CD24 can be used to isolate Lgr5+ putative colonic epithelial stem cells in mice

Abstract

A growing body of evidence has implicated CD24, a cell-surface protein, as a marker of colorectal cancer stem cells and target for antitumor therapy, although its presence in normal colonic epithelium has not been fully characterized. Previously, our group showed that CD24-based cell sorting can be used to isolate a fraction of murine small intestinal epithelial cells enriched in actively cycling stem cells. Similarly, we hypothesized that CD24-based isolation of colonic epithelial cells would generate a fraction enriched in actively cycling colonic epithelial stem cells (CESCs). Immunohistochemistry performed on mouse colonic tissue showed CD24 expression in the bottom half of proximal colon crypts and the crypt base in the distal colon. This pattern of distribution was similar to enhanced green fluorescent protein (EGFP) expression in Lgr5-EGFP mice. Areas expressing CD24 contained actively proliferating cells as determined by ethynyl deoxyuridine (EdU) incorporation, with a distinct difference between the proximal colon, where EdU-labeled cells were most frequent in the midcrypt, and the distal colon, where they were primarily at the crypt base. Flow cytometric analyses of single epithelial cells, identified by epithelial cell adhesion molecule (EpCAM) positivity, from mouse colon revealed an actively cycling CD24(+) fraction that contained the majority of Lgr5-EGFP(+) putative CESCs. Transcript analysis by quantitative RT-PCR confirmed enrichment of active CESC markers [leucine-rich-repeat-containing G protein-coupled receptor 5 (Lgr5), ephrin type B receptor 2 (EphB2), and CD166] in the CD24(+)EpCAM(+) fraction but also showed enrichment of quiescent CESC markers [leucine-rich repeats and immunoglobin domains (Lrig), doublecortin and calmodulin kinase-like 1 (DCAMKL-1), and murine telomerase reverse transcriptase (mTert)]. We conclude that CD24-based sorting in wild-type mice isolates a colonic epithelial fraction highly enriched in actively cycling and quiescent putative CESCs. Furthermore, the presence of CD24 expression in normal colonic epithelium may have important implications for the use of anti-CD24-based colorectal cancer therapies.

Fig. 1.

A–D: mouse colon stained with CD24 antibody. Distal and proximal colon are shown at low (×10) and high (×20) power. E: quantification of CD24 expression as a proportion of total crypt length in proximal (Prox) and distal (Dist) colon. Values are means ± SE (20 crypts per mouse per region, n = 5 mice).

Fig. 2.

Flow cytometric identification of CD24⁺EpCAM⁺ and CD24⁺CD45⁻ fractions in cell preparations from pooled proximal and distal colon. Cells were labeled with epithelial cell adhesion molecule (EpCAM) isotype-phycoerythrin (PE; A), EpCAM-FITC (B), CD45 isotype-PE (C), CD45-FITC (D), CD24 isotype-PE (E), CD24-Pacific Blue (PB; F), CD24-PB and EpCAM-FITC (G), and CD24-PB and CD45-FITC (H). Values represent mean percentage of events falling within their respective quadrants.

Fig. 3.

A–D: mouse proximal and distal colon labeled with 4′,6-diaminido-2-phenylindole (DAPI) and ethynyl deoxyuridine (EdU) at low (×10) and high (×20) power. E: quantification of EdU⁺ colonic epithelial cells by cell position. Values (means ± SE) represent mean number of EdU⁺ cells at each position per 100 colonic crypts (n = 5). Mathematical modeling of data showed significant differences in distribution pattern (P < 0.0001), with proximal colon showing a peak between cell position 6 and 7 and distal colon showing a gradual decline from the crypt base.

Fig. 4.

Flow cytometric identification of a CD24⁺EdU⁺ fraction from colonic epithelium. A and B: cells from proximal and distal colon labeled with CD24-PB and EdU-Alexa 647. Values represent mean percentage of events falling within their respective quadrants (n = 3).

Fig. 5.

A–D: Lgr5-EGFP mouse proximal and distal colon labeled with DAPI and leucine-rich-repeat-containing G protein-coupled receptor (Lgr5)-enhanced green fluorescent protein (EGFP) at low (×10) and high (×20) power.

Fig. 6.

Flow cytometric identification of a CD24⁺Lgr5-EGFP⁺ fraction from colonic epithelium. A and B: epithelial cells isolated from proximal and distal colon of Lgr5-EGFP mice and stained with CD24-PB. Values represent mean percentage of events falling within their respective quadrants (n = 3). C: CD24⁺EGFP⁻, CD24⁻EGFP⁺, and CD24⁺EGFP⁺ single colonic epithelial cells as seen by confocal fluorescence microscopy.

Fig. 7.

A and B: quantitative RT-PCR data showing fold change in gene expression relative to intact colon for CD24⁺ and CD24⁻ fractions of colonic epithelial cells from proximal and distal colon. EphB2, ephrin type B receptor 2; Lrig, leucine-rich repeats and immunoglobin domains; DCAMKL-1, doublecortin and calmodulin kinase-like 1; mTert, murine telomerase reverse transcriptase; Muc2, mucin 2; Chga, chromogranin A; Car2, carbonic anhydrase 2; Acta2, α₂-actin, smooth muscle, aorta. Values are means ± SE (n = 4). nd, Not detected. *P < 0.05 vs. CD24⁻.