Post-surgical adhesions are triggered by calcium-dependent membrane bridges between mesothelial surfaces

Abstract

Surgical adhesions are bands of scar tissues that abnormally conjoin organ surfaces. Adhesions are a major cause of post-operative and dialysis-related complications, yet their patho-mechanism remains elusive, and prevention agents in clinical trials have thus far failed to achieve efficacy. Here, we uncover the adhesion initiation mechanism by coating beads with human mesothelial cells that normally line organ surfaces, and viewing them under adhesion stimuli. We document expansive membrane protrusions from mesothelia that tether beads with massive accompanying adherence forces. Membrane protrusions precede matrix deposition, and can transmit adhesion stimuli to healthy surfaces. We identify cytoskeletal effectors and calcium signaling as molecular triggers that initiate surgical adhesions. A single, localized dose targeting these early germinal events completely prevented adhesions in a preclinical mouse model, and in human assays. Our findings classifies the adhesion pathology as originating from mesothelial membrane bridges and offer a radically new therapeutic approach to treat adhesions.

Fig. 1. Microcarrier model recapitulates physiological adhesions.

a Overview of the bead assay. b 15 min desiccation shock induces carrier-to-monolayer aggregation, which develops as fast as 60 min after injury. Three biological replicates. c, d Nanoluciferase assay to measure adhesion propensity. Desiccation shock and talcum powder both induce carrier-to-monolayer aggregation. Four biological replicates; ***p < 0.001, two-tailed Mann–Whitney (c) and Kruskal–Wallis followed by Dunn’s (d). e Side view of live imaged bead-monolayer adhesion showing exerted pulling forces. Scale bar, 10 µm. Representative images of three biological replicates. f Adhesion severity (see Methods) increases with time. Black arrows, suture sites. Red arrows, secondary organ attachments. Four biological replicates; ***p < 0.001, two-tailed Mann–Whitney. g Immunoblot of lysed Met-5A cells, various time points after a 15 min desiccation shock. GAPDH serves as loading control. Representative images of three biological replicates. h Immunoblot of excised murine adhesion tissue, various time points after injury. Representative images of three biological replicates. i Delayed matrix deposition in vitro in stressed carriers (dark spheres)-to-monolayer (Masson Trichrome staining). Representative images of ten biological replicates. Scale bar, 500 µm. j Masson Trichrome of adhesion tissue section 5–14 days after injury. Representative images of three biological replicates. Scale bar, 100 µm. k Schematic overview of the sequence of events characteristic of adhesion development in both the carrier assay and in vivo model. Error bars represent standard error of the mean.

Fig. 2. Mesothelia produce cytoskeletal protrusions to bind and transmit adhesive potential.

a Phase-contrast image and corresponding silhouette of desiccation-stressed and unstressed Met-5A cells seeded on a matrigel bedding. Representative images of four biological replicates. Scale bar, 30 µm. b Whole-mount 3D reflectance confocal live imaging of two stressed mesothelial cells grown on beads, showing connecting nanotubes. Representative images of three biological replicates; Scale bar, 3 µm. c Transient labeling of actin filaments through Lifeact-mCherry showing protrusion networks at carrier–carrier contacts. Representative images of three biological replicates; Scale bar, 10 µm. d Scanning electron microscopy image of healthy and injured mouse peritoneum. Color overlay based on morphology. Representative images of four biological replicates; Scale bar, 5 µm. e Multiphoton image of whole-mount top and side view of healthy and injured PDPN+ peritoneum. SH second harmonics, showing collagen bed. Representative images of three biological replicates; Scale bar, 15 µm. f Multiphoton live video of tamoxifen-treated Procr^{CreERT2-IRES-tdTomato};Rosa26^mTmG mouse peritoneum, showing individual mesothelial cells. Representative images of three biological replicates; Scale bar, 10 µm. g Confocal image of tamoxifen-treated Procr^{CreERT2-IRES-tdTomato};Rosa26^mTmG mouse adhesion tissue section, 5 days after injury. Representative images of three biological replicates; Scale bar, 1000 µm (left) and 50 µm (right). h Overview of the tamoxifen regime and the Procr-DTA transgene. i Adhesion score of tamoxifen-treated Procr-DTA mice 5 days after injury. Three biological replicates; ***p < 0.001, two-tailed Mann–Whitney.

Fig. 3. Protrusions are capable of membrane fusion and transmission of cytosolic contents.

a Whole-mount multiphoton image of bead clusters seeded with Lifeact-mCherry and -eGFP transfected cells, 24 h after desiccation, showing cells with double GFP and mCherry expression. Representative images of three biological replicates; Scale bar, 10 µm. b Adhesion propagation assay (see Methods) with nanoluciferase expressing Met-5A cells, showing transmittance of the adhesion phenotype 24 h after desiccation, and thereafter every 3 h. Three biological replicates; ***p < 0.001, One-way ANOVA followed by Tukey. c Carrier-monolayer confocal image showing exchange of Cre-protein driving dTomato expression in receiving cells after 24 h. Representative images of ten biological replicates; Scale bar, 50 µm. d Whole-mount antibody stain of CD44 (magenta), and lipid membrane dyes Dio (green) and PKH26 (red) 24 h after stress. Red-colored beads were originally stressed by desiccation, whereas green-colored beads never experienced stress. Representative images of three biological replicates; Scale bar, 50 µm (left) and 10 µm (right). Error bars represent standard error of the mean.

Fig. 4. Single-cell RNAseq identifies cytoskeletal effectors as core mesothelial program.

a t-SNE visualization of >16.000 Met-5A cells colored by group (time after injury) assignment. b Principal component analysis of the top 1644 significantly expressed genes showing separation of the 8 h experimental group. c Violin plots showing normalized expression (scored by UMIs) of cytoskeletal and calcium genes peaking 8 h after desiccation shock. d Volcano plot of top regulated genes 8 h after desiccation shock. e, f Gene pathway and downstream effects analysis (IPA). g Confocal image of human CD44+ abdominal adhesion tissue sections immuno-stained for core adhesion markers. Scale bar, 100 µm.

Fig. 5. Blocking protrusions prevents adhesion development.

a In vitro adhesion assay screening using the Prestwick library consisting of 1280 FDA compounds. b Nanoluciferase adhesion carriers-to-monolayer assay 24 h after desiccation, and after treatment with small molecules targeted against core adhesion genes (10 µM). Four biological replicates; ***p < 0.001, two-tailed Mann–Whitney. c Cre-exchange assay (see Methods) with nanoluciferase expressing Met-5A cells after treatment with small-molecule inhibitors for 24 h (10 µM). Four biological replicates; ***p < 0.001, two-tailed Mann–Whitney. d Epi-fluorescent representative images of Met-5A cells stably expressing membranous GFP, stressed with desiccation and treated with small-molecule compounds for 24 h (10 µM), showing impaired protrusion development. Representative images of ten biological replicates; Scale bar, 10 µm. e Nanoluciferase adhesion carriers-to-monolayer assay 24 h after desiccation, and after treatment with small-interfering RNA against core adhesion genes (1 µg). Four biological replicates; ***p < 0.001, two-tailed t test. f Adhesion score 5 days after injury, of mice treated with small-molecule compounds dissolved in 2% cellulose that was applied topically at the injury site once before closure. Four biological replicates; *p < 0.05, **p < 0.01, One-way ANOVA followed by Tukey. g Adhesion tissue sections derived from (f), immuno-stained for EMT and mesenchymal markers. Representative images of four biological replicates; Scale bar, 500 µm (overview) and 50 µm (inlet). Error bars represent standard error of the mean.

Fig. 6. Proposed model for the early events driving adhesiogenesis.

a Injury to a serosal layer induces a dramatic and rapid shift in mesothelial morphology through the formation of cytoskeletal protrusions. These allow for (1) the physical binding and fusion to neighbouring healthy cells (e.g., at apposing serosal surfaces), and (2) transmission of pathological behavior. This initiates organ tethering and rapid spread of adhesions through serosal surfaces. Once established, mesothelia commit EMT and deposit matrix to form a macroscopic scar.