Intact function of Lgr5 receptor-expressing intestinal stem cells in the absence of Paneth cells

Abstract

Lifelong self-renewal of the adult intestinal epithelium requires the activity of stem cells located in mucosal crypts. Lgr5 and Bmi1 are two molecular markers of crypt-cell populations that replenish all lineages over time and hence function as stem cells. Intestinal stem cells require Wnt signaling, but the understanding of their cellular niche is incomplete. Lgr5-expressing crypt base columnar cells (CBCs) reside deep in the crypt, mingled among mature Paneth cells that are well positioned for short-range signaling. Partial lineage ablation previously had implied that Paneth cells are nonessential constituents of the stem-cell niche, but recently their absence was reported to interfere with Lgr5(+) CBCs, resurrecting an appealing idea. However, previous mouse models failed to remove Paneth cells completely or permanently; defining the intestinal stem-cell niche requires clarity with respect to the Paneth cell role. We find that Lgr5(+) cells with stem-cell activity cluster in future crypts early in life, before Paneth cells develop. We also crossed conditional Atoh1(-/-) mice, which lack Paneth cells entirely, with Lgr5(GFP) mice to visualize Lgr5(+) CBCs and to track their stem-cell function. In the sustained absence of Paneth cells, Lgr5(+) CBCs occupied the full crypt base, proliferated briskly, and generated differentiated progeny over many months. Gene expression in fluorescence-sorted Lgr5(+) CBCs reflected intact Wnt signaling despite the loss of Paneth cells. Thus, Paneth cells are dispensable for survival, proliferation, and stem-cell activity of CBCs, and direct contact with Lgr5-nonexpressing cells is not essential for CBC function.

Fig. 1.

Characterization of Lgr5⁺ CBCs during development. (A–D) Native GFP staining in Lgr5^GFP-CreER mouse duodenum at P1 (A), P5 (B), P10 (C), and P21 (D). Lgr5⁺ cells localize to intervillus regions from the earliest stages. High-magnification images in the lower row emphasize the clustering of GFP⁺ cells without intervening GFP⁻ cells at P5 and P10. (E–G) Immunostaining with cryptdin Crs4c1 antibody in Lgr5^GFP-CreER mouse duodenum at P10 (E), P15 (F), and P21 (G) showing progressive appearance of Paneth cells. (H) Electron micrograph of an intervillus region at P10 showing clustering of CBCs with characteristic ultrastructural features: a broad base, narrow apical cytoplasm, wedge-shaped basal nucleus (white arrows), supranuclear mitochondria, and absence of secretory granules. Dashed lines delineate five adjacent CBCs. (I and J) YFP staining throughout intestinal villi in Lgr5^GFP-CreER:Rosa26^YFP mice 3 wk after a single tamoxifen injection administered at P5 (H) or P10 (I) indicates stem-cell activity of Lgr5⁺ cells at these early time points, giving rise to villus epithelial cells over several renewal cycles. (Scale bars: 50 μm in A–G, I, and J; 2 μm in H).

Fig. 2.

Lgr5⁺ CBCs occupy the Paneth cell zone in the absence of Paneth cells. (A–I) Costaining of Lgr5-GFP (green) and cryptdin Crs41c (red) in control Lgr5^GFP-CreER;Atoh1^+/+ mice (A–C) shows a typically alternating arrangement of Lgr5⁺ and Paneth cells, whereas uniformly GFP-expressing Lgr5⁺ cells are present in Lgr5^GFP-CreER;Atoh1^flx/flx mice (D–F) 3 mo after tamoxifen injections and in Villin^CreER;Lgr5^GFP-CreER;Atoh1^flx/flx mice (G–I) 3 wk after tamoxifen exposure. Images in A, B, D, E, G, and H show longitudinal sections; images in C, F, and I show cross-sectional views. The images in B, E, and H correspond to those in A, D, and G, respectively, with DAPI stain added to visualize all nuclei. Dotted lines in D–F demarcate Paneth cell-depleted and Paneth cell-replete areas as a result of clonally mosaic Cre expression in Lgr5^GFP-CreER intestines. (J and K) In situ hybridization analysis of Olfm4 showing intact Olfm4⁺ stem cells occupying the full base of Paneth cell-depleted crypts in Villin^CreER;Lgr5^GFP-CreER;Atoh1^flx/flx mice 1 d after five daily tamoxifen injections. (Scale bars: 50 μm; Insets show crypts at 10× higher magnification.)

Fig. 3.

Increased proliferation of Lgr5⁺ CBCs in the absence of Paneth cells. (A) Only occasional Ki67⁺ CBCs are mingled among Paneth cells in Atoh1^+/+ tissue. (B) Ki67⁺ proliferating cells occupy the Paneth cell zone in Villin^CreER;Atoh1^flx/flx intestines 2 wk after tamoxifen injection. This difference is highlighted in the Insets. (C–F) Sequential imaging of GFP expression (C and D) and BrdU immunostains (E and F) reveal an increased S-phase fraction among Lgr5⁺ CBCs in Villin^CreER;Lgr5^GFP-CreER;Atoh1^flx/flx intestines compared with Lgr5^GFP-CreER controls. Dotted lines outline the crypts that express GFP in the CBC compartment in mosaic Lgr5^GFP-CreER intestines. (Scale bars: 50 μm in A–F.) (G) Quantitation of three independent intestines from each group of mice; at least 250 Lgr5⁺ CBCs per sample were analyzed. Error bars represent SD.

Fig. 4.

Stem-cell activity of Lgr5⁺ CBCs in the absence of Paneth cells. (A and B) YFP staining of Lgr5^GFP-CreER;Atoh1^flx/flx;Rosa26^YFP mice 3 mo after tamoxifen exposure shows stem-cell activity of Lgr5⁺ cells, which gave rise to all villus epithelial cells over the long term. Staining with Crs4c1 Ab (red) shows persistent Paneth cells in adjacent crypts that lacked Lgr5-Cre expression and absence of Paneth cell in crypts where YFP tracing occurred; the dotted lines demarcate these Paneth cell-depleted and undepleted zones. (C) YFP staining throughout villi in Villin^CreER;Lgr5^GFP-CreER;Atoh1^flx/flx mice 3 mo after tamoxifen administration indicates intact stem-cell activity over 3 mo in the absence of Paneth cells. (Scale bars: 50 μm.)

Fig. 5.

Intact Wnt target genes in the absence of Paneth cells. (A) Flow cytometry profiles of Lgr5⁺ CBCs (blue boxes) among total crypt cells in Atoh1^+/+ (Left) and Villin^CreER; Atoh1^flx/flx (Right) mice, each carrying the Lgr5^GFP allele. (B) qRT-PCR analysis on Lgr5⁺ CBCs isolated from control Lgr5^GFP-CreER and tamoxifen-treated Villin^CreER;Lgr5^GFP-CreER;Atoh1^flx/flx intestines. Wnt target genes, tested because they are associated with cell proliferation, including CD44, Myc, and Ccnd1, were largely unaffected or even were enhanced. The experiment was done on two independent samples from each group; bars represent mRNA levels relative to the control, which is normalized to 1. Error bars represent SD. (C and D) Immunohistochemical corroboration that Wnt target genes, including proliferation-associated CD44 (C), are preserved in the absence of Paneth cells, although Sox9 (D) levels are subtly reduced. (Arrowhead in D points to a representative cell in the CBC position). Insets show single crypts at high magnification, highlighting the conclusions. (Scale bars: 50 μm.)