Enrichment of Lymphatic Vessel Endothelial Hyaluronan Receptor 1 (LYVE1)-Positive Macrophages Around Blood Vessels in the Normal Human Sclera

Abstract

Purpose.: To investigate whether the normal adult human sclera contains lymphatic vessels and to study their relation to immune cells and blood vessel anatomy.

Methods.: Scleral tissue probes from 35 adult human donor bulbi were analyzed by immunohistochemistry and confocal microscopy for blood vessels (CD31+), lymphatic vessels (lymphatic vessel endothelial hyaluronan receptor 1 [LYVE1]+, podoplanin+), and macrophages (CD68+) at 12 locations (anterior, equatorial, and posterior at 3, 6, 9, and 12 o'clock positions of the eye) in all three scleral layers (episclera, stroma, and lamina fusca). Approval for scientific examination was obtained.

Results.: CD31+ blood vessels were detectable in the human sclera, where the percentage area covered by CD31+ blood vessels was highest in the anterior episclera, followed by equatorial and posterior episclera, and was lowest in the scleral stroma (regardless of location). LYVE1+ podoplanin+ lymphatic vessels were not detectable in any location investigated, although there was a high number of LYVE1+ CD68+ macrophages. These macrophages were concentrated around blood vessels. In contrast, in the episclera, the number of detected LYVE1+ CD68+ macrophages was comparable in all locations; within the stroma, their number increased toward the posterior part of the eye.

Conclusions.: The adult sclera contains blood vessels but lacks, as revealed by immunohistochemistry and confocal microscopy, true lymphatic vessels. LYVE1+ CD68+ macrophages are located adjacent to the longitudinal axis of blood vessels. The function of these cells needs further investigation, but could be a next step toward a better understanding of pathological disorders such as inflammation, tumor, trauma, or glaucoma.

Introduction

The sclera is the rigid outer cover of the eye, protecting intraocular structures, providing attachment sides for extraocular muscles, and maintaining intraocular pressure. To obtain this function, the sclera mainly consists of collagen type I, and a smaller amount of collagen types III, IV, V, VI, and VIII. ^1,2 Elastin accounts for less than 2% of the scleral extracellular matrix. ^3,4  

Anatomically, three scleral layers can be identified: the superficial episclera, facing the orbit; the middle scleral stroma; and the deep lamina fusca, a thin pigmented layer in direct contact with the choroid. The episclera is well supplied with blood vessels, deriving from the anterior as well as the short and long posterior ciliary arteries. ⁴ Except for some perforating vessels, vortex veins, and Schlemm's canal, the scleral stroma is mainly avascular. Observations of these anatomic features stem mainly from the first half of the 20th century, revealed by India ink injections ⁵ and vascular casting. ^6–8 Later angiographic studies were performed to examine (epi)scleral circulatory dynamics. ^9–12 Confocal microscopy for scleral whole mounts has not been used so far. 

Unlike that for blood vessel distribution, investigation of the scleral lymphatic vessels has been poor, and moreover the scleral distribution of immune cells under physiological conditions has not been investigated thoroughly. The inner part of the human eye is known to be devoid of lymphatic vessels, ¹³ which contributes to the ocular immune privilege. However, outside of the eye, lymphatics have been shown in the conjunctiva, limbus, extraocular muscles, and lacrimal gland. ^14–17 Under physiological conditions, lymphatic vessels are involved in transport of fluid, proteins, and immune cells, ¹⁸ while under pathological circumstances they can grow through the sclera into the intraocular space after trauma ¹⁹ and during tumor invasion of ocular melanomas, ^20–22 as we could demonstrate recently. This corroborates findings in the equally alymphatic cornea, where, induced by trauma or infection, pathological lymphatic ingrowths have also been described. ^19,23 Such disease-related lymphangiogenesis has a functional relevance. In the cornea, it leads to higher rejection rates after corneal transplantations, which result in deteriorated visual function. ^24–26 In uveal melanomas with extraocular extension, intraocular lymphatic vessels are associated with higher metastasis rates and reduced survival. ^15,16 Secondary invasion of lymphatic vessels into the eye after open globe injuries could be important for wound healing, immunologic defense against intruding microorganisms, and autoimmune reactions against intraocular antigens. ¹⁹  

To answer the question whether the normal adult human sclera contains lymphatic-like structures, we analyzed scleral tissue probes from human donor bulbi for the presence of specific markers typically expressed in lymphatic vascular endothelium, such as lymphatic vessel endothelium-specific hyaluronan receptor-1 (LYVE1) and podoplanin (D2-40). We also analyzed the relation of immune cells to the scleral blood vessels. 

Methods

Human Bulbus Donors

Human scleral specimens were obtained from 35 eyes of bulbus donors (mean age 65.1 ± 17.3 years, of both sexes, maximum postmortem time 24 hours) at the Eye Bank of the Department of Ophthalmology, University of Cologne, Germany, in accordance with the Declaration of Helsinki and with approval of the local ethics committee and approval for scientific examination. 

Included eyes showed no pathological alterations, as revealed by slit-lamp examination; furthermore, clinical records did not show any history of eye diseases. 

Formalin-Fixated Paraffin-Embedded Scleral Probes

Immunohistochemistry was performed as described previously. ¹⁹ Briefly, bulbi were fixed in phosphate-buffered saline (PBS) containing 4% formaldehyde, rinsed in PBS, and subsequently dehydrated via a series of graded alcohol. Sclera was isolated, and probes were taken from preselected areas (horizontal or vertical in the equatorial level) and embedded in paraffin. Sagittal cross sections of sclera were taken from anterior to posterior, or alternatively, frontal sections from lateral to medial, both at the level of the bulbar equator. Serial sections were prepared for immunohistochemistry against CD31 (monoclonal mouse anti-human antibody [ab] from Dako [Dako, Hamburg, Germany], diluted 1:20 in PBS containing 2% fetal calf serum [FCS]), LYVE1 (rabbit anti-human ab, diluted 1:50; Zytomed, Berlin, Germany) or podoplanin D2-40 (mouse anti-human ab, diluted 1:20; Acris, Herford, Germany), and CD68 (monoclonal mouse anti-human ab, diluted 1:100; Dako) with an incubation time of 30 minutes at 20°C. For detection of CD31, LYVE1, and podoplanin, a standard antigen retrieval protocol was used (98°C for 30 minutes in citrate buffer) while CD68 was pretreated with FastEnzyme (DCS Diagnostics, Hamburg, Germany), according to manufacturer's instructions. Negative controls were performed by omission of the primary antibodies and resulted in no staining. Conjunctiva was taken as positive control and revealed specific staining. 

Our definition of lymphatic vessels required at least the colocalization of two established lympathic markers (LYVE1 and podoplanin) and a vessel lumen. ²⁶ Our definition of blood vessels required immunoreactivity for the endothelial marker CD31 and a vessel lumen. 

Confocal Microscopy

Scleral probes sized 1 cm² were excised from defined anterior (beginning from the corneoscleral trepanation), equatorial, and posterior (0.5 cm from the optic nerve) locations in all four quadrants (Fig. 1A). Conjunctival specimens were employed as positive control. Probes from all 12 locations were fixated in 99% ethanol (Merck Chemicals, Darmstadt, Germany) for 15 minutes, and then washed three times for 5 minutes in PBS on the shaker. Probes were then kept in PBS containing 2% FCS at 4°C overnight. On the next day, the full-thickness scleral probes were laminated using a No. 10 scalpel or ophthalmic scalpel Micro Feather No. 7657BR (both FEATHER Safety Razor Co., Ltd., Osaka, Japan), comparably to standard surgical techniques, resulting in 30- to 40-μm-thick layers. Anti-CD31 (monoclonal mouse anti-human; Dako), LYVE1 ab (rabbit anti-human; Zytomed), anti-podoplanin clone D2-40 (mouse anti-human; AbD Serotec, Puchheim, Germany), or anti-CD68 (monoclonal mouse anti-human; Dako) was added in a dilution of 1:50 to 1:100 in PBS (containing 2% FCS) and incubated at 4°C overnight. The next day, probes were washed three times for 5 minutes with PBS on the shaker. Primary antibodies were detected with corresponding secondary antibodies: Goat anti-mouse fluorescein isothiocyanate (FITC) (Sigma-Aldrich, Steinheim, Germany), goat anti-rabbit Cy5 (Dianova, Hamburg, Germany), goat anti-rat Alexa Fluor 555 (Invitrogen Life Technologies GmbH, Darmstadt, Germany), or goat anti-rabbit IgG Alexa Fluor 647 (Abcam, Cambridge, UK) was added and incubated overnight at 4°C. Probes were rinsed again three times for 5 minutes on the shaker and then transferred on Superfrost slides (Thermo Scientific, Braunschweig, Germany), embedded in fluorescent mounting medium (Dako), and stored at 4°C. 

Figure 1

 

Immunostaining of laminated human sclera demonstrates that CD31+ blood vessels are located mainly in the anterior episclera. (A) Schematic illustration of the position of scleral flat mounts used for analysis. (B) Human scleral cross sections were immunohistochemically stained, showing CD31+ vessel allocation anteriorly, equatorially, and posteriorly. Blood vessels can be detected mainly in the episclera (arrowheads); in the posterior part also stroma-piercing blood vessels are detectable (arrow). (C) Laminated sclera, as revealed by CD31 immunostaining, emphasizes location-dependent differences in vessel size and shape. Whereas the anterior episclera shows a tight vessel net, the amount of vessels decreases toward the posterior part of the episclera. (D) Analysis of the scleral area covered by blood vessels as revealed by whole-mount preparations: Blood vessel density decreases from anterior to posterior as well as from superficial to deeper parts of the sclera. Data are derived from 15 bulbi for cross sections and 20 bulbi for confocal microscopy.

Slides were examined with confocal microscopy (LSM Meta 510; Carl Zeiss AG, Jena, Germany); z-stacks were taken in some of the probes. 

Morphometric and Statistical Analyses

To analyze frequencies in the distribution of anterior, equatorial, and posterior blood vessels, we measured the percentage area covered by blood vessels (PABV) as described previously. ²⁷ Briefly, laminated whole mounts were analyzed with the aid of the confocal microscope (LSM Meta 510; Carl Zeiss AG), and digital pictures of the center of each probe layer were taken using software (Zen 2009; Carl Zeiss AG) at 20-fold magnification. From each location and layer (episclera, stroma, and lamina fusca level) an 0.8-mm²-sized central picture was taken, and the area covered by CD31+ FITC+ vessels was detected with an algorithm established in the imaging Cell-F software (Olympus Deutschland GmbH, Hamburg, Germany). Prior to analysis, gray value images of the whole-mount pictures were modified by several software filters, and vessels were detected by the appropriate threshold setting including the bright vessels and excluding the dark background. 

Additionally to the aforementioned frequency analysis, LYVE1-immunoreactive cells were quantified. For this purpose, in 25 randomly chosen pictures 0.8 mm² in size, only those LYVE1+ cells also displaying a 4′,6-diamidino-2-phenylindole (DAPI)-positive nucleus were counted. 

Statistical analysis was performed using Student's t-test in SPSS software (IBM Corporation, Armonk, NY). P values of <0.05 were considered statistically significant. 

Results

Blood Vessel Location in the Healthy Human Sclera

In cross sections prepared for light microscopy, a clear CD31-positive signal was detectable in all of the 12 locations investigated (Fig. 1A). This CD31-positive signal was concentrated in the vascular endothelium, and CD31+ blood vessels were mainly located in the episclera but lacking in the deeper layers, except for perforating vessels. In the cross sections, a clear enrichment of vessels was detected in the episclera (Fig. 1B, arrowhead). In the stroma and the lamina fusca, blood vessels were infrequently seen, except for the perforating vessels (Fig. 1B, arrow). 

In the longitudinally laminated scleral probes, blood vessels were detected by confocal microscopy (Fig. 1C). This technique allowed for a better vessel analysis compared to cross sections. Vessels, usually 50 μm or smaller in size, were detected in all 12 locations. 

While the anterior sclera showed a tight net of blood vessels with many branches, the equatorial blood vessels were mainly straight, displaying almost no branches. When the PABV was calculated (Fig. 1D), an increase in area was evident in the anterior episclera compared to equatorial or posterior episclera (P < 0.05) or the scleral stroma (P < 0.001 and P < 0.01). This increase is caused by blood vessel frequency, not by vessel size (Fig. 1C). Toward the equatorial and posterior episclera, the number of vessels decreased significantly (P < 0.05). In the scleral stroma, blood vessels were almost lacking except for perforating vessels. Statistically significant differences between superior, inferior, lateral, and medial areas concerning the PABV were not detectable, although there was a tendency toward an increased vessel amount inferior- and superior-anteriorly. 

LYVE1+ Cells but No Classical Lymphatic Vessels Are Detectable in the Healthy Human Sclera

Scleral cross sections and whole mounts were analyzed using the lymphatic markers LYVE1 and podoplanin. Within the sclera, cross sections displayed LYVE1+ cells, mainly in the episclera (Fig. 2A). These LYVE1+ cells were usually lined up behind each other adjacent to the longitudinal axis of blood vessels. The highest number could be detected in close contact with blood vessels, but single LYVE1+ cells were also detectable between blood vessels. In all four quadrants of the sclera, LYVE1+ cells but no typical lymphatic vessels with lumen were observable. 

Figure 2

 

LYVE1+ cells but no typical lymphatic vessels are detectable in the human sclera. (A) LYVE1 or podoplanin immunohistochemistry in formalin-fixated, paraffin-embedded scleral or conjunctival (positive control) probes: LYVE1+ cells (arrowhead), but no typical lymphatic vessels, are located mainly in the episclera. Podoplanin immunoreactivity was absent in all scleral layers (see enlargements of the episclera and lamina fusca). Podoplanin immunoreactivity is clearly detectable in the perineurium of perforating posterior nerves, while podoplanin+ or LYVE1+ lumina indicating lymphatic vessels were not detectable. (B) LYVE1+ or podoplanin+ lymphatic vessels (arrowhead) are detectable in the conjunctiva, as well as LYVE1+ cells (magnification box). (C) In whole mounts of laminated episclera tissue positive for LYVE1 and podoplanin, high amounts of LYVE1+ cells, mainly in the episclera, were detectable, while podoplanin immunoreactivity was absent. (D) In whole mounts of superficial conjunctiva layer, LYVE1+ vessels displaying a lumen were detectable. (E) Whole mounts of laminated scleral stroma tissue stained for LYVE1 and podoplanin show LYVE1+ cells in the stroma (arrowhead), whereas again, podoplanin immunoreactivity was absent. Arrow indicates pseudopodia. (F) Whole mounts of profound conjunctiva show LYVE1+ cells within the conjunctival tissue.

Podoplanin immunoreactivity was not detectable in any of the scleral layers (Fig. 2A), and a classically lymphatic lumen was not definable in the sclera. This finding was in contrast to observations in the conjunctiva, where podoplanin+ vessels with lumen were common. Likewise, the neuronal cover of piercing nerves in the posterior part of the eye stained clearly positive for podoplanin (Fig. 2A), but not for LYVE1. 

Within conjunctival cross sections, which served as positive control, LYVE1+ lymphatic vessels and singular LYVE1+ cells (see Fig. 2B, magnification box), as well as podoplanin+ lymphatic vessels with lumen, were detectable. 

In the laminated whole mounts, lymphatic vessels with lumen were lacking, but LYVE1+ cells, mainly in the episclera (Fig. 2C), were detectable. This is in contrast to the conjunctiva, where lymphatic vessels were detectable mainly in the superficial part (Fig. 2D). LYVE1+ cells in the episclera displayed different morphology depending on their location: Adjacent to blood vessels, they were up to 50 μm long with approximately 2- to 5-μm soma diameter, while in the periphery they were rather round with somal diameter of approximately 15 to 20 μm. Podoplanin immunoreactivity was absent in the episclera (Fig. 2C) or scleral stroma (Fig. 2E). 

Within the profound layers of the scleral stroma (Fig. 2E) and the profound conjunctival stroma (Fig. 2F), sporadic singular LYVE1+ cells were detected. However, compared to what was seen in the episclera, the amount of these cells was clearly reduced (Fig. 3), and the lined-up assembly of LYVE1+ cells in the scleral stroma could not be seen. In the scleral stroma, some LYVE1+ cells showed more pseudopodia (Fig. 2E, arrow). Toward the posterior part of the eye the LYVE1+ cells increased in the more profound layers, whereas their amount in the episclera stayed stable (Fig. 3). 

Figure 3

 

Analysis of LYVE1+ cells: These were mainly found in the episclera, and their amount increased in the stroma toward the posterior part of the eye. LYVE1+ DAPI+ cells were counted in 25 randomly selected pictures and analyzed for the three scleral layers.

LYVE1+ CD68+ Cells in the Episclera Surround Blood Vessels

In order to define macrophages, a CD68 antibody was used. In the conjunctiva as well as the episclera, CD68+ cells were clearly detectable in cross sections (Fig. 4A). In the conjunctiva these cells closely surrounded blood vessels, but they were also detected within the vascular lumen. In the scleral cross sections, CD68+ cells were detected, mainly in the episclera and here also mainly surrounding the blood vessels. 

Figure 4

 

CD68+ LYVE1+ macrophages surround episcleral blood vessels. (A) CD68+ cells (black arrowheads) are detectable in the sclera, as well as in the conjunctiva of formalin-fixated, paraffin-embedded cross sections. In the sclera, CD68+ cells are detectable mainly in the episclera, where a high number surround blood vessels. However, some cells are detectable not in relation to blood vessels. In the conjunctiva, CD68+ cells can be detected within and around blood vessels (black arrowheads). (B) LYVE1 and CD68 immunoreactivity in whole mounts of laminated episcleral tissue: Cells colocalizing for both markers were identified as macrophages (white arrowheads and insets in upper row). These LYVE1+ macrophages were concentrated around CD31-positive blood vessels (lower row).

Double immunohistochemistry against LYVE1 and CD68 revealed that nearly all LYVE1+ cells of the episclera were also CD68+ (Fig. 4B). In CD31+ LYVE1+ double staining, the blood vessel–surrounding character of these LYVE1+ macrophages was obvious (Fig. 4B). The alignment of these cells next to blood vessels in the episclera was constant through different depths within the tissue (Supplementary Video S1). 

Discussion

Our study revealed three main results: 

1.  
   Confocal microscopy of the human sclera is feasible with a laminating technique and is a promising tool for further investigation of sclera involving pathological disorders, such as uveitis, scleritis, tumor, and trauma.
2.  
   The normal adult human sclera contains blood vessels but lacks classical lymphatic vessels. CD31+ blood vessels were mainly found in the anterior part of the episclera, where they build a tight net, decreasing toward the posterior part of the eye.
3.  
   LYVE1+ CD68+ macrophages are found in the episclera and stroma of the adult human sclera. These macrophages mainly surrounded scleral blood vessels.

Different markers are known to detect lymphatic vessels. However, the exact relevance to lymphatic vessels of each of these markers is the subject of controversial discussions. On the one hand, LYVE1 is widely accepted as a reliable marker to distinguish lymphatic vessels from blood vessels in different human cancers ^28–30 ; on the other hand, LYVE1 expression was absent in tumor-associated lymphatics. ^31,32 Podoplanin stains lymphatic vessels but also podocytes, mesothelioma cells, or the optic nerve sheath. ^33–37 Therefore at least two positive lymphatic markers and a vessel lumen should be used to unequivocally define lymphatic vessels. 

Our observations of the absence of lymphatic vessels and the existence of LYVE1+ cells within the human sclera are in line with work by Xu et al., ³⁸ who showed large amounts of LYVE1+ cells in normal murine eyes; in the sclera, 75% of these LYVE1+ cells also were CD11b positive (a macrophage marker in mice). Our work yields similar results in humans and gives additional information, especially with regard to the relationship of LYVE1+ cells and blood vessels. This work supports the current understanding that the intraocular part of the eye is devoid of lymphatic vessels in physiological conditions. ³⁹  

Our results also support work by Schrödl et al., ⁴⁰ who described LYVE1+ cells in the human choroid but no lymphatic vessels. Similar to observations in the choroid, numerous CD68+ cell networks were detected in the sclera, as we show here. Due to the positive colocalization with CD68, the scleral LYVE1+ cells here were identified as macrophages. 

Their high number and special alignment within the sclera suggests a functional role for these macrophages in physiological functions, perhaps indicating cell migration or interstitial fluid transport toward nearby lymphatic vessels into the conjunctiva or the limbus. In our morphologic study we could not analyze the limbus-close sclera; such specimens were not available, as the respective corneoscleral buttons were designated for corneal grafting. 

Another possible function of LYVE1+ cells may be their participation in lymphangiogenesis under pathological conditions. Lymphangiogenesis has been shown in the inflamed cornea, ^23,41,42 and macrophages are known to transdifferentiate into lymphatic endothelial cells ⁴³ and stimulate these cells. ^44,45  

With the relaunch of scleral lenses, the sclera gained renewed interest, and a detailed knowledge of the interaction of blood vessels, lymphatics, and immune cells in physiological conditions is required. 

In conclusion, blood but no lymphatic vessels can be detected in the healthy human sclera. The blood vessels are surrounded by CD68+ LYVE1+ cells. The function of these cells needs further investigation, but could be a next step toward a better understanding of pathological conditions involving the sclera, such as inflammation, tumor, trauma, or glaucoma. 

Supplementary Materials

mov S1 

doc S2 

Acknowledgments

We thank Gottfried O. H. Naumann for helpful discussions, Brigit Regenfuss for advice on image analysis, and Martina Becker for outstanding support in the field of immunohistochemistry. 

Partially presented at the annual meeting of the Association for Research in Vision and Ophthalmology, Seattle, Washington, May 2013. 

Supported by Deutsche Forschungsgemeinschaft (DFG) Grant HE 6743/2-1 (LMH), DFG Grant CU47/6-1, Bayer Graduate School of Pharmacology, Ruth und Helmut Lingen Stiftung Cologne (CC), the Cologne Gerok program (SLS), and the Austrian Research Promotion Agency FFG-830770 (FS). 

Disclosure: S.L. Schlereth, None; B. Neuser, None; A. Caramoy, None; R.S. Grajewski, None; K.R. Koch, None; F. Schrödl, None; C. Cursiefen, None; L.M. Heindl, None