Inhibition of Vascular Endothelial Growth Factor Protects against the Development of Oxaliplatin-Induced Sinusoidal Obstruction Syndrome in Wild-Type but Not in CD39-Null Mice

Abstract

(1) Background: Sinusoidal obstruction syndrome (SOS) after oxaliplatin-based chemotherapy is associated with unfavorable outcomes after partial hepatectomy for colorectal liver metastases (CLM). Bevacizumab, a monoclonal antibody against vascular endothelial growth factor (VEGF), may prevent SOS development. We investigated the impact of VEGF-inhibition on the development of SOS in a murine model. (2) Methods: Male wild-type and CD39-null mice received oxaliplatin, additional anti-VEGF (OxAV), or controls, and were sacrificed or subjected to major partial hepatectomy (MH). Specimen were used for histological analysis of SOS. Liver damage was assessed by plasma transaminases. The VEGF pathway was elucidated by quantitative PCR of liver tissue and protein analysis of plasma. (3) Results: Mice treated with oxaliplatin developed SOS. Concomitant anti-VEGF facilitated a reduced incidence of SOS, but not in CD39-null mice. SOS was associated with increased plasma VEGF-A and decreased hepatocyte growth factor (HGF). After OxAV treatment, VEGF-R2 was upregulated in wild-type but downregulated in CD39-null mice. Oxaliplatin alone was associated with higher liver damage after MH than in mice with concomitant VEGF-inhibition. (4) Conclusions: We established a murine model of oxaliplatin-induced SOS and provided novel evidence on the protective effect of VEGF-inhibition against the development of SOS that may be associated with changes in the pathway of VEGF and its receptor VEGF-R2.

Keywords: bevacizumab; colorectal liver metastases; sinusoidal obstruction syndrome.

Figure 1

Project overview. wt (n = 116) and CD39-null mice (n = 70) received weekly intraperitoneal injections of their respective treatment (oxaliplatin alone, oxaliplatin and anti-VEGF, anti-VEGF alone, or glucose) over a time period of five weeks. Afterwards, wt mice were subjected to MH or sacrifice based on their general health. Instead of MH, some received a sham operation. Mice were sacrificed after 24, 36, 48, or 72 h after MH. All CD39-null mice were sacrificed after the five-week treatment period.

Figure 2

Histopathological analysis. (A) Characteristic H&E slides of liver tissue of wt and CD39-null mice after treatment with oxaliplatin or oxaliplatin and anti-VEGF: The typical sinusoidal dilation around the central veins is visible in all groups except wtOxAV. Magnification 20×. (B) Development of sinusoidal dilation as evaluated by H&E slides: Sinusoidal dilation was found in 90% and 58% in the wtOx and wtOxAV group, respectively (***, p = 0.001). In contrast, all CD39-null mice treated with oxaliplatin or oxaliplatin and anti-VEGF developed sinusoidal dilation. Hence, anti-VEGF did not protect against the development of sinusoidal dilation in CD39-null mice receiving oxaliplatin compared to wtOxAV (100% vs. 58%, p = 0.005).

Figure 3

Analysis of quantitative PCR data in resected liver specimen. Median fold change and standard error. wtGlu and cd39Glu served as controls. (* p ≤ 0.05, ** p ≤ 0.01, **** p ≤ 0.0001) (Left column). VEGF-A. Expression of VEGF-A was significantly up-regulated in mice receiving oxaliplatin-containing regimens (wtOx: p = 0.004, wtOxAV: p < 0.0001, cd39Ox: p < 0.0001, and cd39OxAV: p < 0.0001). However, VEGF-A was equivalently expressed between groups receiving oxaliplatin (wtOx/wtOxAV: p = 0.786, cd39Ox/cd39OxAV: p = 0.990). (Middle column). VEGF-R1. In wt mice, VEGF-R1 expressions were comparable. In cd39-null mice compared to controls, VEGF-1 was equivalently expressed in cd39Ox (p = 0.137), up-regulated in cd39OxAV (p < 0.0001), and down-regulated in cd39AV (p = 0.007). (Right column). VEGF-R2. Most importantly, VEGF-R2 was up-regulated in wtOxAV (p = 0.045) compared to controls, while it was comparable among cd39-null mice (p = 0.112).

Figure 4

Results of magnetic bead-based assay of blood plasma. Medians and interquartile range. (* p ≤ 0.05, ** p ≤ 0.01) (A,B) VEGF-A (p < 0.0001, A) and HGF (p = 0.030, B) were significantly different among wt mice. The highest values for VEGF-A were found in the wtOxAV group (A). For HGF, lowest values were found in the wtOx group (B). In the subgroup comparison of wtOx mice with sinusoidal changes and wtOxAV mice without any histological changes, VEGF-A (p = 0.003, A) and HGF (p = 0.040, B) were significantly increased for wtOxAV. (C) cd39OxAV showed the highest values for VEGF-A; however, it was equivalent to cd39AV (p = 0.193). (D) Lowest values for HGF among cd39-null mice were measured for cd39Ox, while control mice (cd39Glu) showed significantly higher values (p < 0.0001). (E,F) cd39-null mice receiving oxaliplatin (cd39Ox and cd39OxAV) showed the lowest values for MMP-9 (E) and the highest values for TIMP-1 (F).

Figure 5

Course of treatment and results after major partial hepatectomy. (* p ≤ 0.05, ** p ≤ 0.01, **** p ≤ 0.0001) (A) Relative change of body weight of animals during the course of treatment (day of first treatment until MH or sacrifice) was −4%, −11%, +27%, and +36% in wtOx, wtOxAV, wtAV, and wtGlu, respectively (p < 0.0001), and −17%, −17%, +15%, and +13% in cd39Ox, cd39OxAV, cd39AV, and cd39Glu, respectively (p < 0.0001). Hence, loss of body weight was observed in all animals receiving systemic chemotherapy. (B,C) Plasma was obtained to compare liver parameters. Statistically significant interaction between intervention group and time after MH was found for ALT (p = 0.003; B) and AST (p = 0.002; C). Highest values for ALT and AST at 36 h after MH were found for wtOx (ALT: p = 0.035, AST: p = 0.004). (D) Hyperbilirubinemia after MH was observed in 58%, 17%, 8%, 17%, 0% in wtOx, wtOxAV, wtAV, wtGluMH, and wtGluSham, respectively (p = 0.004). (E) Liver mass regrowth after MH (p < 0.0001). Mice treated with oxaliplatin showed significantly reduced liver regrowth in comparison to wtGlu at 48 (p = 0.003) and 72 h (p < 0.0001). wtOxAV did not promote improved liver regrowth after MH. (F) Immunohistochemistry for Ki-67. Indices for wtOx and wtOxAV were equivalent at all times after MH. However, Ki-67 indices were reduced for wtOx after 36, 48, and 72 h after MH in comparison to wtAV (p = 0.089, p = 0.004 and p = 0.021, respectively) and wtGlu (p = 0.002, p < 0.0001 and p = 0.012, respectively). In contrast, Ki-67 indices of wtOxAV were only inferior compared to wtGlu at 36 (p = 0.008) and 48 h (p < 0.0001).