Ciliated epithelial cell lifespan in the mouse trachea and lung

Abstract

The steady-state turnover of epithelial cells in the lung and trachea is highly relevant to investigators who are studying endogenous stem cells, manipulating gene expression in vivo, or using viral vectors for gene therapy. However, the average lifetime of different airway epithelial cell types has not previously been assessed using currently available genetic techniques. Here, we use Cre/loxP genetic technology to indelibly label a random fraction of ciliated cells throughout the airways of a cohort of mice and follow them in vivo for up to 18 mo. We demonstrate that ciliated airway epithelial cells are a terminally differentiated population. Moreover, their average half-life of 6 mo in the trachea and 17 mo in the lung is much longer than previously available estimates, with significant numbers of labeled cells still present after 18 mo.

the rate of turnover of different cell types in the steady-state adult lung has not yet been assessed using modern genetic techniques. These data would be highly relevant to investigators who are studying endogenous lung stem cells, the response of the lung epithelium to injury, manipulating gene expression in vivo, or using viral vectors for gene therapy in the adult airways. Currently available estimates for population turnover time (also known as population renewal time, defined as the number of days required to replace the entire population of cells) vary greatly. For example, total epithelial turnover time in the mouse trachea has been estimated in different studies at between 2 and 267 days (1, 3). These estimates have been generated using different methodologies based on measurements of cell cycle time and the assumption that all new cells generated over the course of the experiment are destined to replace old or injured cells. Parameters used to estimate cell cycle time include the labeling index at intervals after a pulse of tritiated thymidine ([³H]thymidine) and the percentage of observed mitotic cells after colchicine treatment to induce metaphase arrest (9). There are multiple possible reasons for the wide variation in estimates of population turnover time for the airways. These include the differing methodologies employed; differences in age, nutrition, strain, and health status of the rodents; diurnal and seasonal variations in proliferation rates; and the difficulties in identifying the different cell types before the availability of cell type-specific antibodies (9). Given these multiple possible sources of variation, the average lifetime of the different epithelial cell types of the lung would be a more useful value than population turnover time.

Studies of airway epithelial cell turnover at homeostasis have concluded that ciliated cells are a terminally differentiated cell population (for example, Refs. 4, 5). We have used the Cre/loxP lineage-labeling system to indelibly label ciliated cells throughout the airways of a cohort of transgenic (FOXJ1CreER^2T; Rosa26R-eYFP/+) mice. In FOXJ1CreER^2T animals, all ciliated cells express a version of Cre recombinase fused to a modified estrogen receptor (13). This fusion protein is activated by exposing the animals to the synthetic steroid tamoxifen (tmx) (7). On binding to tmx, Cre is transported to the nucleus where it recombines the DNA at the Rosa26R locus to turn on the expression of a reporter protein. Many studies have shown that not all CreER-expressing cells activate the reporter; rather, only a random subpopulation does so (for example, Refs. 6, 7, 16). The reporter is activated by a genetic change and is therefore stably expressed by the cells, and any descendents they may have, for the rest of their life (reviewed in Ref. 8). This means that the average lifetime of the cells can be assessed independently of any variation in proliferation rates or cell cycle length. A similar analysis demonstrated that the β-cells of the adult mouse pancreas are a self-renewing population during homeostasis (6). Using this lineage-labeling technique, we have conclusively demonstrated that at steady-state ciliated airway epithelial cells are a terminally differentiated population. Moreover, we find that the average lifetime of a ciliated cell in the mouse airways is considerably greater than previous estimates based on epithelial population turnover time.

MATERIALS AND METHODS

Mouse strains.

FOXJ1CreER^2T and Rosa26R-eYFP [Gt(ROSA) 26Sor^tm1(EYFP)Cos] were maintained on a C57BL/6 background and used at the N5 and N3 backcross generations, respectively (13, 14). Both male and female 8- to 12-wk-old FOXJ1CreER^2T; Rosa26R-eYFP/+ animals were used. A cohort of 28 mice was given five intraperitoneal injections of tmx (Sigma) dissolved in corn oil at the dose of 0.15 mg of tmx/g body wt over the course of 2 wk. Animals were killed in groups of 3 or more at 3 days, 3 mo, 7 mo, 1 yr, and 18 mo after the final tmx injection. Animals were housed in individually ventilated cages in a specific pathogen-free facility with continuous access to food and water. All animal experiments were approved by the Duke University Institutional Animal Care and Use Committee.

Immunohistochemistry.

Lungs were inflated with 1 ml of 4% paraformaldehyde (PFA). Lungs and trachea were then placed in 4% PFA and fixed at 4°C for 4 h on a rocking platform. Tissue was washed in PBS, 15%, 20%, and 30% sucrose and then incubated overnight in 30% sucrose mixed 1:1 with optimum cutting temperature compound (Tissue-Tek OCT) at 4°C before embedding in OCT. Frozen sections (12-μm) were cut and dried 2 h at room temperature before staining. Immunohistochemistry for green fluorescent protein (GFP) (1:500 dilution, rabbit anti-GFP, ab290; Abcam) and β-tubulin (1:200 dilution, mouse anti-β-tubulin, MU178-UC; BioGenex) was performed according to standard procedures. Alexa Fluor-coupled secondary antibodies (Invitrogen) were used at 1:500 dilution.

Confocal microscopy and cell counting.

Cells were counted manually. All images used for scoring cells consisted of a z-series of optical sections collected on a Zeiss LSM 510 Meta laser scanning confocal microscope. Multiple optical sections were examined to distinguish cell boundaries. In the trachea, cells were counted along the entire proximal-distal axis of a sagittal section cut from the middle of the trachea. In the distal lung, cells were counted from two nonadjacent sections cut from the center of the left lobe. Within the lung, cells for which the apical surface was not visible due to the plane of the section were not counted. All airways with a visible junction with the alveoli were scored as terminal bronchioles. Airways were scored as bronchioles if they were clearly located directly proximal to a terminal bronchiole or if they were of a larger diameter than visible terminal bronchioles. Cells were not counted in airways that could not be classified according to these criteria.

Statistics.

All errors given are SE. Data were fitted to exponential curves or linear functions by regression analysis (Microsoft Excel). The average half-life of the ciliated cells was estimated by solving these equations.

RESULTS AND DISCUSSION

Twenty-eight adult FOXJ1CreER^2T; Rosa26R-eYFP/+ mice were exposed to a pulse of tmx to label some of the ciliated cells with enhanced yellow fluorescent protein (eYFP). Previous experiments demonstrated that the FOXJ1 promoter fragment drives expression in all ciliated cells of the mouse airways and that there is no activation of the reporter in the absence of tmx exposure in the FOXJ1CreER^2T strain (13). Therefore, counting large numbers of cells in multiple mice gives an accurate overall picture of the extent of ciliated cell labeling in this mouse strain. Mice were killed, in groups of at least three animals, at intervals over the course of 18 mo, and the numbers of labeled ciliated cells (eYFP⁺, β-tubulin⁺), nonlabeled ciliated cells (eYFP⁻, β-tubulin⁺), and labeled nonciliated cells (eYFP⁺, β-tubulin⁻) were counted. This analysis allowed us to distinguish between three theoretically possible scenarios. 1) The proportion of labeled ciliated cells decreases over time, and the label does not appear in any other cell type. From this result, we would conclude that the ciliated cells are terminally differentiated and be able to calculate the average lifespan of a ciliated cell at homeostasis. 2) The proportion of labeled ciliated cells remains constant throughout the experiment, and the label is not seen passing into any other cell type. From this result, we would conclude that ciliated cells self-renew at homeostasis. 3) The lineage label is observed in other cell types over the course of the experiment. From this result, we would conclude that ciliated cells can act as a stem, or transiently amplifying, cell population.

Three days after the final tmx injection, the proportion of lineage-labeled ciliated cells (∼70%) was the same in the trachea, bronchioles, and terminal bronchioles (Table 1). Previously published experiments have shown that the fraction of CreER-expressing of cells that activate the reporter is proportional to the dose of tmx (7, 13). In each region of the conducting airways, the proportion of lineage-labeled ciliated cells decreased over time. Both males and females were killed at each time interval, and no difference was observed in the decrease in number of lineage-labeled ciliated cells between the sexes. Moreover, the label was never observed in other cell types (Fig. 1). This result was confirmed by staining slides for Clara and basal cell markers (data not shown). The small number of (eYFP⁺, β-tubulin⁻) cells observed did not change over the course of the experiment (Table 2). This fraction most likely represented ciliated cells that had either lost their cilia during processing or were sectioned at an oblique angle so that the cilia were not visible. These data demonstrate that ciliated cells are indeed a terminally differentiated population at homeostasis. This is consistent with previously published results and confirms what we have observed using this lineage-labeling technique during airway repair (13). The overall number of ciliated cells in the airways remained constant throughout these experiments. Therefore, ciliated cells must be replenished by a stem cell population. Current evidence suggests that these stem cell populations are most likely to be basal cells in the trachea and Clara cells in the lung (reviewed in Refs. 11, 12).

Table 1.

Mean percentage of lineage-labeled ciliated cells ± SE in different regions of the airways after a pulse of tamoxifen and varying length chase period

Time	Trachea	Bronchioles	Terminal Bronchioles
3 days	72.1%±3.2% (n = 1,684 ciliated cells, N = 7 mice)	62.5%±4% (n = 2,184 ciliated cells, N = 7 mice)	64.0%±3.5% (n = 1,615 ciliated cells, N = 7 mice)
3 mo	45.0%±5.6% (n = 1,124 ciliated cells, N = 5 mice)	53.1%±4.5% (n = 1,273 ciliated cells, N = 5 mice)	60.1%±2.9% (n = 951 ciliated cells, N = 5 mice)
7 mo	22.1%±6.8% (n = 1,316 ciliated cells, N = 3 mice)	42.2%±2.1% (n = 1,118 ciliated cells, N = 3 mice)	39.0%±1% (n = 816 ciliated cells, N = 3 mice)
12 mo	12.5%±2.1% (n = 1,176 ciliated cells, N = 3 mice)	34%±1.5% (n = 978 ciliated cells, N = 3 mice)	38.5%±4.3% (n = 601 ciliated cells, N = 3 mice)
18 mo	9.4%±2.9% (n = 2,029 ciliated cells, N = 3 mice)	31%±0.6% (n = 1,553 ciliated cells, N = 3 mice)	26.7%±3.6% (n = 975 ciliated cells, N = 3 mice)

Fig. 1.

The percentage of lineage-labeled ciliated cells in the airways decreases over time. A–G: representative FOXJ1CreER^2T; Rosa26R-eYFP/+ frozen sections posttamoxifen. Green, anti-green fluorescent protein (GFP) (lineage label); red, anti-β-tubulin (cilia). A–E: trachea sections. F and G: terminal bronchiole sections. Bar = 50 μm in all panels.

Table 2.

Mean percentage of lineage-labeled nonciliated cells ± SE in different regions of the airways after a pulse of tamoxifen and varying length chase period

Time	Trachea	Bronchioles	Terminal Bronchioles
3 days	2.4%±1.4% (n = 1,248 GFP+ cells, N = 7 mice)	1.5%±0.4% (n = 1,352 GFP+ cells, N = 7 mice)	5%±0.1% (n = 1,087 GFP+ cells, N = 7 mice)
3 mo	4.2%±1.2% (n = 505 GFP+ cells, N = 5 mice)	2.6%±1.0% (n = 710 GFP+ cells, N = 5 mice)	6%±2% (n = 614 GFP+ cells, N = 5 mice)
7 mo	1.2%±0.9% (n = 297 GFP+ cells, N = 3 mice)	0.5%±0.2% (n = 351 GFP+ cells, N = 3 mice)	0%±0% (n = 317 GFP+ cells, N = 3 mice)
12 mo	0.8%±0.8% (n = 153 GFP+ cells, N = 3 mice)	0.4%±0.2% (n = 351 GFP+ cells, N = 3 mice)	0%±0% (n = 164 GFP+ cells, N = 3 mice)
18 mo	1.4%±0.7% (n = 193 GFP+ cells, N = 3 mice)	0.6%±0.3% (n = 485 GFP+ cells, N = 3 mice)	0.9%±0.9% (n = 262 GFP+ cells, N = 3 mice)

GFP, green fluorescent protein.

Our data, for both the trachea and bronchioles, fit very well to first-order exponential decay functions (Fig. 2, A–C). These curves are converted to straight lines when the natural log of the data is plotted against time (Fig. 2D). This indicates that, analogous to the decay of a radioactive isotope, ciliated cells are lost from the trachea and bronchioles at a constant rate. By solving the fitted equations, we estimate the half-life of ciliated cells to be 6 mo in the trachea and 17 mo in the bronchioles. This is in general agreement with previous estimates of whole epithelial turnover time that found turnover to be slower in the more distal airway regions than proximal. The data for the terminal bronchioles are more variable and do not fit precisely to an exponential decay curve (Fig. 2C). This variation is probably due to the smaller number of ciliated cells in these regions. However, the estimated ciliated cell half-life for the terminal bronchioles (14 mo) is very similar to that of the bronchioles, and it is likely that ciliated cell lifespan is identical in these two areas. Our values for ciliated cell half-life are not consistent with many of the previous estimates of epithelial turnover times (2–267 days in the trachea and 10–59 days in the bronchioles; Refs. 1, 3). Our figures are likely to be more accurate for two reasons. 1) The genetic technology that we have employed estimates ciliated cell half-life independently of cell cycle kinetics, which probably accounts for much of the variation in previous reports. 2) Our experiments were performed on mice housed in a specific pathogen-free facility. It has been previously established that rodents exposed to respiratory pathogens have higher rates of cell turnover in the airways (15). It should be remembered that the other airway epithelial cell types (basal, Clara, neuroendocrine, type I, and type II) do not necessarily have the same lifespan as the ciliated cells.

Fig. 2.

The percentage change of lineage-labeled ciliated cells fits to a first-order exponential decay function. A–C: the data points are the mean percentage of lineage-labeled ciliated cells ± SE. Exponential decay functions (curves) were fitted to these data by regression analysis. Half-life of the ciliated cells was estimated from the fitted equations to be 6 mo in the trachea (A), 17 mo in the bronchioles (B), and 14 mo in the terminal bronchioles (C). D: plotting the natural log of the percentage of lineage-labeled ciliated cells against time converts the exponential curves to straight lines.

Interestingly, the majority of ciliated cells in the ducts of the submucosal glands in the mice killed 18 mo after tmx exposure were lineage-labeled (data not shown), even though very few lineage-labeled cells could be identified in the tracheal epithelial surface at the end of the chase period (Fig. 1E). This supports previous suggestions that the submucosal gland ducts provide an environment where cells are protected from damage, possibly due to the presence of mucus (2).

Our analysis demonstrates that, at homeostasis, airway epithelial ciliated cells are a terminally differentiated cell population with a much longer half-life than previously supposed. A potential drawback to our approach for estimating ciliated cell lifespan is that the life of the cells could be artificially prolonged by expression of Cre or eYFP or recombination of the Rosa26R locus. However, a recent report has shown that expression of Cre DNA recombinase can trigger death in some cell types, making it more likely, if indeed there is any effect of Cre in our experiments, that ciliated cell lifespan would be artificially decreased (10). Rather, given that we observed significant numbers of lineage-labeled cells 18 mo after tmx exposure, any effects of Cre on cell lifespan in the airway epithelium are likely to be negligible. It is theoretically possible that the 30% of ciliated cells that were not labeled in our experiments are stem cells or a self-renewing population rather than terminally differentiated cells. However, since a different random fraction of the total ciliated cell population would be labeled in each of our mice, we consider this to be a highly unlikely scenario. The long lifespan of ciliated cells has important implications for researchers in several different areas of airway biology. For example, experiments to conditionally modify genes that are predicted to affect cell fate, rather than proliferation rates, in adult mouse airway epithelial stem cells during homeostasis will need to be performed over the course of 1–2 yr to allow for sufficient epithelial cell turnover to assess a phenotype.

GRANTS

This work was supported by National Institutes of Health Grant NIH-3036424 to B. L. M. Hogan and a Francis Family Foundation Parker B. Francis Fellowship to E. L. Rawlins.