| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Center for Lung Regeneration, Department of Environmental and Occupational Health, University of Pittsburgh, and 2University of Pittsburgh Cancer Institute, Department of Radiation Oncology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania; 3First Department of Medicine, Hokkaido University School of Medicine, Kita-ku, Japan; and 4Department of Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania
Submitted 12 March 2006 ; accepted in final form 22 November 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
clonogenic airway cells
Development of methods for purification of region-specific stem cells and functional testing in vitro would significantly advance the field of lung stem cell biology. To date, lung stem cell purification has been limited by the fact that molecular markers for tracheobronchial and bronchiolar stem cells are intracellular proteins that are redundantly expressed by cells at multiple levels of the stem cell hierarchy (14, 18, 19, 20). Thus, identification of novel markers, biochemical activities, and/or biophysical properties that discriminate region-specific stem cells from their progeny must precede analysis of mechanisms regulating stem cell differentiation.
ATP-dependent efflux of the DNA dye Hoechst 33342 or the mitochondrial dye rhodamine 123 has been identified as a biochemical property of cells derived from various tissues and is dependent on activity of the ATP-binding cassette transporters (30). Interest in this cell subset (commonly termed the side population, SP) originates from the demonstration that the bone marrow SP was highly enriched with cells that were phenotypically and functionally similar to hematopoietic stem cells (16). These studies raised the possibility that phase III metabolism could serve as a protective mechanism common to stem cells from diverse tissues and that analysis of cells within the SP would lead to identification of cell surface markers useful in distinguishing cells at various levels of the stem cell hierarchy.
Previous analysis of Hoechst efflux in lung cell preparations has identified a verapamil-sensitive SP in embryonic lung (34), in central and peripheral regions of the adult lung (16), and in airway-enriched preparations (14). Immunophenotypic and end-point PCR analysis indicated that the lung SP was heterogeneous and contained cells of hematopoietic (CD45+) and non-hematopoietic (CD45–) origin (1, 14, 33). Further analysis of non-hematopoietic SP cells provided consistent evidence that these cells expressed epithelial and mesenchymal cell markers (14, 33, 34), whereas representation of platelet-endothelial cell adhesion molecule (CD31)-positive endothelial cells and Sca1-positive cells varied with cell preparation method (14, 34). Functional analysis of lung SP cells demonstrated the presence of smooth muscle and endothelial precursors in vitro (34). Even though the physical, molecular, and functional attributes of the remaining epithelial- and mesenchymal-like SP cell subsets were not determined, these data suggested that Hoechst efflux would serve as a good starting point for fractionation of lung cell subsets and for development of cell culture approaches needed to assess self-renewal and differentiation potential.
As indicated above, a further constraint to purification and functional analysis of epithelial stem cells has been the lack of assays for reconstitution of solid tissues in vivo. This deficiency has been addressed in part through analysis of clonogenic frequency (36) and differentiation potential in vitro (35). Although these approaches have limited capacity to recapitulate the cellular and matrix interactions thought to be essential for maintenance and regulation of tissue-specific stem cells, some insight into mechanisms regulating stem cell behavior has been realized (10). In particular, determination of clonogenic frequency allows direct comparison of the number of cells capable of colony formation under a specific set of culture conditions. Application of this methodology to analysis of epithelial (22) and mesenchymal (7) populations has permitted development of purification strategies for cell subsets that express a common set of molecular markers and has served as a first step toward development of improved methods for in vitro propagation and analysis of cellular properties.
In this study, we used clonogenic frequency as an outcome measure to test the hypothesis that clonogenic lung epithelial cells are enriched in the SP. A preparation representing all lung compartments was assayed to provide continuity with previous analysis of lung SP cells (total lung cells) and was compared with epithelial cell-enriched preparations from trachea and the alveolus. The biophysical and molecular characteristics of Hoechst-effluxing cells from each cell preparation were determined by flow cytometry, and gene expression was analyzed by quantitative RT-PCR. Clonogenic frequency was assessed in both high-serum and epithelial-selective media, and the morphological characteristics of the resultant colonies were determined. Finally, to determine whether the number, characteristics, and function of SP cells were altered in fibrotic lung disease, SP cells were selected from bleomycin-treated animals and assayed as indicated above. Results of these analyses demonstrate molecular and functional similarity of SP cells derived from proximal and distal compartments of the mouse lung and supported the conclusion that clonogenic lung cells are highly enriched within the Hoechst 33342-effluxing fraction. However, the clonogenic cells did not express classic epithelial cell markers and raised the possibility that lung SP cells were a transitional cell type of either epithelial or mesenchymal origin. Development of methods for isolation and quantitative assessment of clonogenic cells permitted comparison of such cells from steady state and injured lung and suggested that alveolar SP cell function was dysregulated in fibrotic lung disease.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Bleomycin treatment. Mice were treated with 4 U/kg bleomycin in saline by intratracheal instillation using previously reported methods (27). Briefly, mice were anesthetized, and the trachea was exposed by blunt dissection. Saline or bleomycin in saline was injected (25 µl) into the lumen of the trachea. The tracheal wound was sealed with tissue glue, and the skin incision was closed with a surgical staple. Animals were recovered in the supine position. Animals were recovered for 2 or 7 days.
Mouse cell isolation. Tracheal cells were prepared by the method of You and colleagues (37) with modifications based on methods used for preparation of human cells (11). Briefly, tracheal tissue was recovered by blunt dissection, and cells were released by overnight treatment with 1.5% pronase (Roche) in Ham's F-12. Protease activity was quenched with 10% fetal bovine serum, and the cells were recovered by gentle shaking. Rapidly adherent cells were removed by panning on tissue culture grade 100-mm plates in DMEM/10% fetal bovine serum for 2 h at 37°C. The resultant cell preparation was 95% viable, and 3.04 x 105 ± 0.51 x 105 (n = 25, range = 2–4 x 105) cells were recovered from each trachea. Similar numbers of SP cells were recovered from all mouse strains used. Total lung cells were prepared by a variation of the Chichester protocol (8) with the following minor modifications. Lungs were perfused via the vasculature with HBSS and digested with 0.15% pronase (Roche) in Ham's F-12, and erythrocytes were lysed with Red Blood Cell Lysis Solution (Sigma). Cells prepared by this method were 90% viable, and 2–4 x 106 cells were recovered from each lung. Alveolar cells were prepared by the method of Corti and colleagues (9). Briefly, lungs were perfused with HBSS, and dispase (Roche) was instilled followed by low-melting temperature agarose. Lungs were incubated at 37°C for 1 h, and cells recovered after mincing. Hematopoietic cells were depleted by panning on mouse IgG-coated Petri plates. Nonadherent cells were cultured overnight in small airway growth medium (Cambrex) on 100-mm bacterial plastic Petri dishes. Viability was 90% after overnight culture. Approximately 5 x 106 cells were recovered from each lung. All cell preparations were filtered through 100-µm cell strainers, and cells were resuspended at 1 x 106 cells/ml in DMEM/10 mM HEPES, pH 7.4/2% FBS (DMEM+).
Hoechst staining.
Cells were stained with 5 µg/ml Hoechst 33342 (Sigma) in DMEM+ or 5 µg/ml Hoechst 33342 plus 50 mM verapamil (Sigma) in DMEM+ for 85 min at 37°C with intermittent mixing (http://www.bcm.edu/labs/goodell/protocols/goodell_hoechst.pdf). Stained cells were washed twice with ice-cold DMEM+, resuspended at 1 x 106 cells/ml, counterstained with 1 µl/106 cells FITC
CD45 (Becton Dickenson) or a similarly labeled isotype control for 30 min, and washed twice with ice-cold DMEM+. Cells were resuspended at 5 x 106 cells/ml and filtered through 100 µm cell strainers. Immediately before analysis, cells were stained with 2 µg/ml propidium iodide. Cells that were incubated without Hoechst dye were 
85% viable after the 130-min staining and washing period. Cell viability was reduced to 
50% following Hoechst incubation (total lung and tracheal preparations) and 10–30% (alveolar preparations). Cell viability did not change during the course of FLOW cytometric analysis.
Flow cytometry. Cells were analyzed and sorted using a MoFlo high-speed cell sorter (DakoCytomation, Fort Collins, CO) that was capable of simultaneous detection of eight colors using three laser lines. The instrument was equipped with subsystems of SortMaster Droplet Control and a CyCLONE Automated Cloner. SP cells were identified as described in the text and were based on the work of Goodell and colleagues (16). Emissions of Hoechst 33342 (as Hoechst-red and Hoechst-blue) were detected using 670/30-nm band-pass and 450/65-nm band-pass filters. Immunophenotypic analysis of SP cells was performed using 530/40-nm, 585/30-nm, and 630/30-nm band-pass filters to measure the emissions of FITC, phycoerythrin (PE), and propidium iodide (PI), respectively. Compensation was set manually using single color controls.
Cells used for gene expression analysis were sorted directly into Small Volume RNA Lysis Buffer (Promega), and homogenates were stored frozen before RNA purification. Cells for clonogenicity assays were sorted directly into uncoated 96-well plates or the same coated with collagen type 1 (39). Plates were prefilled with 100 µl of culture medium (see below) and brought to 200 µl following cell deposition.
Quantitative PCR analysis.
Messenger RNA abundance was assessed by quantitative RT-PCR (Q-PCR). Only non-hematopoietic (CD45–) cells were analyzed. Total RNA from 2.5 x 104 viable non-hematopoietic cells, 50,000 total Hoechst-stained or non-SP cells, 5,000–15,000 SP, upper SP, or lower SP cells was purified with an SV RNA Kit (Promega) according to the manufacturer's directions. First-strand cDNA was synthesized using Superscript II (Invitrogen, Carlsbad, CA) according to the manufacturer's directions (RT), and a second reaction that lacked reverse transcriptase (NRT) was used as a negative control. Triplicate aliquots of the RT reaction and duplicate aliquots of the NRT reaction served as templates. Assays-on-Demand gene expression probes (Applied Biosystems, Foster City, CA) included Clara cell secretory protein (CCSP) (Mm00442046_m1), CyP4502F2 (Mm00484087_m1), cytokeratin 14 (Mm00516875_m1), surfactant protein C (Mm00488144_m1), and vimentin (Mm00449201_m1). Cycle conditions were 95°C for 12 min (95°C for 15 s, 60°C for 1 min) x 40 cycles. Differential gene expression presented in 
Fig. 5 was determined using an ABI PRISM 7000 Sequence Detection System, and values were calculated by the 
CT method (17). RT and NRT reactions were performed on 500 ng of commercially prepared lung RNA (Ambion, Austin, TX) and used as the calibrator. 
-actin (Mm00607939_s1) was used as the control mRNA. Q-PCR for the target and control genes was performed in separate tubes to avoid possible competition and/or interference. At least two RNA samples from each cell population were assayed. Differential gene expression presented in Fig. 6 was assessed as detailed above with the exception that an ABI PRISM 7700 Sequence Detection System was used. Values were calculated by the linear regression analysis method (3). A standard curve was generated using fourfold serial dilutions of whole lung cDNA from a normal mouse, and the relative amount of target mRNA in each sample was determined by interpolation of threshold cycles. Values for each sample were normalized to glyceraldehyde-3-phosphatase dehydrogenase (Applied Biosystems). Each PCR assay was independently performed three times to confirm reproducibility. Representative data are shown. Differences in gene expression were assessed by Student's t-test.
|
|
|
|
|
|
Clonogenic frequency analysis. For analysis of clonogenic frequency in unstained viable non-hematopoietic cells, known numbers of cells were sorted directly into DMEM/10% FCS (all cell types), MTEC (airway and tracheal), or small airway growth medium (SAGM, alveolar) and cultured and stained as described above. In experiments utilizing airway cells, the number of cells deposited per well was: no Hoechst, 50, 100, 200, or 400; total Hoechst and non-SP, 200, 500, 1,000, or 2,000; SP, 10, 25, or 50. In experiments utilizing tracheal cells, the number of cells deposited per well was: no Hoechst, total Hoechst, and non-SP, 10, 25, 50, or 100; SP, upper SP, and lower SP, 10, 25, or 50. In experiments utilizing alveolar cells, the number of cells deposited per well was: no Hoechst, total Hoechst, and non-SP, 500, 1,000, 2,000, or 4,000; SP, 10, 50, 100, or 200. Clonogenic frequency was calculated according to the linear regression analysis method of Taswell (36). Wells were scored as either positive or negative for cell growth, and the percent of wells that was negative for cell growth was calculated. The natural log of the percent negative value was plotted against the cell input, and a trendline was inserted. The equation for the best fit line was used to calculate the clonogenic frequency: the cell input leading to 1/Ln(37% negative wells). R2 values, which represent the statistical power of this method, ranged from 0.8662 to 0.9929 airway cell studies, 0.9976 to 0.9985 for tracheal cells (no Hoechst R2 was 0.2090), and 0.9712 to 1 for alveolar cells. Clonogenicity assays were carried out at least twice for each cell type.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
CD45 and propidium iodide. Viable CD45– (non-hematopoietic) cells were assessed for Hoechst efflux as detailed in METHODS. Total lung cell preparations contained a well-defined SP that was 0.93% ± 0.21% (n = 4) of cells analyzed (Fig. 1A, A-1), and coexposure to verapamil resulted in 95% ± 1.1% (n = 4) inhibition of the SP band formation (Fig. 1A, A-2). Tracheal cell preparations also contained a distinct SP that was 1.23% ± 0.15% (n = 25) of cells analyzed (Fig. 1B, B-1), and coexposure to verapamil resulted in 94.4% ± 1.37% (n = 17) inhibition of SP band formation (Fig. 1B, B-2). Finally, alveolar cell preparations (Fig. 1C, C-1 and C-2) contained verapamil-inhibited SP that was 0.41% ± 0.09% (n = 8) of viable, non-hematopoietic cells. Backgating the total lung cell SP onto the propidium iodide vs. FITC plot (Fig. 1A, A-3) demonstrated that these cells were low-to-moderately autofluorescent, and backgating the total lung cell SP onto the forward scatter vs. side scatter plot (Fig. 1A, A-4) indicated that these cells tended to be small and agranular compared with the non-SP (gate R2, Fig. 1A, A-5 and A-6). Comparable analysis of the tracheal SP (Fig. 1B, B-3 through B-6) and alveolar SP (Fig. 1C, C-3 through C-6) corroborated these findings and indicated that viable non-hematopoietic SP cells from all airway compartments were relatively small, agranular, and non-autofluorescent compared with the bulk cell population. Tracheal SP cells are enriched in the low-moderate autofluorescence fraction. To determine whether representation of SP cells could be enhanced by selection of low-to-moderately autofluorescent cells, tracheal cells were prepared and stained as indicated above. Gates demarcating all viable/non-hematopoietic cells (white gate, Fig. 2A) or the low (red region), moderate (blue region), and highly (green region) autofluorescent subsets were established (Fig. 2A). Gene expression, relative size, and granularity, or Hoechst-blue vs. Hoechst-red profile of cells within these gates was determined. Quantitative PCR analysis (data not shown) demonstrated that vimentin and cytokeratin 14 mRNAs were relatively more abundant in the low and moderate autofluorescence fractions, whereas Clara cell secretory protein mRNA was enriched in the high-autofluorescence fraction. The total viable/non-hematopoietic population contained a typical SP that was 0.57% of the gated population (Fig. 2B). Low-autofluorescence cells were predominantly small and agranular (Fig. 2C), contained SP cells at a frequency of 15.61% ± 6.36% (n = 8), and were 64.63% ± 9.94% (n = 8) of the total SP (Fig. 2D). Moderately autofluorescent cells were relatively larger and slightly more granular (Fig. 2E) and contained SP cells (Fig. 2D) at a frequency of 1.22% ± 0.45% (n = 8). Because of the relative abundance of moderately autofluorescent cells, the SP derived from this fraction comprised 33.39% ± 10.05% (n = 8) of the total SP. Highly autofluorescent cells were the largest cells selected and were highly granular (Fig. 2G). This population contained a small number of SP cells [0.13% ± 0.06% (n = 8)] and was a minor contributor to the total SP [1.99% ± 0.86% (n = 8), Fig. 2H]. These results suggested that tracheal secretory cells were a minor component of the SP derived from this compartment. Since CCSP mRNA was previously detected in airway SP cells (15), these results raised the possibility that tracheal secretory cells were functionally distinct from airway secretory cells.
Colony formation by tracheal cells. Analysis of clonogenic frequency requires development of methods for growth of cells at clonal density in vitro. Since gene expression analysis suggested that tracheal cell preparations represent a mixture of cell types, we tested the ability of the low-, medium-, and high-autofluorescence cell fractions to form colonies on uncoated and collagen I-coated tissue culture plastic in the presence of either DMEM/10% FCS or mouse tracheal epithelial culture medium. Known numbers of viable/non-hematopoietic cells from the low-, moderate-, and high-autofluorescence regions (gates 1, 2, and 3; Fig. 3A) were plated under the four conditions described above and cultured for 1 wk, and the number and morphological characteristics of resultant colonies were determined (Fig. 3, B–E). Colonies were not detected on uncoated tissue culture plastic in either growth medium (data not shown). Low-autofluorescence region cells grew preferentially in DMEM (Fig. 3D), and the majority of colonies was composed of spindle-shaped cells (Fig. 3C and gray bars in Fig. 3D) as opposed to cuboidal-shaped cells (Fig. 3B and black bars in Fig. 3D). In contrast, low-autofluorescence region cells cultured in MTEC formed cobblestone colonies preferentially (Fig. 3E). Moderate- and high-autofluorescence region cells grew most effectively in epithelial-selective medium but showed a propensity for formation of cobblestone colonies in either medium. This analysis demonstrated an unanticipated preference of various tracheal cell subsets for specific cell culture medium and highlighted the importance of testing clonogenic frequency of SP cells in both types of media.
Clonogenic cells are enriched in the SP. The functional properties of SP cells from total lung, trachea, and alveolar cell preparations were determined though analysis of clonogenic frequency (36). A tight correlation between colony formation and the number of input cells was demonstrated by R2 values that were greater than or equal to 0.9 for all analysis done in DMEM. SP cells from total lung, trachea, and alveolar preparations did not grow effectively in epithelial cell-selective medium, and only minimal differences in clonogenic frequency were detected in pairwise comparisons between the various subpopulations (gray bars, Fig. 4, A-1, B-1, and C-1). In contrast, culture in DMEM permitted growth of SP from all three cell preparations (black bars, Fig. 4, A-1, B-1, and C-1). Clonogenic frequency for total lung SP cells was 6.16-fold greater than unstained cells and 12.78-fold or 86.25-fold greater than total Hoechst-stained or non-SP cells, respectively (Fig. 4, A-1). Clonogenic frequency for tracheal SP cells was 20.74-fold greater than unstained cells and 394-fold greater than unfractionated Hoechst-stained cells (Fig. 4, B-1). No colonies were detected for non-SP cells in this experiment but were detected in other experiments in which the number of non-SP cells plated ranged from 200 to 2,000 rather than the 10–50 cell range used in the current analysis (Fig. 6). Alveolar SP cells were 45.53-fold more clonogenic than unstained cells and 42.69-fold or 37.9-fold more clonogenic than total Hoechst-stained or non-SP cells, respectively (Fig. 4, C-1). The clear enrichment of clonogenic cells within the SP compared with unstained cells suggested that the low clonogenic frequency of non-SP cells relative to unstained cells or total Hoechst-stained cells was due to depletion of clonogenic cells rather than Hoechst toxicity. The average clonogenic frequency for all SPs analyzed was 47 ± 0.79 (n = 6) per thousand cells, and a Student's t-test indicated that clonogenic frequency did not vary among the three types of cell populations tested.
Morphological characteristics of SP cell colonies.
Previous analysis of unselected tracheal cells indicated that these cells generated colonies composed of either cuboidal- or spindle-shaped cells (Fig. 3). To determine whether selection of total lung, tracheal, or alveolar SP cells resulted in enrichment of progenitors for specific colony type(s), colonies formed in DMEM were classified as cobblestone (cuboidal cells, Fig. 4, A-2, B-2, and C-2), stellate (spindle cells, Fig. 4, A-3, B-3, and C-3), or abnormal (lightly staining, bubbly cytoplasm, high cytoplasmic-to-nuclear ratio, Fig. 4, A-4, B-4,and C-4), and the representation of each colony type was calculated as a percent of the total number of colonies evaluated (Table 1). Colonies generated from all test populations varied in size from small (4–32 cells, 2–5 doublings) to large (
250 cells, 8 doublings). The majority of colonies derived form total lung SP cells (90%) contained tightly packed abnormal cells with a diffusely staining cytoplasm (Fig. 4, A-4, and Table 1), whereas the remaining colonies (10%) were evenly distributed between the cobblestone (Fig. 4, A-2) and stellate (Fig. 4, A-3) categories. In contrast, the majority of tracheal SP (80%) and alveolar SP cells (73%) appeared to be viable (relatively few abnormal colonies were noted), although colony type varied with the cell population analyzed. The majority of tracheal SP colonies contained spindle-shaped cells (stellate colony type, Fig. 4, B-3, Table 1), whereas nearly all alveolar SP colonies were composed of cuboidal cells (cobblestone colony type, Fig. 4, C-2, Table 1). Detection of colonies with stellate or cobblestone morphology suggested that a heterogeneous population of colony-forming cells contributed to the tracheal SP, whereas the preferential formation of epithelial-like colonies by alveolar SP cells suggested that this population was more homogeneous. However, these experiments could not determine whether selection of specific colony-forming cell types occurred at the level of Hoechst efflux or culture conditions.
|
To determine whether vimentin expression was a consistent characteristic of lung SP cells, the gene expression profiles of tracheal and alveolar cell fractions were determined by Q-PCR. Analysis of unstained viable/non-hematopoietic tracheal cells demonstrated that this cell preparation was highly enriched with cytokeratin 14-expressing cells. Little to no expression of CCSP, surfactant protein C, or vimentin was detected (Fig. 5A). Expression of cytokeratin 14, CCSP, and vimentin was equivalent in unstained control and tracheal non-SP cells, suggesting that all cell types survived Hoechst exposure (Fig. 5C). However, cytokeratin 14 and CCSP mRNAs were depleted in the tracheal SP. Vimentin mRNA was enriched 130-fold in the SP when compared with either unstained cells or the non-SP (P < 0.001). Quantitative PCR analysis of freshly isolated alveolar cells demonstrated the expected enrichment of surfactant protein C-expressing cells with minor contamination from CCSP-expressing cells (Fig. 5A). Analysis of alveolar non-SP cells revealed a significant depletion of all mRNAs analyzed and suggested that cells from this compartment were highly susceptible to Hoechst toxicity (typically 70–90%, see above, Fig. 5D). Despite this high rate of cell attrition, one marker, vimentin, was highly enriched in the alveolar SP. In these experiments, vimentin was nearly 6 times more abundant in the SP fraction compared with freshly isolated cells and 60 times more abundant in the SP than in the non-SP (P < 0.001). These data support the conclusion that vimentin is an abundant mRNA within each of the SP fractions studied.
Phenotypic and functional subsets of tracheal SP cells. Previous analysis of hematopoietic SP cells demonstrated that long-term repopulating hematopoietic stem cells were concentrated at the tip of the SP band (6, 16). To determine whether tracheal SP cells could be further purified using this approach, the tracheal SP band was divided into two regions, the upper SP and the lower SP that represented 36% and 64% of the SP, respectively (Fig. 6A). Known numbers of cells from each subfraction as well as the non-SP, the entire viable/non-hematopoietic Hoechst-stained population, and unstained viable/non-hematopoietic cells were deposited directly into collagen-coated wells of 96-well plates, and clonogenic frequency was calculated (Fig. 6B). As previously demonstrated, tracheal cells grew preferentially in DMEM (black bars, Fig. 6B) compared with epithelial-selective medium (gray bars, Fig. 6B). Unstained cells and the total Hoechst-stained population had similar but low clonogenic frequencies of 3.3 and 2.8 clonogenic cells per thousand, respectively. Comparison of clonogenic frequency in the non-SP, upper SP, and lower SP demonstrated that clonogenic cells were depleted from the non-SP and were enriched 81-fold in the upper SP and 139-fold in the lower SP. Clonogenic cells were 1.7 times more abundant in the lower SP than in the upper SP. When differences in the absolute number of cells within each fraction were accounted for, clonogenic cells were three times more numerous in the lower SP. Colonies formed from the lower SP fraction tended to be composed of spindle-shaped cells (stellate colony morphology, 95.6%), and all other colonies were abnormal in morphology (4.6%, Table 1).
To determine whether gene expression varied as a function of position within the SP band, vimentin, cytokeratin 14, and CCSP mRNA abundance was assayed as previously described. In addition, the Clara cell marker CyP450-2F2 was assayed to account for the possibility that CCSP gene expression was downregulated under these experimental conditions (28). Vimentin mRNA (Fig. 6C) was enriched 28x in the upper SP compared with either the non-SP or the total cell population and was enriched an additional 1.5-fold in the lower SP (42x relative to the non-SP). Enrichment of vimentin mRNA in the SP fractions relative to unstained cells was statistically significant (P < 0.001). Cytokeratin 14 mRNA (Fig. 6D) was fivefold more abundant in the upper SP than the non-SP or the total cell population (P < 0.001) and was not detected in the lower SP. Very low levels of CCSP (Fig. 6E) and CyP450-2F2 mRNA (Fig. 6F) were detected in the upper and lower SPs and were 2–5% of levels in the non-SP and total populations. Differences in CCSP and CyP450-2F2 mRNA abundance in the total or non-SP cell fractions and the SP fractions were statistically different (P < 0.001). These data demonstrated that vimentin-expressing cells contribute to all portions of the tracheal SP band and that the gene expression profile of the upper SP was distinguished from that of the lower SP by the presence of the epithelial cell marker cytokeratin 14. Overall, these data demonstrate that Hoechst effluxing capacity of tracheal cells was positively correlated with clonogenic frequency and with enrichment of vimentin mRNA. These results suggested that the approximately twofold enhancement of clonogenic frequency in lower SP was the result of depletion of a cytokeratin 14-expressing cell type rather than differential susceptibility of upper and lower SP cells to Hoechst toxicity.
Contribution of alveolar SP cells to lung fibrosis. Injury scenarios that result in disruption of epithelial/mesenchymal interactions are commonly associated with a fibrotic response that can result in scar formation, and, in the case of the lung, severe decrements in lung function (32). To determine whether the colony-forming ability of alveolar SP cells varies in the context of fibrosis, alveolar cells were prepared from groups of mice recovered 2 or 7 days after bleomycin treatment. Total cell recovery and overall viability did not vary between control and treated groups (data not shown). The frequency of S-phase cells was slightly increased on recovery days 2 and 7 (39% in treated cells vs. 23% in control, data not shown); however, the frequency of SP cells did not vary between control and treated groups (data not shown). As previously demonstrated for control cells, clonogenic frequency of alveolar SP cells from bleomycin-treated animals was sevenfold greater in DMEM than in SAGM (Fig. 7A), and the majority of clonogenic cells was distributed to the SP fraction at both the 2- and 7-day recovery time points (Fig. 7B). The clonogenic frequency of SP cells was 8.3% of control cells on recovery day 2 and 28.8% of control on recovery day 7 (Fig. 7B), and all colonies exhibited an abnormal morphology (Table 1). These results demonstrated that the clonogenic potential of alveolar SP cells was altered in the context of bleomycin-induced lung injury, and the shift from cobblestone to abnormal colony phenotype suggested that mitotic potential and survival in vitro were negatively impacted by bleomycin-induced lung injury.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Based on the successful use of clonogenic frequency to assess the functional properties of various isolated cell types (36), we chose this method to evaluate and compare the colony-forming capacity of lung SP cells. These studies demonstrated that the most clonogenic tracheal SP fraction was highly enriched in vimentin mRNA and preferentially formed stellate colonies. Alveolar SP cells were equally clonogenic and highly enriched in vimentin mRNA but were distinguished by preferential generation of cobblestone colonies. These observations reinforce the notion that phase III metabolism, mediated by the ATP binding cassette transporter BCRP1 (ABCG2), is a unifying characteristic of clonogenic progenitors from diverse tissue types. However, biased generation of stellate colonies by tracheal SP cells and cobblestone colonies by alveolar SP cells raised the possibility that the present analysis did not equivalently and appropriately evaluate all cell types within the various SPs tested. Of particular concern to the field of stem cell biology is the possibility that this method may be adequate for analysis of cells with intrinsic clonogenic potential but that it may have failed to detect cells that require extrinsic interactions/factors for clonal growth. The validity of this concern was supported by the finding that colony formation in epithelial-selective medium cells was density dependent and the implication that cellular communication was an important determinant of colony forming potential. Recognizing the importance of cellular interactions in airway homeostasis, modification of the current in vitro methods to include use of feeder layers is warranted. In addition, analysis of SP cell properties in tracheal xenografts and following transplantation to the lung or ectopic sites is needed to fully assess the functional characteristics of lung SP cells. Despite these limitations, development of a method for enrichment of lung colony-forming lung cells represents a technical advancement that will permit development of improved methods for analysis of cellular interactions in vitro.
The gene expression profile of SP cells collected in this study is in contrast with previous reports that detected epithelial differentiation markers (14) and epithelial transcription factors (33) in SP cells from airway or total lung preparations. Based on the mitotic potential and self-renewal capability of Clara and alveolar type 2 cells, we hypothesized that selection of clonogenic progenitors from the airway and alveolar compartments would result in parallel enrichment of CCSP and surfactant protein C mRNAs. However, use of Q-PCR demonstrated that CCSP and surfactant protein C mRNA abundance was decreased in the total Hoechst-stained fractions and non-SP cells relative to unstained cells. Although CCSP and surfactant protein C gene expression is regulated by certain stress conditions (4, 28), a similar lack of the Clara cell differentiation marker CyP450-2F2 in tracheal cells total Hoechst and non-SP fractions suggests that cell type-specific Hoechst toxicity rather than altered gene expression resulted in selective depletion of CCSP and surfactant protein C-expressing cells. This finding highlights a limitation of the Hoechst efflux method for analysis of conditionally differentiated lung epithelial cell subsets. In contrast, cytokeratin 14, a marker for clonogenic tracheobronchial cells (19, 20), was detected within the tracheal SP but was limited to cells with Hoechst-blue fluorescence intensity that was similar to that of the non-SP. These results highlight the potential for cells within the shoulder of the non-SP to be misidentified as SP cells. Based on the use of quantitative molecular and functional criteria for assessment of SP cell phenotype, we conclude that lung SP cells represent a highly purified population of clonogenic cells but that the Hoechst efflux technique is not a useful method for selection of previously identified multipotential epithelial cells and their progeny.
An unanticipated outcome of this study was the finding that SP cells derived from various lung compartments were relatively similar with respect to autofluorescence characteristics and molecular phenotype as well as their propensity for clonal growth in serum-containing medium. One plausible explanation for the relatively uniform phenotype of lung SP cells is that Hoechst efflux defined a cell type that was held in common among the various compartments analyzed. Previous analysis of lung SP cells indicates that candidate cell types include endothelial and mesenchymal cell populations (15, 34). Such cells are frequently identified as vimentin-expressing cells and are known to proliferate in response to injury, alter their phenotype in chronic lung disease, and contribute to aberrant wound-healing processes associated with fibrotic lung disease (32). An alternative identity for lung SP cells may be that of a transitional phenotype induced by cell isolation and related to that of epithelial cells involved in restitution of an injured epithelium (2, 13). Lung injury that results in selective depletion of Clara cells initiates a cellular response characterized by squamation of residual ciliated cells and is associated with alterations in ultrastructural characteristics (24). In contrast, full thickness injury such as that resulting from a scrape wound is associated with flattening of cells at the border between the uninjured and injured regions and movement of these cells into the injured zone (23, 25). Acquisition of a migratory phenotype in skin, kidney, and mammary gland is associated with upregulation of the intermediate filament vimentin and downregulation of epithelial cell type-specific markers (reviewed by Kalluri and Neilson, Ref. 21). Thus, vimentin gene expression, which defines lung SP characterized in this study, may be associated with epithelial-to-mesenchymal transitions that are an important component of normal cell activities that include gland formation, wound healing, and tumor metastasis, as well as abnormal processes leading to fibrosis (37).
Previous analysis of Bcrp1 mRNA and protein distribution in the lung detected expression in bronchial and vascular smooth muscle and a subset of cells within the alveolar region. However, expression of Bcrp1 did not correlate with the biochemical property of Hoechst efflux (33), and subsequent studies suggested that Bcrp1 was translocated to the plasma membrane of hematopoietic stem cells under stress conditions (26). These studies support the concept that the capacity to efflux Hoechst dye is a component of an induced phenotype and raise the possibility that the alveolar SP may include epithelial cells that have undergone an epithelial-to-mesenchymal transition. Identification of clonogenic alveolar cells that likely express a wound-repair phenotype and are compromised in the context of fibrotic lung disease suggests that alveolar SP cells may be a valuable test population for comprehensive analysis of gene expression changes associated with fibrotic lung disease and for development of reconstitution assays that test the importance of extrinsic factors in regulation of cellular phenotype and function.
|
GRANTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
in vitro and in vivo. Am J Physiol Lung Cell Mol Physiol 284: L108–L118, 2003.This article has been cited by other articles:
|
|
 |
|
  T.-L. Hackett, F. Shaheen, A. Johnson, S. Wadsworth, D. V. Pechkovsky, D. B. Jacoby, A. Kicic, S. M. Stick, and D. A. Knight Characterization of Side Population Cells from Human Airway Epithelium Stem Cells, October 1, 2008; 26(10): 2576 - 2585. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. J. Weiss, J. K. Kolls, L. A. Ortiz, A. Panoskaltsis-Mortari, and D. J. Prockop Stem Cells and Cell Therapies in Lung Biology and Lung Diseases Proceedings of the ATS, July 15, 2008; 5(5): 637 - 667. [Full Text] [PDF]
|
|
|
|
|
|
S. D. Reynolds, A. C. Zemke, A. Giangreco, B. L. Brockway, R. M. Teisanu, J. A. Drake, T. Mariani, P. Y.P. Di, M. M. Taketo, and B. R. Stripp Conditional Stabilization of {beta}-Catenin Expands the Pool of Lung Stem Cells Stem Cells, May 1, 2008; 26(5): 1337 - 1346. [Abstract] [Full Text] [PDF]
|
|
