Multipotent Basal Stem Cells, Maintained in Localized Proximal Niches, Support Directed Long-Ranging Epithelial Flows in Human Prostates

Abstract

Sporadic mitochondrial DNA mutations serve as clonal marks providing access to the identity and lineage potential of stem cells within human tissues. By combining quantitative clonal mapping with 3D reconstruction of adult human prostates, we show that multipotent basal stem cells, confined to discrete niches in juxta-urethral ducts, generate bipotent basal progenitors in directed epithelial migration streams. Basal progenitors are then dispersed throughout the entire glandular network, dividing and differentiating to replenish the loss of apoptotic luminal cells. Rare lineage-restricted luminal stem cells, and their progeny, are confined to proximal ducts and provide only minor contribution to epithelial homeostasis. In situ cell capture from clonal maps identified delta homolog 1 (DLK1) enrichment of basal stem cells, which was validated in functional spheroid assays. This study establishes significant insights into niche organization and function of prostate stem and progenitor cells, with implications for disease.

Keywords: DLK1; Notch; basal; branch; epithelium; luminal; niche; organoids; prostate; prostate cancer; stem cell.

Graphical abstract

Figure 1

Transmission of mtDNA Identifies Long-Ranging Clones Spanning the Entire Prostate from Proximal Juxta-Urethra Ducts to Distal Acini (A) The prostate comprises 12–18 paired glandular subunits independently draining into the urethra. (B) Two-color enzyme histochemistry simultaneously detects activity of the mtDNA-encoded CCO and nuclear-DNA-encoded succinate dehydrogenase (SDH), with CCO-deficient cells appearing blue and CCO-proficient cells appearing brown. Scale bars, 50 μm. (C) Serial sections are aligned in reconstruction software to generate a 3D wire-frame reconstruction of glandular subunit. The reconstruction is converted to a topographical representation to more clearly illustrate spread of the CCO-deficient patch through ductal epithelium (solid brown line represents homogeneous CCO proficiency; blue-brown dashed line, mosaic CCO-deficiency; solid blue line, homogeneous CCO-deficiency). “X” marks common duct opening onto the urethra. Scale bar, 200 μm. (D–F) Three types of clone pattern distributions were identified: long-ranging clones (95%) composed of proximal to distal contiguous patches (75%) (left), occasional proximal to distal clones with fragmentation (20%) (center), and a single rare example of a distal-only clone (5%) (right).

Figure 2

Directed Flow of Coherent Streams Reveals Proximal Juxta-Urethral Stem Cell Niche Domains (A) Constant circumferential widths of individual clonal patches are generated in the common trunks following a short region of transient expansion (n = 5 clones; n = 3 prostates; error bars represent SEM). A representative histology “filmstrip” is shown, capturing the start of a CCO-deficient patch in sequential z-plane images. Scale bar, 100 μm. (B) Two patterns of clone transmission are seen at duct branching, either unilateral or bilateral flow, with the clone width in parent duct always shared proportionately into the daughter ducts, consistent with directed flow. Scale bars, 100 μm. (C) Proportion of clone width in parent duct compared with the proportion in daughter ducts is maintenance throughout the branching tree. A ratio of cumulative clone width before and after branching is presented. Error bars, SD. (D) Passive and random flow of streams into daughter branches was theoretically modeled according to duct size and fraction of parent duct occupied by the duct. The predicted probabilities of outcome upon branching were consistent with observed data, showing that streams behave and branch neutrally. (E) Laser capture microdissection of a fragmented clone (arrow marking a break in the mosaic competent of the clone) and mitochondrial genome sequencing from “disconnected” CCO-deficient regions (areas 1 and 3) show identical mtDNA mutations affecting components of the respiratory chain that would be consistent with measurable CCO: m.7059G

Figure 3

Multiplicity of Stem Cells and Peripheral Monoclonal Conversion (A) The ductal fraction occupied by a clonal stream saturates to a constant value along the main trunk length (n = 5 clones; n = 3 prostates; error bars represent SD; shaded area shows combined SD across all prostates). (B) The ductal fractions occupied by clonal streams were also relatively constant across patients of different ages (n = 7 clones; n = 7 prostates; error bars, SD). (C) A mean clone width of 2.4% of the total duct equates to 43 (37–50) circumferentially distributed active stem cells generating epithelial streams in cohesive longitudinal “laminar flow.” (D) Despite initial narrow ribbons of clonal streams in the main trunk, peripheral parts of a prostatic subunit can show entire monoclonal conversion. (E) Schematic describing the flow pattern of a clonal stream throughout the prostate and spatially restricted monoclonal enforcement. (F) Total clonal fraction is maintained before and after branching, but the ever-diminishing diameter of ducts proximally causes successive rounds of clonal enrichment in individual branches and thus monoclonal conversion. Scale bars, 100 μm. (G) Theoretic modeling of random redistribution of streams into identical daughter ducts mirrored the observed data, tending toward monoclonality peripherally as a function of increasing branch generation. Error bars, SD.

Figure 4

Clonal Mapping Reveals Proximal Multipotent Basal Stem Cells and Unipotent Luminal Stem Cells, as well as Basal Progenitors Located throughout the Prostate (A) Proximal clones in the truck start within the basal layer and then expand to basal and luminal compartments (n = 37/42). Scale bars, 50 μm. (B and C) Rarer patterns of clonal patch distribution exclusively in the basal (B; n = 3/42) (scale bars, 50 μm) or luminal layers (C; n = 2/42) were also observed and noted to be restricted to the proximal trunk. Scale bars, 50 μm. (D) Proliferation (Ki76) is predominantly confined to basal cells and displays a proximodistal gradient. Some proliferative activity remained in the distal regions, indicative of progenitor activity. Error bars, SD. Scale bars, 50 μm. (E) Cleaved caspase-3 measures showed apoptosis almost exclusively in the luminal compartment cells. Error bars, SD. Scale bars, 50 μm. (F) Proposed lineage hierarchy of homeostasis in the prostate. The progeny of proximal juxta-urethral basal stem cells generate a “conveyer belt” of migratory bipotent basal progenitors, which divide and differentiate to compensate the continual loss of luminal cells through apoptosis. A minor luminal unipotent stem cell pool maintains the most proximal luminal cells, proving only a small contribution to the entire duct, limited by the predominating apoptotic program of daughter luminal cells.

Figure 5

DLK1 Marks Basal Prostate Stem Cells and Defines Niche Microarchitecture In Situ (A) An example of focal cell laser capture from the proximal and distal end of a clonal patch. Scale bars, 50 μm. (B) Pooled RNA sequencing (n = 3) revealed that the proximal start of the clone is associated with marked upregulation of known stem cell markers, including a previously documented putative candidate marker for prostate stem cells (DLK1). (C) A systematic review of published literature of top upregulated gene expressions highlighted stem cell-pathway-associated markers in prostate studies. DLK1 is a cell-surface marker and was therefore selected as candidate for live-cell sorting. (D) Immunofluorescence of CD49f, DLK1, and NOTCH1 expression in juxta-urethra trunk (which co-localized with the start of CCO-deficient clone), an intermediate duct, and terminal acinus reveals distinct patterns of expression within basal and luminal cells. Dashed line indicates the epithelial basement membrane. Scale bars, 20 μm. (E) The juxta-urethral prostate ducts show variable encroachment of urothelium along the longitudinal axis, marked by 34betaE12 (expressed in all layers of urothelium but basal only in the prostate epithelium) and, in the next sequential slide in the z-plane, PSA (prostate luminal cells only). Scale bars, 100 μm. (F and G) Two consecutive sections of the same gland illustrate the urothelial-prostate epithelium boundary, described by (F) 34betaE12 and (G) PSA immunofluorescence in the radial axis, and demonstrate an interdigitating pattern on which DLK1^+ve basal prostate stem cells are positioned. Scale bars, 20 μm. (H) Sketch of the spatial arrangement of cells types at the niche (cross-section along the longitudinal axis of the proximal truck). Prostate stem cells are localized in between urethral and prostatic epithelial interdigitation, giving rise to transiently expanding clonal streams.

Figure 6

DLK1 Basal Cells Demonstrate Stem Cell Function by Generating Prostate Spheroids Ex Vivo (A) Schema outlining the 3D growth potential of selected cells from whole-prostate epithelium sorted into basal DLK1^+ve and DLK1^−ve populations. Sorted basal DLK1^−ve cells were viable for 6 weeks before becoming exhausted. (B) Sphere-forming capacity of whole-prostate basal DLK1^+ve and DLK1^−ve sorted cells, error bars, SEM (^∗∗p ≤ 0.01). (C) Size of spheroid regeneration of basal cells through serial passage. Error bars, SEM. (D) Number of spheres from basal cells through serial passage; P1, first-generation spheroids; P2, second-generation spheroids; P3, third-generation spheroids. Error bars, SEM. (E) Histological patterning of DLK1^+ve basal cell-derived spheroids. Scale bars, 100 μm. (F–Q) Bright-field and immunofluorescence for DLK1, basal (CK5), luminal (PSA, AR, CK8), and apoptosis (cleaved caspase-3) markers for DLK1^+ve basal cells at 8 weeks, scale bars, 50 μm (F–K) and DLK1^−ve basal cells at 6 weeks (L–Q). Scale bars, 50 μm. (R) In the spheroids derived from unsorted cells and sorted fractions DLK1^−ve basal cells (6 weeks) and DLK1^+ve basal cells (8 weeks), AR mRNA was confirmed, albeit at low levels, and is associated with downstream readout of PSA expression. Error bars, SEM. (S) To further validate the functionality of AR, dose-dependent induction of PSA was demonstrated in transiently androgen-starved spheroid culture. Error bars, SEM.