Actin in polygonal arrays of microfilaments and sequestered actin bundles (SABs) in lens epithelial cells of rabbits and mice

The pattern of localization of actin filaments was compared in whole mounts of lens epithelium of rabbit and mouse using the fluorescently-labeled actin-specific probe, rhodamine-phalloidin. In the adult rabbit lens, fluorescent polygonal arrays consisting of central vertices and interconnecting filaments were present in the apical end of each epithelial cell. Electron microscopy confirmed that these arrays lined the cytoplasmic side of the apical membrane. In the mature adult mouse, polygonal arrays were not seen either with fluorescence or electron microscopy. Instead, the actin was packaged in a single, elongated, often curved bundle near the epithelial cell apex, referred to as a "sequestered actin bundle" or SAB. The SAB often appeared attached to the plasma membrane and to approach the perinuclear basket of microfilaments. The significance of the differences in these two patterns of actin is discussed in terms of differences in the accommodative ability and static lens shape in these two animals.     

Comparative study of actin filament patterns in lens epithelial cells. Are these determined by the mechanisms of lens accommodation?

Actin filament patterns in lens epithelia from animals of various taxonomic groups were studied using rhodamine phalloidin fluorescence microscopy of epithelial whole mounts and transmission electron microscopy of tangential sections. The results were compared with the accommodative mechanism operating in each case as reported in the literature. Lenses that accommodate by deformation of the anterior surface, in squirrel, chipmunk, rabbit, monkey and human, showed polygonal arrays (PAs) at the apical end of the epithelial cells. Lenses that translate as a whole, in shark, bony fish and frog, showed stress fibers (SFs) at the basal or apical end of the cells. No specialized actin pattern was seen in turtle and bird, which have lenses that are squeezed into an anterior lenticonus; cat, where the lens is translated forward; or rat, cow and most mice, which have no defined accommodation. In exception, certain strains of laboratory mice did show sequestered actin bundles (SABs) and/or PAs. Based on our findings, we conclude that PAs, which resemble geodesic domes, do not take an active part in near-point accommodation; but like SFs, may serve to resist overextension by internal pressure of the fiber mass or by zonular tension.     

Development of actin polygonal arrays in rabbit lens epithelial cells

In searching for a clue to the role of actin filament bundles organized into polygonal arrays, or geodomes, in lens epithelial cells, we examined several physical events occurring in the young rabbit lens which may initiate their formation. We used NZW rabbits between the ages of 24 days gestation and 50 days postnatal. Data were obtained from TEM, SEM and fluorescence microscopy. Parameters measured were lens weights, apical surface areas of cells in epithelial whole mounts, epithelial cell thickness, and timing of eyelid opening, breakdown of the tunica vasculosa lentis (TVL) and formation of the ciliary zonules; these findings were correlated with the first signs of development of the arrays. Polygonal arrays formed slowly beginning at one to two days after birth, and with advancing time these thickened and made more numerous connections with the lateral plasma membranes. Development of the arrays was not correlated with onset of vision or disappearance of the TVL or a sudden increase in cell area, since these events occur postnatally at about 9-10 days, nor with the development of zonular fibers since these are already in place at 24 days of gestation. Only lens weights showed a dramatic increase between 24 days gestation and birth. It is surmised that the expanding lens mass may be involved in some way in signaling the organization of actin filaments into geodomes.     

Polygonal arrays of microfilaments in epithelial cells of the intact lens

Polygonal arrays of microfilaments have been discovered to line the inner apical plasma membrane of anterior epithelial cells of the intact rabbit lens. When tangential sections are studied with the electron microscope, the polygonal arrays are seen to consist of central vertices interconnected by rays of filaments. The rays near the cell periphery insert into the lateral plasma membrane. The vertices are spaced about 1 micron apart and appear to be attached to the apical plasma membrane. The polygonal arrays have little depth as judged by stereo-pairs and are incorporated within the dense band of microfilaments seen in cross-section at the epithelio-fiber junction. The diameter of the filaments and their similarity to actin-containing polygonal arrays described by other investigators in cultured cells suggest that these structures contain actin in lens epithelial cells. The function of the polygonal arrays in relation to maintenance of lens shape or to changes in lens shape in accommodation is discussed.     

Immunogold-EM localization of actin and vimentin filaments in relation to polygonal arrays in lens epithelium in situ

An indirect immunogold technique for transmission electron microscopy was used for localizing two cytoskeletal proteins, actin and vimentin, in the epithelium of freshly removed rabbit lens, especially in relation to the polygonal array structures located at the apices of the epithelial cells. Antibody specificity was determined on semi-pure chicken breast muscle actin and bovine lens vimentin using Western blotting of these proteins and extracts of rabbit lens epithelium separated by SDS-PAGE. Whole lenses of rabbits were lightly fixed in glutaraldehyde and embedded in LR White resin. Tangential sections were taken at 70 to 80 nm and at 0.25 micron and used for single-labeling, and double-labeling with antibodies raised in different hosts and treated with appropriate second antibodies conjugated with non-overlapping sizes of gold particles. Routine and stereomicroscopy were used to analyze gold-label patterns. The study shows that the rays of the polygons project deeply into the cell from the vertices lying on the inner apical membrane. Actin is located on the filaments of rays, but vimentin is not associated with the polygons at the level in the cell that we studied. Vimentin filaments are found in deeper regions of the epithelial cell. Stereopairs were useful in differentiating where the gold-label was located and in fact, this technique demonstrated that most of the label is on the surface of sections where the filaments are exposed.     

Formation of after-cataract by regeneration of human and rabbit lens epithelium in tissue culture

Regeneration of lens epithelium on the lens capsule was studied in tissue culture. The entire capsule with attached epithelium was taken from rabbit lenses and from human lenses with cataract. Generally, the epithelium grew in a monolayer but multilayered masses of cells were also seen. Most lens fibers degenerated during the first days and formed spherical membrane enclosed vesicles containing cytoplasm but no nuclei. The lens fiber remnants, together with regenerating epithelium, created structures that were similar in many ways to the clinical appearance of after-cataract.     

Immunocytochemical evidence for an actin-myosin system in lens epithelial cells

Since filamentous actin had been shown earlier to exist in lens epithelial and fiber cells, we inquired whether this could represent a contractile system with myosin and other actin-associated proteins. We resolved this question in freshly removed or organ-cultured rabbit and squirrel lens epithelial whole mounts using immunocytochemical techniques and by immunoblots of extracts separated by electrophoresis. In the former, methods were developed using long fixation times and long incubation in primary antibodies and biotinylated second antibodies visualized by streptavidin immunofluorescence and by diaminobenzidine peroxidase. Myosin was found to be localized along the filamentous rays and at central vertices of polygonal arrays situated at the apices of epithelial cells. It was not clear whether myosin and actin occurred together along the same or adjacent filaments in a bundle. Tubulin and vimentin were found deeper in the cells and were not aligned with actin and myosin filaments. Control lens epithelia treated similarly except for deletion of the primary antibodies showed no staining. As positive controls, pieces of glycerinated sartorius muscle exhibited characteristic cross-banded patterns of actin and myosin when incubated with the same reagents used on the lens epithelium. Denatured extracts of rabbit lens epithelium and of cortical fiber cells separated by electrophoresis and transferred to nitrocellulose paper, stained specifically with the same myosin and tubulin antibodies used in the immunocytochemistry experiments. The molecular weight profile of the myosin polypeptide indicated that lens tissue has myosin II. We conclude that a contractile system exists in lens epithelial and cortical fiber cells, although the function is not understood at this time. We conjecture that the system may act to stabilize lens shape by providing contractile tone.     

Tropomodulin and tropomyosin mediate lens cell actin cytoskeleton reorganization in vitro

PURPOSE: To determine the role of the actin cytoskeleton regulatory proteins tropomyosin and tropomodulin (Tmod) in the reorganization of the actin cytoskeleton during lens epithelial cell differentiation. METHODS: Primary cultures of chick lens epithelial cells were allowed to differentiate in vitro to form lentoid bodies. Localization of F-actin, Tmod, and tropomyosin were determined by immunofluorescent staining followed by confocal microscopy. Tropomyosin and Tmod isoform expression was determined by immunoprecipitation and western blot analysis. RESULTS: In undifferentiated epithelial cells F-actin was organized in polygonal arrays of stress fibers and was also associated with the adherens belt. In contrast, F-actin in differentiated cells was predominantly associated with membranes in a reticular or fibrillar pattern and was organized in curvilinear fibrils in the cytoplasm. Tmod was not detected in the undifferentiated epithelial cells but was expressed upon cell differentiation and assembled into F-actin and non-F-actin structures. Tmod isoforms expressed in the lens cell cultures were identical with those expressed in the embryonic chick lens fiber cells. Tropomyosin was associated with the polygonal arrays of stress fibers in the undifferentiated epithelial cells and was recruited to cortical F-actin at the cell periphery during differentiation. This occurred coincident with a shift in tropomyosin isoform expression. CONCLUSIONS: Expression and sequential assembly of low-molecular-weight tropomyosin and Tmod into the cortical actin cytoskeleton of differentiated lens cells may help to reorganize the actin cytoskeleton during morphogenetic differentiation. Moreover, lens epithelial cell differentiation may include the generation of novel Tmod-containing, non-F-actin cytoskeletal structures.     

The effects of near-UV radiation on elasmobranch lens cytoskeletal actin.

The role of near-UV radiation as a cytoskeletal actin-damaging agent was investigated. Two procedures were used to analyse fresh smooth dogfish (Mustelus canis) eye lenses that were incubated for up to 22 hr in vitro, with elasmobranch Ringer's medium, and with or without exposure to a near-UV lamp (emission principally at 365 nm; irradiance of 2.5 mW cm-2). These were observed histologically using phalloidin-rhodamine specific staining and by transmission electron microscopy. In addition, solutions of purified polymerized rabbit muscle actin were exposed to the same UV conditions and depolymerization was assayed by ultracentrifugation and high-pressure liquid chromatography. While the two actins studied do differ very slightly in some amino acid sequences, they would react physically nearly identically. The results showed that dogfish lenses developed superficial opacities due to near-UV exposure. Whole mounts of lens epithelium exhibited breakdown of actin filaments in the basal region of the cells within 18 hr of UV exposure. TEM confirmed the breakdown of actin filaments due to UV exposure. SDS-PAGE and immunoblotting positively identified actin in these cells. Direct exposure of purified polymerized muscle actin in polymerizing buffer led to an increase in actin monomer of approximately 25% in the UV-exposed solutions within 3-18 hr, whether assayed by ultracentrifugation or HPLC. The above indicates that elasmobranch lens epithelial cells contain UV-labile actin filaments, and that near-UV radiation, as is present in the sunlit environment, can break down the actin structure in these cells. Furthermore, breakdown of purified polymerized muscle actin does occur due to near-UV light exposure.(ABSTRACT TRUNCATED AT 250 WORDS)     

Immunologic conservation of the fiber cell beaded filament

Lenses were obtained from the eyes of four different classes of Chordates, including Mammalia (rat, mouse, cow, human), Aves (chicken), Amphibia (tiger salamander), and Osteichthyes (steelhead), as well as from one Mollusca (squid). Buffer soluble, urea soluble and urea insoluble fractions were prepared from each, and probed by western blot analysis for the presence of the lens fiber cell 115 and 49 kD beaded filament proteins. Application of both polyclonal and monoclonal antibodies revealed that an immunologic homologue to the bovine fiber cell 115 kD protein is present in all examples of Chordates tested, and that this homologue possessed properties very similar to those of its bovine counterpart. Both monoclonal and polyclonal antibodies revealed an immunologically cross-reactive homologue in squid as well, but suggested that the squid protein had a native molecular weight of closer to 70-80 kD. A monoclonal antibody to the bovine 49 kD beaded filament protein was successful at identifying an immunologic homologue to this protein in mouse, chicken, and tiger salamander. Ultrastructural analysis of rat, human, and fish lenses showed that a beaded filament was present in these lenses, which was indistinguishable from that seen in the bovine lens. In the squid a filamentous, beaded structure was observed, but it differed from that seen in the bovine lens. We conclude from the data presented that the beaded filament, and its constituent proteins, are well-conserved. This data should facilitate the identification of lens cytoskeletal proteins and structure in a wide range of animal models, and establish that probes for these proteins may be of broad applicability.     

Expression and activity of the signaling molecules for mitogen-activated protein kinase pathways in human, bovine, and rat lenses

PURPOSE: The mitogen-activated protein kinase (MAPK) pathways play distinct roles in the lens. However, the expression patterns and activity levels of various components for these pathways have not been well-documented in vertebrate lenses, especially human lens. In the present study, the expressions and activities of extracellular signal-regulated kinase (ERK)-1/2/3, c-Jun NH2-terminal kinase (JNK)-1/2, p38 kinase, mitogen-activated protein kinase kinase (MEK)-1/2, and RAF1 were recorded in human, bovine, and rat lenses. METHODS: Human, bovine, and rat lenses were isolated from intact eyes. The epithelia and different layers of fiber cells were isolated from these lenses. Total proteins extracted from these samples were subject to analysis of the expression patterns and activity levels of the MAPKs and the activating kinases of ERK1/2. RESULTS: ERK1 and ERK2 were the most abundant MAPKs in terms of both protein and activity levels in all lenses. JNK1 and JNK2 were highly expressed in bovine lens, which differed from the pattern shared by human and rat lenses. p38 kinase was similarly expressed in bovine and rat lenses, but different from that in human lens. However, p38 kinase activity was exclusively detected in the epithelia. All lenses had MEK1/2 activity in their epithelia but the expression patterns of MEK1 and MEK2 differed in these lenses. RAF1 was expressed in the epithelia of all lenses, but its activity was detected only in rat lens. CONCLUSIONS: ERK1 and ERK2 are the most abundant MAPKs in the ocular lens, providing the basis for their multiple functions in lens development and pathogenesis. The dominant epithelial distribution of JNK1/2 and p38 kinase suggests that the lens epithelium is a major site for stress response. ERK1, p38 kinase, and PKCalpha can be used as molecular markers for aging.     

Histological differentiation in fetal organ-cultured and human cataractous eye lenses

Swine fetal organ-cultured lenses were studied histoautoradiographically and histologically, and human cataractous lenses were examined histologically to investigate the continuous differentiation of epithelial cells into fiber cells in the lens equatorial region. Both lens groups studied presented abnormal differentiation: the equatorial epithelium proliferates; nucleus-containing cortical fibers turn into bladder cells (of Wedl), and the formation of new cortical fibers stops.     

Human lens hexokinase

Clear crystalline lenses were obtained post-mortem from human prenates, neonates, young and old adults. Senile cataracts removed surgically were also studied. The specific activity of soluble fetal lens hexokinase is equal to that of the rat lens. There is a progressive decrease in specific activity as the lens size increases. In the fetal lens cortex and nucleus there is high hexokinase activity; young adult and cataractous lens nuclei are practically devoid of hexokinase activity. The epithelium and most superficial cortex possess high hexokinase levels even in the senile cataract. There is a large amount of insoluble hexokinase in the epithelium and most superficial cortex of the senile cataract. This can be solubilized with salts and detergents. No soluble or insoluble hexokinase is found in the nucleus of the senile cataract. Starch gel electrophoresis reveals the presence of Type I and probably also Type II hexokinase in the human lens. Types I and II were definitely found in the lens of the Rhesus monkey. There is a thermolabile form (presumably Type II hexokinase) and a thermostable form (presumably Type I) of hexokinase in the human lens. Michaelis constants were measured. In the fetal lens: K_M glu = 0·11 mm, K_{M ATP} = 0·74 mm; in the senile cataract: K_{M glu} = 0·045 mm, K_{M ATP} = 0·385 mm. There is great similarity between the hexokinases of the rat, rabbit and human lens. Further investigation employing lenses of these experimental animals is likely to yield results relevant to the human lens.     

Lens epithelium and fiber Na,K-ATPases: distribution and localization by immunocytochemistry.

PURPOSE: To use immunofluorescence and immunogold techniques to identify the catalytic subunits of the Na,K-ATPases of the lens and to determine their location in the cells of the epithelium and cortex of bovine and human lenses. METHODS: Frozen sections of capsulated and decapsulated bovine and human lenses were prepared, blocked, and treated with affinity-purified polyclonal rabbit antibodies to the Na,K-ATPase catalytic subunit isoforms with subsequent treatment with fluorescein isothiocyanate-labeled goat anti-rabbit IgG and visualization of the fluorescence by light microscopy. An immunogold-labeled goat anti-rabbit IgG was used to detect, by electron microscopy, the binding of the same affinity-purified polyclonal antibodies to thin sections of decapsulated lenses that had been fixed and embedded in Lowicryl K4M. The results were confirmed by staining of western blot analysis of sodium dodecyl sulfate-polyacrylamide gel separations of enriched membrane preparations from bovine and human lenses. RESULTS: The three common catalytic subunits of the Na,K-ATPases are present in the plasma membranes of lens epithelium, lens fibers, or both. The data indicate a polarized distribution of the alpha1 and alpha3 catalytic subunit isoforms in central epithelium. In the cortical fibers, the alpha2 isoform is present around the interdigitations. The alpha3 isoform is found in the interdigitation-free regions of human cortical fibers. CONCLUSIONS: This unique distribution of Na,K-ATPases precludes the popular pump-leak model for lens monovalent cation homeostasis. The functional significance of the distribution of Na,K-ATPases in the lens epithelium and superficial fibers is currently under investigation.     

Age-related changes in the rat lens: morphological and morphometrical studies

Age-related changes in the lenses of 90 Wistar male rats, whose ages ranged from 7 to 26 months, were studied morphologically and morphometrically. Morphological studies were performed using biomicroscopy, light microscopy, scanning electron microscopy and transmission electron microscopy. Early cataract formation was observed biomicroscopically in rats of 19 months of age and the number of cataractous lenses increased with aging. Age-related morphological changes in the lenses were observed before cataract formation and these histological changes showed various forms with aging. Flat preparations of the lens epithelium were examined with a computer-assisted cell-analyzer. The mean cell density decreased and the mean cell area enlarged in both central and germinative zones of the lens epithelium with aging. These results suggest that age-related changes of lens epithelial cells do occur and may be associated with the formation of Wistar rat cataract.     

Expression of alphasmooth muscle actin in lens epithelia from human donors and cataract patients

In order to re-evaluate functional implications of alphasmooth muscle actin (alphaSMA) expression in lens epithelial cells (LECs), we assessed its presence in donor lenses without visible opacities (DON), lenses with mature cataract (CAT), and cataractous lenses with posterior subcapsular opacities (PSO) or anterior subcapsular fibrosis (ASF). The levels of alphaSMA and transforming growth factor-beta2 (TGFbeta2) mRNAs were measured by classical and real-time PCR. Expression and structural organisation of alphaSMA protein and beta-catenin were monitored by Western blotting and confocal microscopy. All DON analysed contained measurable amounts of alphaSMA mRNA. In CAT without and with PSO, mRNA expression was increased and, again more than doubled in ASF. TGFbeta2 mRNA expression varied widely between the individual samples but was slightly increased in ASF. No correlation existed between alphaSMA or TGFbeta2 expression and the age of the donors in any of the lens categories. Confocal microscopy revealed that, in DON and CAT, alphaSMA was preferentially expressed in a simple granular pattern in single or small clusters of LECs within a normally shaped cobblestone epithelium. Locally, the granules were merged into short stretches at the cell margin. In CAT, a few abnormally shaped cells contained polygonal alphaSMA structures and short stress fibres. In CAT with PSO and ASF, polygons and stress fibre bundles predominated in spindle-shaped cells. Expression patterns of different complexity were often present in the same epithelium. Apical polygons and basal stress fibres were detected within the same cell and may reflect instability of the interface between epithelium and cortical fibres and changes in adhesion to the capsule, respectively. High levels of betacatenin mRNA and protein were present in all lens types. However, with increasing complexity of alphaSMA organisation, betacatenin staining disappeared from the cell margin and basal infoldings and was shifted towards the cytoplasm and nucleus. The presence of alphaSMA in DON, the absence of any correlation between mRNA level and age, and the modest increase in complexity of alphaSMA-containing structures in CAT argue against an inevitable link between alphaSMA expression and the development of age-related cataract. Low levels of alphaSMA expression may reflect repair of normal wear and tear. In pathologic situations such as PSO and ASF, persisting stimulation and additional incentives may induce increased alphaSMA expression and more elaborate patterning, eventually leading to completion of EMT.     

波形纤维蛋白在老年性白内障晶状体上皮细胞的表达

目的观察波形纤维蛋白在老年性白内障晶状体上皮细胞的变化.方法用卵白素-生物素过氧化物酶法对22例老年性白内障患者晶状体前囊膜上皮细胞进行波形纤维蛋白染色,采用包埋前免疫酶电镜技术处理6个晶状体前囊膜标本,并观察其超微结构;利用十二烷基硫酸钠聚丙烯酰胺凝胶电泳及Westemnblot法分析4个晶状体表层组织(前囊膜、上皮细胞和表层皮质)中的波形纤维蛋白.结果老年性白内障晶状体上皮细胞的波形纤维蛋白表达减弱,与对照组比较差异有显著性(t=2.0948,P<     

Decreased caspase-3 activity in human lens epithelium from posterior subcapsular cataracts

Apoptosis has been implied in normal lens development in the embryo as well as in lens fibre differentiation. It has also been suggested to play a role in non-congenital cataract and in the formation of posterior subcapsular opacification, but data on the presence of apoptosis in human lens epithelium from cataractous lenses are scarce and conflicting. The present study aimed to investigate apoptosis in lens epithelium from patients undergoing cataract surgery. The amount of apoptosis detected was correlated to age, gender, type of cataract, medications and disease. Moreover, the ability of human lens epithelial cells in culture to respond to the apoptosis-inducing agent staurosporin by activation of caspase-3 was investigated. Human lens capsulotomy specimens were collected immediately after surgery, frozen and later analysed with respect to caspase-3 activity, using the fluorogenic substrate Ac-DEVD-AMC. Generally, the activity of caspase-3 detected in this manner was very low and in 23% of the specimens it was non-detectable. However, there were differences in caspase activity between lens epithelial cells from different types of cataract, where samples from lenses with posterior subcapsular cataract exhibited significantly lower caspase-3 activity than lenses with a clear subcapsular zone. Age, gender or medications did not show any correlation with caspase activity but human capsulotomy specimens from diabetic patients exhibited significantly lower caspase-3 activity. Staurosporin caused a concentration-dependent increase in caspase activity in cultured human lens epithelial cells and the amount of apoptotic nuclei was also increased as viewed by staining with Hoechst 33342, showing chromatin condensation and nuclear fragmentation. Similar results were obtained when fresh human lens capsulotomy specimens were exposed to 1000 nM staurosporin for 24 hr. To conclude, the present data indicate that human lens epithelial cells have the ability to respond to apoptosis-inducing agents with caspase-3 dependent apoptosis, and that even though the general level of apoptosis in human lens epithelium in vivo is low, there are differences in caspase-3 activity levels in lenses with or without posterior subcapsular cataract. The latter finding supports previous studies indicating that this type of cataract may result from defective differentiation, in which apoptosis may play an important role.     

Lenses of SPARC-null mice exhibit an abnormal cell surface-basement membrane interface

SPARC (secreted protein acidic and rich in cysteine) is a matricellular protein involved in cell-matrix interactions. We have shown previously that mice deficient in SPARC develop posterior cortical cataract early in life that progresses to a mature opacity and capsule rupture. To evaluate the primary effects of SPARC deficiency in the lens, we examined the lenses of SPARC-null and wild-type mice by electron microscopy and immunohistochemistry to investigate whether ultrastructural abnormalities occur at the basement membrane (capsule)-lens cell interface in SPARC-null mice. The most notable feature in the lenses of SPARC-null mice, relative to wild-type animals, was the modification of the basal surface of the lens epithelial and fiber cells at the basement membrane (capsule) interface. Electron microscopy revealed numerous filopodial projections of the basal surface of the lens epithelial and fiber cells into the extracellular matrix of the anterior, posterior, and equatorial regions of the lens capsule. In 1 week old precataractous lenses, basal invasive filopodia projecting into the capsule were small and infrequent. Both the size and frequency of these filopodia increased in precataractous 3-4 week old lenses and were prominent in the cataractous 5-6 week old lenses. By rhodamine-phalloidin labeling, we confirmed the presence of basal invasive filopodia projecting into the lens capsule and demonstrated that the projections contained actin filaments. In contrast to the obvious abnormal projections at the interface between the basal surface of the lens epithelial and fiber cells and the lens capsule, the apical and lateral plasma membranes of lens epithelial cells and lens fibers in SPARC-null mice were as smooth as those of wild-type mice. We conclude that the absence of SPARC in the murine lens is associated with a filopodial protrusion of the basal surface of the lens epithelium and differentiating fiber cells into the lens capsule. The altered structures appear prior to the opacification of the lens in the SPARC-null model. These observations are consistent with one or more functions previously proposed for SPARC as a modulator of cell shape and cell-matrix interactions.