Dissemination via the lymphatic or angiogenic route impacts the pathology, microenvironment and hypoxia-related drug response of lung metastases.

Complications associated with the development of lung metastases have a detrimental effect on the overall survival rate of many cancer patients. Preclinical models that mimic the clinical aspects of lung metastases are an important tool in developing new therapy options for these patients. The commonly used intravenous models only recapitulate dissemination of cancer cells to the lungs via the haematological route. Here we compared spontaneous and intravenous lung metastases of the highly metastatic KHT mouse fibrosarcoma cells after injecting KHT cells into the subcutaneous layer of the skin or directly into the tail vein. In contrast to the intravenous model, metastases spontaneously arising from the subcutaneous tumours disseminated most consistent with the lymph nodes/lymphatics route and were more hypoxic than the metastases observed following tail-vein administration and haematological spread. To ascertain whether this impacted on drug response, we tested the effectiveness of the hypoxia-sensitive cytotoxin AQ4N (Banoxantrone) in both models. AQ4N was more effective as an anti-metastatic drug in mice with subcutaneous KHT tumours, significantly reducing the metastatic score. Complementing the KHT studies, pathology studies in additional models of spontaneous lung metastases showed haematological (HCT116 intrasplenic implant) or mixed haematological/lymphatic (B16 intradermal implant) spread. These data suggest that preclinical models can demonstrate differing, clinically relevant dissemination patterns, and that careful selection of preclinical models is required when evaluating new strategies for targeting metastatic disease.