Breast cancer cells chemotact to specific regions of lymph node in an ex vivo model of lymph node metastasis

Introduction::
Metastatic disease remains the underlying cause of mortality in breast cancer patients. Located downstream from the primary cancer site, the sentinel lymph node (LN) is the first organ encountered by cancer cells and is a common site of the initial metastatic seeding. In breast cancer, LN involvement is clinically indicative of progressive disease and poor prognosis. While it is clinically established that breast cancer often engages the LNs, the understanding of the mechanisms required for metastatic seeding of the LN is currently limited. The migration and spread of metastatic cancer cells resemble trafficking of leukocytes, which is controlled by chemokine signaling. Even though several chemokines, including CXCL12, CXCL13, CCL1 and CCL21, have been suggested to play a role in breast cancer metastasis, the exact mechanisms that dictate the targeted migration and invasion of cancer cells into LNs have yet to be determined. It is unknown whether tumor cells are chemotactically guided specifically to and within the LN by hijacking chemokine pathways evolved for recruitment of circulating leukocytes. We postulate that mapping tumor cell migration within the LN while controlling chemokine signaling pathways may reveal the fundamental mechanisms of metastatic spread. However, existing experimental models fail to simultaneously capture the sequential steps of LN metastasis and identify the factors controlling tumor cell invasion into the complex tissue architecture. Here, we developed an ex vivo model of breast cancer cells invading spatially organized LN tissue slices to enable analysis of cancer cell migration within the LN.




Materials and Methods:: To establish an ex vivo model that enables breast cancer cell seeding and examination of cancer spread within the LN ex vivo, we utilized a live murine LN slice model developed by the Pompano Lab. To assess the impact of biochemical factors secreted by LN tissue on cancer cell migration, LN conditioned media collected after 20h of culture and individual chemokines were used in a 3D transwell migration assay. ELISA was used to measure the concentration of chemokines secreted by the LN tissue. To seed breast cancer cells on the LN slice, a suspension of fluorescently labeled BRPKp110 breast cancer cells (C57Bl/6 strain background) was added onto the surface of naïve LN slices from female C57Bl/6 mice and incubated for 1h. Following 1h seeding, free-floating cancer cells were removed by media replenishment and the LN slices were cultured for 40h. Immunofluorescent antibody staining of LN tissues was used to define the general anatomy of LN slices and matrix-bound fraction of chemokines. Using Fiji ImageJ software, LN regions were defined by image segmentation relative to fluorescent immunostaining. Cancer cell distribution was quantified as area of cancer cells relative to area of LN region, normalized to the total area of each LN slice.


Results, Conclusions, and Discussions::
First, we assessed the potential for tumor cells to be chemoattracted to media that was conditioned by naïve LN tissue slices. Conditioned media promoted significant increase in BRPKp110 migration in comparison to control media (Fig.1a,b). This effect was abolished in cancer cells pretreated with Pertussis toxin, suggesting it was mediated by chemokines. Consistent with this, live LN tissue slices secreted detectable levels of CXCL12, CXCL13, and CCL21 chemokines, while CCL1 secretion was below the level of detection (Fig.1 c).




Furthermore, BRPKp110 cells cultured in media expressed surface receptors CCR8, CXCR4 and CXCR5, although not CCR7 (Fig.1d). Interestingly, surface expression of CCR7 and CXCR5 receptors was upregulated by BRPKp110 by 20h culture in conditioned media (not shown). We found that BRPKp110 cells migrated towards CCL21, CCL1, CXCL12 and CXCL13 across a range of concentrations (Fig. 1e). Collectively, these data demonstrated that BRPKp110 cells express functional chemokine cognate receptors and were chemoattracted to chemokines secreted by LN tissue.









Next, we assessed invasion of breast cancer cells in ex vivo LN tissue, by seeding suspensions of BRPKp110 cells onto live LN slices. Cells invaded rapidly, and we observed that after 20h culture, the shape of cancer cells changed from rounded to elongated, characteristic of cell adhesion and spread (Fig.1 f). Immunofluorescence staining showed that cells preferentially accumulated in the LN cortex and B cell follicles (not shown). To determine whether BRPKp110 cells were enriched in chemokine rich regions, BRPKp110 cell invasion in chemokine-rich regions was compared to regions of the LN where chemokine immunofluorescence was absent. We found that BRPKp110 invasion was significantly increased in the CXCL13+ region of the LN (Fig. 1g, h).









In summary, transwell assays and invasion into intact ex vivo LN tissue slices revealed that BRPKp110 migrated to chemokines secreted by the LN and preferentially accumulated in the CXCL13+ regions of the LN tissue, suggesting involvement of chemokine signaling in LN targeted invasion. We are currently testing the requirement for chemokine signaling for invasion into LN slices and region-specific enrichment. Future work will investigate breast cancer chemokine-mediated invasion into tumor-draining LNs.




Acknowledgements (Optional): :
This work was supported in part by the National Institute of Allergy and Infectious Diseases (NIAID) under award number R01AI131723 to RRP, as well as the UVA Carter Immunology Center under a collaborative Research Award to RRP and MRR, and the UVA Cancer Center 2022 Trainee Fellowship to KM.




References (Optional): :