Store-operated Ca(2+) inflow in Reuber hepatoma cells is inhibited by voltage-operated Ca(2+) channel antagonists and, in contrast to freshly isolated hepatocytes, does not require a pertussis toxin-sensitive trimeric GTP-binding protein.

The treatment of H4-IIE cells (an immortalised liver cell line derived from the Reuber rat hepatoma) with thapsigargin, 2, 5-di-(tert-butyl)-1,4-benzohydroquinone, cyclopiazonic acid, or pretreatment with EGTA, stimulated Ca(2+) inflow (assayed using intracellular fluo-3 and a Ca(2+) add-back protocol). No stimulation of Mn(2+) inflow by thapsigargin was detected. Thapsigargin-stimulated Ca(2+) inflow was inhibited by Gd(3+) (maximal inhibition at 2 microM Gd(3+)), the imidazole derivative SK&F 96365, and by relatively high concentrations of the voltage-operated Ca(2+) channel antagonists, verapamil, nifedipine, nicardipine and the novel dihydropyridine analogues AN406 and AN1043. The calmodulin antagonists W7, W13 and calmidazolium also inhibited thapsigargin-induced Ca(2+) inflow and release of Ca(2+) from intracellular stores. No inhibition of either Ca(2+) inflow or Ca(2+) release was observed with calmodulin antagonist KN62. Substantial inhibition of Ca(2+) inflow by calmidazolium was only observed when the inhibitor was added before thapsigargin. Pretreatment of H4-IIE cells with pertussis toxin, or treatment with brefeldin A, did not inhibit thapsigargin-stimulated Ca(2+) inflow. Compared with freshly isolated rat hepatocytes, H4-IIE cells exhibited a more diffuse actin cytoskeleton, and a more granular arrangement of the endoplasmic reticulum (ER). In contrast to freshly isolated hepatocytes, the arrangement of the ER in H4-IIE cells was not affected by pertussis toxin treatment. Western blot analysis of lysates of freshly isolated rat hepatocytes revealed two forms of G(i2(alpha)) with apparent molecular weights of 41 and 43 kDa. Analysis of H4-IIE cell lysates showed only the 41 kDa form of G(i2(alpha)) and substantially less total G(i2(alpha)) than that present in rat hepatocytes. It is concluded that H4-IIE cells possess store-operated Ca(2+) channels which do not require calmodulin for activation and exhibit properties similar to those in freshly isolated rat hepatocytes, including susceptibility to inhibition by relatively high concentrations of voltage-operated Ca(2+) channel antagonists. In contrast to rat hepatocytes, SOCs in H4-IIE cells do not require G(i2(alpha)) for activation. Possible explanations for differences in the requirement for G(i2(alpha)) in the activation of Ca(2+) inflow are briefly discussed.

Maitotoxin activates an endogenous non-selective cation channel and is an effective initiator of the activation of the heterologously expressed hTRPC-1 (transient receptor potential) non-selective cation channel in H4-IIE liver cells.

The structures and mechanisms of activation of non-selective cation channels (NSCCs) are not well understood although NSCCs play important roles in the regulation of metabolism, ion transport, cell volume and cell shape. It has been proposed that TRP (transient receptor potential) proteins are the molecular correlates of some NSCCs. Using fura-2 and patch-clamp recording, it was shown that the maitotoxin-activated cation channels in the H4-IIE rat liver cell line admit Ca(2+), Mn(2+) and Na(+), have a high selectivity for Na(+) compared with Ca(2+), and are inhibited by Gd(3+) (half-maximal inhibition at 1 microM). Activation of the channels by maitotoxin was inhibited by increasing the extracellular Ca(2+) concentration or by inclusion of 10 mM EGTA in the patch pipette. mRNA encoding TRP proteins 1, 2 and 3 at levels comparable with those in brain was detected using reverse transcriptase-polymerase chain reaction in poly(A)(+) RNA prepared from H4-IIE cells and freshly-isolated rat hepatocytes. In H4-IIE cells transiently transfected with cDNA encoding hTRPC-1, the expressed hTRPC-1 protein was chiefly located at intracellular sites and at the plasma membrane. Cells expressing hTRPC-1 exhibited a substantial enhancement of maitotoxin-initiated Ca(2+) inflow and a modest enhancement of thapsigargin-initiated Ca(2+) inflow (measured using fura-2) and no enhancement of the highly Ca(2+)-selective store-operated Ca(2+) current (measured using patch-clamp recording). In cells expressing hTRPC-1, maitotoxin activated channels which were not found in untransfected cells, have an approximately equal selectivity for Na(+) and Ca(2+), and are inhibited by Gd(3+) (half-maximal inhibition at 3 microM). It is concluded that in liver cells (i) maitotoxin initiates the activation of endogenous NSCCs with a high selectivity for Na(+) compared with Ca(2+); (ii) TRP proteins 1, 2 and 3 are expressed; (iii) maitotoxin is an effective initiator of activation of heterologously expressed hTRPC-1 channels; and (iv) the endogenous TRP-1 protein is unlikely to be the molecular counterpart of the maitotoxin-activated NSCCs nor the highly Ca(2+)-selective store-operated Ca(2+) channels.

In vitro adenovirus mediated gene transfer to the human cornea.

BACKGROUND/AIMS: Replication deficient adenovirus is an efficient vector for gene transfer to the cornea. The aim was to optimise the transduction of human corneal endothelium with adenoviral vectors and to measure transgene production from transduced corneas. METHODS: Adenoviral vectors (AdV) encoding enhanced green fluorescent protein (eGFP) or a transgenic protein (scFv) were used to transfect 34 human corneas. Reporter gene expression was assessed after 72-96 hours of organ culture. The kinetics of scFv production was monitored in vitro for 1 month by flow cytometric analysis of corneal supernatants. RESULTS: Transduction of human corneas with high doses (5 x 10(7)-3 x 10(8) pfu) of AdV caused eGFP expression in 12-100% of corneal endothelial cells. Corneas were efficiently transduced following up to 28 days in cold storage. Very high AdV doses (2 x 10(9) pfu) reduced endothelial cell densities to 98 (SD 129) nuclei/mm(2) (compared to 2114 (716) nuclei/mm(2) for all other groups). Transgenic protein production peaked at 2.4 (0.9) microg/cornea/day at 2 weeks post-transduction, and decreased to 1.2 (0.4) microg/cornea/day by 33 days, at which time endothelial cell density had decreased to 431 (685) nuclei/mm(2). CONCLUSION: Human corneas can be efficiently transduced by AdV following extended periods of cold storage, and transgene expression is maintained for at least 1 month in vitro.

Influence of format on in vitro penetration of antibody fragments through porcine cornea.

AIM: Antibody fragments, appropriately formulated, can penetrate through the ocular surface and thus have potential as therapeutic agents. The aim was to investigate the influence of protein fragment format on the kinetics and extent of ocular penetration in vitro. METHODS: Immunoglobulin single chain variable domain fragments of a murine monoclonal antibody with specificity for rat CD4 were engineered with a 20 or 11 amino acid linker by assembly polymerase chain reaction, expressed in Escherichia coli and purified by chromatography. Fab fragments of the parental antibody were prepared by papain digestion. Antibody fragments were formulated with a penetration and a viscosity enhancer and were applied to the surface of perfused pig corneas for up to 10 hours in vitro. Penetration was quantified by flow cytometry on rat thymocytes. RESULTS: 20-mer antibody fragments formed natural monomers and dimers following purification that could be separately isolated, while 11-mer fragments were dimeric. All formats of fragment (20-mer monomers and dimers, 11-mer dimers, Fab) showed penetration through the pig cornea after 6 hours of intermittent topical administration. CONCLUSION: Antibody fragments of different shapes and sizes can penetrate the cornea after topical administration, thereby increasing the potential of this class of proteins for topical ophthalmic use.

Novel variants of voltage-operated calcium channel alpha 1-subunit transcripts in a rat liver-derived cell line: deletion in the IVS4 voltage sensing region.

Using reverse transcriptase-PCR and Northern analysis, we have shown that the H4IIE cell line, derived from the Reuber H35 rat hepatoma, contains significant amounts of transcripts for the CaCh3 (neuroendocrine) and CaCh1 (skeletal muscle) L-type voltage-operated calcium channel alpha 1-subunits. Two of the CaCh3 transcripts have a 45 bp deletion in the IVS4 membrane-spanning region which is the result of a mutation in genomic DNA. The deduced amino acid sequences of the PCR-derived clones of CaCh3 indicate that the mutation causes the loss of 15 amino acids from the IVS4 region, including three of the six positively charged residues, which are thought to be part of the voltage-sensing mechanism of voltage-operated Ca2+ channels. Quantitative-PCR and Northern analysis indicate that one of the novel CaCh3 transcripts is present in sufficient amounts to imply it could play a functional role in Ca2+ inflow. RT-PCR analysis of hepatocytes isolated from rat liver detected transcripts of CaCh3 (without the IVS4 mutation) and CaCh2, but at considerably lower levels than observed for the isoforms in the H4IIE cell line. Transcripts of CaCh1 and CaCh2 were also detected at low levels in Jurkat T lymphocytes. Fluorimetric studies with the Ca(2+)-sensitive probe, Fluo-3, have shown that H4IIE cells exhibit receptor-activated and store-activated (thapsigarin-induced), but not depolarisation (extracellular KCl)-induced Ca2+ inflow. The mutant transcripts are unlikely to produce Ca2+ channels that are opened by membrane depolarisation. The idea that they may be opened by other mechanisms is briefly discussed.

Use of DNA polymorphisms and the polymerase chain reaction to examine the survival of a human limbal stem cell allograft.

PURPOSE: The extent to which limbal epithelial stem cell allografts will repopulate the human corneal ocular surface, and the time frame over which such cells survive, are uncertain. We investigated the survival of donor-derived epithelial cells after limbal stem cell allotransplantation in a patient with bilateral limbal stem cell failure by using short tandem-repeat DNA polymorphisms to distinguish donor and recipient cells. METHODS: Epithelial cells were harvested by impression cytology from the grafted eye before and at various times after transplantation. DNA was extracted and amplified by the polymerase chain reaction at an informative locus, D8S264. RESULTS: Cells of donor genotype were present over the grafted areas at the time of surgery but were not detected in the central cornea until 12 weeks postoperatively, indicating that repopulation of the epithelial surface from transplanted limbal stem cells took considerable time. However, by the 20th postoperative week, only recipient-type cells were detected in the grafted eye, despite systemic immunosuppression of the recipient with azathioprine and cyclosporine. CONCLUSIONS: Discrimination between donor and recipient cells on the ocular surface after limbal allotransplantation was possible using genotypic variation at DNA polymorphic sites (microsatellites). Long-term survival of donor cells after limbal transplantation did not occur in this patient. Detection of DNA polymorphisms amplified by the polymerase chain reaction is a simple, rapid, and noninvasive method of following the course of transplanted cells at the ocular surface.

Prospects for genetic modulation of corneal graft survival.

Irreversible immunological rejection is the major cause of clinical corneal graft failure. Ex vivo gene therapy directed at the donor cornea has been shown to prolong orthotopic corneal allograft survival significantly in experimental animal models including the mouse, rat, rabbit, and sheep. Transgenes effective in prolonging corneal graft survival include immunomodulatory cytokines and cytokine receptors, an inhibitor of neovascularisation, a blocker of antigen-presenting cell-T cell co-stimulation, nerve growth factor, a dominant negative regulator of apoptosis, and the enzyme indoleamine 2,3-dioxygenase. Although many viral and non-viral vectors have been shown to transduce the corneal endothelium efficiently, allograft survival has so far been prolonged only following transduction of the donor cornea with adenoviral and lentiviral vectors. The degree of graft prolongation, although promising, is still insufficient for immediate translation to the clinic. Increasing the time that the therapeutic gene is expressed in the eye with an integrative, non-immunogenic viral vector is likely to be one way to achieve long-term graft survival. Simultaneous targeting of multiple pathways of graft rejection with more than one transgene is likely to be another. We suggest that the use of an adeno-associated viral or lentiviral vector combined with multiple transgenes may provide the key to future clinical trials.

A steroid-inducible promoter for the cornea.

AIM: Topical glucocorticosteroids are administered to virtually every corneal transplant recipient, but irreversible immunological rejection remains the leading cause of graft failure. Ex vivo gene therapy of the donor cornea has been shown to modulate graft rejection in experimental models. The efficacy of a glucocorticosteroid-inducible promoter was assessed in controlling transgene expression following lentivirus-mediated gene transfer to ovine and human corneas. METHODS: A glucocorticosteroid response element (GRE5) was cloned into a lentiviral vector (LV-GRE-IL10) encoding the model transgene interleukin 10. Transgene expression by LV-GRE-IL10-transduced A549 cells, ovine corneas and human corneas cultured with or without dexamethasone was quantified by an IL10-specific enzyme-linked immunosorbent assay. RESULTS: IL10 levels were 30-40-fold higher in supernatants from LV-GRE-IL10-transduced A549 cells cultured with dexamethasone than in controls. Dexamethasone withdrawal resulted in restoration of baseline IL10 levels. Supernatants from LV-GRE-IL10-transduced ovine and human corneas cultured in dexamethasone contained nine to 10 times more IL10 than supernatants from transduced corneas cultured without dexamethasone. CONCLUSION: The GRE5 promoter in a lentiviral vector drove rapid, sustained and inducible transgene expression in both ovine and human corneas in the presence of dexamethasone. A steroid-inducible promoter may be useful for controlling transgene expression in gene-modified donor corneal allografts.

Lentivirus-mediated gene transfer to the rat, ovine and human cornea.

Gene therapy of the cornea shows promise for modulating corneal transplant rejection but the most appropriate vector for gene transfer has yet to be determined. We investigated a lentiviral vector (LV) for its ability to transduce corneal endothelium. A lentivector expressing enhanced yellow fluorescent protein (eYFP) under the control of the Simian virus type 40 early promoter (LV-SV40-eYFP) transduced 80-90% of rat, ovine and human corneal endothelial cells as detected by fluorescence microscopy. The kinetics of gene expression varied among species, with ovine corneal endothelium showing a relative delay in detectable reporter gene expression compared with the rat or human corneal endothelium. Vectors containing the myeloproliferative sarcoma virus promoter or the phosphoglycerate kinase promoter were not significantly more effective than LV-SV40-eYFP. The stability of eYFP expression in rat and ovine corneas following ex vivo transduction of the donor cornea was assessed following orthotopic corneal transplantation. Following transduction ex vivo, eYFP expression was maintained in corneal endothelial cells for at least 28 days after corneal transplantation in the sheep and >60 days in the rat. Thus, rat, ovine and human corneal endothelial cells were efficiently transduced by the LV, and gene expression appeared stable over weeks in vivo.

Topically applied antibody fragments penetrate into the back of the rabbit eye.

AIMS: Antibody fragments have been shown to penetrate into the anterior chamber when applied to the cornea. The aim of this study was to investigate whether such fragments could penetrate into the vitreous cavity after topical administration to the ocular surface of rabbits. METHODS: An engineered single-chain variable-domain antibody fragment with specificity for an irrelevant rat determinant was applied as a 50 microl eye drop to the eyes of live rabbits at 20-min intervals over 12 h. Eye drops contained 0.8-1.1 mg/ml protein in a buffered salt solution supplemented with penetration and viscosity enhancers. Samples were collected by paracentesis from the vitreous cavity immediately postmortem. Antibody fragments in these samples were quantified by measuring the binding activity to specific antigen, using flow cytometry. RESULTS: Topically applied antibody fragments were detectable in the vitreous of rabbit eyes after 4-12 h but had cleared at 12 h following the final eye drop. Concentrations of the antibody fragment in the vitreous samples were estimated to be 50-150 ng/ml at 12 h. Penetration of the parental whole antibody into the vitreous was not observed. CONCLUSION: Antibody fragments penetrate into the vitreous chamber of the rabbit eye after topical administration to the ocular surface. Such fragments may have therapeutic potential for diseases affecting the posterior segment.

Origins of polymorphism at a polypurine hypervariable locus.

We present characterisation of a hypervariable locus, D8S210, mapped to the telomeric region of the short arm of chromosome 8. The locus is highly polymorphic with alleles varying in size from 1.8 kb to 24 kb. Sequence data from 7 alleles shows that the variable region is entirely polypurine on one strand with a tetranucleotide repeating unit GGAA at the margins and diverged versions of this motif internally. The margins are conserved between alleles; polymorphism occurring in the internal regions of the repeat. Alleles are inherited in a Mendelian manner and one new mutation has been observed in analysis of 51 meioses. Use of single copy flanking sequences to elaborate the polymorphism revealed loss of single copy DNA in 3 unrelated families and in 2 other unrelated individuals. Restriction mapping shows that this loss is similar for different sized alleles in all three families suggesting that it was an early event that may have involved a flanking Alu sequence. We present evidence that the polypurine region can adopt triplex conformations in vitro. Such structures may facilitate loss or gain of unique sequences in the genome, contribute to mutation at conformation transition points and drive the hypervariability (> 99% heterozygosity) of this locus.

A diacylglycerol-activated Ca2+ channel in PC12 cells (an adrenal chromaffin cell line) correlates with expression of the TRP-6 (transient receptor potential) protein.

The structures, and mechanisms of activation, of plasma membrane intracellular-messenger-activated, non-selective cation channels in animal cells are not well understood. The PC12 adrenal chromaffin cell line is a well-characterized example of a nerve cell. In PC12 cells, 1-oleolyl-2-acetyl-sn-glycerol (OAG), a membrane-permeant analogue of diacylglycerol, initiated the inflow of Ca(2+), Mn(2+) and Sr(2+). Acetylcholine and thapsigargin initiated the inflow of Ca(2+) and Mn(2+), but not of Sr(2+). The activation of bivalent cation inflow by OAG: (i) was mimicked by another membrane-permeant diacylglycerol analogue, 1,2-dioctanoyl-sn-glycerol, but not by the membrane-impermeant analogue 1-stearoyl-2-arachidonyl-sn-glycerol; (ii) was not blocked by staurosporin or chelerythrine, inhibitors of protein kinase C; (iii) was enhanced by RHC80267 and R50922, inhibitors of diacylglycerol lipase and diacylglycerol kinase respectively; and (iv) was inhibited by extracellular Ca(2+). When OAG was added after acetylcholine, the effect of OAG on Ca(2+) inflow was over-and-above that induced by acetylcholine. 2-Aminoethyl diphenylborate (2-APB) inhibited Ca(2+) inflow initiated by either acetylcholine or thapsigargin, but not that initiated by OAG. Flufenamic acid inhibited OAG-initiated, but not acetylcholine-initiated, Ca(2+) and Mn(2+) inflow. OAG-initiated Ca(2+) inflow was less sensitive to inhibition by SK&F96365 than acetylcholine-initiated Ca(2+) inflow. In polyadenylated RNA prepared from PC12 cells, mRNA encoding TRP (transient receptor potential) proteins 1-6 was detected by reverse transcriptase (RT)-PCR, and in lysates of PC12 cells the endogenous TRP-6 protein was detected by Western blot analysis. It is concluded that PC12 cells express a diacylglycerol-activated, non-selective cation channel. Expression of this channel function correlates with expression of the TRP-3 and TRP-6 proteins, which have been shown previously to be activated by diacylglycerol when expressed heterologously in animal cells [Hofmann, Obukhov, Schaefer, Harteneck, Gudermann, and Schultz (1999) Nature (London) 397, 259-263].

Evidence that the TRP-1 protein is unlikely to account for store-operated Ca2+ inflow in Xenopus laevis oocytes.

The role of the TRP-1 protein, an animal cell homologue of the Drosophila transient receptor potential Ca2+ channel, in store-operated Ca2+ inflow in Xenopus laevis oocytes was investigated. A strategy involving RT-PCR and 3' and 5' rapid amplification of cDNA ends (RACE) was used to confirm and extend previous knowledge of the nucleotide and predicted amino acid sequences of Xenopus TRP-1 (xTRP-1). The predicted amino acid sequence was used to prepare an anti-TRP-l polyclonal antibody which detected the endogenous oocyte xTRP-1 protein and the human TRPC-1 protein expressed in Xenopus oocytes. Ca2+ inflow (measured using fura-2) initiated by 3-deoxy-3-fluoroinositol 1,4,5-trisphosphate (InsP3F) or lysophosphatidic acid (LPA) was completely inhibited by low concentrations of lanthanides (IC50 = 0.5 microM), indicating that InsP3F and LPA principally activate store-operated Ca2+ channels (SOCs). Antisense cRNA or antisense oligodeoxynucleotides, based on different regions of the xTRP-1 cDNA sequence, when injected into Xenopus oocytes, did not inhibit InsP3F-, LPA- or thapsigargin-stimulated Ca2+ inflow. Oocytes expressing the hTRPC-1 protein, which is 96% similar to xTRP-1, exhibited no detectable enhancement of either basal or InsP3F-stimulated Ca2+ inflow and only a very small enhancement of LPA-stimulated Ca2+ in-flow compared with control oocytes. It is concluded that the endogenous xTRP-1 protein is unlikely to be responsible for Ca2+ inflow through the previously-characterised Ca2+ -specific SOCs which are found in Xenopus oocytes. It is considered that xTRP-1 is likely to be a receptor-activated non-selective cation channel such as the channel activated by maitotoxin.

Familial atopy in Australian pedigrees: adventitious linkage to chromosome 8 is not confirmed nor is there evidence of linkage to the high affinity IgE receptor.

Atopy frequently displays autosomal dominant inheritance and recent studies have favoured genetic linkage between atopy and the human chromosome 11q13. We have studied 12 extended families with aggregation of atopy consistent with autosomal dominant inheritance. The families have been studied for linkage of asthma and atopy to loci on chromosome 8p following the observation that one family suggested preliminary evidence of linkage to an anonymous hypervariable locus cloned from a DNA fingerprint and mapped to 8pter-p22. Subsequent analysis shows this putative linkage to be adventitious as the remaining 11 families do not support linkage between atopy and 8p. We have analysed the same families for evidence of linkage of atopy to loci on 11q13. In these families there is no evidence of association between atopy and the 11q loci stronger than that expected by chance alone; furthermore there is no suggestion that a subpopulation of these families display linkage between atopy and the loci. In addition neither the 8p loci nor the 11q loci exhibit evidence of linkage to atopy by affected sib-pair analysis. This also conflicts with previously published data for 11q.

Plasma membrane Ca2+ release-activated Ca2+ channels with a high selectivity for Ca2+ identified by patch-clamp recording in rat liver cells.

Repetitive waves of increased cytoplasmic Ca2+ concentration play a central role in the process by which hormones regulate liver function. Maintenance of these Ca2+ waves requires Ca2+ inflow through store-operated Ca2+ channels. The properties and mechanism(s) of activation of these channels are not well understood. Store-operated Ca2+ channels (SOCs) in the H4-IIE rat liver cell line were studied by whole-cell patch clamping. Depletion of Ca2+ in intracellular stores by intracellular perfusion with either inositol 1,4,5-trisphosphate (InsP(3)) or thapsigargin in the presence of 10 mmol/L ethylene glycol-bis(beta-aminoethyl ether)-N,N-tetraacetic acid (EGTA), or with 10 mmol/L EGTA alone, activated an inward current that reversed at a membrane potential above +40 mV. In physiologic extracellular medium, this inward current was carried exclusively by Ca2+ and was blocked by a variety of di- and trivalent cations. In the absence of extracellular Ca2+ and Mg2+, the inward current was carried by monovalent cations. This current was 10 to 30 times larger than that observed in the presence of extracellular Ca2+, and permitted the detection of single-channel events that corresponded to a single-channel conductance of about 40 pS. Both the Ca2+ and Na+ inward currents were blocked by the calmodulin antagonist, N-(6-amino hexyl)-5-chloro-1-naphthalenesulphonamide (W7), but not by calmidazolium or calmodulin-dependent protein kinase II fragment 290-309. It is concluded that liver cells possess plasma membrane Ca2+ channels that have a high selectivity for Ca2+, are activated by a decrease in the concentration of Ca2+ in intracellular stores through a mechanism that is unlikely to involve calmodulin, and are involved in re-filling intracellular Ca2+ stores during Ca2+ signaling.

Penetration of engineered antibody fragments into the eye.

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.