NK trafficking, like other lymphocytes, is controlled by chemokine receptors and by chemokines gradients. We have demonstrated that the chemokine receptor repertoire is different between the CD56dimCD16+ (more cytotoxic) and CD56brightCD16- (more cytokine-secreting) subsets . Once in the tumor microenvironment, increased proliferation of the CD56bIightCD16- compared with the CD56dimCD16+ subset might occur by interactions with both mDCs and pDCs [23, 24] . Indeed, while the CD56brightCD16-population consists less than 10% of the NK in the blood, inside tumors infiltrates this subset was found to be the predominant NK population . The complexity of examining NK cell phenotype inside tumors is further illustrated by the observations suggested that there are at least 48 different phenotypic NK subsets in human blood . Hence, it is most likely that detailed characterization of tumor infiltrating NK cells for the expression of additional markers will reveal novel NK subsets distribution.
In mice, NK are commonly found in many organs, including bone-marrow, spleen, thymus, lung, lymph nodes and liver [4, 27, 28]. NK cells were also found in the human skin . However, NK cells are not equally distributed in all organs. Thus, when tumors develop, NK might be recruited to the tumor site from the blood stream by mechanisms such as induced fractalkine expression and its recognition by the CX3CR1 receptor which is found on most NK cells [30, 31]. Importantly, it was suggested that the major difficulty in adoptive transfer of NK cells in tumor patients is to bring the effector killer cells to the tumor site . The tumor microenvironment probably plays a major role in NK cell recruitment or preclusion.
After exiting the blood vessels, NK cells needs to pass through the extracellular matrix in order to directly contact their targets. Multiple matrix metalloproteinase
(MMPs) , urokinase plasminogen activator (uPA) and its receptor are expressed by NK cells , and can be augmented upon stimulation. Intriguingly, the journey through extra cellular matrix was suggested not to wear out NK cells but rather to further activate them for killing [33-35]. Such activation might be enhanced in vivo due to the release of cytokines, including IL-2, from collagens by the proteases activity in the extracellular matrix . Histological examinations revealed that most tumor infiltrating NK cells reside in the stromal area, and are not in direct contact with the tumor cells . This suggest for another protective mechanism, by which the extracellular matrix and stroma around the tumor absorb the infiltrating NK cells and reduce their trafficking toward the tumor cells . The characterization of NK cells behavior at the tumor territory would lead to better tumor targeting. A direct imaging of NK cell movements in vivo was recently demonstrated and NK cells showed to have a slower-speed in the lymph node (2.75 ^m/min) compared to T cell (9.6 rcm/min) . It would be interesting to perform such an intravital imaging of NK cells in the tumor microenvironment.
Was this article helpful?