A transit-amplifying progenitor with biphasic behavior contributes to epidermal renewal

Abstract

Transit-amplifying (TA) cells are progenitors that undergo an amplification phase followed by transition into an extinction phase. A long postulated epidermal TA progenitor with biphasic behavior has not yet been experimentally observed in vivo. Here, we identify such a TA population using clonal analysis of Aspm-CreER genetic cell-marking in mice, which uncovers contribution to both homeostasis and injury repair of adult skin. This TA population is more frequently dividing than a Dlx1-CreER-marked long-term self-renewing (e.g. stem cell) population. Newly developed generalized birth-death modeling of long-term lineage tracing data shows that both TA progenitors and stem cells display neutral competition, but only the stem cells display neutral drift. The quantitative evolution of a nascent TA cell and its direct descendants shows that TA progenitors indeed amplify the basal layer before transition and that the homeostatic TA population is mostly in extinction phase. This model will be broadly useful for analyzing progenitors whose behavior changes with their clone age. This work identifies a long-missing class of non-self-renewing biphasic epidermal TA progenitors and has broad implications for understanding tissue renewal mechanisms.

Keywords: Lineage tracing; Mathematical modeling; Mouse; Skin; Stem cells; Transit-amplifying progenitor.

Fig. 1.

Aspm is expressed in a distinct subset of highly proliferative basal cells. (A) scRNA-seq feature plots of Sca1⁺/α6-integrin⁺ basal layer (BL) cells sorted from mouse tail skin at PD52. *Data extracted from Ghuwalewala et al., 2022 scRNA-seq database. Shades of blue show expression levels. Dlx1 and Aspm mark the BL subsets employed in this study, K14 marks the BL, K10 and Involucrin mark differentiating cells, and Ki67 marks proliferative cells (Hsu and Fuchs, 2022). (B) The scRNA-seq BL data in A were analyzed to determine the fractions of Dlx1⁺ cells that co-expressed the indicated genes. (C) As in B, except that the Aspm⁺ cells were analyzed. (D) Cell cycle phase PCA analysis of BL cells. (E) Fractions of distinct BL populations in the cell cycle phases. (F,G) Differentially expressed genes (DEGs) that were upregulated (UP) in the indicated cells as analyzed using GO Ontology (2023) identifiers. DEGs in G were extracted by comparing each population with all BL Ki67⁺ cells. Enrichr combined scores (indicated by color) (Edward et al., 2013) and gene expression ratio (indicated by ellipse size) are shown. The x-axis in F shows the negative log₁₀ of the P-value adjusted for multiple testing (Benjamini-Hochberg correction).

Fig. 2.

Aspm-CreER genetic marking of basal cells demonstrates contribution of the Aspm⁺ progenitors to tail epidermis homeostasis and wound repair. (A) Schematic of high tamoxifen (TM) dose lineage tracing in Aspm-CreER mice. Each arrow indicates one injection/day of 100 µg/g body weight TM. (B) Images of immunofluorescent (IF)-stained tail skin in tissue sections (upper panel) and whole mount (lower panel) from mice in A show robust epidermal tdTomato labeling. No leaky tdTomato expression was observed in TM negative (TM⁻) control mice analyzed at the beginning of the chase (left panel) or at subsequent time points (Fig. S3A). Asterisks indicate autofluorescence from the cornified envelope seen in images taken at high exposure. Arrowheads indicate tdTomato⁺ basal cells. Dashed line encircles scales (white) and interscales (yellow). Starting at 2 weeks, images were taken at low exposure to avoid saturation due to the bright accumulation of tdTomato signal. (C) Schematic of high dose TM injections for Aspm- and Dlx1-CreER lineage traced mice (top) used for Ki67 IF staining (bottom). Each arrow indicates one injection/day. Quantification of images like those in C (right panel). Error bars represent s.d. and P-values were calculated using a two-tailed unpaired Student's t-test from n=3 mice and 5-8 images per mouse. (D) Schematic of the tail punch wound experiment. Each arrow indicates one injection/day. (E) Top-view images of tail wounds showing tracks of the tdTomato signal from the Aspm- and Dlx1-CreER lineages at indicated times after injury. (F) TdTomato signal intensities within the wounded areas shown in E (broken outlines) relative to the intensities at day 0. N=4 mice/group, with low TM 3 week (n=2) and 5 week (n=2) mice combined into one group. P-values were calculated by a two-tailed unpaired Student's t-test. Scale bars: 100 µm (B); 50 µm (C); 200 µm (E).

Fig. 3.

Clonal genetic lineage tracing for Dlx1-CreER and Aspm-CreER epidermal progenitors. (A) Schematic of low dose tamoxifen (TM) induction in mice injected once at PD35 and sacrificed at the chase times indicated. (B) xyz orthogonal projections through optical z-stacks from whole-mount skin collected from the chased mice stained with β4-integrin to mark the basal layer show the basal and suprabasal tdTomato⁺ clones and cells. (C,D) Beeswarm plots of tdTomato⁺ clone and cell counts in images like those shown in B from comparable tail skin areas of Aspm-CreER^{low TM} and Dlx1-CreER^{low TM} lineage traced mice (Table S1). Numbers at the top indicate the total number of clones counted.

Fig. 4.

Dlx1-CreER and Aspm-CreER clonal long-term lineage tracing data. (A,B) The total (solid) and BL (dashed) relative labeled cell fractions computed from the data in Fig. 2C,D. (C,D) The average numbers of labeled BL cells/clone (solid; left ordinate) and surviving fractions of labeled BL clones (dashed; right ordinate). The arrows in B and D point to the Aspm-CreER^{low TM} transition between early and late phase. (E) Summary of data in A-D. P-values testing the similarity of the Dlx1- and Aspm-CreER^{low TM} relative cell fractions, BL clone survival, and average BL clone size were all <10⁻¹⁰. Best-fit, standard error, and P-value calculations are described in supplementary Materials and Methods, ‘Analysis of Lineage Tracing Data’ and ‘Statistical Methods’.

Fig. 5.

Generalized birth-death analysis of Aspm-CreER^{low TM}-labeled cells and biological BL clone dynamics in homeostasis. (A) Simplified schematic of biological processes considered in ‘birth-death’ modeling of clonal lineage tracing. The effect on each process on the observed data on the timescale of lineage tracing is shown; e.g., B0/T+1 means that BL clone size and cell fraction is unchanged while total cell fraction increases (for an expanded version see Fig. S8). (B,C) Cumulative labeled clone size distributions at the indicated chase times (see also Fig. S9). is the fraction of clones with size ≤n, where . is the average clone size at the specified chase-time. Dots indicate experimental values. Solid lines are the neutral competition predictions of Eq. 2. The dashed black lines (identified by arrows) mark the asymptotic ‘neutral drift’ limit of Eq. 1, which is reached by the large Dlx1-CreER^{low TM}-marked clones but not by the Aspm-CreER^{low TM}-marked clones. (D) Basal labeled cell fraction (Fig. 4B) and basal labeled clone size data (Fig. 4D) were used together to compute the best-fits for Aspm-CreER^{low TM} progenitor birth and death rates, with a transition time at 45 days (supplementary Materials and Methods, ‘Analysis of Lineage Tracing Data’). The predictions for three models encompassing potential CreER labeling biases are compared with the data (see Fig. S10 for model-predicted labeled BL cell fractions). Likelihood ratio tests reject the ‘post-transition’ cell labeling model, which is equivalent to a monophasic model, with (1−P-value) confidence levels of >0.994. (E) Predicted average number of BL Aspm-CreER^{low TM} cells descended from a single nascent cell as a function of clone age (i.e. time nascent cell introduction of the clone's nascent cell founder) for the models indicated. The amount of pre-transition progenitor amplification and the percentage of cells in the amplification phase during homeostasis are shown for the two labeling models that are consistent with the data. (F) Summary of stem and progenitor cell behavior during homeostasis of adult epidermis.