Losartan treatment enhances chemotherapy efficacy and reduces ascites in ovarian cancer models by normalizing the tumor stroma

Abstract

In ovarian cancer patients, tumor fibrosis and angiotensin-driven fibrogenic signaling have been shown to inversely correlate with survival. We sought to enhance drug delivery and therapeutic efficacy by remodeling the dense extracellular matrix in two orthotopic human ovarian carcinoma xenograft models. We hypothesized that targeting the angiotensin signaling axis with losartan, an approved angiotensin system inhibitor, could reduce extracellular matrix content and the associated "solid stress," leading to better anticancer therapeutic effect. We report here four translatable findings: (i) losartan treatment enhances the efficacy of paclitaxel-a drug used for ovarian cancer treatment-via normalizing the tumor microenvironment, resulting in improved vessel perfusion and drug delivery; (ii) losartan depletes matrix via inducing antifibrotic miRNAs that should be tested as candidate biomarkers of response or resistance to chemotherapy; (iii) although losartan therapy alone does not reduce tumor burden, it reduces both the incidence and the amount of ascites formed; and (iv) our retrospective analysis revealed that patients receiving angiotensin system inhibitors concurrently with standard treatment for ovarian cancer exhibited 30 mo longer overall survival compared with patients on other antihypertensives. Our findings provide the rationale and supporting data for a clinical trial on combined losartan and chemotherapy in ovarian cancer patients.

Keywords: angiotensin inhibition; antifibrotic miRNAs; ascites; drug delivery; ovarian cancer.

Fig. 1.

Losartan treatment reduces matrix content, fibroblast infiltration, and solid stress in SKOV3ip1 and Hey-A8 ovarian cancer models. Control and losartan-treated human ovarian SKOV3ip1 and Hey-A8 tumors were stained for (A) collagen by Sirius Red staining, (B) hyaluronan by hyaluronic acid binding protein (HABP) staining, and (C) fibroblasts by αSMA staining. The fractions of Sirius Red (A, red), HABP (B, brown) and αSMA (C, red) positive areas were quantified using ImageJ software. All representative images shown are from SKOV3ip1 tumors. Data presented are mean ± SD. For each stain, n = 12 sections, with three sections per tumor. (D) Representative maps and quantification of compressive and tensile solid stresses in size-matched control and losartan-treated peritoneal SKOV3ip1 tumors. (E) The equilibrium modulus (stiffness) in control and losartan-treated peritoneal tumors.

Fig. 2.

Losartan treatment improves vessel perfusion, relieves tumor hypoxia, and increases drug delivery. Control and losartan-treated human ovarian SKOV3ip1 and Hey-A8 tumors were stained for (A) perfused (FITC-lectin⁺, green) and total blood vessels (CD31⁺ endothelial cells, red) and (B) hypoxic tumor tissues by pimonidazole stain (brown). The fraction of FITC-lectin⁺/CD31⁺ and hypoxic area was quantified using ImageJ software. (C) Representative images of doxorubicin intratumoral distribution and quantification of the fraction of tumor area positive for doxorubicin. Green, FITC-lectin labeled perfused vessels; red, fluorescent doxorubicin; blue, DAPI. All representative images shown are from SKOV3ip1 model. Data presented are mean ± SD. For each stain, n = 12 sections, with three sections per tumor.

Fig. 3.

Mathematical model based on losartan’s mechanism of improved therapy. Model results for (A) the spatial intratumoral distribution of the cytotoxic drug doxorubicin following i.v. injection with and without losartan treatment. A 2D slice of the 3D model is presented; the dashed line denotes the interface of the tumor with the peritoneum. Model results for the relative change in the average (B) solid stress levels, (C) functional vascular density, (D) doxorubicin delivery, (E) delivery of i.p. administered paclitaxel, (F) fluid pressure at the tumor center, and (G) average tumor oxygen concentration with and without losartan treatment. (H) Predictions of the relative change in tumor volume for the treatment groups employed in the experimental study.

Fig. 4.

Parametric analysis for the spatial intratumoral distribution of the drug following i.p. administration. The elastic modulus (A), hydraulic conductivity of the tumor (B), and the diffusion coefficient of the drug (C)—parameters that are known to affect drug delivery—were varied and model predictions from a slice of the 3D tumor model are presented.

Fig. 5.

Combined losartan treatment enhances the efficacy of paclitaxel. Mice were injected i.p. with SKOV3ip1 (A–C) and Hey-A8 (D) tumors, and peritoneal tumor growth was monitored by G-luc value. Between 7 and 10 d after implantation when peripheral blood G-luc value reached 2 × 10⁶ RLU, mice were randomized into four treatment groups receiving (i) control (saline), (ii) losartan, (iii) paclitaxel, or (iv) losartan combined with paclitaxel. When mice became moribund, all peritoneal tumors were collected and weighed. The incidence (B) and volume (C) of ascites in SKOV3ip1 model were measured. Representative of at least three independent experiments (n = 10 each), data presented are mean ± SEM. (E) PCNA⁺ tumor cells (per 0.041 mm²), and (F) TUNEL⁺ tumor cells (per 0.329 mm²) were manually counted in 10 random fields in frozen sections of SKOV3ip1 and Hey-A8 tumors.

Fig. 6.

Losartan reduces collagen content in the tumors invading the diaphragm and “normalizes” diaphragm lymphatic vessel morphology. Diaphragms from non-tumor-bearing mice and from mice bearing SKOV3ip1 tumors treated with control or losartan were collected. (A) Representative immunofluorescent staining images of collagen I (red) in cross-sectioned diaphragm. (B) The fraction of collagen I-positive area in the diaphragm was quantified using ImageJ software. (C) Representative images of fluorescent lymphangiography. In non-tumor-bearing mice, and mice bearing SKOV3ip1 tumors treated with control or losartan, FITC-dextran (green) were injected into the peritoneum to label lymphatic vessels on the pleural and peritoneal side of diaphragm. (D) Lymphatic vessel diameter on the pleural side of the diaphragm was quantified using ImageJ. (E) Representative immunofluorescent staining images of lymphatic vessels (LYVE-1-green and CD31-red) in whole-mounted diaphragm. Data presented are mean ± SD. For each staining, n = 12 sections, with three sections per tumor.

Fig. 7.

Losartan improves diaphragm lymphatic vessel drainage. In non-tumor-bearing mice, and mice bearing SKOV3ip1 tumors treated with control or losartan, fluorescent beads (green) were injected into the peritoneum to observe their drainage. Representative images of (A) the diaphragm and (B) the CMLN frozen sections under confocal microscope. Blue, DAPI. (C) The amount of fluorescent beads drained to the CMLN was quantified by measuring the fluorescence intensity of homogenized CMLNs using a plate reader. Data presented are mean ± SD, n = 12 diaphragms and CMLNs each.

Fig. 8.

Losartan treatment increases miR-133 level, which regulates collagen levels. (A) miRNAs differentially regulated in control and losartan-treated SKOV3ip1 cells as evaluated by miRNA array. (B, Top) Sequence alignment of miR-133 with the 3′UTR of the COL1A1, COL5A3, and COL6A3 genes. The noncoding RNA sequence of each gene and the seed sequence of hsa-miR-133-3p are shown. (B, Bottom) The seed sequence of miR-133 in COL1A1 is highly conserved among different species, including humans (Homo sapiens), mice (Mus musculus), and rats (Rattus norvegcus). (C) Western blot of collagen I in parental and miR-133–overexpressing SKOV3ip1 cells. (D) Schematic representations of the luciferase reporter construct with the locations of the seed sequence (SS) in the COL1A1 3′-UTR. Potential base pairs between hsa-miR-133 and the target site are indicated in the wild-type and mutated seed sequence. (E) Relative luciferase activity. Parental, mock, and miR-133–overexpressing SKOV3ip1 cells were transiently cotransfected with wild-type (pmirGLO-COL1-wt) or mutant (pmirGLO-COL1-mut) firefly luciferase reporter genes and Renilla luciferase genes. Firefly luciferase activities were normalized to Renilla luciferase activity. Data presented are mean ± SD. (F) Potential mechanisms by which losartan treatment enhanced chemotherapy efficacy and reduced ascites. Losartan treatment significantly up-regulates miRNAs that target collagen molecules, leading to reduced matrix content. Reduced matrix content can alleviate compression on vessels. As a result (i) blood vessel perfusion is significantly improved, leading to increased delivery of oxygen and drug and enhanced chemotherapy efficacy of blood-borne drugs and (ii) diaphragm lymphatic vessels drainage function is enhanced, leading to improved peritoneal fluid drainage and decreased accumulation of ascites.

Fig. 9.

ACEi/ARB adjunctive treatment improves survival in women with ovarian cancer receiving standard of care. (A) Inverse probability of treatment-weighted survival curves for patients with advanced ovarian cancer who were users of ACEis or ARBs (ACEi/ARB, blue line) compared with users of any other antihypertensive medication (No ACEi/ARB, red line) at the time of cytoreductive surgery. Hazard of death from any cause was significantly lower among women receiving an ACEi or ARB compared with controls (hazard ratio 0.55; 95% CI 0.36–0.95). (B) Inverse probability of treatment-weighted survival curves patients with advanced ovarian cancer who were users of ARBs (red line) compared with users of ACEis (blue line) at the time of cytoreductive surgery. Hazard of death from any cause was significantly lower among women receiving an ARB compared with ACEi (hazard ratio 0.38; 95% CI 0.15–0.91).