Differentiated Troy+ chief cells act as reserve stem cells to generate all lineages of the stomach epithelium

Abstract

Proliferation of the self-renewing epithelium of the gastric corpus occurs almost exclusively in the isthmus of the glands, from where cells migrate bidirectionally toward pit and base. The isthmus is therefore generally viewed as the stem cell zone. We find that the stem cell marker Troy is expressed at the gland base by a small subpopulation of fully differentiated chief cells. By lineage tracing with a Troy-eGFP-ires-CreERT2 allele, single marked chief cells are shown to generate entirely labeled gastric units over periods of months. This phenomenon accelerates upon tissue damage. Troy(+) chief cells can be cultured to generate long-lived gastric organoids. Troy marks a specific subset of chief cells that display plasticity in that they are capable of replenishing entire gastric units, essentially serving as quiescent "reserve" stem cells. These observations challenge the notion that stem cell hierarchies represent a "one-way street."

Fig. 1. Troy is expressed in chief and parietal cells at the base of corpus glands

(A) Confocal image showing Troy-eGFP expression at the base of corpus glands in a Troy-ki mouse. Projection of six 1µm spaced z-stacks. (B) Troy-eGFP is expressed at gland bottoms throughout the corpus. Black arrows point to the base of the epithelial lining, white arrows to the muscle layer. (C–E) Confocal microscopy reveals that Troy-eGFP⁺ cells are co-expressing either the parietal cell marker H⁺K⁺-ATPase (C) or the chief cell marker gastric intrinsic factor (D). Basal enteroendocrine cells marked by chromogranin A are Troy-eGFP⁻ (E). (F) Electron microscopy of cryo immunogold labeled Troy-eGFP⁺ cells. 15 nm Gold label corresponding to eGFP, visible as black dots. Both chief and parietal cells at the gland base express eGFP and positive cells possess characteristics of maturation specific to that lineage. See also Fig. S1.

Fig. 2. Troy⁺ cells generate all lineages

(A) Expression of the Rosa-LacZ reporter gene in Troy⁺ cells one day p.i. and followed for 1, 3, 6 month and 1.5 years p.i. (B) Whole-mount LacZ staining of Troy-ki stomach 6 month p.i. Lineage tracing is evident in the corpus; pylorus is negative. (C) The position of LacZ⁺ cells quantified for 100 tracing clones at 1 day, 1 month and 3 month p.i.. (D) Numbers of LacZ⁺ cells in 100 tracing clones counted at 1 day, 1 month, 3 month and 1.5 years p.i. as percentage of clones with <3, 3–15 and >15 cells. (E–H) Double/Triple immuno-fluorescence stainings for corpus lineage markers on Troy-ki;Rosa-YFP mice: Chief cells (Pepsinogen C, E), mucus neck cells (GSII, F), parietal cells (VEGFB, G) and isthmus cells (Ki67, H). See also Fig. S2 and S3.

Fig. 3. Mist1⁺ cells generate all lineages

(A) One and three days following Mist1-CreERT2 induction (p.i.), typical corpus gastric units show 1–3 Rosa-mTmG reporter positive epithelial cells, which label with the chief cell marker gastric instrinsic factor (GIF). (B) One month p.i., the vast majority of Mist1 lineage traced cells (green) are basal, GSII (purple) negative chief cells. (C) Occasionally, at 1 month p.i. Mist1⁺ cells generate AAA positive pit cells, GSII positive mucus neck cell and VEGFB positive parietal cells. (D) At 6 months p.i., entirely marked corpus units can be detected, encompassing the GSII-marked mucous neck cells (staining of luminal cells is background), the isthmus and the pit region. Units are visualized as Z-stack reconstructions from thick tissue sections using Zeiss ApoTome (left) and multiphoton (right) technology.

Fig. 4. Transcriptional profile of Troy⁺ chief cells

(A) Unsupervised hierarchical clustering joins the Troy⁺ chief cell arrays with a chief cell array prepared from laser-captured chief cells, while pit and parietal cell arrays cluster in a separate tree (data taken from Gene Expression Omnibus dataset GSE5018 (Ramsey et al., 2007). (B) Log2 ratio of selected genes comparing the Troy⁺ chief cell arrays to arrays from whole corpus glands. Troy⁺ chief cells show enrichment of marker genes for chief, digestive tract stem cells and Wnt target genes. (C) qPCRs performed on a separately sorted set of Troy⁺ chief cells and compared to corpus glands (set to 1) confirms enrichment of all marker genes depicted in B in Troy⁺ chief cells. Data are represented as mean +/− SEM of 3 qPCRs. (D) Comparison of the expression level of the Wnt target gene Axin2 and the stem cell marker Lgr5 between small intestine, gastric corpus and pylorus. Data are represented as mean +/− SEM of 3 qPCRs. (E) LacZ staining of the Axin2-LacZ reporter mouse documents expression of Axin2 in a few gland bottom cells in the gastric corpus (nuclear red counter stain). (F) Endogenous eGFP expression in the Lgr5-DTR:eGFP reporter mouse visualizes Lgr5 expression at the bottom of gastric corpus glands. See also Fig. S4 and Table S1.

Fig. 5. Lineage tracing from in vivo to in vitro reveals Troy⁺ cells as the origin of corpus organoids

(A) Freshly isolated corpus unit 4h after seeding. The unit already starts to degenerate. Note the small cystic structure at the bottom of the gland (arrow). (B) Tracing was induced in Troy-ki/Rosa-YFP reporter mice three days before isolation of corpus glands. Two days after seeding, GFP and YFP signal still overlap, indicating that the tracing clone was still restricted to the Troy⁺ population. (C) One week after seeding, the YFP clones had expanded and occupied large parts of the organoid.

Fig. 6. Single Troy⁺ cells generate gastric organoids in vitro

(A) Two different populations of Troy-eGFP⁺ cells can be distinguished by FACS (gate A and gate B). (B) Sorted 'gate A' Troy-eGFP⁺ cells stain with chief cell marker pepsinogen C,. Gate B (large) Troy-eGFP⁺ cells express parietal cell marker H⁺K⁺-ATPase. (C) Colony formation efficiency of Troy⁺ chief and parietal cells as well as Troy negative cells. The number of organoids was counted at day 7 after seeding. Data are represented as mean ± SEM of 180 seeded wells. (D) Representative example of a single sorted and cultured Troy⁺ chief cell. Organoids where passaged weekly and grown for >6 months. (E) Organoids are highly proliferative. Edu: proliferative cells; cell borders marked by E-cadherin. (F) Dividing (Ki67 positive) chief cells (gastric intrinsic factor (GIF) positive) occur frequently. (G) Mucus neck cells (GSII positive) cells form a coherent distinct domain. (H) Organoids from single sorted Troy-GFP⁺ and Troy-GFP⁻ cells from Troy-ki/Rosa-YFP mice were induced by Tamoxifen during the first 16h of culture. Troy-GFP⁺ derived organoids were homogenously YFP positive, while Troy-GFP⁻ derived organoids were YFP negative. (I) Troy-GFP⁺ consistently showed a more complex growing behavior then Troy-GFP⁻ derived organoids. See also Fig. S5.

Fig. 7. Troy⁺ chief/stem cells activated by depletion of cycling isthmus cells

(A–F) 5-FU treatment ablates proliferating cells in the isthmus. Control (A, B, C) and 5-FU treated (D, E, F) mice analyzed for Ki-67 (A, D, for proliferating cells), GFP (B, E, for Troy⁺ cells), and cleaved caspase 3 (C, F, for apoptotic cells). 5-FU induces apoptosis of cycling isthmus cells. (G–H) Accelerated expansion of Troy initiated lineage tracing upon tissue damage. Control (G) and 5-FU treated (H) mice were analyzed 1 month p.i.. Representative examples of Troy tracings (i, ii) and whole mount pictures from the gastric lumen (iii). Circles indicate LacZ⁺ glands reaching the lumen. (I) The number of tracings that reach the lumen was counted one month p.i. in 5-FU treated vs. untreated mice and shows a 6-fold increase (p<0.0001, t-test). Data are represented as mean +/− SEM of ten sections. (J) The number of LacZ⁺ cells in 100 tracing clones counted 1 month after 5-FU treatment. Percentage of clones with one, two or more than two cells is represented. (K) Time scheme of 5-FU experiment. Troy⁺ cells were labeled on day 0 by tamoxifen induction. Proliferative cells were subsequently ablated by 5-FU treatment three days later and tissues analyzed at 1 month p.i.