Xyloside-Lnduced Disruption of Inferphororeceptor Matrix Proteoglycans Results in Retinal Detachment

Xyloside-lnduced Disruption of Inferphororeceptor Matrix Proteoglycans Results in Retinal Detachment

Howard 5. Lazarus and Gregory S. Hageman

Unique domains of the retinal interphotoreceptor matrix (IPM), termed cone matrix sheaths, are composed largely of chondroitin 6-sulfate proteoglycan in most higher mammalian species. Recent investigations suggest that cone matrix sheaths participate in the maintenance of normal retinal attachment. To investigate the potential functional roles of IPM proteoglycans further, the synthesis of cone matrix sheath chondroitin 6-sulfate proteoglycan was perturbed in vivo. Intravitreal injections of p-nitrophenyl-0-D-xylopyranoside (xyloside), a sugar that inhibits chondroitin sulfate proteoglycan synthesis, were administered to Yucatan micropigs. Their eyes were examined funduscopically and electroretinographically. At selected times, the eyes were enucleated and examined histochemically and immuno- histochemically with various probes directed against cone photoreceptor cells and cone matrix sheaths. The IPM was affected selectively after xyloside administration; no inner retinal pathology or dysfunc- tion was detected morphologically or electroretinographically. The degree of xyloside-induced perturbation was dependent on the duration of xyloside exposure and dose. It was classified into three stages, based on morphologic and histochemical criteria. Although all three stages could be observed in a given retina, a single stage typically predominated, depending on the particular dosage regimen. The early stage was characterized by IPM disruption, as evidenced by disorganization of chondroitin 6-sulfate- and peanut agglutinin (PNA)-binding glycoconjugates. Cone photoreceptor cell outer segment degeneration and markedly decreased chondroitin 6-sulfate immunoreactivity distinguished the middle stage. During the late stage, there was a near complete absence of both immunoreactive chondroitin 6-sulfate and PNA-binding glycoconjugates in the IPM. Shallow retinal detachments that appeared funduscopically as patches of retinal whitening frequently were observed after moderate durations of xyloside exposure; these progressed peripherally with continued xyloside exposure. Histologically, the areas of retinal whitening corresponded to regions in which cone matrix sheaths were split transversely (ie, in a plane perpendicular to the longitudinal axes of the photoreceptor cell outer segments) or were separated completely from cone outer segments. Similar effects were not observed in control eyes. These results suggest that adhesion between the neural retina and retinal pigmented epithelium may be dependent, in part, on continuous synthesis of cone matrix sheath-associated proteoglycans and, potentially, other IPM proteoglycans. In addition, these proteoglycans appear to be necessary for the maintenance of cone photoreceptor cell outer segment integrity. Invest Ophthalmol Vis Sci 33:364-376,1992

The molecules that occupy the retinal interphoto- lia, the neural retina, and the retinal pigment epithe- receptor space, collectively termed the interphoto- lium (RPE). There is a consensus that these molecules receptor matrix (IPM), are situated uniquely between mediate interactions between (1) photoreceptor, the apices of two neuroectodermally derived epithe- RPE, and Muller cells and (2) the choroidal vasculature.1"3 However, the specific functional roles of most of these constituents, especially those of recently de- 1 3 From the Department of Ophthalmology, Bethesda Eye Insti- scribed proteoglycans, are not understood fully. " tute, St. Louis, University School of Medicine, St. Louis, Missouri. Evidence for compartmentalization of some IPM- Supported by National Institutes of Health (Bethesda, Maryland) associated molecules has been provided by lectin his- grant EY06463 (GSH), by Research to Prevent Blindness, Inc. tochemical and antibody immunohistochemical stud- ([New York, New York] an Olga Keith Wiess Scholar Award to GSH and an unrestricted grant to the BEI), and by the Bethesda Eye ies. Photoreceptor cell-specific domains of IPM glyco- Institute Resident Research Fund (HSL). conjugates have been identified based on selective 4 6 Presented in part at the annual meeting of the Association for binding of peanut agglutinin (PNA) " and wheat Research in Vision and Ophthalmology, Sarasota, Florida, and the germ agglutinin.7"9 The PNA-binding molecules are Ninth International Congress of Eye Research, Helsinki, Finland. located specifically in IPM domains termed "cone Submitted for publication: July 5, 1990; accepted September 5, matrix sheaths," that surround cone, but not rod, 1991. photoreceptor outer segments and ellipsoids in Reprint requests: Dr. Gregory S. Hageman, PhD, Bethesda Eye 3 10 Institute, 3655 Vista Avenue, St. Louis, MO 63110. various species, including humans. " Immunohisto-

364

Downloaded from iovs.arvojournals.org on 09/30/2021 No. 2 XYLOSIDE-INDUCED PERTURBATION OF IPM / Lozorus and Hageman 065

chemical and biochemical investigations have shown Materials and Methods that chondroitin 6-sulfate glycosaminoglycan is also a specific structural component of cone matrix Animals sheaths34" and that it is part of a larger, membrane- We used 11 female Yucatan micropigs (age range, associated proteoglycan or proteoglycan aggregate 4-6 months old; weight range, 9.6-14.2 kg) obtained that cannot be extracted by aqueous buffers.3'4'12'13 from Charles River (Wilmington, MA) in these stud- Based on our understanding of the roles of proteo- ies. The micropigs were maintained on Purina (St. glycans in other tissues and organ systems,14"16 several Louis, MO) Mini-Pig Grower or Mini-Pig Breeder putative roles, including retinal adhesion, can be pos- chows. Yucatan micropigs were used in these studies tulated for the proteoglycans associated with cone because (1) the porcine retina contains numerous matrix sheaths. Recent observations from a number cone photoreceptor cells23 with well-defined cone ma- of laboratories suggest a potential role for cone matrix trix sheaths;6 (2) their retinal vasculature24 and vitre- sheaths in retina-RPE adhesion.317"20 For example, ous body25 are similar to those of humans; and (3) a in experiments in which porcine, monkey, or human library of probes specific for various porcine retinal retinas were peeled away gently from the RPE, cone cell types and both cone matrix sheath- and noncone matrix sheaths remain attached to the RPE and be- matrix sheath-specific IPM components is available. come elongated up to four to six times their normal Although these animals do not have a true macula or length before (1) splitting along their transverse fovea, there is an area centralis with decreased vascu- axes,313 (2) detaching the RPE from Bruch's mem- larity that is populated densely by cone photorecep- brane,3' '3 or (3) tearing apical RPE microvilli.'9 Collec- tors.23 The animals were treated in conformity with tively, these studies suggest that the strength of attach- the ARVO Resolution on the Use of Animals in Re- ment of cone matrix sheaths to the RPE and cone search, the National Institutes of Health (Bethesda, photoreceptor cell outer segments is greater than that MD) "Guide for the Care and Use of Laboratory Ani- of the cone matrix sheaths themselves or than that mals" (NIH publication 86-23), and the guidelines set between the RPE and Bruch's membrane. Other stud- forth by the St. Louis University Department of Com- ies show that retinal adhesion is weakened signifi- parative Medicine. cantly without concomitant loss of photoreceptor function after intravitreal or subretinal injections of Reagents protease-free chondroitinase ABC.20 Taken together, these studies suggest that cone matrix sheaths may We obtained p-nitrophenyl-/3-D-xylopyranoside provide a molecular bond of considerable strength from Research Products International (Mt. Prospect, that participates in the adhesion between the RPE and IL) and Koch-Light (Suffolk, England). Antibodies di- neural retina. rected against chondroitin-6-sulfate (ADi-6S) were obtained from ICN (Lisle, IL). A cone photoreceptor To investigate further the potential functional role cell-specific monoclonal antibody, designated CSA-1, of cone matrix sheath-associated chondroitin 6-sul- has been generated and characterized.26 Fluorescein- fate proteoglycan in retinal adhesion, studies directed conjugated goat anti-mouse immunoglobulins G and toward perturbing the synthesis of this molecule by M antibodies were purchased from Sigma (St. Louis, administering intravitreal injections of p-nitro- MO) and Cappel (Durham, NC). Protease-free chon- phenol-/3-D-xylopyranoside (xyloside) to Yucatan droitinase ABC (Proteus vulgaris) was obtained from micropigs were conducted. Xyloside inhibits chon- Seikagaku America (St. Petersburg, FL). The fluores- droitin sulfate proteoglycan (also dermatan sulfate cein-conjugated PNA lectin, Arachis hypogaea agglu- and, to some extent, heparan sulfate proteoglycan, nei- tinin, was obtained from Vector (Burlingame, CA) ther of which is present in the IPM) synthesis by com- and EY (San Mateo, CA). Dimethyl sulfoxide peting with the xylosylated proteoglycan core protein (DMSO), globulin-free bovine serum albumin (BSA), as an acceptor for the first galactosyltranferase reac- MgCl , and CaCl were purchased from Sigma. Acryl- tion.21-22 As such, the proteoglycans synthesized con- 2 2 amide and other reagents used for embedding were tain less than the normal compliment of glycosamino- purchased from Bethesda Research Laboratories (Be- glycan side chains and may be transported inade- thesda, MD). quately into the extracellular matrix.22 To our knowledge, this is the first attempt to perturb IPM Injections proteoglycan synthesis either in vivo or in vitro. Our findings support the hypothesis that cone matrix Micropigs were immobilized in a Panepinto sling sheath-associated proteoglycans and, potentially, and anesthetized with isoflurane administered other IPM proteoglycans, participate in normal reti- through a face mask. In those instances where pro- nal attachment. longed anesthesia was required (eg, fundus photogra-

Downloaded from iovs.arvojournals.org on 09/30/2021 066 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / Februory 1992 Vol. 33

phy and electroretinography), the animals were intu- ately before and after each injection to ensure that the bated. One or two drops of proparacaine (Alcon, Fort injection was atraumatic to the retina and/or vitreous. Worth, TX) were administered topically to the eye The fundus was examined at least once a week thereaf- before injection. To prepare the injections, xyloside ter. Fundus photographs were recorded with Kowa was dissolved in DMSO and subsequently suspended RC-2 (Torrance, CA) or Cannon CF 60 U (Itasca, IL) under sterile conditions in balanced salt solution fundus cameras. (BSS; Alcon) with a DMSO/BSS ratio of 1:9. Optimal xyloside doses were determined by preliminary dose- Electroretinography (ERG) ranging studies in which doses resulting in vitreous Retinal responses to light-adapted white flash and concentrations from 1-150 raM were tested in a series 30-Hz flicker were monitored on selected experimen- of 12 Yucatan micropig eyes. Retinal gliosis was ob- tal and control eyes using Burian-Allen bipolar con- served with vitreous xyloside doses exceeding 20 mM. tact lenses and a Nicolet Compact Four electroretino- As such, doses resulting in a vitreous concentration graphic processor (Racine, WI). We recorded ERGs from 1-20 mM (0.7-14 mg of xyloside delivered in a before xyloside treatment and on a weekly basis there- volume of 100 /xl; vitreous volume of 2.5 ml) were after. The micropigs were restrained and anesthe- used in our experiments (Table 1). The injections tized, and their pupils were dilated with one to two were administered through the temporal pars plana, 3 drops each of phenylephrine HC12.5% and cyclopen- mm posterior to the temporal limbus, using a 30- tolate HC1 1%. Photopic ERGs were recorded in re- gauge needle. Care was taken to avoid penetration sponse to single flashes (100-msec duration and 62- into the vitreous cavity greater than 1 cm. Both single foot-lamberts illumination) of white light presented and multiple injection regimens were used (Table 1). in a Ganzfeld with 10 foot-lamberts of background For the latter, injections were administered every 3 illumination. Cone flicker responses were recorded days (based on a half-life of cone matrix sheath asso- similarly using a 30-Hz white stimulus (100-msec du- ciated-chondroitin 6-sulfate turnover of approxi- ration and 62 foot-lamberts illumination) without mately 3-4 days; Hageman, unpublished observa- background illumination. tion). Injections of carrier without xyloside were administered to control eyes by the same protocol. Fixation, Embedding, and Antibody-Lectin Binding Additional controls included the administration of other proteoglycan perturbants (ie, tilorone and sura- The animals were killed with Sleep-Away euthana- min) by similar protocols; these agents did not pro- sia solution (Fort Dodge Laboratories, Fort Dodge, duce any of the effects observed after xyloside admin- IA). After death, their eyes were enucleated, the ante- istration. rior segments removed, and the eyecups fixedfo r 2-4 hr by immersion in formaldehyde 4.0% (freshly generated from paraformaldehyde) in 100 mM sodium Funduscopic Examination cacodylate buffer, pH 7.4. After fixation, the eyecups The eyes were dilated with a single drop each of were rinsed for a minimum of 6 hr in 100 mM sodium phenylephrine HC1 2.5% and cyclopentolate HC1 1% cacodylate buffer, embedded in acrylamide, and sec- (Alcon). Indirect ophthalmoscopy was done immedi- tioned to a thickness of approximately 5-6 nm on a

Table 1. Summary of dosages and treatment effects

Predominant stage

Early stage Middle stage Late stage

Vitreous xyloside No. of Injection No. of Injection No. of Injection concentration eyes regimen eyes regimen eyes regimen

1 mM 1 Once, sacrifice 1 Every 3 days for 2 Every 3 days for 6 after 3 days 2 weeks weeks 1 Every 3 days for 9 weeks 5 mM 1 Once, sacrifice 1 Every 3 days for 2 Every 3 days for 6 after 2 weeks 2 weeks weeks 20 mM 1 Every 3 days for 1 Every 3 days for 2 6 days weeks

More severe stages of xyloside-induced perturbation are observed with in- single treatment stage may be present in a given retina, one stage predomi- creasing dose and/or duration of xyloside exposure. Although more than a nates with a given dosage regimen.

Downloaded from iovs.arvojournals.org on 09/30/2021 No. 2 XYLOSIDE-INDUCED PERTURBATION OF IPM / Lazarus and Hageman 367

cryostat at -20°C, as described previously.27 All four series of ethyl alcohols, passed through propylene ox- quadrants of each eye were sectioned meridionally ide, and embedded in Epon resin (Balzers, Hudson, from the pars plana to the optic nerve. Serial sections NH).31 We cut l-2-/um sections with a Reichert Ultra- of each region were cut, and the adjacent sections cut microtome (Cambridge, Buffalo, NY) and stained each were incubated with one of a battery of probes. them with Richardson's stain.32 All four quadrants of For lectin histochemical studies, the sections were each eye were examined by light microscopy and pho- exposed to fluorescein-conjugated PNA at a concen- tographed on Kodak Technical Pan or Ektachrome tration of 50 Mg/ml in 10 raMphosphate-buffere d sa- 50 professional films using an exposure index of 50. line containing 1 mM MgCl2 and 1 mM CaCl2 (PBS/ M/C) and 1 mg/ml globulin-free BSA and processed as described previously.6'27 For immunohistochemi- Results cal studies, antibodies directed against chondroitin-6- sulfate (AC6S) and cone photoreceptor cell plasma A continuum of biochemical and structural alter- membranes (CSA-1) were applied to sections at the ations was observed after intravitreal injection of xy- appropriate dilution in PBS/M/C for 60 min at room loside. The degree of xyloside-induced effects varied temperature or for 12 hr at 4°C.26 Sections incubated with the duration of exposure and dose (Table 1). Low with AC6S were exposed first to 0.2 units/ml of pro- doses and short durations of xyloside exposure re- tease-free chondroitinase ABC in 50 mM Tris HC1 in sulted in the disruption of the structural and molecu- 0.03 mM sodium acetate, at pH 8.0, for 20 min at lar integrity of cone matrix sheaths. After longer dura- 37 °C. (Preexposure to chondroitinase ABC was used tions of xyloside exposure and/or increased doses, loss because the antibody recognizes unsaturated uronic of chondroitin 6-sulfate immunoreactivity, degenera- acid residues which remain attached to core proteins tion of cone photoreceptor cell outer segments, and after digestion with chondroitinase ABC.28) The slides shallow retinal detachments were observed. Based on subsequently were rinsed in PBS/M/C and exposed to clinical examination, retinal detachments most often the appropriate fluorescein-conjugatedsecondar y an- originated in the central retina and progressed periph- tibody (GAMIgG for CSA-1 and GAMIgM for AC6S) erally with increasing duration of treatment. Despite in PBS/M/C for 30-60 min at room temperature in changes in the distribution and composition of cone the dark. Controls included (1) incubation of sections matrix sheath constituents, no indication of inner reti- with secondary antibodies alone; (2) deletion of chon- nal or RPE pathology was observed at any stage (Fig. droitinase digestion for AC6S; and (3) incubation of 1). The choroid, optic nerve, lens, ciliary body, iris, sections with PNA in the presence of the competitive and cornea also were unaffected by xyloside exposure. hapten o-methyl-a-D-galactopyranoside. The sec- The ERG data provided further evidence that xylo- tions were examined by epifluorescence microscopy side exposure did not result in damage to the inner on an Olympus Vanox VH-2 light microscope (New retina (Fig. 2). The amplitudes of the photopic and Hyde Park, NY) and photographed on Kodak Ek- cone flickerERG s progressively diminished to approx- tachrome 400 or Tri-X films (Rochester, NY) using imately 70% of normal after longer durations of xylo- an exposure index of 800. side treatment (Fig. 2). However, the ratio of the photopic a- to b-wave amplitudes and the a- and b-wave implicit times were not affected significantly. Routine Histologic Examination For descriptive purposes, xyloside-induced effects were classified into three stages (designated early, For routine histologic examination, the tissue was middle, and late), based primarily on the distinct mor- fixed overnight in paraformaldehyde 4%, progressively dehydrated to absolute alcohol, cleared with xy- phologic hallmarks we found. The observed treat- lene, and infiltrated with paraffin wax. We cut 5-6- ment effect typically varied from one retinal region to nm thick sections on a microtome and stained them another; therefore, more than one stage often was ob- with hematoxylin and eosin.29 Tissue for l-/*m sec- served in a single eye, especially after long durations tions was fixed in formaldehyde 2.0% and glutaralde- of xyloside exposure. However, the predominant hyde 2.5% in 100 mM cacodylate buffer containing stage depended on the particular dosage regimen (Ta- ble 1). Each stage is described in detail subsequently. CaCl2 0.025%, pH 7.4, for 2-4 hr, briefly rinsed in 100 mM sodium cacodylate buffer, pH 7.4, and second- arily fixed in 2.0% osmium tetroxide in 100 mM so- Early Stage dium cacodylate buffer for 1-2 hr. En bloc staining The early stage was heralded by an obvious disor- was done using magnesium uranyl acetate 0.5% in ganization of the molecular constituents associated 30 saline 0.9% for 24 hr at 4°C. After another short with cone matrix sheaths, as evidenced by comparing buffer rinse, the tissues were dehydrated in a graded the distribution of PNA-binding constituents in con-

Downloaded from iovs.arvojournals.org on 09/30/2021 RPE

ILM B

Downloaded from iovs.arvojournals.org on 09/30/2021 No. 2 XYLOSIDE-INDUCED PERTURBATION OF IPM / Lazarus and Hageman 369

Fig. 1. Light micrographs depicting sections of a control porcine retina (A) and a xyloside-exposed porcine retina at a late stage of interphotoreceptor matrix (IPM) perturbation (B). At late stages of xyloside exposure, photoreceptor cell outer segments (OS) are disoriented and disrupted, and there is widening of the interphotoreceptor space, interposed between the external limiting membrane (ELM) and retinal pigmented epithelium (RPE) (B). The effects of xyloside treatment are limited to the interphotoreceptor matrix and photoreceptor outer segments. The retinal pigmented epithelium and inner retina appear unaffected following xyloside exposure (B). IS, inner segments; OPL, outer plexiform layer; ILM, internal limiting membrane.

trol and xyloside-treated retinas (Figs. 3A-B, 5A). segments, disorganization and loss of IPM-associated The disruption of cone matrix sheath constituents of- PNA-binding molecules, and a near complete ab- ten was associated with a detectable widening of the sence of AC6S immunoreactivity in the IPM (Figs. interphotoreceptor space (Fig. 3B). The distribution 6A-C). Cone photoreceptor cell outer segment degen- of cone matrix sheath-associated chondroitin 6-sul- eration, as distinguished by alterations in the binding fate was disorganized and decreased in intensity at pattern of CSA-1, was observed only after a signifi- this stage. In most instances, a near complete loss of cant loss of AC6S immunoreactivity in the IPM (Figs. antichondroitin 6-sulfate binding constituents was 3C-D, 6B-C). Rod photoreceptor cell outer segments observed (Figs. 3C, 5B), but not reflected in degree, by were unaffected at this stage, based on light micro- loss of PNA-binding constituents (Fig. 3B). Despite scopic histologic findingsan d conventional transmis- these alterations of cone-associated IPM components, sion electron microscopy (data not shown). cone photoreceptor cell outer segments appeared normal as monitored by monoclonal antibody CSA-1, which specifically labels porcine cone photoreceptors Late Stage (Figs. 3D, 5C). Furthermore, ERG recordings, obtained just before death from eyes with predomi- The late stage was characterized by a generalized nantly early effects, were not changed significantly loss of IPM-associated PNA-binding glycoconjugates compared with pretreatment control eyes. (Figs. 7A-C). Morphologically, cone photoreceptor cell outer segments were disrupted severely and/or shortened (Fig. 7C). Although rod photoreceptor cell Middle Stage outer segments were disoriented severely by this stage, The middle stage was characterized by degenera- they were not disrupted as were the cone outer seg- tion of cone, but not rod, photoreceptor cell outer ments, based on light microscopic histologic findings

XYLOSIDE-INDUCED PERTURBATION ERGs

Fig. 2. Representative white flash and 30- WHITE FLASH 30 Hz FLICKER Hz flicker ERG recordings from the same pig before xyloside treatment (control) and 3 (mid) and 6 (late) weeks after initiation of 5 CONTROL mM intravitreal injections of xyloside, administered every 3 days. Amplitudes of the 30-Hz flicker recordings are decreased by approximately 20% at 6 weeks; however, the a- to-b wave ratio and implicit times are not affected, suggesting widespread cone photoreceptor cell degeneration without significant inner retinal pathology. Photopic ERGs, re- MID corded in response to single flashes (100-msec duration, 62 ft-Lamberts) of white light presented in a Ganzfeld with 10 ft-Lamberts background illumination. Cone flicker responses were recorded in a similar manner using a 30-Hz white stimulus (100-msec duration, 62 ft-Lamberts) without background illumination. LATE

Downloaded from iovs.arvojournals.org on 09/30/2021 070 INVESTIGATIVE OPHTHALMOLOGY b VI5UAL SCIENCE / February 1992 Vol. 33

Downloaded from iovs.arvojournals.org on 09/30/2021 No. 2 XYLOSIDE-INDUCED PERTURBATION OF IPM / Lazarus and Hagemon 371

Fig. 3. Fluorescence light micrographs depicting sections of control (A) and xyloside-exposed (B-D) porcine retina at the early stage. In normal retinae, domains of the interphotoreceptor matrix, termed cone matrix sheaths, are specifically bound by fluorescein-conjugatedPN A (A, arrows). The earliest detectable effect on retinal structure, following exposure to xyloside, is a marked disruption of the normal distribution of PNA-binding constituents associated with cone matrix sheaths (B, asterisks). By the end of this stage a near complete loss of anti-chondroitin 6-sulfate immunoreactivity is observed, although scant immunoreactivity is detectable in some instances (C, arrows). Cone photoreceptor cell morphology is unaffected at this stage based on labeling with the cone-specific antibody CSA-1 (D, arrows). R, retinal pigmented epithelium; arrowheads indicate the location of the external limiting membrane. Fig. 4. Fundus photograph depicting a characteristic area of retinal whitening (A, arrows) that is typically observed in middle and late stage xyloside-treated eyes. Histologically, these regions correspond to areas of shallow splitting of the interphotoreceptor matrix that are easily visualized in sections incubated with fluorescein-conjugatedPN A (B, arrows). Cone matrix sheaths typically, although not always, split along their transverse axes and the interphotoreceptor space (arrows) widens (also see Fig. 1B). R, retinal pigmented epithelium; arrowheads indicate the location of the external limiting membrane.

(Fig. 1) and transmission electron microscopy (data adhesion. In more recent studies, experiments were not shown). done on postmortem human donor eyes in which the neurosensory retina was separated gently from the 19 Retinal Detachment RPE. The microvilli of the RPE often remain embedded in the distal tips of cone matrix sheaths, sug- Shallow retinal detachments often were observed in gesting that the adhesive force between cone matrix eyes with predominantly middle- and late-stage ef- sheaths and the RPE is sufficient to tear the microvilli fects. They were visible funduscopically as areas of from their apical surface. In similar experiments, us- retinal whitening (Fig. 4A); these occurred initially in ing freshly enucleated monkey and pig eyes, cone ma- the central retina and expanded peripherally with in- trix sheaths were found to be firmly adherent to both creasing duration of treatment. In some instances, reti- cone photoreceptor cell outer segments and the apical nal elevations and retinal folds were observed fundu- surface of the RPE, and these attachments were scopically. Histologic examination of these regions re- stronger than those between the RPE and Bruch's vealed that cone matrix sheaths were split transversely membrane.313 Taken together, these studies provide (Figs. 4B, 8) or had separated completely from cone outer segments (Figs. 9A-B). Most often, the proxi- evidence that cone matrix sheaths participate in reti- mal portions of transversely split cone matrix sheaths nal adhesion. retained their association with cone photoreceptor Recent biochemical studies have increased our un- cells, and the distal portions remained adherent to the derstanding of the molecular basis by which cone ma- apical surfaces of the RPE (Figs. 4B, 8). In other in- trix sheaths could participate in retinal adhesion. stances, entire cone matrix sheaths retained their at- Cone matrix sheaths consist largely of a large chon- tachment to the apical RPE but separated away from droitin 6-sulfate proteoglycan, the core protein of which appears to be associated with the cone photore- intact cone photoreceptor cell outer segments (Figs. 3 9A-B). In both cases, the expanded subretinal space ceptor cell outer segment plasma membrane. Cell was devoid of both PNA- and AC6S-binding constitu- membrane-associated chondroitin sulfate proteogly- ents (Fig. 8). cans are associated with other cell types including melanoma cells,36 neuronal cells,37 glial cells,37 epithelial cells,38 and chondrocytes.39 Although the specific Discussion function of these molecules has not been determined For years, it has been speculated that IPM constitu- in most instances, it has been suggested that the mela- ents might participate in retinal adhesion. The force noma- and chondrocyte-associated proteoglycans par- required to peel the retina from the RPE in vitro was ticipate in cell-cell adhesion.39*40 Based on this infor- found to be significantly reduced after treatment with mation, it is reasonable to speculate that cone matrix testicular hyaluronidase, an enzyme that degrades sheath-associated chondroitin 6-sulfate proteoglycan both chondroitin sulfate and hyaluronic acid, suggest- might participate in retinal adhesion. This supposi- ing that IPM proteoglycans and/or glycosaminogly- tion was supported by the results of a recent study in cans might participate in retinal adhesion.33-34 By con- which subretinal and intravitreal injections of pro- trast, in vitro experiments using aqueous-soluble IPM tease-free chondroitinase ABC weakened retinal ad- components suggested that neither cell-surface nor sol- hesion, caused a loss of PNA-binding glycoconju- uble interphotoreceptor proteoglycan-glycoprotein gates, and induced shallow retinal detachment in rab- 35 molecules play a role in retinal adhesion. These stud- bits20 and monkeys (Hageman, unpublished results). ies, however, did not address adequately the potential In this study, the synthesis and normal distribution role of aqueous-insoluble IPM constituents in retinal of cone matrix sheath-associated chondroitin 6-sul-

Downloaded from iovs.arvojournals.org on 09/30/2021 072 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Februory 1992 Vol. 33

Downloaded from iovs.arvojournals.org on 09/30/2021 No. 2 XYLOSIDE-INDUCED PERTURBATION OF IPM / Lazarus and Hageman 373

Figs. 5-7. Fluorescence light micrographs depicting early (Fig. 5), middle (Fig. 6), and late (Fig. 7) stages of xyloside-exposed retinae incubated with fluorescein-conjugated PNA, AC6S, and CSA-1 (A, B, and C, respectively). At the early stage of disruption there is a mottled distribution of PNA-binding interphotoreceptor matrix constituents (5A) and a marked disorganization of chondroitin 6-sulfate proteoglycan (5B, asterisk) in contrast to control retina (see Fig 3A) Cone outer segments, labeled with CSA-1 (5C, arrow), are unaffected at the early stage. The middle stage of disruption is characterized by shortening and degeneration of cone outer segments (6C, arrow). There is also a marked loss of cone matrix sheath-associated PNA-binding constituents (6A) and near complete loss of chondroitin 6-sulfate immunoreactivity (6B, asterisk). Note that chondroitin 6-sulfate immunoreactivity was not present in any retinas in which there was significant cone photoreceptor cell outer segment degeneration. At the late stage of xyloside-induced disruption, a near complete loss of PNA-binding (7A) and anti-chondroitin 6-sulfate-binding constituents (7B, asterisk) is noted. Cone outer segment degeneration is well advanced at this stage; only short fragments of cone outer segment material, bound by CSA-1, remain (7C, arrow). RPE, retinal pigmented epithelium; ELM, external limiting membrane.

fate proteoglycan was perturbed in vivo after intra- tire cone matrix sheath from cone photoreceptor cell vitreal injections of 1-20 mM xyloside, an inhibitor outer segments. It is reasonable to suggest that, in xy- of chondroitin sulfate (also of dermatan sulfate and, loside-treated eyes, cone matrix sheaths are perturbed to some extent, heparan sulfate) proteoglycan synthe- structurally either by improper incorporation of the sis.2122 Similar doses inhibit chondroitin sulfate syn- altered cone matrix sheath proteoglycan core protein thesis/secretion in vitro.2122 In the early stage of xylo- into the cone outer segment plasma membrane or by side-induced perturbation, there was disorganization a functional deficiency of this molecule. This would and loss of cone matrix sheath-associated chondroitin result in transverse splitting of cone matrix sheaths 6-sulfate proteoglycan and PNA-binding glycoconju- and/or separation of cone matrix sheaths from cone gates. Shallow retinal detachments and specific de- outer segments. When xyloside-treated retinas are ex- generation of cone photoreceptor cell outer segments amined electron microscopically (Hageman and Laza- were observed in eyes with predominantly middle- rus, unpublished data), alterations of cone photore- and late-stage treatment effects. The results of these ceptor cell outer segment plasma membranes typi- studies provide additional evidence that cone matrix cally are observed, suggesting that cone matrix sheath sheath-associated and, potentially, other IPM proteo- proteoglycans are not being intercalated properly into glycans, participate in retinal adhesion and may be the cone outer segment plasma membranes. Because important in the maintenance of cone photoreceptor the apical portions of the cone matrix sheaths always cell outer segment viability. remain firmly attached to the apical surfaces of the The results obtained after intravitreal injection of RPE cells, it appears that cone matrix sheaths may be xyloside into micropig eyes can be explained based on attached to the RPE by molecules, such as integrins,41 the known mechanism of action of xyloside.21'22 The that are unaffected by xyloside. early loss of chondroitin 6-sulfate immunoreactivity Weakening of the structural integrity of cone in the IPM observed in xyloside-treated eyes probably matrix sheaths and their attachment to cone photore- represents either (1) a decreased addition of chondroi- ceptor cells does not per se explain the retinal detach- tin sulfate to secreted core protein or (2) decreased ments that commonly occurred after xyloside admin- secretion and/or incorporation of core protein into istration because additional mechanisms that main- the IPM or plasma membrane. The more generalized tain retinal adhesion also would have to be overcome. loss of PNA-binding glycoconjugates, < observed at These include active transport of subretinal fluid, pas- later stages of xyloside treatment, may represent (1) sive hydrostatic forces, osmotic forces, and physical impaired mobility and secretion of cone matrix interdigitations between outer segments and RPE mi- sheath-associated proteoglycan into the IPM; (2) crovilli.42 Conceivably, vitreal-retinal traction, caused disruption of the normal association between other by mechanical perturbation of the vitreous during xy- PNA-binding glycoconjugates and chondroitin 6-sul- loside injection, could result in retinal detachment; fate proteoglycan; and/or (3) degeneration of cone however, detachment was never observed in control photoreceptor cell outer segments, resulting in a con- eyes. Furthermore, funduscopic examination done sequent loss of cone matrix sheath-associated proteo- immediately after intravitreal injections of xyloside glycan secretion into the interphotoreceptor space. did not reveal any new onset of retinal detachment or Current studies are being directed toward determin- change in existing detachments. Alternatively, the ac- ing the specific biochemical alterations of cone matrix cumulation of free, xyloside-conjugated chondroitin sheath composition induced by xyloside. 6-sulfate glycosaminoglycan in the subretinal space Shallow retinal detachments, commonly occurring could alter significantly the osmolality of the subreti- after xyloside administration and typically originating nal space, resulting in fluid accumulation and retinal in the posterior pole, correlated histologically with detachment. The propensity for retinal detachment to splitting of cone matrix sheaths along their transverse occur centrally (area centralis) in xyloside-exposed axes or, in some instances, with separation of the en- eyes remains speculative. However, it is likely that

Downloaded from iovs.arvojournals.org on 09/30/2021 374 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / February 1992 Vol. 33

Downloaded from iovs.arvojournals.org on 09/30/2021 No. 2 XYLOSIDE-INDUCED PERTURBATION OF IPM / Lazarus and Hageman 375

Figs. 8,9. Fig. 8. Fluorescence light micrograph of a section of xyloside-exposed retina incubated in fluorescein-conjugatedPN A at the early stage. Following loss and disorganization of PNA-binding interphotoreceptor matrix constituents, transverse splitting of cone matrix sheaths commonly occurs, often resulting in detachment of the neurosensory retina from the retinal pigmented epithelium, and the concomitant formation of a retinal detachment (asterisks) (also;see Fig. 4B). Adjacent to the detached region, the distribution of PNA-binding molecules is mottled; however, separation has not occurred. ELM, external limiting membrane. Fig. 9. Fluorescence light micrographs of two fields(A , B) of the same section of xyloside-exposed retina at the middle stage, incubated with fluorescein-conjugated PNA. The neurosensory retina (B) has separated from the retinal pigmented epithelium (A) and there is complete separation of cone matrix sheaths (A, arrows) from cone outer segments (B, arrows); cone matrix sheaths remain adherent to the apical retinal pigmented epithelium (RPE), and have not split transversely.

this cone-dense region of retina is affected preferen- ings is evident because events that alter the structure tially because of the increased amount of cone matrix and/or composition of the IPM could predispose an sheath proteoglycan in this region. eye to retinal detachment, loss of photoreceptor func- The specificity of xyloside-induced proteoglycan tion, and ultimately, blindness. It also follows that the perturbation is a concern when interpreting our re- reestablishment of a normal IPM-RPE-photorecep- sults. Conceivably, the synthesis of IPM and retinal tor interface after retinal reattachment surgery may proteoglycan molecules, other than those associated be essential for optimal visual recovery. Further eluci- with cone matrix sheaths, might be affected by xylo- dation of the mechanisms governing the metabolic side treatment because other classes of proteoglycans, structure and function of the molecular constituents including chondroitin sulfates, dermatan sulfates, of the RPE-IPM-photoreceptor cell interface will no and heparan sulfates, contain the xylose-serine link- doubt lead to a better understanding of the pathogene- age competed for by xyloside. Currently, however, nei- sis of retinal detachment and retinal degenerations ther heparan sulfate- nor dermatan sulfate-containing and to improved therapeutic interventions. proteoglycans have been identified in the IPM.3 An Key words: retina, interphotoreceptor matrix, cone photor- aqueous-soluble chondroitin 4-sulfate proteoglycan is eceptor, xyloside, retinal adhesion present in the IPM-surrounding rod photoreceptor 3 cells. It seems unlikely, however, that this proteogly- Acknowledgments can, which does not appear to be membrane intercalated, would participate in retinal adhesion. Bruch's The authors thank Drs. Don Anderson, Steven Fliesler, 43 44 Tyl Hewitt, Lincoln Johnson, Michael Marmor, and Ste- membrane and Miiller cells, each of which contain phen Russell for their helpful suggestions and discussions; significant quantities of chondroitin sulfate proteogly- Mr. Kaj Anderson, Mr. Tim Cornell, Ms. Denise Kelly, Ms. can, did not appear to be affected by xyloside treat- Margie Kirchoff, and Ms. Angela Zeller for their expert ment, based on the histologic, funduscopic, and ERG technical assistance and dedication; Ms. Nicola Sater, Ms. results obtained. It is conceivable that the turnover of Tracy Underwood, and Ms. Sharon Skyles for clerical assistance; and Ms. Anene Tressler-Hauschultz for editorial as- proteoglycans in these regions is significantly slower sistance. than those associated with cone matrix sheaths or that alterations of these proteoglycans do not affect retinal References function. In addition to their probable role in retinal adhe- 1. Hewitt AT: Extracellular matrix molecules: Their importance in the structure and function of the retina. In The Retina: A sion, our results suggest that cone matrix sheaths may Model for Cell Biology Studies, Part II, Adler R and Farber D, be important for maintaining cone photoreceptor cell editors. London, Academic, 1986, pp. 170-201. outer segment viability. Along similar lines, others 2. Hewitt AT and Adler R: The retinal pigment epithelium and identified changes in the distribution of IPM glycos- interphotoreceptor matrix: Structure and specialized function. aminoglycans at a time before photoreceptor cell In Retina, Vol 1, Ryan, SJ, editor. St. Louis, CV Mosby, 1989, 45 47 pp. 56-64. death in the Royal College of Surgeons rat. " The 3. Hageman GS and Johnson LV: Structure, composition and ERG data obtained in this study suggests widespread function of the retinal interphotoreceptor matrix. In Progress cone or cone outer segment degeneration without con- in Retinal Research, Vol 10, Osborne N and Chader J, editors. comitant damage to the inner retina. Furthermore, Oxford, Pergamon, 1991, pp. 207-249. alterations in the binding pattern of CSA-1 were ob- 4. Hageman GS, Kirchoff MA, and Anderson DA: Biochemical characterization and distribution of retinal interphotoreceptor served only after significant loss of AC6S immunore- matrix glycoconjugates. Glycoconjugate 7:572, 1990. activity, suggesting that cone photoreceptor outer seg- 5. Johnson LV, Hageman GS, and Blanks JC: Restricted extracel- ment degeneration occurred subsequent to loss of lular matrix domains ensheath cone photoreceptors in verte- AC6S-binding constituents. brate retinae. In The Interphotoreceptor Matrix in Health and In conclusion, our results strongly implicate the Disease, Bridges CDB and Adler AJ, editors. New York, Alan R. Liss, 1985, pp. 36-46. IPM as a participant in normal retinal attachment 6. Johnson LV, Hageman GS, and Blanks JC: Interphotoreceptor and/or maintenance of cone photoreceptor outer seg- matrix domains ensheath vertebrate cone photoreceptor cells. ment integrity. The clinical significance of these find- Invest Ophthalmol Vis Sci 27:129, 1986.

Downloaded from iovs.arvojournals.org on 09/30/2021 076 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / Februory 1992 Vol. 33

7. Sameshima M, Uehara F, and Ohba N: Specialization of the 28. Couchman JR, Caterson B, Christner JE, and Baker JR: Map- interphotoreceptor matrices around cone and rod photorecep- ping by monoclonal antibody detection of glycosaminoglycans tor cells in the monkey retina, as revealed by lectin cytochemis- in connective tissue. Nature 307:650, 1984. try. Exp Eye Res 45:845, 1987. 29. Sheehan H and Hrapchak B: Theory and Practice of Histology. 8. Hollyfield JG, Rayborn ME, Landers RA, and Myers KM: In- Columbus, Battelle, 1980, pp. 143-144. soluble interphotoreceptor matrix domains surround rod pho- 30. Jamieson JD and Palade KE: Intracellular transport of secre- toreceptors in the human retina. Exp Eye Res 51:107, 1990. tory proteins in the pancreatic exocrine cell: I. Role of the pe- 9. Fariss RN, Anderson DH, and Fisher SK: Comparison of pho- ripheral elements of the Golgi complex. J Cell Biol 34:577, toreceptor-specific matrix domains in the cat and monkey ret- 1967. inas. Exp Eye Res 51:473, 1990. 31. Luft JH: Improvements in epoxy embedding methods. 10. Kim CK, Kirchoff-Rempe MA, Anderson K, and Hageman Biophys Biochem Cytol 9:409, 1961. GS: Interspecies biochemical heterogeneity of cone matrix 32. Richardson JC, Jarret L, and Finke EH: Embedding in epoxy sheath-associated proteoglycans. ARVO Abstracts. Invest Oph- resins for ultrathin sectioning in electron microscopy. Stain thalmol Vis Sci 32(Suppl):1217, 1991. Technol 35:313, 1960. 11. Hageman GS and Johnson LV: Chondroitin 6-sulfate glycos- 33. Berman ER: Mucopolysaccharides (glycosaminoglycans) of aminoglycan is a major constituent of primate cone photore- the retina: Identification, distribution and possible biological ceptor cell matrix sheaths. Curr Eye Res 6:639, 1987. role. Mod Probl Ophthalmol 8:5, 1969. 12. Hageman GS, Hewitt AT, KirchoffM, and Johnson LV: Selec- 34. Marmor MF and Yoon YH: Effects on retinal adhesion of tem- tive extraction and characterization of cone matrix sheath-spe- perature, cyclic AMP, cytochalasin and enzymes. Invest Oph- cific molecules. ARVO Abstracts. Invest Ophthalmol Vis Sci thalmol Vis Sci 29:910, 1988. 28(Suppl):341, 1989. 35. Adler AJ and Klucznik KM: Interaction of bovine pigment 13. Hageman GS and Johnson LV: The cone matrix sheath: Struc- epithelium cells, photoreceptor outer segments, and interpho- tural and functional analysis. ARVO Abstracts. Invest Ophthal- toreceptor matrix: A model for retinal adhesion. Curr Eye Res mol Vis Sci 29(Suppl):108, 1988. 1:579, 1981. 14. Ruoslahti E: Structure and biology of proteoglycans. Annu Rev 36. Bumol TF, Wang QC, Reisfeld RA, and Kaplan NO: Monoclo- Cell Biol 4:229, 1988. nal antibody and an antibody-toxin conjugate to a cell surface 15. Heinegard D and Paulsson M: Structure and metabolism of proteoglycan of melanoma cells can suppress in vivo tumor proteoglycans. In Cell Biology of the Extracellular Matrix, Piez growth. Proc Natl Acad Sci USA 80:529, 1983. K and Reddi A, editors. New York, Elsevier, 1984, pp. 277- 37. Stallcup WB and Beasley L: Bipotential glial precursor cells of 328. the optic nerve express the NG2 proteoglycan. J Neurosci 16. Fransson LA: Structure and function of cell associated proteo- 7:2737, 1987. glycans. Trends Biochem Sci 12:406, 1987. 38. Sanderson RD and Bernfield M: Molecular polymorphism of a 17. Lazarus H and Hageman G: Xyloside-induced perturbation of cell surface proteoglycan: Distinct structures on simple and "cone matrix sheath" structure and composition in vivo. stratified epithelia. Proc Natl Acad Sci USA 80:529, 1988. ARVO Abstracts. Invest Ophthalmol Vis Sci 30(Suppl):490, 39. Noro A, Kimata K, Olike Y, Shinomura T, and Maeda N: 1989. Isolation and characterization of a third proteoglycan (PG-Lt) 18. Lazarus H, Hageman G, and Johnson L: Xyloside-induced from chick embryo cartilage which contains disulfide-bonded perturbation of cone matrix sheaths. Proc Int Cong Eye Res collagenous polypeptide. J Biol Chem 258:9323, 1983. 6:179, 1990. 40. Harper JR and Reisfeld RA: Inhibition of anchorage-indepen- 19. Hollyfield JG, Varner HH, Rayborn ME, and Osterfeld AM: dent growth of human melanoma cells by a monoclonal anti- Retinal attachment to the pigment epithelium: Linkage body to a chondroitin sulfate proteoglycan. J Natl Cancer Inst through an extracellular sheath surrounding cone photorecep- 71:259, 1983. tors. Retina 9:59, 1989. 41. Anderson DH, Guerin CJ, Matsumoto B, and Pfeffer BA: 20: Yao X-Y, Hageman GS, and Marmor MF: Retinal adhesive- Identification and localization of a £-1 receptor from the inte- ness is weakened by enzymatic modification of the interpho- grin family in mammalian retinal pigment epithelial cells. In- toreceptor matrix in vivo. Invest Ophthalmol Vis Sci 31:2051, vest Ophthalmol Vis Sci 31:81, 1990. 1990. 42. Marmor MF: Mechanisms of normal retinal adhesion. In Ret- 21. Schwartz NB: Regulation of chondroitin sulfate synthesis: Ef- ina, Vol 3, Ryan SJ, editor. St. Louis, CV Mosby, 1989, pp. fect of/3-xylosides on synthesis of chondroitin sulfate proteogly- 71-87. can, chondroitin sulfate chains and core protein. J Biol Chem 43. Hewitt AT, Nakazawa K, and Newsome DA: Analysis of newly 252:6316, 1977. synthesized Bruch's membrane proteoglycans. Invest Ophthal- 22. Schwartz NB: Synthesis and secretion of an altered chondroitin mol Vis Sci 30:478, 1989. sulfate proteoglycan. J Biol Chem 254:2271, 1979. 44. Yamagata M, Kimata K, Oike Y, Tani K, Maeda N, Yoshida 23. Vestre WA: Porcine ophthalmology. Vet Clin North Am K, Shimomura Y, Yoneda M, and Suzuki S: A monoclonal 6:667, 1984. antibody that specifically recognizes a glucuronic acid 2-sul- 24. Bloodworth JMB, Gutgesell HP, and Engerman RL: Retinal fate-containing determinant in intact chondroitin sulfate vasculature of the pig: Light and electron microscope studies. chain. J Biol Chem 262:4146, 1987. Exp Eye Res 4:174, 1965. 45. La Vail MM, Pinto LH, and Yasumura D: The interphotore- 25. Sedlacek GM, Gloor BD, and Meszaros J: Anterior bulbus per- ceptor matrix in rats with inherited retinal dystrophy. Invest foration and vitrectomy in the minipig. Graefes Arch Clin Exp Ophthalmol Vis Sci 21:658, 1981. Ophthalmol 218:141, 1982. 46. Porrello K and La Vail M: Immunocytochemical localization 26. Johnson LV and Hageman GS: Characterization of molecules of chondroitin sulfates in the interphotoreceptor matrix of the bound by the cone photoreceptor-specific monoclonal anti- normal and dystrophic rat retina. Curr Eye Res 5:981, 1986. body CSA-1. Invest Ophthalmol Vis Sci 29:550, 1988. 47. Porrello K, Yasumura D, and La Vail MM: Immunogold local- 27. Johnson LV and Blanks JC: Application of acrylamide as an ization of chondroitin 6-sulfate in the interphotoreceptor ma- embedding medium in studies of lectin and antibody binding trix of normal and RCS rats. Invest Ophthalmol Vis Sci 30:638, in the vertebrate retina. Curr Eye Res 3:969, 1984. 1989.

Downloaded from iovs.arvojournals.org on 09/30/2021