SEARCH

SEARCH BY CITATION

Keywords:

  • Esophagus;
  • Stem cells;
  • Transferrin receptor;
  • α6 Integrin;
  • Keratinocyte;
  • Epithelial cell

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

The identification and characterization of esophageal stem cells are critical to our understanding of the biology of the esophageal epithelium in health and disease. However, the proliferative compartment within the mouse esophageal epithelium remains poorly characterized. Here, we report that the basal cells of the mouse esophagus can be separated into three phenotypically and functionally distinct subpopulations based on the expression of α6 integrin and transferrin receptor (CD71). Cells that express high levels of α6 integrin and low levels of CD71, termed α6briCD71dim, are a minor subpopulation of small and undifferentiated cells that are enriched for label-retaining cells and thus represent a putative esophageal stem cell population. Conversely, cells expressing high levels of both α6 integrin and CD71 (α6briCD71bri), the majority of basal esophageal cells, are enriched for actively cycling cells and therefore represent a transit-amplifying population. Kinetic analyses revealed that a third cell population, which is α6 integrin-dim and CD71-bright (α6dim), is destined to leave the basal layer and differentiate.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

In recent years, much attention has been devoted to resolving different classes of proliferative cells, particularly stem cells in a variety of mammalian tissues. Continuously renewing tissues like the hemopoietic system, the intestinal epithelium, and stratified epithelia are organized in a hierarchical fashion with the undifferentiated, quiescent, long-lived, multipotent, and self-renewing stem cells at one end of the hierarchy and fully mature, terminally differentiated cells at the other [1, 2]. Stem cells are important not only because of their potent ability to regenerate their own tissue but because their extensive capacity for proliferation implicates them in wound healing and tumorigenesis [3].

It is well-accepted that tissue renewal in the hemopoietic system and the skin is achieved through the combined proliferative activity of a minor population of stem cells and a large pool of actively dividing, short-lived, transit-amplifying cells and that these populations can be discerned by cell kinetic analyses and/or cell surface markers [4, [5]6]. However, little work has been done on the esophageal epithelium despite the clinical importance of this tissue.

The epithelia of the mouse and rat esophagus can be divided into two compartments, the basal layer composed of densely packed columnar cells and the superficial spinous and granular cell layers composed of large polyhedral and flattened keratinizing cells, respectively. Tritiated thymidine-labeling studies revealed that all cells in the basal layer, and only cells in the basal layer, are capable of undergoing cell division [7]. Furthermore, cell division in the basal layer, and cell migration from it (i.e., differentiation), appears to be stochastic. Therefore, it was deduced that “the basal cells are the stem cells of the esophageal epithelia” [7]. This contention remains to be proven experimentally, although it seems unlikely that all basal cells represent stem cells. Indeed, in virtually all other stratified epithelia, it has been possible to demonstrate that stem cells are relatively rare cells within the basal layer. For example, we have previously shown that basal cells of the mouse and human epidermis can be divided into different populations on the basis of α6 integrin (α6) and transferrin receptor (CD71) expression. Using these markers, we clearly identified three populations in the interfollicular epidermis: α6bri CD71dim cells, which are rare, undifferentiated, quiescent, long-lived cells with high proliferative potential in vitro, representing putative stem cells; α6bri CD71bri cells, which are a rapidly cycling population containing the majority of basal cells and representing transit-amplifying cells; and α6dim cells, which are the early differentiating cells [4, 8]. Given the structural similarity between interfollicular epidermis and esophagus, we have investigated the usefulness of these markers in fractionating murine basal esophageal cells into primitive stem cells and their more committed progeny. We initially sought to distinguish esophageal cells using in vivo cell turnover studies and then endeavored to confirm these differences using in vitro, and tissue-reconstitution, assays.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Isolation of Primary Murine Esophageal Keratinocytes

To obtain esophageal keratinocytes, female C57/B6 mice, 7–10 weeks of age, were culled at 6:00 p.m. (unless otherwise indicated), and the esophagi were removed and incubated in dispase II (6 mg/ml; Roche Diagnostics Australia Pty. Ltd., Castle Hill, Australia, http://www.rochediagnostics.com.au) in phosphate-buffered saline (PBS) with 12 μg/ml penicillin, 160 μg/ml gentamicin, and 0.6 μg/ml fluconazole for 6–16 hours at 4°C. The mucosa was then removed with sterile forceps and incubated in 5 ml of prewarmed (37°C) trypsin (2.5 mg/ml in PBS; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) for 4 minutes with agitation on a magnetic stirrer with a sterile stir bar. The trypsin was then quenched with 200 μg/ml soybean trypsin inhibitor (Sigma-Aldrich), and the suspension was passed through a 40-μm cell strainer (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com). Cells were centrifuged and resuspended in keratinocyte basal medium (KBM) (Cambrex Bio Science Walkersville, Inc., Walkersville, MD, http://www.cambrex.com) with 20 mg/ml bovine serum albumin (BSA) (Sigma-Aldrich) in PBS. A small proportion of the cells were found to be CD45-positive (0.34% ± 0.05%, mean ± SEM; n = 7) by fluorescence-activated cell sorting (FACS) analysis, and these were gated out in all subsequent experiments. The remaining cells stained positive for keratin-14 [8], confirming their epithelial origin.

Analysis of Cell Surface Markers

Freshly isolated primary murine esophageal keratinocytes (1 × 107 cells per milliliter) were stained with α6-fluorescein isothiocyanate (FITC) (GoH3, 1:100; Becton, Dickinson and Company) and CD71-phycoerythrin (CD71-PE) (C2; 2 μg/ml; Becton, Dickinson and Company) for 30 minutes on ice, washed once, and resuspended at 1 × 107 cells per milliliter in KBM containing 20 mg/ml BSA. Cells were either analyzed on a Becton Dickinson LSRII flow cytometer or sorted using a Becton Dickinson FACSDiva (see supplemental online data for FACS setup). Nonviable cells were excluded using 1 μg/ml viability dye, either 7-amino-actinomycin D or fluorogold (both from Invitrogen Corporation, Carlsbad, CA, http://www.invitrogen.com), and CD45-PE-CY7 (30-F11, 0.4 μg/ml; Becton, Dickinson and Company) was used to exclude CD45-positive cells. Isotype-matched negative controls and single-color controls were included with each experiment.

Bromodeoxyuridine Labeling and Cell Cycle Analysis

C57/B6 mice (7–9 weeks old) were pulse-labeled by a single injection of 100 μg/kg bromodeoxyuridine (BrdU) at 6:00 a.m. At the indicated times after labeling, mice were sacrificed and the esophageal keratinocytes were isolated, sorted based on α6, CD71, and CD45 expression, and fixed in 70% (vol/vol) ethanol. The cells were then washed with PBS and incubated in 10 mg/ml BSA for 1–2 hours. DNA hydrolysis was performed by incubating the cells in 2 M HCl at room temperature for 30 minutes followed by neutralization of the acid with 0.1 M sodium tetraborate (Sigma-Aldrich). The cells were then washed once in PBS and stained with anti-BrdU (B44; 1:100, 25°C, 30 minutes; Becton, Dickinson and Company) and detected with goat anti-mouse immunoglobulin G (IgG)-Alexa Fluor 647 (1 μg/ml; Invitrogen Corporation). DNA was stained using 2 μg/ml 4′,6-diamidino-2-phenylindole (Sigma-Aldrich) so as to simultaneously determine cell cycle status and BrdU incorporation using an LSRII analyzer. Esophageal cells from animals that had not been injected with BrdU were used as controls.

For label-retention studies, groups of mice were injected intraperintoneally with BrdU (100 μg/kg) daily for 7 days, and labeled cells were visualized in tissue sections using a BrdU In Situ Detection Kit (Becton, Dickinson and Company) and a Tyramide Signal Amplification Kit (PerkinElmer Life and Analytical Sciences, Boston, http://www.perkinelmer.com). At various times (1–200 hours) after labeling, individual mice were sacrificed, esophageal keratinocytes were isolated, and the loss of BrdU from sorted cell populations was assessed by flow cytometric analysis using anti-BrdU. Overton histogram subtraction was used to establish the percentage of BrdU-positive cells in experimental versus unlabeled control animals.

Limiting Dilution Assay

Mouse esophageal epithelial cells were sorted directly into 96-well plates coated with 20 μg/ml collagen IV (Sigma-Aldrich) using single-cell mode and doublet discrimination on the FACSDiva. They were plated at 3, 10, and 30 cells per well in replicates of 24 or 48 (except for the CD71+++ population, which was plated at 30, 100, and 300 cells per well). The plates were centrifuged at 200g for 5 minutes to encourage cell adherence. Cells were then cultured in KBM supplemented with 5 μg/ml insulin (Sigma-Aldrich), 0.5 μg/ml hydrocortisone (Sigma-Aldrich), 70 μg/ml bovine pituitary extract (Hammond Cell Tech, Windsor, CA, https://secure2.wahju.com/∼hammond), 10 ng/ml epidermal growth factor (Sigma-Aldrich), 2 mg/ml BSA, 100 ng/ml cholera toxin (Sigma-Aldrich), 5 μg/ml transferrin (Sigma-Aldrich), 20 pM triiodothyronine (Sigma-Aldrich), 2.4 μg/ml penicillin, 32 μg/ml gentamicin, and 0.6 μg/ml fluconazole for 7 days at 37°C, 5% CO2. The plates were then washed, fixed with 5% (vol/vol) formalin, and stained with toluidine blue. Wells were scored as positive when there was at least one colony with at least 32 cells. According to Poisson distribution, at a limiting dilution of one colony, 37% of the wells are expected to be negative. Extrapolation of the 37% negative point on the y-axis intersects the x-axis at the number of cells required to produce a colony [9].

In Vivo Tissue Reconstitution

The ability of the cell populations to regenerate a stratified epithelium was assessed using an in vivo transplant model as described previously [10]. Briefly, 1 × 105 freshly sorted cells from each population were resuspended in 30 μl of KBM with supplements and inoculated into devitalized rat tracheas, which were then implanted under the dorsal skin of SCID (severe combined immunodeficient) mice. The trachea were retrieved after 3–5 weeks, processed as previously described [10], and stained with hematoxylin and eosin.

Immunofluorescence Staining

Cryostat sections of mouse esophagus were fixed in acetone for 10 minutes at −20°C, blocked, and stained with rat anti-mouse CD71-FITC (clone R17 217.1.4; Caltag Laboratories, South San Francisco, CA, http://www.caltag.com), α6-FITC (both at 1:100), or mouse anti-K4 (1:1,000). Sections were counterstained with propidium iodide.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Phenotyping of Mouse Esophageal Keratinocytes

The α6-CD71 phenotype of primary mouse esophageal keratinocytes is shown in Figure 1. It contains four discernable populations: the α6briCD71bri population, which constitutes the majority of isolated cells (72.3% ± 2.1%, mean ± SEM; n = 7), and three smaller populations, α6briCD71dim (7.8% ± 1.2%), α6dimCD71bri (referred to hereafter as α6dim; 8.6% ± 0.9%), and CD71 very bright (CD71+++; 11.2% ± 2.6%).

thumbnail image

Figure Figure 1.. Phenotype of mouse esophageal keratinocytes. Primary mouse esophageal epithelial cells were double-labeled with anti-CD71 (phycoerythrin) and anti-α6 integrin (fluorescein isothiocyanate) and analyzed by flow cytometry. The dot-plot indicating the identified subpopulations is representative of at least seven independent experiments. Abbreviations: α+/71, α6briCD71dim cells; α+/71+, α6briCD71bri cells; α/71+, α6dim cells; 71+++, CD71+++ cells.

Download figure to PowerPoint

Flow cytometry was used to investigate the basic characteristics of the cells in these populations. The α6briCD71dim cells were found to have both the lowest forward scatter (size) and low side scatter (internal complexity) (supplemental online Fig. 8A, 8B). The CD71+++ population contained the largest and most complex cells. This was confirmed by hematoxylin staining of these populations, which demonstrated that the CD71+++ population was enriched for large polyhedral cells with a high cytoplasmic-to-nuclear ratio. The other three populations all contained small round cells (supplemental online Fig. 8C). Consistent with this, flow cytometric analysis of expression of keratin-4, a known marker of differentiation in the human esophagus [11], revealed that increased cell size was correlated with greater differentiation, with the CD71+++ cells containing the highest proportion of keratin-4bri cells and the α6briCD71dim population containing the lowest level of keratin-4 (supplemental online Fig. 9).

Analysis of Cell Cycle

As has been demonstrated in previous studies [12], we detected a strong circadian rhythm to cell proliferation in the mouse esophagus, with 16.7% ± 1.2% (mean ± SEM) of cells cycling at 6:00 a.m. compared with only 1.8% ± 0.4% of cells cycling at 6:00 p.m. (n = 7; p < .001) (supplemental online Fig. 10A, 10B). Assessment of the cell cycle was therefore undertaken at 6:00 a.m., near the peak in the circadian rhythm. As shown in Figure 2A, the α6briCD71bri population was clearly enriched for cycling cells (cells in S, G2, or M phase), whereas the α6briCD71dim, α6dim, and CD71+++ populations had reduced levels of cycling cells. Consistent with this result, the α6briCD71bri population had the greatest percentage of cells incorporating BrdU, thus confirming that this population has the greatest proportion of cycling cells (Fig. 2B).

thumbnail image

Figure Figure 2.. Analysis of BrdU-labeled esophageal keratinocytes. Mice were pulse-labeled by a single intraperitoneal injection of BrdU (100 μg/kg) at 6:00 a.m., and esophageal keratinocytes were collected after 1 hour. Bivariate flow cytometric analysis was used to simultaneously assess (A) the percentage of cells in S + G2M phase of the cell cycle via DNA content and (B) the percentage of BrdU-positive cells in α6-CD71-defined subpopulations (mean ± SEM; n = 7). The percentages of BrdU-positive cells in each of the α6-CD71-defined subpopulations assessed at 1, 12, 24, and 36 hours after BrdU injection are shown in (C) (mean ± SEM; n = 3). Dotted line connected by ○ indicates α6briCD71dim cells; dotted-and-dashed line connected by ♦ indicates α6briCD71bri cells; dashed line connected by ⋄ indicates α6dimcells; unbroken line connected by • indicates CD71+++ cells. Abbreviations: α+/71, α6briCD71dim cells; α+/71+, α6briCD71bri cells; α/71+, α6dim cells; 71+++, CD71+++ cells; BrdU, bromodeoxyuridine; hrs, hours.

Download figure to PowerPoint

Movement of Cells Between Phenotypic Compartments

We pulse-labeled the DNA of cycling cells by injecting animals with a single dose of BrdU and examined the fate of the labeled cells over the next 36 hours. Twenty-four hours after the initial pulse of BrdU, there was a second wave of cell proliferation, in keeping with the circadian rhythm. Using bivariate DNA/BrdU flow cytometric analysis of the total cell population, we were able to show that of the cells in cycle (i.e., in S + G2/M) at this time most (83.5%) were BrdU-negative, indicating that they had not been cycling at the time of the pulse-labeling 24 hours earlier (supplemental online Fig. 10D). This suggests that most cells in the mouse esophagus have a cell cycle time of at least 48 hours. Consistent with this, most (85%) of the BrdU-labeled cells were found in G0/G1 at 24 hours after pulse. Thus, the total amount of BrdU label in the tissue could be expected to remain relatively constant during the 36 hours after the labeled cells have divided, and we would predict that any fluctuations in the proportion of BrdU-labeled cells within the phenotypically discrete fractions over this time period could be attributed to the movement of cells between the subpopulations. Therefore, we examined the percentage of BrdU-labeled cells in the different populations of the mouse esophageal epithelium at 12, 24, and 36 hours after labeling (Fig. 2C). At 12 hours, the percentage of BrdU-positive cells in the α6briCD71dim population approximately doubled, consistent with the cells that incorporated label having now completed traversing the G2/M phase and undergone mitosis. In contrast, the proportion of labeled cells in the α6dim population at 12 hours increased fivefold to 58.8% ± 0.9% (mean ± SEM; n = 4), suggesting that this fraction had accumulated cells that had been cycling at the time of pulse-labeling. In contrast, the percentage of BrdU-positive cells in the actively cycling α6briCD71bri population had not increased but remained essentially constant during this time. This is consistent with the net effect of both an increase in labeled cells due to mitosis and a loss due to movement of cells into the α6dim compartment.

During the interval from 12 to 36 hours after the initial pulse, the percentage of labeled cells in the α6dim population decreased dramatically to 17.1% ± 5.2%; concomitantly, the proportion of BrdU-positive cells in the CD71+++ population increased from less than 10% to more than 40%. Given that the absolute number of cells in both the α6dim and the CD71+++ populations are approximately equivalent and that the level of label in the other two populations, α6briCD71bri and α6briCD71dim, remains essentially constant over this time period, we infer that the loss of labeled cells from the α6dim population is due to the migration of labeled cells into the CD71+++ population.

Label-Retention Studies

Essentially complete labeling of esophageal keratinocytes was achieved by the injection of mice with BrdU daily for 7 days (supplemental online Fig. 11). The proportion of BrdU-positive cells in each population after chase periods ranging from 12 hours to 8 days was then used to establish the turnover time for each population. Using this technique, we were able to show that the α6briCD71dim population retained a greater proportion of the BrdU compared with both the α6briCD71bri and α6dim populations (Fig. 3A). Linear regression analysis demonstrates that the t1/2 (time required for 50% of the cells to become BrdU-negative) is approximately 100 hours for the α6briCD71dim population compared with approximately 75 and 70 hours for the α6briCD71bri and α6dim populations, respectively (Fig. 3B, 3C).

thumbnail image

Figure Figure 3.. Assessing cell turnover by label retention. Mice were labeled by daily injections of BrdU for 7 days. (A): The percentage of BrdU-positive cells in α6-CD71-defined subpopulations of esophageal keratinocytes freshly isolated after a 5-day chase period (mean ± SEM; n = 5). (B, C): Plots illustrating the kinetics of label retention in mouse esophageal cells after various chase times, as indicated, comparing the α6briCD71dim cell population (•, dashed lines) with (B) α6briCD71bri and (C) α6dim cell populations (○, dotted and dashed lines). Data shown are derived from at least three independent mice at each time point. p < .004 (B); p < .001 (C) (paired t test). Abbreviations: α+/71, α6briCD71dim cells; α+/71+, α6briCD71bri cells; α/71+, α6dim cells; BrdU, bromodeoxyuridine; hrs, hours.

Download figure to PowerPoint

Assessment of Clonogenic Potential

The ability of mouse esophageal keratinocytes to form colonies in in vitro culture was assessed using a limiting dilution assay. This assay determines the proportion of cells from each phenotypic fraction that are capable of forming clonal units of proliferation under stringent conditions (i.e., at low density). Using this approach, approximately 1 in 10 cells from the α6briCD71dim, α6briCD71bri, and α6dim populations was capable of forming a colony as compared with 1 in 175 of the CD71+++ cells (Fig. 4).

thumbnail image

Figure Figure 4.. Clonogenicity of α6 integrin-CD71-defined subpopulations of mouse esophageal cells. The indicated numbers of cells per well were plated, and after 7 days, wells were scored as positive if at least one colony with at least 32 cells was present. The number of cells required to produce a colony was calculated from the graph as the point at which 37% of the wells are negative (dotted lines). Limiting dilution graphs of (A) CD71+++ cells (○) compared with basal cells (all other populations) (•) and (B) α6briCD71dim (•) compared with α6briCD71bri (♦) and α6dim (⋄) cells. Data shown are mean ± SEM; n = 3 for CD71+++ and n = 5 for all others. Abbreviations: α+/71, α6briCD71dim cells; α+/71+, α6briCD71bri cells; α/71+, α6dim cells; 71+++, CD71+++ cells.

Download figure to PowerPoint

In Vivo Tissue Reconstitution

We characterized the tissue-reconstituting abilities of the four kinetically and phenotypically distinct populations described above using an in vivo reconstitution model [4, 13]. The α6briCD71dim and α6briCD71bri fractions exhibited excellent tissue regeneration with a thicker epithelium displaying the full range of differentiation seen in normal esophagus, including a polarized basal layer, granular and spinous layers, and finally a cornified layer, whereas the α6dim cells performed poorly (Fig. 5). Notably, the CD71+++ population was not capable of maintaining an epithelium in this experiment.

thumbnail image

Figure Figure 5.. Esophageal tissue reconstitution in vivo. Transverse histological sections of rat trachea lumens lined with epithelium derived from 1 × 105 mouse esophageal cells and harvested at 5 weeks. (A): Total cells (magnification ×2.5). (B): Total cells (magnification ×10). (C): α6briCD71dim cells. (D): α6briCD71bri cells. (E): α6dim cells. (F): CD71+++ cells. Negative control tracheas not inoculated with cells do not form an epithelium (not shown). Scale bar = 100 μm. Abbreviations: c, trachea cartilage; e, epithelium; l, trachea lumen; s, submucosal layer.

Download figure to PowerPoint

In Situ Localization

We used immunofluorescence staining in an attempt to localize the α6/CD71 populations in the mouse esophageal epithelium. Bright staining for α6 integrin was seen throughout the epithelium (supplemental online Fig. 12), consistent with the flow cytometric data. In contrast, we observed differential expression of CD71 along the basal layer of the esophageal epithelium as shown in Figure 6. Notably, bright patches of CD71 staining alternated with regions of CD71dim staining within the basal layer. Interestingly, the CD71bri regions were associated with the presence of suprabasal cells exhibiting even brighter CD71 staining.

thumbnail image

Figure Figure 6.. In vivo localization of CD71 in the murine esophageal epithelium. Immunofluorescence images of a cross-section of mouse esophagus stained with anti-CD71-fluorescein isothiocyanate (green) and counterstained with propidium iodide (red) to identify nuclei. (A): Low-magnification image. (B): High-magnification image. Scale bar = 100 μm.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

The mouse esophageal epithelium can be divided into two compartments: a superficial stratified layer consisting of large, differentiated, and keratinising squamous cells and a basal layer composed of densely packed columnar cells. By analogy with other stratified epithelia, the stem cells responsible for maintaining the esophageal epithelium are thought to reside in the basal layer. Previous work has suggested that all basal cells in the mouse esophagus are equivalent with respect to the probability of undergoing cell division and leaving the basement membrane; indeed, it has been suggested that all basal cells are effectively stem cells [7]. On the other hand, there is evidence from chimeric animal studies consistent with the notion that there are clonal populations within the esophageal basal layer [14], implying that esophageal renewal originates from single stem cells at the apex of a proliferative hierarchy within the basal layer.

We have previously used the cell surface markers α6 and CD71 to identify putative stem cells in the mouse epidermis [4, 8]. We now show that these markers can identify distinct cell populations in the mouse esophageal epithelium and have been able to elucidate a kinetic and functional hierarchy.

The CD71+++ Population Is Enriched for Differentiated, Suprabasal Cells

Cell morphology, flow cytometry data (forward and side scatter), and staining for the differentiation marker keratin-4 are consistent with the conclusion that CD71+++ cells are derived from the suprabasal compartment. This conclusion is supported by immunofluorescence studies demonstrating the highest CD71 staining in the suprabasal layers of the esophageal epithelium (Fig. 6). As would be expected of cells from the differentiated suprabasal compartment, the CD71+++ cells exhibited the lowest clonogenicity in vitro and no tissue-reconstituting activity.

The α6dim Population of Esophageal Epithelial Cells Fulfils Criteria Expected of an EarlyDifferentiating Compartment

By definition, the early differentiating compartment is characterized by cells that have left the cycling pool but are not yet demonstrating all the characteristics of fully mature differentiated cells. Like the CD71+++ cells, the α6dim cells appear to have a poor proliferative activity and a reduced capacity to reconstitute a normal esophageal epithelium in a tissue-reconstitution model (Fig. 7), suggesting that most have left the cell cycle and committed to differentiation. Indeed, although the α6dim cells express relatively low levels of the differentiation marker keratin-4, cell-tracking experiments suggest that these cells will eventually move into the CD71+++ compartment. Thus, the α6dim population appears to be enriched for cells that, having just undergone a cell division (possibly their last), are now destined to leave the basement membrane and differentiate into suprabasal cells. This would be consistent with their cell surface phenotype, as downregulation of α6 expression has been linked to detachment of keratinocytes from the basement membrane in other tissues [4]. Interestingly, the CD71+++ population appears by flow cytometry to be α6 integrin-bright; however, because this technique measures total cell-associated fluorescence, the apparent increase in total α6 may not indicate an upregulation in the density of α6 expression on the cell surface but rather may be a reflection of the increase in cell size.

thumbnail image

Figure Figure 7.. Proposed model of stem cell hierarchy in the esophageal epithelium of the mouse. Diagrammatic representation of the α6-CD71 phenotype illustrating the proposed parent-progeny/differentiation relationships between the α6-CD71-defined subpopulations. Arrows indicate movement of cells from the stem cell compartment, to the transit-amplifying, early differentiating, and finally the suprabasal differentiating compartments in the course of epithelial cell renewal and differentiation. Abbreviations: ED, early differentiating; SC, stem cell; SD, suprabasal differentiating; TA, transit-amplifying.

Download figure to PowerPoint

The α6briCD71bri Population of Esophageal Epithelial Cells Fulfills Criteria Expected of a Transit-Amplifying Compartment

The transit-amplifying cells are the direct progeny of stem cells and form a pool of rapidly dividing cells that expand their own numbers to provide cells to the differentiating suprabasal compartment. The α6briCD71bri population represents the largest proportion of esophageal basal cells and is characterized by a high percentage of cells in S, G2, and M phases of the cell cycle and a high level of incorporation of BrdU following a single pulse. These data are consistent with the α6briCD71bri population representing the transit-amplifying compartment. The α6briCD71bri cells are able to form an epithelium in a tissue-reconstitution assay, indicating that they are not yet committed to differentiation. However, cell-tracking experiments demonstrate that cells from this population will eventually move to the early differentiated population (α6dim) and, ultimately, differentiate into mature suprabasal cells.

The α6briCD71dim Population of Esophageal Epithelial Cells Fulfills Criteria Expected of a Stem Cell Compartment

The stem cell compartment is ultimately responsible for the renewal of tissues undergoing continuous turnover. The defining characteristics of stem cells are sustained self-renewal and tissue regeneration, although these criteria are often difficult to demonstrate experimentally.

Typically, epithelial stem cells tend to be small, quiescent, undifferentiated blast-like cells [8, 15]. Such characteristics most closely describe the α6briCD71dim population. The α6briCD71dim cells retain BrdU significantly longer after pulse-labeling than the other populations (Fig. 3), indicating a slower turnover rate. However, because the entire tissue is turning over very rapidly, even this putative stem cell population has a relatively high rate of cell turnover as compared with the hematopoietic system [16] or the epidermis [15]. The α6briCD71dim cells efficiently regenerate a complete epithelium in the tissue-reconstitution assay (Fig. 5), demonstrating their potency. Furthermore, by serially passaging α6briCD71dim cells through the tissue-reconstitution model, we not only were able to demonstrate the capacity for self-renewal but also increased long-term proliferative capacity compared with the other cell populations (data not shown).

Esophageal Stem Cell Hierarchy

Based on the data presented above, we propose a model of the esophageal stem cell hierarchy (outlined in Fig. 7) in which the α6briCD71dim population represents the stem cell compartment. Cells then sequentially progress through the α6briCD71bri (transit-amplifying) and α6dim (early differentiating) compartments, eventually emerging as mature, differentiated cells that constitute the suprabasal layers of the esophageal epithelium (CD71+++).

Consistent with our model, CD71 expression in the basal layer of the mouse esophagus alternates between areas of low expression and regions of high expression, and the transition between the two appears to be gradual rather than abrupt (Fig. 6). This is consistent with the flow cytometric data, which suggest that there is a continuum of expression levels. We propose that the areas of lowest CD71 expression contain the putative stem cells and that these regions merge into regions of high expression, which contain progressively more committed transit-amplifying (α6briCD71bri) and early postmitotic (α6dim and also CD71bri) cells that ultimately transit into the suprabasal layers where CD71 expression is the highest.

Although our in vivo cell turnover studies were clearly able to differentiate between the α6briCD71dim (stem) and α6briCD71bri (transit-amplifying) cell populations, it has proven difficult to separate these populations using the tissue-reconstitution assay. Similarly, we have previously reported that extensive tissue-regenerative capacity measured over the course of 6–10 weeks can be elicited from keratinocyte stem, transit-amplifying, and early differentiating cells isolated from human foreskin [13]. Most likely, this can be attributed to the gradual loss of “stem-ness” as cells progress through the proliferative hierarchy prior to irreversible commitment to terminal differentiation. This notion was first proposed by Potten and Loeffler in the “diminishing stem-ness spiral model” for epithelia [17] and suggests that the basal cells of the esophagus are likely to be organized in a biological continuum whereby tissue-regenerating potency is lost gradually as cells progress from being stem cells to functional end cells. Furthermore, it is also important to recognize that liberation of cells from their normal tissue constraints, which is required to assay subsets of epithelial cells in vitro and in vivo, may result in revealing their potential ability for proliferation or tissue regeneration and may not necessarily reflect their actual properties in situ.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

We thank Ralph Rossi for invaluable help and advice with flow cytometry and cell sorting and Ivan Bertoncello for valuable discussions. D.C. is the recipient of a Surgeon-Scientist Fellowship from the Royal Australasian College of Surgeons and a postgraduate research scholarship from the National Health and Medical Research Council of Australia.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
Supp_Fig_8.pdf57KSupplemental Figure 1
Supp_Fig_9.pdf71KSupplemental Figure 2
Supp_Fig_10.pdf52KSupplemental Figure 3
Supp_Fig_11.pdf35KSupplemental Figure 4
Supp_Fig_12.pdf73KSupplemental Figure 5
Supp_Legends.pdf64KSupplemental Legends
Supp_Methods.pdf60KSupplemental Methods

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.