Module 4 · The Blanpain Lab

Skin & Epithelial Stem Cells

The skin is the most regenerative organ in the body. The basal keratinocyte layer turns over every ~30 days; the hair follicle cycles through growth (anagen), regression (catagen), and rest (telogen) thousands of times across a lifetime. Both processes depend on self-renewing stem-cell populations whose identification and characterisation has reshaped adult stem-cell biology. This module treats the two best-characterised epidermal compartments, the interfollicular epidermis and the hair-follicle bulge, and the revisionist committed-progenitor model that has come out of lineage tracing experiments in the Blanpain and Simons labs.

Featured Lecture — Cédric Blanpain

Blanpain’s iBiology lecture is the field-defining overview of skin stem cell biology. He covers the interfollicular epidermis, the hair-follicle bulge, the mammary gland, and the therapeutic promise for regenerative medicine.

1. The Epidermal Architecture

The epidermis is a stratified squamous epithelium with four layers in thin skin (palms/soles have a fifth):

Basal layer (stratum basale): proliferative, attached to basement membrane via integrin α6β4 and hemidesmosomes.
Spinous layer: committed keratinocytes, rich in desmosomes.
Granular layer: keratohyalin granules, loss of nucleus beginning.
Cornified layer (stratum corneum): dead anucleate corneocytes, cross-linked cornified envelope, ~15–30 cell layers thick. The body’s mechanical barrier.

Cells are continuously produced in the basal layer, committed to terminal differentiation as they leave it, and shed from the corneum roughly 4–6 weeks later. The basal layer is therefore permanently in division; this is where the stem cells must be.

2. The Classical (Hierarchical) Model

From the 1970s through the early 2000s, the epidermis was modelled with a hierarchical architecture: rare, slowly-cycling stem cells produced transit-amplifying (TA) cells that underwent several divisions before terminally differentiating. The “epidermal proliferative unit” was a vertical column of ~10 basal cells supporting one column of suprabasal cells. Label-retention experiments with tritiated thymidine supported this picture: a small fraction of basal cells retained label long-term.

This model fit the mouse tail epidermis best and was extrapolated to other body sites. It also matched the clinical experience of burn surgery: the Green/Rheinwald culture method for expanding epidermal stem cells underwrote the first stem-cell therapy — cultured autologous grafts for severely burned patients (Gallico 1984; O’Connor 1981). This remains an FDA-approved treatment today.

3. The Revisionist (Committed-Progenitor) Model

Clayton et al. (Nature 2007), Lopez-Garcia et al. (2010), and follow-up work from the Klein, Simons, and Blanpain labs used Cre-lox lineage tracing and quantitative clone-size distributions to fit the data. Result: in mouse tail and ear interfollicular epidermis, the observed clone-size dynamics are not fit by the hierarchical model — they are fit by a single population of committed progenitors undergoing stochastic fate choice with balanced outcomes.

Specifically: each division of a basal progenitor yields, with probabilities that sum to one:

Two basal progenitors (self-renewal), probability r;
One basal + one suprabasal (asymmetric), probability 1 − 2r;
Two suprabasal (differentiation), probability r.

Because the self-renewal and differentiation probabilities are equal, thepopulation is homeostatic but individual clones undergo neutral drift. The clone-size distribution scales as ~√t, a prediction quantitatively confirmed. The “slow-cycling stem cell” of the old model was, in this reinterpretation, an artefact of label-retention experiments.

Simulation: Clonal Dynamics (Clayton Model)

Python

script.py72 lines

import numpy as np, matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt

# Epidermal clonal dynamics (Clayton et al. 2007; Klein & Simons 2011)
# Committed-progenitor model: single progenitor population with stochastic fate
# Each division produces:
#   - 2 progenitors (prob r)
#   - 1 progenitor + 1 differentiated (prob 1-2r)
#   - 2 differentiated (prob r)
# With r = 0.08 gives stable population, characteristic clone-size scaling

np.random.seed(42)
n_clones = 5000
t_weeks = np.arange(0, 53, 2)
clone_sizes = np.ones((n_clones, len(t_weeks)))

r = 0.08
division_rate = 1.2   # per week
for ti, t in enumerate(t_weeks):
    if ti == 0: continue
    dt = t - t_weeks[ti-1]
    for c in range(n_clones):
        n = int(clone_sizes[c, ti-1])
        if n == 0:
            clone_sizes[c, ti] = 0
            continue
        # Each cell divides with prob division_rate * dt
        surviving_prog = 0
        for _ in range(n):
            if np.random.random() < division_rate * dt:
                outcome = np.random.random()
                if outcome < r: surviving_prog += 2
                elif outcome < 1 - r: surviving_prog += 1
                # else +0
            else:
                surviving_prog += 1
        clone_sizes[c, ti] = surviving_prog

# Statistics
n_surviving = np.sum(clone_sizes > 0, axis=0)
avg_size_surviving = np.array([np.mean(clone_sizes[clone_sizes[:, ti]>0, ti])
                                if np.any(clone_sizes[:, ti]>0) else 0
                                for ti in range(len(t_weeks))])

fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5.5), facecolor='#0a0a1a')
for ax in (ax1, ax2):
    ax.set_facecolor('#111827'); ax.tick_params(colors='#cbd5e1')
    for s in ax.spines.values(): s.set_color('#334155')
    ax.grid(True, color='#334155', alpha=0.3)

ax1.semilogy(t_weeks, n_surviving, color='#fbbf24', lw=2.6, marker='o', markersize=5)
ax1.set_xlabel('Time (weeks)', color='#cbd5e1')
ax1.set_ylabel('Surviving clones', color='#cbd5e1')
ax1.set_title('Clayton 2007: ~exponential clone loss',
              color='#fda4af', fontweight='bold')
# Fitted: n(t) ~ n(0)/sqrt(t)
ax1.semilogy(t_weeks[1:], 0.8*n_clones/np.sqrt(t_weeks[1:]+0.5), 'r--', alpha=0.6, label='~1/sqrt(t)')
ax1.legend(facecolor='#1e293b', edgecolor='#334155', labelcolor='#cbd5e1')

ax2.plot(t_weeks, avg_size_surviving, color='#2dd4bf', lw=2.6, marker='o', markersize=5)
ax2.plot(t_weeks, np.sqrt(t_weeks+1), 'r--', alpha=0.6, label='~sqrt(t) scaling')
ax2.set_xlabel('Time (weeks)', color='#cbd5e1')
ax2.set_ylabel('Mean size of surviving clones', color='#cbd5e1')
ax2.set_title('Stochastic clonal growth: population-asymmetric',
              color='#fda4af', fontweight='bold')
ax2.legend(facecolor='#1e293b', edgecolor='#334155', labelcolor='#cbd5e1')

plt.tight_layout()
plt.savefig('output.png', dpi=120, bbox_inches='tight', facecolor='#0a0a1a')
print('Clayton/Klein/Simons: single committed-progenitor model fits epidermis')
print('~1/sqrt(t) clone loss + sqrt(t) size growth are universal scaling predictions')

Click Run to execute the Python code

Code will be executed with Python 3 on the server

4. The Hair-Follicle Bulge

The hair follicle contains several stem-cell populations, but the bulge — a pocket of cells at the base of the permanent portion of the follicle — is the best-characterised. Bulge cells are marked by K15⁺, Lgr5⁺ (lower bulge), and Lhx2⁺. They are quiescent for most of the hair cycle and activate at anagen onset, fuelling the production of hair matrix cells that extrude as the growing hair shaft.

Bulge cells also contribute to interfollicular epidermis regeneration after injury — a plasticity that is not deployed during normal homeostasis. Similarly, mammary stem cells are heterogeneous and context-dependent: the same cell can act as a unipotent progenitor during homeostasis and a multipotent stem cell after injury (Van Keymeulen 2011).

This plasticity is the central insight of the Blanpain–Fuchs 2014 Sciencereview: stem-cell identity is not a fixed property, and many cells that behave as differentiated progenitors in homeostasis can revert to a stem-cell phenotype under stress. It is also directly relevant to cancer biology — tumour-initiating cells can emerge from differentiated cells via the same plasticity (Module 7).

5. Hair-Follicle Cycling

The hair cycle in mouse has three phases: anagen (growth, ~3 weeks), catagen (regression, ~3 days), telogen (rest, variable). Transitions are controlled by WNT/BMP balance: WNT high + BMP low = anagen activation; WNT low + BMP high = telogen. Hair-cycle dysregulation — by chemotherapy (affecting rapidly dividing matrix cells) or by autoimmune attack on the bulge (alopecia areata) — is a major clinical concern. JAK inhibitors have recently achieved FDA approval for severe alopecia areata, the first effective immunological therapy for the condition.

6. Regenerative-Medicine Applications

The first stem-cell therapy in clinical use — predating iPSCs by 40 years — was the expansion of autologous keratinocytes for burn victims. It remains gold-standard for >80% TBSA burns. More recent applications: Holoclar (autologous limbal stem cell expansion for corneal epithelial reconstruction) was approved in Europe in 2015 for alkali burns of the cornea. Ex vivo gene correction of epidermal stem cells has restored skin function in a child with junctional epidermolysis bullosa (Hirsch & De Luca 2017, Bochum): the patient’s entire epidermis was replaced with a corrected clone from a single skin biopsy. This is the clinical ceiling of the field: a single biopsy, gene correction, and a lifelong reconstituted organ.

← Module 3 Module 5: Haematopoietic →