The Effects of Novel Amphipathic Block Copolymers on Stabilization of the Rat Tear Film

Abstract

purpose. To determine whether various novel amphipathic polymers could be used to stabilize the tear film of the rat. The rheologic properties of these polymers were examined to investigate whether particular structural or physical characteristics improve the stability of the tear film.

methods. Amphipathic polymers or particular phospholipids were mixed with a test solution of tears and saline and applied to the clean, dry corneal surface of a rat. The specular reflection of the tear film was observed at high magnification and recorded. For each of the polymers or lipids, the effects on surface regularity and tear break-up time were compared. After the experiments, histologic sections of the tested eyes were prepared and examined for acute cytotoxic effects on the cornea and ocular conjunctiva.

results. Tear film break-up time was markedly affected by differences in polymer structure. Copolymers consisting of separate hydrophobic and hydrophilic regions appeared to be the best stabilizers. No acute cytotoxic effects were observed in histologic sections of corneas to which the polymers had been applied.

conclusions. Amphipathic polymers can be designed to increase tear film stability. Increased tear film stability occurred more readily with copolymers, possibly through their interaction with both lipid and aqueous tear components.

The tear film forms a complex and dynamic structure that lubricates and protects the exposed surface of the eye. Our basic understanding of its structure recognizes three major components secreted by different ocular tissues: mucins secreted by corneal epithelial cells and conjunctival goblet cells, the aqueous layer and its associated proteins secreted by the lachrymal and accessory glands, and lipids secreted by the meibomian glands. ^{1 2 3 4 5} An anomaly in any one of these components can lead to instability of the tear film, resulting in a condition known as dry eye. ^{6 7 8 9}  

The lipid layer is a complex mixture of fatty acids, phospholipids, waxes, sterols and cholesterol esters. ¹⁰ Viewed simply, it provides a hydrophobic barrier that prevents both the spillage of tears and the contamination of the tear film by sebum present on the lid margin. It also prevents evaporation from the aqueous and mucous layers beneath it. ^{10 11} More detailed analysis indicates that the efficiency of this barrier function is critically dependent on the lipid subcomponents and how they interact with proteins in the aqueous component of the tears. Although phospholipids make up a relatively small proportion of the lipid layer, their amphipathic nature (having both hydrophilic and hydrophobic domains) allows them to interact with both the hydrophobic superficial regions (e.g., nonpolar waxes) of the lipid layer, and with the hydrophilic aqueous layer. ^{12 13 14} It has been shown that there is a significant difference between the components of the tear lipid layers of dry eyes and normal eyes. ¹⁵ In particular, the absence of anionic and neutral phospholipids in the tear film correlates with an increase in the incidence of evaporative dry-eye disorders. ^{13 14 16 17 18} It is now believed that the lipid layer is not simply floating on the surface of the tear film but interacts with proteins, such as tear lipocalins. This interaction is believed to assist in maintaining lipid layer stability ^{19 20 21} by interacting with lipids at the junction between the lipid and aqueous layers. 

One strategy for developing artificial tears or tear supplements for treatment of dry eye has been to incorporate polymers, such as methylcellulose, that increase the viscosity of the tear film. ²² In the current study we adopted a different strategy based on the use of amphipathic polymers with the intention of improving the stability of the tear lipid layer, and hence the tear film, as a whole. This concept arises from studies exploring the usage of amphipathic polymers to enhance the stability of liposomes. ^{23 24 25} The most successful of these have comprised hydrophilic and hydrophobic blocks. ^{23 26 27} There are different views as to how this increased stability occurs. One is that the polymers coat the bilayer of the liposome and the other is that they are incorporated into the bilayer. ²⁸ Although the lipid layer of the tear film is regarded as a structured monolayer, ²⁹ the block copolymers could nevertheless bind to or be integrated into the structured monolayer and thereby enhance the stability of the tear film. Therefore the purpose of this study was to investigate whether amphipathic polymers could be used to stabilize the tear film, using the rat as a model. 

Materials and Methods

Polymers and Their Synthesis

Non–cross-linked polymers were synthesized that had different hydrophilic and hydrophobic patterns or differed in size (Table 1) . Some had a polar head group and hydrophobic tail and others had blocks of hydrophilic and hydrophobic groups along their lengths. Reversible addition fragmentation chain transfer (RAFT) ³⁰ was used to synthesize homopolymers and block copolymers of styrene, methyl acrylate, and acrylic acid (Figs. 1 2 3 4) . A homopolymer of polystyrene was also synthesized using anionic polymerization so that a terminal carboxyl group could be introduced. The molecular weights of the polymers were determined by gel permeation chromatography (tetrahydrofuran mobile phase, Mixed Gel A and E columns; Polymer Laboratories, Amherst, MA) and the structures were confirmed using nuclear magnetic resonance (NMR) and Fourier transform infrared spectroscopy (FTIR). 

Testing the Polymers in Eyes

The method used was based on that described by Tragoulias and Anderton. ³¹ All animal experiments adhered to the tenets of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Animal Ethics Committee of the University of New South Wales. Male and female Wistar rats (250 g) were anesthetized with an intraperitoneal injection of ketamine (80 mg/kg) and xylazine (10 mg/kg), and the head was stabilized by placing the animal in a nontraumatic head holder and positioned so that the corneal limbus was horizontal. If eye movements or a blink reflex to light was observed, the anesthetic was boosted to maintain deep anesthesia. Small retractors applied to the eyelashes maintained an open palpebral aperture. 

Tears were collected using a 10-μL glass capillary tube applied to the inner canthus tear reservoir. To enable several experiments to be performed with a single tear sample, the very small tear volumes collected were diluted with an equal volume of 0.9% saline (saline for injection; Astra Pharmaceuticals, North Ryde, Australia). Previous experience has shown that this mixture of tears plus saline performs very similarly to tears alone. For testing, the polymers or phospholipids were first dissolved in an appropriate solvent (Table 1) at 2 mg/mL. These were then mixed 1:1 with the tears-plus-saline mixture. For control subjects, vehicle alone was mixed 1:1 with the tears-plus-saline mixture, or saline alone (not mixed with tears) was used. 

Before the application of experimental solutions, the corneal surface was prepared by removal of the existing tear film. This was accomplished by applying 5 μL of 0.9% saline to the eye, after which a wick of tissue (Kim-wipe; Kimberly-Clark, Milsons Point, NSW, Australia) was gently applied to the tear pool of the inner canthus without touching the eye. When the displaced fluid was removed by capillary action, the cornea rapidly dried through evaporation of the surface saline film. Then, 3 μL of the tears-plus-saline solution was applied to the dry cornea to observe whether a stable tear film would form on the ocular surface that broke up in a manner similar to normal tears. After a stable tear film was observed to form, the tear film was displaced again with saline and a wick. To the freshly prepared surface, 3 μL of the test solution or vehicle control was applied to the eye and observed until it dried. At least two rats (four eyes) were used for each polymer or control. 

The performance of the various solutions in the eye was monitored using a digital video camera (model NV-MX7; Panasonic, Osaka, Japan) attached to a specular microscope (WILD-M3C; Leica, Heidelberg, Germany; light source, Leica Intralux 5000; Volpi, Zurich, Switzerland). In particular, the manner of tear break-up and the tear drying time were recorded. 

Analysis of Results

Still frames of the tear film were captured at 30-second intervals after solution application to the eye and were transferred to a computer with video capture and editing software (Ulead Systems, Torrance, CA). To compare data, an in-house grading scale of ocular dryness ³¹ was used. The scale consisted of a series of six standard images ranging from a grade of 0 being a normal tear film and 5 being dry (Fig. 5) . 

Histology of Treated Eyes

At the completion of experiments, anesthetized rats were killed by intraperitoneal injection of pentobarbital sodium (0.5 mL/kg, Lethobarb; Virbac, Sydney, Australia) and then both eyes were removed and fixed for 24 hours in 10% formalin made up in PBS (pH 7.4). The maximum amount of time an ocular surface was exposed to a polymer before fixation was never more than 2 hours. The eyes were then washed in several changes of PBS, and the lens was removed through a slit in the sclera and the cavity filled with optimal cutting temperature (OTC) compound (Tissue-Tek; Sakura, Torrance, CA). The eye was then frozen into a block of OCT and frozen sections (8 μm) cut with a cryostat (Shandon, Rowcorn, UK), mounted on microscope slides, and stained with hematoxylin and eosin. Sections were examined for evidence of cell death (karyorrhexis, karyolysis, swelling) lymphocytic invasion, or any irregularities in the morphology of the cornea or conjunctiva that may have been attributable to the treatment. 

Results

A summary chart describing the effects of all samples applied to the eye is shown in Figure 6 using the in-house grading scale (Fig. 5) . 

Control Tears

Application of the tears-plus-saline solution (1:1) to the eye surface resulted in the formation of a stable and smooth layer. This layer appeared to break down gradually, and as it thinned, small nodules appeared that were surface irregularities of epithelial cells emerging above the tear layer (Fig. 7) . With time, these nodules became more numerous and apparent on the surface as the tear film thinned. It took approximately 420 seconds for maximum breakdown (grading scale level 4), which compares favorably with the reported normal blink rate for rats of 300 seconds (Swarbrick HA, et al. IOVS 1994;35:ARVO Abstract 1690). This differed markedly from saline alone, which resulted in a smooth ocular surface on application, but the surface lost its integrity rapidly, within 10 seconds (Fig. 7) . Mixing of ion exchange purified water (18.2 MΩ) with tears plus saline (a control for the water-soluble polymers) resulted in a thick layer that remained stable on the ocular surface after application. The coating was observed to thin gradually in the minutes after application until it ruptured and the surface dried through evaporation. 

The methanol control (vehicle for C-PS, PTBA, PMA-PAA) was unable to form a film on the eye surface (Fig. 6) . After it was applied, irregularities in the tear film began to appear immediately, with several dry islands visible within seconds of application, and these dry spots rapidly expanded until the surface was completely dry after 90 seconds. Possible mechanisms for this could be that methanol damaged the ocular surface, precipitated the tear film proteins during the mixing process, or interacted with the lipids. 

The chloroform control (vehicle for PS, PNBA, PTBA-PS) appeared thin and watery, but remained stable on the ocular surface for twice the time of the methanol control (120 seconds compared with 60 seconds to grade 5). The film appeared to break down gradually to a point of complete dryness. In this case it is likely that the small amount of chloroform evaporated from the surface during the mixing of the tears with the chloroform before application to the ocular surface. 

Phospholipids

Phosphatidylcholine (PC) and phosphatidylinositol (PI) were dissolved in chloroform before being mixed with the tears plus saline. Both lipids formed thin and unstable layers when applied to the ocular surface. PC dried in the same time as its solvent control (120 seconds) indicating it failed to interact with the tear film lipids, and PI yielded a longer time on the ocular surface before drying (210 seconds; Fig. 6 ). 

Polymer-Treated Tears

There was a large variation in the ability of tears with added synthetic polymers to maintain a stable tear film. PAA-5 and PNBA were the worst performers and then there was an increasing order of performance PVA-1, PAA3, PTBA, PAA2, and C-PS which all performed worse than tears plus saline. PS performed similarly to tears plus saline, except that it maintained a greater level of stability in a state of partial breakdown (grading scale level 3; Fig. 6C ) than did the control. An interesting observation was that the block copolymers PTBA-PS and PMA-PAA far outperformed the normal tears. 

The water-soluble polymers PVA-1, PAA-3, and PAA-5 were similar to one another in their appearance on the ocular surface, breaking up totally within 150 seconds of application (Fig. 8) . The tear film appeared to thin, and dry spots were visible within 15 seconds of application. In all cases, the tear film collapsed within 120 seconds, and any remaining fluid evaporated. PAA-2 provided the most stable film that was produced by the water-soluble polymers. It appeared to dry in a similar fashion, although more rapidly than the purified water control (Fig. 8) . 

Tears-plus-saline solutions containing PTBA (methanol), PNBA (chloroform), and C-PS (chloroform)-containing tear solutions appeared thin and watery, and there was little increase in tear film stability over the control. PTBA formed an irregular film on application, but despite its irregularity, its decline was gradual, with the film taking 270 seconds to reach total dryness (grade 5) on the ocular surface. The structurally similar PNBA yielded the same drying time as the chloroform control (120 seconds). For PNBA, drying occurred as a rapid collapse of the film, which differed from the continual thinning of PTBA. This may have been due to the different vehicles rather than the structure of the polymer. PS mixed with the tears readily, and as far as we could judge, the solution appeared identical with the control tears. The break-up time was similar to that of control tears, taking 510 seconds to reach a state of total ocular dryness. The main difference between the PS and the control tears was that it maintained grade 3 stability longer (120 seconds versus 90 seconds; Fig. 6 ). 

Addition of PTBA-PS to the tears plus saline resulted in the formation of a gel-like, highly viscous mixture, the rheology of which no longer resembled normal tears. Its application to the eye resulted in a layer that appeared thicker than that obtained with the other polymers and controls (Figs. 6D 9) . It was highly stable, with the only disruptions occurring after several minutes, which appeared due to evaporation rather than a loss of integrity of the polymer-containing layer. No nodules of epithelial cells were observed during this time. Unlike PTBA-PS, PMA-PAA formed a stable layer of the same appearance and thickness as the tears-plus-saline mixture. Similar to PTBA-PS, the tear film appeared to evaporate before the disruption of the polymer containing film after approximately 660 seconds (Figs. 6D 9) . 

Histology Results

Histologic examination of ocular sections of the corneal epithelium and conjunctiva indicated that none of the test solutions or control solutions applied had initiated an acute cytotoxic response or had caused any lymphocytic or morphologic changes in the ocular tissue. This was also unexpectedly the case for the chloroform control, which might be expected to elicit toxic effects. However, because the chloroform was premixed with tears plus saline before application and we were using very small volumes, a large amount of the chloroform might have evaporated before application. Figure 10 displays a selection of conjunctival and corneal sections, illustrating the typical morphology observed in the sections. 

Discussion

These studies indicate that polymers can be used to stabilize the tear film and that small differences in the structure of the polymers can have a large affect on their fluid properties. It appears that hydrophobic polymers, particularly copolymers, were more effective than the more hydrophilic polymers, which actually had a destabilizing effect on the tear film. 

The most effective tear-film stabilizers were the block copolymers PTBA-PS and PMA-PAA, which have distinct hydrophobic and hydrophilic regions. PTBA-PS, having a more sterically shielded charged region than PMA-PAA was the more hydrophobic of the copolymers, and performed better than PMA-PAA (15 minutes compared with 11.5 minutes). The two polymers are of similar size, so the difference in performance was probably due to structural differences between them. This concept could be explored further by systematically synthesizing polymers that have similar hydrophobic regions, altering the charge and conformation of the hydrophilic section, and determining their interaction with tear film lipids. Such a strategy may also enable development of polymers that have similar properties that stabilize the tear film, but could be dispersed in water rather than in chloroform or methanol, which, even though they did not show acute toxic effects in our studies, would be of great concern if they were to be used as clinical vehicles. 

By contrast, the very hydrophilic polymers PAA and PVA actually accelerated disruption of the tear film. Structurally, they contrasted markedly from the co-block polymers, in that the hydrophilic components were distributed along their entire length. It also appeared that the higher the molecular weight, the more disruptive they were to the tear film. This is in contrast to studies of humans with dry eye where PAA and PVA have been found to increase tear break-up time and improve ocular comfort. ^{32 33} One major difference is that our polymers were mixed with normal rat tears before application, whereas in the human studies the polymers were applied directly to the eye where they mixed with the presumably abnormal tears of the patients with dry eye. The concentration of PVA used in the clinical trials was approximately 10 times more than we used on the rats, but the PAA concentrations used were similar. The concentrations that we chose were known to have surface activity at the air–aqueous interface as determined in-house (by author GRD) using a Langmuir trough. Our data from the experiments also indicate that the size of the PAA molecules is important, and because we do not know the form of the PAA used in the clinical trials, whether it was a single or mixed molecular weight polymer, or whether there was any cross bridging, it is very difficult to judge the similarity between the human and the rat studies. 

That C-PS was more destabilizing than PS gives further support to the notion that hydrophobicity promotes of the tear film stability (Fig. 6) . This was interesting, given that the added hydrophilic component of the molecule was the very small carboxyl terminus. However, because PS gave virtually the same performance as tears plus saline and hence did not increase the tear stability as much as the co-block polymers, it is not simply the hydrophobicity of molecules that leads to greater tear stability. 

Although we have no direct evidence, a possible mechanism for the ability of PTBA-PS and PMA-PAA to stabilize the tear film could be an interaction with the tear film lipids. Some evidence to support this comes from the observation that liposomes are most effectively stabilized by tri-block polymers with both hydrophobic and hydrophilic regions. ^{34 35} Although this interaction is with a phospholipid bilayer, it indicates that such molecules can interact with and affect the structure of lipid complexes. Their possible stabilization effect through the meibomian lipids could be tested directly in the future by examining their effects on the eye after the meibomian orifices are blocked and also by testing their effects on surface pressure changes when mixed with meibum. Another possibility is that they also interact with the proteins or mucous in the tears. We have found that if PTBA-PS is mixed with a saline solution in the same manner as it was mixed with the tears-plus-saline mixture, then there is distinct partitioning with the PTBA-PS locating on the surface and no gel formation. Indeed, that PTBA-PS forms a gel with tears, but not with saline, intuitively suggests that a component of the tear bulk rather than that the tear lipids could be leading to this phenomenon. 

In summary, these experiments demonstrate that amphipathic polymers can be designed to interact with the normal tear film and alter its stability. There is an interesting trend toward more hydrophobic polymers affording greater stability. The availability of methods to control the size and the charge of the blocks used gives enormous scope not only for developing other compounds, but also for subtly changing their structures to maximize their performance. The concept then arises that these polymers could be developed to enhance the performance of normal tears and increase normal ocular comfort. This may be of benefit in professions where long periods of staring are necessary (e.g., looking at a computer screen), or in activities in which airflow could disrupt the tear film. If such polymers were to be considered as possible treatments for increasing ocular comfort or for the treatment of dry eye, their chronic effects on tissue would have to be evaluated. Our initial efforts look promising, because the histology implies that in acute applications to the rat eye, these polymers are not toxic. 

 Supported by the Australian Government through the Co-operative Research Program.

 Submitted for publication September 9, 2002; revised December 20, 2002; accepted January 14, 2003.

 Disclosure: K. Peters, None; G.R. Dennis, None; P.J. Anderton, None; T.J. Millar, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

List of Polymers Used and Their Characteristics

Polymer	Abbreviation	MW	Surfactant Class/Solvent
Carboxyl terminated polystyrene	CPS	2,850	Nonpolar/methanol
Thioester terminated polystyrene	PS	26,634	Nonpolar/chloroform
Poly n-butyl acrylate	PNBA	9,197	Polar/chloroform
Poly tert-butyl acrylate	PTBA	9,293	Polar/methanol
Poly methyl acrylate-block-polyacrylic acid	PMA-PAA	774,530	Polar/strongly polar/methanol
Poly tert-butyl acrylate-block-polystyrene	PTBA-PS	163,245	Polar/nonpolar/chloroform
Poly(vinyl alcohol)	PVA-1	85,000–146,000	Strongly polar/water
Poly(acrylic acid 2)	PAA-2	2,000	Strongly polar/water
Poly(acrylic acid 3)	PAA-3	4,000	Strongly polar/water
Poly(acrylic acid 5)	PAA-5	50,000	Strongly polar/water

Figure 1.

 

Carboxyl-terminated polystyrene (A) has a long hydrophobic chain terminated by a small hydrophilic carboxyl group giving resemblance to a very large fatty acid. Polystyrene (B) is an uncharged molecule, although the terminal dithiobenzoate thioester group is surface active. In this and subsequent figures black represents oxygen, grey represents carbon, and white represents hydrogen.

Figure 2.

 

The copolymer PMA- PAA (A) is polar block linked to a strongly polar block, which contrasts with PTBA-PS (B) which has a polar block linked to a very nonpolar block.

Figure 3.

 

PAA (A), and PVA (B), which are both strongly hydrophilic, with PAA being negatively charged through the negative charge on the carboxyl group. This is pH-dependent but mostly ionic at physiological pH. PVA is hydrophilic due to its uncharged hydroxyl groups.

Figure 4.

 

The ester head group of PNBA (A) is shielded by an n-butyl group. The structurally similar PTBA (B) differs, in that the ester head group is shielded by tertiary butyl groups, which should have different steric shielding.

Figure 5.

 

Specular microscopy of ocular surface of rat eye illustrating the grading scale of ocular dryness from 0, complete tear film, to 5, dry ocular surface.

Figure 6.

 

Plot comparing the effects of different substances on ocular drying, using the grading scale illustrated in Figure 5 . (A) Compiled data in rank order for all solutions applied to the eye, from the least to the most stable. (B–D) Full data for related sets of solutions. Comparisons are shown of (B) control solutions, (C) polystyrene with carboxyl terminal polystyrene, and (D) the co-block polymers PMA-PAAC and PTBA-PS.

Figure 7.

 

Specular microscopy of ocular surface of rat eye comparing the break-up of saline plus normal tears (left) with saline alone (right). Time after initial application is indicated in each panel in seconds.

Figure 8.

 

Specular microscopy of the ocular surface of rat eye comparing the effects on tear break-up time of the water-soluble polymers PAA-2 (left) and PAA-5 (right). The two polymers differ from one another only by molecular weight. Note the difference in the break-up and apparent instability of the higher molecular weight PAA-5 tear solution. Time after initial application is indicated in each panel in seconds.

Figure 9.

 

Specular microscopy of the ocular surface of rat eye comparing the effects on tear break-up time of the copolymers PMA-PAA (left) and PTBA-PS (right). Note that the film remained flat and unbroken after a long period. Time after initial application is indicated in each panel in seconds.

Figure 10.

 

Micrographs of frozen sections of rat cornea stained with hematoxylin and eosin. Panel 1: peripheral limbus after application of ion exchange purified water (control eye). Panels 2 and 3: the corneal epithelium after C-PS and PTBA were applied to the ocular surface. There were no obvious morphologic changes that indicated acute inflammation or toxicity.

We would like to thank Sophia Tragoulias for assistance and advice on setting up the animal model.