Animal Model for the Absorption of Lipophilic Substances from Tear Fluid by the Epithelium of the Nasolacrimal Ducts

Abstract

purpose. To compare the nasolacrimal tissues of several species to see how closely they resemble the human and to measure nasolacrimal absorption of a substance, to show that an absorption pathway exists for substances placed in the external eye, other than directly through the cornea or conjunctiva.

methods. The nasolacrimal systems of six different vertebrates were investigated by light microscopy to find a species with a nasolacrimal system comparable to that of humans, for use in absorption experiments. In addition to primates, rabbits were revealed by histology to have a lacrimal system closely comparable to that of humans. The rabbit lacrimal system had a stratified epithelium consisting of two layers. Subepithelially, the lamina propria was composed of two strata: loose connective tissue containing elastic fibers and lymphatic cells and a rich venous plexus comparable to a cavernous body. Rabbits were therefore chosen for the absorption experiments. ³H-cortisol was dropped into the eyes of female rabbits. After 21, 43, or 146 minutes, the rabbits were killed, the blood collected, and the nasolacrimal systems prepared and embedded for histologic examination. Serum was obtained from the clotted blood, and radioactivity was counted. Autoradiographs of sections of rabbit nasolacrimal duct were also prepared.

results. Uptake of radioactivity into the serum was high and increased with time. After 21 minutes, maximum incorporation of the applied radioactivity into the blood the level was 7.1%; after 43 minutes, 12.4%; and after 146 minutes, 15.5%. Transport of radioactivity was visualized in autoradiographs of rabbit nasolacrimal systems.

conclusions. ³H-cortisol is incorporated from the nasolacrimal ducts into the blood of rabbits. The comparable morphology of rabbits and humans suggests that absorption of cortisol would also take place in humans. Future investigations of the nasolacrimal passage are needed to understand whether absorption of normal tear fluid components in the nasolacrimal ducts is a physiological function that also plays a role in pathologic conditions such as dry eye. The similarities between rabbit and human nasolacrimal ducts support the use of the rabbit for such studies.

From each drop applied to any eye, only a small portion is taken up into the eye by the conjunctiva, sclera, or cornea. A larger part is drained through the upper and lower lacrimal puncta into the nasopharyngeal space, where an enlarged mucosal surface is available for absorption. The remaining part is swallowed, absorbed in the gastrointestinal tract, and may be metabolized during liver passage, whereas drugs absorbed in the nasopharynx are not subjected to this mechanism. They bypass the liver to enter the systemic circulation directly. The administration of eye drops has therefore been compared with a slow intravenous injection. Compression of the lacrimal puncta can reduce systemic absorption considerably. ¹ This concept disregards an important aspect: What happens to the applied tear drop inside the nasolacrimal ducts? What happens to tears inside the efferent lacrimal passage? 

Tucker and Codere ² showed the median transit time of an applied tear drop containing fluorescein dye to be 8 minutes until its appearance in the nose. As early as 1964, Rohen ³ suggested absorption of tear fluid through the epithelial lining of the human efferent lacrimal passage. Recent studies involving human nasolacrimal ducts have confirmed this speculation. It has been demonstrated that the epithelial cells of the human lacrimal sac and nasolacrimal duct are double-layered in most areas and characterized by microciliation. ^{4 5} Moreover, human nasolacrimal ducts are surrounded by a cavernous body ^{6 7} that may enable closure and opening of the lumen of the lacrimal passage by swelling and shrinking, with consequent regulation of tear outflow. ⁶ The retention time of tear fluid in the efferent tear ducts may be prolonged by this mechanism. Also, formation of dacryoliths in the human efferent tear ducts may be a hint of the role of absorption in the lacrimal sac and nasolacrimal duct. ^{8 9} Therefore, we tested the hypothesis of Rohen ³ —that absorption of tear fluid takes place in the efferent tear ducts—in an animal experiment. Based on the recent finding that the rat is unsuitable for studying absorptive processes in the nasolacrimal ducts, ¹⁰ and because there were practically no further data available on vertebrate histology of the efferent tear ducts, we first performed comparative anatomy studies to determine which animal species would serve best for the absorption experiments. 

Materials and Methods

Animals

All the animals used in the comparative studies had been killed elsewhere. Heads of rabbits and deer were obtained from hunters. Heads of apes (Macaca mulatta) were obtained from the German Primate Center (Göttingen, Germany). Heads of rats were obtained from the animal house of the Preclinical Institutes, University of Kiel, and heads of cats from the Institute of Physiology, University of Kiel. Heads of pigs were obtained from a local abattoir. All tissues were freshly obtained (no more than 24 hours after death). 

Living Chinchilla bastard rabbits were used for the absorption experiments. The test was approved by the Ministry of Environment, Nature and Forests (Schleswig-Holstein, Germany; permit number X330a-72,241.121-1) and was performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Human Samples

For purposes of comparative anatomy, four lacrimal systems (two male, two female, aged 51 to 76 years) were obtained no more than 24 hours after death from cadavers donated to the Institute of Anatomy, Christian Albrecht University of Kiel. Limited information was available on the specimens; however, the specimens were obtained from individuals free of recent trauma, eye or nasal infections, or diseases potentially involving or affecting lacrimal function. 

Ocular Sample Preparation and Light Microscopy for Comparative Anatomy

Heads of apes, rabbits, rats, cats, deer, and pigs were immediately fixed in 4% formalin. Heads of rats were fixed for 1 week in 4% formalin and then decalcified in 20% EDTA at 37°C for several days until x-ray examination revealed complete decalcification. Human nasolacrimal systems were removed from the bony channel before fixation and then also fixed in 4% formalin. After 4 weeks of fixation, the nasolacrimal systems of apes, rabbits, cats, deer, and pigs were removed from the heads and postfixed in 4% formalin. All human, primate, and vertebrate nasolacrimal systems were finally embedded in paraffin and sectioned in a horizontal plane. Sections (7 μm) were stained with toluidine blue O (pH 8.5), alcian blue (pH 1.0), resorcin-fuchsin-thiazin-picric acid, azan, and by the Goldner method. 

Scanning Electron Microscopy

For scanning electron microscopy, four nasolacrimal ducts of rabbits and two of humans that had been fixed previously in 4% formalin were cut longitudinally and halved to examine the lining epithelium. All tissue blocks were then impregnated with 2.5% tannic acid for 2 days. A counterfixation in 2% OsO₄ for 4 hours was followed by dehydration in ethanol and drying in a critical-point dryer. The preparations were coated with gold and examined with a scanning electron microscope (Philips, Kassel, Germany). 

Determination of ³H-Cortisol Uptake

[1,2,6,7-³H]-cortisol (250 μCi; specific activity, 100 Ci/mmol) was obtained from Amersham Pharmacia Biotech UK, Ltd. (Buckinghamshire, UK; code 117 B). The toluene-ethanol solution supplied was dried under a layer of nitrogen, and the residue was dissolved in 80 μL 1% fatty acid–free bovine serum albumin in phosphate-buffered saline (0.14 M NaCl and 20 mM HEPES buffer [pH 7.4]). Twenty microliters (containing approximately 60 μCi ³H-cortisol) was dropped by syringe (Hamilton, Reno, NV) into the eyes (10 μL into each eye) of female rabbits weighing 2300 to 2400 g. Then, the eyelids were carefully closed three times by the investigator. After 21, 43, and 146 minutes the rabbits were killed with the barbituate Narcoren (Merial GmbH, Hallbergmoos, Germany) and decapitated. Blood (10 mL) was collected and the heads saved for autoradiography and histologic examination. Serum was obtained from the clotted blood (30 minutes) by centrifugation (12,000 g; 10 minutes) and 2-mL aliquots were analyzed for radioactivity by liquid scintillation using a quench standard. 

Light Microscopy and Autoradiography

The nasolacrimal systems were removed from the heads of the rabbits. One system each was immediately fixed in 4% formalin, embedded in paraffin, and sectioned in a horizontal plane. Sections (7 μm) were stained with toluidine blue O (pH 8.5), alcian blue (pH 1.0), resorcin-fuchsin-thiazin-picric acid, according to Goldner, and with mild periodic acid Schiff (mPAS). The other nasolacrimal system in each case was divided into two parts: upper and lower. For autoradiography, these parts were placed on x-ray film (Hyperfilm ECL; Amersham Pharmacia Biotech) for 3 weeks. 

Results

Comparative Anatomy

Except for the human and ape nasolacrimal systems composed of upper and lower canaliculi, the lacrimal sac, and the nasolacrimal duct, all lacrimal systems from the vertebrates investigated consisted solely of the upper and lower canaliculus leading directly into the nasolacrimal duct. 

In light microscopy, the human, ape, rabbit, cat, deer, and pig tissues revealed a pseudostratified, columnar epithelium with double layering in most areas, a basal cell layer, and a superficial columnar layer (Figs. 1A 1B 1C 1D 1E 1F 1G 2C 2D 2E 2F 2G 2H) . The rat showed a multilayered epithelium (Figs. 2A 2B) . The upper cell layers consisted of larger squamous elements over several layers of more or less cuboidal cells. Goblet cells were integrated in the epithelium of human, rat, and cat as solitary cells (Figs. 1A 1B 2A 2B 2C 2D) and in the epithelium of human and rat as intraepithelial mucous glands (Figs. 1A 1B 2A 2B) . The secretory product of these cells reacted strongly positive with alcian blue (pH 1.0; Fig. 1H ). By contrast, the epithelium of ape, rabbit, deer, and pig contained no goblet cells (Figs. 1C 1D 1E 1F 1G 2E 2F 2G 2H) . However, the epithelium of ape, rabbit, and pig contained many epithelial cells that showed mildly positive staining with alcian blue (pH 1.0) in the upper cytoplasm. The cell surface reacted strongly positive with alcian blue. The cells with the mild staining were mostly arranged in cell groups (Figs. 1D 1F 1J 1K) . There were also epithelial areas without such cells or cell groups (Fig. 1G) . Subepithelially, the lamina propria of the human lacrimal passage was composed of two strata: loose connective tissue containing elastic fibers and lymphatic cells and a rich venous plexus comparable to a cavernous body (Figs. 1A 1B) . A surrounding cavernous system of blood vessels was also found in ape, rabbit, deer, and pig (Figs. 1C 1E 2E 2G) , but was absent in rat and cat (Figs. 2A 2C) . Small seromucous glands opening their excretory ducts into the lacrimal passage were integrated in the lamina propria of human and pig. In the pigs, seromucous glands were distributed along the whole nasolacrimal duct (Fig. 2G) , whereas in the human specimens the glands were present only in the lacrimal sac in small numbers. None of the other animals possessed seromucous glands. The results are summarized in Table 1 . 

Scanning Electron Microscopy in Human and Rabbit

Scanning electron microscopy of human and rabbit nasolacrimal systems revealed the outer surface of the epithelial cells to be covered with microvilli (Fig. 3) . The length of the microvilli was between 400 and 600 nm (human) and 200 and 400 nm (rabbit). The diameter was between 60 and 70 nm in both humans and rabbits. Goblet cells were clearly distinguishable from epithelial cells on account of their differing cell surfaces or secretions (Fig. 3A) . 

Kinetics of Uptake of Radioactive Cortisol into Blood and Serum in the Rabbit

Radioactivity detected in the serum increased in a time-dependent manner (Fig. 4) . After only a short time, a readily measurable blood level was reached: at 21 minutes the level was 7.1% and at 43 minutes 12.4%. At 146 minutes, 15.5% of the applied radioactivity was incorporated into the blood, assuming a total blood volume of 180 mL. 

Light Microscopy and Autoradiography in the Rabbit

Light microscopy of the nasolacrimal ducts of laboratory rabbits (data not shown) used in the absorption experiments was comparable to that observed and described earlier in wild rabbits in the comparative study. After autoradiography for 23 days, all six samples showed blackening of the x-ray film in a time-dependent manner. Tear duct samples of the rabbit in the 146-minute samples revealed the clearest marks, followed by the 43-minute samples and the 21-minute samples (Fig. 5) . No differences were observed between the upper and lower segments of the nasolacrimal ducts, except in the nasolacrimal duct of the 21-minute sample, where the lower segment showed the least blackening (Fig. 5) . 

Discussion

Absorption of substances, their uptake, and their transport into the blood and lymph are controlled by morphologic and functional barriers. Functional barriers include enzymes of the epithelium that may destroy the substances before they can be absorbed, as well as intrinsic properties of the substances including solubility, lipophilicity, and charge. The morphologic barriers are represented by the anatomic structures of the mucosa. ¹¹  

The purpose of our research was to ascertain whether, as suggested by Rohen, ³ the lining of the efferent lacrimal passage is able to absorb parts of the tear fluid or tear fluid components before reaching the nose. To answer this question in an animal experiment, comparative anatomy must be the first step. A recent absorption experiment in rats with iodinated albumin showed that the rat is unsuitable for absorptive processes in the efferent tear ducts. ¹⁰ In rats, only traces of iodinated albumin were incorporated from the nasolacrimal ducts into the blood. A higher proportion of the radioactivity was taken up as the proteolytic degradation product of bovine serum albumin than as free amino acids, and 96% of the radioactivity incorporated was free iodine, probably as a contaminant of the iodinated preparation. ¹⁰ An explanation for the failure of the absorption experiment in the rat may be seen in our present histologic results. Microscopy of the rat nasolacrimal system showed a multilayered epithelium and the absence of a surrounding cavernous body. Based on these findings, absorptive processes seem improbable in the rat. 

By contrast, the present comparative investigations demonstrate that the nasolacrimal system of the rabbit is most suitable for absorptive studies, because it shows the highest homology of all animals compared with the ape nasolacrimal system and also has a very high homology with the human nasolacrimal system, with the exception of the distribution and integration of goblet cells in the epithelium (Table 1) . Moore et al. ¹² showed that the dog is most suitable for studies of the mucous system, because there is high similarity between canine and human conjunctival goblet cell distribution. Unfortunately, dogs were not available for our comparative studies. Also, it is known, and a recent study by May et al. ¹³ indicates, that the pig is closer to the human than the rabbit. However, our comparative studies show that this is not the case for the nasolacrimal system, in that there are differences in the epithelial structure and the distribution of goblet cells as well as subepithelial seromucous glands between human and pig (Table 1) . We will not exclude that the pig also would be a valuable animal model for absorption studies; nevertheless, the rabbit is a handy model and allows inferences concerning absorptive processes that may take place in human efferent tear ducts. In the absorption experiment, we demonstrated that lipophilic steroids can cross the morphologic and functional barriers of the nasolacrimal ducts in rabbits. 

Although the morphology of the lining epithelium and the surrounding cavernous body of the nasolacrimal ducts is very similar in rabbits and primates, as shown in the present study, absorptive processes in the nasolacrimal ducts of the rabbit are not transferable to humans without further ado. It is well known that the rabbit has a very low eye-blink rate, with an average of one blink every 20 to 30 minutes, ¹⁴ and they can also manage for long periods without drinking water, ^{15 16} suggesting special enabling mechanisms. Maurice ¹⁷ questioned the effect of the low blink rate in rabbits on topical drug penetration through the cornea and showed that for most drugs the epithelial permeability of the cornea is so high that all the material present in the tear film penetrates into the cornea in a few minutes. Thus, he concluded, considerations of contact time become irrelevant. Also in humans, both Maurice ^{18 19 20} and Macdonald and Maurice ²¹ have shown that a labeled substance is absorbed across the conjunctival and corneal surfaces. However, the rate of loss in humans seems to be much lower than in rabbits, leading to an average overestimate of tear turnover of 25%. ²¹ We largely avoided the effect of losing ³H-cortisol through the cornea and conjunctiva by closing the rabbit’s eyes manually several times after instilling the substance. We also showed by autoradiography of the nasolacrimal ducts that ³H-cortisol is absorbed through the lining of the efferent tear ducts in the upper part of the nasolacrimal duct as well as in the lower part. Nevertheless, partial absorption of the measured radioactivity by the nasal mucosa cannot be excluded. However, as demonstrated in the present study by scanning electron microscopy and supported by the previous finding that the epithelium of the human nasolacrimal sac and nasolacrimal duct possess microcilia, ⁴ our experiment shows that absorption takes place in the nasolacrimal passage. In contrast to rabbits, the eye-blink rate in adult humans is many times higher (averaging 17 blinks per minute). Thus, most of the tear fluid and applied substances, such as drugs, enter the nasolacrimal ducts, where they have an average of 8 minutes to be absorbed, as shown by Tucker and Codere. ²  

Recent studies on the nasolacrimal ducts have pointed mainly to secretory functions of the epithelial lining. It has been shown that the mucosa plays a major role in immune defense by producing a mucus layer containing various carbohydrates, ⁴ trefoil factor (TFF) peptides, ²² and antimicrobial peptides ^{23 24} and is part of the mucosa-associated lymphoid tissue (MALT). ^{25 26 27 28}  

Our data suggest that the rabbit can serve as an animal model for the analysis of absorption of substances from tear fluid by the epithelium of the nasolacrimal ducts. Such investigations can include normal tear components, such as freely water-soluble small molecules (e.g., urea or amino acids) or perhaps smaller model tear proteins, such as lysozymes. These investigations would be useful for extrapolation to the human situation, because an understanding of the exact mechanism of absorption and regulation of these processes at the mucosa of the lacrimal passage could provide further insights into the occurrence of dacryoliths, which often lead to residual functional impairment with epiphora. Moreover, some factors such as old age, changes in hormonal status (postmenopausal women), immunodeficiency or the sicca syndrome may be associated with downregulation of the transporter systems necessary for absorption. Further, the hypothesis that the normally constant absorption of tear fluid components into the blood vessels of the surrounding cavernous body ^{6 7} connected to the blood vessels of the ocular surface ⁶ may be a feedback signal for tear fluid production could be investigated. 

It can be concluded that the epithelium of the nasolacrimal passage of the rabbit is able to absorb lipophilic substances, and it is most likely that this is also true of the epithelium of the human lacrimal sac and nasolacrimal duct. We therefore suggest that lipophilic components of tear fluid or drugs from tear drops are also readily absorbed by the nasolacrimal ducts, thus bypassing primary metabolism in the liver and therefore exerting direct systemic effects. The absorption of cortisol in the rabbit nasolacrimal ducts and the similarities that exist between rabbit and human nasolacrimal ducts support the use of the rabbit as a model species for studying absorption of substances from tear fluid in the nasolacrimal ducts. 

 Supported by German Research Foundation Program Grants Pa 738/1-2 and Pa 738/1-3.

 Submitted for publication February 12, 2002; revised April 15, 2002; accepted May 30, 2002.

 Commercial relationships policy: N.

 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.

 

Figure 1.

 

(A) Horizontal section of a human lacrimal sac. The lumen (l) of the lacrimal passage is surrounded by a cavernous body rich in blood vessels with wide lumina ( ). (B) Higher magnification of (A). The lacrimal sac contains a double-layered epithelium. Goblet cells are integrated as solitary cells (g) or show a characteristic arrangement of several cell groups forming mucous glands (i). (C) Horizontal section through the nasolacrimal duct of an ape. The lumen (l) of the lacrimal passage is surrounded by a cavernous body rich in blood vessels with wide lumina ( ). (D) Higher magnification of (C) revealing the epithelium. Some of the epithelial cells have a goblet cell–like appearance with pale supranuclear cytoplasm (arrowheads). (E) Horizontal section through the nasolacrimal duct of a rabbit. The lumen (l) of the lacrimal passage is surrounded by a cavernous body rich in blood vessels with wide lumina ( ). (F) Higher magnification of (E) revealing the epithelium. Some of the epithelial cells have a goblet cell–like appearance with pale supranuclear cytoplasm (arrowheads). These cells are mostly arranged in cell groups. (G) High magnification of the epithelium inside the nasolacrimal duct of a rabbit revealing a pseudostratified columnar epithelium without goblet cells. (H) Section through the lining epithelium of a human nasolacrimal duct. The secretory product of goblet cells and intraepithelial mucous glands (i) reacted strongly positive to alcian blue staining. (J) Section through the lining epithelium of an ape’s nasolacrimal duct. The epithelium contains epithelial cells that showed mildly positive staining with alcian blue in their upper cytoplasm (arrowheads). The cell surface reacted strongly positive to alcian blue. (K) Section through the lining epithelium of a rabbit nasolacrimal duct. The epithelium contained epithelial cells arranged mostly in cell groups (arrowheads) that showed mildly positive staining with alcian blue in the upper cytoplasm. The cell surface reacted strongly positively to alcian blue. Goldner staining (A–G); alcian blue staining, pH 1.0 (H–K); magnification: (A, C) ×59; (E) ×118; (B, D, F–K) ×372.

Figure 2.

 

(A) Horizontal section through the nasolacrimal duct of a rat. A surrounding cavernous body is absent. (B) Higher magnification of the lining epithelium of the rat nasolacrimal duct with large squamous cells at the epithelial surface and cuboid cells in deeper epithelial layers. Goblet cells are integrated as mucous glands in the epithelium (i). (C) Horizontal section of a nasolacrimal duct from a cat. The lumen is surrounded by strong connective tissue in which some blood vessels are visible. However, a surrounding cavernous body is absent. (D) Higher magnification of Figure 1C revealing the lining epithelium of the nasolacrimal duct. The pseudostratified columnar epithelium contains many goblet cells (g). (E) Horizontal section through the nasolacrimal duct of a deer. The lumen of the lacrimal passage (l) is surrounded by a cavernous body rich in blood vessels with wide lumina ( ). (F) Higher magnification of Figure 1E revealing a pseudostratified columnar epithelium without goblet cells. (G) Horizontal section through the nasolacrimal duct of a pig. The lumen (l) of the lacrimal passage is surrounded by a cavernous body rich in blood vessels with wide lumina ( ). The lamina propria contains many seromucous glands (arrows). (H) Higher magnification of Figure 1G . The duct contains a pseudostratified columnar epithelium without goblet cells. All sections: Goldner staining. Magnification: (A, E, G) ×30; (B, D, F, H) ×372; (C) ×59.

Table 1.

 

Comparison of Nasolacrimal Ducts of Humans with Those of Various Vertebrates

Specimen	Epithelium	GC	IMG	SCB	SSMG
Human	Double-layered	Yes	Yes	Yes	Yes, but small
					Several, in lacrimal sac
Ape	Double-layered	No, but mucin-secreting epithelial cells	No, but mucin-secreting epithelial cells organized in cell groups	Yes	No
Rabbit	Double-layered	No, but mucin-secreting epithelial cells	No, but mucin-secreting epithelial cells organized in cell groups	Yes	No
Rat	Multilayered	Yes	Yes	No	No
Cat	Double-layered	Yes, many	No	No	No
Deer	Double-layered	No	No	Yes	No
Pig	Double-layered	No, but mucin-secreting epithelial cells	No	Yes	Yes, throughout entire nasolacrimal duct

 GC, goblet cells; IMG, intraepithelial mucous glands; SCB, surrounding cavernous body; SSMG, subepithelial seromucous glands.

Figure 3.

 

Scanning electron micrographs of the surface of epithelial cells in the nasolacrimal ducts of human (A, C) and rabbit (B, D). Cell borders are clearly distinguishable in all sections. A surface covering of epithelial cells consisting of a trimming with microvilli is present in both human and rabbit. (A, arrows) Secretory products of goblet cells. (C, D) Higher magnification of (A) and (B), respectively. Bar: (A, B) 10 μm; (C, D) 5 μm.

Figure 4.

 

Kinetics of uptake of radioactive ³H-cortisol into blood/serum of rabbits. After 21 minutes 41,470 disintegrations per minute (dpm)/mL; after 43 minutes, 72,320 dpm/mL; and after 146 minutes, 91,020 dpm/mL was measured.

Figure 5.

 

Autoradiography of left rabbit nasolacrimal ducts after uptake of radioactive ³H-cortisol. Nasolacrimal systems reveal a time-dependent darkening of the x-ray film.

The authors thank Harald Deger for expert help with the animal experiments and Michael Beall for editing the English.