Aflibercept and Ang1 supplementation improve neoadjuvant or adjuvant chemotherapy in a preclinical model of resectable breast cancer

Abstract

Phase III clinical trials evaluating bevacizumab (an antibody to the angiogenic ligand, VEGF-A) in breast cancer have found improved responses in the presurgical neoadjuvant setting but no benefits in the postsurgical adjuvant setting. The objective of this study was to evaluate alternative antiangiogenic therapies, which target multiple VEGF family members or differentially modulate the Angiopoietin/Tie2 pathway, in a mouse model of resectable triple-negative breast cancer (TNBC). Neoadjuvant therapy experiments involved treating established orthotopic xenografts of an aggressive metastatic variant of the MDA-MB-231 human TNBC cell line, LM2-4. Adjuvant therapies were given after primary tumor resections to treat postsurgical regrowths and distant metastases. Aflibercept ('VEGF Trap', which neutralizes VEGF-A, VEGF-B and PlGF) showed greater efficacy than nesvacumab (an anti-Ang2 antibody) as an add-on to neoadjuvant/adjuvant chemotherapy. Concurrent inhibition of Ang1 and Ang2 signaling (through an antagonistic anti-Tie2 antibody) was not more efficacious than selective Ang2 inhibition. In contrast, short-term perioperative BowAng1 (a recombinant Ang1 variant) improved the efficacy of adjuvant chemotherapy. In conclusion, concurrent VEGF pathway inhibition is more likely than Ang/Tie2 pathway inhibition (e.g., anti-Ang2, anti-Ang2/Ang1, anti-Tie2) to improve neoadjuvant/adjuvant chemotherapies for TNBC. Short-term perioperative Ang1 supplementation may also have therapeutic potential in conjunction with adjuvant chemotherapy for TNBC.

Figure 1. Aflibercept is more potent than paclitaxel chemotherapy or Ang/Tie2-targeted agents in terms of inhibiting primary breast tumor growth.

14 days after orthotopic implantation of 2 × 10⁶ LM2-4 cells, mice bearing ~150-mm³ primary breast tumors were randomized and administered with either the controls (PBS vehicle or IgG1 isotype), aflibercept (anti-VEGF-A/VEGF-B/PlGF), nesvacumab (anti-Ang2), BowAng1, or an anti-Tie2 antibody, with or without paclitaxel chemotherapy, for 2 weeks. End-point tumor mass, in grams, is plotted with mean ± SEM depicted. Predefined comparisons were subjected to two-sampled unpaired t tests (n = 5 to 11). Mice given PBS vehicle alone (green solid circles) versus a non-specific IgG1 antibody (black solid circles) did not have significantly different mean terminal tumor weights (P = 0.11); these mice were considered as a single ‘untreated control’ group (n = 11) in subsequent analyses for treatment effects.

Figure 2. Differential treatment effects on primary breast tumor blood vessels.

Orthotopic primary LM2-4 breast tumors harvested after 2 weeks of therapy were analyzed for tumor vascularity by CD31 staining. (A) Automated quantification of CD31-positive pixels normalized to total pixels at 100×, with mean ± SEM depicted. Predefined comparisons (control vs. monotherapies; paclitaxel monotherapy vs. paclitaxel-containing combinations) were subjected to two-sampled unpaired t tests (n = 5 to 11 mice per group). (B) Representative microscopy images of CD31-stained breast tumor sections.

Figure 3. Differential treatment effects on primary breast tumor invasiveness into the abdominal wall.

14 days after orthotopic implantation of 2 × 10⁶ LM2-4 cells, mice bearing ~150-mm³ primary breast tumors were randomized and administered with either the controls (PBS vehicle or IgG1 isotype), aflibercept (anti-VEGF-A/VEGF-B/PlGF), nesvacumab (anti-Ang2), BowAng1, or an anti-Tie2 antibody, with or without paclitaxel chemotherapy, for 2 weeks. On day 29 post-implantation, all mice were sacrificed and their primary breast tumors were examined during necropsy and confirmed histologically for signs of invasion into the abdominal wall. (A) The incidence (%) of invaded tumors per treatment group is plotted. P > 0.05 by Fisher’s exact test. (B) Representative microscopy images from the histological analysis of primary LM2-4 breast tumors for invasions into the adjacent abdominal wall by hematoxylin and eosin staining. “M” denotes abdominal wall muscle. “T” denotes tumor cells. Black arrows mark regions where tumor cells are infiltrating into the abdominal wall and separating muscular fascicles.

Figure 4. Addition of adjuvant aflibercept or perioperative BowAng1 improves adjuvant paclitaxel chemotherapy for resected breast cancer.

18 days after orthotopic implantation of 2 × 10⁶ LM2-4 cells, mice bearing roughly 400-mm³ primary breast tumors were randomized into nine treatment groups. Primary tumor resections by complete mastectomies were performed at 22 days post-implantation (DPI). BowAng1 was given as a 10-day perioperative therapy beginning one day before surgery (21 DPI). Aflibercept (anti-VEGF-A/VEGF-B/PlGF), nesvacumab (anti-Ang2), the anti-Tie2 antibody, paclitaxel chemotherapy, and combinations thereof, were given as 4-week-long adjuvant therapies starting two days after surgery (24 DPI). Dosing schedule is depicted above survival curves. N = 7–9 mice per treatment group. Kaplan-Meier analyses of overall survival: (A) comparison of monotherapy groups; (B) comparison of combination therapies; and (C) showing only those single-agent and combination therapies that led to a statistically significant overall survival benefit compared to vehicle control and chemotherapy alone respectively (P < 0.05, log-rank test).

Figure 5. Triple combination of adjuvant aflibercept, perioperative BowAng1 and adjuvant paclitaxel therapies for resected breast cancer.

18 days after orthotopic implantation of 2 × 10⁶ LM2-4 cells, mice bearing approximately 200-mm³ primary breast tumors were randomized into four treatment groups. Primary tumor resections by complete mastectomies were performed at 20 days post-implantation (DPI). BowAng1 was given as a 10-day perioperative therapy, beginning one day before surgery (19 DPI). Aflibercept (anti-VEGF-A/VEGF-B/PlGF) and paclitaxel chemotherapy were given as 4-week-long adjuvant therapies, starting two days after surgery (22 DPI). Dosing schedule is depicted above Kaplan-Meier survival curves (n = 10 mice per treatment group). The log-rank statistical test was used to assess differences in overall survival between treatment groups at two timepoints: 80 DPI (interim) and 115 DPI (final) respectively.