We have examined whether changes in versican levels, or in the sulfation pattern of its chondroitin sulfate (CS) side chains, are associated with the reduction in perialveolar tissue volumes that characterize lung maturation in late-gestation fetal sheep. Lung tissue was collected from fetuses [90–142 days gestational age (GA)] and lambs (2 wk after term birth). The level and distribution of versican and CS glycosaminoglycans (GAG) were determined using immunohistochemistry, whereas fluorophore-assisted carbohydrate electrophoresis was used to determine changes in CS sulfation patterns. Versican was the predominant CS-containing proteoglycan in the lung and decreased from 19.9 ± 2.7 arbitrary units at 90 days GA to 6.0 ± 0.5 arbitrary units at 142 days GA, in close association (P < 0.05) with the reduction in tissue volumes (from 66.0 ± 4.6 to 25.3 ± 1.5% at 142 days); similar reductions occurred for both chondroitin-6-sulfate and chondroitin-4-sulfate CS side chains. Hyaluronic acid levels decreased from 3,168 ± 641 pmol/μg GAG at 90 days GA to 126 ± 9 pmol/μg GAG at 142 days GA, and the predominant sulfated disaccharide changed from Δ-di-6S at 90 days GA to Δ-di-4S at term. These data indicate that structural development of the lung is closely associated with marked changes in versican levels and the microstructure of CS side chains in perisaccular/alveolar lung tissue.
- lung maturation
- versican mRNA
to successfully assume the role of respiratory gas exchange, the lung must be mechanically compliant and have a large surface area for gas exchange, which is closely apposed to a large capillary network by the time of birth. The development of these features is closely associated with marked structural remodeling of the perialveolar parenchyma late in gestation in precocial species such as humans and sheep. Tissue distances between terminal airways decrease from ∼50 μm at ∼90 days of gestation (1, 23) to ∼6 μm at ∼130 days of gestation ((21) term is ∼147 days) in fetal sheep. As a result, the percentage of space occupied by lung tissue decreases from ∼80% at 90 days to ∼20% at 130 days of gestation (1). These structural changes greatly contribute to increased lung tissue compliance and to the reduction in the alveolar/capillary tissue barrier (air/blood-gas barrier) and can be induced by antenatal corticosteroid administration (28, 29). Although the process of interairway tissue thinning must principally include remodeling of the extracellular matrix (ECM) within the perialveolar/saccular region, the mechanisms and the specific matrix components involved are largely unknown.
Proteoglycans (PGs) are an important component of the ECM (17, 30) and are likely candidates for the regulation of lung tissue volumes in late gestation. They consist of a protein core with one or more covalently bound glycosaminoglycan (GAG) side chains, which are linear polymers of repeating disaccharides that carry a number of sulfate and/or carboxyl groups that contribute to the polyanionic charge density of the molecule (13). Variations in the degree and position of the sulfate groups on each of the disaccharides that form the GAG side chains confer immense structural diversity to these molecules (13). Furthermore, some PG protein cores are expressed as alternatively spliced variants, which alter the sizes of the GAG attachment domains, thereby contributing further to structural and functional diversity. A major PG in lung matrix is the large chondroitin sulfate (CS)-rich species, versican (8). Versican is a member of the “hyalectan” family of PGs that share highly homologous NH2- and COOH-terminal globular domains, but have distinctive CS-bearing middle portions. Large PGs such as versican are thought to serve important physiological roles in tissue by controlling interstitial tissue hydration and solute permeability and influencing viscoelastic properties of tissue; they are likely, therefore, to have a major influence on lung tissue mechanics. The high anionic charge density of GAG side chains attracts mobile counterions to maintain electroneutrality and, in turn, generate the osmotic swelling pressure that provides tissue resilience and influences tissue mechanics (7). PGs are also thought to act as a reservoir for growth factor binding and regulate cell migration, as well as cell-cell and cell-matrix binding. Recent evidence indicates that CS-PGs play a critical role in the early growth and branching morphogenesis of the lung (26).
Our aim was to investigate changes in versican and the sulfation pattern of CS chains on versican, and other CS-PGs, in the lung over the last one-third of gestation in fetal sheep. This gestational period coincides with the exponential-like reduction in parenchymal tissue volumes, relative to total lung volume, in fetal sheep (1). We hypothesized that this reduction in perialveolar tissue volumes is caused by a reduction in versican content, changes in the versican core protein structure, or changes in the pattern or degree of sulfation on CS side chains, leading to a reduction in charge density and water-retaining capacity of the tissue. To test this hypothesis, we have used immunohistochemistry to examine lung tissue from fetal sheep between 90 and 142 days of gestation. Antibodies that recognize versican, or CS chains sulfated at the carbon-4 or carbon-6 position on hexosamine, were used to demonstrate changes in tissue distribution of these components. Fluorophore-assisted carbohydrate electrophoresis (FACE) analysis was used to identify and measure changes in CS sulfation within lung tissue over this gestational period. In addition, we have used ovine-specific cDNA probes to determine whether there is a development-related change in the expression pattern of the alternatively spliced variants of versican in late gestation.
Lung tissue was collected from fetal sheep at 90 days (n = 4), 111 days (n = 4), 126 days (n = 5), 138 days (n = 5), and 142 days (n = 4) of gestational age (GA) and from lambs at 2 wk postnatal age (PA; n = 4); term is 147 days of gestation. Ewes, fetuses, and lambs were all painlessly killed by an intravenous injection of pentobarbital sodium and were weighed before their lungs were removed and weighed. The left main bronchus was ligated, the left lung removed, and portions of the left lung were snap frozen for biochemical and histological analysis. The right lungs of all fetuses (but not postnatal lambs) were fixed at 20 cmH2O via the lung lumen, using 4% paraformaldehyde, to measure percentage of air space and tissue space volumes. The lungs collected from lambs at 2 wk of PA were not fixed at a constant airway pressure and, therefore, were not used for the stereological studies. All ewes, fetuses, and lambs used in this study were either sham operated or unoperated controls that were not subjected to experimental manipulation. The Monash University Committee for Ethics in Animal Experimentation approved all procedures on animals.
The primary antibody used to detect the versican core protein was a rabbit polyclonal antibody, which was a kind gift from Dick Heinegard, Lund, Sweden. The primary antibodies used to detect chondroitin-4-sulfate (C-4-S) and chondroitin-6-sulfate (C-6-S) were the mouse monoclonal antibodies 2030 and 2035, respectively (Chemicon International). All immunohistochemical analysis of versican and the sulfation pattern of CS side chains (C-4-S and C-6-S) were performed on frozen lung sections (10 μm) chosen at random from both upper and lower lobes of the left lung. The sections were fixed for 20 min in 4% paraformaldehyde, washed in phosphate-buffered saline (PBS) for 5 min (×3), then incubated with 0.02 units chondroitinase avidin-biotin complex [in 0.1 M Tris·HCl, pH 8, 30 m sodium-acetate, 1 mM EDTA; containing trans-epoxysuccinyl-L-leucylamido-(4-guanidino)butane, 4-(2-aminoethyl)-benzenesulfonylfluoride, and pepstatin protease inhibitors; Seikagaku, Japan] for 60 min at 37°C. Sections were then washed in PBS (5 min, ×3) and incubated in blocking/permeabilization buffer (2% normal goat serum; 0.1% Triton X-100 in PBS) in a humidity chamber (30 min at room temperature). Sections were incubated with the versican primary antibody (1:700 dilution in 1% BSA/PBS) overnight at 4°C, washed in PBS/0.1% Tween 20 for 5 min (×3), and then incubated with a goat anti-rabbit (IgG) secondary antibody conjugated with Alexa Fluor 488 (Invitrogen; at 1:700 diluted in 1% BSA/PBS) for 1 h at room temperature. The sections were again washed in PBS/0.1% Tween 20 for 5 min (×3), and the process was repeated with either the C-4-S (1:15,000 dilution in 1% BSA/PBS) or the C-6-S (1:700 dilution in 1% BSA/PBS) primary antibody, followed by a goat anti-mouse secondary antibody conjugated with Alexa Fluor 594 (Invitrogen; 1:800 dilution in 1% BSA/PBS).
Tissue sections were viewed under a fluorescent microscope (Nikon, Japan) using the excitation and emission wavelengths specified. Multiple fields of view (>3) from at least four sections obtained from different regions of the lungs were analyzed; care was taken to avoid fields of view that included major airways and blood vessels. Digital images were captured and stored electronically, and the mean density of staining was quantified using ImagePro plus image analysis software (Media Cybernetics).
Tissue/air space volume.
To determine the percentage of tissue and air space volumes for each fetus at each GA, the fixed right lung was separated into the upper, middle, and lower lobes, and each lobe was accurately cut into 5-mm slices. Every second slice from each lobe was further subdivided into three sections. Six slices were then chosen at random from each lobe, and the tissue was cut into ∼1-cm × 1-cm sections (×5 mm thick) and embedded into paraffin blocks. Blocks were selected at random, and tissue sections were cut at 5 μm, stained with hematoxylin and eosin, and viewed under the light microscope to calculate percent air space and tissue space volumes using a point-counting technique, as previously described (21); at least five sections and two fields of view per section were examined. Lung tissues collected from lambs used in this analysis were not fixed at a constant airway pressure, and, therefore, no measurements of air space or tissue space volumes were made.
Versican gene expression.
Versican mRNA levels in lung tissue were measured from fetal sheep and newborn lambs by Northern blot analysis, as previously described (18, 19). Total RNA was extracted from lung tissue (40 μg), denatured, and electrophoresed in a 0.8% agarose gel containing 2.2 M formaldehyde. The gel was hydrolyzed by washing in 50 mM sodium hydroxide (10 mM sodium chloride) for 20 min and then neutralized in 0.1 M Tris·HCl, pH 7.5, for 20 min. The RNA was transferred to a nylon membrane (Duralon, Stratagene, La Jolla, CA) by capillary action and cross-linked to it using ultraviolet light (Hoeffer UVC 500, Amrad). The membrane was incubated in hybridization buffer overnight at 42°C, followed by hybridization with an ovine-specific 32P-labeled cDNA probe for versican (2 × 106 counts·min−1·ml−1) for 24 h at 42°C; radioactivity bound to the membrane was detected by exposure to a storage phosphor screen (Molecular Dynamics, Sunnyvale, CA). To standardize the amount of total RNA loaded onto each lane, the blot was stripped by washing in 0.01× standard saline-sodium citrate, containing 0.5% SDS, and was reprobed with a 32P-labeled cDNA probe for 18S rRNA. The relative levels of versican mRNA and 18S rRNA were quantified by measuring the total integrated density of each band using a phosphor-imager and ImageQuant software (Molecular Dynamics).
The ovine-specific versican cDNA probe was isolated by RT-PCR using forward (5′-CTCCCTCTCTGGAAAAGTC-3′) and reverse (5′-CCGCCCTGTAGTGAAAC-3′) primers to amplify a 371-bp sequence corresponding to the G1 domain of versican, which is present in all isoforms of versican. The isolated cDNA fragment had 88% homology over nucleotides 368–735 of the human V0-V2 sequence and 88% homology over nucleotides 207–574 of the published V3 sequence. The fragment was subcloned into PGEMT-easy for amplification and was labeled with [α-32P]dCTP by the random-priming technique (Oligolabelling Kit, Pharmacia).
CS analysis by FACE.
CS and hyaluronan (HA) levels in microdissected fetal lung tissue were analyzed by FACE, as previously described (22), with a modification to remove genomic DNA. Briefly, fetal lung tissue (∼1 g) was microdissected to remove all visible blood vessels and airways using a dissecting microscope, as previously described (14). Dissected lung tissue (20–25 mg) was placed in 300 μl of 100 mM ammonium acetate, pH 7.0, and digested with 100 μg/ml proteinase K at 60°C. The enzyme was then inactivated by boiling, the digested tissue was centrifuged, and the supernatant digested for 1 h at 37°C with 125 μg/ml DNase I. The samples were desalted by centrifugation in MicroCon YM-3 filter devices and GAGs recovered from the filter in the retentate. They were freeze-dried and resuspended in a small volume of 100 mM ammonium acetate, pH 7.0, and the concentration of sulfated GAGs was measured by the 1,9-dimethylmethylene blue assay (11) using CS C as a standard (shark cartilage, Sigma-Aldrich). Aliquots containing 5 μg of GAGs were digested in the ammonium acetate buffer (pH 7.0) with 10 mU of chondroitinase avidin-biotin complex (Seikagaku) for 16 h at 37°C. The cooled samples were then centrifuged through MicroCon YM-3 filter devices, and the disaccharides recovered in the filtrate and freeze-dried. Freshly prepared 2-aminoacridone (5 μl; Bioscientific; 25 mg/ml dissolved in dimethylsulfoxide-acetic acid 85:15 vol/vol) was added to the dried disaccharides, vortexed, and allowed to incubate for 15 min at room temperature. Five microliters of 1 M sodium cyanoborohydride were then added to each tube, and fluorotagging was allowed to occur at 37°C for 18–20 h. The samples were cooled to room temperature and mixed with 10 μl of 37.5% glycerol, and 3- to 5-μl aliquots were analyzed immediately by electrophoresis on Mono Composite gels (Glyko or Epitope Technologies) or stored at −70°C. To enable quantitation, for each experiment, monosaccharide standards of known picomolar amounts were fluorotagged and analyzed on the same gels. Fluorescently tagged samples and standards were imaged and quantitated using Quantity One software (BioRad).
Data are expressed as the means ± SE, and the level of significance is P < 0.05, unless otherwise stated. All GA-related changes in versican, C-4-S, and C-6-S, as well as the proportions of HA (measured as Δ-di-HA) as well as monosulfated CS (Δ-di-4S and Δ-di-6S) and unsulfated CS disaccharides (Δ-di-0S) were analyzed by a one-way ANOVA, followed by a Fisher least squares difference test to identify differences between values at different gestational ages.
Versican was widely distributed throughout the perialveolar region of the fetal lung at all gestational ages studied; staining was highest around blood vessels, but dense staining also occurred in the perialveolar parenchyma (Fig. 1). The mean density of versican staining per unit area of tissue within the perialveolar parenchyma was similar at 90 days [29.9 ± 2.3 arbitrary units (AU)], 111 days (29.4 ± 0.4 AU), 126 days (26.6 ± 3.0 AU), and 138 days GA (28.7 ± 1.8 AU), but was significantly reduced to 23.7 ± 1.5 AU at 142 days GA (compared with 90-day value); the reduction between 138 and 142 days GA just failed to reach significance (P = 0.06). In lambs at 2 wk of age, the mean density of versican staining per unit area of perialveolar tissue increased to 36.2 ± 0.9 AU and was significantly higher than the values measured at each fetal age.
To account for the reduction in tissue volume between 90 days (66.0 ± 4.6%) and 142 days GA (25.3 ± 1.5%), the mean density of versican staining for each individual fetus was adjusted for the tissue fraction, calculated for that animal, to determine relative versican levels per unit volume of lung. Per unit lung volume, versican levels were reduced (P < 0.05) from 19.9 ± 2.7 at 90 days GA to 13.3 ± 1.9 at 111 days and to 7.8 ± 1.0, 9.1 ± 0.8, and 6.0 ± 0.5 at 126, 138, and 142 days GA, respectively (Fig. 2, top). This reduction in versican level was closely correlated with the reduction in lung tissue volumes (r2 = 0.94, P < 0.01) over late gestation (Fig. 3, top). Due to the lack of tissue space volume data, measurements of versican density could not be calculated for lambs at 2 wk PA.
C-4-S and C-6-S.
The mean density of C-4-S staining per unit area of perialveolar tissue increased from 57.0 ± 1.0 AU at 90 days to 73.2 ± 0.9 AU at 142 days GA. However, after accounting for the decrease in lung tissue volume, C-4-S level per unit lung volume decreased from 37.5 ± 2.2 at 90 days GA to 28.2 ± 4.2 at 111 days GA and then to 17.6 ± 1.2, 21.0 ± 1.0, and 18.5 ± 1.3 at 126, 138, and 142 days GA, respectively (Fig. 2, middle). As for C-4-S, the uncorrected mean density of C-6-S staining per unit area of tissue increased from 42.8 ± 2.0 AU at 90 days GA to 60.2 ± 1.3 AU at 111 days GA and 58.4 ± 1.8 AU at 126 days GA and then increased further to 67.5 ± 4.3 and 66.9 ± 1.4 AU at 138 and 142 days GA, respectively; C-6-S values remained at these values in lambs at 2 wk after birth (65.7 ± 1.2 AU). However, the corrected values for C-6-S levels per unit lung volume decreased from 28.3 ± 2.8 and 27.4 ± 4.3 at 90 and 111 days GA, respectively, to 17.0 ± 0.6, 21.3 ± 1.2, and 16.9 ± 0.8 at 126, 138, and 142 days GA, respectively (Fig. 2, bottom). The reductions in both C-4-S and C-6-S levels were correlated with the reduction in lung tissue volumes (r2 = 0.97, P < 0.001 and r2 = 0.75, P < 0.01, respectively) over late gestation (Figs. 3, middle and bottom).
Colocalization of versican with C-4-S and C-6-S.
The majority of versican was found to be colocalized with both C-4-S and C-6-S (Fig. 1) at all gestational ages examined, with only a very small proportion (1–3%) of versican not colocalizing with either of these CS groups. The proportion of versican that was not colocalized with C-4-S tended to decrease from 90 to 142 days GA, although the decrease was not quite significant (P = 0.07) when all data were analyzed together; however, the value measured at 111 days GA (3.1 ± 0.1%) was almost threefold greater than the value measure at 142 days GA (1.1 ± 0.2%). On the other hand, the proportion of versican that was not colocalized with C-6-S significantly increased from 0.7 ± 0.6% at 90 days GA to 3.2 ± 0.6% at 142 days GA (P < 0.05).
Colocalization of C-4-S and C-6-S with PGs other than versican.
The proportion of C-4-S that was not colocalized with versican significantly increased from 17.3 ± 2.1% at 90 days GA to 42.9 ± 4.3% at 142 days GA, indicating increased localization with nonversican PGs (Fig. 4). Similarly, the proportion of C-6-S that was not colocalized with versican significantly increased from 3.5 ± 1.2% at 90 days GA to 25.1 ± 2.6% at 142 days GA (Fig. 4).
The relative proportions of monosulfated (Δ-di-4S and Δ-di-6S) and unsulfated disaccharides (Δ-di-0S) that comprised the CS GAGs isolated from fetal lung tissue were detected by FACE analysis (Fig. 5); nonsulfated HA disaccharide units (Δ-di-HA) were also detected.
Relative to the total amount of GAG, the amount of Δ-di-0S was unchanged throughout the last one-third of gestation in fetal sheep, being similar at 91 days (93.2 ± 2.2 pmol/μg GAG) and 142 days GA (80.2 ± 4.0 pmol/μg GAG) (Fig. 6, top). Similarly, the amount of Δ-di-4S did not change throughout gestation and was not different at 91 days (111.3 ± 6.0 pmol/μg GAG) and 142 days GA (109.1 ± 4.5 pmol/μg GAG) (Fig. 6, top). In contrast, the amount of Δ-di-6S significantly decreased from 224.0 ± 13.8 pmol/μg GAG at 91 days GA to 154.6 ± 13.4 pmol/μg GAG at 111 days GA, 125.2 ± 15.0 pmol/μg GAG at 126 days GA, to 103.4 ± 6.9 pmol/μg GAG at 138 days, and to 99.2 ± 3.5 pmol/μg GAG at 142 days (Fig. 6, top). At 2 wk of PA, the amounts of Δ-di-0S and Δ-di-6S were significantly increased to 108.9 ± 9.6 pmol/μg GAG and 154.2 ± 10.4 pmol/μg GAG, respectively; the amount of Δ-di-4S tended to increase to 136.1 ± 5.2 pmol/μg GAG.
When expressed as a proportion of all CS disaccharides, Δ-di-6S accounted for 52.1 ± 1.0% of the total disaccharide units at 91 days GA, which decreased to 44.6 ± 1.2% at 111 days GA and to 39.3 ± 0.1% at 126 days GA (Fig. 6, middle). At 138 days GA (38.1 ± 0.5%), the proportion was similar to the 126-day value, but then decreased to 34.4 ± 0.4% at 142 days GA. In contrast, Δ-di-4S accounted for 26.0 ± 0.4% of the total disaccharide units at 91 days GA, and its proportion increased to 31.2 ± 1.2% at 111 days GA (P < 0.05; Fig. 6, middle). The proportion of Δ-di-4S remained at this level at 126 days (33.7 ± 1.0%) and 138 days GA (35.0 ± 0.8%) and then increased again at 142 days GA (to 37.8 ± 0.4%; P < 0.05). The proportion of Δ-di-0S accounted for 21.9 ± 1.1% of total disaccharide units at 91 days GA, which remained unchanged at 111 days GA (24.2 ± 0.6%) before increasing to 27.0 ± 1.2 at 126 days GA (P < 0.05); values remained at this level during late gestation (138 days: 27.3 ± 1.0%, 142 days: 27.8 ± 0.3%) and after birth (27.2 ± 0.8%). Following birth, the proportion of Δ-di-4S decreased to 34.2 ± 1.1%, whereas the proportion of Δ-di-6S increased to 38.6 ± 0.2% (P < 0.05; Fig. 6, middle).
The concentration of HA within the distal airway parenchyma was very high at 90 days GA (1,397.6 ± 258.8 pmol/μg GAG) and was markedly reduced to 134.5 ± 22.3 pmol/μg GAG at 111 days GA (P < 0.05), 99.0 ± 20.8 pmol/μg GAG at 126 days GA, 83.8 ± 6.5 pmol/μg GAG at 138 days GA, and 92.9 ± 9.9 pmol/μg GAG at 142 days GA (Fig. 6, bottom). At 2 wk of PA, HA levels (135.3 ± 13.7 pmol/μg GAG) were not different from levels just before birth (142 days GA).
Versican mRNA Levels
Northern blot analysis was used to detect two large transcripts, which encode the two versican isoforms, V0 (∼12.2 kb) and V1 (9.5 kb; Fig. 7A). As the density of each of these isoforms varied in an identical manner across gestation, the densities of both bands were summed to give a combined value (Fig. 7B). Expressed as a percentage of the mean values obtained from fetuses at 105 days GA (100.0 ± 13.7%), versican mRNA levels were significantly increased at 114 days GA (157.8 ± 25.6%, P < 0.05). Compared with the 114-day value, versican mRNA levels were significantly (P < 0.05) reduced at 126 days (89.8 ± 8.9%), 138 days (82.4 ± 10.3%), and 142 days GA (92.1 ± 8.3%), as well as 2 wk after birth (63.4 ± 2.1%) (Fig. 7B).
Versican is a member of the hyalectan family of PGs and is an abundant PG in the lung ECM (8). It is a large CS-rich species, which underpins its important physiological role in controlling interstitial tissue hydration, solute permeability, and influencing viscoelastic properties of tissue (2). The high anionic charge density of PGs, such as versican, attracts mobile counterions, which, in turn, generate an osmotic gradient promoting the retention of water. As a result, PGs contribute to interstitial tissue volumes, provide tissue resilience, and influence tissue mechanics (6, 9, 20, 25, 30). We hypothesized that a reduction in versican content or a change to its microstructure, particularly to the CS side chains leading to a decrease in anionic charge density, may account for the decrease in lung tissue volumes that characterize structural development of the lung in late gestation. In this study, we investigated the ontogenic changes in expression and localization of versican, as well as changes in the microstructure of CS PGs, within the lung. The stage of development examined included the period during which terminal respiratory sacs and alveoli develop, as well as the period when parenchymal tissue volume surrounding the saccules/alveoli markedly declines; tissue volumes decline from 66.0 ± 4.6% of total lung volume at 90 days to 25.3 ± 1.5% at 142 days. Our results demonstrate that the reduction in perisaccular/alveolar tissue volumes is closely associated with marked changes in versican and HA content, as well as in CS microstructure.
Versican was widely distributed throughout the interstitial tissue of the terminal airways at all development stages examined. However, the level of versican in perisaccular/alveolar lung tissue markedly decreased between 90 and 126 days GA, which closely corresponded with the greatest change (56% reduction) in lung tissue volumes over this developmental period. As a result, the level of versican within the perisaccular/alveolar tissue was closely correlated with the fractional tissue volume of the lung. This finding is consistent with the suggestion that the level of versican retained within the interstitium could be a major determining factor of lung tissue volumes in late gestation. Other studies performed within our laboratory (data not shown) have shown that elastin (another ECM protein) increases as a proportion of total tissue mass with increasing gestation, indicating that not all ECM proteins follow the same pattern. Indeed, versican decreases at an earlier GA than the increase in elastin, which increases in association with alveolarization. The factors that could regulate versican levels within the fetal lung, leading to a reduction in tissue volumes, are unknown, although alterations in HA levels may be involved (see below). It is well established that endogenous corticosteroids are important regulators of prenatal lung development, including remodeling of the terminal respiratory units. However, the major ontogenic changes in versican levels that we observed occurred well before the expected preparturient increase in fetal plasma cortisol levels begin ∼135 days GA (4).
HA plays a major role in binding and anchoring versican (and other hyalectans) within the interstitial tissue space (13). It is possible, therefore, that the large reduction in HA levels we detected in lung tissue (Fig. 6, bottom) is responsible for the reduction in versican levels with advancing GA, particularly as versican expression was not reduced. HA levels decreased by 96% between 90 and 126 days GA, with most of the decrease occurring before 111 days (Fig. 6, bottom); these findings are consistent with previous studies in the developing rabbit lung (3). Lower HA levels in lung tissue would provide fewer anchoring sites for versican, thereby limiting the ability of the tissue to retain versican. Since HA-PG aggregates contribute to the overall charge density of the interstitial space and hence play an important role in “tissue hydration,” changes in the amount of HA-versican aggregates are likely to have a large effect on the viscoelastic properties of the lung. For instance, loss of CS PGs (via degradation) decreases the viscoelasticity of the lung (2), whereas degradation of CS GAGs in lung parenchymal strips increases mechanical friction, leading to increased energy dissipation and stiffness; this may be related to changes in hydration and confirms the importance of CS GAGs for lung tissue viscoelasticity (2). Thus it is likely that large reductions in HA-versican aggregates will reduce the charge density of the tissue, leading to a reduction in hydration and perialveolar tissue volumes, in preparation for breathing at birth. The mechanism leading to such a profound reduction in lung tissue HA levels is currently unknown, although it is likely that increased hyaluronidase activity, leading to increased depolymerization of HA, is a major contributing factor.
As for versican, C-4-S and C-6-S disaccharides were widely distributed throughout the perisaccular/alveolar tissue of the ovine lung at all gestational ages examined. The distribution of C-4-S and C-6-S in the lung has not been previously studied in fetal sheep, although studies on the distribution of total CS throughout lung development in rats have yielded similar results (27). When the decrease in tissue volume was accounted for, the level of both C-4-S and C-6-S was found to decrease with increasing GA. It is likely that this decrease resulted from the decrease in versican levels, as only a very small proportion of versican (1–3%) was not colocalized with either C-4-S or C-6-S, and versican was the predominant sulfated CS PG up until term. Indeed, colocalization with versican accounted for ∼65% of C-4-S and ∼75% of C-6-S in lung parenchymal tissue at term, decreasing from ∼80 and ∼95%, respectively, at 90 days GA. As the proportion of C-4-S and C-6-S not colocalized with versican markedly increased with increasing age, colocalization of these CS side chains with other CS PGs must increase over this developmental period. Furthermore, the finding that the colocalization of versican with C-4-S tended to decrease, whereas the colocalization with C-6-S increased with increasing developmental age, indicates that the sulfation pattern of the CS side chains on versican is changing and may be developmentally regulated. However, the net effect of this change in CS sulfation pattern is currently unknown.
An analysis of the amount and proportion of nonsulfated and monosulfated CS disaccharides in perisaccular/alveolar lung tissue, as assessed by FACE, indicated significant changes in the ratio of C-4-S to C-6-S; this analysis includes all CS-containing PGs, not just versican. At 90 days of gestation, Δ-di-6S accounted for the greatest proportion of total CS disaccharides present in the ovine fetal lung, which markedly decreased with advancing GA. On the other hand, although the amounts of Δ-di-4S did not change over the last one-third of gestation, expressed as a proportion of total CS disaccharides, the proportion of Δ-di-4S increased toward term and eventually became the predominant CS disaccharide at term. As a result, there was a marked gestation-dependent decrease in the ratio of 6-sulfated to 4-sulfated CS residues in perisaccular/alveolar lung tissue. An age-related decrease in the proportion of C-6-S to C-4-S has not previously been reported in the developing lung, but has been demonstrated in rat skin (12), chicken brain (16), and chick epiphyseal cartilage (24). Although the biological significance of the changes in CS sulfation patterns within perisaccular/alveolar lung tissue is unclear, they may alter the ligand-binding properties of versican, which could have a number of biological consequences, including alterations in growth, cellular differentiation, as well as structural development of the terminal respiratory units (15). It is also important to consider the potential contribution of the low-abundant disulfated and trisulfated CS species, which we did not measure in this study; changes in these species may alter the osmotic properties of the tissue.
As versican is the predominant C-6-S containing PG in the perisaccular/alveolar region of the lung, colocalizing with ∼95% of all C-6-S at 90 days GA, the reduction in Δ-di-6S levels over the developmental stage examined is likely due to the decrease in versican levels. On the other hand, although versican is also the predominant C-4-S containing PG at 90 days GA, colocalizing with ∼80% of all C-4-S, the reduction in versican content with increasing gestation was not reflected by a decrease in Δ-di-4S levels. Consequently, it is likely that the maintenance of Δ-di-4S levels, despite a reduction in versican levels with increasing gestation, is due to an increase in other PGs that contain Δ-di-4S. This is consistent with the finding that the proportion of C-4-S not colocalized with versican increased from <20% to almost 50%, again with most of the change occurring before 126 days of GA. Following birth, the amounts of Δ-di-6S, Δ-di-4S, and Δ-di-0S were all markedly increased at 2 wk PA. The mechanisms and significance of these changes are unclear, but could relate to the change in interstitial tissue pressure associated with lung aeration and the formation of surface tension in the lung after birth.
Messenger RNA encoding two isoforms of versican, V0 and V1, were detected in ovine lung tissue by Northern blot analysis. These isoforms contain the largest domains for CS GAG attachment (10) and, therefore, are likely to have the greatest influence on tissue hydration within the perialveolar region of the lung. Other studies using RT-PCR have determined that V3, a splice variant of versican that carries no CS chains, is also detectable within the human lung (5). However, using RT-PCR, we were unable to detect the presence of V3 in fetal lung tissue in sheep (data not shown). The mRNA levels for V0 and V1 were found to vary in an identical manner over the last one-third of gestation, indicating that remodeling of the perialveolar parenchyma is unlikely to involve differential expression of the versican splice variants. Except for a significant increase at 114 days GA, total versican mRNA levels were unchanged, compared with the 105-day value, up until near term (142 days GA). Similarly, although at 2 wk PA there was a tendency for versican mRNA levels to be reduced, this was not statistically significant compared with the 105-day value. These data indicate that alterations to versican expression are unlikely to account for the reduction in versican levels and perialveolar tissue volumes with advancing development. However, it is important to note that the mRNA was not extracted from microdissected lung tissue, and, therefore, all tissue compartments in the distal lung, not just perisaccular/alveolar regions, were included in the Northern analysis. As a result, high versican expression in structures such as arteries may have obscured subtle changes in V0 and V1 expression patterns in perisaccular/alveolar lung tissue.
In this study, we demonstrate that the reduction in perialveolar/saccular tissue volumes that characterizes lung maturation during late gestation in fetal sheep is closely associated with a reduction in versican levels, as well as marked changes in the sulfation pattern of the CS side chains. In view of the role that PGs such as versican play in tissue hydration, we propose that the loss of versican from the perialveolar interstitial tissue is a major contributing factor to the decrease in parenchymal tissue volumes that characterizes fetal lung development. Furthermore, we suggest that the reduction in versican levels may have resulted from a reduction in anchoring sites within the interstitium due to a marked reduction in HA levels, which is believed to be responsible for retaining versican within the ECM. In addition to the reduction in versican levels, we detected marked changes in the sulfation pattern of CS disaccharides within perisaccular/alveolar lung tissue. Together, these findings indicate that structural development of the lung is closely associated with marked changes in versican levels and microstructure of the CS side chains in the perisaccular/alveolar lung tissue. It is possible, therefore, that these molecular structures play an integral role in regulating parenchymal tissue volumes, which, in turn, determines the structural characteristics of the gas exchange units, particularly the dimensions of the air/blood-gas barrier.
This work was funded by the National Health and Medical Research Council of Australia and the Murdoch Childrens Research Institute.
We are indebted to Alison Thiel and Valerie Zahra for expert technical assistance.
↵* J. Faggian and A. J. Fosang contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society