Skip to main content
NCBI home page
Clinical and Experimental Immunology logoLink to Clinical and Experimental Immunology
. 2002 Apr;128(1):67–74. doi: 10.1046/j.1365-2249.2002.01808.x

Penetration of engineered antibody fragments into the eye

M A THIEL *, D J COSTER *, S D STANDFIELD *, H M BRERETON *, C MAVRANGELOS †, H ZOLA †, S TAYLOR ‡, A YUSIM ‡, K A WILLIAMS *
PMCID: PMC1906367  PMID: 11982592

Abstract

Antibodies are powerful immunotherapeutic agents but their use for treating ocular disorders is limited by their poor penetration into the eye. We hypothesized that antibody fragments of relatively small size might penetrate the cornea more readily. Monovalent single chain variable region (scFv) antibody fragments and divalent miniantibodies were engineered from existing monoclonal antibodies, expressed in a bacterial expression system, and purified by metal ion affinity chromatography. Corneoscleral preparations from normal pig and cat eyes were mounted in a corneal perfusion chamber. Intact antibodies and antibody fragments were applied topically to the anterior corneal surface over 12-h periods, and samples were collected from the artificial anterior chamber. Similar experiments were performed with whole enucleated pig and human eyes. Penetration of antibodies and fragments was quantified by high-sensitivity flow cytometry on appropriate target cells. Both monovalent scFv and divalent miniantibody fragments (but not whole immunoglobulin molecules) passed through de-epithelialized and intact corneas after topical administration, and could be detected by antigen binding. Addition of 0·5% sodium caprate facilitated penetration through intact corneas. Topically-applied scFv was found to penetrate into the anterior chamber fluid of rabbit eyes in vivo. The engineered fragments were stable and resistant to ocular proteases. Monovalent and divalent antibody constructs of molecular weight 28 kD and 67 kD, respectively, can penetrate through intact corneas into the anterior chamber, with retention of appropriate antigen-binding activity. Such constructs may form novel therapeutic agents for topical ophthalmic use.

Keywords: engineered antibody fragments, cornea, immunotherapeutics

INTRODUCTION

Antibodies are important therapeutic agents in clinical medicine [1,2]. Polyclonal and monoclonal antibodies are known to modulate corneal graft rejection [3,4], ocular autoimmune disease [5,6] and some ocular infections [7] when administered systemically in animal models. However, topical rather than systemic therapy is the preferred route for drug administration to the cornea and anterior segment of the eye. Extremely high local drug concentrations can be achieved with minimal risk of systemic side effects. While topical therapy for ophthalmic conditions is conceptually attractive and ethically appropriate, it requires that the drug of choice be able to penetrate into the eye at a clinically useful rate [8]. Topical administration of antibody to modulate immunopathological conditions of the cornea or anterior segment is ineffective because such molecules are too large to penetrate the cornea rapidly [9].

Over the past 10 years, techniques of genetic engineering have allowed the isolation of the genes encoding the variable light (VL) and heavy (VH) domains of antibody-binding sites and their covalent linkage into the molecule known as a single chain variable-region fragment, scFv [10]. ScFvs lack all of the constant-region domains of whole antibodies and are relatively small (28 kD). Two scFv fragments can be linked with a highly flexible hinge peptide to form a divalent miniantibody (67 kD) [11]. Monovalent and divalent fragments have shown excellent tissue penetration and some therapeutic potential when administered systemically [12,13]. Were antibody fragments to be able to penetrate the ocular surface, they might form a useful new class of therapeutic drugs for ophthalmic purposes.

Almost all drugs in preparations formulated for topical ophthalmic use have a molecular weight of less than 400 D [8]. Strategies to enhance ocular drug penetration have been extensively reviewed [8,14,14,15,16,. In general, the bioavailability of topically-applied drugs can be improved by either increasing drug lipophilicity [14,17], or increasing permeability of the superficial epithelial barrier through co-administration of penetration enhancers [18–22].

The aims of this in vitro study were to examine the in vitro penetration of engineered antibody fragments through the cornea of the pig, cat and human, to evaluate the effect of various penetration enhancers, and to investigate the stability of scFvs and miniantibodies on the ocular surface and in the aqueous humour. In vivo penetration of antibody fragments into rabbit eyes was then examined.

MATERIALS AND METHODS

Parent antibody and engineered antibody constructs

ScFvs and miniantibodies were generated from the OX38 hybridoma (European Collection of Animal Cell Cultures, Porton Down, Wiltshire, UK), which secretes mouse anti-rat CD4 IgG2a antibody [23], using techniques described by Krebber et al.[24]. Briefly, mRNA isolated from the hybridoma was amplified by reverse transcription polymerase chain reaction (PCR). The genes for VL and VH domains were amplified separately and joined with a lipophilic 20-amino acid linker sequence (gly4ser)4. The scFv gene was cloned into the pAK100 vector and transformed into Escherichia coli strain JM83. Culture supernatant fluids were tested for the presence and activity of scFvs by slot-blot and flow cytometry, respectively. Clones expressing functional scFv were transferred to pAK400 for scFv expression, and into pAK500 for miniantibody expression. ScFvs and miniantibodies were expressed in E coli and active protein extracted from the bacterial periplasm [25]. Antibody fragments were purified over an immobilized metal affinity column using the poly histidine tag engineered into the expression vectors [26]. For in vivo experiments, bacterial endotoxin was removed by passage over Q-Sepharose resin (Pharmacia Biotech, Uppsala, Sweden) [27].

Antibody and antibody fragments for topical application

OX38 hybridoma culture supernatant fluid containing IgG at a concentration detectable at a dilution of 1 in 30 000 by flow cytometry was used as the control eye drop. ScFvs were dialysed against Dulbecco’s A phosphate-buffered saline (PBS) solution, pH 7·2, and filter-sterilized. For some experiments, pH was adjusted to 8·0 or 3·5. Miniantibodies were dialysed against 150 mm NaCl, 20 mm HEPES buffer, pH 8·0. ScFv and miniantibody concentrations were 0·2 mg/ml and 1·85 mg/ml, respectively, as determined by optical density at 280 nm. Benzalkonium chloride, dimethyl sulphoxide, imidazole, dihydrocytochalasin B, digitonin and capric acid sodium salt were purchased from Sigma Chemical Company, St Louis, MO, USA. All penetration enhancers were added to the solution containing antibody fragments just prior to the experiment. To investigate the risk of degradation of OX38 scFv and miniantibodies by rat serum proteases, dilution series of each construct were prepared in fresh normal rat serum and incubated at 37°C for 72 h. For in vivo administration to rabbits, the scFv was dissolved in 150 mm NaCl, 10 mm HEPES buffer, pH 7·5, supplemented with 0·5% sodium caprate and 1·5% hydroxypropyl methylcellulose (Dow Chemical Pacific Ltd, Marleston, South Australia, Australia) to a final concentration of 0·8 mg/ml. The endotoxin level of this preparation was less than 2·2 endotoxin units/ml as measured by the Limulus Amebocyte lysate test (Bio Whittaker, Walkerville, MD, USA).

Ocular tissues and experimental animals

Twenty-six normal pig eyes obtained from a local abattoir were used within 2–3 h of enucleation. Some pig corneas were de-epithelialized by mechanical debridement with a number 11 blade. Two cat eyes were harvested from normal laboratory animals killed by a separate group of investigators and were used within 1–2 h. Adult Dutch-Belted rabbits obtained from Nanowie Small Animal Production Unit, Bellbrae, Victoria, Australia were housed individually in approved cages and allowed unlimited access to rabbit chow and water. All experimentation was carried out with approval from the institutional Animal Welfare Committee. Two human eyes retrieved for human transplantation but found unsuitable for the purpose were used for research with the informed consent of the donor families, with institutional approval, and in accordance with the tenets of the Declaration of Helsinki. Slit lamp examination revealed normal, phakic eyes with an intact corneal epithelium and no signs of oedema.

Penetration experiments on perfused, isolated corneoscleral preparations

Corneoscleral preparations were dissected from whole eyes and mounted in a corneal perfusion chamber as previously described [28]. Isolated corneas were perfused with ocular irrigating solution (BSS-Plus, Alcon Laboratories, Frenchs Forest, NSW, Australia) at a flow-rate of 1 ml/min. The perfusate from the outflow tube was collected in a reservoir and recirculated back into the artificial anterior chamber. The total fluid volume recirculating through the chamber, tubing and reservoir was 4 ml. The reservoir was elevated 25 cm above the level of the clamped cornea to create a positive pressure of 18 mmHg (2·4kPa) inside the artificial anterior chamber. The perfusion chamber was heated to maintain a chamber temperature of approximately 35°C. One 50 μl drop of BSS-Plus (negative control) or solution of whole antibody or antibody fragments was applied topically to the corneal surface every 20 min over the time course of the experiment. Every hour, 220 μl of the perfusate was removed from the perfusion reservoir for testing and replaced with the same volume of fresh BSS-Plus.

Penetration experiments on intact eyes

The whole eye was positioned on a Teflon block with a depression tooled into the surface, such that the downward-facing cornea was placed into a well [29]. The well was partly filled with 100 μl of the antibody construct to be tested. The eye was covered with a plastic container to form a humid chamber and placed on a 37°C heating block. At the end of the experiment, the globe was washed six times with BSS-Plus. Anterior chamber fluid was collected by paracentesis using a 27 G needle at one time-point per eye, and the eye was then discarded. Samples of vitreous were also collected from the human eyes through a posterior paracentesis.

Penetration experiments in live rabbits

The left eye of each of six rabbits was treated with an antibody preparation (scFv or OX38 antibody) by topical application of one 50 μl drop to the cornea every 20 min for up to 12 h. Peripheral blood was collected from the marginal ear vein at the end of the experiment and the serum separated. Rabbits were killed at various time points by intravenous administration of 5 ml 325 mg/ml veterinary sodium pentobarbitone (Arnolds of Reading Pty Ltd, Boronia, Victoria, Australia) through the marginal ear vein. Samples of anterior chamber fluid were collected by paracentesis after irrigation of the surface of the eye with sterile 20 ml normal saline. In one instance, a 50 μl drop of scFv was applied to the otherwise untreated eye of a rabbit immediately after euthanasia. After 30 s, the eye was irrigated and sampled as before. All samples were supplemented with 20 mm sodium azide.

Measurement of antibody and antibody fragment concentration

Binding activity of OX38 scFv or OX38 miniantibody was measured by high-sensitivity immunofluorescence detected by flow cytometry [30]. Binding of OX38 IgG antibody was measured by standard indirect immunofluorescence detected by flow cytometry. Target cells were human Jurkat cells transfected for surface expression of full-length rat CD4 antigen (line J5.CD4) or the untransfected control line (both from the Medical Research Council, Oxford, UK), or normal rat thymocytes purified over a Lymphoprep (Nycomed Pharma, Oslo, Norway) cushion. For the scFv or miniantibody, 50 μl of cell suspension at 2 × 107 cells/ml in PBS containing 20 mm sodium azide (PBS-azide) were incubated with 50 μl antibody or perfusate at 4°C for 30 min. Cells were washed with PBS-azide and incubated in 50 μl 1/250 dilution of anti-PolyHIS antibody (Sigma) in PBS-azide for 30 min at 4°C. Cells were washed as before and resuspended in 50 μl 1/100 dilution biotinylated goat anti-mouse antibody (DAKO Corporation, Carpinteria, CA, USA) in PBS-azide for 30 min at 4°C. After washing, the cells were incubated with 50 μl 1/50 dilution streptavidin-phycoerythrin (Sigma) in PBS-azide for 30 min at 4°C. Finally, cells were fixed in 50 μl 5% v/v formaldehyde, 10 mm glucose in PBS-azide and analysed within 24 h in a FACScan flow cytometer (Becton Dickinson, Mountain View, CA, USA). All assays were performed in duplicate. Mean fluorescence intensity (MFI) was used as a relative quantitative measure for antibody or antibody fragment concentration after penetration through the cornea by comparison with titrations of purified preparations of known concentration.

Measurement of corneal viability

Tissue viability was assessed every hour during all perfusion experiments by measuring corneal thickness with a hand-held ultrasonic pachymeter (B.V. International, Clermont-Ferrand, France) to ensure that corneal thickness did not vary from baseline levels by more than 15% during perfusion. At the end of each experiment, corneo-scleral buttons or whole eyes were fixed in 10% buffered formalin in PBS and paraffin-embedded. Multiple 10 μm sections taken across the cornea were stained with periodic acid-Schiff (PAS).

RESULTS

Reactivity and stability of engineered antibody fragments with target cells

By flow cytometry, the OX38 scFv and miniantibody reacted with CD4-positive rat thymocytes (Fig. 1a) and with a CD4-positive transfected cell line, but not with a control, non-transfected cell line (Fig. 1b). The scFv fragment was stable on storage in vitro for at least 12 months (Fig. 2a), and was resistant to rat serum proteases for periods of up to 72 h (Fig. 2b), in that binding activity to target cells remained essentially unchanged.

Fig. 1.

Fig. 1

Open in a new tab

Representative flow cytometry histograms of reactivity of antibody constructs with (a) normal rat thymocytes and (b) untransfected or rat CD4-transfected Jurkat cells. OX38: parental whole anti-rat CD4 IgG; scFv: engineered monovalent anti-rat CD4 scFv; miniab: engineered divalent anti-rat CD4 miniantibody; PBS: Dulbecco’s A phosphate-buffered saline; MFI: mean fluorescence intensity.

Fig. 2.

Fig. 2

Open in a new tab

Stability of engineered antibody constructs over time, and in the presence of normal rat serum. Antibody constructs were titrated against CD4-positive normal rat thymocytes by flow cytometry. Mean fluorescence intensity was used as a relative index of stability. (a) The scFv exhibited stable reactivity for at least 12 months during storage at room temperature. (•), At start; (○) after 6 weeks; (▾) after 3 months; (▿) after 4 months; (■) after 6 months; (□) after 12 months. (b) Both the scFv and the miniantibody (miniab) remained active when stored for 72 h at 37°C in fresh normal rat serum. Each point represents the mean of duplicate determinations. (▾) scFv in PBS, 4°C; (▿) scFv in rat serum, 37°C; (•) miniab in PBS, 4°C; (○) miniab in rat serum, 37°C.

Antibody fragment penetration through de-epithelialized pig and cat corneas

Penetration of OX38 scFv, OX38 miniantibody and whole OX38 IgG through three de-epithelialized pig corneas was tested using corneal perfusion chambers. Each construct was applied topically three times per hour for up to 14 h and the perfusate sampled regularly for antigen-binding activity on rat thymocytes by flow cytometry. In every experiment, rat thymocytes were also tested against PBS (negative control) and OX38 scFv (positive control) and reacted appropriately. ScFv and miniantibody fragments penetrated through the pig cornea into the artificial anterior chamber and retained their specific activity (Fig. 3a). There was a 4 h lag-time between the first topical application of fragments to the cornea and the detection of antigen-binding activity in the perfusate. Antigen-binding activity of each fragment increased steadily with time and in the case of the miniantibody, the detection system was saturated after approximately 8 h. In contrast, whole OX38 IgG remained undetectable inside the penetration chamber throughout the 14 h observation period. The experiment was repeated with two de-epithelialized cat corneas (Fig. 3b). Results were almost identical. ScFv binding activity after penetration through the cat cornea became detectable after 3 h. Whole OX38 antibody did not penetrate through the cornea.

Fig. 3.

Fig. 3

Open in a new tab

Whole OX38 IgG antibody and OX38 scFv penetration through (a) de-epithelialized pig corneas and (b) de-epithelialized cat corneas. Constructs were applied topically three times per hour. Penetration was measured as antigen-binding activity on rat thymocytes by flow cytometry. Results were expressed as mean fluorescence intensity (MFI), where a MFI = 50 corresponds to approximately 40 ng OX38 scFv. Each line represents data from a single corneoscleral preparation, and each point represents the mean of duplicate determinations. (■) scFv; (▴) miniantibody; (•) OX38 IgG.

Antibody fragment penetration through pig corneas with intact epithelium

Intact pig corneas with the epithelium present were then perfused. In an initial experiment with OX38 scFv and with the miniantibody at pH 7·2, no penetration of either construct was observed (data not shown). In an attempt to facilitate epithelial penetration of scFv fragments, the penetration enhancers 0·04% w/v benzalkonium chloride, 5% v/v dimethyl sulphoxide, 0·015% w/v imidazole, 0·1 mm dihydrocytochalasin B alone or in combination with 0·1 mm digitonin, and 0·5% w/v sodium caprate were investigated. The pH of the scFv solution was also increased from 7·2 to 8·0, and decreased to 3·5. A separate pig cornea was used for each formulation. With most formulations, no antigen-binding activity could be detected in the artificial anterior chamber following topical application of scFvs, three times per hour for up to 12 h (Fig. 4). The binding activity data for scFv formulated in 0·04% benzalkonium chloride, 5% dimethyl sulphoxide, 0·015% imidazole, 0·1 mm dihydrocytochalasin B, 0·1 mm dihydrocytochalasin B plus 0·1 mm digitonin, or the scFv at pH 3·5, were thus pooled. Best results were obtained with the addition of 0·5% sodium caprate, with binding activity in the artificial anterior chamber first being observed after 6 h and increasing steadily thereafter (Fig. 4). Increasing the pH from 7·2 to 8·0 also resulted in detection of some scFv binding activity in the perfusate after 8 h (Fig. 4). Whole OX38 IgG failed to penetrate into the artificial anterior chamber at any time (data not shown).

Fig. 4.

Fig. 4

Open in a new tab

Penetration of OX38 scFv through pig corneas with intact epithelium. ScFv fragments were formulated at pH 3·5, or with 0·04% benzalkonium chloride, 0·5% dimethyl sulphoxide, 0·015% imidazole, or 0·1 mm dihydrocytochalasin B alone or with 0·1 mm digitonin (pooled data). ScFv were also formulated at pH 8·0 or in 0·5% sodium caprate (shown separately). (•) Corneas + scFv + enhancer (pooled data, n = 6); (■) cornea + scFv pH = 8·0; (▴) cornea + scFv pH = 8·0 + 0·5% sodium caprate.

Penetration of miniantibody into the intact pig and human eye

Penetration of the miniantibody construct into 12 whole pig eyes was investigated. For these experiments, only one datum point per eye was possible because the eye had to be punctured to collect the aqueous humour sample from the intact anterior chamber. In five pig eyes without epithelium, antigen-binding activity was detected in the aqueous humour at 3 h and thereafter (Table 1). Seven pig eyes with intact epithelium were then incubated with miniantibody supplemented with 0·5% sodium caprate. Miniantibody binding activity in the anterior chamber fluid of these eyes became detectable after 5 h and increased rapidly thereafter (Table 1).

Table 1.

Penetration of OX38 miniantibody into the intact pig eye in vitro

Mean fluorescence intensity in aqueous humour after topical application for cumulative time (h)
Specimen Formulation applied 0 2 3 5 6 7 8
Whole pig eye, de-epithelialized miniantibody, pH 8·0 0* NT 16 NT 24 27 41
Whole pig eye, epithelium intact miniantibody with0·5% sodium caprate 0 2 1 5 25 23 34
Open in a new tab
*

Each datum point represents a separate eye.

NT, not tested.

Only two human eyes were available for study. One de-epithelialized eye was incubated with OX38 antibody to confirm that whole IgG did not penetrate the cornea, even in the absence of epithelium. The MFI measured in the aqueous humour after 5 h was <2. The second eye, which was not de-epithelialized, was incubated for 5 h with miniantibody supplemented with 0·5% sodium caprate. Binding activity in the anterior chamber was within the range of test saturation, with an MFI of 30 (data not shown). Titration of the aqueous humour sample indicated a miniantibody concentration of 0·8 μg/ml when compared with a standard of known concentration. No binding activity was detected in the vitreous sample collected from either eye.

In vivo penetration of topically-applied scFv into the intact rabbit eye

Penetration of the scFv construct into eyes in vivo was tested in the rabbit (Table 2). Antibody formulations were applied topically three times per hour for up to 12 h to one eye only of any rabbit, and anterior chamber fluid samples were collected immediately after euthanasia. ScFv was detectable in anterior chamber fluid of treated rabbit eyes 4 h after initiation of treatment, and thereafter to 12 h. The relatively poor penetration observed in the single animal treated for 8 h was otherwise unaccountable. The concentration of scFv in aqueous at 12 h was estimated to be approximately 3 μg/ml by titration of the sample. One rabbit was treated every 20 min for 12 h and then treatment ceased, but the animal was not euthanazed. This rabbit was killed at 12 h after cessation of treatment; little or no binding activity was detected in anterior chamber fluid at this time. To control for possible contamination of ocular fluid samples during the collection process, one otherwise untreated eye was treated topically with the scFv preparation immediately after euthanasia of the rabbit, and the sample collected 30 s later, at a time at which penetration of the scFv into the anterior chamber would not be anticipated. No binding activity was detected in anterior chamber fluid. No binding activity was detected in anterior chamber fluid collected from any contralateral untreated eye, and endpoint serum samples were also uniformly negative. One eye of one rabbit was treated for 12 h with whole OX38 IgG; no activity was found in the anterior chamber fluid of this eye. All rabbits tolerated repeated topical application of the antibody formulations very well without any signs of irritation or pain, and all corneas remained crystal clear throughout the experiment.

Table 2.

Penetration of OX38 scFv into the intact rabbit eye in vivo

Mean fluorescence intensity in sample after topical application for cumulative time (h)
Formulation applied Specimen 0 30 s 4 6 8 12 24
ScFv anterior chamber 3 4 266 300 23 450 12
fluid serum 4 10 15 15 10 20 12
Whole IgG anterior chamber NT NT NT NT NT 5 NT
fluid serum NT 6 NT NT NT 6 NT
Open in a new tab

End-point histology on perfused corneas and on whole globes

End-point histology was performed on all corneas. In perfused pig corneas with epithelium intact at the start of the experiment, the superficial epithelial lamellae were, in most instances, well preserved (Fig. 5a,b). In particular, the histological appearance of the pig corneal epithelium was largely unaffected by 0·5% sodium caprate (Fig. 5c). Similarly, end-point histology of the human cornea showed no signs of toxic or mechanical damage to the epithelium. Histology of rabbit corneas harvested after in vivo administration of the scFv showed normal corneal architecture (not shown).

Fig. 5.

Fig. 5

Open in a new tab

End-point histology of representative corneas after 12 h perfusion and exposure to antibody fragments applied topically three times per hour. (a) Pig cornea treated with topical BSS (control); (b) pig cornea treated with OX38 scFv, pH 8·0 without penetration enhancer; (c) pig cornea treated with scFv with 0·5% w/v sodium caprate. All sections were stained with PAS, ×500 magnification.

DISCUSSION

The corneal epithelium, stroma and endothelium all form bar-riers to intraocular drug penetration. Molecular shape and hydrophobicity, as well as molecular size, influence the intraocular penetration of proteins. Molecules as large as dextran (60– 90 kD) have been shown to penetrate the de-epithelialized cornea [31]. However, the corneal endothelial tight junctions form a strong penetration barrier for the approximately spherical human albumin molecule (64 kD) [9]. That antibody fragments of 28– 67 kD would penetrate through the cornea was thus not intuitively obvious.

Our in vitro data indicated that both an scFv (28 kD) and a miniantibody (67 kD) could penetrate through the de-epithelialized cornea into the anterior chamber within 4–8 h after topical application to the ocular surface, a useful time-frame for an ocular therapeutic. In contrast, a whole IgG antibody (146 kD) used as a control for leakage of macromolecules through the cornea showed no detectable penetration through the cornea over 14 h. Each of the two scFv fragments that form a miniantibody has a rod shape of approximately 3·0 × 3·5 nm [11]. Miniantibody penetrated the cornea in a time-frame similar to that of the smaller scFv fragment, suggesting that the helical hinge region of the miniantibody is flexible enough to allow the individual scFv domains to penetrate the cornea.

ScFv and miniantibody fragments also penetrated through corneas with intact epithelium, albeit slowly. Various penetration enhancers were investigated for their ability to increase the epithelial permeability of the antibody fragments. Benzalkonium chloride, dimethyl sulphoxide, imidazole, dihydrocytochalasin B and digitonin were virtually without effect. Lowering the pH of the formulation to 3·5 was also ineffective, although raising the pH to 8·0 enhanced penetration to some degree. The best enhancement was observed with 0·5% sodium caprate. An eye drop formulation which included 0·5% sodium caprate resulted in penetration of the scFv through the perfused pig cornea, and of the miniantibody into the anterior chamber of whole pig eye. Capric acid (decanoic acid) is a medium chain (C10) fatty acid and a normal constituent of milk fat and plant oils [32]. It has previously been shown to improve the penetration of macromolecules through the cornea [33–35]. End-point histology of corneas exposed to sodium caprate indicated that the corneal epithelium was well preserved.

The stability of proteins designed for therapeutic use is an important practical consideration. Our scFv construct proved remarkably stable on storage. Furthermore, the scFv withstood the presence of serum proteases for at least 72 h at 37°C. Binding activity was easily measured in perfusate fluid after 12 h, and the aqueous humour samples obtained at the end of the penetration experiments using whole eyes were stored overnight at room temperature before antigen-binding activity was measured. The detection of antigen-binding activity after this time showed that the constructs must have been stable on the ocular surface and within the corneal stroma, and were resistant to degradation by aqueous humour proteases.

Our in vitro studies established that appropriately-formulated scFv and miniantibody fragments could penetrate through the cornea after topical application. However, we could not extrapolate from our in vitro data to estimate the ocular bioavailability of scFv or miniantibody in vivo. The concentration after penetration was measured indirectly by assessing antigen-binding activity by flow cytometry. This provides stringent evidence of the presence of an intact, functional, antigen-binding site. However, the dilution factor resulting from the relatively large volume of perfusate in experiments in which the perfusion chamber was used was substantial, and mean fluorescence intensities were correspondingly low. The complex detection system together with factors such as potential protein degradation by the peristaltic pump, as well as protein adhesion to the chamber and tubing, made it impossible to estimate bioavailability reliably from these experiments. Topical application of the scFv to rabbit eyes in vivo according to a clinically-relevant protocol, however, indicated that antibody fragments (but not whole antibody) could pass into the eye and could be readily detected in anterior chamber fluid over a reasonable time-span, despite continuous turnover of aqueous.

In summary, our data show that scFv and miniantibodies will penetrate the corneal epithelium, stroma and endothelium in vitro, and through the anterior surface of the eye in vivo. Penetration can be enhanced by sodium caprate, while commonly-used penetration enhancers such as benzalkonium chloride are ineffective. ScFv fragments and miniantibodies resist enzymatic degradation on the ocular surface, within the cornea and in anterior chamber fluid. Engineered antibody fragments are thus potentially useful therapeutic agents for topical use in the eye.

Acknowledgments

The authors thank Mrs M. Philpott and Ms K. Marshall (Department of Ophthalmology, Flinders University) for provision of human corneas, and for technical assistance with experimental animals, respectively. The authors acknowledge Ms L Kirk, Mr G. Dumsday and Mr C. Cook for production of antibody fragments, Dr A. Meister and Dr D. Gearing for expert advice (all from CSL Research & Development), and Professor A Plückthun (University of Zurich) for generous provision of some vectors. This work was supported by the NH & MRC of Australia, the Ophthalmic Research Institute of Australia, the Flinders Medical Centre Foundation, the Swiss National Science Foundation, the Swiss Foundation for Medical and Biological Scholarships, EMDO Foundation and the Swiss Foundation for the Prevention of Blindness.

REFERENCES

  • 1.Cosimi AB, Burton RC, Colvin RB, et al. Treatment of acute renal allograft rejection with OKT3 monoclonal antibody. Transplantation. 1981;32:535–9. doi: 10.1097/00007890-198112000-00018. [DOI] [PubMed] [Google Scholar]
  • 2.Breedveld FC. Therapeutic monoclonal antibodies. Lancet. 2000;355:735–40. doi: 10.1016/s0140-6736(00)01034-5. [DOI] [PubMed] [Google Scholar]
  • 3.Ayliffe W, Alam Y, Bell EB, et al. Prolongation of rat corneal graft survival by treatment with anti-CD4 monoclonal antibody. Br J Ophthalmol. 1992;76:602–6. doi: 10.1136/bjo.76.10.602. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.He YG, Ross J, Niederkorn JY. Promotion of murine orthotopic corneal allograft survival by systemic administration of anti-CD4 monoclonal antibody. Invest Ophthalmol Vis Sci. 1991;32:2723–8. [PubMed] [Google Scholar]
  • 5.Thurau SR, Wildner G, Reiter C, et al. Treatment of endogenous uveitis with anti-CD4 monoclonal antibody: first report. German J Ophthalmol. 1994;3:409–13. [PubMed] [Google Scholar]
  • 6.Isaacs JD, Hale G, Waldmann H, et al. Monoclonal antibody therapy of chronic intraocular inflammation using Campath-1H. Br J Ophthalmol. 1995;79:1054–5. doi: 10.1136/bjo.79.11.1054. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Su YH, Yan XT, Oakes JE, et al. Protective antibody therapy is associated with reduced chemokine transcripts in herpes simplex virus type 1 corneal infection. J Virol. 1996;70:1277–81. doi: 10.1128/jvi.70.2.1277-1281.1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Prausnitz MR, Noonan JS. Permeability of cornea, sclera, and conjunctiva: a literature analysis for drug delivery to the eye. J Pharm Sci. 1998;87:1479–88. doi: 10.1021/js9802594. 10.1021/js9802594. [DOI] [PubMed] [Google Scholar]
  • 9.Allansmith M, de Ramus A, Maurice D. The dynamics of IgG in the cornea. Invest Ophthalmol Vis Sci. 1979;18:947–55. [PubMed] [Google Scholar]
  • 10.Hudson PJ. Recombinant antibody fragments. Curr Opin Biotechnol. 1998;9:395–402. doi: 10.1016/s0958-1669(98)80014-1. [DOI] [PubMed] [Google Scholar]
  • 11.Plückthun A, Pack P. New protein engineering approaches to multivalent and bispecific antibody fragments. Immunotechnology. 1997;3:83–105. doi: 10.1016/s1380-2933(97)00067-5. 10.1016/s1380-2933(97)00067-5. [DOI] [PubMed] [Google Scholar]
  • 12.Yokota T, Milenic DE, Whitlow M, et al. Rapid tumor penetration of a single-chain Fv and comparison with other immunoglobulin forms. Cancer Res. 1992;52:3402–8. [PubMed] [Google Scholar]
  • 13.Larson SM, El-Shirbiny AM, Divgi C, et al. Single chain antigen binding protein (sFv CC49). First human studies in colorectal carcinoma metastatic to liver. Cancer. 1997;80:2458–68. doi: 10.1002/(sici)1097-0142(19971215)80:12+<2458::aid-cncr17>3.3.co;2-t. 10.1002/(sici)1097-0142(19971215)80:12+<2458::aid-cncr17>3.3.co;2-t. [DOI] [PubMed] [Google Scholar]
  • 14.Sasaki H, Yamamura K, Mukai T, et al. Enhancement of ocular drug penetration. Crit Rev Ther Drug Carrier Syst. 1999;16:85–146. [PubMed] [Google Scholar]
  • 15.Aiache JM, el Meski S, Beyssac E, et al. The formulation of drug for ocular administration. J Biomater Appl. 1997;11:329–48. doi: 10.1177/088532829701100305. [DOI] [PubMed] [Google Scholar]
  • 16.Mishima S. Clinical pharmacokinetics of the eye. Invest Ophthalmol Vis Sci. 1981;21:504–41. [PubMed] [Google Scholar]
  • 17.Maurice D, Mishima S. Ocular pharmacokinetics. In: Sears M, editor. Pharmacology of the eye. Berlin: Springer-Verlag; 1984. pp. 19–116. [Google Scholar]
  • 18.Keller N, Moore D, Carper D, et al. Increased corneal permeability induced by the dual effects of transient tear film acidification and exposure to benzalkonium chloride. Exp Eye Res. 1980;30:203–10. doi: 10.1016/0014-4835(80)90114-1. [DOI] [PubMed] [Google Scholar]
  • 19.Sasaki H, Nagano T, Yamamura K, et al. Ophthalmic preservatives as absorption promoters for ocular drug delivery. J Pharm Pharmacol. 1995;47:703–7. doi: 10.1111/j.2042-7158.1995.tb06726.x. [DOI] [PubMed] [Google Scholar]
  • 20.Marsh RJ, Maurice DM. The influence of non-ionic detergents and other surfactants on human corneal permeability. Exp Eye Res. 1971;11:43–8. doi: 10.1016/s0014-4835(71)80063-5. [DOI] [PubMed] [Google Scholar]
  • 21.Tang Liu DD, Richman JB, Weinkam RJ, et al. Effects of four penetration enhancers on corneal permeability of drugs in vitro. J Pharm Sci. 1994;83:85–90. doi: 10.1002/jps.2600830120. [DOI] [PubMed] [Google Scholar]
  • 22.Sasaki H, Igarashi Y, Nagano T, et al. Different effects of absorption promoters on corneal and conjunctival penetration of ophthalmic beta-blockers. Pharm Res. 1995;12:1146–50. doi: 10.1023/a:1016203725128. [DOI] [PubMed] [Google Scholar]
  • 23.Jefferies WA, Green JR, Williams AF. Authentic T helper CD4 (W3/25) antigen on rat peritoneal macrophages. J Exp Med. 1985;162:117–27. doi: 10.1084/jem.162.1.117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Krebber A, Bornhauser S, Burmester J, et al. Reliable cloning of functional antibody variable domains from hybridomas and spleen cell repertoires employing a reengineered phage display system. J Immunol Methods. 1997;201:35–55. doi: 10.1016/s0022-1759(96)00208-6. [DOI] [PubMed] [Google Scholar]
  • 25.Pack P, Kujau M, Schroeckh V, et al. Improved bivalent minianti-bodies, with identical avidity as whole antibodies, produced by high cell density fermentation of Escherichia coli. Biotechnology. 1993;11:1271–77. doi: 10.1038/nbt1193-1271. [DOI] [PubMed] [Google Scholar]
  • 26.Plückthun A, Krebber A, Krebber C, et al. Antibody engineering: a practical approach. Oxford: Oxford University Press; 1996. Producing antibodies in Escherichia coli: from PCR to fermentation; pp. 203–52. [Google Scholar]
  • 27.Colangeli R, Heijbel A, Williams AM, et al. Three-step purification of lipopolysaccharide-free, polyhistidine-tagged recombinant antigens of Mycobacterium tuberculosis. J Chromatogr B Biomed Sci Appl. 1998;714:223–35. doi: 10.1016/s0378-4347(98)00094-2. [DOI] [PubMed] [Google Scholar]
  • 28.Thiel MA, Morlet N, Schulz D, et al. A simple corneal perfusion chamber for drug penetration and toxicity studies. Br J Ophthalmol. 2001;85:450–3. doi: 10.1136/bjo.85.4.450. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Mishima S, Kudo T. In vitro incubation of rabbit cornea. Invest Ophthalmol. 1967;6:329–39. [Google Scholar]
  • 30.Zola H, Neoh SH, Mantzioris BX, et al. Detection by immunofluorescence of surface molecules present in low copy numbers. High sensitivity staining and calibration of flow cytometer. J Immunol Methods. 1990;135:247–55. doi: 10.1016/0022-1759(90)90278-4. [DOI] [PubMed] [Google Scholar]
  • 31.Kim JH, Green K, Martinez M, et al. Solute permeability of the corneal endothelium and Descemet’s membrane. Exp Eye Res. 1971;12:231–8. doi: 10.1016/0014-4835(71)90143-6. [DOI] [PubMed] [Google Scholar]
  • 32.Soderholm JD, Oman H, Blomquist L, et al. Reversible increase in tight junction permeability to macromolecules in rat ileal mucosa in vitro by sodium caprate, a constituent of milk fat. Dig Dis Sci. 1998;43:1547–52. doi: 10.1023/a:1018823100761. [DOI] [PubMed] [Google Scholar]
  • 33.Morimoto K, Nakamura T, Morisaka K. Effect of medium-chain fatty acid salts on penetration of a hydrophilic compound and a macromolecular compound across rabbit corneas. Arch Int Pharmacodyn Ther. 1989;302:18–26. [PubMed] [Google Scholar]
  • 34.Sasaki H, Igarashi Y, Nagano T, et al. Different effects of absorption promoters on corneal and conjunctival penetration of ophthalmic beta-blockers. Pharm Res. 1995;12:1146–50. doi: 10.1023/a:1016203725128. [DOI] [PubMed] [Google Scholar]
  • 35.Kato A, Iwata S. Studies on improved corneal permeability to bunazosin. J Pharmacobiodyn. 1988;11:330–4. doi: 10.1248/bpb1978.11.330. [DOI] [PubMed] [Google Scholar]

Articles from Clinical and Experimental Immunology are provided here courtesy of British Society for Immunology

RESOURCES