Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The collagen-elastin-proteoglycan matrix is the key constituent of lung parenchyma and plays a major role in the mechanical behavior of lung tissues (1). Although the collagen and elastin components of the lung extracellular matrix have been widely studied, the exact composition and distribution of the proteoglycan (PG) components within the matrix of lungs have not yet been fully determined. PGs are macromolecules composed of a protein core and glycosaminoglycans side chains that have a number of known functions. PGs regulate the water balance of the extracellular matrix; influence tissue biomechanics; facilitate cellular adhesion, proliferation, and migration; and modulate growth factor and cytokine activities (2).

Lumican is a keratan sulfate-PG that belongs to the family of relatively small, leucine-rich repeat (LRR)-PGs. The other well-characterized members of this PG family include biglycan, decorin, and fibromodulin. These small PGs possess similar core proteins characterized by the presence of a central region possessing ten adjacent LRRs, which are flanked by N-terminal and C-terminal disulfide-bonded regions (6). In the case of decorin and biglycan, chondroitin sulfate or dermatan sulfate chains are attached to the extreme N-terminal region of the core protein (8), whereas for fibromodulin and lumican, keratan sulfate may be attached to the central LRR region. Lumican was first described in the cornea but is present in the extracellular matrix of many tissues, such as cartilage, aorta, liver, skin, muscle, and intestine (9). Details on the organization and chromosomal location of the lumican gene have recently been published (13). All members of the small-PG family, including lumican, interact with fibrillar collagen and may influence the interaction of the collagen fibrils with other components of the extracellular matrix, thus participating in the maintenance of the extracellular milieu (3, 14).

In the present study we report the expression of lumican in adult human lungs and its distribution in peripheral lung tissue.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Tissue Extracts and Enzyme Digestion

Peripheral lung tissue was obtained from seven patients (age range: 54 to 73 yr) undergoing therapeutic lung resections for central lung tumors (the size of samples ranged from 0.3 to 0.5 g). The tissue was frozen in 10 mM acetate buffer, pH 6.0, and cut into 20-µm sections using a cryostat (18). Samples were extracted with 10 vol 4 M guanidinium chloride and 100 mM sodium acetate (pH 6.0) containing proteinase inhibitors for 48 h at 4°C (19) and then dialyzed into 10 mM sodium acetate and 10 mM Tris/HCl, pH 7.3, overnight. To estimate the amount of PG extracted, the dimethylmethylene blue assay (20) was used to measure the sulfated glycosaminoglycan content of separate samples from the same patient. The sulfated glycosaminoglycan content in samples extracted with guanidinium was compared with that obtained in samples with papain at 60°C (20). The dialyzed extracts were then incubated with 0.1 unit/ml chondroitinase ABC (Sigma, Oakville, ON, Canada) for 4 h at 40°C. One sample was dialysed against 10 mM sodium acetate, pH 6.0, and treated with endo--galactosidase (ICN, Montreal, PQ, Canada) or keratanase II (Seikagaku Kogyo, Tokyo, Japan) at 0.01 unit/100 µl of extract overnight at 37°C.

Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis and Immunoblotting

Tissue extracts were then analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting using polyclonal antipeptide antisera to human lumican and fibromodulin (13, 21) and polyclonal antipeptide IgG preparations specific to human biglycan and decorin (22, 23). Specifically, the samples (20 µl) were analyzed by SDS-PAGE on 10% polyacrylamide slab gels, as previously described (22). After electrophoresis, the fractionated proteins were electrophoretically transferred to nitrocellulose membranes (24). Electroblotting was performed in 20% (vol/vol) methanol/25 mM Tris/190 mM glycine, pH 8.3, at 100 V for 4 h. The transfer membrane was incubated in blocking solution overnight, then incubated with a 1:50 dilution of anti-lumican or anti-fibromodulin serum, or with a 1:2,500 dilution of anti-decorin or anti-biglycan IgG, in TBST (10 mM Tris/HCl, pH 8.0; 150 mM NaCl; and 0.05% Tween 20) for 30 min at room temperature. After washing with TBST, the nitrocellulose membranes were incubated for 30 min with a 1:7,500 dilution of an alkaline phosphatase-conjugated goat antirabbit second antibody (Promega, Madison, WI) in TBST buffer. The nitrocellulose was then washed in TBST buffer and in one change of alkaline phosphatase buffer, before being incubated for 5 min in alkaline phosphatase substrate solution at room temperature. Control experiments were performed using extracts of adult human articular cartilage, which is known to be a source of all members of the LRR-PG family.

Immunohistochemistry

Immunohistochemical staining was performed in subpleural lung strips (10.0 × 2.0 × 2.0 mm) obtained from 15 patients undergoing therapeutic lung resections. The tissue was fixed with 4% paraformaldehyde, embedded in paraffin, and cut into 5-µm-thick sections. Sections were deparaffinized, hydrated, and incubated in 2% normal human serum (NHS) for 1 h at room temperature. Sections were then rinsed with TBS (0.5 M Tris, pH 7.6, and 1.5 M NaCl) and incubated with anti-lumican antiserum (1:400 in TBS) overnight at 4°C. After washing with TBS, the tissue was incubated with a biotin-labeled swine antirabbit IgG (Dako, Mississauga, ON, Canada) (1:30 in 20% NHS) for 1 h, washed again, and incubated with alkaline phosphatase-conjugated avidin (Dako) (1:30 in 20% NHS) for 1 h. After further washing, sections were developed with Fast Red salt (Sigma) (1 mg/ml in alkaline phosphatase substrate) for 10 min at room temperature. Sections were counterstained with Harris Haematoxylin for 1 min. Negative controls were made by substitution of the primary antibody with normal rabbit serum (1:400 in TBS).

Morphometric Study

A semiquantitative analysis was performed on the slides stained for lumican by applying point-counting. Using a 121-point grid, we calculated the volume proportion of lumican in airways, vessels, and parenchyma as the relation between the number of points falling on lumican-stained and nonstained tissue. Measurements were performed in 20 fields per slide, using a magnification of ×400. Positive staining for lumican was established by comparing lumican-stained slides with controls in order to determine a color threshold. Lumican stained bright red, whereas negative controls stained a faint red or pink color.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

The dimethylmethylene blue assay estimated that at least 80% of the tissue glycosaminoglycan was extracted from 20-µm lung sections by 4 M guanidinium chloride. These lung extracts were then analyzed for LRR-PG content.

In all samples analyzed with SDS-PAGE and immunoblotting, a single major component was detected for lumican (Figure 1). The molecular weight of this component varied from 65 to 90 kD, with different individuals showing some heterogeneity in the molecular weight range of the lumican present in their lung samples. Relatively smaller amounts of lower molecular-weight components were also observed in the lung extracts, some of which could be indicative of proteolytic processing occurring within the lumican core protein. The major lumican component had a mobility that was dependent upon endogalactosidase or keratanase treatment, and therefore represented a PG with attached keratan sulfate chains (Figure 2). All of the lumican in this component was present in a PG form because prior to glycosidase treatment there was no material present in the region where the core protein elutes. The molecular weight of the lumican core protein after endogalactosidase or keratanase treatment was about 57 kd. This was about the same size as the largest of the minor components observed in some lung extracts (Figure 1), and this component could therefore represent a glycoprotein form of lumican devoid of keratan sulfate chains.

View larger version (88K): [in this window] [in a new window]   Figure 1.   Western blot analysis of lumican in extracts of human lungs. In all samples (lanes 1-6) a broad component within the molecular weight range of 65 to 90 kD was detected. Blots were incubated with anti-lumican antiserum, and subsequent color development was for 5 min.

View larger version (99K): [in this window] [in a new window]   Figure 2.   Western blot analysis of lumican in human lung extracts following treatment with keratanase II or endo--galactosidase. Samples were analyzed following incubation with endo--galactosidase (lane 1), incubation with keratanase II (lane 2), no incubation (lane 3), or incubation in buffer without glycosidase (lane 4). The digested samples had a mobility dependent upon glycosidase treatment. Blots were incubated with anti-lumican antiserum, and subsequent color development was for 5 min.

Whereas lumican was readily observed in the lung extracts, the other LRR-PGs were not detectable under the same conditions. A small amount of biglycan could be observed when color development was extended from 3 to 50 min, but even these conditions failed to reveal decorin or fibromodulin (Figure 3). In contrast, all four of the LRR-PG were readily detected in the extracts of adult human articular cartilage with only 3 min of color development. It was also apparent that the lumican in the lung was less heterogeneous in size than that in the articular cartilage. This was due mainly to the absence of lower molecular-weight components that constitute the glycoprotein form of lumican lacking keratan sulfate chains. Such components are the predominant form in adult cartilage and are of identical size to the core protein of the lung lumican that is generated upon treatment with endogalactosidase or keratanase.

View larger version (53K): [in this window] [in a new window]   Figure 3.   Western blot analysis of chondroitinase-treated extracts from human articular cartilage (lanes A) and lung (lanes B). Blots were incubated with antibodies against lumican (1), biglycan (2), fibromodulin (3), and decorin (4). Subsequent color development was for 3 min for lung exposed to antibodies against lumican; but in the case of lung exposed to the antibodies against decorin, biglycan, and fibromodulin, color development was for 50 min.

Immunohistochemical staining revealed that lumican was found mainly in the extracellular matrix of vessel walls, but it was also observed in airway walls and in alveolar septa (Figure 4). In small vessels, lumican was observed uniformly beneath the endothelial layer (Figure 4A). It was found in patchy distribution in alveolar walls (Figure 4B). We also observed small amounts of lumican in airway walls, localized either in the basement membrane region or within the submucosal layer (Figure 4C). Table 1 shows the volume proportion of lumican in the different anatomical components of peripheral lung tissue. Lumican represented 9% of the whole tissue, including 6.6% of parenchymal tissue, 18.8% of the vessel wall, and 4.4% of the airway wall. The volume proportions of each anatomic element as a percentage of the whole tissue were: parenchyma, 78 ± 2%; vessel, 17 ± 1%; and airway, 5 ± 1%. We counted 2,420 points per slide, which resulted in a mean error of 11%.

View larger version (72K): [in this window] [in a new window]   Figure 4.   Immunolocalization of lumican in peripheral human lung tissue. Immunohistochemistry shows positive staining (red) for lumican in blood vessel (A), alveolar wall (B), and small airway (C). (D) Negative control. (Magnification: ×400.) V = blood vessel; AW = airway wall; L = lumen.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The LRR-PGs have relatively small molecular sizes with core proteins of approximately 40 kD (25). They share similar core protein structures but differ in their glycosaminoglycan content and in their distribution in tissues. Decorin and lumican are found in many extracellular matrices (23, 13), while fibromodulin and biglycan show a more limited distribution in connective tissues (21, 23, 26). Histologic examination of many connective tissues has revealed that decorin is widely present in tissues rich in collagen types I, II, and III, and resides at the surface of collagen fibrils. Since all LRR-PGs can bind to specific regions of collagen fibrils through their core proteins, and delay the formation of collagen fibrils in vitro, it has been proposed that they function as regulators of collagen fibril formation (3, 14, 25).

In lungs, relatively little has been reported about the PG content of the extracellular matrix. In an ultrastructural analysis of developing lungs of Macaca nemestrina, Juul and colleagues (27) observed that the immature alveolar interstitium was rich in large chondroitin sulfate-PGs (CSPG). As maturation proceeded, the amount of large CSPG declined and was replaced by dermatan sulfate-PGs (DSPG). Fetal alveolar basement membrane contained heparan sulfate-PGs (HSPG). Juul and coworkers have also shown that large CSPG are upregulated in lungs of neonatal rats exposed to hyperoxia (28). Veness-Meehan and associates (29) reported the presence of biglycan in normal rat alveolar walls which is increased after exposure to chronic hyperoxia, whereas Westergren-Thorsson and colleagues (30) reported the presence of both biglycan and decorin in adult rat lungs. Finally, Sannes and coworkers (31) reported immunoreactivity for CSPG and HSPG in alveolar, vascular, and airway basement membranes in developing and adult rat lungs.

Less information is available concerning human lung. Van Kuppevelt and colleagues have reported that HSPG are present in basement membranes and DSPG are associated with collagen fibrils (32). Bianco and coworkers have described biglycan and decorin in the "small" and "large" interstitium of developing human lungs of 14 to 17 wk gestational age (33). More recently, Bensadoun and coworkers (34) have performed immunohistochemistry on human tissue obtained at the time of open lung biopsy or autopsy in patients with adult respiratory distress syndrome, bronchiolitis obliterans organizing pneumonia, and idiopathic pulmonary fibrosis. Patients undergoing lung resection for tumor served as control subjects. In normal lung, nonspecific staining for PGs (with alcian blue) was positive only in bronchial cartilage and the media of some pulmonary arteries. Specific immunostaining for versican, decorin, biglycan, and hyaluronan revealed these molecules to be present in blood vessel walls and the subepithelial layer of airways. Alveolar walls showed only a trace amount of staining for these PGs. With fibrotic disease, versican deposition increased markedly.

The present study was performed to further characterize the PG matrix composition in mature peripheral human lungs. We used antipeptide antisera specific to human lumican and fibromodulin and anti-peptide IgG specific to human biglycan and decorin that have previously been shown to interact with these PGs in human cartilage samples (13, 21). PGs were extracted from lungs using the same protocol as has been employed with cartilage (22), which is known to disrupt the noncovalent interactions that retain these molecules in the tissue. The systematic positive results for lumican in all samples analyzed with SDS-PAGE and immunoblotting indicate that lumican is a major important component of the PG-based extracellular matrix in adult human lungs, an observation not previously reported in the literature. Of note, Juul and associates (28) reported upregulation of a keratanase-sensitive PG in neonatal rat pups subjected to hyperoxia challenge.

Decorin and fibromodulin were not detected in the human lung extracts, and biglycan was minimally detected. This result suggests that these PGs are either not extractable with the methods applied, or are relatively minor components in mature human lung peripheral tissue. It is possible that proteolytic processing in the region of the C-terminal peptide (the region containing the epitope the antibody detects) occurred during extraction. However, in all other tissues we have examined previously, there has been no evidence of proteolytic processing in this region. Moreover, the data from Bensadoun and coworkers (34) describing a lack of these molecules in the alveolar wall is consistent with our observation. The predominance of CSPG and DSPG in other species (27, 28, 30, 32) may be due to interspecies or age-related differences in the type of PGs in the lung or, alternately, may reflect the fact that in these other studies, more central airways and vessels were included in the analysis. In the current study, we sampled peripheral adult lung that contained only a modest amount of vessels and airways. In adult human lung periphery our results indicate that lumican is the major LRR-PG present.

The ability of specific PGs to interact with collagen and regulate the interaction of the collagen fibrils with one another and with other components of the extracellular matrix argues for the importance of these macromolecules in the maintenance of the extracellular milieu in physiologic and pathologic conditions (3, 14). Immunohistochemistry was performed in order to localize lumican in the lung. The results show that the distribution of lumican was not uniform, even though it was present in all structural components of lung tissue. Lumican was found mainly in peripheral blood vessels, in patchy distribution in the alveolar interstitium, and in small amounts in peripheral airway walls. Since collagen types I and III are present in these areas, it is reasonable to postulate that lumican interacts with the collagen fibrils and may play a role in regulating their structure in the lung extracellular matrix.

The ubiquitous presence of lumican in lung tissue suggests that it may have an important role in lung extracellular matrix function, though further studies are necessary to determine the precise role of this PG in lung function.