Actomyosin-mediated cellular tension drives increased tissue stiffness and β-catenin activation to induce epidermal hyperplasia and tumor growth

Abstract

Tumors and associated stroma manifest mechanical properties that promote cancer. Mechanosensation of tissue stiffness activates the Rho/ROCK pathway to increase actomyosin-mediated cellular tension to re-establish force equilibrium. To determine how actomyosin tension affects tissue homeostasis and tumor development, we expressed conditionally active ROCK2 in mouse skin. ROCK activation elevated tissue stiffness via increased collagen. β-catenin, a key element of mechanotranscription pathways, was stabilized by ROCK activation leading to nuclear accumulation, transcriptional activation, and consequent hyperproliferation and skin thickening. Inhibiting actomyosin contractility by blocking LIMK or myosin ATPase attenuated these responses, as did FAK inhibition. Tumor number, growth, and progression were increased by ROCK activation, while ROCK blockade was inhibitory, implicating actomyosin-mediated cellular tension and consequent collagen deposition as significant tumor promoters.

Figure 1. ROCK activation increases epidermal tissue stiffness and the density and depth of collagen deposition within the skin

(A) Western blot shows Mlc2 Ser19phosphorylation in 4HT treated WT, KD:ER and ROCK:ER keratinocytes. Immunohistochemical analyses show pMypt1 staining in 4HT or Vehicle treated WT, KD:ER and ROCK:ER skin. (B) AFM Force maps of 4HT treated WT and ROCK:ER skin areas enclosed by yellow square as well as picrosirius red staining viewed through parallel or orthogonally polarizing filters. (C) Quantification of Young’s elastic Modulus and % area threshold of collagen in 4HT treated WT and ROCK:ER skin. (D) Maximum projections along three axes of SHG collagen signal from representative 500 μm × 500 μm × 200 μm volumes of 4HT treated WT and ROCK:ER skin. Red double headed arrows show relative thicknesses of collagen layer in 4HT treated WT and ROCK:ER skin. (E) Graphs of collagen coverage values derived from SHG signal by depth (line graph) and at peak value (histogram inset), measured over three tissue volumes each from three mice per cohort. Results are mean ± SEM. Scale bars, 100μm See also Figure S1 and Movies S1, S2 and S3.

Figure 2. ROCK activation within murine skin causes epidermal hyperplasia

(A) Hemotoxylin and eosin (H&E) stained skin sections derived from 4HT-treated WT, KD:ER and ROCK:ER mice. Change in ROCK:ER interfollicular epidermis thickness (arrows) compared to WT and KD:ER skin following five daily topical 4HT treatments is shown in histogram. Measurements were made over five fields with three measurements taken per field. Scale bars, 20μm (B) BrdU incorporation in skin sections derived from 4HT-treated WT, KD:ER and ROCK:ER mice. Histogram shows number of BrdU-labeled cells per mm basement membrane (BM) following 4HT treatment. Scale bars, 50 μm. (C) Histogram shows fold change in number of colonies formed in soft agar by 4HT-treated ROCK:ER keratinocytes relative to KD:ER keratinocytes. Representative fields of colonies are shown. Scale bars, 100 μm. (D) K1, K6 and K14 expression in skin sections derived from 4HT-treated WT, KD:ER and ROCK:ER mice, without or with BAPN treatment to inhibit lysyl oxidase. Scale bars, 100μm. Histogram shows percentage K6 positive cells within epidermis derived from 4HT-treated WT, KD:ER and ROCK:ER mice as measured by FACS (See Figure S2C for profiles). See also Figure S2

Figure 3. ROCK activation in the murine epidermis is associated with Tcf/Lef target gene activation

(A) Expression of CD44 variants v3 and v3-v10 in 4HT-treated KD:ER and ROCK:ER skin by immunofluorescence. Scale bars indicate 100 μm. Histogram shows CD44v3 mRNA expression within WT, KD:ER and ROCK:ER epidermis. (B) Cyclin D1 expression in 4HT-treated ROCK:ER (arrow) and KD:ER skin. Histogram shows Cyclin D1 mRNA expression within WT, KD:ER and ROCK:ER epidermis. (C) Ectopic expression (yellow arrows) of Lgr5 promoter-driven EGFP:Cre within interfollicular epidermis in 4HT-treated KD:ER and ROCK:ER skin. Boxed areas are enlarged in insets. Histogram shows Lgr5 mRNA expression within WT, KD:ER and ROCK:ER epidermis. (D) Fresh-frozen skin from 4HT-treated TOPGAL mice on KD:ER and ROCK:ER backgrounds in which β-Galactosidase activity was visualized using fluorescein di-galactoside. A red filter has been added to monochromatic channel to aid visualization. (E) Tcf/Lef transcriptional activity in ex vivo cultured KD:ER and ROCK:ER keratinocytes assayed using TOPFlash reporter assay. GSK3β inhibitor BIO was used as positive control. See also Figure S3

Figure 4. β-catenin is stabilized and relocalized upon ROCK induction within the interfollicular epidermis

(A) Confocal slice showing β-catenin (red) and E-cadherin (green) distribution in 4HT-treated KD:ER and ROCK:ER skin. (B) Pixel intensity of red (β-catenin), green (E-cadherin) and blue (DAPI) fluorescence measured along red arrows in Figure 4A to assess co-localization. Membrane and nuclear regions are denoted. Red fluorescence is observed within cytoplasm (red arrowheads) and nucleus (blue arrowheads) in 4HT treated ROCK:ER but not KD:ER skin. (C) Pixel intensity of nuclear or total β-catenin staining in 4HT treated KD:ER and ROCK:ER skin sections. Measurements were carried out on 5 fields each of sections derived from three mice per genotype. Western analysis shows β-catenin protein level in pooled epidermal lysates from 4HT treated KD:ER and ROCK:ER mice (3 mice per genotype). (D) FACS of β-catenin staining versus forward scatter of epidermal cells derived from WT, KD:ER and ROCK:ER skin. Histogram shows percentage β-catenin positive cells in each sample. (E) Unphosphorylated (Ser37/Thr41)-β-catenin (red) and phosphorylated (Ser9) GSK3β (green) distribution in 4HT treated KD:ER and ROCK:ER skin, either without or with BAPN treatment to inhibit lysyl oxidase.

Figure 5. ROCK induced epidermal hyperproliferation is β-catenin dependent

Vehicle or 4HT-treated β-catenin^Fl/Fl; K14-Cre:ER mice expressing KD:ER or ROCK:ER as indicated were stained or scored for: (A) β-catenin (B) β-catenin target genes CD44 (top panels) and Cyclin D1 (bottom panels). (C) Ki-67 (top panels) and H&E (bottom panels). Skin thickness differences are indicated by white arrow pairs in H&E panels. (D) Skin thickness (E) Ki67 positive cells per mm basement membrane (F) Collagen area coverage by SHG (G) K14-Cre:ER, Ctnnb1^tm1Mmt/+ mice were treated as indicated with vehicle or 4HT to activate β-catenin. Panels show β-catenin staining and skin thickness (arrows); histogram indicates fold change in skin thickness. Scale bars, 100 μm. See also Figure S4

Figure 6. ROCK induced epidermal hyperproliferation is dependent on actomyosin contractility

(A) Skin sections from ROCK:ER mice treated with Vehicle or 4HT, 4HT+LIMKi or 4HT+Blebbistatin, stained with H&E or for K6, K14, β-catenin or pCfl. Scale bars indicate 100 μm. (B) Fold change in ROCK:ER skin thickness relative to KD:ER skin following 4HT treatment in absence or presence of LIMKi, Blebbistatin or BAPN. (C) BrdU incorporation scored as positive cells per mm BM from untreated, 4HT-treated 4HT+LIMKi or 4HT+Blebbistatin treated ROCK:ER skin. (D) Collagen area coverage in ROCK:ER skin following 4HT treatment in absence or presence of LIMKi or Blebbistatin. See also Figure S5

Figure 7. FAK, Akt and GSK3β phosphorylation increased in ROCK-activated skin

WT, KD:ER and ROCK:ER skin treated with 4HT alone or 4HT plus FAK inhibitor PF-562271 stained for: (A) K14 (red) and FAK pTyr397 (green) (B) K14 (red) and Akt pSer473 (green) (C) GSK3β pSer9 (green) and unphosphorylated β-catenin Ser37/Thr41 (red) (D) Total β-catenin (green) (E) Fold change in ROCK:ER skin thickness relative to KD:ER skin following 4HT treatment in absence or presence of FAK inhibitor PF-562271. Scale bars, 100μm See also Figure S6

Figure 8. ROCK activation accelerates papilloma growth and progression

(A) Timeline of 15 week chemical carcinogenesis and ROCK activation in cohorts of 6 mice of ROCK:ER and KD:ER genotypes. Treatments were: orange arrow – DMBA, black arrows – TPA, green arrows – 4HT or Vehicle. Papilloma to carcinoma conversions in individual ROCK:ER mice denoted by red arrows. (B) Box and whisker plot of DMBA/TPA-induced papillomas per mouse in ROCK:ER and KD:ER mice treated with vehicle or 4HT as indicated. (C) Pie charts of papilloma size distribution at completion of 15 week treatment period in ROCK:ER and KD:ER mice treated with vehicle or 4HT as indicated. Carcinomas are not included. (D) Histogram of DMBA/TPA papilloma burden and carcinoma conversion rate in ROCK:ER and KD:ER mice treated with vehicle or 4HT as indicated. (E) Collagen area coverage measured by SHG in 4HT treated KD:ER or ROCK:ER skin. (F) Pixel intensity of nuclear or total β-catenin staining in 4HT treated KD:ER or ROCK:ER DMBA/TPA papillomas. Measurements were carried out on 7 fields per condition. (G) Timeline of 20 week chemical carcinogenesis and ROCK inhibition in cohorts of 6 WT FVB/N mice. Treatments were: orange arrow – DMBA, black arrows – TPA, grey arrows – Y27632 or Vehicle. Papilloma to carcinoma conversion events are denoted by red arrows. (H) Box and whisker plot of DMBA/TPA-induced papillomas per mouse in vehicle or Y27632 treated WT FVB/N mice. (I) Pie charts of papilloma size distribution at completion of 20 week treatment period in Y27632 or vehicle treated WT FVB/N mice. Carcinomas are not included. (J) Histogram of DMBA/TPA papilloma burden and carcinoma conversion rate in vehicle or Y27632 treated WT FVB/N mice. (K) Collagen area coverage measured by SHG in DMBA/TPA treated skin along with vehicle or Y27632. (L) Pixel intensity of nuclear or total β-catenin staining in vehicle or Y27632 treated DMBA/TPA papillomas. Measurements were carried out on 7 fields per condition. (M) Results of positive pixel analyses of ROCK expression and MYPT phosphorylation in a sample of 40 human epidermal squamous cell carcinomas. (N) Correlation curve of pMYPT and ROCK positive pixel analyses. The 95% confidence interval is denoted by dotted lines. Statistical significance was calculated using Chi-square test. A Pearson product-moment correlation coefficient (r) of 0.7 confirms a large correlation between ROCK and pMYPT positive pixels. See also Table S1