Type II pneumocytes in mixed cell culture of human lung: a light and electron microscopic study.

Alveolar Type II epithelial cells dedifferentiate rapidly in vitro. Studies with animal tissue suggest that cell-cell and extracellular matrix-cell interactions are important in the retention of Type II cell morphology in vitro. Thus, in this study with human tissue, alveolar Type II cells, alveolar macrophages, and spindle cells were prepared from the same sample of lung (obtained following lobectomy for cancer, n = 3), cocultured on glass cover slips or tissue culture plastic, and studied by light microscopy with scanning (SEM) and transmission (TEM) electron microscopy for 8 days. The primary cell isolates contained approximately 45% Type II cells; the remainder were macrophages or unidentifiable cells. Clusters, made up of a single layer of cuboidal Type II cells around a central core of connective tissue (largely collagen and some elastic tissue), formed above a monolayer of spindle cells. The Type II cells were morphologically similar to those seen in vivo. The cells were still cuboidal at 8 days but had lost their lamellar bodies, which were released into the medium via the apical surface. The clusters increased in size with time (area, microns 2: day 1, 29(5-143) x 10(2); day 8, 63(10-311) x 10(2); mean(range); p less than 0.02) without changing in number per culture, suggesting Type II cell proliferation. This may have been due to factors produced by the other cells and adherence to the extracellular matrix (ECM); (free collagen fibers, present in the original preparation, spindle cells, and/or Type II cells could be responsible for presence of ECM). We propose this as a useful model for the study of human Type II epithelial cells in vitro.


Introduction
The alveolarType II pneumocyte is important in maintaining normal lung function, preventing alveolar collapse by the synthesis and secretion ofsurface active material (1,2), acting as a progenitor of alveolar Type I cells that are damaged during lung irjury (3,4), controlling fluid and electrolyte transport (5,6), and producing components of the extracellular matrix (7)(8)(9). The Type II cell, therefore, plays an essential role in the homeostasis of the alveolar unit in health and disease.
Although it is possible to isolate and culture pure preparations of Type II cells from rodent (10,11) and human lung tissue (12), the study of isolated Type II cells is limited since these cells do not retain their cuboidal morphology when plated onto plastic. They lose lamellar body inclusions, their metabolic functions, and they dedifferentiate into intermediate or Type I cells (10,(12)(13)(14). Neither do they divide in isolated cell culture (13,14). The normal morphology of Type II cells in vitro can be prolonged by plating onto exogenous extracellular matrices (ECM), such as laminin (15), collagen gels (16) or basement membrane derived from cultured cells (15,17), and other sources (18,19). Furthermore, macrophageor fibroblast-derived factors are thought to influence the differentiated function of Type H cells in vitro (17,20).
Such studies illustrate the interdependence of different cell types and ECM in normal cell morphology and function. Thus, the development of an in vitro model to study lype II pneumocytes must consider the close proximity of the alveolar epithelial cell to mesenchymal and endothelial cells and the ECM produced by these cells in situ. Although some of these factors have recently been investigated using rat Type H cells (7)(8)(9)15,17,(19)(20)(21), there are few similar studies using human tissue (12). In this preliminar study with human lung tissue, we have studied Type II pneumocytescocultured with spindle cells and macrophages obtained from the same parent tissue-by light and electron microscopy over a period of 8 days Lght Mlicroscopy Cells were identified at the light microscopic level by staining for alkaie phosphatase (AP), cytokeratin, and by the nitroblue tetrazolium (NBT) method.

Alkaline Phosphatase
The modified method of Miller et al. (22) was used for identification of epithelial Type II cells. Cytospun or cultured cells were alowed to dry in air then stained with naphthol phosphate and fast red (10 mg naphthol AS phosphate dissolved in 40 tiL dimethyl sulfoxide then diluted in 10 mL 0.125 M 2-amino-2-methyl-1propanol buffer, pH 8.9, containing 10 mg fast red) for 15 min at room temperature. The stain was washed off with distilled water, and the cells were counterstained in I1% methylene green for 30 sec and mounted in aqueous media.

N'itroblue Tetrazoliuni
The method descnibed by Devereux et al. (23) was used to identify Clara cells. Cytospun or cultured cells were fLxed in 10% buffered formalin for 40 sec. The slides were rinsed in Hepes P13S and incubated in 0.1 % NBTand 0.1% NADPH in PBS at 370C for 10 min. The cells were washed in distffled water, counterstained with 1% methylene green for 2 min, and mounted in glycerol.

Cytoke rati
In order to identify epithelial cells, the cultures were stained for cytokeratin. Cultured cells were fixed for 5 min in methanol and washed in Tween saline (0.15% Tween, 0.15 M NaCI). The cells were incubated with the anticytokeratin antibody (diluted 1:100 with 1 mg/mL bovine serum albumin [BSA] in PBS), for 45 min at room temperature then washed with Tween saline. The preparation was incubated with antimouse IgG conjugated to horseradish peroxidase (diluted 1:100 with I mg/mL BSA in PBS) for 45 min at room temperature and then washed with Tween saline. Following a PB$ wash, the cell preparation was incubated with the substrate (0. 01% H202 and diaminobenzdene in PBS) for 5 to 10 min, until a brown color developed and the reaction te inated by washing in distWled water.

IYansniission Electron Mlicroscopy
Cells were scraped off coverslips and centrifuged at 2509 for 20 min at 4°C. The cell pellet was resuspended in 2% naaraformaldehyde and 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.2, and was left ovemight at 40C. The fixative was removed after centrifugation and the pellet resuspended in cacodylate sucrose buffer. The tissue was postfixed in 1%o osmium tetroxide in 0.1 M sodium cacodylate buffer for 1 hr at 40C. The tissue pellets were then stained with 2% aqueous uranyl acetate for 30 min at room temperature, followed by dehydration to absolute alcohol. Propylene oxide was

Materials and Methods
Tissue culture reagents were obtained from Gibco Ltd.

Humaan Lung Cell Culture
Tissue was obtained from lobectomy specimens removed for lung cancer (n = 3). Tissue wasobtained from sites as far away from the tumor as possible. One or more airways were cannulated and 0.15 M NaCl instilled until the lung was fully inflated. The saline lavage was poured off and retained, and the process was repeated until the number of cells in the last lavage had fallen to less than 5°/% of those in the first lavage. The lung was reinflated by infusion with 0.5°/ trypsin, the cannula was fixed into position, and the lung suspended in 0.15 M NaCI at 370C for 30 min. Any seepage of trypsin through the tissue during the incubation period was replaced via the cannula. Following this, the tissue was chopped into 1to 2-mm pieces and suspended in fetal calf serum (FsCS) and deoxyribonuclease (DNase; 250 mg/mL) to neutralize the enzyme and prevent cell aggregation. The suspension was filtered through sterile guze and-then 150 zm and 30,gm nylon mesh; this resulted in a cell suspension that was then layered onto a Percoll density gradient (1.04-1.09 g/mL) and centrifuged at 250g for 20 min at 4°OC. Fraction 4 from the gradient was collected and diuted with a balanced sat solution containing 5014g/mL DNase and centrifuged (11). The resulting cell pellet was resuspended in 10% I;CS ill low protein hybridoma media containing 100 U/m.L penicili, 100,gg/mL streptomycin, 0.05 mg/mL gentamycin, and 0.3 mg/mL glutamine.
A small portion of the cell suspension was cytospun for staining; the remaining cells were plated at 2 x 106 cells/well into 12-well (25 mm diameter) plastic culture dishes-some of which contained glass coverslips for up to 8 days. The cells were maintained in an atmosphere containing 5/ %CO2 and air, and the medium was changed every 24 hr. At regular intervals the cells were prepared for light and electron microscopy as described below. We were also able to plate cells from one of the tissue samples onto confluent, unpassaged, spindleshaped cells obtained from a previous subject. These were stained for light microscopy at 24 and 48 hr. used as a link reagent before embedding the cells in Taab epoxy resin.
Semithin sections (1-2 jAm) were cut using an ultramicrotome with a glass knife and stained with toluidine blue in borax forligt microscopic exaination. Ultrathin sections (80 nm) of selected areas were mounted on copper grids, electron contrasted with Reynolds lead citrate, and examined with a Philps 201 transmission electron microscope.

ScanniLng Electron Mticroscopy
The cultures that were on tissue culture plastic were placed in Karnovsky fixative overnight at 4°1C, then they were washed in cacodylate-sucrose rinse and dehdae to absolute ethanol via graded alcohols. The preparations were critically point-dried using liquid C02, mounted on stub!s, and sputtercoated with gold. The samples were examined with a Cambridge S200 scanning electron microscope.

Imag,e Anaysis
The slides stained for alkline phosphatase (AP) were photographed, and 2 in. x 2 in. transparencies were prepared. These were projected onto a translucent, sensitized screen of a computerized image analysis system. A digitized pen was used to trace around the perimeter of the aLlkaline phosphatase positive (AP+) clusters of cels and the number of clusters/MM2 of cover slip and cluster area were calculated by reference to a 2 in. x 2 in. transparency of a micrometer scale that had been photographed on the same microscope and at the same magnification as the stained cells.
At 24 hr the AP + cells were forming clusters that had a mean area of 29 x 102,gM2 that increased in size to a mean area of 63 x 102,gM2 by day 8 (Fig lb,c,d; Table   1). The mean cluster size was significantly larger on day 8 compared to days 1, 2, and 3 (p < 0.01; 'Pble 1), but it was not different from day 5. The mean cluster size on day 5 was significantly larger than that on day 2 (p < 0.02; Table 1).
The clusters of cells also stained for cytokeratin. The clusters staining for cytokeratin and AP appeared to overlay negative staining cels which were composed of particle-laden macrophages and spindle-shaped cells (Fig lb-d). The spindle cells increased in number, often reaching confluence. The mean number of clusters per MM2 Of culture remained relatively constant (Table 1).
When confluent human lung spindle cell cultures were plated with the freshly isolated cells used in this study, similar clusters of AP+ cells formed on the surface of the spindle cells. After 2 days the areas of these AP+ cell clusters had signifilcantly increased compared to the areas recorded on day 1 (p < 0.001; lhble 1); single cells and some of the cells on the edges of these clusters appeared to be spreading (I?ig. le, f). Most of the NBT+ cells were located in the same areas as the AP+ cells. Some macrophages also stained with NBT.
TEM of fraction 4 from the density gradient showed the presence of alveolar Type II pneumocytes (-50%n) -often in small chains (Fvig, 2a) -macrophages, some ciliated cels and other cells that could not be identified.
Clara cells were not found. Free collagen fibers were also present (Fvig. 2a). The Type II pneumocytes contained a few recognizable lamellar bodies, as well as vacuoles containing remnants of lamellar bodies. There were many lipid-filed vacuoles (Fsig. 2a). At 24 hr most of the Type II cells were cuboidal with numerous microvill, and the lamellar bodies were large, well-formed, and full of myeli figures. By day 2, some cells were releasing myeli bodies into the medium (Fig. 2b). The lamel.lar bodies were dense and usualy monocentric until day 3 when, although stffl very dense, the number of multicentric lamellar bodies steadily increased (Mg 2c).
The lamellar bodies of the Type II cells were still present at day 5 (Fig 2d), but the lamellar bodies were greatly decreased by day 8 and there were numerous vacuoles, some of which contained remnants of myelin figures (Fig 2e). Lipid vacuoles similar to those observed in the cells prior to culture were present at day 8     FIGURE3. Scanning electron microscopy at day 3 showing cluster of Type II cells, covered in microvilli, above the spindle cell (SC) monolayer.
Some alveolar macrophages (AM) are also present; x 1500. (Figs. 2b,c). In places there was basement membrane that separated the cell from the underlying connective tissue (Fig. 2j). In some Iype II cells, microtubules were seen, and the collagen fibers were intimately associated with the surface membrane of the Type II pneumocyte (Fig. 2g). Throughout the 8 days of culture macrophages were present, but these deteriorated in structure so that by day 5 they were mainly degenerate vacuolated cells (Fig. 2d). Within the clusters of Type II cells there were cells without myelin bodies, that had microvilli, and were cuboidal in shape (Figs. 2d,e). Only one Clara cell was positively identified by electron microscopy in this study.
Scanning electron microscopy (SEM) showed clusters of rounded cells covered with microvilli characteristic of Type II pneumocytes. These were situated above flattened monolayers of spindle shaped cells that were spreading on the culture vessel. Macrophages were also observed above the spindle cells in the proximity of the clusters (Fig. 3).

Discussion
In this study of cocultures of human alveolar Type II pneumocytes, alveolar macrophages, and spindle cells isolated from the same parent tissue, we observed the formation of clusters of lype II cells around a central core of connective tissue and above monolayers of spindle cells. Initially the Type II cells contained lamellar bodies, the contents of which were released via the apical surface into the culture medium. Most of these cells retained their cuboidal morphology during the 8 days of the study, although the number of lamellar body inclusions decreased during this time. As the clusters increased in size without a decrease in cluster number, it is possible that the Type II cells were undergoing cell division in vitro. However, no mitoses were seen, and no further studies have been undertaken to confirm or refute this hypothesis.
The appearance by light microscopy of clusters of rounded AP+ cells above negatively stained spindleshaped cells suggested that the Type l cells had preferentially migrated to these regions and avoided glass and plastic. SEM, which provided a clear three-dimensional picture, confirmed the relationship of the clusters to the spindle cells. TEM confirmed that the cellular component of the clusters largely consisted of cuboidal lype II pneumocytes. Type II pneumocytes have been shown to retain their cuboidal morphology when cultured on various ECM substrates, in contrast to the loss of morphology and spreading seen on tissue culture plastic. In previous studies the substrates used were pure single components of ECM (9,16) or complete ECM from cells of known origin (15)(16)(17)(18)(19). However, in this study the exact source and composition of the extracellular matrix is unclear. The ECM may originally have been the fragments of connective tissue in the primary cell preparation. Alternatively, the spindle type cells, which adhere within 3 or 4 hr of plating, may have laid down sufficient extracellular matrix to later attract and encourage Type II cell aggregation. Spindle cells present in the core of the clusters as unidentifiable cells (Fig. 2d) may have continued to produce ECM. It is also likely that the Type II cells themselves synthesized and secreted connective tissue components. Collagen fibers were intimately associated with the surface of the Type II cells that contained abundant microtubules indicating the synthesis and secretion of macromolecules (Fig. 2g). Furthermore, it appeared that the central core of connective tissue within the clusters was getting larger with time; basement membrane formation, found in the later cultures, was not present at an earlier stage (Fig. 2e). Synthesis of basement membrane (16,19), as well as Type IV procollagen and fibronectin has been demonstrated in cultured rat ype II pneumocytes (7-9).
All these molecules are essential components of the alveolar matrix.
There are at least three possible reasons for an increase in the size of clusters (lible 1) apart from the increase in the central connective tissue core: a) aggregation of Type II cells; b) proliferation of Type II cells; and c) flattening, without dividing, of Type II cells over the surface of the core. There was no evidence that aggregation (i.e., decreased cluster number with an increase in cluster size) was occurring (Thble 1), nor that the Type II cells were becoming flattened (Fig. 2). There is circumstantial evidence that cell division may account for an increase in the size of clusters. That is, over 8 days there was a doubling in the size of the clusters that were usually covered only by a single continuous layer of largely cuboidal cells. The increase in cluster size was more marked when cells were plated onto monolayers, with a 2-fold increase in size within 24 hr. Although some of the latter increase in cluster size may reflect spreading of some of the cells on the edge of the clusters, most of the cells remained cuboidal. In neither situation was there a decrease in the number of clusters per unit area of culture, supporting the suggestion that the cells were dividing. The finding that isolated Type II pneumocytes retain their cuboidal morphology longer when cultured on laminin or more complex ECM of known origin suggests that such compounds may be important in Type II cell proliferation, since cell morphology is thought to be directly related to cell function. In addition, there is evidence that the differentiated function and proliferation of Type II cells is stimulated by factors from macrophages, fibroblasts, and possibly other cell types (17,20,24). In the present study, a combination of coculture of human Type II pneumocytes with alveolar macrophages and spindle cells obtained from the same tissue sample, together with adherence of the Type II cells onto a connective tissue substratum, may have provided a situation in favor of Type II cell proliferation in vitro.
A recent study of rat lype II cells suggests that a pliable substrate assists in the retention of normal Type II cell morphology (17). In the present model, the core of connective tissue may be relatively pliable, being free of the culture vessel and possibly manipulated by the Type II cells that have been shown to degrade ECM in vitro (9). Thus, the Type II cells themselves remodel the connective tissue core that, in turn, promotes their normal morphology. Nevertheless, it is important to remember that a number of studies have shown a retention of Type II cell morphology on nonpliable extracellular matrices, implying that the nature of the substrate is a dominating factor in the expression of cell phenotype.
The NBT stain has been used to differentiate Clara cells from Type II cells isolated from rabbits (23). This present study, using the same modified technique of Devereux et al. (23), shows that human Type II cells and macrophages stain with this dye. NBT+ macrophages may reflect the reduction of NBT by the products of high levels of diaphorases and superoxides produced by activated macrophages from smokers' lungs (25). It is also possible that products of the reaction of high levels of alkaline phosphatase with endogenous or exogenous substrates would reduce NBT to give a positive result (26). This would explain the present observation of NBT+ Type II pneumocytes and macrophages.
Summary An in vitro model of human Type II pneumocytes cultured with spindle cells and macrophages from the same sample of human lung has been established. Clusters of cuboidal Type II cells are formed around a core of connective tissue on the surface of spindle cells. The Type II cells remain cuboidal for 8 days in culture, although the number of lamellar body inclusions decreases. The clusters increase in size; we suggest that this might be due to cell division, possibly under the influence of factors released by the other cells. This model needs to be studied further to establish whether active metabolism is occurring in these cells and whether they are undergoing division.