Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The microvessels of the normal adult lung (vessels  100 µm in diameter) consist of a mixed population of segments that have muscular, partially muscular, or nonmuscular walls. Typically, new smooth-muscle cells develop in these segments in clinical pulmonary hypertension (PH) (1). Increasing in the normally muscular segments they contribute to change in vasoreactivity, and appearing in segments that are normally nonmuscular they contribute to restriction of the microcirculation, profoundly influencing pulmonary vascular resistance and pressure. By the development of these cells, and by endothelial cell hypertrophy and hyperplasia, these relatively thin-walled segments are converted to "resistance" vessels in which wall thickness is high for diameter. Experimental studies of hyperoxia- induced PH have shown that these new smooth-muscle cells develop by hypertrophy and hyperplasia of pre-existing smooth-muscle cells or from precursor smooth-muscle cells (Figures 1A to 1D). These precursor cells include intermediate cells (2), cells that lie immediately beneath the endothelium of the normal vessel wall (Figure 1B), and fibroblasts that are recruited to the vessel wall (Figure 1C) from the adjacent interstitium (3, 4). In many of the smallest vessels, ones 20 to 35 µm in diameter, the recruited fibroblasts form an important source of new vascular cells, migrating through the widened interstitial space of the injured lung to align around the vessel wall (3, 4). Pericytes, normally found within capillaries, thicken capillary walls in the hyperoxic lung (3, 4), where adjoining fibroblasts may also be recruited (Figure 1D). Although microvascular fibroblasts and intermediate cells proliferate at a similar time (i.e., early hyperoxic PH, Days 4 to 7), the proliferation of fibroblasts is the greater (5). Either singly or in combination, these cells proceed to build new vessel walls by acquiring a contractile phenotype and by the de novo formation of elastic laminae. In those microvessels at the entrance to the acinus where smooth-muscle cells are normally present (Figure 1A), these cells proliferate relatively late (late hyperoxic PH, Day 28) and peak activity is less than for fibroblasts or intermediate cells (5), indicating that these existing cells contribute little to the increasing microvascular smooth-muscle cell population.

View larger version (40K): [in this window] [in a new window]   Figure 1.   Cellular pathways of vessel wall remodeling in PH. Illustration (not to scale) of normal precapillary vessels (top vessels, A-C ) and cell pathways of remodeling in PH (indicated by arrows). In normally muscular segments (A), where preexisting smooth-muscle cells are defined by an internal and external elastic lamina, smooth-muscle cell hypertrophy and hyperplasia increase wall thickness. Similarly, vessel walls are thickened by preexisting intermediate cells that are normally found adluminal to a single elastic lamina (B); an internal lamina then forms to separate these cells from endothelium. In small and normally nonmuscular segments (20 to 35 µm), where there is no elastic lamina and the wall consists of endothelial cells (C ), interstitial fibroblasts are an important source of new smooth-muscle cells; external and internal laminae form de novo. In capillary segments (D), the wall is thickened by hypertrophy and hyperplasia of preexisting pericytes and adjacent interstitial fibroblasts. On occasion, a single lamina forms to enclose these cells. The morphology of recruited fibroblasts (including cells in the process of migrating and aligning) and of intermediate cells and pericytes in distal vessels and capillaries of the hyperoxic lung is as previously described (3, 4).

Although the sequence of morphologic changes that lead to neomuscularization of the microvessels has been established, the contractile proteins expressed as precursor smooth-muscle cells (PSMCs) acquire a smooth-muscle phenotype that has not been systematically studied. Clearly, identifying the type and sequence of expression of the filament proteins associated with this phenotypic shift is important to understand well remodeling in the hypertensive lung and the pathogenesis of this vascular lesion. In this first in a series of studies to characterize the evolving phenotype of these cells, we studied -smooth-muscle actin (SMA) expression, a recognized first marker of developing smooth-muscle cells (6), during microvessel wall remodeling in an in vivo model of hyperoxic PH (4). Other studies are under way to determine the expression of proteins needed to characterize smooth-muscle cell maturation (e.g., smooth-muscle myosin heavy chains [SMMHCs], calponin, h-caldesmon, and metavinculin) (6). The most abundant isoform of a family of actin isoenzymes (, , ), SMA is also the most abundant cellular protein of smooth-muscle cells (representing up to 40% of total protein) (7). It has been widely used to identify the initial stage of expression of a smooth-muscle phenotype in vascular development and of myofibroblast development in response to injury and in disease (8, 9). We first established the distribution of vessels with SMA expressing cells in a quantitative light microscopy study, and then used high-resolution immunogold labeling techniques to relate SMA expression to PSMC phenotype. We found that SMA was abundantly expressed within the microvessels of the hypertensive lung; that expression is filament-associated relatively early within intermediate cells; and that in fibroblasts, expression is not filament-associated until late-stage wall remodeling. The data support the presence of dual pathways of PSMC development.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Tissue Preparation

Adult male Sprague-Dawley viral antibody-free rats (220 to 240 g body weight; Charles River Laboratories, Kingston, NY) breathed air or 87% oxygen at normobaric pressure for 1 to 28 d. The vessels and airways of the lungs of control and hyperoxic rats (Days 0, 1, 4, 7, 14, and 28; n = 4 at each time point except Day 21, when n = 2) were simultaneously inflated with 3% paraformaldehyde/ 0.1% glutaraldehyde in phosphate-buffered saline (PBS, pH 7.23, at 100 cm H₂O and 23 cm H₂O, respectively). Tissue blocks selected from the upper, middle, and basal regions of the left lung (15 × 6 × 4 mm) were washed in PBS, suspended in 6.8% sucrose in PBS (pH 7.4), dehydrated in 100% acetone (60 min at 4°C), and embedded in Historesin Plus (3 h at 4°C; no. 7022 2224-861; Leica, Deerfield, IL); or 1 mm³ cubes were washed in PBS, dehydrated in cold ethanol at 4 to 20°C, embedded in Unicryl Resin (British BioCell, Cardiff, UK), and polymerized by ultraviolet light (48 h at 10 to 20°C). Two-micrometer Historesin Plus sections were stored at 4°C until use; 90-nm Unicryl sections were picked up onto formvar-coated nickel grids and stored at room temperature until use.

Immunocytochemistry

A monoclonal SMA antibody was used (the NH₂-terminal synthetic decepeptide of SMA being used as the immunogen; mouse immunoglobulin G2a isotype; Clone 1A4, no. A2547; Sigma, St. Louis, MO). Details of the specificity of the antibody as determined by Western analysis and demonstrated in immunohistochemical studies are as reported (8). We routinely tested the specificity of immunostaining by treated additional sections with 0.5% bovine serum albumin (BSA) in PBS and omitting the primary antibody. No staining was observed in these negative control sections.

All sections were washed well in PBS or distilled water (dH₂O) between all steps of the immunocytochemical and immunogold procedures. For the Labeled-[strept]Avidin-Biotin technique we used a Histostain SP kit (95-6543 AEC Mouse Kit; Zymed Laboratories, Inc., South San Francisco, CA) appropriate for the primary antibody. Two-micrometer sections were hydrated in PBS, and sites of endogenous peroxide were quenched in 3% hydrogen peroxide in absolute methanol (10 min at room temperature [RT]). Nonspecific background staining was blocked with 10% normal goat serum (10 min), and the sections were incubated with the primary antibody diluted 1:400 in 0.5% BSA in PBS (16 h at 4°C). After incubation in the biotinylated secondary antibody (Zymed Histostain-SP kit, 20 min at RT), they were treated with the enzyme conjugate (Zymed Histostain-SP kit, 20 min at RT) and placed in the substrate-chromogen mixture (aminoethylcarbazole, 3 to 5 min at RT, the development of the reaction product being monitored under a microscope). They were then washed and stained with 50% Meyers hematoxylin (BioGenex, San Ramon, CA) and mounted in GVA (Zymed kit) or aqueous mounting medium (BioGenex). The chromogen produces a red reaction product at immunopositive sites, whereas nuclei stain blue.

For immunogold studies, 90-nm sections were treated with 1% BSA in PBS (5 min at RT), incubated with the primary antibody diluted 1:400 in 0.5% BSA in PBS (2 h at RT or 16 h at 4°C), treated with 0.5% BSA in PBS (5 min at RT), and incubated (60 min at RT) with protein-A gold (Auroprobe AG10) diluted 1:50 in PBS. All sections were finally stained with 7.5% uranyl magnesium acetate in dH₂O (20 min at RT) and 0.2% lead citrate in dH₂O (5 min at RT).

Quantitative Analysis of Vessels by Light Microscopy

Vessels from the lungs of each animal (for Days 0, 1, 4, 7, 14, and 28, n = 3 at each time point; for Day 21, n = 2) were landmarked by their location, that is, associated with bronchioli, terminal or respiratory bronchioli, or alveolar ducts, or lying in the alveolar wall. When immunopositive cells forming a single or multiple subendothelial layer encompassed the circumference fully or in part, a vessel was termed muscular or partially muscular, respectively. When immunopositive cells were absent, the wall consisted only of endothelial cells and the vessel was termed nonmuscular. In 2-µm sections, vessels associated with bronchioli, terminal bronchioli, and respiratory bronchioli were assessed qualitatively because all were immunopositive in the normal and hypertensive lung; alveolar duct and alveolar wall vessel populations were analyzed quantitatively because each contains segments with muscular, partially muscular, and nonmuscular walls. We recorded the wall structure, external diameter (ED), and size of 50 alveolar vessels per animal, and measured the wall thickness of muscular and partially muscular vessels. Because elastic laminae (which usually define the limits of vessel walls) were not stained by the immunocytochemical techniques used in this study, the wall thickness of these vessels was measured between the abluminal edge of immunopositive cells and the adluminal edge of adjacent endothelium. All of the muscular and partially muscular vessels of the hyperoxic lung had immunopositive cells. Their diameters included wall thickness and lumen diameter; that of nonmuscular vessels included the lumen diameter and endothelium. Percent wall thickness (%WT) was calculated as [2 × 100 × WT]/ED.

Continuous data were analyzed using a one-way analysis of variance followed by a post hoc Scheffe Test. Categorical data were compared using a cross-tabulation continency table (²-test). All continuous data were shown as means ± SEM (Figure 2). Categorical data were expressed as frequency distribution (Figures 3 and 4). A probability of P < 0.05 was taken as significant.

View larger version (18K): [in this window] [in a new window]   Figure 2.   Percent wall thickness in hyperoxic PH. %WT of partially muscularized (shaded bars) and muscularized (closed bars) vessels expressing SMA in normal lung and between Days 4 and 28 of hyperoxia. At Day 28, in both partially muscular and muscular vessels the %WT is significantly increased as compared with earlier time points (*P < 0.001). Values are means ± SEM.

View larger version (23K): [in this window] [in a new window]   Figure 3.   Distribution of vessels by wall structure in hyperoxic PH. The distribution of nonmuscular, partially muscular, and muscular vessels is significantly different (P 2 < 0.02) from Day 4 in alveolar wall vessels (A) and from Day 14 in alveolar duct vessels (B); muscular and partially muscular vessels increase at the expense of nonmuscular vessels.

View larger version (15K): [in this window] [in a new window]   Figure 4.   Predominant remodeling of small vessels. Distribution of vessels with wall muscle by size, expressing SMA in hyperoxic PH, including vessels with partially and fully muscular walls. Note that the muscularity of vessels in the diameter range of 15 to 25 µm and 26 to 50 µm increased with time. At Day 28, the distribution of these vessels was significantly different from earlier time points (P 2 < 0.001).

Vessel Analysis by Electron Microscopy

We mapped the location of alveolar vessels and measured their size in 1-µm-thick Unicryl sections stained with toluidine blue (0.1% in 1% sodium borate). We identified the same vessels in 90-nm sections from each block and photographed consecutive segments of the wall around the circumference of each vessel. Working prints (×2.5 negative) provided an enlarged composite of a vessel and its cells with immunopositive sites identified by gold label. The typical morphology of the intermediate cell and of fibroblasts at stages in their recruitment to microvessel walls, and the pericytes of capillaries, has been described in detail (3, 4). Briefly, as part of the normal vessel wall, intermediate cells always lie in close apposition to the endothelial cell basement membrane and contain filament arrays, whereas fibroblasts recuited to the wall are identified by their morphology and their location in relation to the interstitium and endothelium (3, 4). Initially, the morphology of recruited cells is that of typical fibroblasts, that is, extensive rough endoplasmic reticulum, Golgi, and no filaments. Migrating cells resemble cells in vitro, with an extended leading pseudopodia and trailing cell body; aligning cells are characterized by the development of filaments and formation of lamellapodia along their adluminal cell edge as this approaches the endothelial basement membrane; whereas aligned cells are oriented circumferentially around vessels (3, 4).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Analysis of Lung Vessel Populations

In the normal and hyperoxic lung, bronchial smooth-muscle cells and the smooth-muscle cell of preacinar vessels associated with bronchioli and terminal bronchioli expressed SMA (Figures 5A, 5C, 6A, and 6C). In the normal lung, in the thick-walled oblique muscular artery (TWOMA; where an additional oblique muscle coat is present along the middle two-thirds of the axial pathway), and in vessels associated with respiratory bronchioli at the entrance to the acinus, cells expressing SMA were evident (Figure 5B). Alveolar duct and alveolar wall vessels with cells expressing SMA were rare. The thin processes of immunopositive cells were seen around an occasional vessel, including ones with a partially muscular wall (Figures 5B and 5D).

View larger version (163K): [in this window] [in a new window]   View larger version (165K): [in this window] [in a new window]   View larger version (182K): [in this window] [in a new window]   View larger version (161K): [in this window] [in a new window]   View larger version (102K): [in this window] [in a new window]   Figure 5.   SMA expression in normal lung and in early-stage hyperoxic PH. In the normal lung, SMA is expressed by the smooth-muscle cells of bronchioli (A and B) and terminal bronchioli (C, tb) and in associated vessels (B, tw = TWOMA, see ANALYSIS OF LUNG VESSEL POPULATIONS in RESULTS; C, tbv = terminal bronchiolar vessel); expression is low or absent in alveolar vessels of normal lung (D, alveolar wall vessel at double arrow and alveolar wall vessel, ED 29 µm at single arrow) and in the hyperoxic lung at Day 4 (E, alveolar wall vessel, ED 45 µm). Bars = 50 µm.

Initially in hyperoxia (Day 1), the distribution of SMA cells was similar to that throughout the normal lung. By Day 4, increased numbers of SMA cells were evident in preacinar vessels of the hyperoxic lung but most intraacinar vessels were still negative (Figure 5E). At Day 7, alveolar vessels with SMA cells increased and septal cells at the entrance to alveolar ducts were also positive. This pattern persisted at Day 14. By Day 28, the number and intensity of SMA cells had increased and small thick-walled alveolar wall and duct vessels with these cells were evident, including vessels with muscular walls (Figures 6D and 6E) as well as partially muscular vessels (Figure 6F). All of the distal thick-walled vessels had SMA cells. At this time, immunoreactive cells also were clearly evident in capillary walls, and septal smooth-muscle cells retained their immunoreactivity. Little evidence of SMA cells was detected within the interstitium. Only in rare small foci, where alveolar architecture was disrupted and interstitial fibroblasts were seen infiltrating alveoli (with Boutons de Masson clearly evident), were these cells weakly immunopositive.

View larger version (131K): [in this window] [in a new window]   View larger version (144K): [in this window] [in a new window]   View larger version (206K): [in this window] [in a new window]   View larger version (104K): [in this window] [in a new window]   View larger version (114K): [in this window] [in a new window]   View larger version (138K): [in this window] [in a new window]   Figure 6.   SMA expression in alveolar vessels in late-stage hyperoxic PH. In the hyperoxic lung at Day 28, as in normal lung, SMA is expressed by airway smooth-muscle cells and in associated vessels (A, bv = bronchial vessel; B, tw = TWOMA, ED 138 µm; C, tbv = terminal bronchiolar vessel, ED 50 µm). In contrast to normal lung, many alveolar vessels are thick-walled and have SMA-positive cells (D, muscular alveolar vessel, ED 26.4 µm; E, muscular alveolar duct vessel, ED 32 µm; F, partially muscular alveolar wall vessel, ED 33 µm). Bars = 25 µm.

The increase in %WT (reflecting the relationship between vessel wall thickness and diameter) in alveolar wall and duct vessels developing SMA cells in hyperoxic PH is illustrated in Figure 2, increasing (P < 0.001) from 3.4 ± 0.2% to 11.4 ± 2.4% in partially muscular vessels by Day 28, and from 7.8 ± 0.6% to 40.2 ± 1.9% in muscular vessels. The frequency of partially muscular and muscular vessels, including alveolar wall and duct vessels, increased at the expense of nonmuscular vessels (Figures 3A and 3B). By Day 4 the distribution of alveolar wall vessels (Figure 3A) had changed significantly (P < 0.02), and by Day 14 so had the distribution of alveolar duct vessels (P < 0.02, Figure 3B). Analysis of these by size (Figure 4) showed that cells expressing SMA appeared earliest in the smallest vessels (15 to 25 and 26 to 50 µm ED). By Day 28, the distribution of vessels with these cells was significantly different (P < 0.001) from that at the earlier time points (Figure 4).

Analysis of Cell Phenotype in Alveolar Wall Vessels

In all sections examined by high-resolution microscopy the gold label localized only to cell processes; that is, no label localized to the extracellular matrix or to nontissue areas. This technique readily identified cells that had low as well as high levels of SMA expression. Levels were especially low in those cells where the gold localized to the cytoplasm but was not associated with filaments. These studies also detected expression of protein over the nucleus and cytoplasm of endothelial cells, and developing PSMCs.

In the smallest vessels of the normal lung (20 to 35 µm), PSMCs were thin and filament arrays rarely detected (Figures 7A and 7C); when present in larger vessels ( ED 50 µm), they contained organized arrays of fine filaments expressing -SMA (Figure 7B). An example of a small vessel with SMA localized to an endothelial cell and an adjacent PSMC is shown in Figure 7C. Interstitial fibroblasts did not express SMA.

View larger version (103K): [in this window] [in a new window]   View larger version (106K): [in this window] [in a new window]   View larger version (85K): [in this window] [in a new window]   Figure 7.   Alveolar wall vessels in normal lung. (A) Small, thin-walled alveolar wall vessel (ED 35 µm) with the thin processes of PSMCs (arrowheads), alveolus (alv), and endothelial cells (e). (B) High magnification showing PSMC (asterisk) in apposition to endothelium in larger alveolar wall vessel (ED 54 µm). The PSMC process contains fine arrays of filaments decorated with SMA. (C ) High magnification showing PSMC (asterisk) in apposition to endothelium in small alveolar wall vessel (ED 25 µm). The PSMC contains fine arrays of filaments decorated with SMA, whereas in the associated endothelial cell, SMA is localized to the cell nucleus. Bars: (A) 5 µm; (B and C ) 0.5 µm.

At an early stage of hyperoxic PH (Day 4), PSMCs with well-developed arrays of SMA filaments were evident in some of the smallest vessels (20 to 35 µm) (Figure 8). In others, these cells lacked filaments and SMA expression (Figure 9A). This was followed, at Day 7, by the appearance of vessels surrounded by fibroblasts that by their location, orientation, and morphology were defined as at a stage in the process of migrating through the adjacent interstitium toward the abluminal surface of the endothelial cell basement membrane. These cells expressed SMA but not filaments. By Day 14, the fibroblasts were in the process of aligning circumferentially around the endothelial basement membrane, and in those cells lying in close apposition to the endothelium the filaments were well developed and decorated with SMA (Figure 9B and inset). Cells aligning circumferentially but forming part of the outer layer of the vessel wall rather than lying adjacent to the endothelium expressed SMA but no filaments at this time (Figure 9B and inset).

View larger version (142K): [in this window] [in a new window]   Figure 8.   Alveolar wall vessel in early-stage hyperoxic PH. At Day 4, PSMCs (asterisks) in close apposition to endothelium (e) contain extensive filament arrays decorated with SMA (ED 30 µm). Their location at this time indicates that they are intermediate cells. Note specificity of labeling and lack of background scatter. Bar = 0.5 µm.

View larger version (107K): [in this window] [in a new window]   View larger version (138K): [in this window] [in a new window]   Figure 9.   Alveolar wall vessel in early- and mid-stage hyperoxic PH. (A) Early vessel wall remodeling (Day 4, ED 29 µm). PSMCs (asterisks) in close proximity to endothelium (e) and aligning fibroblast (fb) are negative for SMA at this time. (B) Segment of vessel at mid-stage wall remodeling (Day 14, ED 26.5 µm) with the process of a PSMC (asterisk) expressing fine arrays of SMA-positive filaments along its adluminal edge. This cell is lying between the endothelium (e) and a second cell that is an aligning fibroblast (fb). The aligning cell has extensive gold label throughout the cytoplasm in region indicated by arrow (see inset) but no filaments. Bars: (A and B) 1 µm; inset, 0.5 µm.

At the late stage of hyperoxic PH (Day 28), PSMCs formed the walls of the smallest vessels (Figures 10A, 10B, and 12). All of these cells expressed SMA, and in most there were well-developed filaments (Figures 10A and 11). At this time, filaments decorated with SMA either were arranged as parallel arrays localized to cytoplasmic domains (Figures 10 and 11) or appeared as a dense mesh (Figures 12 and 13). Interstitial fibroblasts were hypertrophied and some of these cells expressed SMA over the course of development of PH, but at no time were filaments detected in these cells.

View larger version (135K): [in this window] [in a new window]   View larger version (106K): [in this window] [in a new window]   Figure 10.   Alveolar wall vessel in late-stage hyperoxic PH. (A) Alveolar wall vessel in late-stage wall remodeling (see inset, ED 31 µm, Day 28), endothelial cell (e), PSMCs (asterisks). High magnification of wall region showing PSMC with dense arrays of SMA filaments (double asterisks), lying abluminal to the endothelium and adluminal to a second PSMC (single asterisk) that is binucleated, indicating incomplete cytokinesis (A and B, see arrow). Bars: (A and B) 5 µm; inset, 10 µm.

View larger version (175K): [in this window] [in a new window]   Figure 11.   Alveolar wall vessel in late-stage hyperoxic PH. Higher magnification of PSMC in wall of vessel in Figure 10, showing dense arrays of SMA-positive filaments (single asterisk) and SMA expression in the abluminal PSMC in the absence of filaments (double asterisks). Bar = 1 µm.

View larger version (173K): [in this window] [in a new window]   Figure 12.   Alveolar wall vessel in late-stage hyperoxic PH. Thick-walled vessel in late-stage wall remodeling (Day 28, ED 20 µm). PSMCs (asterisks) lie abluminal to endothelium (e). These cells express SMA filaments and are organized between elastic laminae (iel = internal elastic lamina, eel = external elastic lamina). Aligning fibroblasts are associated with the wall (fb). Bar = 5 µm.

View larger version (156K): [in this window] [in a new window]   Figure 13.   Alveolar wall vessel in late hyperoxic PH. Higher magnification of PSMC (double asterisks) in the vessel shown in Figure 12 with dense filament arrays expressing SMA arranged as a mesh. Bar = 1 µm.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The sequence of wall thickening by the de-differentiation and proliferation of existing smooth-muscle cells typical of large vessels in the hypertensive lung is not mirrored in the distal microvessels. New smooth-muscle cells arise at this site by the migration, proliferation, and differentiation of precursor cells. The sparse amount of data currently available in relation to the contractile proteins expressed by these differentiating cells in part reflects the need for special fixation and embedding procedures to preserve antigenic sites. Using a dilute fixative and a hydrophilic resin requiring a low polymerization temperature, we achieved our goal of preserving antigenic sites and filament structures, and retained the detail needed to determine the structural relationship between vascular wall cells and vessel location within the alveolar-capillary membrane. This approach provides exceptional structural detail when compared with the use of frozen tissue sections obtained by ultracryotomy.

We found SMA, taken as the earliest marker of expression of a smooth-muscle phenotype (6, 9), expressed by intermediate cells (the PSMCs of existing vessel walls), in both the normal and hypertensive lung. Fibroblasts recruited to form new vessel walls in the hypertensive lung also expressed this protein, including cells in the process of migrating and aligning. All vessels undergoing wall remodeling had SMA-positive cells. Initially, SMA expression within the thin layer of cells forming the vessel wall early in the hypertensive lung varied in degree and uniformity, but later this became uniform within the developing PSMC population as the wall thickened. A gradient persisted, however, in which the PSMCs closest to endothelial cells demonstrated the greatest degree of filament expression. Although intermediate cells typically had filaments organized in a centrally located domain, the filaments of evolving fibroblasts either were arranged in a central domain or formed a dense mesh.

In the adult hypertensive lung, the recruitment and differentiation of fibroblasts to form vascular smooth-muscle cells resembles that of mesenchymal cells during vessel formation in lung development (12). The expression of SMA alone in these cells, however, does not identify a differentiated smooth-muscle cell phenotype (which requires the sequential expression of other proteins associated with the contractile apparatus). Thus, SMA is followed by expression of SM-22 calponin, h-caldesmon, -tropomyosin, and SMMHC-SM1, with SMMHC-SM2 appearing after birth (6, 13). Additional studies are under way to establish, in relation to cell staging and vessel wall formation, when proteins characteristic of a mature smooth-muscle phenotype are expressed by differentiating PSMCs. A previous light microscopy study has demonstrated SMMHC expression by PSMCs in a model of hypoxic pulmonary hypertension (14). Our preliminary high-resolution data indicate that SMMHC expression is both temporally and spatially regulated within the PSMC population and, of note, we have recently demonstrated that an isoform typical of airway and gastrointestinal smooth-muscle cells (SMMHC-B) is preferentially synthesized (15). The expression of desmin, a cytoskeletal intermediate-filament protein characteristic of vascular smooth-muscle cells, is similarly regulated within the population (16). Early expression of -actin, a nonmuscle actin isoform characteristic of the cytoskeleton of smooth-muscle cells (17), normally highly expressed in the fetal lung and decreased in adult lung, suggests that the sequential expression of proteins by these cells will resemble that in vascular development (18).

In granulation tissue, and in other pathologic states, elegant morphologic and immunocytochemical studies report the development of cells termed myofibroblasts (9). These cells, with their characteristic filament bundles, highly indented nuclei, attachment plaques, and, in some instances, basement membrane, are thought to promote contraction in healing wound tissue. In normal lung, specialized interstitial cells termed "contractile interstitial cells," in which filament bundles traverse the cell process, tethering the cell to the alveolar surface of each side of the alveolar-capillary membrane, have been described (19, 20). These cells normally express desmin but they do not express SMA, and whether these cells or normal fibroblasts are the source of myofibroblasts that develop in lung fibrosis is not known. Expression of cytoskeletal markers demonstrates that in a variety of physiologic and pathologic conditions fibroblasts give rise to myofibroblast subsets that express vimentin alone, vimentin and SMA, vimentin and desmin, or each of these proteins (21). Although the recruited fibroblasts described in this study resemble neither the contractile interstitial cell nor a typical myofibroblast in morphology, our preliminary data indicate that by their expression of contractile and cytoskeletal filament proteins they resemble myofibroblasts by expressing SMA alone, SMA with vimentin or desmin, or each of these proteins (16, 19). Our data indicate, however, that although the differentiating fibroblasts express these proteins, they quickly shift to express a smooth-muscle cell phenotype by the synthesis of SMMHC isoforms. This is consistent with the reported differentiation pathway of mesenchymal cells at other sites (22).

The difference in expression of SMA between fibroblasts remaining in the interstitium of the hypertensive lung and those recruited to the vessel wall is striking. Clearly, the injury that induces vessel wall remodeling via the expression of specific receptor/ligand interactions is differentially orchestrated within the interstitial fibroblast population. Whether, as in development, this is regulated by endothelium, and if so how, remains to be determined. Endothelial cell-derived platelet-derived growth factor (PDGF) and basic fibroblast growth factor (bFGF) stimulate fibroblast chemotaxis and proliferation, and we have demonstrated increased expression of these ligands in our model (23, 24). Release of other mediators, such as endothelin-1, may also stimulate chemotaxis and alignment of these cells, whereas transforming growth factor- (TGF-) expression may induce the expression of a smooth-muscle phenotype as fibroblasts align in close apposition to the endothelium (25). A recent in vitro study of the regulation by endothelial cells of fibroblast (CH310T1/2 cells) chemotaxis, proliferation, and differentiation to a smooth-muscle cell phenotype support this concept (28), although direct evidence that similar events regulate lung cells is needed. The early increase in SMA filament expression in intermediate cells already in close apposition to endothelium may be similarly regulated.

In vitro studies demonstrate that TGF- induces endothelial cell expression of SMA, treatment over several weeks resulting in cells morphologically indistinguishable from smooth-muscle cells (29). In the present study, endothelial cells expressed SMA but expression levels did not increase in the hypertensive lung as in the PSMC population. Although the stress fibers that develop in vitro in endothelial cells, and in certain circumstances in vivo, typically consist of SMA, no stress fibers developed in the vessels we studied. The evident translocation of this protein to the nucleus of these cells, and to the nucleus of PSMCs and interstitial fibroblasts, was an unexpected finding. We initially considered this localization of the gold label an artifact, but consistent labeling in the absence of nonspecific background binding suggested that it was specific, if somewhat difficult to interpret. The import and export of specialized proteins from the nucleus is a well-described process, but less is understood of the expression of other recently described proteins such as bFGF, PDGF, and tumor necrosis factor- within nuclear chromatin, although these are increasingly being identified (30). Even more intriguing is the recent finding that nuclear chromatin is linked via the cytoskeleton to the cell membrane in ways not previously recognized (33), and that the cytoskeletal filaments previously considered to play a role in the structural support of the cell play a role in signal transduction (in the mechanical control of DNA and gene expression and regulation). Given the distribution of SMA we describe, it may be that, as proposed by Desmouliere and Gabbiani (34, 35), this protein has a role in addition to the formation of contractile filaments.