Pharmacological HIF-inhibition attenuates postoperative adhesion formation

Abstract

Peritoneal adhesions represent a common complication of abdominal surgery, and tissue hypoxia is a main determinant in adhesion formation. Reliable therapeutic options to reduce peritoneal adhesions are scarce. We investigated whether the formation of postsurgical adhesions can be affected by pharmacological interference with hypoxia-inducible factors (HIFs). Mice were treated with a small molecule HIF-inhibitor, YC-1 (3-[5'-Hydroxymethyl-2'-furyl]-1-benzyl-indazole), or vehicle three days before and seven days after induction of peritoneal adhesions or, alternatively, once during induction of peritoneal adhesions. Pretreatment or single intraperitoneal lavage with YC-1 significantly reduced postoperative adhesion formation without prompting systemic adverse effects. Expression analyses of cytokines in peritoneal tissue and fluid and in vitro assays applying macrophages and peritoneal fibroblasts indicated that this effect was cooperatively mediated by various putatively HIF-1α-dependent mechanisms, comprising attenuated pro-inflammatory activation of macrophages, impaired recruitment and activation of peritoneal fibroblasts, mitigated epithelial-mesenchymal-transition (EMT), as well as enhanced fibrinolysis and impaired angiogenesis. Thus, this study identifies prevention of postsurgical peritoneal adhesions as a novel and promising field for the application of HIF inhibitors in clinical practice.

Figure 1

YC-1 reduces peritoneal adhesion formation. (a) Experimental schedules. Pretreatment with YC-1 was performed for three days before surgical adhesion induction, and continued for seven days until evaluation (left). Alternatively, single intraoperative lavage with YC-1 was performed upon adhesion induction (middle). To assess short-term effects, pimonidazole injection and subsequent sampling of tissue along with peritoneal fluid were performed 24 h after adhesion induction and intraoperative lavage with YC-1 (right). (b) Left: Representative images of ischemic buttons (arrowheads) and postoperative adhesions (dotted white lines) on postoperative day (POD) 7 following pretreatment with YC-1 or vehicle (Ctrl). Right: Quantification of postoperative adhesions on POD7 after pretreatment or single intraoperative lavage with YC-1 or vehicle (Ctrl) (***P ≤ 0.001, **P ≤ 0.01, n = 9). (c) Left: Representative Masson-Trichrom-Goldner stainings, revealing the thickness of collagen capsules (arrowheads) in between the necrotic core (#) and adjacent peritoneum (asterisks) of ischemic buttons in mice pretreated with YC-1 or vehicle (Ctrl). Right: Histomorphometric quantification, revealing reduced maximal capsule thickness in YC-1-pretreated mice (*P ≤ 0.05, n = 7). Scale bars = 50 μm. (d) Representative pimonidazole immunostainings (left) and histomorphometric quantification (right), revealing comparable extents of hypoxic tissue areas within ischemic buttons of vehicle- (Ctrl) and YC-1-treated animals 24 h after adhesion induction (P > 0.05, n = 4). Scale bars = 50 μm.

Figure 2

YC-1 ameliorates peritoneal inflammation in vivo. (a) Representative CD45 immunostainings (left) and histomorphometric quantification (right) of leukocytes within ischemic buttons of mice pretreated with YC-1 or vehicle (Ctrl), seven days after adhesion induction. (b) Representative F4/80 immunostainings (left) and histomorphometric quantification (right) of macrophages within ischemic buttons of mice pretreated with YC-1 or vehicle (Ctrl). Asterisks in (a,b) indicate positions of sutures (extracted) within ischemic buttons (see also Supplementary Fig. 5; HPF high power field; *P ≤ 0.05, n = 7 in (a,b). Scale bars = 200 μm. (c,d) ELISA of IL-6 (c) and TNFα (d) in tissue lysates of ischemic buttons, harvested 24 h after adhesion induction from Ctrl- and YC-1-treated animals (*P ≤ 0.05, n = 5) (e) ELISA-based quantification of IL-6 levels in peritoneal lavage fluid, harvested 24 h after adhesion induction (*P ≤ 0.05, n = 6). (f,g) Antibody array, revealing the relative abundance of M1- (f) and M2-macrophage-associated cytokines (g) in peritoneal fluid of mice pretreated with YC-1 or vehicle (Ctrl), one day after adhesion induction (***P ≤ 0.001, n = 4; ANOVA with Bonferroni post-hoc tests).

Figure 3

YC-1 attenuates pro-inflammatory differentiation of macrophages. (a,b) qRT-PCR analyses, revealing mRNA expression of the pro-inflammatory M1-differentiation markers NOS2 (a) and IL-6 (b) in murine J774A.1 macrophages under normoxia (left bars) and hypoxia (right bars). LPS exposure was carried out to induce M1-polarization in presence of YC-1 or vehicle control (DMSO; **P ≤ 0.01, n = 9). (c,d) Western blots of nuclear extracts, revealing protein levels of HIF-1α (c) and HIF-2α (d) in J774.A1 macrophages treated with vehicle (DMSO) or YC-1 under normoxic (left) or hypoxic (right) culture conditions. Simultaneous LPS-treatment was performed to induce pro-inflammatory activation. Note LPS-induced stabilization of HIF-1α- (c), but not HIF-2α-levels (d), and reversal of LPS-induced HIF-1α-stabilization by YC-1 (c). Histone H3 was used as a loading control. (e,f) qRT-PCR analyses, revealing mRNA expression of the pro-inflammatory M1-differentiation markers NOS2 (e) and IL-6 (f) in BMDMs under normoxia (left bars) and hypoxia (right bars). LPS exposure was carried out to induce M1-polarization in presence of YC-1 or vehicle control (DMSO; ***P ≤ 0.001, **P ≤ 0.01, n = 6).

Figure 4

YC-1 attenuates EMT and fibroblast activation. (a) qRT-PCR analysis, revealing mRNA expression levels of EMT-associated genes (Twist2, Snail1, TGF-β) and myofibroblast markers (αSMA, Vimentin) in peritoneal adhesion tissue from mice pretreated with YC-1 or vehicle (Ctrl), seven days after adhesion induction (**P ≤ 0.01, *P ≤ 0.05, n = 8). (b) Western blot (top) and densiometric quantification (bottom), revealing TGF-β protein abundance in tissue lysates of ischemic buttons derived from YC-1- or Ctrl-treated mice, Vinculin was used as loading control (*P ≤ 0.05, n = 4). (c) Representative αSMA immunostainings (left) and histomorphometric quantification (right) of myofibroblasts within ischemic buttons of mice pretreated with YC-1 or vehicle (Ctrl), seven days after adhesion induction (HPF high power field; *P ≤ 0.05, n = 6). Scale bars = 200 μm. (d) qRT-PCR, revealing mRNA expression of the myofibroblast marker, αSMA, in peritoneal fibroblasts treated with vehicle (DMSO), or different dosages of YC-1 in hypoxic culture conditions. Note dose-dependent reduction of hypoxia-induced αSMA expression upon YC-1 treatment (*P ≤ 0.05, n = 6). (e) Representative immunocytochemistry for αSMA protein (red) and DAPI (cell nuclei, blue) in primary peritoneal fibroblasts treated with vehicle (DMSO) or YC-1 at normoxic (left panels) or hypoxic (right panels) culture conditions. Note reduction of hypoxia-induced αSMA protein expression upon YC-1 treatment. Scale bars = 50 μm. (f) qRT-PCR, revealing suppression of hypoxia-induced TGF-β mRNA expression in primary peritoneal fibroblasts upon treatment with YC-1 (**P ≤ 0.01, *P ≤ 0.05, n = 6). (g) Immunoblot revealing TGF-β protein levels in primary peritoneal fibroblasts treated with vehicle (DMSO) or YC-1 under normoxic (left) or hypoxic (right) culture conditions. Note hypoxic up-regulation of TGF-β, and reversal by YC-1. Vinculin (bottom) served as loading control. (h) qRT-PCR, revealing that expression of TGF-β mRNA in NIH-3T3 fibroblasts is suppressed by siRNA-mediated interference with HIF-1α (siHIF-1α), but not HIF-2α (siHIF-2α) under normoxic (left bars) and hypoxic (right bars) culture conditions (*P ≤ 0.05, **P ≤ 0.01, n = 6).

Figure 5

YC-1 enhances fibrinolysis and attenuates angiogenesis. (a,b) ELISA measurements, revealing concentrations of PAI-1 protein (a) and active tPA (b) in peritoneal fluid of mice treated with YC-1 or vehicle (Ctrl), one day after adhesion induction (*P ≤ 0.05, n = 6). (c) qRT-PCR, revealing suppression of hypoxia-induced PAI-1 mRNA levels in primary peritoneal fibroblasts upon treatment with YC-1 (*P ≤ 0.05, **P ≤ 0.01, n = 6). (d) qRT-PCR, revealing that hypoxia-induced expression of PAI-1 mRNA in NIH-3T3 fibroblasts is suppressed by siRNA-mediated interference with HIF-1α (siHIF-1α), but not HIF-2α (siHIF-2α) (***P ≤ 0.001, **P ≤ 0.01, n = 6; ANOVA). (e) Representative CD31 immunostainings (left) and histomorphometric quantification (right) of blood vessels within ischemic buttons of mice pretreated with YC-1 or vehicle (Ctrl), seven days after adhesion induction (HPF high power field; **P ≤ 0.01, n = 6). Scale bars = 200 μm. (f) qRT-PCR, revealing suppression of hypoxia-induced VEGF mRNA levels in primary peritoneal fibroblasts upon treatment with YC-1 (***P ≤ 0.001, *P ≤ 0.05, n = 6; ANOVA). (g) qRT-PCR, revealing that hypoxia-induced expression of VEGF mRNA in NIH-3T3 fibroblasts is suppressed by siHIF-1α, but not siHIF-2α (*P ≤ 0.05, n = 6).

Figure 6

Putative mechanisms mediating anti-adhesive effects of YC-1. Intraperitoneal YC-1 counteracts adhesion-triggering inflammation by diverting pro-inflammatory M1-differentiation of macrophages towards immuno-modulatory M2-differentiation (left). In addition, YC-1 attenuates EMT-dependent recruitment and activation of peritoneal fibroblasts (middle). Reduced formation of activated, pro-fibrotic myofibroblasts may result in attenuated secretion of plasminogen activator inhibitor (PAI-1) and vascular endothelial growth factor (VEGF), thus enhancing fibrinolysis via active tissue plasminogen activator (tPA), and mitigating angiogenesis, respectively (right). These effects cooperatively counteract postoperative adhesion formation.