Age-dependent changes in the structure, composition and biophysical properties of a human basement membrane

Matrix Biology 29 (2010) 402–410

Contents lists available at ScienceDirect

Matrix Biology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / m a t b i o

Age-dependent changes in the structure, composition and biophysical properties of a human basement membrane Joseph Candiello a, Gregory J. Cole c, Willi Halfter b,⁎ a b c

Department of Bioengineering, University of Pittsburgh, North Carolina Central University, United States Department of Neurobiology, University of Pittsburgh, North Carolina Central University, United States Julius L. Chambers Biomedical Research Institute, North Carolina Central University, United States

a r t i c l e

i n f o

Article history: Received 19 January 2010 Received in revised form 17 March 2010 Accepted 24 March 2010 Keywords: Basement membrane Inner limiting membrane Proteoglycans Laminin Collagen IV Eye

a b s t r a c t Basement membranes (BMs) are considered to be uniform, approximately 100 nm-thin extracellular matrix sheets that serve as a substrate for epithelial cells, endothelial cells and myotubes. To find out whether BMs maintain their ultrastructure, protein composition and biophysical properties throughout life the natural aging history of the human inner limiting membranes (ILM) was investigated. The ILM is a BM at the vitreal surface of the retina that connects the retina with the vitreous. Transmission electron microscopy (TEM) showed that the ILM steadily increases in thickness from 70 nm at fetal stages to several microns at age 90. By the age of 20, the ILM loses its laminated structure to become an amorphous and very irregular extracellular matrix layer. Atomic force microscopy (AFM) showed that the native, hydrated ILMs are on average 4-fold thicker than the dehydrated ILMs as seen by TEM and that their thickness is prominently determined by its water-binding proteoglycans. The morphological changes are accompanied by age-related changes in the biochemical composition, whereby the relative concentrations of collagen IV and agrin increase, and the concentration of laminin decreases with age. Force-indentation measurements by AFM also showed that ILMs become increasingly stiffer with advancing age. The data suggest that BMs from other human tissues may undergo similar age-related changes. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Basement membranes (BMs) are extracellular matrix sheets that are found at the base of every epithelium, at the surface of all muscle fibers and at the basal surface of vascular endothelial cells. BMs are composed of extracellular matrix proteins that include members of the laminin, nidogen collagen IV and proteoglycan families (Timpl and Brown, 1996; Erickson and Couchman, 2000). The commonly accepted model proposes that two networks of polymerized laminins and crosslinked collagen IVs constitute the basic frame-work of every BM. Deletion or mutations of BM proteins result in embryonic lethality (Smyth et al., 1999; Willem et al., 2001; Costell et al., 1999; Arikawa-Hirasawa et al., 1999; Poeschl et al., 2004; Bader et al., 2005) or, postnatally, in muscular dystrophy, paralysis (Gautam et al., 1996), vascular disruption (Gould et al., 2005; 2007), and eye and brain abnormalities (Sertie et al. 2000; Fukai et al., 2002; Halfter et al., 2002; 2005a; Lee and Gross, 2007). The fact that many of the BM protein mutations cause vascular disruptions with increasing blood pressure during later embryogenesis showed that BMs have an important role in the mechanical stability of tissue walls (Candiello et al., 2007).

⁎ Corresponding author. Tel.: +1 412 648 9424; fax: +1 412 648 1441. E-mail address: [email protected] (W. Halfter). 0945-053X/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.matbio.2010.03.004

Most studies on BMs were conducted in mice, rats, fish and Drosophila, laboratory animals that have a very limited life span. They show that BMs are rather uniform and thin ECM sheets of less than 100 nm thickness. In the longer-lived humans, however, many extracellular matrix tissues undergo age-related structural and compositional changes that only become obvious after decades of life. It is well conceivable that human BMs also undergo age-related structural and compositional changes and that these changes are only minor or even undetectable in most laboratory animals. To this end, we investigated the age-related structural, compositional and biomechanical changes of a human BM, the inner limiting membrane (ILM). The inner limiting membrane is a BM that is located at the border between the retinal neuroepithelium and the vitreous body (VB). Proteome analysis of the chick ILM showed that it is composed of the BM-typical ECM proteins that are found in other BMs as well (Candiello et al., 2007; Balasubramani et al., in press). Analysis of mice, zebra fish and humans with mutations or deletions of BM proteins (Zenker et al., 2004; Halfter et al., 2005a; Semina et al., 2006; Lee and Gross 2007) or their receptors (Georges-Labouesse, 1998; Satz et al., 2008) has shown that the ILM is essential in retinal histogenesis by establishing a substrate for the neuroepithelial endfoot processes to attach to. In cases where the ILM was defective, the Muller cell processes were retracted, and the histogenesis of the retinal cell layers was greatly disrupted, eventually resulting in retinal dysplasia, retinal ectopia and a massive loss of

J. Candiello et al. / Matrix Biology 29 (2010) 402–410

ganglion cells and axons (Halfter et al., 2005a, Semina et al., 2006; Lee and Gross 2007). The ILMs can be isolated, and it is one of the few BMs that are accessible for direct biochemical analysis. In addition, the ILM can be flat-mounted on glass slides and is thereby suitable for biomechanical measurements using atomic force microscopy (AFM; Candiello et al., 2007). Thus, the ILM provides unique experimental advantages over most other BMs that are inseparable from adjacent connective tissues and that are difficult to mount for biophysical examination. The present study shows that the human ILM undergoes agedependent alterations that include a dramatic increase in thickness, a loss of the typical BM ultrastructure, an increase in stiffness and agerelated changes in its biochemical composition. We propose that BMs from other tissues undergo similar age-related changes that may contribute to an increase in disease frequency at older age. 2. Results 2.1. Location, protein composition and morphology of the fetal human ILM The ILM is located at the vitreal border of the retina (Fig. 1A). It is one of 6 BMs of the eye, which include the lens capsule, the BMs of the cornea, the BM of the pigment epithelium and the BMs of the intravitral hyaloidal vessels. The protein composition of the ILM from human fetal eyes was analyzed by immunocytochemistry using antibodies against basement membrane-specific proteins. Consistent with earlier reports, labeling of the ILM and the other ocular BMs was recorded after staining with antibodies to collagen IV (Fig. 1A), laminin-1 (Fig. 1B, C), nidogen1, perlecan, collagen XVIII and agrin (Halfter et al., 2005b, 2008). The immunocytochemistry was reproduced with 6 fetal human eyes and with 10 adult human eyes from ages 22 to 86. Transmission electron microscopy (TEM) of 10 and 20-week fetal human retina samples showed that the ILM of the fetal human eye has the typical ultrastructure of a standard BM with a central lamina densa and an outer and inner lamina rara (Fig. 2A). The ILM of the 10 and 20-week fetal human eye has a thickness of 70± 7 nm (n = 6 pairs of fetal eyes).

403

2.2. Structure and composition of the adult human ILM To find out whether the ILM undergoes age-related structural changes thin sections of the retina from fetal and adult human eyes were examined by TEM. Since the structure of the adult human ILM varies depending on the location of the retina (Foos, 1972), all samples were taken from the dorso-central retina (Fig. 2E). Different to the 70 nm-thin and very even ILM at 20 weeks of gestation (Fig. 2A), the adult human ILM loses its distinctly layered ultrastructure and becomes gradually thicker and more irregular with advancing age. By age 22, the thickness of the ILM had increased to 300–350 nm and had lost its lamina densa and rara substructures (Fig. 2B). By age 50, the ILM had further increased in thickness and the retinal surface of the ILM had long extensions deep into the retina (Fig. 2C). These indentations were even more pronounced at older age. The vitreal surface of the ILM remained smooth at all ages. By age 83, the ILM thickness had increased to over 1500 nm (Fig. 2D), over 20 fold relative to the fetal ILM. The graph in Fig. 2F shows the age-dependent increase in human ILM thickness. The increasingly larger standard deviations at older age accounted for the more frequent and more extensive indentation of the ILM at its retinal surface. 2.3. Isolation of the adult human ILM and the protein composition of the aging human ILM To analyze the protein composition and to probe the biomechanical qualities of the adult human ILM we devised a method to isolate the ILM from adult human retina. The procedure takes advantage of the fact that BMs are insoluble in detergent (Meezan et al., 1975; Duhamel et al., 1983; Halfter and von Boxberg, 1992): following the dissolution of the retinal cells in Triton-X-100 and deoxycholate the ILMs were collected from the detergent solution under a dissecting microscope using darkfield illumination (Fig. 3A, insert). TEM micrographs (Fig. 3A) showed that the isolated ILMs were sheets of extracellular matrix, free of cellular debris and organelles. Further, the vitreal surface of the sheets was smooth and even, whereas the retinal surface had a very irregular appearance, identical to the retinal surface of the ILM in situ (Fig. 3B).

Fig. 1. BMs in the fetal human eye. The low power overview of an 10-week fetal human stained for collagen IV (red) shows the BM of the lens, the hyaloids vessels in the vitreous (HV), the pigment epithelium (PE), the ILM, and the vascular BMs in the eye lids. The corneal stroma is also labeled. High power micrographs of the 10 week fetal retina (B) show strong labeling of the ILM and the BM of the pigment epithelium and the choroid for laminin-111. Staining of the adult human retina for laminin-1 (1C) shows the ILM and the BMs of the retinal blood vessels (RV). Bars: a: 200 µm; B, C: 50 µm.

404

J. Candiello et al. / Matrix Biology 29 (2010) 402–410

Fig. 2. Age-dependent increase in ILM thickness. TEM micrographs show the vitreal surface of the retina from a 16-week fetal (A) and from 22, 54 and 83 year-old adult human eyes (B–D). The ILM from the fetal eye has the typical morphology of a BM (A): it is 70 nm thick with a typical lamina densa. Between 22 and 83 years of age (B–D), the ILM dramatically increases in thickness, becomes highly irregular at its retinal surface and is no longer sub-structured with a distinct lamina densa. The black bars in (A) to (D) indicate the thickness of the ILM. Panel (E) shows an adult human retina with the location in the dorso-central retina of a right eye where the samples (S) were taken from. The graph in panel (F) shows the age-dependent increase in thickness. Note the increasing variation of ILM thickness with advancing age, demonstrated by the large standard deviations. The data point for the fetal eyes is the average of measurements from 6 pairs of eyes between 10 and 20 weeks of gestation. Each of the remaining data points represents averages from measurements of one pair of eyes. F: fovea. OP: optic papilla; D: dorsal; V: ventral. Scale bar: 500 nm.

The ILM sheets were readily differentiated from the vascular BMs that were isolated from the detergent extracted human retina as well (Fig. 3C). To further prove the purity of the isolated ILM preparations, mice were immunized with the BMs, and the antisera from the mice were tested by immunocytochemistry. As shown in Fig. 3D, the antisera labeled only the ILM and vascular BMs of human retina, identical to the labeling obtained by using anti-collagen IV or anti-laminin antibodies (Fig. 3E). Previous reports have shown that the adult human ILM is comprised of laminin 111 (Fig. 1C), nidogen 1, collagen IV, agrin, perlecan and collagen XVIII (Libby et al., 2000; Halfter et al, 2005b; 2008), ECM proteins that had been detected in the fetal ILM as well (Fig. 1A, B). To determine to what percentile collagen IV and laminin 1 contribute to the ILM from an 88-year-old human eye we determined the concentration of both proteins ILM by western blotting (Fig. 4A–D). Reference samples of purified collagen IV and laminin 111 were used to establish calibration curves. Densitometry and extrapolation from the calibration curves showed that laminin γ1 accounted for 7.5% of the total ILM protein (Fig. 4A, B). The concentration of collagen IV was 57% of the total ILM protein (Fig. 4C, D) showing that collagen IV is the dominant protein of the aged human ILM. To determine if the human ILM undergoes a compositional change with advancing age we compared the relative concentrations of collagen IV, laminin and agrin core protein in ILMs from middle-aged eyes from 40 to 50 years of age and ILMs from eyes of over 80 years of age (Fig. 4E, F). Results showed that the relative concentration of collagen IV and agrin was higher in the older than in younger eyes, whereas the concentration of laminin was slightly lower in older than in younger eyes. The western blotting data were obtained

with each 2 pairs of young and old eyes, and the results for each of ILM sample pairs were reproduced at least 5 times. 2.4. Age-dependent changes in the biophysical properties of the adult human ILM To investigate the biomechanical qualities of the adult human ILM, segments of human ILMs were mounted on glass slides (Fig. 5A, B) and probed by atomic force microscopy (AFM). It is important pointing out that the measurements by AFM were taken while the ILM was submerged in PBS, thus the preparations were fully hydrated. Force indentation by the AFM probe has previously been used to determine the elasticity of cells (Binnig et al., 1986; Laney et al., 1997; A-Hassan et al., 1998; Quist et al., 2000; Begley and Mackin, 2004), and the thickness and stiffness of the native ILM from chick embryos (Candiello et al., 2007). Two parameters were determined: thickness and stiffness. Both data sets were plotted according to the age of the donor eyes. AFM measurements showed that the thickness of the ILM and the standard variation in thickness increased with advancing age (Fig. 5C, D). The Young's modulus of the adult human ILMs was measured between 1.5 and 5 MPa (Fig. 5E), and the graph in Fig. 5F showed that the Young's modulus increased with advancing age. It is of note that Young's modulus is a measure for the material quality of the BM and independent of its thickness. The Young's modulus was also determined for human lens capsule by using the same AFM probing technique as for the ILM. The primary goal of these experiments was to find out how the AFM measurements compare with elasticity measurements of lens capsule probed by

J. Candiello et al. / Matrix Biology 29 (2010) 402–410

405

Fig. 3. Isolation of the inner limiting membrane (ILM) of the adult human retina. TEM micrograph (A) shows an isolated and folded human ILM with a smooth vitreal (Vit) and an irregular retinal surface (RET). The insert shows a rolled-up human ILM as seen under a dissecting microscope using dark-field illumination. The isolated ILMs (A) are structurally very similar to ILMs in situ (B). The bars in (B) indicate the alterations in the thickness of the ILM. The morphology of vascular BMs from human retina, as shown by scanning electron microscopy, is shown in (C) for comparison. An antiserum from a mouse immunized with purified human ILMs exclusively labeled the ILM and the BMs of the blood vessels (BV) in sections of human retina (D). The staining pattern was similar to that obtained by staining with anti-collagen IV (E). The sections in (D) and (E) were counter stained with Sytox Green. Bar scales: A; 2 µm; B: 500 nm; C: 100 µm; D, E: 50 µm.

conventional hydrostatic measurements (Fisher, 1969; Krag and Andreassen, 2003; Danysh et al., 2008). The Young's modulus of human lens capsule, as measured by AFM, was 3.92, 2.70 and 4.37 MPa for 3 lens capsules, similar to the data obtained by hydrostatic measurements (Fisher, 1969; Krag and Andreassen, 2003; Danysh et al., 2008). Further, the Young's modulus of the human lens capsule was in the same range as for the adult human ILM and for the embryonic chick ILM, despite the fact that the lens capsule is about 10-times thicker than the human and 100 times thicker that the embryonic chick ILM, emphasizing again that the Young's modulus is a measure for the biomechanical property of BMs and independent of the thickness of the sample. Finally, the fact that the lens capsules had never been in contact with detergent yet resulted in a similar Young's modulus as the detergent-treated human and chick ILMs also showed that Triton/ deoxycholate treatment of BMs does not affect their biomechanical quality. Comparing thickness measurement of AFM and TEM recordings showed that the hydrated ILMs were by average four times thicker than the dehydrated ILMs as prepared for TEM. The major difference in the thickness of the hydrated ILM vs. the dehydrated ILM (Fig. 6A) indicated that some of the ILM components bind large quantities of water to the native BM. The most likely candidates for binding high amounts of water are the GAG side chains of the proteoglycans perlecan, agrin and collagen XVIII that are present in the ILM (Halfter et al., 2005b). To test whether the removal of the GAG side chains has an effect on the thickness and biomechanical properties of BMs human ILMs were treated with heparitinase/chondroitinase, flat-mounted, and probed by AFM. The measurements from the enzyme-treated ILM samples were compared with those from non-treated control samples. Results showed that the removal of the GAG side chains resulted in a reduction in ILM thickness that ranged between 30 and 50% depending on the age of the

eye from which the ILM was obtained (Fig. 6B). Further, the Young's modulus of ILMs increased by at least 30% after GAG removal (Fig. 6C), showing an increase in ILM stiffness after the GAG side chains had been removed. The decrease in thickness and the increase in the Young's modulus after GAG removal were less pronounced with advancing age (Fig. 6B, C). The selectivity of the enzyme treatment was confirmed by staining retinal sections from human eyes for heparan sulfate side chains (a-HS) or for the core proteins of agrin, one of the proteoglycans. In control sections, an antibody to HS-side chains labeled the ILM and the BM of the retinal blood vessels (Fig. 6D; cont). After glycosidase treatment, the HS staining was no longer detectable (Fig. 6E; −HS). When ILMs were labeled for the agrin core protein, the ILM labeling of the non-treated control (Fig. 6F) and the glycosidase-treated retinal sections (Fig. 6G) was indistinguishable. The data showed that heparitinase/chondroitinase treatment led to the removal of the GAG side chains of the proteoglycans, but did not lead to the degradation of their protein cores. Further confirmation came from western blots (Fig. 6H): ILM samples were incubated with BSA or with BSA plus heparitinase/chondroitinase. The western blots were labeled for agrin (a-Ag) and laminin (a-Ln). In control samples treated with BSA alone, agrin appeared as a long smear of 400kD, typical for highly glycosylated proteoglycans (Fig. 6H; cont). In samples treated with heparitinase/ chondroitinase, the agrin immunoreactivity appeared as a relative narrow band at 200 kD, the predicted molecular weight of the agrin core protein minus the heparan sulfate side chains (Fig. 6H; −HS). When blots with the same samples were labeled for laminin, the same prominent 200 kD band of laminin γ1 appeared in the control and the enzyme-treated samples (Fig. 6H). The data combined demonstrate that the heparitinase/chondroitinase treatment led to the removal of the GAG side chains from the BM proteoglycans, but did not lead to a generic proteolysis of BM proteins.

406

J. Candiello et al. / Matrix Biology 29 (2010) 402–410

Fig. 4. Western blots stained for laminin (A) and collagen IV (C) to determine the concentration of both proteins in adult human ILM. The ILM samples were run out on SDS PAGE in parallel with either laminin (A) or collagen IV references (C). The calibration curves (B, D) were established by densitometry of the laminin and collagen proteins bands indicated by arrows. The laminin concentration of the 88-year-old human ILM was estimated at 15 µg/ml, which equals 7.5% at a total ILM protein concentration of 200 µg/ml. The concentration of collagen IV was estimated for the same sample at 115 µg/ml, with equals 57% of the total ILM protein. Age-dependent changes in the concentrations ECM proteins in adult human ILM are shown in panels (E) and (F). Samples of 47 and 83-year-old ILMs were probed by western blotting for collagen IV (Coll4), laminin (Ln) and agrin (Ag). The ILM samples from young and old eyes were adjusted to identical total protein concentrations. The bar graphs below are based on densitometry of the labeled bands (density × pixel number). The graphs show that there is more (% relative to younger ILM sample) collagen IV and agrin core proteins in the old ILM and similar or slightly less laminin in the younger ILMs (E). Very similar data were obtained with another set of 52 and 85-year-old ILM samples (F). The bands outlined by the bars were scanned for collagen IV.

3. Discussion 3.1. Age-dependent changes in the morphology of the ILM The ILM of fetal human eyes is a 70 nm thick, sub-structured ECM sheet that morphologically resembles a standard BM. It is comprised of laminin 111, collagen IV, nidogen 1, perlecan, agrin and collagen XVIII (Libby et al., 2000; Halfter et al., 2005b; 2008), the typical proteins that are found in other BMs as well (Timpl and Brown, 1996; Erickson and Couchman, 2000). A recent mass spectrometry analysis showed that the ILM from chick embryos consists of 26 proteins, whereby laminins and nidogens are the dominant components, whereas collagen IVs account for only 3% of the total BM proteins (Balasubramani et al., in press). We expect a similar extended list of proteins and a similar stoichiometry of constituents for the fetal human ILM. The adult human ILM is structurally very different from the fetal ILM: it is irregular with long indentations into the retinal tissue, no longer sub-structured with a central lamina densa and much thicker. The thickness and the extensions of the ILM increase continuously with advancing age, and thickness and the degree of indentations can be used to estimate the age of the donor eye. The current data are consistent with earlier publications on the ultrastructure of ILM from adult humans and monkeys (Gartner, 1970; Foos, 1972; Matsumoto et al., 1982). The massive increase in thickness and the irregular structure of the ILM as seen in older humans and monkeys are not detectable in the eyes of aged rats, rabbits, chickens and pigs. Agedependent thickening of other BMs, however, has been reported for the epithelial BM of the choroid plexus and seminal vesicles in rats (Serot et al., 2001; Richardson et al., 1995) and for the vascular BMs of the inner ear in gerbils (Thomopoulos et al., 1997). The rate of thickening in these

cases ranged between 50 and 90%, much less than the over 2000% increase of the human ILM over 80 years of lifetime. Except for primates, laboratory animals are, therefore, less than perfect models for studying the normal aging process of BMs. Several reports have shown that BMs from other human tissues become thicker at old age as well, including BMs from capillaries and seminephros tubules (Xi et al., 1982) and the glomerular BMs of the kidney (Bloom et al., 1959; Anderson and Brenner, 1986), yet a detailed time-course study combined with biochemical and biophysical experiments as presented here has not been reported for any of these cases. A connection of BM thickening to age-related pathogenesis has been proposed, including a causal relationship to the age-dependent loss of neurons in the brain, to hearing loss, to loss of male fertility and to a deteriorating filtration rate in kidney glomeruli and the choroid plexus. Studies in chick, mice and human have demonstrated that ILM proteins are synthesized by the lens, the ciliary body and the intraocular vasculature and that the proteins are shed into the vitreous prior to their incorporation into the ILM (Sarthy and Fu, 1990; Dong and Chung, 1991; Sarthy 1993; Dong et al., 2002; Halfter et al., 2005b; 2008). The fact that the ILM protein concentrations in the human vitreous are very high during fetal development and sharply decline within 2 years after birth (Halfter et al., 2005b; 2008) indicated that the ILM proteins in the human eyes are prominently synthesized during fetal and early postnatal life and that the ILM has little turnover in the adult (Halfter et al., 2005b; 2008). Yet, the continuous increase in ILM thickness with advancing age, as reported here, indicates that ILM protein synthesis in the adult human eye does occur, yet at a very low rate. We speculate that that the continued growth in ILM thickness is in part due to a slow rate of ILM protein synthesis, but more importantly, due to an even slower decay of the existing BM constituents. The half-life of collagens in

J. Candiello et al. / Matrix Biology 29 (2010) 402–410

407

Fig. 5. Flat-mount preparations of the human ILM (A, B) and thickness (C, D) and elasticity measurements of ILMs using AFM (E, F). Adult human ILMs were immobilized onto glass slides and stained for collagen IV (A) or processed for SEM (B). Both ILM samples were folded showing the smooth vitreal surface (V), and the irregular retinal surface (R). The arrows in (A) indicate single-layered segments of the ILM, where AFM thickness and elasticity scans were done. Actual AFM measurements of a 44 and an 88-year-old ILM are shown in panel (C): the black and the red bars show the average thickness of each of the samples; the baseline is the glass slide on the left of the scan. The graph in (D) shows AFM thickness measurements of the human ILM according to age. AFM elasticity scans of a 44 and 88-year-old ILM are shown in panel (E). The graph in (F) shows the age-dependent increase in Young's modulus.

vitreous and cartilage ranges between 11 and 100 years (Maroudas et al., 1992; Verzijl et al., 2000; 2001; Bishop et al., 2003), and we expect a similar long half-life for proteins in BMs. We also speculate that the down-regulation of the ILM protein synthesis in early human life is essential, since a continuous high rate of synthesis of ILM proteins with a very slow degradation would eventually lead to a very thick extracellular matrix sheet that would interfere with the transparency of the eye. It is of note that soluble proteins in the vitreous, such as transferrin and a2 macroglobulin that are also synthesized by the ciliary body (Bertazolli-Filho et al., 2003; Kubota et al., 2004) are not downregulated in the adult eye, thus, it appears that only the slowly metabolized ECM proteins of the eye have a developmentally regulated synthesis. There are no data on the abundance of laminin and collagen IV in the ILM of the fetal human eye, but extrapolating from the chick embryonic ILM (Balasubramani et al., in press), we expect that laminins and nidogens are the dominant proteins and collagen IV is only a minor

component in the fetal human ILM. Since the adult human ILM consists of over 50% of collagen IV it appears that the human ILM undergoes an age-dependent compositional change from a laminin to a collagen IVdominated BM. Western blotting data showed that human ILMs from eyes of advanced age have a higher concentration of collagen IV, and a slightly lower concentration of laminin than ILMs of younger eyes. We propose that the shift in the composition of the ILM is due to the fact that collagens have a slower turnover than non-collagenous BM proteins. Our data are consistent with reports showing that vascular BMs from older human brain have a higher concentration of collagen IV as compared to vascular BMs from younger brains (Uspenskaia et al., 2004). 3.2. Age-dependent changes in the biomechanical qualities of the ILM We found that the thickness of adult human ILMs, as measured by AFM, was on average 4-times greater than measured by TEM. Sample

408

J. Candiello et al. / Matrix Biology 29 (2010) 402–410

Fig. 6. ILM thickness measured by AFM and TEM. Thickness measurements (A) of the hydrated ILM by AFM (red) resulted in values that were four times higher than thickness measurements based on TEM of the dehydrated samples (black). The AFM and TEM samples for each age group are from the same retina. AFM measurements (B) showed that when the GAG side chains of the BM proteoglycans were removed by heparitinase/chondroitinase, the ILMs shrank by 30 to 50% relative to untreated controls. Elasticity measurements (C) showed that the enzyme treatment led to a 20–30% increase in the Young's modulus. The reduction in ILM thickness and the increase in stiffness after heparitinase treatment were less at older age. The p-values between control and heparitinase-treated (H) ILM samples are indicated. The selectivity of the heparitinase treatment was confirmed by immunocytochemistry. Staining human control sections (Cont) with an antibody against the HS carbohydrate epitope (a-HS) labeled the ILM (red; D). When the sections had been incubated with the glycosidases to remove the heparan/chondroitin sulfate side chains (–HS), the HS-labeling was no longer detectable (E). When sections were stained for the agrin core protein, staining of the non-treated (F) and the enzyme-treated sections (G) was indistinguishable showing that the glycosidases did not degrade the core proteins of the proteoglycans. The selective removal of the HS chains was also confirmed by western blots (H). Control ILM sample (Cont) labeled for agrin (a-Ag) showed the typical smear of the agrin proteoglycan at 400 kD (star), whereas in the glycosidases-treated ILM sample (−HS) the agrin immunoreactivity had shifted to a narrow band at 200 kD (arrow head). When blots with the same samples were labeled with an antibody to laminin (a-Ln) the same 200 kD laminin bands appeared in control and enzyme-treated samples showing that the enzymes did not lead to proteolysis of the peptide chains. The sections in (D) to (G) were counterstained with Sytox Green.

preparation for TEM always requires dehydration, whereas the AFM probing of the ILM samples was performed in PBS. Dehydration of tissues leads to shrinkage between 10 and 20%. The over 50% difference of AFM and TEM thickness measurements indicated that BMs contain a much greater amount of water than other tissues. In cartilage, the GAG chains of the prominent proteoglycans are well known to bind large quantities of water and determine the elastic qualities of this extracellular matrix (Poole, 1997). Proteoglycans are, therefore, the prime candidates for the water-binding constituents in BMs as well. Indeed, when the GAG chains of ILM samples were enzymatically removed, the ILMs underwent a major compaction. The reduction in ILM thickness was accompanied by a several-fold increase in stiffness. The data are in perfect agreement with studies in embryonic chick ILM, where the removal of the GAG side chains led to a 50% reduction in ILM thickness and a four-fold increase in ILM stiffness (Balasubramani et al., in press). In the chick study, we also showed that the glycosidase treatment did not lead to a solubilization of ILM proteins and did not change the ultrastructure of the ILM. Based on these findings, we propose that the GAG side chains of proteoglycans are essential for the hydration status of BMs and determine to a large degree BM thickness and elasticity. The data also demonstrate that water is the most prominent component of BMs. The finding that the adult human ILM loses elasticity and that the shrinkage of the ILM after GAG removal is less at older age indicates that the water content of the older ILMs is lower than in younger ILMs. We assume this is, at least in part, due to shortened or a reduced number of GAG side chains in older BMs. It is conceivable that increased stiffness and a higher compaction of BMs at older age has physiological consequences, including a lower rate of diffusion and filtration through BMs, a loss of vascular elasticity that may contribute to chronic high blood pressure, and a higher incidence of

vascular herniation and rupture. We propose that the alteration of the BMs contributes to the normal aging process of the human body and contributes to the higher susceptibility to disease at old age. 4. Experimental procedures 4.1. Human eyes and antibodies Adult human eyes were obtained from CORE, the “Center of Organ Recovery and Education” of the University of Pittsburgh. Fetal human eyes were obtained from the Brain and Tissue Bank for Developmental Disorders of the University of Maryland (Baltimore, MD). The use of the human eyes was approved by the Institutional Review Board of the University of Pittsburgh under the IRB Protocol number #0312072. The total number of pairs of eyes that were used for the present study is listed in Table 1. the number of pairs of eyes used for TEM, AFM and Western blots is listed as well. Thirty two of the adult donors had died from cardiac failure, 9 from accidents, 6 from stroke, 3 from pneumonia, 2 from cancer, and one from a gun short wound, CO-poisoning, or

Table 1 Total number of pairs of eyes used for the present study listed according to age. The table also includes the number of eyes used for TEM, AFM and PAGE and western blotting experiments.

Total #61 TEM #21 AFM # 13 Western #6

Fetal

20 s

40 s

50 s

60 s

70 s

80 s

6 6 – –

1 1 – –

4 3 4 1

10 2 2 1

7 1 2 –

18 5 1 –

15 3 4 4

J. Candiello et al. / Matrix Biology 29 (2010) 402–410

infection. Twenty eight of the donors were male and 31 were female. None of the eyes used in this study were from patients with type I or type II diabetes. Polyclonal antisera to collagen IV, laminin-1 were obtained from Rockland Immunochemicals (Gilbertsville, PA), Invitrogen (Carlsbad, CA), and Santa Cruz Biotechnology (Santa Cruz, CA). A monoclonal IgM antibody to the heparan sulfate (HS) side chains of proteoglycans was purchased from Seikagaku (clone F58-10E4; Japan). The polyclonal rabbit anti-human agrin antiserum was raised a against a fusion protein from the N-terminal part of human agrin (Cotman et al., 2000). SytoxGreen (Molecular Probes, Eugene, OR) was used as a nuclear counter stain. 4.2. Histology For immunocytochemistry, fetal eyes or adult retinal samples were fixed in 4% paraformaldehyde and processed for cryostat-sectioning as described (Halfter, 1998). The sections were mounted in 90% glycerol and examined with an epifluorescence or a confocal microscope (Flowview, Olympus). For transmission electron microscopy, samples from the dorso-central retina (see Fig. 2E) were fixed in 2.5% glutaraldehyde and 2.5% paraformaldehyde overnight. The samples were osmicated and embedded in EPON according to standard procedures. Thin sections were examined by a JEOL electron microscope at 20,000×. TEM measurements of ILM thickness are based on the analysis of 21 pairs of human eyes. The adult donors had died from cardiac failure (6), stroke (4), pneumonia (2), gun shot wound (1), sepsis (1) and accident (1). Seven of the adult donors had been female, 8 were male. The eyes were enucleated from donors less than 30 h after expiration with expiration–enucleation intervals of approximately 10 h. Retinal samples were taken from both eyes, and two thin sections per retina sample were surveyed at 20,000×. At least twenty ILM measurements per retinal sample were taken at random, and the average value± standard deviation was calculated for the pairs of eyes. 4.3. ILM preparation ILMs were prepared by incubating segments of adult human retina overnight in 2% Triton-X-100 in water. The detergent-insoluble ILMs were transferred with a Pasteur pipette under a dissecting microscope and under dark-field illumination into new Triton-X-100 and 1% deoxycholate. This transfer was repeated 3-times. The washed ILMs are pelleted by centrifugation at 10,000 rpm. To facilitate the solubilization of the ILM proteins for SDS PAGE, the ILM pellets were incubated with 100 µL of each 0.1 µU/µl heparitinase and chondroitinase for 1 h (Seikagaku Corporation, Japan). 10× SDS sample buffer and solid urea are added to a final concentration of 1× and 8 M, respectively. The samples were boiled for 10 min, centrifuged and loaded onto 3.5–15% SDS gradient gels. Lens capsules were obtained by dissection. 4.4. Atomic force microscopy (AFM) of the ILM ILMs and lens capsule flat mounts for AFM probing were prepared in the following way: glass slides were coated with 10 µg/ml poly-lysine (P-1524; Sigma) and washed 3 times with PBS. The coated area of the slide was encircled with a PAP-Pen (Research Products International, Mt. Prospect, IL) to contain the droplet of PBS that holds the BM samples. Segments of ILM or lens capsule (see above) were transferred into the PBS droplet, excessive PBS was removed, and the slides were centrifuged at 1000 rpm for 5 min to firmly attach the ILMs to the poly-lysine-coated slides. The immobilized BMs were washed 3 times in PBS. Some of the preparations were immuno-stained for laminin or collagen IV for better visualization of the BMs. Some of the ILM samples were treated with 500 µl of 0.1 µU/µl of each heparitinase and chondroitinase (Seikagaku) and 1 mg/ml BSA in PBS for 2 h at 37 C in suspension to test for the importance of glycosaminoglycans (GAGs) for thickness and elasticity of the BMs. For controls, samples were incubated

409

for the same period of time with 1 mg/ml BSA in PBS. The samples were washed twice with PBS and mounted onto glass slides by centrifugation. The samples were washed again prior to AFM probing. The imaging and force-indentation experiments were done by an MFP-3D AFM (Asylum Research, Santa Barbara, CA) mounted on top of the Olympus IX-71 fluorescence microscope (Olympus, Tokyo, Japan). For all experiments, 100 μm long silicon-nitride triangular cantilevers with pyramidal tips were used (Veeco, Inc, Santa Barbara, CA) that have a nominal spring constant of ∼0.8 N/m. The spring constant was individually calculated for each cantilever prior to experimentation using the thermal fluctuation method (Hutter and Bechhoefer, 1993). The images of the ILM used for thickness measurements were done in intermittent contact mode (AC, tapping) and the elasticity was determined by the forceindentation method as previously described (Candiello et al., 2007). For the thickness and elasticity measurements, 13 ILM and 3 lens capsules were probed. For thickness characterization, 25 image cross-sections were measured at 4 different locations on each sample. Measurements of the elastic properties were taken at 100 random locations on the BM sample. For the hepartinase/chondroitinase experiments, 3 samples for enzyme-treated ILM and 3 samples for the control samples were probed. In thickness measurements 25 measurements were taken at 4 randomly selected locations. Elastic properties were again measured at 100 random locations. Average values with ±standard deviation were calculated. Statistical analysis was performed using Student's t-test. Acknowledgements We would like to thank Robert Johnson from the Brain and Tissue Bank for Developmental Disorders (University of Maryland, Baltimore, MD) for the shipping of the fetal eyes, and Janice Anderson and Lenny Coll from the Center for Organ Recovery and Education (CORE) of Pittsburgh for providing and delivering adult human eyes. The study was funded by a grant from the National Science Foundation (IBN# 0240774). We would also like to acknowledge Jonathan Sager, Mara Snell, Ming Sun and Simon Watkins for their help and support with the electron microscopy. Andrew Feola, University of Pittsburgh, Pittsburgh, PA, helped in developing computational method to measure elasticity. References A-Hassan, E., Heinz, W.F., Antonik, M.D., D'Costa, N.P., Nageswaran, S., Schoenenberger, C.A., Hoh, J.H., 1998. Relative microelastic mapping of living cells by atomic force microscopy. Biophys. J. 74, 1564–1578. Anderson, S., Brenner, B.M., 1986. Effects of aging on the renal glomerulus. Am. J. Med. 80, 435–444. Arikawa-Hirasawa, E., Watanabe, H., Takami, H., Hassell, J.R., Yamada, Y., 1999. Perlecan is essential for cartilage and cephalic development. Nat. Genet. 23, 354–358. Bader, B.L., Smyth, N., Nedbal, S., Miosge, N., Baranowsky, A., Mokkupat, S., Murshed, M., Nischt, R., 2005. Compound genetic ablation of nidogen 1 and 2 causes basement membrane defects and perinatal lethality in mice. Mol. Cell Biol. 25, 6846–6856. Balasubramani, M., Candiello, J., Schreiber, E., Balasubramani, S., Kurtz, J., Halfter, W. in press. Proteomic analysis of basement membranes; a role of proteoglycans for the elasticity of BMs. Matrix Biology. Begley, M.R., Mackin, T.J., 2004. Spherical indentation of freestanding circular thin films in the membrane regime. J. Mech. Phys. Solids 52, 2005–2023. Bertazolli-Filho, R., Laicine, E.M., Haddad, A., 2003. Synthesis and secretion of transferrin by isolated ciliary epithelium of rabbit. Biochem. Biophys. Res. Comm. 305, 820–825. Binnig, G., Quate, C.F., Gerber, C., 1986. Atomic force microscope. Phys. Rev. Lett. 56, 930–933. Bishop, P.N., Holmes, D.F., Kadler, K.E., McLeod, D., Bos, K.J., 2003. Age-related changes on the surface of vitreous collagen fibrils. Investig. Ophthalmol. Vis. Sci. 45, 1041–1046. Bloom, P.M., Hartmann, J.F., Vernier, R.L., 1959. An electron microscopic evaluation of the width of normal glomerular basement membrane in man at various ages. Anat. Rec. 133, 251–263. Candiello, J., Balasubramani, M., Schreiber, E.M., Cole, G.J., Mayer, U., Halfter, W., Lin, H., 2007. Biomechanical properties of native basement membranes. FEBS J. 274, 2897–2908. Costell, M., Gustafsson, E., Aszodi, A., Moergelin, M., Bloch, W., Hunziger, E., Addicks, K., Timpl, R., Faessler, R., 1999. Perlecan maintains the integrity of cartilage and some basement membranes. J. Cell Biol. 147, 1109–1122. Cotman, S., Halfter, W., Cole, G.J., 2000. Agrin binds to amyloid (Aβ), accelerates Aβ fibril formation, and is localized to Aβ deposits in Alzheimer's disease brain. Mol. Cell Neurosci. 15, 183–198.

410

J. Candiello et al. / Matrix Biology 29 (2010) 402–410

Danysh, B.P., Czymmek, K.J., Olurin, P.T., Sivak, J.G., Duncan, M.K., 2008. Contributions of mouse genetic background and age on anterior lens capsule thickness. Anat. Rec. 291, 1619–1627. Dong, L.-J., Chung, A.E., 1991. The expression of the gene for entactin, laminin A, laminin B1 and laminin B2 in murine lens morphogenesis and eye development. Differentiation 48, 157–172. Dong, S., Landfair, J., Balasubramani, M., Bier, M.E., Cole, G.J., Halfter, W., 2002. Expression of basal lamina protein mRNA in the early embryonic chick eye. J. Comp. Neurol. 447, 261–273. Duhamel, R.C., Meezan, E., Brendel, K., 1983. Selective solubilization of two populations of polypeptides from bovine retinal basement membranes. Exp. Eye Res. 36, 257–267. Erickson, A.C., Couchman, J.R., 2000. Still more complexity in mammalian basement membranes. J. Histochem. Cytochem. 48, 1291–1306. Fisher, R.F., 1969. Elastic constants of the human lens capsule. J. Physiol. 201, 1–19. Foos, R.Y., 1972. Vitreoretinal juncture; topographic variations. Investig. Ophthalmol. 11, 801–810. Fukai, N., Eklund, L., Marneros, A.G., Oh, S.P., Keene, D.R., Tamarkin, L., Niemela, M., Ilves, M., Li, E., Pihlajaniemi, T., Olsen, B.R., 2002. Lack of collagen XVIII/endostatin results in eye abnormalities. EMBO J. 21, 1535–1544. Gartner, J., 1970. Electron microscopic observations on the cilio-zonular border area of the human eye with particular reference of the aging changes. Z. Anat. Entwicklungsgesch. 131, 263–273. Gautam, M., Noakes, P.G., Moscovo, L., Rupp, F., Scheller, R.H., Merlie, J.P., Sanes, J.R., 1996. Defective neuromuscular synaptogenesis in agrin-deficient mutant mice. Cell 85, 525–535. Georges-Labouesse, E., 1998. Essential role of α6 integrins in cortical and retinal lamination. Curr. Biol. 8, 983–986. Gould, D.B., Phalan, F.C., Breedveld, G.J., van Mil, S.E., Smith, R.S., Schimenti, J.C., Aguglia, U., Van der Knaap, M.S., Heutink, P., John, S.W., 2005. Mutations in Col4a1 cause perinatal cerebral hemorrhage and porencephaly. Science 308, 1167–1171. Gould, D.B., Marchant, J.K., Savinova, O.V., Smith, R.S., John, S.W., 2007. Col4a1 mutation causes endoplasmic reticulum stress and genetically modified ocular dysgenesis. Hum. Mol. Genet. 16, 798-80. Halfter, W., 1998. Disruption of the retinal basal lamina early in embryonic development leads to a retraction of vitreal endfeet, and increased number of ganglion cells and aberrant axonal growth. J. Comp. Neurol. 397, 89–104. Halfter, W., von Boxberg, Y., 1992. Axonal growth on solubilized and reconstituted matrix from the embryonic chicken retina inner limiting membrane. Eur. J. Neurosci. 4, 840–852. Halfter, W., Dong, S., Yip, Y.P., Willem, M., Mayer, U., 2002. A critical role of the pial basement membrane in cortical histogenesis. J. Neurosci. 22, 6029–6040. Halfter, W., Willem, M., Mayer, U., 2005a. Basement membrane-dependent survival of retinal ganglion cells. Investig. Ophtalmol. Vis. Sci. 46, 2202–2209. Halfter, W., Dong, S., Schurer, B., Ring, C., Cole, G.J., Eller, A.W., 2005b. Embryonic synthesis of the inner limiting membrane and vitreous body. Investig. Ophtalmol. Vis. Sci. 46, 2202–2220. Halfter, W., Dong, S., Dong, A., Eller, A.W., Nischt, R., 2008. Origin and turnover of ECM proteins from the inner limiting membrane and vitreous body. Eye 22, 1207–1213. Hutter, J.L., Bechhoefer, J., 1993. Calibration of atomic force microscope tips. J. Rev. Sci. Instrum. 64, 1868–1873. Krag, S., Andreassen, T.T., 2003. Mechanical properties of the human lens capsule. Prog. Retin. Eye Res. 22, 749–767. Kubota, R., McGuire, C., Dierks, B., Reh, T.A., 2004. Identification of ciliary epithelialspecific genes using subtractive libraries and cDNA arrays in the avian eye. Dev. Dyn. 229, 529–540. Laney, D.E., Garcia, R.A., Parson, S.M., Hansma, H.G., 1997. Changes in the elastic properties of cholinergic synaptic vesicles as measure by atomic force microscopy. Biophys. J. 72, 806–813. Lee, J., Gross, J.M., 2007. Laminin beta1 and gamma1 containing laminins are essential for basement membrane integrity in zebrafish eye. Investig. Ophtalmol. Vis. Sci. 48, 2483–2490.

Libby, R.T., Champliaud, M.-F., Claudepierre, T., Xu, Y., Gibbons, E.P., Koch, M., Burgeson, R.E., Hunter, D.D., Brunken, W.J., 2000. Laminin expression in adult and developing retinae: evidence of two novel CNS laminins. J. Neurosci. 20, 6517–6528. Maroudas, A., Palla, G., Gilav, E., 1992. Racemization of aspartic acid in human articular cartilage. Connect. Tissue Res. 28, 161–169. Matsumoto, B., Blanks, J.C., Ryan, S.J., 1982. Topographic variations in the rabbit and primate internal limiting membrane. Investig. Ophtalmol. Visc. Sci. 25, 71–82. Meezan, E., Hjelle, J.T., Brendel, K., Carlson, E.C., 1975. A simple, versatile non-disruptive method for the isolation of morphologically and chemically pure basement membranes. Life Sci. 17, 1721–1732. Poeschl, E., Schlotzer-Schrehardt, U., Brachvogel, B., Saito, K., Ninomiya, Y., Mayer, U., 2004. Collagen IV is essential for basement membrane stability but dispensable for initiation of its assembly during early development. Development 131, 1619–1628. Poole, C.A., 1997. Articular cartilage chondron: form, function, and failure. J. Anat. 191, 1–13. Quist, A.P., Rhee, S.K., Lin, H., Lai, R., 2000. Physiological role of gap-junctional hemichannels. Extracellullar calcium-dependent isosmotic volume regulation. J. Cell Biol. 148, 1063–1074. Richardson, L.L., Kleinman, H.K., Dym, M., 1995. The effect of aging on basement membrane in the testis. J. Andol. 16, 118–126. Sarthy, V., 1993. Collagen IV mRNA expression during development of the mouse retina: an in situ hybridization study. Investig. Ophtalmol. Vis. Sci. 34, 145–152. Sarthy, P.V., Fu, M., 1990. Localization of laminin B1 mRNA in retinal ganglion cells by in situ hybridization. J. Cell Biol. 110, 2099–2108. Satz, J.S., Barresi, R., Durbeej, S., Willer, T., Turner, A., Moore, S.A., Campbell, K.P., 2008. Brain and eye malformations resembling Walker–Warburg syndrome are recapitulated in mice with dystroglycan deletion in the epiblast. J. Neurosci. 28, 10567–10575. Semina, E.V., Bosenko, D.V., Zinkevich, N.C., Soules, K.E., Hyde, D.R., Vihtelic, T.S., Willer, G.B., Gregg, R.G., Link, B.A., 2006. Mutations in laminin alpha 1 result in complex, lens-independent ocular phenotypes in zebrafish. Dev. Biol. 299, 63–77. Serot, J.M., Foliguet, B., Bebe, M.C., Faure, G.C., 2001. Choroid plexus and aging in rats: a morphometric and ultrastructural study. Eur. J. Neurosci. 14, 794–798. Sertie, A.L., Sossi, V., Camargo, A.A., Zatz, M., Brabe, C., Passos-Bueo, M.R., 2000. Collagen XVIII, containing an endogenous inhibitor of angiogenesis and tumor growth, plays a critical role in the maintenance of retina structure and in neural tube closure. Hum. Mol. Genet. 9, 2051–2058. Smyth, N., Vatansever, N.S., Murray, P., Meyer, M., Frie, C., Paulsson, M., Edgar, D., 1999. Absence of basement membranes after targeting the LANC1 gene results in embryonic lethality due to failure of endoderm differentiation. J. Cell Biol. 144, 151–160. Thomopoulos, G.N., Spicer, S.S., Gratton, M.A., Schulte, B.A., 1997. Age-related thickening of basement membrane in stria vascularis capillaries. Hear. Res. 111, 31–41. Timpl, R., Brown, J.C., 1996. Supramolecular assembly of basement membranes. Bioassays 18, 123–132. Uspenskaia, O., Liebetrau, M., Herms, J., Danek, A., Hamann, G.F., 2004. Aging is associated with increased collagen IV accumulation in the basal lamina of human cerebral microvessels. BMC Neurosci. 5, 37–48. Verzijl, N., DeGroot, J., Thorpe, S.R., Bank, R.A., Shaw, J.N., Lyons, T.J., Bijlsma, J.W., Bijlsma, J.W.J., Lafeber, F.J.G., Bayne, J.W., TeKoepele, J.M., 2000. Effect of collagen turnover on the accumulation of advanced glycation end products. J. Biol. Chem. 275, 39027–39031. Verzijl, N., DeGroot, J., Bank, R.A., Shaw, J.N., Lyons, T.J., Bijlsma, J.W.J., Lafeber, F.J.G., Baynes, J.W., TeKoepele, J.M., 2001. Age-related accumulation of the advanced glycation endproduct of pentosie in human articlar cartilage aggrecan: the use of pentosine levels as a quantitative measure of protein turnover. Matrix Biol. 20, 409–417. Willem, M., Miosge, N., Halfter, W., Smyth, N., Janetti, I., Burghart, E., Timpl, R., Mayer, U., 2001. Specific ablation of the nidogen binding site in the laminin γ1 chain interferes with kidney and lung development. Development 129, 2711–2722. Xi, Y., Nette, E.G., King, D.W., Rosen, M., 1982. Age-related changes in normal m basement membranes. Mech. Ageing Dev. 19, 315–334. Zenker, M., Aigner, T., Windler, O., Tralau, T., Muntefering, H., Fenski, R., Pitz, S., Schumacher, V., Royer-Pokora, B., Wuhl, E., 2004. Human laminin beta2 deficiency causes congenital nephrosis, mesangial sclerosis, and distinct eye abnormalities with microcoria: an autosomal recessive syndrome. Am. J. Genet. A130, 138–145.