A basal cell defect promotes budding of prostatic intraepithelial neoplasia

Abstract

Basal cells in a simple secretory epithelium adhere to the extracellular matrix (ECM), providing contextual cues for ordered repopulation of the luminal cell layer. Early high-grade prostatic intraepithelial neoplasia (HG-PIN) tissue has enlarged nuclei and nucleoli, luminal layer expansion and genomic instability. Additional HG-PIN markers include loss of α6β4 integrin or its ligand laminin-332, and budding of tumor clusters into laminin-511-rich stroma. We modeled the invasive budding phenotype by reducing expression of α6β4 integrin in spheroids formed from two normal human stable isogenic prostate epithelial cell lines (RWPE-1 and PrEC 11220). These normal cells continuously spun in culture, forming multicellular spheroids containing an outer laminin-332 layer, basal cells (expressing α6β4 integrin, high-molecular-weight cytokeratin and p63, also known as TP63) and luminal cells that secrete PSA (also known as KLK3). Basal cells were optimally positioned relative to the laminin-332 layer as determined by spindle orientation. β4-integrin-defective spheroids contained a discontinuous laminin-332 layer corresponding to regions of abnormal budding. This 3D model can be readily used to study mechanisms that disrupt laminin-332 continuity, for example, defects in the essential adhesion receptor (β4 integrin), laminin-332 or abnormal luminal expansion during HG-PIN progression.

Keywords: Integrin; Laminin; Neoplasia; Prostate; Spheroids.

Fig. 1.

Human HG-PIN in tissue and the focal absence of α6β4 integrin expression. (A,B) Human prostate tissue was stained for HMWCK (brown stain) to mark basal cells and α-methylacyl CoA racemase (P504S, red stain) to mark luminal cells (Kumaresan et al., 2010). (A) Representative image showing continuous distribution of basal cells at the base of normal prostate gland (N), discontinuous distribution of basal cells at the base of the gland, and expansion of cells into the lumen in high-grade PIN (HG-PIN) and prostate carcinoma (Ca). Note the loss of basal cell layer in cancer. (B) Higher magnification of region in A illustrating basal cell (brown stain) attenuation and continuity gaps in HG-PIN as compared to normal gland. (C,D) Human prostate tissue stained for integrin β4, showing basal cell distribution as in A and B. Note the budding of HG-PIN into muscle stroma (asterisk). Scale bars: 100 µm.

Fig. 2.

3D morphogenesis of PIN arising from single cells. (A) RWPE-1 acini growth from a single cell after the indicated number of days (d) in culture. Scale bars: 20 µm. (B) A day-14 RWPE-1 spheroid stained for β-catenin (red), and ZO1 (green). Scale bar: 10 µm. (C) Quantification of the number of RWPE-1 spheroids with lumens on day 10. Data are mean±s.d. from three experiments, n=100. **P=0.0088 (two-tailed t-test). (D) Spheroids from RWPE-1 cells on day 14 (top row) and PREC 11220 cells on day 10 (bottom row) stained for with laminin-332 (L332, red), integrin β4 (red), HMWCK (green) and p63 (red). Scale bars: 10 µm. (E) Day-14 RWPE-1 spheroids without DHT (left) and with DHT (right) stained for cytokeratin 5 and 14 (CK, green) and PSA (red). Scale bars: 10 µm. (F) DHT treatment induces a twofold increase in PSA production as determined by a quantification of signal intensity for cytokeratin 5 and 14 distribution (green) and PSA distribution (red) across the images in E. The PSA intensity signal relative to cytokeratin signal is indicated by the size of double-headed arrow. (G) Single-plane images of day-10 spheroids. Staining for CEP135 (left panel, red) and α-tubulin (left panel, green) reveals the spindle orientation. The right-hand panel shows spheroids stained for α-tubulin (green), laminin-332 (red) and DNA (blue) to highlight spindle orientation relative to laminin-332. Note that cell division parallel to basement membrane (laminin-332) results in two cells that are located side by side, whereas cell division perpendicular to laminin-332 results in cells located in the lumen. Scale bars: 10 µm (main images), 1 µm (inset in top left). (H) Radial histograms (rose plots) showing mitotic spindle angle relative to the laminin-332 layer of dividing cells in RWPE-1 spheroids on day 10; n=33 spindles. Nuclei in all panels were stained with Hoechst 33342 (blue).

Fig. 3.

3D PIN Model – absence of α6β4 integrin expression and discontinuous laminin-332 layer. (A) Cell lysates from shRNA-expressing RWPE-1 cells were probed for β4 integrin and α-tubulin; the position of molecular mass markers is shown in kDa. (B) Day-10 WT RWPE-1 spheroids immunostained for laminin-332 (L332, red) and β4 integrin (β4, green) showing colocalization of laminin-332 and β4 integrin (yellow). (C) Day-10 RWPE-1 spheroids immunostained for laminin-332 (red) showing a continuous layer in WT RWPE-1 spheroids (top panel) and a discontinuous distribution of laminin-332 (white arrowheads) in β4 integrin-depleted spheroids (middle and bottom panels). DNA (blue). Scale bars: 10 µm. (D) Detection of the β3 chain of laminin-332 and α-tubulin in wild-type (WT), shRNA control (sh-Ctrl) and β4-integrin-depleted RWPE-1 lysates. (E) Percentage of continuous laminin-332 assembly in day-10 WT and β4 integrin-depleted spheroids. Data are mean±s.d. from three experiments, n=100. **P<0.001 (two-tailed t-test).

Fig. 4.

Active budding of basal cells through the discontinuous laminin-332 layer occurs in the absence of β4 integrin expression in RWPE-1 spheroids. (A) Day-10 β4 integrin-depleted (sh-β4) spheroid showing areas of budding (white arrowheads). Basal cells are immunolabeled for HMWCK (green) and the laminin-332 layer (L332, red). DNA, blue. Scale bar: 10 µm. (B) Day-10 WT and β4-integrin-depleted (sh-β4-1, sh-β4-2) spheroids. Cells were immunostained for α6 integrin (α6, red). Areas of budding are indicated by white arrowheads. DNA, blue. Scale bar: 10 µm. (C) Time series of WT (top row, Movie 4) and β4-integrin-depleted spheroids (middle and bottom rows, Movies 5 and 6) showing examples of active cell budding time stamps as hours:minutes. Scale bars: 40 µm. (D) Quantification of the percentage of spheroids with the invasive budding phenotype in WT and β4-integrin-depleted (sh-β4) spheroids. Data are mean±s.d. from three experiments, n=100. **P=0.0016 (two-tailed t-test).