Whether lymph node metastases can be a source of cancer cells for distant metastases has been widely debated. Now, two studies published in Science have used mouse models to show that metastatic tumour cells can colonize distant organs by invading lymph node blood vessels to enter the blood circulation.

figure a

Lara Crow/Macmillan Publishers Limited

tumour cells can colonize distant organs by invading lymph node blood vessels

Brown et al. and Pereira et al. employed different methods to trace the fate of tumour cells in lymph nodes. Brown et al. mimicked the formation of lymph node metastases in the absence of a primary tumour by microinfusing mouse 4T1 mammary cancer cells into the afferent lymphatic vessel of popliteal lymph nodes located in the legs of mice. Pereira et al. orthotopically implanted 4T1 and mouse B16F10 melanoma cells stably expressing the photoconvertible fluorescent fusion protein Dendra2–H2B into mice, followed by primary tumour resection. A laser diode was then used to convert Dendra2–H2B-expressing cancer cells from green to red fluorescence in the draining lymph node.

Both studies observed cancer cells initially traversing the floor of the subcapsular sinus before later invading through to the cortex of the lymph node. Here, the metastatic cancer cells were found in close association with blood vessels and often within the lumen of high endothelial venules (HEVs), which enable lymphocyte trafficking from the blood into the lymph node. As a similar association between cancer cells and lymphatic vessels in lymph nodes was not found, these data indicate intravasation across blood vessels being the predominant route by which metastatic cancer cells exit the lymph node. Importantly, comparison of these mouse models with samples of lymph node micrometastases from patients with breast cancer and head and neck cancer demonstrated similar histology and tumour cell–blood vessel interactions.

Next, the two groups of researchers used their respective models to determine whether cancer cells that had entered the bloodstream via lymph nodes could result in systemic metastases. Brown et al. intralymphatically infused the lymph nodes of mice with fluorescent 4T1 cells and noted metastatic seeding in the lung as early as 3 days after infusion. To rule out the possibility that cancer cell escape via the efferent lymphatic vessel could contribute to lung metastatic colonization, this vessel was surgically ligated before intralymphatic infusion. Whilst the procedure prevented the recirculation of injected T cells into the spleen, it did not affect the capacity of 4T1 cells to seed lung metastases. To account for the influence of pre-metastatic conditioning on the metastatic process, the authors adapted their model by implanting unlabelled 4T1 cells above the ankle to form a peripheral tumour 8 days before infusing fluorescent 4T1 cells into lymph nodes. This experiment revealed that priming did not increase the frequency of lung metastases over that of the unprimed condition.

Similarly, Pereira et al. found photoconverted red 4T1 and B16F10 cells present in the lungs of mice following light exposure of lymph node metastases. However, these authors also showed that lung metastases could originate directly from the primary tumour as well as the lymph node. This was achieved by implanting green fluorescent 4T1 cells directly in the axillary lymph node of mice and, on the contralateral side, removing this same lymph node before implanting red fluorescent 4T1 cells in the mammary fat pad. Upon resection of both the tumour-bearing lymph node and the mammary fat pad primary tumour, both red and green metastases had formed in the lung. In another variation of the model, the authors excised the sentinel lymph nodes before implanting Dendra2–H2B-expressing 4T1 cells into the mammary fat pad. After primary tumour resection, lung metastases were detected in the absence of lymph nodes. Yet, mice with intact lymph nodes had higher numbers of lung metastases, perhaps implying that metastatic dissemination via the lymph node is more efficient than a direct route from the primary tumour.

The next step will be to confirm whether this dissemination route occurs in human cancers. If identified, its discovery is likely to affect treatment decisions in cancer patients.