c l 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 efferentneurons in four different mammalian species. A 2% solution of HRP-WGA was applied to the centralcornea 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 ophthalmicregion of the ipsilateral trigeminal ganglion of the rat, cat and monkey. In marked contrast, labeledcells in the rabbit trigeminal ganglion were clustered in a sharply defined longitudinal column locatedin the midregion of the ophthalmic area. Occasional cells in some cats and monkeys were observednear, and possibly within, the maxillary region of the ganglion. There was no evidence for a dorsoven-tral somatotopy of corneal afferent neurons in any species. The majority of the labeled afferent somatawere small or medium in size, although some larger diameter neurons were also observed. Modestnumbers of labeled neurons were observed in the ipsilateral superior cervical ganglion (SCG) of therabbit and cat; however, only occasional labeled neurons were observed in the SCG of the rat, and nonewere seen in the monkey. The labeled SCG cells, when present, were concentrated in the rostral half ofthe ganglion, although many cells in the cat SCG were also found further caudally. No labeled neuronswere found in the middle cervical or stellate ganglia of any animal. The results of this study haverevealed the existence of subtle interspecies differences in the organization of the mammalian cornealafferent and efferent innervations. Invest Ophthalmol Vis Sci 30:461-472,1989

The mammalian cornea is richly supplied by sen-sory nerve fibers from the trigeminal ganglion (TG)and, to a lesser extent, by sympathetic fibers from thesuperior cervical ganglion (SCG). Both populationsof nerve fibers play unique roles in maintaining theoverall structural and functional integrity of the cor-nea. Corneal afferent nerves, in addition to signallingsensory information, exert important trophic influ-ences on the corneal epithelium1 and stimulatewound healing following epithelial lesions.2 Cornealsympathetic nerves, in turn, regulate epithelial iontransport processes,3 exert inhibitory influences onepithelial mitogenesis and wound healing4"6 and may

From the Departments of *Anatomy and ^Physiology, North-west Center for Medical Education, Indiana University School ofMedicine, Gary, Indiana, and the fDepartment of Physiology,South Bend Center for Medical Education, Indiana UniversitySchool of Medicine, University of Notre Dame, Notre Dame, In-diana.

Supported in part by PHS grants EY-05717 to CFM andHL37223 to SFE.

Submitted for publication: June 24, 1988; accepted October 5,1988.

Reprint requests: Dr. Carl F. Marfurt, Northwest Center forMedical Education, 3400 Broadway, Gary, IN 46408.

modify corneal sensitivity.7 The mechanism bywhich corneal nerves exert these influences is un-known, although the release of axonally transported,biologically active substances such as peptides andneurotransmitters may be involved.

Information about the cell bodies that give originto these different components of the corneal innerva-tion is limited. The corneal sensory fibers in mice,rats, cats and monkeys originate from neurons lo-cated, with rare exceptions, in the ophthalmic (me-dial) region of the ipsilateral trigeminal ganglion.8"13

However, the distribution of corneal afferent somatain the rabbit, a popular animal model in ophthalmo-logical research, has not been described. In addition,there is little information concerning the relativenumbers of sensory cells that contribute to the cor-neal innervation in different species. While it iswidely reported that comeal afferent perikarya aresmall or medium in size, the cell size spectra for theseneurons'415 in most species remain unknown.Knowledge of the origin(s) of the corneal sympatheticinnervation is even more fragmentary,716 and analy-sis of the limited available data suggests that theremay be significant interspecies differences.

461

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933147/ on 07/08/2018

462 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / March 1989 Vol. 30

In the current investigation, we have used themethod of retrograde transport of horseradish perox-idase-wheat germ agglutinin (HRP-WGA) to studythe afferent and efferent corneal innervation of fourmammalian species commonly used in ophthalmo-logical research: the rat, rabbit, cat and monkey. Thespecific goals of the current study were to determine:(1) the distribution within the trigeminal and superiorcervical ganglia of corneal sensory and sympatheticnerve cell bodies; (2) the relative numbers of trigemi-nal and SCG cells that contribute to the corneal in-nervation in different species; and (3) the somata di-ameters of the corneal afferent neurons. The datagenerated from this study will increase our under-standing of the anatomy of the mammalian cornealinnervation and will provide a useful database of in-formation for future studies by ourselves and otherworkers in such areas as nerve cell loss following cor-neal wounding, the development of experimental an-imal models of herpes virus infections, and the role ofthe sympathetic nervous system in the regulation ofcorneal physiology.

Materials and Methods

A total of ten rats, six rabbits, four cats and fourmonkeys were used in this investigation. Some of thedata from the rat and monkey experiments have beenpresented in preliminary fashion as parts of largerstudies on corneal sympathetic innervation in therat16 and brainstem termination sites of corneal affer-ent neurons in monkeys.13 AH experiments were per-formed in accordance with the guidelines outlined inthe ARVO Resolution on the Use of Animals in Re-search.

Each animal was anesthetized with ketamine (10mg/kg, I.M.) or sodium pentobarbital (35-45 mg/kg,I.P. or I.V.), and the central area of the corneal epi-thelium was gently scratched in a checkerboard fash-ion with a sterile number 11 scalpel blade. Partialdisruption of the epithelium was found in prelimi-nary experiments to be necessary in order to permitsatisfactory penetration of the tracer substance intothe cornea and to obtain good retrograde transportwithin the corneal nerve fibers. The scratched areawas in every case restricted to the central one-fourthto one-half of the cornea and never involved the pe-ripheral cornea or the coraeoscleral limbus. The pre-pared corneal areas in cats were, on the average,smaller than those in the other species due to thetendency for the cat eyeballs to be drawn back intothe orbit by contraction of the retractor bulbi mus-cles. The procedure was performed unilaterally in atleast two animals from each species, and bilaterally inthe remaining animals. '

In the rabbits, cats and monkeys, the scratched areawas encircled with a 2 mm high vaseline wall con-structed by squeezing a thin strand of vaselinethrough a 25-gauge syringe needle. Thin wafers ofgelfoam surgical sponge were then saturated with 5-8n\ of 2% HRP-WGA in saline, placed on thescratched corneal surface, and manipulated until theycompletely filled the area inside the vaseline retainingwall. The preparation was continuously monitoredthrough a dissecting microscope for 15-30 min, afterwhich the gelfoam pledgets were removed and thecorneal surface was rinsed for 2-3 min with sterilesaline. The eyes were then gently dabbed with cottonapplicator sticks and the eyelids were closed withbutterfly bandages or sutures. In the rats, the methodof tracer application was similar to that used in thelarger animals, except that a modified corneal tre-phine with an inside diameter slightly larger than thatof the scratched corneal surface was used to hold theHRP-WGA solution in place. Postoperative recoveryfor all of the animals was unremarkable and dailymonitoring of their progress revealed no evidence ofepiphora, photophobia or ocular irritation of anysort.

At least one eye from one animal of each specieswas used as a control to check the specificity of theexperimental design and to test for possible leakage oftracer away from the central corneal application site.In the latter cases, 3-5 n\ of 2% HRP-WGA wasapplied in dropwise fashion to the central areas ofnormal, nondamaged corneas and was permitted toflow freely over the entire cornea and surroundingextraocular tissues. Thirty minutes later, the tracerwas removed by rinsing the eye repeatedly with sterilesaline, dabbing it gently with cotton applicator swabs,and closing the eyelids with butterfly bandages or su-tures as described above.

Forty-eight to 96 hr later, all animals were reanes-thetized and perfused through the left ventricle orretrogradely through the descending aorta with200-500 ml of warm, phosphate-buffered saline, pH7.4, containing 5000 units/kg heparin and 3% pro-caine-HCl, followed in sequence by 1-2 1 of warmfixative containing 1% paraformaldehyde-2% glutar-aldehyde in 0.1 M phosphate buffer, and 1 1 of ice-cold 0.1 M phosphate buffer with 10% sucrose. Thetrigeminal ganglia, superior cervical ganglia, corneasand irides were immediately removed and placed infresh 0.1 M phosphate buffer. In at least two animalsof each species, the ciliary, middle cervical and stel-late ganglia were also removed. Sections through eachganglion were cut serially at 40 pm in a cryostat,mounted directly on chrome alum-gelatin-coatedslides, air-dried and reacted for the presence of HRPactivity according to the tetramethylbenzidine


No. 3 CORNEAL INNERVATION / Morfurr er al 463

(TMB) procedure of Mesulam.17 All tissues were thencarefully examined with darkfield optics for the .pres-ence of HRP labeled neurons.

The corneas and irides from every animal were alsoexamined in order to observe the application sites.Rat and monkey corneas and the irides of all fourspecies were reacted whole with TMB histochemistryas free-floating preparations; the thicker cat and rab-bit corneas were split along natural cleavage planes inthe corneal stroma into two or three layers prior toTMB processing to permit better penetration of theincubation media components. Four radial slits weremade in each tissue and they were then flat-mountedon chrome alum-gelatin-coated slides and air-driedunder weighted coverslips.

The total number of labeled neurons in each gan-glion was determined by counting labeled somatacontaining visible nuclei. Other cells were alsocounted if their maximum profiles were containedentirely within the thickness of a section but were soheavily filled with reaction product that the cell nu-cleus was obscured. Corrections for double-countingof split neurons in the trigeminal ganglia were calcu-lated by using the formula of Abercrombie,18 how-ever, double-counting of the relatively small numbersof SCG neurons was not a problem since it was possi-ble to identify the same cell in adjacent serial sections.The locations of every HRP-labeled neuron in threerandomly selected trigeminal and superior cervicalganglia from each species were carefully plotted ontoline drawings made by using a drawing tube attachedto the microscope. Measurements were determinedfor 500 HRP-labeled trigeminal ganglion neurons ineach species by using an ocular micrometer and aver-aging the maximum and minimum cell diametersthrough the equatorial region of each cell. To insurerandomness of sample, cell diameters were obtainedfrom at least four different ganglia for each species,and within each ganglion the labeled cells were mea-sured in every third section to eliminate the possibil-ity of measuring the same cell twice in adjacent sec-tions.

ResultsCorneal Injection Sites

Examination of the corneas and irides from theexperimental animals revealed a dense staining of theepithelium and stroma of the central scarified cornea,with no reaction product in the peripheral cornea,limbus or iris (Fig. 1). The percentage of the cornealarea stained with TMB reaction product varied fromanimal to animal, but in general covered the centralone-quarter to one-half of the cornea, and was, on theaverage, slightly less in the cats than in the other spe-

Fig. 1. Brightfield light micrograph of a rabbit corneal wholemount illustrating the central location and restricted size of a typi-cal HRP-WGA application site (arrow). The application site in thisparticular animal covers approximately 48.5% of the total cornealarea.

Trigeminal Ganglia

HRP-WGA application to the central cornea ofthe experimental animals resulted in the productionof numerous well labeled somata and fiber processesin the ipsilateral trigeminal ganglia of all animals(Fig. 2). Labeled cells were never observed in the con-tralateral trigeminal ganglia. An average of 143HRP-labeled neurons were observed in the trigemi-nal 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 cor-neal afferent perikarya were 23.2 /im for the rat, 30.0fim for the rabbit, 32.9 fim for the cat, and 30.8 jumfor the monkey (Fig. 4). The cell frequency distribu-tion of the cat was relatively symmetrical with respectto the mean (skewness = 0.86); however, the fre-quency of distributions of the rat, rabbit and monkeywere skewed to the left of the mean (skewness = 1.58,1.36 and 1.62, respectively).

The labeled sensory neurons in all four specieswere located in the ophthalmic region of the ipsilat-eral trigeminal ganglion (Fig. 5) and were intermin-gled extensively among other neurons in the ophthal-mic region that did not contain reaction product. Inrats, cats and monkeys, the labeled cells were distrib-uted relatively uniformly throughout both the medio-lateral and anteroposterior dimensions of the oph-thalmic region, except that cells in the ventral part ofthe cat and monkey trigeminal ganglia tended to bemore numerous posteriorly. In marked contrast, la-beled cells in the rabbit trigeminal ganglion weremuch more restricted in their distribution and


464 INVESTIGATIVE OPHTHALMOLOGY G VISUAL SCIENCE / Morch 1989 Vol. 30

Fig. 2. Darkfield light micrograph illustrating several HRP-WGA-labeled corneal afferent somata (large arrows) in the oph-thalmic region of a rabbit trigeminal ganglion, 72 hr after HRP-WGA was applied to the central cornea. The labeled cell in theupper center of the figure gives origin to an axon that bifurcates(open arrowhead) into a peripheral and a central process.

formed a well denned, longitudinal column approxi-mately 0.4 mm in width that did not, except near theventral surface of the ganglion, extend all the way tothe medial edge of the ganglion (Fig. 5). There was nostrong and consistant evidence for a somatotopic dis-tribution of corneal afferent neurons within the dor-soventral axis of the trigeminal ganglion in any of thespecies examined (Fig. 6).

Occasional labeled cells were observed in all fourspecies within the transitional zone representing theapprpximate border between the ophthalmic andmaxillary regions of the ganglion (Fig. 5), and it wasdifficult on the basis of somata location alone to clas-sify these neurons as belonging to one region of theganglion of the other. However, by serial analyses ofadjacent sections, it was determined that the majorityof the "border cells" were connected to fibers thatentered the ophthalmic division of the trigeminalnerve. In one monkey, however, six neurons gaveorigin to peripheral axons that clearly entered themedial part of the maxillary nerve.

Superior Cervical Ganglia

Following HRP-WGA application to the centralcornea, modest numbers of labeled neurons were ob-served in the ipsilateral SCGs of the rabbits (x = 53)and cats (x = 48), however, only occasional labeledneurons were observed in the rats (x = 1.1), and nolabeled SCG cells were observed in the monkeys (Fig.7). The labeled SCG neurons, when present, werestellate or spindle shaped and possessed variablenumbers of radiating dendrites (Fig. 8). In all threespecies in which labeled cells were observed, the cellswere concentrated in the rostral half of the ganglion;however, in cats, many cells were also observed inmore caudal regions of the ganglion (Fig. 9).

No labeled neurons were observed in the middlecervical or stellate ganglia of any of the animals. Onefaintly labeled neuron was found in the ciliary gan-glion of one cat; all other ciliary ganglia were devoidof labeled neurons.

Controls

HRP-WGA application to the corneal surfaces ofnondamaged, control eyes of each species labeled 23and 35 neurons, respectively, in the trigeminal gan-

O 500-

C

bO

bO

4 0 0 -

100-

RAT CAT MONKEY

Fig. 3. Histogram illustrating the mean number of HRP-WGA-labeled neurons ± SEM observed in the trigeminal ganglia of eachspecies following tracer application to the central cornea.


No. 3 CORNEAL INNERVATION / Morfurr er ol 465

RAT

0)

g

X=23.3t5.7

RABBIT

32 3 6 4 0 4 4 4 8 52 5 6 6 0 6 4 68

CAT

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

X=30.8t9.0

1 2 1 6 2 0 2 4 2 8 3 2 3 6 4 0 4 4 4 9 5 2 5 6 6 0 6 4

100 -, MONKEY

X=30.8t6.3

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 thetrigeminal ganglia of the rabbit, and one cell in thetrigeminal ganglion of one cat. Labeled cells wereabsent from the trigeminal ganglion of the secondcontrol cat, the trigeminal ganglia of the monkeys,and the SCGs of all control animals.

Discussion

The results of the present investigation provide de-tailed information on the origins, numbers and affer-ent somata diameters of the sensory and sympatheticneurons that innervate the central corneas in fourspecies of mammals: rat, rabbit, cat and monkey. Thedata reported here extend the observations of'dtherworkers in this area and provide several important,new pieces of information, including: (1) the locationof the corneal afferent neurons within the trigeminalganglion of the rabbit; (2) somatotopic maps of thetrigeminal ganglia of the rat, cat and monkey thatprovide more detail than those published pre-

viously9"13 by virtue of the fact that in the currentstudy the location of all labeled corneal afferentneurons, and not just those in representative sections,have been meticulously plotted; (3) detailed analysesof the dorsoventral distribution of corneal afferentneurons in the trigeminal ganglia of all four species,(4) comprehensive cell size analyses of 500 randomlyselected corneal afferent somata from each species;and (5) comparative data on the numbers and loca-tions of SCG neurons that provide sympathetic in-nervation to the mammalian cornea.

Technical Considerations

Specificity of the experimental design: Examina-tion of the TMB-reacted corneal whole mounts andirides from the experimental animals (eg, Fig. 1) re-vealed that HRP-WGA applied to the central cornearemained in the central one-fourth to one-half of thecornea without spreading into surrounding tissues. Ifany tracer diffused into the peripheral cornea, the



Monkey

Cat

Rabbit

Fig. 5. Schematic illustra-tions showing the locationsof HRP-WGA-labeled cor-neal afferent neurons (solidcircles) in the trigeminalganglia of the monkey, cat,rabbit and rat. The figuresare arranged in series withthe most dorsal section tothe left and progressivelymore ventral sections to theright. Each drawing repre-sents a composite of six orseven consecutive sectionsand, taken collectively, eachseries of drawings accuratelydepicts the total number ofcells seen in that ganglion.See text for details of celldistribution. Oph, ophthal-mic nerve; max, maxillarynerve; mand, mandibularnerve.

Rat

corneoscleral limbus or iris, it did so in quantitities sosmall that we were unable to demonstrate its presenceby using the sensitive TMB technique. The absenceof labeled cell bodies from all but one of the ciliaryganglia examined (and in the latter case, only onefaintly labeled cell was detected) further suggests thattracer leakage from the cornea into the neighboringtissues of the iris and ciliary body was negligible.Thus, it may be concluded that the labeled neuronsobserved in the trigeminal and superior cervical gan-

glia of the current investigation resulted predomi-nantly if not solely from the uptake and retrogradetransport of tracer by nerve fibers that terminated inthe central cornea.

Controls: The small number of labeled neuronsobserved in some of the trigeminal ganglia of thecontrol animals were probably labeled via HRP-WGA uptake and retrograde transport in corneal af-ferent fibers that terminated near the epithelial sur-face. The control-labeled neurons were in all cases


No. 3 CORNEAL INNEIWATION / Morfurr er ol 467

number of cells

Fig. 6. Histogram illus-trating the relatively uni-form dorsoventral distribu-tion of the corneal afferentneurons in the trigeminalganglia of the species exam-ined in the current study.

bO

10 20 30 40I I i I

0 1 L

0.5

1.0-

RAT

0.5

1.0

0 - .

0 5

i n

10I

» w* ¥ V

* rt (ft * 1 X

IT' V

1

J

201

ir

i m

,^t,r,M;

',: i

J

*

30 401 1

JRABBIT

. • . : • • - : . • • Y I

I]

, , J

±"4:-:::i

40J

rCAT

.i..A.i:.:..M

i

—i—I

10 20 30 400 1 I I I

0.5-

MONKEY

.:.,..;.;^,j,,,

-i i uyi^r n̂ ni i rj^jfTru i iJTS "

very faintly labeled and their somata diameters andlocation within the trigeminal ganglia were identicalto the labeled corneal cells observed in the experi-mental animals.

Cell counts: The numbers of trigeminal and supe-rior cervical ganglion neurons labeled in the currentinvestigation undoubtedly represent conservative es-timates of the total number of cells that innervate themammalian cornea since, as mentioned above, thetracer applications were restricted to the central cor-nea and did not include more peripheral areas of thetissue. However, since the aim of the current investi-gation was to study the sensory and sympathetic in-

nervation of the cornea, we intentionally restrictedthe size of the application area in order to preventnonspecific tracer uptake by nerve fibers in the cor-neoscleral limbus, sclera, ciliary body and iris. Never-theless, since most large corneal nerve bundles con-verge radially on the central cornea and the centralcornea is more densely innervated than the periph-eral cornea,19 it is likely that HRP-WGA as appliedin this study reached the terminal or preterminalportions of a large percentage of the nerve fibers thatsupply the cornea.

The lack of labeled SCG cells in the monkey ex-periments merits special comment. It is doubtful that



100'

u.£

uT3

%

^3 4oHM-Hoa!

6 0 -

RAT RABBIT CAT MONKEY

Fig. 7. Histogram illustrating the mean number of HRP-WGA-labeled neurons ± SEM observed in the superior cervical ganglia ofeach species following tracer application to the central cornea.

these "negative" data are due to inadequacies of theHRP-WGA retrograde tracing technique becauseHRP-WGA applied to the monkey central cornearesulted in strong labeling of a large number ofneurons in the trigeminal ganglion (Fig. 3) and oftheir central projections to the trigeminal brainstemnuclear complex.13 In addition, similar methods ofcorneal preparation and tracer application producedlabeled cells in both the trigeminal and superior cer-vical ganglia of all other species of animals used inthis study. Thus, the absence of labeled monkey SCGcells in the current studies probably reflects a truelack of sympathetic innervation in this species andnot a "false negative" due to technical insufficiencies.

Trigeminal Ganglion

Cell counts and somatotopic localization: The cur-rent study provides evidence that there is a positivecorrelation between corneal size and numbers of cor-neal primary afferent neurons in the trigeminal gan-glion. More HRP-labeled cells are found in the tri-geminal ganglia of animals with large corneas (eg,rabbits and cats) than of animals with smaller corneas(eg, rats), suggesting that larger corneas requiregreater numbers of primary afferent neurons to ade-quately subserve the sensory and trophic needs of thetissue.

We have also shown that corneal afferent neuronsin all of the species examined here are concentratedin the ophthalmic region of the ipsilateral trigeminalganglion.9"13 Similar medial localizations of cornealneurons have been reported in other mammals, in-cluding, mice8 and guinea pigs (Keller et al, personalcommunication, 1988). However, the results of thisand other studies provide no strong evidence for theexistence within the ophthalmic region of a refinedsomatotopy according to the particular peripheraltarget organ innervated. For example, cornealneurons in cats overlap extensively with neurons thatgive rise to the supraorbital10 and ethmoidal20 nerves,and monkey corneal neurons overlap extensivelywith cells that innervate the extraocular muscles.21

Even in the rabbit, in which corneal neurons areclustered together to a larger degree than in otherspecies, individual labeled corneal cells are separatedfrom one another by unlabeled neurons of unknownperipheral distribution.

A very small number of corneal afferent neurons insome trigeminal ganglia of cats12 and monkeys13 giverise to axons that enter the cornea via the maxillarydivision of the trigeminal nerve. Theoretically, these

Fig. 8. Darkfield micrograph illustrating two HRP-WGA-la-beled neurons (arrows) in the superior cervical ganglion of a rabbit.


No. 3 CORNEAL INNEIWATION / Morfurr er ol 469

fibers may provide important sources of preservedcorneal sensitivity following total or subtotal ophthal-mic nerve injury, although the available evidence1213

suggests that the contribution made by these fibers tothe overall corneal innervation, when present, is verysmall. In contrast to the findings of Morgan and co-workers'2 who found that 4% of all corneal-innervat-ing cells in the cat are located in the maxillary divi-sion, we found that all labeled neurons in the trigemi-nal ganglia of cats were located in the ophthalmicdivision. This variability may be the result of differ-ences in the type of tracer used, methods of tracerapplication and percentage of the cornea exposed tothe solution.

The results of this study have the revealed that cor-neal afferent neurons are distributed relatively uni-formly in the dorsoventral axis of the trigeminal gan-glia of all four species examined. According to onecurrent theory of trigeminal ganglion somatotopy,cells that innervate oral and perioral regions of theface are found in greater numbers in ventral regionsof the trigeminal ganglion, and cells that supply pe-ripheral regions of the head are concentrated dor-sally.10'22-23 Thus, the distribution of corneal afferentneurons observed in this study may reflect the "cen-tral" position of this tissue in the orofacial somato-topic map. Interestingly, the uniform dorsoventraldistribution of corneal neurons that we observed inthe rat trigeminal ganglia differs from the pro-nounced dorsal concentration of corneal cells ob-served in an earlier study in this species1' and pro-vides evidence for a previously unrecognized interan-imal variability as regards the distribution of thesecells.

Corneal afferent somata diameters: Our resultshave shown that corneal afferent neurons in all fourspecies examined constitute a population of predomi-nantly small and medium sized neurons. The meandiameter of the cat corneal afferent neurons (33.1(un) is nearly identical to the figure reported by Ni-shimori and coworkers14 (33.3 nm), however, themean diameter of the rat corneal afferent neurons(23.3 um) is approximately 3-4 fim larger than therange reported by Sugimoto et al.15 Since the meandiameter of normal adult rat trigeminal ganglionneurons is 26.1 nm24 corneal afferent cells in the lat-ter species are among the smaller neurons in the gan-glion.

The small sizes of the corneal afferent somata re-ported here are consistant with electron microscopicstudies showing that the corneal innervation is de-rived almost exclusively from unmyelinated andfinely myelinated axons.25 Electrophysiological ob-servations have shown that the majority of cornealsensory axons conduct in the lower A delta fiber

Cat Monkey

Fig. 9. Schematic illustrations showing the distribution of HRP-WGA-labeled neurons (solid circles) in representative superior cer-vical ganglia from each species.

range, with some cold-sensitive units conducting inthe C fiber range.2627

In all of the animals examined in the current in-vestigation, a small percentage of the labeled somatawere of relatively large size (greater than 40 fim indiameter), suggesting that they may serve a mechano-receptive function. Clinical testing suggests that sometactile sensibility is present in the human eye, but thatit is normally masked by the dominant sensation ofcorneal pain.2829 A definite tactile component of cor-neal innervation has been described in patients withtrigeminal tractotomies; cutting the spinal tract of Vat the level of the obex in these individuals producescorneal analgesia while maintaining tactile sensibil-ity.3031 Alternatively, the large diameter corneal af-ferent neurons may, like the more numerous smallerdiameter neurons, signal some form of corneal pain.In the latter case, the size of the perikarya may becorrelated with the total volume of the peripheral andcentral nerve branches supported by the cell, and notby the sensory modality subserved by that cell.



Sympathetic Innervation of the Cornea

Our results have shown that the cornea of the rat,rabbit, and cat is innervated by small to moderatenumbers of cells located in the ipsilateral superiorcervical ganglion. These observations confirm andextend previous reports of corneal sympathetic nervefibers in these species as revealed by fluorescentstaining techniques32"35 and HRP-WGA anterogradetransport procedures16 and the resultant disappear-ance of these fibers following superior cervical gangli-onectomy.36 The negative findings reported here forthe monkey also substantiate previous reports of alack of a sympathetic innervation to the primate cor-nea, as evidenced by the inability to label cornealadrenergic fibers in this species via fluorescent stain-ing35-3738 or neuropeptide Y immunocytochem-istry.39

We have also shown that the SCG neurons thatsupply the cornea are located primarily in the rostralhalf of the ganglion. The mammalian SCG is soma-totopically organized; cells that project through theinternal carotid nerve are concentrated near the ros-tral pole of the ganglion, whereas cells that projectthrough the external carotid nerve aggregate furthercaudally.40"42 However, we found no evidence for so-matotopic localization of SCG neurons according tothe particular peripheral target tissue innervated. Inthe rat SCG, for example, neurons that supply thepineal gland,43'44 anterior eye chamber,45 Miiller'ssmooth muscle46 and cornea (current investigation)intermingle extensively with one another in the ros-tral region of the ganglion.

Based on estimated numbers of neurons in the ratSCG of 29,000,47 in the rabbit, 76,000,48 and in thecat, 66,000,49 the number of labeled neurons thatsupply the cornea as indicated by our results (Fig. 7)represent only 0.01%, 0.1% and 0.2%, respectively, ofthe total SCG cell population. The data presentedhere also show that neither the middle cervical northe stellate ganglia contribute to the corneal innerva-tions in the species examined. The latter observationis compatible with what is known of the organizationof the cervical sympathetic chain ganglia, ie, somecells in the middle cervical and stellate ganglia sendtheir axons through the SCG and into the externalcarotid nerve to reach target tissues in the head andface,4150 but only cells in the SCG send their axonsthrough the internal carotid nerve to supply the orbit.

Functional considerations: The functional signifi-cance of the sympathetic innervation of the mamma-lian cornea is unclear, although several interestinghypotheses have been proposed. The literature in thisarea has recently been reviewed in detail16 and willonly be commented on here briefly. There is a good

deal of experimental evidence to suggest that cornealsympathetic nerve fibers in rabbits modulate iontransport processes in the corneal epithelium,3 and,therefore, may assist the endothelium in regulatingstromal hydration and corneal transparency. Sympa-thetic fibers in rats and rabbits have been shown toexert antimitogenic effects on the corneal epithe-lium4'551 and inhibitory effects on corneal woundhealing.6-52 Finally, corneal sympathetic fibers mayplay a role in modulating the sensitivity of the corneaby interacting with the terminal or preterminal por-tions of trigeminal sensory nerve fibers.733 However,solid experimental evidence in support of the latterhypothesis is lacking. In humans with Homer's syn-drome (in which the sympathetic outflow from theSCG to the eye has been interupted), the corneal epi-thelium of the affected eye shows abnormal function:epithelial thickness in the Homer's eye is slightly in-creased, and when subjected to an hypoxic "stresstest" it shows significantly more epithelial greying,microcytic edema and a slower rate of desweUing.53

The significance of the pronounced interspeciesdifferences in SCG innervation of the mammaliancornea as revealed in the current investigation is un-certain and will require additional study. Perhaps an-imals with very large corneas (eg, rabbits and cats)require greater numbers of SCG neurons than ani-mals with small corneas (eg, rats) in order to maintainsimilar degrees of neural control over target tissues ofdisproportionate size. However, this theory is unten-able when one considers that the monkey cornea,intermediate in size between that of the rat and therabbit, is totally lacking in sympathetic innervation.Alternatively, sympathetic fibers may play more"important" roles in regulating certain aspects ofcorneal physiology and metabolism in some speciesthan in others. However, it is unclear how the corneasof monkeys and humans, which contain few or nosympathetic nerve fibers,37'54 differ anatomically,metabolically or biochemically from the corneas ofrabbits and cats, which contain modest numbers ofsympathetic fibers.

Key words: corneal innervation, trigeminal ganglion, supe-rior cervical ganglion, herpes simplex virus, ocular nerves

Acknowledgments

The authors would like to thank Mr. Mark A. Jones forexcellent technical assistance rendered during the course ofthese experiments and Ms. Kathleen Drajus for caring forthe cats used in these experiments.

References

1. Duke-Elder S and Leigh AG: Neurotrophic keratitis. In Systemof Ophthalmology, Vol. VIII, Diseases of the outer eye, Part 2,


No. 3 CORNEAL INNERVATION / Marfurr er ol 471

Duke-Elder S and Leigh A, editors. St. Louis, The C.V. MosbyCo., 1965, pp. 803-810.

2. Beuerman RW and Schimmelpfennig B: Sensory denervationof the rabbit cornea affects epithelial properties. Exp Neurol69:196, 1980.

3. Klyce SD and Crosson CE: Transport processes across therabbit corneal epithelium: A review. Curr Eye Res 4:323, 1985.

4. Friedenwald JS and Bushke W: The effects of excitement, ofepinephrine and of sympathectomy on the mitotic activity ofthe corneal epithelium in rats. Am J Physiol 141:689, 1944.

5. Mishima S: The effects of the denervation and the stimulationof the sympathetic and trigeminal nerve on the mitotic rate ofthe corneal epithelium in the rabbit. Jpn J Ophthalmol 1:65,1957.

6. Perez E, Lopez-Briones L, Gallar J, and Belmonte C: Effects ofchronic sympathetic stimulation on corneal wound healing.Invest Ophthalmol Vis Sci 28:221, 1987.

7. Morgan C, de Groat WC, and Jannetta PJ: Sympathetic in-nervation of the cornea from the superior cervical ganglion. AnHRP study in the cat. J Auton Nerv Syst 20:179, 1987.

8. Arvidson B: Retrograde axonal transport of horseradish per-oxidase from cornea to trigeminal ganglion. Acta Neuropathol(Berlin) 38:49, 1977.

9. Morgan CW, Nadelhaft I, and de Groat WC: Anatomical lo-calization of corneal afferent cells in the trigeminal ganglion.Neurosurgery 2:252, 1978.

10. Marfurt CF: Somatotopic organization of the cat trigeminalganglion as determined by the horseradish peroxidase tech-nique. Anat Rec 201:105, 1981.

11. Marfurt CF and Del Toro DR: Corneal sensory pathway in therat: A horseradish peroxidase tracing study. J Comp Neurol261:450, 1987.

12. Morgan C, Jannetta PJ, and deGroat WC: Organization ofcorneal afferent axons in the trigeminal nerve root entry zonein the cat. Exp Brain Res 68:411, 1987.

13. Marfurt CF and Echtenkamp SF: Central projections and tri-geminal ganglion somatotopy of corneal afferent neurons inthe monkey, Macaca fascicularis. J Comp Neurol 272:370,1988.

14. Nishimori T, Sera M, Suemune S, Yoshida A, Tsuru K, TsuikiY, Akisaka T, Okamoto T, Dateoka Y, and Shigenaga Y: Thedistribution of muscle primary afferents from the masseternerve to the trigeminal sensory nuclei. Brain Res 372:375,1986.

15. Sugimoto T, Takemura M, and Wakisaka S: Cell size analysisof primary neurons innervating the cornea and tooth pulp ofthe rat. Pain 32:375, 1988.

16. Marfurt CF: Sympathetic innervation of the rat cornea as dem-onstrated by the retrograde and anterograde transport ofhorseradish peroxidase-wheat germ agglutinin. J CompNeurol 268:147, 1988.

17. Mesulam MM: Tetramethylbenzidine for horseradish peroxi-dase neurohistochemistry: A non-carcinogenic blue reaction-product with superior sensitivity for visualizing neural affer-ents and efferents. J Histochem Cytochem 26:106, 1978.

18. Abercrombie M: Estimation of nuclear population from mi-crotome sections. Anat Rec 94:239, 1946.

19. Rozsa AJ and Beuerman RW: Density and organization of freenerve endings in the corneal epithelium of the rabbit. Pain14:105, 1982.

20. Lucier GE and Egizii R: Central projections of the ethmoidalnerve of the cat as determined by the horseradish peroxidasetracer technique. J Comp Neurol 247:123, 1986.

21. Porter JD, Guthrie BL, and Sparks DL: Innervation of mon-key extraocular muscles: localization of sensory and motor

neurons by retrograde transport of horseradish peroxidase. JComp Neurol 218:208, 1983.

22. Beaudreau DE and Jerge CR: Somatotopic representation inthe Gasserian ganglion of tactile peripheral fields in the cat.Arch Oral Biol 131:247, 1968.

23. Kerr FWL and Lysak W: Somatotopic organization of trigemi-nal ganglion neurones. Arch Neurol 11:593, 1964.

24. Forbes DJ and Welt C: Neurogenesis in the trigeminal gan-glion of the albino rat: A quantitative autoradiographic study. JComp Neurol 199:133, 1981.

25. Hoyes AD and Barber P: Ultrastructure of the corneal nervesin the rat. Cell Tissue Res 172:133, 1976.

26. Belmonte C and Giraldez F: Responses of cat comeal sensoryreceptors to mechanical and thermal stimulation. J Physiol321:355, 1981.

27. Tanelian DL and Beuerman RW: Responses of rabbit comealnociceptors to mechanical and thermal stimulation. ExpNeurol 84:165, 1984.

28. Giraldez F, Geijo E, and Belmonte C: Response characteristicsof corneal sensory fibers to mechanical and thermal stimula-tion. Pain 7:1, 1979.

29. Lele PP and Weddell G: Sensory neurons of the cornea andcutaneous sensibility. Brain 79:119, 1959.

30. Grant FC, Groff RA, and Lewy FH: Section of the descendingspinal root of the fifth cranial nerve. AMA Arch Neurol Psy-chiatry 43:498, 1940.

31. Rowbotham GF: Observations on the effects of trigeminal de-nervation. Brain 62:364, 1939.

32. Ehinger B: Distribution of adrenergic nerves to orbital struc-tures. Acta Physiol Scand 62:291, 1964.

33. Laties A and Jacobowitz D: A histochemical study of theadrenergic and cholinergic innervatidn of the anterior segmentof the rabbit eye. Invest Ophthamol 3:592, 1964.

34. Ehinger B: Distribution of adrenergic nerve in the eye andsome related structures in the cat. Acta Physiol Scand 66:123,1966.

35. Laties AM and Jacobowitz D: A comparative study of theautonomic 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 theanterior eye segment of some primates. Z Zellforsch 116:157,1971.

38. Ehinger B and Sj6berg N-O: Development of the ocularadrenergic neve supply in man and guinea-pig. Z Zellforsch118:579, 1971.

39. Stone RA, Laties AM, and Emson PC: Neuropeptide Y andthe ocular innervation of rat, guinea pig, cat, and monkey.Neuroscience 17:1207, 1986.

40. Bowers CW and Zigmond RE: Localization of Neurons in therat superior cervical ganglion that project into different post-ganglionic trunks. J Comp Neurol 185:381, 1979.

41. Dail WG, Khoudary S, Barraza C, Murray HM, and BradelyC: The fate of adrenergic fibers which enter the superior cervi-cal ganglion. In Histochemistry and Cell Biology of Auto-nomic Neurons, SIF Cells, and Paraneurons, Eranko O, Soin-ila S, and Paiva'rinta H, editors. New York, Raven Press, 1980,pp. 287-297.

42. Jacobowitz D and Woodward JK: Adrenergic neurons in thecat superior cervical ganglion and cervical sympathetic nervetrunk. A histochemical study. J Pharmacol Exp Ther 162:213,1968.



43. Bowers CW, Dahm LM, and Zigmond RE: The number anddistribution of sympathetic neurons that innervate the rat pi-neal gland. Neuroscience 13:87, 1984.

44. Patrickson JW and Smith TE: Innervation of the pineal glandin the rat: An HRP Study. Exp Neurol 95:207, 1987.

45. Arvidson B: Retrograde transport of horseradish peroxidase insensory and adrenergic neurons following injection into theanterior eye chamber. J Neurocytol 8:751, 1979.

46. Smith PG, Brucker JW, and Mills E: Reinnervation of Muller'ssmooth muscle by atypical sympathetic pathways followingneonatal ganglionectomy in the rat: Structural and functionalinvestigations of enhanced neuroplasticity. Neuroscience23:781, 1987.

47. Smolen AJ, Wright LL, and Cunningham TJ: Neuron num-bers in the superior cervical sympathetic ganglion of the rat: Acritical comparison of methods for cell counting. J Neurocytol12:739, 1983.

48. Purves D, Rubin E, Snider WD, and Lichtman J: Relation ofanimal size to convergence, divergence, and neuronal numberin peripheral sympathetic pathways. J Neurosci 6:156, 1986.

49. Wolf G: The ratio of preganglionic neurons to postganglionic

neurons in the visceral nervous system. J Comp Neurol75:235, 1941.

50. Thomander L, Aldskogius H, and Arvidsson J: Evidence for asympathetic component in motor branches of the facial nerve:A horseradish peroxidase study in the cat. Brain Res 301:380,1984.

51. Butterfield LC and Neufeld AH: Cyclic nucleotides and mi-tosis in the rabbit cornea following superior cervical ganglion-ectomy. Expfcye Res 25:427, 1977.

52. Beuerman RW, Klyce SD, Vigo MC, Duruy BM, and CrossonCE: Modulation of latency and velocity in cornea! epithelialwound healing. ARVO Abstracts. Invest Ophthalmol Vis Sci26(Suppl):91, 1985.

53. Sweeney DF, Vannas A, Holden BA, Tervo T, and TelarantaT: Evidence for sympathetic neural influence on human cor-neal epithelial function. Acta Ophthalmol (Copenh) 63:215,1985.

54. Toivanen M, Tervo T, Partanen M, Vanna A, and HervonenA: Histochemical demonstration of adrenergic nerves in thestroma of human cornea. Invest Ophthalmol Vis Sci 28:398,1987.


Documents

Sensory and sympathetic innervation of the …iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/...Sensory and Sympathetic Innervation of the ... completely filled the area