Permeability of blood-tear barrier to fluorescein and albumin after application of platelet-activating factor to the eye of the guinea pig

One of the inflammatory responses of the eye to local application of platelet-activating factor (PAF) is oedema of the conjunctiva, caused by extravasation of plasma. Aim of the study was to investigate if fluorescein would leak from the blood into the tears together with plasma protein after application of PAF to the eye. Fluorescein was given intraperitoneally 30 min prior to application of 25 μl of 0.1% solution of PAF. Thirty min after PAF the tear film was collected by washing the surface of the eye with 25 μl of phosphate buffered saline (PBS). Fluorescein in eye washings and in plasma was measured by fluorophotometry and albumin by immunodiffusion. Both fluorescein and albumin appeared in a related fashion in tears, being absent in washings of placebo-treated control eyes. Extravasation of fluorescein can be used as a measure for plasma leakage in the conjunctiva with the advantage over the Evans Blue method that the former is a non-invasive method.


Introduction
The blood ±tear barrier, causing the large concentration differences between tear and blood components, is formed by the acinar and ductal cells of the lacrimal gland and by the conjunctival and corneal epithelium at the surface of the eye.
Breakdown of the blood ±tear barrier occurs in allergic conjunctivitis, where increased vascular permeability has been described as an important feature. 1 Besides histamine and prostaglandins, platelet activating factor (PAF) takes part in the in ammatory process. 2 On a molar basis PAF appeared to be 10 3 to 10 4 times more potent in inducing vascular permeability changes in skin compared with histamine. 3 In guinea pigs topical application of PAF on the eye produces within 30 min a dose-dependent increase of the vascular permeability of the conjunctiva, as determined by measurement of the serum albumin concentration in lavage uid of the surface of the eye. 1 The aim of the present study is to investigate if application of PAF to the eye might also cause leakage of intraperitoneal injected uorescein from blood into the tears together with serum albumin. Both uorescein and albumin appeared signi cantly correlated in the eye washings after application of PAF.

Materials and Methods
Eight female Hartley strain guinea pigs (weight range 350 ±450 g) were intraperitoneally injected with 0.25 ml/ kg body weight of a 20% solution of uorescein sodium and 30 min later 25 ml of 0.1% solution of PAF in phosphate buffered saline (PBS) was applied to the right eyes. The left eyes served as controls and received 25 ml of PBS.
Thirty min after application of PAF or PBS tears were collected by washing the surface of the eye with 25 ml of PBS, using a polypropylene micropipette, avoiding touching the eye. After three forced blinks the lavage uid was collected and stored at -208 C until use. Blood was collected by heart puncture of the animals, anticoagulated with heparin and centrifuged for separation of the plasma.
Albumin concentration in lavage uid was determined using a radial immunodiffusion method in agar plates (1.5%), containing guinea pig albumin antiserum in 1:100 dilution. Various concentrations of guinea pig albumin were used as standard. Samples (5 ml) were tested in appropriate dilutions.
Fluorescein was measured in lavage uid and plasma by uorophotometry, using a uorophotometer (Fluorotron Master, Ocumetrics Inc., Mountain View, CA, USA). For measure-ment in lavage uid 20 ml was diluted in 0.5 ml of PBS in a 5 mm-cuvette and plasma 20 ml was diluted in 4 ml of PBS in a 10 mm-cuvette. Fluorescence was measured in the range 510 ± 600 nm, using an excitation wavelength of 430 ±490 nm.
PAF was purchased from Cayman (Ann Arbor, MI, USA), guinea-pig albumin antiserum and guinea-pig albumin from Nordic (Tilburg, The Netherlands). Animals were housed and cared in accordance with rules and regulations of the European Community Board concerning the use of experimental animals (86/ 609/ CEE). Fig. 1 shows a signi cant correlation of the uorescein concentration, expressed in percentage of the plasma concentration, and the albumin concentration in the lavage uid of eyes, treated with PAF, in eight animals. The concentrations of uorescein and albumin in lavage uid of PBS treated eyes were below the detection limit of respectively 3 ng/ ml and 15 mg/ ml. The measured concentrations of uorescein in plasma varied from 11.2 to 25.0 mg/ ml (mean 16·3 1·7) and in lavage uid from 3 to 257 ng/ ml (mean 147·5 38). Albumin concentrations varied from 0 to 376 mg/ ml (mean 188 48).

Discussion
The variation of the appearance of albumin in the lavage uid of the eyes in response to local application of PAF is in correspondence with our observations, described earlier. 1 The great variation in response may be explained by variation in the penetration of the outer eye. The resorption depends on the ocular contact time of the eye drop, containing PAF, and on enzymatic hydrolysis of PAF to lyso-PAF by an acetylhydrolase, present in ocular tissue. 4 Several ocular tissues themselves, subsequent to in ammation, are able to produce PAF in ng quantities. 5 Albumin is not produced by the lacrimal gland of the guinea pig 6 and the appearance of albumin on the ocular surface is a result of the ow of plasma protein containing tissue uid across the epithelium to the ocular surface under the in uence of PAF. The lavage uid from the surface of the eye at this stage contains a mixture of leaking plasma and lacrimal gland uid from the lacrimal gland.
The amount of uorescein leaking out of the blood vessels is dependent on the rapidly changing concentration of uorescein in the blood. Therefore uorescein concentrations are calculated as percentage of the plasma concentration. The same should apply for serum albumin, however the plasma concentration of about 45 g/ l is not likely to change in short-time experiments as carried out in this study. The greater part, more than 90% of uorescein in blood is bound to protein and this explains the relationship in lavage uid found for albumin and uorescein. The determination of albumin in lavage uid as a measure for conjunctival in ammation is a relatively simple method.
The Evans Blue method for measuring extravasation of plasma in in amed conjunctival tissue is based on the same principle of comigration of dye with plasma proteins. However it is an invasive procedure where the animals have to be killed for isolation of the conjunctival tissue and extraction of extravasated dye. The concentration of Evans Blue in the tears is too low to be detectable by colorimetry compared with uorescein, which can be measured uorimetrically with a detection limit of 1 ng/ ml. Measurement of uorescein in lavage uid of the eyes is an alternative for Evans Blue with the advantage that it is not necessary to kill the animals.