STRUCTURE AND FUNCTION OF THE TEAR FILM

P.N. Dilly

Department of Anatomy
St. George's Hospital Medical School
Cranmer Terrace, Tooting
London SW17 ORE

INTRODUCTION

The existence of the tear film is well known but its structure is less well understood.
It is unwise to consider the tear film in isolation from blinking. Blinking has a profound
influence upon the structure, stability, and function of the tear film. Many glands from
different sites contribute to the clear fluid that bathes the surface of the eye. Chemically
tears are very similar to dilute blood, with a reduced protein content. (The pH of tears
approximates to that of blood plasma but it has a slightly greater osmotic pressure.) The
film covers the exposed surface of the eye and provides an optically smooth interface with
the atmosphere. Lacrimation is well known throughout the animal kingdom, but crying
with sorrow and laughter are probably confined to man. This is probably an adjunct to the
vast range of facial expressions available to man. The parasympathetic nerve fibers that are
secretomotor to the lacrimal gland are distributed for much of their course with the facial
nerve, the motor nerve of facial expression.
It has been known for some time that the tear film has three layers (Wolff, 1954).
There is a mucus layer close to the epithelium, an aqueous layer outside that and the
external surface of the aqueous layer is covered with an oily layer. The oily and aqueous
layers are easy to demonstrate. The aqueous being obvious, and the oily layer can be made
to produce a whole host of diffraction colors depending upon its thickness. The mucus
layer is more difficult to visualize simply because it has an identical refractive index with
the aqueous phase of the tear film. It can be deduced from the vast numbers of goblet cells
that discharge onto the surface of the eye from the conjunctival surface and also from the
mucus thread that is found in the lower canthus. Postmortem it can be demonstrated by a
whole host of mucus revealing stains. Electron microscopic investigations have confirmed

Lacrimal Gland, Tear Film, and Dry Eye Syndromes

Edited by D.A. Sullivan, Plenum Press, New York, 1994 239
its presence and subsequent workers have reported an increasing thickness for this layer as
the methods of preservation have advanced.
Many protective substances are present in tears including lactoferrin, lysozyme, non-
lysozyme antibacterial factor, complement and anti-complement factor and interferon, as
well as the immunoglobulins. There are also lymphocytes in tears. Lactoferrin works by
chelating iron and thus deprives microorganisms of iron. It modulates complement activity,
at least in vitro. It is thought that it is synergistic with a specific antibody, and may also
enhance the action of lyzozyme.
Lysozyme is lytic to glycosaminoglycans and in the presence of complement it
facilitates IgA bacteriolysis. The immunoglobulin A neutralizes viruses and inhibits
bacterial adherence to the epithelial surface. IgO promotes phagocytosis and complement-
mediated bacteriolysis. IgE is increased in allergic responses.
The closed eye presents a special set of problems in the understanding of the tear
film. When the eye is closed, the precorneal tear film is in osmotic equilibrium with the
aqueous humor, and no osmotic flow occurs; as a consequence the corneal stroma thickens.
During extended periods of lid closure such as sleep the concentrations of
immunoglobulins increase, only to decrease again in the tear film of the open eye. An
increase in the concentration of plasmin and a considerable increase in the numbers of
polymorphonuclear leucocytes is also associated with the closed eye of sleep.

BLINKING

Blinking functions above all as a protective mechanism, at its crudest interposing the
tissue of the eyelid between a potential insult to the eye and the eye. There are other known
functions of blinking. It has been shown that a blink is associated with a discharge of
secretion from the meibomian glands. The normal human blinks between two and ten times
a minute in a normal environment. These blinks have been shown to change the
distribution of the lipid layer covering the aqueous layer of the tears. The mucus for the eye
comes mainly from the conjunctival goblet cells. Although there is some autonomic control
of mucus secretion, it is difficult to see how this control is effected since the goblet cells
are not associated with myoepithelial cells. It is known that a blink exerts a force of at least
5 grammes on the globe, and that firmly closing the eye can displace the eye backwards in
the orbit several millimeters. It is possible that this squeezing could cause mucus to be
released from the goblet cells. Any painful object is probably painful because it has
penetrated the mucus layer and is stimulating the superficial nerve endings of the cornea
and conjunctiva directly, evoking the response of tightly screwing the eyelids together.
Perhaps this is a mechanism to replace rapidly the contaminated mucus layer with a fresh
layer from beneath and thus removing the foreign body from contact with the globe. The
increased load applied to the tear film during this powerful lid activity probably also serves
to distribute the fresh mucus evenly. Prydal (1992) has shown that it takes about 30
minutes for the tear film mucus to be replaced after it has been destroyed by acetyl cystein.
This probably represents the physiological maximum rate at which the tear film mucus can
be replaced from the goblet cell source by normal unconscious blinking mechanisms. The
similar well known vast increase in aqueous flow is effective in washing foreign material
from the surface of the mucus phase of the tear film.

240
 AIR

~=-~~~~;{~-t~~;~{i~ ~t Oily Layer

~_ =_::_=_=_=:. =- ~-iI-_-:'::==7.
--------------
Aqueous

~~~~~~~~}~=-=-J.}

.·0'
..... '.. .'
.

. '. .
. .J: •

Mucus
. lIf.·. ~ .
. . . .

~.
. ~:

tt
Glyc.ocalyx

Microvilli

CORNEAL EPITHELIUM
Figure 1. Diagram of the structure of the tear film. The film is 30-40 /lm thick. The immunoglobulins are
represented by stars, and the glycoproteins by the long lines with crosses. It is probably a much more stable
structure than has previously been proposed, and measurable flow probably only occurs after stimulation. The
components are known to be secreted from many different sources. How their relative proportions and
distribution is controlled is unknown. Most components probably exist in varying concentrations in most
layers.

241
 Another function of the blink is to smooth the anterior surface of the mucus layer and
thus improve its optical properties. Indeed the physical forces applied to the tear film
during blinking are probably essential not only for spreading the oily layers, but also for
maintaining the structural integrity of the deeper layers.

OILY LAYER

The thickness of the oily layer varies with the width of the palpebral fissure and in
varying places across the exposed part of the globe. It is made up of many oily substances
derived from the meibomian glands. It is said that meibomian secretions help stabilize the
tear film, but when meibomian secretion is added to a thin water film it seems to have little
effect. Meibomian secretions collected directly from the glands are solid or semisolid,
having a melting point above body temperature, tear temperature is about 3 or 4°C below
this, so it must need some additional factors to allow melting at this lower temperature and
spreading across the aqueous tears.
The oily layer behaves as if it is a sheet suspended from the upper lid, thickening as
the eyelid closes and thinning again on opening. It is an inhomogenous layer containing
complexes of lipids and other substances. The layer is not of uniform thickness, but these
variations seem to have little effect on visual activity. Part of the lipid layer's function is to
reduce the evaporation of the aqueous tears. The normal rate of evaporation from the tear
film is 0.085 Ill/minute, the superficial oily layer is probably responsible for this low value.
Mishima has estimated that without the oily layer the rate would be between 0.85 to 1.7
Ill/minute. It is known that this layer dispels and breaks up if a drop of sebum is added and
the rate of evaporation increases. The other probable effect is to reduce the surface tension
of the aqueous phase and thus to reduce wave formation as the aqueous is propelled over
the cornea by movements of the eye or eyelids.
The oily part of the tear film is extremely efficient in protecting the eye from small
dust particles. This is easily demonstrated by examining the tear film in one eye and the
contact lens surface of the other eye in volunteers, who have been exposed to a dusty
environment. After as little as ten minutes examination with the slit lamp reveals few
particles in the contact lens free eye, whereas the contact lens that destroys the lipid
aqueous integrity is heavily contaminated on its exterior surface. It is surprising that the
considerable inhomogeneity of the lipid layer, unless gross, seems to have little effect on
visual acuity.

AQUEOUS

The aqueous layer is about 7 11m deep in man. Normally most of it is produced by the
accessory lacrimal glands. Its more important functions are to provide a lubricating layer
between the moving surfaces of the eye and its adnexa, to remove foreign material and to
nurture the corneal and conjunctival epithelia. There is probably an aqueous layer
separating the mucus covering of the bulbar and palpebral conjunctivae under the eyelids.

242
It would be desirable that such a layer exists since otherwise the sheer stress of lid
movements would be transmitted via the mucus to the epithelium. Since the lacrimal gland
discharges its watery secretions into the upper canthus, it would seem reasonable to assume
that the secretions would be released into the aqueous part of the tear film. The osmolarity
of the tear film seems to be critical. Gilbard (1978) has shown that concentrations of over
310 milliOsmols are associated with dry eye syndromes. Part of the functions of the
aqueous phase may be to supply water to keep osmolarity below this critical level. It is
known from Gilbard's work with keratoconjunctivitis sicca that a rise in osmolarity is
associated with aqueous tear deficiency. There are many soluble mucins contained in the
aqueous layer.
It is a function of aqueous to provide the water for mucus hydration by inward flow
from the aqueous into the mucus layer.
The lubrication of the lid movements over the globe and vice versa is provided by the
aqueous part of the tear film. The gliding surfaces themselves are covered with mucus. This
gliding is enhanced by the exclusion of the lipid layer between lid and globe. The relative
velocity of the sliding surfaces is about 15-25 cm/sec and the sheer rate 20,000 sec -1. The
viscosity of the aqueous is low about 1.1 cps. The sheer stress at the mucus/aqueous
interface is 150 dynes/cm 2 . Because the mucus is so much more viscous, the sheer will
decrease rapidly in mucus and will be negligible at the ocular cellular surface. Thus, this is
an excellent protective mechanism stopping the blinking eyelids from damaging the
corneal and conjunctival epithelium.
A stable aqueous phase must exit under the lid since the structure of the tear film and
the nature of the lid/globe movements are such that a hydrodynamic lubrication mechanism
is required. If it were a mucus/mucus interface, the boundary lubrication would be
inadequate to prevent ocular surface tissue damage.
Wearing contact lenses destroys the integrity of the oily layer, and Maurice (1961)
has shown that the tear evaporation rate then becomes significant. Mishima's (1965) work
on evaporation rates from the normal eye show a miniscule loss of water. It would seem
wasteful to have a basic tear flow rate greater than this, as the eye would rapidly overflow,
or be drained via the nasolacrimal system. It would seem unreasonable to have a basic flow
rate ten times that required to maintain the aqueous volume against evaporation, when there
exists a massive reflex flow rate for emergencies. There has been much debate about the
amount and significance of any basic tear flow in the unstimulated eye (Jordan & Baum
1980). Such flow, if it exists, is very difficult to measure as almost any investigation
produces some reflex tearing. Patients are known who have a normal tear film and a
blocked nasolacrimal system who do not suffer epiphora. As methods of investigation have
become less and less invasive, estimates of a basal secretion rate for tears have reduced in
volume. It is now probably worthwhile to reappraise if there is a basal secretion rate. Many
investigators using fluorophotometry have assumed that the fluorescein mixing in the tears
was homogeneous. Recently Maurice (1993) has shown that after the installation of a micro
drop of fluorescein, it can take up to several minutes before it is evenly distributed. As yet
the feed back mechanisms that must signal the state and composition of the tear film are
virtually unknown. There must be some clues in the distribution of the paccinian-like
corpuscles within the eye. They are concentrated along the superior limbus, a site passed

243
over by the upper eyelid during each blink. The fact that the eye is so sensitive to changes
and abnormalities in the tear film suggests that there are other receptors sensitive to very
subtle changes in the tears.

MUCUS LAYER

The mucus layer in the eye is unlike the mucus layer lining the lung, in that it is
probably less mobile, it is not being moved along by the action of cilia, rather it is static
and anchored to the microvilli. The cleansing role of lung mucus is undertaken in the eye
by the aqueous layer of the tear film flowing over the mucus layer. There must however be
some movement of mucus from the conjunctiva where it is secreted, to the precorneal tear
film. It is probably spread across the surface of the eye by the action of the lids. Although
the mucus layer is anchored to the cells beneath it, it is not in accurate register with their
boundaries. It is possible to observe mucus spreading out from goblet cell apertures, the
spread between adjacent goblet cells is confluent, and often is not in register with the
underlying cell outlines. The goblet cell mucus has variable staining properties, suggesting
some chemical non homogeneity (Adams & Dilly 1989).
The tear film is rich in nutrients, but it is not known how much if any are used for
metabolism by the corneal epithelium. It does have the vital function of allowing diffusion
of oxygen and other gasses to and from the epithelium. Obstruction of this pathway by an
impermeable contact lens rapidly leads to corneal damage.
Whereas the aqueous provides the cleavage plain for movements of the eyelids, it is
the mucus layer that protects the underlying epithelium from sheer damage.
The mucus layer itself heals rapidly and effectively. Holes and other defects made in
the mucus layer of the tear film repair almost immediately, provided that the underlying
cells are intact. The edges of defects are probably brought together by attractive forces
between the long chain molecules. It is probable that these attractive forces are enhanced
by the compression forces of the eyelids during blinking. The rapid restoration of this layer
is vital for the protection of the eye against drying and bacterial invasion. It is the external
surface of the mucus layer that is the first solid layer encountered by incoming material. It
is known that mucus has an inhibiting effect on bacterial adhesion.
Mucus is a spongelike material with fluid in a meshwork of glycoprotein molecules.
Mucus is known to be viscoelastic (Kaura & Tiffany 1986), which probably explains the
dents left in the precorneal tear film mucus when tonometer cones or glass fibres are
pressed against them (Prydal et aI, 1993).
The mucus layer, because of its micellar structure, probably acts as a reservoir for
immunoglobulins that allows only the slow release of immunoglobulins during the day
when the eye is open and its vulnerability to air borne pathogens and antigens is increased.
The overnight increase in concentration probably represents the normal physiological rate
of secretion of immunoglobulins by the untroubled eye. When the eye is closed it is
unlikely that a layer of lipid exists, its function being taken over by the physical barrier of
the opposed eyelids. The aqueous layer would need to be maintained in order to protect the
surface epithelium from sheer damage during the rapid eye movements associated with
sleep.

244
 Further evidence for a micellar structure of the mucus layer of the tear film is that
microdrops of fluorescein introduced into the aqueous rapidly disperse and disappear,
whereas drops introduced more deeply take several minutes to become evenly spread
throughout the tear film in spite of normal blinking (Maurice 1993).
The mucus layer is sufficiently rigid and elastic for a contact lens to float on it. The
movements induced in a contact lens by the upper eyelid would soon destroy the corneal
epithelium if it was not for this protective buffer. The potential effects of a contact lens
moving in the eye upon the corneal epithelium must be similar to that of an eyelid during a
blink and both of them are prevented from damaging the eye by the barrier of the mucus
layer.
Doane (1980) has shown that there is a retraction of the eyeball of between 0.7 -1.6
mm during a blink. Riggs (1987) showed that the amount of retraction increased the more
powerful the blink. It is known that the eyelids can exert a force of several grammes on the
globe and it is likely that it is this push from the eyelids that displaces the globe backwards.
Such a force if applied directly to the epithelial cells once every 2-10 seconds while awake
would surely disrupt the cells. It is the viscoelastic nature of the mucus layer that prevents
this damage from occurring. Besides the eyelids as a potential source of disruption of the
tear film, movements of the globe will also cause stresses in the film and the elastic
behaviour of the tears will also resist the distortion of the tear film during eye movements.
The mucus of the tears contains very long molecules of over 3 microns in length and
with molecular weights of 50 million daltons or more. These long molecules are probably
glycoproteins and hyaluronan, and are responsible for trapping the water in the mucus
layer, the self repair of this layer and also for the viscoelastic properties of the mucus layer
of the tear film. They have the classical 'lampbrush' structure, but there are spaces and
naked areas along the backbone that permit aggregation through the disulphide bonds. It is
the combination of this molecular structure and the enclosed and trapped water that gives
the viscoelastic properties to the ocular mucus layer.
The mucus is probably linked together by hydrophobic bonding. This bonding is
weak and probably forms and breaks with minor local changes, the molecular lattice
zipping open and shut. This bonding is probably part of the mechanism that produces the
self repair properties of the mucus. The stability of this mucus bonding is probably
enhanced by hyaluronan. Hyaluronan has the ability to form gels of randomly entangled
molecules. It is known to be responsible for the main opposition to water flow in the matrix
of connective tissue. It can achieve this at extremely low concentrations of less than 1
Ilg/ml. The aggregation is probably driven by the hydrophobic bonding between the
glycosaminoglycan polymer backbone, which with additional hydrogen bonds balances the
electrostatic repulsion between the polyionic aggregants (Scott 1992).
In connective tissues these bondings can take on a semipermanent nature. Such a
feature in tear glycosaminoglycans would support the idea of a more permanent nature for
tear film mucus than has previously been suggested. If this is so, then hyaluronan should be
an effective treatment for some types of dry eye syndromes.
Mucus much reduces the ability of some bacteria to adhere to the surface of the eye.
The bacteria seem inhibited from penetrating the mucus layer, and thus prevented from
reaching the vulnerable cell layers beneath it. The size of bacteria would suggest that a
submicron thick layer of mucus would not have the dimensions to prevent bacterial

245
penetration and that the thicker layer found by Prydal (1992 a & b) and his colleagues
would be better suited for this role.
The tears are known to contain many biologically active molecules. Lysozyme, first
discovered by Fleming in 1922 and renowned for its bacteriocidal activity, may have
another much more important function in tears. It is known that lysozyme destroys bacteria
by digesting their mucopolysaccharide coats. A similar activity involving the
glycosaminoglycans in the mucus layer might reduce the rate of drying and solidification
of the mucus layer of the tear film. It is known that in some cases dry eyes in which there is
a mucus abnormality the concentration of lysozyme is reduced. It is a common experience
that mucus removed from the eye rapidly solidifies. Just adding water does not cause it to
liquify for many hours. This process can be speeded up by adding lysozyme. Lysozyme
may have a role in the management of some forms of dry eyes. Perhaps the lysozyme has a
role in maintaining the physical properties of the complex chemical mixture that is tear
film mucus. The other property of lysozyme that has a role in tear film stability is its
molecular shape and electric charge. Besides helping to keep the mucus fluid, lysozyme is
known also to increase the viscosity of tear mucus.

GLYCOCALYX

The attachment of the mucus layer to the surface of the epithelial cells is crucial for
the stability of the whole tear film. The epithelial cells upon which the tear mucus layer
abuts are covered with a dense layer of microvilli. These microvilli are decorated with long
chains of material that appear anchored to the microvilli and extend out into the mucus
layer. They are revealed by tannic acid and ruthenium red staining. I have proposed that
these filaments are the same material as is found in the sub-surface vesicles, and that their
function is to anchor the mucus layer to the cell surface (Dilly 1985). In diseases of the
epithelial cells, it is this anchorage system that is destroyed with the consequent
destabilisation of the tear film.
The presence of a glycocalyx at the epithelial surface would suggest that the surface
is strongly polar, and hence readily wettable by aqueous solutions. This is required for all
cellular surfaces according to modem theories of membrane structure with dense arrays of
heterogeneous polar and highly hydrated carbohydrate groups attached to membrane
glycoproteins giving a very low interfacial tension against aqueous solutions.

CONCLUSION

I am proposing therefore that the tear film is much thicker than the 7 /lm usually
quoted. The figure is nearer 30-40 /lm, with the 'extra' thickness being contributed by the
thick mucus layer. This thick layer had not been detected previously because it is invisible
to standard optical methods. I propose that the mucus layer is stable and anchored to the
epithelium. It contains many vital biologically active substances that can probably be 'slow
released' from its micelles. The aqueous phase contains dissolved mucins and provides the
cleavage plane for lid movements and the more viscous mucus layer stops the forces

246
generated by the movements from disrupting the ocular epithelium. It is probably more
accurate to describe the mucus and aqueous layers of the tear film as phases with more or
less mucus, respectively.

REFERENCES

Adams, G.G.W., and Dilly, P.N., 1989, Differential staining of ocular goblet cells, Eye 3:840.
Cope, c., Dilly, P.N., Kaura, R., and Tiffany, J.M., 1986, Wettability of the corneal surface: a reappraisal,
Curro Eye Res. 5:777.
Dilly, P.N., 1985, On the nature and role of the subsurface vesicles in the outer epithelial cells of the
conjunctiva,Brit. J. Ophthalmol. 69:447.
Dilly, P.N., 1985, Contribution of the epithelium to the stability of the tear film, Trans Ophthalmol Soc. UK.
104:381.
Doane, M.G., 1980, Interaction of eyelids and tears in corneal wetting and the dynamics of the normal human
eyeblink, Amer. J. Ophthal. 89:507.
Fleming, A., 1922, On a remarkable bacteriolytic element found in tissues and secretions, Proc. Roy. Soc. B
93:306.
Gilbard, J.P., Farris, R.L., and Santamaria, J., 1978, Osmolarity of tear microvolumes in keratoconjunctivitis
sicca, Arch. Ophthalmol. 96:677.
Holly, FJ., 1973, Formation and stability of the tear film. In The preocular Tear Film and Dry Eye
Syndromes, FJ. Holly and M.A. Lemp (Eds.) International Ophthalmology Clinics, Boston, Little,
Brown.
Jordan, A., Baum, J., 1980, Basic Tear Flow Does It Exist? Ophthalmology 87:920.
Kaura, R., Tiffany, J.M., 1986, The role of mucus glycoproteins in the tear film, In The Precorneal Tear
Film, Dry Eye Institute, Lubbock Texas, Holly FJ., Editor, 728-732.
Kijlstra, A., and Veerhuis, R., 1981, The effect of an anticomplementary factor on normal human tears, Amer
1. Ophthalmol. 92:24.
Mishima, S., 1965, Some physiological aspects of the precorneal tear film, Arch. Ophthalmol. 73:233.
Mishima, S., and Maurice, D.M., 1961, The oily layer of the tear film and evaporation from the corneal
surface, Exp. Eye Research 1:39.
Prydal, J.1., Campbell, F.W., 1992, Study of tear film thickness and structure by interferometry and confocal
microscopy'/nvest. Ophthalmol Vis Sci 33:1996.
Prydal, 1.1., Arlal, P., Woon, H., Campbell, F.W., 1992, Study of human tear film thickness using laser
inteferometry, Invest. Ophthalmol Vis Sci 33:2006.
Prydal, 1.1., Kerr Muir, M.G., and Dilly, P.N., 1993, Comparison of the tear film thickness in three species
determined by the glass fibre method and confocal microscopy, Exp .. Eye Research. In Press.
Riggs, L.A., Kelly, J.P., Manning K.A., and Moore, R.K., 1987, Blink related eye movements, Invest.
Ophthalmol. 28:334.
Scott, J .E., 1992, Supramolecular organisation of extra cellular matrix glycosaminoglycans in vitro and in the
tissues, FASEB J. 62:639.
Wolff, E., 1954, Anatomy of Eye and Orbit, New York Blakiston Co. 4th Edn. 207.

247