Alterations in the lens capsule contribute to cataractogenesis in SPARC-null mice

Targeted disruption of the SPARC (secreted protein, acidic and rich in cysteine) gene has been performed in mice from three different genetic backgrounds (Gilmour et al.,1998; Norose et al.,1998). The mice are viable and fertile and appear grossly similar to their wild-type (wt, ^+/+) counterparts. Unexpectedly, a predominant and consistent phenotype of all SPARC-null (^-/-) mice is the presence of cataracts. The mechanism of cataractogenesis in SPARC-null mice is not known. In this report, we provide a partial explanation for this pathology.

SPARC, which is also termed osteonectin or BM-40, belongs to the matricellular class of secreted glycoproteins (thrombospondins 1 and 2,osteopontin, tenascins C and X) that function as modulators of cell-cell and cell-matrix interactions. SPARC mediates these interactions by (1) binding to extracellular matrix (ECM) proteins, (2) regulating ECM and growth factor production/efficacy and/or 3) modulating matrix metalloproteinase expression(Lane and Sage, 1994;Tremble et al., 1993;Yan and Sage, 1999;Sage, 1997). Both counteradhesive and antiproliferative properties in vitro have been attributed to SPARC, which induces rounding of cultured cells and disassembly of focal adhesions (Lane and Sage,1990; Murphy-Ullrich et al.,1995). SPARC participates in morphogenesis, tissue repair and differentiation by virtue of its regulation of cell cycle, cell shape change,migration, adhesion and ECM production(Lane and Sage, 1994;Yan and Sage, 1999). Therefore, SPARC does not appear to serve a structural role in the ECM but is a functional modulator of various activities attributed to protein/proteoglycan networks (Sage and Bornstein, 1991; Bornstein,1995). Despite its expression as a consequence of development or injury-related remodeling (Brekken and Sage, 2001), the production of SPARC in normal adult tissues is rather limited. Interestingly, SPARC is expressed in lenticular epithelium in both developing and adult mammals (Yan et al., 1998; Yan et al.,2000). The phenotype of early cataract formation in SPARC-null mice indicates that SPARC participates significantly in the function and homeostasis of the normal lens.

The lens is a cellular structure without blood vessels, lymphatics and nerves. It is enclosed by an avascular thick capsule and is nourished by diffusion from the aqueous humor through the lens capsule and epithelium. The capsule, also called the lens basement membrane (BM), is a substantial,acellular and structurally complex ECM, consisting of an orderly meshwork of various glycoproteins and proteoglycans (for example, collagen type IV,laminin, perlecan, nidogen, and fibronectin)(Timpl and Dziadek, 1986;Bosman et al., 1989;Cammarata et al., 1986). BM produced by different cells has been correlated with functions such as proliferation, differentiation, adhesion and permeability(Yurchenco and Schittny,1990). Thus, the composition of the ECM can be expected to be of particular importance for the maintenance of normal morphological and functional properties of the lens. Other important functions of the lens capsule that are dependent upon the organization of its constituent proteins and proteoglycans are filtration and permeability(Fisher, 1977;Lee et al., 1997;Winkler et al., 2001). The capsular ECM of the lens is not a static structure; rather, it is continually produced and remodeled anteriorly by the lens epithelial cells and posteriorly by newly differentiated fiber cells(Johnson and Beebe, 1984). SPARC regulates the production of certain ECM proteins, in addition to its interaction with collagens (Sasaki et al.,1998; Sage et al.,1989; Mayer et al.,1991; Maurer et al.,1995) and growth factors(Raines et al., 1992;Kupprion et al., 1998). In this study we have asked whether SPARC plays a significant role in the organization and deposition of ECM proteins in the lens capsule and in its structural integrity.

This report demonstrates alterations in the structure of the lens capsule,increased dye or radioactive tracer penetration through the capsule and swelling of the lens fiber cells in lenses from mice with a targeted disruption of the SPARC gene. The absence of SPARC disturbed the normal relationship between the capsular ECM and the underlying cells. We propose that damaged capsular integrity contributes significantly to cataractogenesis in SPARC-null mice and that cell-matrix interactions that are sensitive to the presence (or diminution) of SPARC are a major component of cataractogenesis. The SPARC-null mouse appears to be an opportune model for understanding the role of SPARC in the modulation of ECM organization and function.

Targeted disruption of the SPARC gene was performed in embryonic stem cells from 129SvEv mice (Norose et al., 1998), and these cells were injected into C57BL/6J blastocysts to generate chimeras. F1 SPARC heterozygous mice were produced by the mating of male chimeras with C57BL/6J females(Norose et al., 1998). Crosses between heterozygotes produced a Mendelian distribution of homozygous(^-/-), heterozygous (^+/-) and wt animals(^+/+). The homozygous mutants expressed neither SPARC mRNA nor protein. A polymerase chain reaction (PCR) method, as described previously(Bassuk et al., 1999), was used to genotype the transgenic mice. The presence of the vaginal plug was counted as day 0 of gestation (E0). All the mice used in this study were maintained under pathogen-free conditions. The treatment and use of mice followed the guidelines for the Care and Use of Laboratory Animals by the National Institutes of Health and the Association for Research in Vision and Ophthalmology Statement for the use of Animals in Ophthalmology and Vision Research.

Lenses of both wt and mutant mice were examined and photographed by slit-lamp photomicroscopy (Nikon FS-2). The pupils of unanesthetized mice were dilated with 0.1% Mydriacyl tropicanide ophthalmic solution and 10%phenylephrine hydrochloride ophthalmic solution (1:1 by volume). Slit views were taken at a 30° angle to the optic axis with a Nikon electronic flash power supply at a maximum setting.

Pregnant mice were given 100 μg BrdU (5-bromo-2′-deoxyuridine;Sigma, St Louis, MO) suspended in phosphate-buffered saline (PBS) by intraperitoneal injection. One hour after the injection, embryonic eyes were collected. Postnatal mice younger than 2 months of age and weighing less than 20 g were injected with 500 μg BrdU. The eyeballs were collected 2 hours after injection. For mice older than 2 months and weighing more than 20 g,BrdU was loaded into an osmotic minipump (Alza, Palo Alto, CA) implanted under the skin behind the interscapular space. BrdU was delivered at a rate of 2μg/g body weight per hour for 1 week to ensure an identical amount infused per unit of body weight (Li et al.,1997). The collected eyeballs were fixed immediately in 10%neutral buffered formalin (0.1 M sodium phosphate, pH 7.4), dehydrated through a series of ethanol concentrations and embedded in paraffin. Serial 5 μm thick paraffin sections were cut medially and through the optic nerve head. Immunostaining with anti-BrdU antibody was performed as described previously(Li et al., 1997).

BrdU incorporation into lens epithelium was assessed by microscopy. Nine sections of each lens per animal were analyzed, and an average value was determined. Three to six animals were studied for each time point. All counts were performed without knowing the identity of the animals. Wild-type and transgenic mice were compared within the same age group. Significant differences were determined by Student's paired t-test for comparison of two sample means.

Lenses from embryos and postnatal mice were prepared by fixation with methyl Carnoy's solution (60% methanol, 30% chloroform and 10% glacial acidic acid) for 4 hours. The eyeballs were dehydrated in a solution of ethanol and were embedded in paraffin for staining with hematoxylin and eosin and for immunofluorescence (anti-mouse collagen IV(α1/α2) IgG,Collaborative Biomedical Research, Bedford, MA). For reaction with the anti-MIP26 antibody, the eyeballs were fixed with 4% paraformaldehyde for 20 minutes, washed with PBS and soaked for 4 hours in 30% sucrose in PBS. Frozen sections were processed and exposed to anti-MIP (major intrinsic protein) IgG,followed by a secondary antibody conjugated with fluorescein isothiocyanate. For electron microscopy (EM), lenses were fixed in 2.5% glutaradehyde in 0.1M sodium cacodylate buffer and were processed and photographed as described(Wight et al., 1997;Norose et al., 2000).

Before sodium dodecyl sulfate-polyacrylamide gel electrophoresis(SDS-PAGE), each lens nucleus was separated from the cortex. The cortex was extracted with 0.1 M NaCl. The supernatant produced represents the total cortical water-soluble proteins. The insoluble pellet was extracted further with 8 M urea; this supernatant represents the urea-soluble proteins in the cortex, and the final insoluble pellet was termed the urea-insoluble pellet. The same procedure was performed on the lens nucleus, with the resulting three fractions described above. All samples were dissolved in sample buffer (60 mM Tris [pH6.8], 2%SDS, 10% glycerol, and 0.001% bromphenol blue) containing 10 mM dithiothreitol (DTT) and were boiled for 5 minutes. 4-20% precast polyacrylamide gradient gels (Novex) were used for SDS-PAGE, and proteins were stained with Coomassie Brilliant blue.

For immunoblotting, lenses were homogenized in 0.1M NaCl, 0.1 M Na₂HPO₄ (pH 7.4). Soluble and insoluble fractions were collected and dissolved in sample buffer with 10 mM DTT prior to SDS-PAGE and subsequent transfer (Yan et al.,2000).

Total RNA was isolated from lens fiber cells of wt and SPARC-null mice with an RNeasy Mini Kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. Lens fiber cells were homogenized in TRI reagent (Molecular Research Center, Inc., Cincinnati, OH) and were washed initially with chloroform and isopropanol to remove fiber proteins. The quality and yield of recovered RNA were evaluated by absorption at 260 and 280 nm. Total RNA was reverse-transcribed into cDNA by the use of an Omniscript RT kit (Qiagen). MIP cDNA was amplified using the primer pair:GCCTGGCCTTGGCTACATTGGT/TGGCCTTGGCTACATTGGT.

Whole lenses with intact capsules were removed from freshly enucleated eyes under a dissecting microscopy. Lenses from wt and SPARC-null mice were incubated immediately in 0.4% trypan blue dye (Sigma), 1 μCi/ml[methyl-³H]-thymidine (1 mCi/ml; New England Nuclear, Boston, MA)or a 1: 2000 dilution of 1 mCi/g ³H₂O (1 mCi/g; New England Nuclear, Boston, MA) for 20 minutes at room temperature. The lenses were subsequently rinsed briefly in PBS. The lens capsules were removed immediately. The decapsulized lenses and isolated capsules were homogenized separately (1 lens capsule/100 μl PBS; 1 decapsulized lens/100 μl PBS in a 1.5 ml tube). The [³H]-thymidine or ³H₂O that penetrated into the lens capsules or lens masses was quantified by liquid scintillation counting in 3 ml Ecolume (ICN, Irvine, CA). For quantification of trypan blue dye, supernatants were collected after centrifugation of the lens capsules or lens masses at 13,000 g, and the absorbance at 497 nm was determined for samples.

We examined lenses of age-matched wt and SPARC-null mice from 3 weeks to 1 year of age by slit-lamp microscopy. The SPARC-null mice showed 100%penetrance of the cataract, whereas no significant opacification was observed in the lenses of wt animals up to 1 year of age. Lenticular opacity could be detected in some of the SPARC-null mice in the anterior and/or posterior subcapsular cortex as early as 4 weeks after birth[Fig. 1, 1 month (mo), arrow indicates anterior opacity]. The degree and extent of the cortical opacity were progressive and varied among individual mice of the same age, as well as in both lenses of the same mouse (data not shown). By 3 months of age, the SPARC-null lenses often showed both anterior and posterior subcapsular cortical opacity, whereas the nucleus was clear[Fig. 1, 3 months (mo),arrows]. By 5-7 months of age, opacity was present in the nuclear region, with complete loss of lens transparency [i.e., full mature cataract,Fig. 1, 7 months (mo),arrow].

To determine the onset of cataract formation in SPARC-null mice, we examined lenses from SPARC^+/+ and SPARC-null animals before and after the appearance of cataracts. In E14 and E18 embryos, wt and SPARC-null lenses were indistinguishable in size and in general morphology(Fig. 2). The initial induction of the lens vesicle on E11.5 was not affected (data not shown). Elongation of the lens fiber cells appeared to proceed normally, as the cavity of the lens vesicle was filled with postmitotic, differentiated fiber cells(Fig. 2). The nuclei of the lens fiber cells were located in the anterior part of the cytoplasm. Microphthalmia was not detected during the development of SPARC-null mice. At 1 month (the beginning of cortical opacity), the SPARC-null lenses exhibited the same size and symmetrical arrangement of lens fibers, with the lens nucleus centrally located (Fig. 3A,B). However, under higher magnification, a few of the cortical fiber cells of the SPARC-null lenses appeared enlarged/swollen and lost their hexagonal shape (Fig. 3D,arrows). Tiny vacuoles appeared in fibers at the bow region; these fiber cells were loosely packed and exhibited an uneven distribution of cytoplasm(Fig. 3F, arrows). By 3 months of age, the size of wt and SPARC-null lenses remained similar(Fig. 4A,B). However, the SPARC-null lens nucleus was displaced posteriorly(Fig. 4B). All the secondary fiber cells in SPARC-null cortex were swollen, altered in size and shape and disorganized (Fig. 4D,F,H). By 4 months, SPARC-null lens fiber cell swelling was advanced, fiber cell plasma membranes were ruptured and vacuolated fibers appeared at the bow region(Bassuk et al., 1999). At 5 months, the SPARC-null posterior lens capsule had ruptured, and the nucleus was dislocated in the posterior cavity, with a severely disorganized cortex(data not shown) (Norose et al.,1998).

Whether or not the proliferation of lens epithelial cells is abnormal and a causative factor in the formation of cataract in SPARC-null lenses is not known. Lens development and growth depend on normal lens epithelial cell proliferation and terminal differentiation into lens fibers(Wride, 1996). Because SPARC is a potent inhibitor of the cell cycle in vitro, and overproliferation of lens epithelial cells is associated with posterior subcapsular cataract, lens cell proliferation was measured in SPARC^+/+ and SPARC-null lenses. Indirect immunocytochemical staining of BrdU-incorporated S-phase cells was conducted in lenses from E14 to 1 year old SPARC^+/+ and SPARC-null mice. Detection of BrdU incorporated into DNA was used as an indication of replicating cells. From E14-E18, the primary fiber cells formed normally, and cell proliferation was restricted to the anterior lens epithelium, with equivalent numbers of BrdU-positive cells in SPARC^+/+ and SPARC-null animals (Fig. 5A). Inappropriate S-phase entry in lens fiber cells was not observed in SPARC-null embryonic mice (data not shown). Posnatally, there was no significant difference in BrdU labeling between wt and SPARC-null lens cells(Fig. 5B,C); moreover, the fiber cells showed no incorporation of BrdU up to 6-7 months of age. In 7.5 month-old SPARC^+/+ lenses, BrdU labeling was restricted mainly to the germinative zone of the epithelium, with a few in the transition zone, and there were reduced numbers of positive cells in comparison with younger SPARC^+/+ lenses. However, the number of BrdU-labeled cells increased significantly in 7.5 month-old SPARC-null lenses(P<0.01, Fig. 5C),largely because of fibroblast-like cell masses appearing beneath the posterior capsule (normally, nucleated cells are not found under the posterior capsule). The posterior capsule was thin and ruptured at this age. Ocular inflammation was indicated by the presence of neutrophils, lymphocytes and macrophages in the vitreous cavity (data not shown)(Norose et al., 1998). In 7 month to 1 year-old SPARC^+/+ mice(Fig. 5C), the extent of lens cell proliferation decreased. In contrast, SPARC-null lenses showed increased numbers of BrdU-labeled cells, concomitant with ocular inflammation. In summary, these results indicate that SPARC does not exert a significant effect on the cell cycle in the lens. Therefore, abnormal proliferation of lens epithelial cells does not contribute causally to cataractogenesis in animals lacking SPARC.

Disorganization and swelling of lens fiber cells were the primary morphological alterations observed in the cortex of SPARC-null lenses at 1-3 months after birth. Since severe changes in fiber cells were apparent by 3 months of age (Fig. 4), we isolated these cells and analyzed the fiber proteins in wt and SPARC-null lenses. Fiber cells derived from both genotypes showed almost identical protein patterns by SDS-PAGE (Fig. 6). The water-soluble crystallin proteins (α, β andγ crystallins comprise approximately 90% of lens soluble proteins)showed no differences in levels or in degradation by immunoblot analysis of SPARC-null versus SPARC^+/+ lenses (data not shown). The MIP-26(aquaporin 0) is a membrane-bound protein, abundantly and specifically expressed in lens fiber cells, that regulates the transport of water(Benedetti et al., 1974;Shiels and Bassnett, 1996). At either the mRNA or protein level, there were minimal differences between SPARC^+/+ and SPARC-null fiber cells with respect to expression of MIP26 (Fig. 7A,B). However,immunostaining for MIP26, which outlines the fiber cell membrane, revealed a striking disorganization of cortical fiber cells in SPARC-null relative to SPARC^+/+ lenses, which could result from swelling(Fig. 7C,D).

Since the secondary lens fiber cells were swollen by 2-3 months of age, and there was apparently no degradation or alteration of their major proteins, we began to suspect that the filtration barrier of the capsule and underlying cells for maintaining normal homeostasis might be compromised in the SPARC-null lens. Such a functional alteration could lead to an imbalance of water and ions in the lens and might result in swollen fibers as observed in Figs 3 and4.

By light microscopy, the lens capsule showed minimal abnormalities within 3 months (Figs 3 and4). Moreover, EM showed a smooth interface between the lens capsule and the underlying cells in SPARC^+/+ lenses at 1 and 3 months of age(Fig. 8A,C, arrows). However,by 1 month of age, SPARC-null lens fibers immediately posterior to the equator exhibited a few finger-like protrusions into the lens capsule(Fig. 8B, arrows). These protrusions appeared progressive, in that numerous large extensions penetrated deeply into the lens capsule at 3 months of age(Fig. 8D, arrows). These observations are in agreement with a previous EM study(Norose et al., 2000). The extensions of the cell membrane into the anterior capsule were not as obvious as those into the posterior capsule in the lenses of 1 month-old SPARC-null mice (data not shown).

Collagen type IV is a major structural component of the lens capsule(Fitch et al., 1983;Cammarata et al., 1986). Immunohistochemical staining of collagen IV revealed a similar distribution of this ubiquitous BM protein in SPARC^+/+ and SPARC-null anterior lens capsules of 1 month-old mice (Fig. 9A,C), However, SPARC-null lens capsules posterior to the bow region were abnormal (Fig. 9D,arrows). At 3 months, SPARC-null anterior lens capsules exhibited a more mottled appearance, and small, fine protrusions could be observed by light microscopy (Fig. 9G, arrows). The posterior lens capsule was characterized by numerous clefts, which penetrated deeply into the lens capsule and might be predicted to disrupt the permeability of the capsule (Fig. 9H, arrows).

Trypan blue dye (906 Da) and [³H]-thymidine (242 Da) were chosen as tracers to evaluate the penetration of small molecules through the lens capsules of wt and SPARC-null lenses. In lenses from mice of 1-3 month of age,the possibility of incorporation of [³H]-thymidine into lens epithelial cell DNA is highly unlikely during the 20 minute assay [there was no labeling of lens epithelial cells of 1-month-old mice after 1 half hour BrdU delivery; (Q.Y., unpublished)]. InFig. 10 the difference in dye and [³H]-thymidine penetration between wt and SPARC-null lenses is obvious. The dye content measured in SPARC-null decapsulized lenses(1-month-old) was 1.38 times the value of the SPARC^+/+ lenses. The[³H]-thymidine CPM in SPARC-null decapsulized lenses were 1.3 times(1-month-old) and 3.4 times (3 month-old) the values of the corresponding +/+lenses. ³H₂O was also evaluated for its penetration into the lenses. After 20 minutes, the CPM in capsules of wt and SPARC-null mice (1 and 3 months of age) were minimal, with no difference (data not shown). However, there were significant differences in decapsulized lenses: the CPM in the SPARC-null lenses were 1.6 times (1-month-old) and 2.5 times (3-month-old)the values of the corresponding ^+/+ lenses (data not shown). The conclusions are that (1) SPARC-null lens capsules exhibited increased levels of dye or [³H]-thymidine, but not ³H₂O, (2)SPARC-null lens capsules and underlying cells allowed more dye,[³H]-thymidine or ³H₂O to penetrate into the lens mass and (3) the amount of ³H₂O or[³H]-thymidine that penetrated into the lenses was substantially greater in 3-month-old mice than in 1-month-old mice. The normal lens capsule is permeable to water and small macromolecules(Newell, 1996). Trypan blue dye and [³H]-thymidine can diffuse through the capsule and enter the lens mass (Fig. 10). However, within 20 minutes, SPARC-null lenses contained more dye and radioactive tracer relative to wt controls, a result indicating that disorganization of the capsule and its underlying cells is associated with increased penetration into SPARC-null lenses. The dye was distributed mostly in the equatorial region (Fig. 10A, arrowheads). This observation is consistent with our data from EM and immunostaining of collagen type IV, that is, the lens capsule posterior to the bow region is the initial and principal site of the structural abnormality (Figs 8and 9).

Wt and SPARC-null lenses had similar wet weights (4.09 mg versus 4.15 mg,respectively) (Fig. 11). After dehydration, the dry weight of 1-month-old SPARC-null lenses was less than that of the SPARC^+/+ lenses (0.77 mg versus 1.13 mg, respectively),owing to the increased water content of the SPARC-null lenses. This observation is consistent with the ³H₂O penetration assay. Breakdown of the physiological barrier of SPARC-null lenses would allow increased water flux and/or altered ion transport, which would perturb the osmotic balance in the lens fiber cell membranes and lead to the swelling of fibers. The results are the first to indicate the possible influence of SPARC on permeability in the lens.

In this study, we show that cataract formation in SPARC-null mice is associated with compromised structure of the lens capsule as early as 1 month of age. Our results demonstrate that SPARC is a key protein affecting the maintenance and structural integrity of the lens capsule and the regulation of normal cellular homeostasis in lens cells.

The lens capsule is a specialized ECM that (a) is responsible for lens compartmentalization and the maintenance of lens tissue structure, (b)controls fluid and substrate exchange and (c) regulates normal lens epithelial cell growth and differentiation. In the SPARC-null lens, we propose that its compromised lens capsule contributes to altered homeostasis of the lens and the transport of water, small molecules and ions into the strictly aligned lens fiber cells. The result of these changes is swelling of the lens fibers,one of the hallmarks of an osmotic cataract (e.g., the diabetic cataract)(Bond et al., 1996).

SPARC-null mice begin to exhibit cortical alterations at 1 month of age,when secondary lens fibers located in the cortex and bow region appear swollen(Fig. 3). Ultrastructural analysis disclosed defects in the lens capsule and lens cells in 1 month-old SPARC-null mice (Fig. 8A,B). At this early stage, the capsule structure begins to appear disorganized. The structurally compromised lens capsule, especially at the equator, was associated with increased dye and radioactive tracer penetration into the lens(Fig. 10A, arrowheads;Fig. 10), a result indicating that SPARC-null lenses lack normal permeability. The swelling of the fibers in the bow region at early stages (Fig. 3F), and the numerous vacuoles in this region, could be caused by the disruption of normal cell-matrix interactions and initial damage of the posterior lens capsule behind the equator in the SPARC-null lens. The progression of the fiber cell protrusion into the lens capsule is consistent with the amounts of tracers in the lens capsules and lenses(Fig. 10, 1 month versus 3 months). The presence of defective lens capsular structure is consistent with the swelling of the fiber cells.

The lens capsule is an unusally thick basement membrane. Its intrinsic structural components include collagen type IV, laminin, nidogen and heparan sulfate proteoglycan (Cammarata et al.,1986; Starkey et al.,1987; Inoue,1994). The importance of heparan sulfate proteoglycan in the BM for regulation of the permeability of ions and macromolecules has been amply demonstrated in the glomerular BM (Groffen et al., 1999). Heparan sulfate glycosaminoglycans provide a chargesieving function, as the polyanionic chains are associated with large spheres of hydration, which would leave little free water between macromolecules. Since the distribution of glycosaminoglycans appears to be different in normal and cataractous lens capsules(Winkler et al., 2001), it is likely that alteration of anionic sites would affect the permeability of metabolites as well. Disruption of the organization of the macromolecular network of the lens BM is predicted to facilitate the passage of water, ions and possibly larger molecules across the BM. In fact, increased permeability of the lens capsule to water was found in lenses with cortical cuneiform opacities, to the exclusion of the nucleus(Fisher, 1977;Maraini and Mangili, 1973). Moreover, such increases were also found in the rabbit lens, in which the posterior lens capsule had been perforated, and in adult human diabetic lenses, in which edematous lens fibers were correlated with a 37% increase in hydraulic permeability (Fisher,1985; Fisher and Wakely,1976). SPARC-null lenses with cortical opacity exhibited an increased permeability to water, relative to wt lenses, that could disrupt the Na⁺ and K⁺ balance in fiber cells and could cause cell swelling. Dysfunctional cell membrane selectivity to ions has been correlated with increased permeability (Duncan and Croghan, 1969) and is consistent with the association of increases in ion flux with the onset of cataracts(Matsuda et al., 1982). The SPARC-null mice that exhibit cortical lenticular opacity, but no involvement of the nucleus until later stages, might represent another cataract model with compromised permeability of the lens capsule.

Lens capsular BM is not a static structure but is continually produced and remodeled by lens epithelial cells and newly differentiated fiber cells(Johnson and Beebe, 1984). As SPARC regulates the production of ECM proteins and binds to several collagens including the BM collagen type IV (Lane et al., 1992; Kamihagi et al.,1994; Francki et al.,1999; Sage et al.,1989; Sasaki et al.,1998), our data are consistent with the proposal that SPARC affects the organization and assembly of the components in the lens capsular matrix. Additionally, it is highly likely that lens cells without SPARC exhibit abnormal behavior. For example, the protrusion of the fiber cells into the lens capsule indicates not only that the lens capsule is abnormal, but also that these fiber cells without SPARC become invasive and pathological. SPARC-null lens cells are clearly altered in their morphology, but their production of ECM components and heparan sulfate proteoglycan, as well as their migratory, adhesive and differentiation properties, are poorly understood and are currently under investigation in our laboratory.

It is interesting that age-related changes in the lenses of Fischer rats were morphologically similar to those observed in SPARC-null lenses. For example, at 18 months of age, Uga et al. observed posterior invasion by the processes of the cortical fiber cells toward the lens capsule and swelling of anterior cortical fibers (Uga et al.,1996). Whether or not age-related human cataracts exhibit any abnormalities in the lens capsule, as seen in SPARC-null mice, has not been reported and needs to be investigated. Elevated levels of SPARC mRNA and protein have been described in human cataratous lenses, an observation leading Kantorow et al. to propose that SPARC is associated with the process of human cataractogenesis (Kantorow et al.,2000).

In this study, the labeling of proliferating cells in vivo with BrdU tested the hypothesis that accelerated proliferation of lens epithelial cells in SPARC-null mice results in lenticular opacity. No significant differences in cell proliferation between wt and SPARC-null lenses from E14 to 4 months of age were observed, despite the presence of opacity in SPARC-null mice that began at 1 month. Only in SPARC-null mice of 7 months and older was there a significant increase in BrdU-positive cells, which occurred concomitantly with inflammation and tissue necrosis in the ruptured posterior area(Norose et al., 1998). It is likely that inflammatory cells recruited into the injured area stimulated an increase in BrdU-positive cells by their release of cytokines. In addition,there were no significant changes in the levels of MIP 26 and crystallins in SPARC-null lenses. The absence of proteolytic degradation of crystallins and MIP 26 indicate that the activation of proteases for these substrates is minimal in the SPARC-null lenses, at least up to 3 months of age.

SPARC interacts with ECM proteins (e.g., collagen types I, III and IV) and growth factors (platelet-derived growth factor and vascular endothelial growth factor) and influences cell proliferation, migration, adhesion,differentiation and/or barrier function(Yan and Sage, 1999;Brekken and Sage, 2001;Goldblum et al., 1994). If SPARC is important in the regulation of the distribution and organization of lens BM matrix, it follows that this protein is likely to influence lens epithelial cell growth and differentiation. This study indicates that compromised lens permeability is a consequence of the absence of SPARC and contributes to cataract formation. The mechanism(s) that accounts for the defects in the lens capsule will contribute to our understanding of the biological functions of SPARC and the organization of lens BM proteins and proteoglycans, as well as the importance of cell-matrix interactions in the lens.

We are grateful for the excellent technical assistance of D. Blake, C. Ganders and S. Lara. The authors thank X. Gong and J. Horwitz for the antibody against MIP 26. We thank D. Graves and R. Brekken for advice and our colleagues in the Sage lab for helpful discussions. This work was supported by National Science Foundation grant EEC-9814404, National Institutes of Health grant EY13180 and by a National Research Service Award F32-EY-06987 to Q.Y.