Rat alveolar type I cells proliferate, express OCT-4, and exhibit phenotypic plasticity in vitro

Abstract

Alveolar type I (TI) cells are large, squamous cells that cover 95–99% of the internal surface area of the lung. Although TI cells are believed to be terminally differentiated, incapable of either proliferation or phenotypic plasticity, TI cells in vitro both proliferate and express phenotypic markers of other differentiated cell types. Rat TI cells isolated in purities of >99% proliferate in culture, with a sixfold increase in cell number before the cells reach confluence; >50% of the cultured TI cells are Ki67+. At cell densities of 1–2 cells/well, ∼50% of the cells had the capacity to form colonies. Under the same conditions, type II cells do not proliferate. Cultured TI cells express RTI40 and aquaporin 5, phenotypic markers of the TI cell phenotype. By immunofluorescence, Western blotting, and Q-PCR, TI cells express OCT-4A (POU5F1), a transcription factor associated with maintenance of the pluripotent state in stem cells. Based on the expression patterns of various marker proteins, TI cells are distinct from either of two recently described putative pulmonary multipotent cell populations, the bronchoalveolar stem cell or the OCT-4+ stem/progenitor cell. Although TI cells in adult rat lung tissue do not express either surfactant protein C (SP-C) or CC10, respective markers of the TII and Clara cell phenotypes, in culture TI cells can be induced to express both SP-C and CC10. Together, the findings that TI cells proliferate and exhibit phenotypic plasticity in vitro raise the possibility that TI cells may have similar properties in vivo.

the alveolar epithelium, which comprises >99% of the internal surface area of the lung, is composed of two types of cells, type I (TI) and type II (TII) cells. TI cells are large (50- to 100-μm diameter) squamous cells that cover 95–99% of the alveolar surface area (47). TII cells, which cover the remainder of the alveolar surface, are smaller (∼10 μm in diameter), cuboidal cells best recognized as the source of pulmonary surfactant. The currently accepted paradigm regarding alveolar epithelial cell lineage and differentiation is that TII cells have the capacity to proliferate and transdifferentiate into TI cells, whereas TI cells are terminally differentiated and do not have the capacity to proliferate. This paradigm is based on the classic studies of Evans et al. (14) and Adamson and Bowden (1) who used techniques of autoradiography in oxidant-injured rodent lungs to elucidate the events that occurred during lung injury caused by oxidant gases and repair. Following exposure to oxidant gases, TI cells in terminal bronchiolar locations were damaged, and severely injured TI cells were sloughed from the basement membrane. Some TII cells divided; cell progeny either retained morphological characteristics of TII cells or flattened and spread, acquiring morphological characteristics of TI cells. Although in these studies a small percentage of TI cells took up [³H]thymidine, it was concluded that the TI cell is “terminally differentiated,” a concept supported by Weibel (52) from perceived structural constraints on the TI cell. These paradigms have shaped our current thinking about alveolar epithelial repair but need to be reconsidered in view of the experimental design and the advances in experimental methodology that have become available since the time this work was done (for more details, see discussion). Although it has been suggested that TI cells can help to repair lung injury (31), this has not been a widely accepted concept.

Techniques to obtain highly purified TI cells have been developed in the last few years, permitting direct study of this cell type (5, 10, 26, 27). In this article, we report that cultured TI cells have the capacity to proliferate until they are contact inhibited. More than 50% of cultured TI cells are positive for Ki67, a proliferation antigen (16, 17, 43). When plated at cell densities of 1–2 cells/well, ∼50% of the cells had the capacity to form colonies. Freshly isolated and cultured TI cells express OCT-4A, a protein that is involved in establishing and maintaining the undifferentiated pluripotent state. Finally, cultured TI cells do not appear to be terminally differentiated. When cultured on fibronectin in the presence of FBS, TI cells continue to express RTI40 (rat podoplanin) and aquaporin 5, two markers associated with the differentiated TI cell phenotype. However, under other culture conditions, TI cells can be induced to express surfactant protein C (SP-C), a marker of the TII cell phenotype, and CC10, a marker of Clara cells, thereby demonstrating phenotypic plasticity. That TI cells in vitro proliferate and can be induced to express phenotypic markers of other differentiated epithelial cells is consistent with the concept that TI cells may have the capacity to proliferate in vivo and to participate in processes of lung repair after injury. A portion of these results has been reported in abstract form (18).

MATERIALS AND METHODS

Isolation of TI and TII cells by flow cytometry.

Alveolar epithelial cells were isolated from the lungs of pathogen-free Sprague-Dawley rats (100–150 g) by elastase digestion as described previously (10). Digested lungs were minced to ∼1 mm³ pieces in solution A [RPMI-HEPES containing 50% FBS (Hyclone, Cell Culture Facility, UCSF) and DNase (50 μg/ml; Sigma, St. Louis, MO)] at 4°C, and single cells were separated from large aggregates by successive filtration through 100-, 40-, and 20-μm nylon mesh filters (Tetko; Depew, NY). Rat IgG (50 μg/ml, Sigma) and mouse IgG (50 μg/ml, Sigma) were added to the cell suspension and incubated at 4°C for 10 min to saturate cell surface IgG binding sites. All animal protocols were approved by the UCSF Animal Care and Use Committee.

To isolate both cell types by flow cytometry, we used antibodies specific within the lung for TI [anti-RTI40, an IgG₁ monoclonal antibody (13)] or TII cells [anti-RTII70, an IgG₃ monoclonal antibody (11)]. We directly conjugated anti-RTI40 to Alexa 610-RPE using the Zenon technology (Invitrogen, Carlsbad, CA). Because this technology is not available for antibodies of the IgG₃ subclass and because other methods of direct conjugation destroyed the immunogenicity of anti-RTII70, we used Alexa 488 anti-IgG₃ as a secondary antibody to visualize RTII70. Anti-RTII70 was added before centrifugation at 300 g for 12 min on a 50-μl Percoll (Invitrogen) cushion. The cell pellet was resuspended in solution A, and both goat anti-IgG₃ Alexa 488 (Invitrogen) and Alexa 610-RPE anti-RTI40 were added. The antibodies were incubated with the cells for 10 min and centrifuged at 300 g for 12 min. The cell pellet was resuspended in solution A. Differentially labeled TI and TII cells were sorted using a FACS Aria (BD Biosciences, San Jose, CA). Cell sorting results were analyzed using BD Diva software. Alveolar type I and type II cell purities were greater than 99% by immunofluorescence using anti-RTI40 and anti-RTII70. TI cells are large and fragile; viability by fluorescein diacetate was 85–90%. Yields varied between 2 and 6 × 10⁵ TI cells/rat.

Cell culture.

Alveolar TI and TII cells were cultured in DME-H16 containing 20% FBS and 50 μg gentamicin/ml in tissue culture six-well Transwell plates (Corning, NY) previously coated with bovine fibronectin; cultures were maintained in a 10% CO₂/air incubator. For growth curves and for the images in Fig. 2, cells were seeded at a density of 4.4 × 10³ cells/cm². At various time points, duplicate wells were trypsinized, and cells were counted using a hemacytometer.

Fig. 2.

Type I cells gradually adhere to cell surfaces for several days before they begin to proliferate. A–F show Hoffman modulation optical contrast images of TI cells in tissue culture. In the first 24 h of culture, TI cells adhere to the surface of the plate. A: at 24 h, cells are adherent but remain round and compact. B: by 48 h, many cells are still rounded; others have begun to spread. Over the next several days (C–E), TI cells continue to adhere, spread, begin to extend filopodia and lamellipodia (arrows), and become motile. By day 12 (F), cells have divided, and cultures are mostly confluent and contact inhibited. A–D and F are at the same magnification; E is a higher magnification view of the boxed area in D. One can appreciate the large changes in morphology that occur as cells spread, adhere, migrate, and divide.

To determine the temporal course of TI cells in culture, we photographed cells daily for the first 7 days.

To examine the potential for clonal growth, we plated TI cells by limiting dilution in 48-well plates to achieve cell densities of 1 or 2 cells/well. Cells were obtained from two different cell isolations; a total of 576 individual wells were inspected and followed with time in culture. We identified 68 wells containing 1 cell/well and 96 wells containing 2 cells at 48 h. Wells were monitored daily, and the number of wells containing colonies and number of colonies per well were determined at 7 days.

In separate experiments, TI cells were cultured under conditions that favored the expression of the type II cell phenotype, using a modification of conditions described by Shannon et al. (44) and Sugahara et al. (49). For these studies, proliferating TI cells were overlain with EHS matrix and cultured for 5 more days (Collaborative Biomed, Bedford, MA) in DME media (2:1, EHS: DME) containing 1% rat serum, 10 ng KGF/ml (R&D Systems, Minneapolis, MN), 10⁻⁴ M 8-bromocyclic AMP (Sigma), and 50 μg gentamicin/ml.

Mouse ES cells (F1 BL6/129v ES) were cultured on inactivated fibroblasts as described by Hawgood et al. (23).

The X 404 cell line, an adenocarcinoma cell line derived from X-irradiated rat lungs, was the kind gift of Dr. Stephen Belinsky (Lovelace Institute, Albuquerque, NM).

Calculation of cell surface area.

Virtually confluent cultures of TI cells were trypsinized from the surface of tissue culture dishes and counted. The surface area of the tissue culture dish was divided by the number of cells. Because cultures are almost, but not totally, confluent, the calculated surface area may be somewhat smaller than calculated by this method, but should not be markedly different.

Immunocytochemistry.

Freshly isolated or cultured cells were deposited onto glass slides using a Shandon Cytospin (Thermo Fisher Scientific, Waltham, MA). Samples were fixed overnight in freshly prepared 4% paraformaldehyde in PBS (fixative A) at 4°C. Cells were permeabilized by exposure to 0.5% Triton X-100 in PBS for 30 min. Slides were incubated with 2% goat or donkey serum in PBS in all reactions to block nonspecific binding of antibodies. Cells were stained with antibodies against RTI40, RTII70, prosurfactant protein C (AB3786; Chemicon International, Temecula, CA), aquaporin 5 (Alpha Diagnostic, San Antonio, TX), OCT-4 (H-134; Santa Cruz Biotechnology), or CC10 (S-20 and T-18, Santa Cruz Biotechnology). Samples were incubated with primary antibody (30 min at room temperature) followed by the appropriate secondary goat anti-mouse, goat anti-rabbit, donkey anti-rabbit, or donkey anti-goat IgG Alexa 350-, Alexa 488-, or Alexa 594-conjugated secondary antibodies at 1:3,000 (Invitrogen). Nuclei were visualized with DAPI (Sigma). For Ki67 staining, fixative A was directly added to culture well slides (Lab-Tek Chamber Slides; Nunc, Naperville, IL) containing cells, and the samples were fixed overnight at 4°C and then permeabilized with methanol:acetone (2:1) at −20°C for 5 min. Anti-rat Ki-67 antibody (clone MIB-5; Dako, Carpinteria, CA) was added at 1:500 dilution followed by Alexa 488 goat-anti-mouse IgG_2a secondary antibody (Invitrogen). Slides were mounted in Prolong (Invitrogen) and examined on a Leica Orthoplan fluorescence microscope. Each fluorescent signal was captured in its separate channel using a Leica DC500 digital camera. Images were merged in Adobe Photoshop CS to produce multicolored composite images. Living cultures were examined on a Nikon Eclipse TE 300 inverted microscope equipped with Hoffman Modulation Contrast Optics and a heated stage; images were captured with a Diagnostic Instruments Spot RT camera. Controls without primary antibody or with irrelevant primary antibodies (rabbit IgG; Calbiochem, San Diego, CA; purified mouse and goat IgG, nonspecific IgG1) were performed with every reaction and did not display fluorescence that was above background (data not shown). For determination of the percentage of Ki67+ cells, we counted a total of 10,000 cells (freshly isolated), 246 cells (48 h of culture when the density was sparse, >100 separate fields were counted), and 1,000 cells (after 7 days in culture).

Real-time PCR.

We performed real-time PCR as previously described. RNA was purified from individual TI, TII, or cultured TI cell preparations using Qiagen RNeasy Total RNA Isolation Kit (Valencia, CA) and digested with DNase (Qiagen). RNA was reverse transcribed using RETROscript reagents (Ambion, Austin, TX). Quantitative real-time PCR (Q-PCR) amplification of cDNA was performed using an ABI PRISM 7700HT Sequence Detector System with a 384-well block (Applied Biosystems, Foster City, CA). Reaction mixtures consisted of 5 ng of cDNA, TaqMan 2X Universal PCR Master Mix (Applied Biosystems), forward primer, reverse primer, and probes in a reaction volume of 10 μl. Using a two-step PCR program, we heated samples to 95°C for 10 min followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Differences in the amount of cDNA amplified were normalized to endogenous levels of ribosomal RNA (TaqMan Ribosomal RNA control reagents, Applied Biosystems). Standard curves for test genes and 18S ribosomal RNA were constructed on each assay plate from serial log dilutions of stock whole lung cDNA or ES cell cDNA containing the cDNA insert of interest; the relative quantification in triplicate for each experimental sample was obtained by the standard curve method. Control reactions were performed without reverse transcriptase and in the absence of target DNA.

Q-PCR primers and probes.

Rat RT1-40: forward CAGTGTTGCTCTGGGTTTTTGG, reverse AGACCTGGGTTCACCATGTCA, probe FAMATGGCCCCTCCCTGCGCTGATTAMRA; AQP5: forward CCCTGCGGTGGTCATGAA, reverse AGCATGGCCCCCACAATAG, probe FAMCTCTCACTGGGTCTTCT; OCT-4A: forward GGCTGGACACCTGGCTTCAGA, reverse TGGTCCGATTCCAGGCCCA, probe FAMCCTCGAACCTGGCTAAGCTTCCATTAMRA; rat SP-B: forward GCTGAGCGTTACACAGTACTTCTAC, reverse ACCAGGCCACAGACTAGCT, probe FAMACCCAGCAGTGCATCTA; rat SP-C: forward GCTCCTGACCGCCTATAAGC, reverse CTCTCTGGAGCCATCTTCATGAT, probe FAMAGCTCCAGGAACCTAC.

RESULTS

Homogeneity of TI and TII cell isolates.

We used FACS analysis of differentially labeled TI (Alexa 610-RPE anti-RTI40) and TII (Alexa 488 anti-RTII70) cells to obtain very pure populations of cells (Fig. 1). The mixed cell preparation before sorting contained many different kinds of lung cells, including TI cells, TII cells, and Clara cells (Fig. 1A). After FACS sorting, we obtained three different populations of cells: P1, consisting of Alexa 610-RPE+ (TI) cells; P2, composed of Alexa 488+ (TII) cells; and P3, a population that was both Alexa 610-RPE+ and Alexa 488+. The TI cell preparations (Fig. 1C) contained <0.3% TII cells; the TII cell preparations (Fig. 1D) contained <0.2% TI cells. The P3 population expressing both markers consisted of clumps of TI and TII cells (Fig. 1B), rather than cells coexpressing both antigens, which were observed only on rare occasions.

Fig. 1.

Cells isolated by FACS. Lung cells liberated by digestion with elastase were stained with Alexa 610-RPE anti-RTI40 (TI cell marker, red), Alexa 488 anti-RTII70 (TII cell marker, green), and CC10 [Clara cell (CC) marker, blue], as described in text. A: mixed cell population before FACS analysis and sorting, showing alveolar type I (TI) cells (red), TII cells (green), and Clara cells (blue). FACS scatter graph is shown (top right). Three populations were selected: P1, high Alexa 610-RPE- and low Alexa 488-staining (TI cells), P2, low Alexa 610-RPE- and high Alexa 488-staining (TII cells), and P3, high levels of both Alexa 610 and Alexa 488 staining (cell aggregates containing both TI and TII cells). B–D: cytocentrifuged preparations of P1, P2, and P3 populations. B: P3 population, demonstrating that the population of cells expressing both RTI40 and RTII70 is composed of aggregates of TI and TII cells, rather than cells coexpressing both markers. C: P1 population, essentially pure TI cells. D: P2 population, >99% TII cells.

All of the isolated cells expressed AQP-5, another established marker for TI cells in rodent lung (38) (see Fig. 7, A and B). By Q-PCR, the preparation expressed large amounts of mRNAs for aquaporin 5 and RTI-40; the amount of SP-B mRNA was 1.2% and SP-C mRNA was 1.4% of that found in isolated TII cells, values similar to those obtained by comparative immunofluorescence. Isolated TII cells expressed a somewhat higher percentage of AQP-5 and RTI-40 than we would anticipate from the comparative immunofluorescence and cell-sorting data. TII cells are known to express both of these markers when placed in culture (14, 19) and may start to express these from the time the cells are isolated.

Fig. 7.

Both freshly isolated and cultured TI cells express aquaporin 5, a marker of the TI cell phenotype. A and B: paired immunofluorescence and phase-contrast views of freshly isolated TI cells demonstrating expression of aquaporin 5. C and D: TI cells were cultured on fibronectin, and immunocytochemistry for aquaporin 5 was performed as described in text. Paired immunofluorescence and Hoffman modulation contrast optical images of the same group of confluent cells.

TI cells proliferate in vitro.

Figure 2 depicts the behavior of TI cells for the first several days of culture. After 24 h in culture (Fig. 2A), TI cells remain rounded, although adherent to the plates; nonadherent cells were removed. At 48 h in culture (Fig. 2B), some cells remain rounded; others have begun to spread on the tissue culture surface. By 72 h (Fig. 2C), some cells send out long filopodia and form lamellipodia, and by 96 h (Fig. 2D), most cells exhibit these morphological characteristics and become highly motile. By day 7, cells are actively dividing.

Figure 3A shows a growth curve of TI cells in vitro. Cells were plated at a density of 2 × 10⁴ cells/well. After the initial lag period during which cells adhered, spread, and became motile, TI cells started to proliferate. TI cell doubling time estimated from the logarithmic phase of the growth curve was ∼36–60 h until the cells reached confluence, at which time they were contact inhibited. At confluence, cultured TI cells were large and highly pleomorphic, with an average surface area/cell of 5,270 sq μm. The morphological appearance of cultured TI cells (Fig. 2, Fig. 3B, Fig. 4, Fig. 7) was markedly different from TII cells cultured under the same conditions (Fig. 3C).

Fig. 3.

TI cells proliferate in culture under conditions in which TII cells do not proliferate. Cells were plated at densities of 2 × 10⁴ cells/transwell under conditions described in materials and methods. At various time points, duplicate samples were taken, and cells were counted with a hemacytometer. Growth curves of TI and TII cells from 1 representative experiment of 3 are shown in A; TI cells proliferate until they are contact inhibited. Under the conditions of the experiment shown here, the number of TI cells increases ∼6-fold. Hoffman modulation contrast optical images of TI cells (B) and TII cells (C) cultured for 7 days. TI cells are very large pleomorphic cells that become contact inhibited. TII cells are smaller cells that do not proliferate and assume a flattened, “fried-egg” appearance. The marker bar of 100 μm applies to B and C.

Fig. 4.

Mitotic TI cells. TI cells were labeled with RTI40 conjugated to Alexa 610-RPE at the time of isolation and placed in culture under conditions described in text. The fluorescent signal present when cells are isolated persists in culture and can be used to identify and track TI cells over time. At day 5 in culture, dividing TI cells were followed by time-lapse photography over a period of 4 h. A: 4 TI cells with staining for Alexa 610-RPE anti-RTI40. The arrow designates a TI cell that will divide, as shown in B–I. I–N: the same group of cells after cell division, showing: phase-contrast view (I; arrows denote daughter cells); nuclear staining (J; blue, DAPI) demonstrating 5 nuclei; original staining (K; red) for RTI40 from the time of isolation, a punctate pattern consistent with an intracellular location; cultured cells stained with Alexa 488 anti-RTI40 (L; green) after mitosis. The more homogeneous cellular staining pattern clearly demarcates lamellipodia and filopodia. This staining pattern is consistent with a plasma membrane location and is compatible with synthesis of RTI40 in culture. The newly divided TI cells are rounded (see I and M), with fewer lamellipodia; M: higher magnification of image (I) in which cell borders can be more clearly visualized; N: higher magnification view of a merged color image of J–L, with nuclear staining (blue), 610-RPE anti-RTI40 (red) from the time of cell isolation, and Alexa 488 anti-RTI40 (green) staining after cells were cultured for 5 days. The 2 different RTI40 cellular localization patterns are highlighted in the merged image.

Although the initial cell preparations contained >99% TI cells, it is important to ascertain whether TI cells actually proliferated. RTI40 is an integral apical plasma membrane protein that, within the rat lung, is a marker for TI cells (13, 21). We used immunoselection with directly conjugated Alexa 610-RPE anti-RTI40 to isolate TI cells. In freshly isolated cells, the pattern of cellular RTI40 staining is consistent with the known plasma membrane localization of this protein (13) (Fig. 1C). With time in culture, the homogenous cellular staining pattern becomes punctate (Fig. 4, A and K), presumably from internalization of the marker. The continuing presence of fluorescence can be used to identify and track TI cells from the time of isolation for several days. Figure 4 shows a time-lapse series of images of a cluster of TI cells labeled with Alexa 610-RPE anti-RTI40 at the time of isolation; the images capture a TI cell undergoing mitosis (Fig. 4, A–J), resulting in two daughter cells after division (Fig. 4, I–M). TI cells synthesize both RTI40 (Fig. 4L) and aquaporin 5 (Fig. 7, C and D) after division.

Expression of Ki67, a cell proliferation antigen.

Ki67 is a nuclear antigen associated with cell proliferation; it is expressed in G₁, S, G₂, and M phases, but is not expressed in resting cells in G₀ (17, 42). A small percentage (∼0.05%) of freshly isolated TI cells express Ki67 (Fig. 5A). By 48 h in culture, 51% of the adherent cells are Ki67+ (Fig. 5B). After 7 days in culture, >50% of the cultured TI cells were Ki67+ (Fig. 5C). The inset in Fig. 5C shows two TI cells in anaphase stained with Ki67.

Fig. 5.

TI cells express Ki67, a proliferation antigen. Immunocytochemistry of both freshly isolated TI cells (A) and cultured TI cells (B and C) for RTI40 (red) and Ki67 (green) was performed as described in text. The percentage of cells expressing Ki67 increased dramatically in culture: 0.05% of freshly isolated cells (A), 51% of cells after 48 h of culture (B), and 53% of cells after 7 days (C) expressed Ki67. Inset: higher magnification view of cultured cells showing dividing TI cells in anaphase. To determine the percentages of Ki67+ cells, we counted 10⁴ freshly isolated cells, 246 cells at 48 h (the cell density is low, >100 different fields were counted), and 1,000 cells after 7 days in culture.

TI cells exhibit clonal growth capacity.

To determine the capacity of individual TI cells to proliferate, we plated cells by limiting dilution and counted the number of cells in each well at 24–48 h (n = 2 experiments). We followed 288 wells over time, identifying wells that contained 1 or 2 cells/well at 48 h. Individual wells were evaluated for clonal growth. Data are shown in Fig. 6. In 68 wells that contained one cell at 48 h, there were 34 wells (50%) that contained proliferating cells (34 colonies) at 7 days; in 96 wells containing two cells at 48 h, 69 wells (72%) contained 89 colonies (46% of the total number of cells) of proliferating cells at 7 days. An image of representative wells containing clones is shown in Fig. 6. These data are consistent with the data from Ki67 staining that ∼50% of the TI cells proliferate in vitro.

Fig. 6.

Clonal growth of TI cells plated at densities of 1 or 2 cells/well. Cells were plated by limiting dilution, and wells were carefully inspected at 48 h to determine the number of adherent cells/well. After 7 days, we determined the number of colonies in those wells that had previously been identified as containing 1 or 2 cells (shown in Table 1). Right: image of wells containing colonies at 7 days. Cells were fixed and stained with Coomassie blue, and the plates were photographed from the bottom of the wells.

Table 1.

Relative expression by Q-PCR of marker mRNAs for TI and TII cells

Cell Type	SP-B	SP-C	Aquaporin-5	RTI40
TI cells	1.2	1.4	100	100
TII cells	100	100	2.6	4.6

Q-PCR was performed as described in materials and methods. TI, alveolar type I; TII, alveolar type II.

TII cells cultured under the same conditions do not proliferate.

Under the culture conditions we used for TI cells, TII cells do not proliferate (Fig. 3). We (unpublished observations) and others (35) have observed that TII cells cultured on tissue culture plastic in the presence of FBS have a very low proliferative capacity.

Type I cells in culture continue to express both RTI40 and aquaporin 5.

After 5 days in culture and division, TI cells contain detectable Alexa 610-RPE anti-RTI40 (red) in a punctate staining pattern compatible with an intracellular localization (Fig. 4, A, K, N). When the same cells are restained using a different chromophore (Alexa 488, green) conjugated to anti-RTI40, the green staining pattern differs (Fig. 4, L and N) from that of the original Alexa 610-RPE anti-RTI40. This pattern is more homogeneous and diffuse over the entire cell, including the thin lamellipodia; the pattern is similar to the staining in freshly isolated TI cells (Fig. 1C) and consistent with a plasma membrane localization. These results are compatible with the hypothesis that ongoing synthesis of RTI40 occurs in vitro. Cultured TI cells also continue to express aquaporin 5 (Fig. 7C), which is found in rat TI cells in situ (38).

Expression of OCT-4 in freshly isolated and in cultured TI cells.

The OCT-4 gene encodes a protein that belongs to a family of transcription factors containing the POU DNA binding domain. OCT-4 (POU5F1) is found in pluripotent stem cells of early embryos and is important in maintaining embryonic stem cell pluripotency (reviewed in Ref. 4). OCT-4 staining was used by Ling et al. (36) as one characteristic to identify a putative pulmonary stem cell population. In humans, OCT-4 exists as two isoforms, resulting from alternative splicing. The OCT-4A isoform, localized mainly in the nucleus, is believed to confer pluripotency. The OCT-4B isoform, which is cytoplasmic in location, has uncertain biological functions (30). To detect OCT-4, we used the H134 antibody (Santa Cruz Biotechnology) against the 134 amino acid COOH terminus of OCT-4 (32). Most freshly isolated TI cells express OCT-4; expression is predominantly in the nucleus, although there is some cytoplasmic staining as well (Fig. 8, A and B). In cultured TI cells, OCT4A is expressed almost completely in the nucleus (Fig. 8D). Embryonic stem cells are shown for comparison (Fig. 8, E and F). In Western blotting (Fig. 8G), we find a band of ∼50 kDa, consistent with OCT-4A and in contrast with OCT-4B, which has an apparent molecular mass of ∼30 kDa (Ref. 2 and data from manufacturer). By Q-PCR with probes for the COOH terminus region specific for OCT-4A, we demonstrated OCT-4A mRNA expression in both freshly isolated and cultured TI cells; the levels of OCT-4A mRNA in freshly isolated TI cells were approximately one-half that found in X 404 cells (Fig. 8H), an adenocarcinoma line derived from rat lung that expresses RTI40.

Fig. 8.

Both freshly isolated and proliferating TI cells express OCT-4. A and B: paired images of freshly isolated TI cells. In A, freshly isolated TI cells stained for RTI40 (red), DAPI (blue), and OCT-4 (green). B: only the OCT-4 channel is shown. OCT-4 staining is both nuclear and cytoplasmic. C and D show a similar pairing of TI cells cultured for 10 days, as described in text; cultured TI cells are very large. In C, cells are stained for RTI40 (red), DAPI (blue), and OCT-4 (green). D shows the same field with only the OCT-4 channel. In cultured TI cells, OCT-4 is localized primarily to the nucleus. E and F: paired images of cultured embryonic stem cells, stained for DAPI (blue) and OCT-4 (green); only the OCT-4 channel is shown in F. The magnification is the same in all panels. G: Western blot of freshly isolated TI cells showing a band at ∼52 kDa, an apparent molecular weight appropriate for OCT-4A. One of two replicates are shown. H: Q-PCR analysis of OCT-4A expression in freshly isolated TI cells (TID0), TI cells cultured for 7 days (T1D7), and, for comparison, X 404 cells, an adenocarcinoma line derived from X-irradiated rat lungs (n = 2, means ± average).

TI cells cultured in Matrigel express SP-C and CC10, markers of the TII and Clara cell phenotypes.

Matrigel was applied to proliferating TI cells, as described in materials and methods. After 5 days, cells were removed from the matrix and stained for SP-C (Fig. 9, A and B) and CC10 (Fig. 9, C and D). Most of the cells expressed SP-C; a subset of cells also expressed CC10. Controls with an irrelevant primary antibody (normal rabbit and goat IgG) were negative (Fig. 9, E–H).

Fig. 9.

Proliferating TI cells can be induced to express SP-C and CC10, markers of the type II and Clara cell phenotypes. TI cells were allowed to proliferate on a matrix of fibronectin in the presence of FBS and were then overlain with Matrigel matrix with rat serum and cultured for 5 days, as described in the text. Cells were then removed from the matrix and cytocentrifuged before fixation and staining. Fluorescence (A) and gray-scale (B) images of TI cells stained for SP-C. Fluorescence (C) and gray-scale (D) images of TI cells stained for CC10. E–H: staining using normal rabbit and goat IgG as irrelevant primary antibody controls. Fluorescence (E and F) and gray-scale (G and H) images of cultured TI cells, demonstrating background staining. These results demonstrate that TI cells cultured on EHS matrix can be induced to express SP-C and CC10, demonstrating the phenotypic plasticity of this cell type.

DISCUSSION

The currently accepted paradigm regarding alveolar epithelial cell lineage and differentiation is based on the studies of Evans et al. (14, 15) and Adamson and Bowden (1). Utilizing autoradiography in oxidant-injured rodent lungs, these investigators described a series of events that occurred during lung injury and repair. Following exposure to oxidant gases, the majority of cells labeled with tritiated thymidine were TII cells. From autoradiographic ultrastructural data, some TII cells divided. The cell progeny either retained morphological characteristics of TII cells or flattened and spread, acquiring morphological characteristics of TI cells. Because the majority of labeled cells were TII cells and because of the time course of labeling, it was concluded that TII cells were progenitor cells of both TI and TII cells and that TI cells did not proliferate. Although Kyono et al. (31) have suggested that TI cells could help to repair lung injury, this has not been a widely accepted concept.

Although these observations in oxidant-injured lungs cited above have shaped our current thinking about alveolar epithelial cell lineage, the interpretation of the results that precludes a progenitor role for TI cells seems overly restrictive for several reasons. First, a small percentage of TI cells were labeled within 1 h of administration of [³H]thymidine (1, 15), consistent with the concept that some TI cells have proliferative potential. Second, in model systems using these concentrations of oxidant gases, TI cells are severely injured (46), making it difficult to detect a possible role of this cell type in repair. Third, injury is limited to the terminal bronchiolar/alveolar junction area of the lung, with sparing of more distal alveoli (37, 45). It is possible that repair may occur by different mechanisms depending on the microenvironment of the particular anatomic niche. Fourth, the interpretation of these results is limited by the methodologies available at the time the studies were performed. Because the work antedated the discovery of biochemical markers specific for alveolar cells, cell phenotype was determined solely by morphological criteria, which, for TI cells, mandated electron microscopic analysis. The experimental design would have been biased against identifying proliferating or transdifferentiating TI cells because of the small sample area in electron microscopic studies and the extraordinarily large size of TI cells. Finally, cell lineage was determined by autoradiography; any changes in cellular phenotype that occurred without cell division would not have been detected.

Rodent TII cells do not proliferate under conventional tissue culture conditions of a tissue culture plastic or fibronectin matrix and FBS (7, 34, 35, 51), although they can be induced to proliferate under specialized culture conditions. Therefore, we were initially surprised to find that highly purified populations of rat TI cells, which were not thought to have proliferative potential, in fact have the capacity to proliferate in vitro. The number of TI cells in our cultures increased approximately sixfold before cells reached confluence and became contact inhibited (Figs. 3B and 7D). Interestingly, when TI cells reach confluence (Fig. 3B and 7D), their average surface area is ∼5,270 sq μm, similar to the surface area of TI cells in Sprague-Dawley rats in situ (5,320 sq μ) (48). Although TII cells can be induced to proliferate at low density when cultured with a cocktail of various growth factors such as a cocktail of EGF, FGF, cholera toxin, insulin, and concentrated bronchoalveolar lavage (34), TII cells do not proliferate when cultured in FBS in the absence of these factors (7, 34, 35, 51). To culture TI cells, we used a higher concentration of FBS (20%) than the conventional 10% FBS used in the previously cited studies. However, when we cultured TII cells under identical conditions to that used for TI cells, TII cells did not proliferate (Fig. 2, A and C). When cultured on Matrigel in the presence of KGF, Zhang et al. reported TII cells exhibited a 50% increase in DNA/well, which may correlate with a small degree of proliferation (55). This figure is lower than the sixfold increase we observed for TI cells in vitro.

The low mitotic index of freshly isolated TI cells suggests that most TI cells in the lung are quiescent at any one time and would explain, along with the very large surface area of this cell type, why it is difficult to demonstrate mitotic TI cells in situ. There are ∼80 × l0⁶ TI cells in the Sprague-Dawley rat lung (48); if the mitotic index of the TI cell in situ is similar to that of the freshly isolated cells (0.05%), one may infer that ∼40,000 TI cells at any one time are in a phase of the cell cycle other than G₀. In contrast to the low observed mitotic index of cultured TII cells (34), more than 50% of cultured TI cells (Fig. 5, B and C) express Ki67, a proliferation antigen (16, 17, 43). Ki67+ is expressed at 48 h, prior to a time that proliferation is observed. When cultured at densities of 1–2 cells/well, ∼50% of the TI cells can proliferate and form colonies. Based on these data, it appears that at least 50% of the freshly isolated TI cells have proliferative capacity; the percentage of TI cells that can proliferate may be higher than this estimate. Lack of Ki67 expression does not indicate a lack of proliferative potential, but indicates only that a cell was in G₀ at the time of fixation. Ki67 negative cells could be in G₀ either before or following proliferation. The clonal density experiments are subject to the problem that primary epithelial cells plated at very low densities do not appear to do as well in culture as cells plated at higher densities. Finally, it is possible that TI cells are subtly damaged by the isolation procedure and that this is not reflected in vital dye staining. The cell isolation method involves repeated centrifugation steps, and FACs sorting involves passing these large, fragile cells through small orifices with high shear forces. However, it may be that only a subpopulation of TI cells has the capacity to proliferate under the conditions we have examined, although this does not rule out the possibility that such cells that do not divide under one set of conditions could proliferate under different conditions.

TI cells express OCT-4, a member of the family of transcription factors containing the POU DNA binding domain. OCT-4 is found in pluripotent stem cells of early embryos, where it is important in maintaining embryonic stem cell pluripotency. The striking observation that OCT-4-transfected keratinocytes could acquire characteristics of neuronal cells when exposed to the appropriate stimuli (22) appeared to confirm the concept that OCT-4 might be a “master regulator” of the pluripotent state, although OCT-4 by itself was not sufficient to induce pluripotency in fibroblasts (50). Recent experiments have shown that OCT-4 is a critical factor that can mediate, along with other factors, transformation of somatic cells to a stem cell phenotype (25, 40, 41). The extent to which OCT-4 confers pluripotency has been the subject of recent controversy (30, 33, 54). One possible reason for this controversy is that there are two isoforms of OCT-4, OCT-4A and OCT-4B. Human OCT-4B, in contrast to OCT-4A, is mainly found in the cytoplasm rather than in the nucleus. Furthermore, OCT-4B does not bind to a probe carrying the OCT-4 binding sequence, does not activate transcription in OCT-4-dependent promoters, and does not sustain ES cell renewal (32). Therefore, it is important to distinguish OCT-4B from OCT-4A expression, because the biological function of OCT-4B is unknown. Although there may be species differences in isoforms of OCT-4, the rat database contains a sequence that is 83% homologous with the 133-amino acid OCT-4A-specific sequence found in human OCT-4A. To detect OCT-4, we used the H134 antibody (Santa Cruz Biotechnology) against OCT-4 that recognizes an epitope within the first 134 amino acids of OCT-4 specific to human OCT-4A (32). Our results with rat TI cells are similar to the patterns seen for human OCT-4A in that there is predominantly nuclear rather than cytoplasmic staining (32). By Western blotting, the antibody recognizes a band with apparent molecular weight of ∼50 kDa, consistent with OCT-4A rather than OCT-4B. By Q-PCR, using primers specific for the OCT-4A isoform of the gene, both freshly isolated and cultured TI cells express OCT-4A mRNA; the level of expression in freshly isolated cells is approximately one-half that of the X 404 cell line, an adenocarcinoma line derived from rat lungs (Fig. 8). Together, these data support the concept that TI cells express OCT-4A. The observation that TI cells contain a protein consistent with the properties of OCT-4A parallels reports by Kim (28) and Ling (36), who reported its presence in two populations of potential progenitor cells in the lung, and is compatible with a potential proliferative function for TI cells in vivo.

In addition to proliferating, TI cells can exhibit phenotypic plasticity. TI cells cultured on fibronectin in the presence of FBS do not express SP-C, a marker that within the lung is limited to TII cells (data not shown). However, proliferating TI cells cultured with a matrix rich in laminin and type IV collagen can be induced to express markers of other pulmonary epithelial cells. Under these conditions, TI cells express SP-C, a marker of TII cells, and CC10, a marker of Clara cells (Fig. 9).

Cultured TII cells, which have been extensively studied for 35 years, also display phenotypic plasticity. When grown on tissue culture or other flat substrata in the presence of FBS, TII cells cease to express phenotypic markers of the TII cell phenotype; by changing culture conditions, TII cells can be induced to reexpress at least some of the TII cell markers in culture (3, 8). By varying culture conditions, TII cells can be induced to express markers of TI cells (3, 8, 9, 12, 13), mesenchyme (53), and many other mRNAs (19), suggesting a diverse phenotypic response dependent on culture conditions (19). Of course, each in vitro model system needs to be assessed for the relevance to the biological question being asked. From the current studies and previous reports in the literature cited above, in culture, TI and TII cells require different conditions for proliferation. There are other interesting differences in phenotype between the two populations. Freshly isolated rat TI cells, TII cells, and cultured TII cells each display distinct molecular phenotypes (49), with cultured TII cells and freshly isolated TI cells displaying up to a ∼200-fold difference in relative expression of marker molecules (19). Freshly isolated TI cells and TII cells have a different complement of ion channels, which is, in addition, different from that of cultured TII cells (26).

It has been proposed that there are “resident stem cell” populations in the lung, with two recent publications identifying different potential candidate cell types. Kim et al. (28) identified a population of potential local progenitor or “stem” cells in the distal terminal bronchiolar region of mouse lungs, which they called “bronchoalveolar stem cells.” This population expressed SP-C, CC10 (CCA), and stem cell antigen-1 (Sca-1). Ling et al. (36) reported that a different population of cells (“OCT-4+ stem/progenitor cells”) also displayed some stem-like/progenitor characteristics; these cells simultaneously expressed Sca-1, CC10 (CCA), and OCT-4, but did not express either CD34 or markers of either the TI cell phenotype (AQP5) or the TII cell phenotype (SP-C, alkaline phosphatase). TI cells express a set of marker antigens that is different from either the bronchoalveolar stem cell or OCT-4 stem/progenitor cell populations; TI cells are OCT4+, AQP5+, CC10-, and SPC-. [The Sca-1 gene has neither a rat nor a human ortholog; a 500-kb region of the Ly6 locus of the mouse chromosome containing Sca-1 and eight other genes were deleted between mouse and rat speciation (24).] TI cells differ from the bronchoalveolar stem cells in that TI cells do not express CC10 or SP-C proteins unless cultured under specific conditions; TI cells differ from the OCT-4 stem/progenitor cells in that TI cells express AQP5 but do not express CC10. Although our preparations of TI cells did contain a few TII cells (<0.3%), these cells did not express CC10 (data not shown), distinguishing them from the cell populations described by either Kim et al. (28) or Ling et al. (36). Chen et al. (6) reported a putative progenitor cell population that was the major target for SARS coronavirus in human lung, but the relevance to TI cells or epithelial cells is uncertain.

Both the bronchoalveolar stem cell and OCT-4+ progenitor cell populations have been proposed to be pulmonary stem or progenitor cells. The evidence cited to support this hypothesis is that the cells express antigens common to ES cells, the cells proliferate, and in vitro they can be induced to express single markers associated with the phenotypes of various differentiated cell types, such as CC10 (Clara cells) or SP-C (TII cells) (28, 36). It may be more accurate to describe these cell populations as exhibiting some degree of proliferative and phenotypic properties in vitro, raising the possibility that they may play a similar role in vivo. Type I cells meet similar criteria, in that they express OCT-4, they proliferate, and proliferating TI cells can be induced to express single markers associated with other types of differentiated pulmonary epithelial cells, such as SP-C or CC10. However, it is not clear whether any of these cell populations play a significant role as progenitor cells in either lung development or repair, which would entail methods employing cell fate analysis.

It seems likely that there is more than one resident stem cell or progenitor cell population in the lung. The three-dimensional anatomic complexity of the lung and anatomic separation between the airway and alveolar compartments may act as barriers to repopulation of injured lungs by cells in a different anatomic niche. Further complicating the assessment of “stemness” are the observations that differentiated pulmonary epithelial cells can display considerable phenotypic plasticity. After oxidant injury in vivo, TII cells can transdifferentiate (at least morphologically) into TI cells (14), and some alveolar epithelial cells may transdifferentiate to mesenchyme (29). TI cells in vitro can be induced to express markers of the TII cell phenotype (Fig. 9). Clara cells have the capacity to proliferate and transdifferentiate into bronchial ciliated epithelial cells, and bronchial ciliated epithelial cells can transdifferentiate into squamous cells (39). Whether these cell types play a role as progenitor cells, or whether they may be unipotent, multipotent, or pluripotent is unknown.

Although TI cells in vitro fulfill the criteria that have been used previously to identify putative pulmonary stem cells, it is premature to assert that TI cells function in vivo as progenitor cells, pluripotent, or even multipotent stem cells. If TI cells in vivo display similar characteristics to isolated, cultured TI cells, it is conceivable that they participate in processes of lung development and repair after injury. Because TI cells cover >99% of the internal surface area of the lung, this would provide a large reservoir of cells with the potential capacity to participate in lung repair, making it attractive to propose such a function for this cell type. Definitive experiments regarding this will require specific, irreversible marking of TI cells followed by analysis of cell fate in various models of injury. The techniques to perform such experiments are currently in development.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants HL-24075 and HL-57426.