Opposing roles for ADAMTS2 and ADAMTS14 in myofibroblast differentiation and function

Abstract

Crosstalk between cancer and stellate cells is pivotal in pancreatic cancer, resulting in differentiation of stellate cells into myofibroblasts that drives tumour progression. To assess cooperative mechanisms in a 3D context, we generated chimeric spheroids using human and mouse cancer and stellate cells. Species-specific deconvolution of bulk-RNA sequencing data revealed cell type-specific transcriptomes underpinning invasion. This dataset highlighted stellate-specific expression of transcripts encoding the collagen-processing enzymes ADAMTS2 and ADAMTS14. Strikingly, loss of ADAMTS2 reduced, while loss of ADAMTS14 promoted, myofibroblast differentiation and invasion independently of their primary role in collagen-processing. Functional and proteomic analysis demonstrated that these two enzymes regulate myofibroblast differentiation through opposing roles in the regulation of transforming growth factor β availability, acting on the protease-specific substrates, Serpin E2 and fibulin 2, for ADAMTS2 and ADAMTS14, respectively. Showcasing a broader complexity for these enzymes, we uncovered a novel regulatory axis governing malignant behaviour of the pancreatic cancer stroma. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Keywords: 3D in vitro models; TGFβ; cancer invasion; cancer-associated fibroblasts; cellular cross talk; extracellular matrix; myofibroblast; pancreatic cancer; protease.

Figure 1

Chimeric spheres reveal cancer and stellate cell transcriptomes that underpin 3D invasion. (A) Schematic of spheroid invasion model. Stellate and cancer cells are formed into spheres using methylcellulose hanging drops, which are then placed in a 3D matrix and cultured for 3 days. (B) Brightfield and confocal images of chimeric spheres. Top panels, human cancer cells (Miapaca2; H2B‐RFP, purple) cocultured with mouse stellate cells (PSC; H2B‐GFP, yellow). Lower panels, mouse cancer cells (DT6066; H2B‐RFP, grey) mixed with human stellate cells (PS1; H2B‐GFP, green). Images representative of at least three biological replicates. Confocal images are representative collapsed z‐projections. Scale bar, 100 μm. (C) Schematic of transcriptomic approach. Spheroids are processed for bulk RNA sequencing and reads mapped to either human or mouse genome, providing cancer and stellate cell information. Cell information is then compared with the opposing cell type of the same species from corresponding spheroids. (D) Volcano plot of genes differentially regulated between stellate and cancer cells from human dataset. (E and F) Enriched gene sets in (E) human cancer cell and (F) human stellate cell datasets. Schematics created with BioRender.com.

Figure 2

ADAMTS2 and ADAMTS14 have opposing roles in invasion. (A) Heatmap of metzincin expression in human dataset from chimeric spheroids. (B) Volcano plot of metzincins identified in A using RNA sequencing information from laser dissected PDAC tumour and stromal compartments. Data obtained from [23]. (C) Brightfield and confocal images and quantification of invasion and central area from Miapaca2 (H2B‐RFP, purple): mouse stellate cell (mPSC; H2B‐GFP, yellow) spheroids with siRNA knockdown of either Adamts2 (ts2) or Adamts14 (ts14) specifically in stellate cells. Scale bar, 100 μm. (D) Schematic of Boyden chamber migration assay. Fluorescently labelled Miapaca2 cancer and mouse stellate cells were added to the apical chamber and migration to the basal surface monitored. (E) Kinetics and area under the curve measurements of total cell migration (Miapaca2 + stellate cell) with stellate cell‐specific knockdown of either Adamts2 or Adamts14. Images representative of three biological repeats. Confocal images are representative collapsed z‐projections. Individual colours represent distinct biological repeats. ****p < 0.0001, **p < 0.01, *p < 0.05. One‐way analysis of variance (ANOVA) with Dunnett's post hoc test. Schematic created with BioRender.com.

Figure 3

ADAMTS2 and ADAMTS14 have opposing roles in myofibroblast differentiation. (A) Confocal images of actin (purple) and αSMA (white) expression in stellate cells with knockdown of either ADAMTS2 or ADAMTS14. Scale bar, 20 μm. (Ai) Quantification of αSMA fibre intensity per cell from A. (B) Confocal images of αSMA expression in human PS1 stellate cells following knockdown of ADAMTS2 and stimulation with 5 ng/ml TGFβ for 48 h. Nuclei presented in green (H2B‐GFP). (Bi) Quantification of αSMA fibre intensity per cell from B. (Bii) Western blot for αSMA expression in human PS1 stellate cells following knockdown of ADAMTS2 and stimulation with 5 ng/ml TGFβ for 48 h. (C) Confocal images of αSMA in human PS1 stellate cells following knockdown of ADAMTS14 and treatment with 10 μm SB431542 (TGFβR inhibitor) for 48 h. Nuclei presented in green (H2B‐GFP). (Ci) Quantification of αSMA fibre intensity per cell from C. (Cii) Western blot for αSMA expression in human PS1 stellate cells following knockdown of ADAMTS14 and treatment with 10 μm SB431542 (TGFβR inhibitor) for 48 h. (D) SMAD reporter luminescence in 1,089 myoepithelial cells exposed for 24 h with conditioned medium from human PS1 stellate cells with indicated ADAMTS knockdown. Data presented as SMAD Firefly luminescence relative to control Renilla luminescence and normalised to conditioned medium from control stellate cells. (E) Representative z‐stack images of stellate cells with indicated knockdown and expressing a CAGA‐eGFP reporter construct embedded in collagen I: Matrigel hydrogels and cultured for 72 h. eGFP fluorescence presented in grey and cell nuclei in green. (Ei) Quantification of eGFP fluorescence from E. Data are presented as mean fluorescence intensity per cell normalised to respective background and control cells. Images representative of at least two biological repeats. Confocal image quantification preformed on at least five fields of view per sample. Individual colours represent distinct biological repeats. Densitometry of αSMA expression relative to HSC70 and normalised to respective control presented beneath blot. ****p < 0.0001, **p < 0.01, *p < 0.05. One‐way ANOVA with Dunnett's post hoc test. Scale bar, 20 μm.

Figure 4

Loss ADAMTS2 and ADAMTS14 produce distinct matrisomes with enrichment of known substrates. (A) Schematic of proposed role of ADAMTS2 and ADAMTS14 in regulation of myofibroblast differentiation. (B) Schematic of matrisomic approach. (C) PCA plot of matrisome expression following knockdown of either ADAMTS2 or ADAMTS14 in human PS1 stellate cells. (D and E) Volcano plot of differentially expressed matrisome proteins following knockdown of either (D) ADAMTS2 or (E) ADAMTS14. (F and G) Heatmaps of differentially expressed (F) ADAMTS2 and (G) ADAMTS14 substrates identified from matrisome data. Data obtained from at least two biological repeats analysed in technical duplicates. Schematic created with BioRender.com.

Figure 5

The ADAMTS2 substrate Serpin E2 regulates myofibroblast differentiation. (A) Brightfield and confocal images and quantification of invasion and central area from Miapaca2 (H2B‐RFP, purple): mouse stellate cell (mPSC; H2B‐GFP, yellow) spheroids with siRNA knockdown of Adamts2 (ts2) with and without coknockdown of Serpine2. (B) Plasmin activity in mouse stellate cell supernatant 48 h following knockdown of Adamts2 with and without coknockdown of Serpine2. (C) Brightfield images and quantification of invasion and central area of Miapaca2: mPSC spheroids treated with 10 μm aprotinin for 72 h. Scale bar, 100 μm. (D) Representative z‐stack images of stellate cells with indicated knockdown and expressing a CAGA‐eGFP reporter construct embedded in collagen I: Matrigel hydrogels and cultured for 72 h. eGFP fluorescence presented in grey and cell nuclei in green. Scale bar, 20 μm. (Di) Quantification of eGFP fluorescence from (D). Data are presented as mean fluorescence intensity per cell, normalised to respective background and control cells. Images representative of at least two biological repeats. Individual colours represent distinct biological repeats. ***p < 0.001, **p < 0.01, *p < 0.05. One‐way ANOVA with Dunnett's post hoc test. (E) Schematic proposing a role for ADAMTS2 and Serpin E2 in myofibroblast differentiation. ADAMTS2 degrades Serpin E2, which normally inhibits the action of Urokinase Plasminogen Activator (uPA). uPA catalyses the conversion of plasmin from plasminogen, which releases latent‐bound TGFβ. Loss of ADAMTS2 enhances Serpin E2 function, diminishing the release of active TGFβ. Schematic created with BioRender.com.

Figure 6

ADAMTS14 regulates myofibroblast differentiation through fibulin 2. (A) Brightfield and confocal images and quantification of invasion and central area from Miapaca2 (H2B‐RFP, purple): mouse stellate cell (mPSC; H2B‐GFP, yellow) spheroids with siRNA knockdown of Adamts14 (ts14) with and without coknockdown of Fbln2. Scale bar, 100 μm. (B) Area under the curve analysis of cancer and stellate cell migration with mouse stellate cell knockdown of Adamts14 alone or in combination with Fbln2 knockdown. (C) Western blot for αSMA and fibulin 2 expression in human PS1 stellate cells with knockdown of ADAMTS14 alone or in combination with FBLN2 knockdown. Densitometry of αSMA and fibulin 2 expression relative to HSC70 and normalised to respective control is presented beneath the blot. (D) Confocal images of αSMA expression in stellate cells following knockdown of ADAMTS14 alone or in combination with FBLN2 knockdown. Scale bar, 20 μm. (Di) Quantification of αSMA fibre intensity per cell presented in (D). Confocal image quantification performed on at least five fields of view per sample. (E) SMAD reporter luminescence in 1,089 myoepithelial cells exposed for 24 h to conditioned medium from human PS1 stellate cells with FBLN2 knockdown with and without coknockdown of ADAMTS14. Data presented as SMAD Firefly luminescence relative to control Renilla luminescence and normalised to conditioned medium from control stellate cells. (F) Representative z‐stack images of stellate cells with indicated knockdown and expressing a CAGA‐eGFP reporter construct embedded in collagen I: Matrigel hydrogels and cultured for 72 h. eGFP fluorescence presented in grey and cell nuclei in green. Scale bar, 20 μm. (Fi) Quantification of eGFP fluorescence from (F) Data are presented as mean fluorescence intensity per cell normalised to respective background and control cells. Images representative of at least two biological repeats. Individual colours represent distinct biological repeats. ****p < 0.0001, ***p < 0.001, *p < 0.05. One‐way ANOVA with Dunnett's post hoc test. (G) Schematic of possible role for ADAMTS14 and fibulin 2 in myofibroblast differentiation. Fibulin 2 and the TGFβ large latent complex compete for binding to fibrillin. In the absence of ADAMTS14, fibulin 2 outcompetes TGFβ large latent complex binding to fibrillin, releasing active TGFβ into the milieu. Schematic created with BioRender.com.