Antiangiogenic therapy elicits malignant progression of tumors to increased local invasion and distant metastasis

Abstract

Multiple angiogenesis inhibitors have been therapeutically validated in preclinical cancer models, and several in clinical trials. Here we report that angiogenesis inhibitors targeting the VEGF pathway demonstrate antitumor effects in mouse models of pancreatic neuroendocrine carcinoma and glioblastoma but concomitantly elicit tumor adaptation and progression to stages of greater malignancy, with heightened invasiveness and in some cases increased lymphatic and distant metastasis. Increased invasiveness is also seen by genetic ablation of the Vegf-A gene in both models, substantiating the results of the pharmacological inhibitors. The realization that potent angiogenesis inhibition can alter the natural history of tumors by increasing invasion and metastasis warrants clinical investigation, as the prospect has important implications for the development of enduring antiangiogenic therapies.

Figure 1. Increased Invasive Phenotype after Anti-VEGFR2 Therapy

Histological analysis and quantification of tumor invasion in anti-VEGFR2 antibody (DC101)-treated or control immunocompromised tumor-bearing RIP1-Tag2 animals are presented. (A) Histological images of tumors from control, 1-week, and 4-week treatment arms are shown in hematoxylin and eosin staining (H&E; top row), indicating prominent invasive fronts in the treated animals (black arrowheads); high-magnification images of immunofluorescence staining for SV40 T antigen (middle row), where T indicates tumor and Ac indicates surrounding acinar tissue; and immunodetection of the vasculature with anti-CD31 antibody (bottom row), where a central view of the tumor parenchyma is shown. (B) Quantification of tumor invasiveness represented as the percentage of encapsulated islet tumors (IT), microinvasive carcinomas (IC1), and fully invasive carcinomas (IC2) for the three different treatments. Both anti-VEGFR2 treatments show a statistically significant decrease in the percentage of IT and a significant increase in IC2 tumors (**p < 0.01 by Kruskal-Wallis test). (C) Quantification of tumor invasiveness in animals treated for only 1 week with anti-VEGFR2 antibody and then maintained without therapy for 1, 2, or 3 more weeks. Percentage of IT (white bars), IC1 (gray bars), and IC2 (black bars) is shown per treatment and time point. All treatment groups showed statistical differences by Mann-Whitney test (p < 0.01) when compared to the control group at the point when treatment was started. Age-matched treated versus control statistical analysis is shown for each time point. Note that no animals in the control group survived to the 16 week time point. *p < 0.05, **p < 0.01 by Mann-Whitney test. Error bars indicate ± SD.

Figure 2. Increased Tumor Invasion after Tumor-Specific Vegf-A Gene Deletion

(A) Top row: H&E staining of lesions in RIP1-Tag2/Cre;Vegf-A^+/+ (β-VEGF-WT; left) and RIP1-Tag2/Cre;Vegf-A^fl/fl (β-VEGF-KO; middle) mice. Middle row: insulin immunostaining reveals the tumor cells in serial sections to the H&E images shown in the top row. Bottom row: CD31 vessel staining (red) and DAPI nuclear staining (blue) of tumors in a β-VEGF-WT mouse and a β-VEGF-KO mouse. Images at right show a higher magnification of images in the middle column. (B) Microvessel area (MVA) at the invasive front was equivalent to that of exocrine tissue in β-VEGF-KO tumors. MVA of six tumors from four β-VEGF-WT and β-VEGF-KO animals was determined in the tumor “border” as the band-like region 50 μm inside the tumor perimeter (gray bars), the surrounding exocrine tissue (“Exo”; white bars), and the tumor “core” (black bars). MVA values were standardized to those of exocrine tissue distant from the tumors in each section. *p < 0.05. (C) Quantification of invasiveness represented by the frequency of each tumor grade in 27 tumors from 5 β-VEGF-KO mice (gray bars) versus 75 tumors from 5 β-VEGF-WT mice (black bars) at 13 weeks of age. *p < 0.05, **p < 0.01 by Mann-Whitney test. Error bars indicate ± SD.

Figure 3. Increased Incidence of Lymph Node and Liver Metastasis in Anti-VEGFR2-Treated Animals

Histological analysis of lymph node (LN) and liver metastasis (Met) in RIP1-Tag2 animals treated with anti-VEGFR2 for 10 days starting at 10 weeks of age and then left untreated until 16 weeks of age. (A) LN and Met observed by histological H&E staining of tissue sections from anti-VEGFR2-treated animals appear as enlarged hemorrhaging LNs infiltrated with tumor cells and a small tumor nodule in the liver parenchyma (left panels). Immunohistochemical staining for the tumor marker SV40 T antigen (brown) reveals the presence of tumor cells infiltrated into a LN or in the midst of the liver parenchyma (right panels). (B) Top: quantification of the incidence of animals with microscopic liver micrometastasis and macroscopic LN metastasis in the control (gray bars) and anti-VEGFR2-treated (black bars) treatment arms. Bottom: contingency table relating the number and percentage of animals in each treatment/metastasis case. Treated animals show a statistically significant increase in the incidence of liver micrometastasis and LN metastasis by the chi-square test (*p < 0.05; **p < 0.01). (C) Quantification of the number of microscopic liver metastasis in the anti-VEGFR2 treated (black bars) and control (gray bars) treatment arms. Treated animals show a statistically significant increase in the number of liver micrometastases per animal by the Mann-Whitney test (*p < 0.05). Error bars indicate ± SEM.

Figure 4. Increased Life Span and Tumor Reduction in Sunitinib-Treated RIP1-Tag2 Animals

(A) Kaplan-Meier survival curves in tumor-bearing RIP1-Tag2 mice (12 weeks) treated continuously with vehicle control or sunitinib starting at 12 weeks. While vehicle-treated mice showed a median life span of 15.2 weeks, mice receiving continuous sunitinib treatment demonstrated a survival benefit of 7 additional weeks. (Kaplan-Meier statistic: p < 0.01.) (B) Total tumor burden analysis in 5-week treatment trials with sunitinib or vehicle control starting at 10 weeks of age. Gray bar shows tumor burden of vehicle-treated RIP1-Tag2 animals, and black bar shows the statistically significant tumor volume reduction of efficacious sunitinib therapy. Each treatment cohort involved a minimum of ten animals per group. **p < 0.01 by Mann-Whitney test. Error bars indicate ± SEM.

Figure 5. Increased Tumor Invasion Evoked by Treatment with a Multitargeted Angiogenic Kinase Inhibitor

Histological analysis and quantification of tumor invasion in RIP1-Tag2 mice treated for 5 weeks with sunitinib compared to control vehicle-treated mice are shown. (A) Top row: H&E staining of control- and sunitinib-treated tumors showing the invasive front extensively intercalating into the surrounding tissue (dashed white line), producing isolated islands of normal exocrine tissue inside treated tumors (white arrow). Middle row: immunofluorescence staining for SV40 T antigen showing the widely invasive phenotype in treated tumors, with several fronts invading into the surrounding tissue. Bottom row: visualization of blood vessels by intravenous perfusion of FITC-conjugated tomato lectin (green), revealing the effect of the antiangiogenic treatment on the tumor vasculature. (B)Quantification of the percentage of IT, IC1, and IC2 tumors in vehicle- versus sunitinib-treated animals. Sunitinib treatment shows statistically significant decrease in IT and increase in the highly invasive grade of tumor (IC2) by Mann-Whitney test (*p < 0.01; **p < 0.01). Error bars indicate ± SD. (C) Left panels: histological analysis by H&E staining of tissue sections of lymph node and liver metastasis from sunitinib-treated animals. Right panels: confirmation of metastasis is shown by immunofluorescence staining for the tumor marker T antigen (red). The images give evidence of tumor cells infiltrated into the lymph node or the liver parenchyma. (D) Top: quantification of the proportion of animals with liver and lymph node (LN) micrometastasis in the vehicle-treated (gray bars) and sunitinib-treated (black bars) arms. Bottom: contingency table listing the number and percentage of animals in each treatment/metastasis case. Treated animals show a 3.5-fold increase in the incidence of liver micrometastasis by the chi-square test (*p < 0.05). (E) Quantification of the number of liver micrometastases in the vehicle-treated (gray bars) and sunitinib-treated (black bars) arms. The number of liver micro-metastases was 3.7-fold increased in the sunitinib-treated group, which showed a clear but not statistically significant trend compared to control animals. Error bars indicate ± SEM.

Figure 6. Effects of VEGFR-Selective Kinase Inhibitors on an Orthotopic Mouse Model of Glioblastoma Multiforme

(A) Top row: immunohistological analysis of glioblastoma multiforme (GBM) with fluorescent SV40 T antigen staining (red) on whole brain sections counterstained with DAPI (blue). Control wild-type (WT) GBMs appear less invasive than WT GBMs treated with SU10944 or sunitinib, or VEGF-KO GBMs (purple arrowheads). Bottom row: high-magnification images of fluorescent detection of tumor cells (anti-T antigen antibody; red) and the perfused vasculature (FITC-lectin, green). WT GBM cells infiltrate as single cells into the brain parenchyma without associating with blood vessels (white arrowheads) or invade alongside blood vessels in the brain (perivascular invasion; yellow arrowheads). GBMs treated with either SU10944 or sunitinib as well as VEGF-KO GBMs are more invasive and predominantly migrate along blood vessels (yellow arrowheads). (B) Kaplan-Meier survival curves of WT GBM-bearing mice untreated or treated with a relatively selective VEGFR inhibitor (SU10944) or a multitargeted VEGFR inhibitor (sunitinib) starting 3 days after inoculation. The effects on survival observed with SU10944 and sunitinib treatment were minimal or modest, respectively, and were not statistically significant. (C) Top: grading of perivascular invasion was performed on a scale from 1 to 4, where 1 indicates minimal distant spread of tumor cells and 4 indicates substantial and marked distant spread; this parameter indicated a significant increase in perivascular invasion in response to both pharmacological treatments and in the VEGF-KO GBMs (**p < 0.01). Bottom: quantification of perivascular invasive mode by immunohistochemical analysis on tumor sections. Perivascular invasive cells, counted as the number of cells in the tumor periphery tightly associated with vessels, were significantly increased by both antiangiogenic small-molecule treatments and the genetic ablation of VEGF-KO GBM cells (**p < 0.01). Error bars indicate ± SD.

Figure 7. Antiangiogenic Treatment also Provokes Hypoxia in Tumors and Liver Micrometastases

(A) Hypoxia in islet tumors was detected by immunofluorescence staining of pimonidazole adducts in sections of pancreas (Aa-Af) or liver (Ag-Ai) from control untreated animals (Aa and Ag), animals receiving short-term (Ab) or long-term (Ac) anti-VEGFR2 treatment, β-VEGF-WT (Ad) and β-VEGF-KO (Ae) islet tumors, and sunitinib-treated animals (Af, Ah, and Ai). (Aa)-(Ag) show pimonidazole immunodetection (green) with blood vessel CD31 staining (red); (Ah) shows pimonidazole immunodetection (green) and T antigen onco-protein (red); (Ai) shows blood vessel MECA32 staining (green) and T antigen oncoprotein (red). All images show nuclei counterstained with DAPI (blue). (B) Quantitation of the incidence of hypoxic tumors was performed in long-term anti-VEGFR2-treated and sunitinib-treated animals and plotted as the percentage of pimonidazole-positive tumors per animal compared to control animals. **p < 0.01 by Mann-Whitney test. Error bars indicate ± SEM.

Figure 8. Adaptive-Evasive Responses by Tumors to Antiangiogenic Therapies

Schematic summary of adaptive responses to VEGF/VEGFR inhibitors (and likely other angiogenesis inhibitors) that elicit “evasive resistance.” Tumors respond to VEGF/VEGFR pathway inhibition with tumor stasis or regression and a loss of blood vessels, but mechanisms of evasive resistance to the antiangiogenic treatment are then induced that can variously enable revascularization via alternative proangiogenic signals, increased local invasiveness, and/or enhanced distant metastasis.