Figure 1.

NRP2 is associated with TICs and mediates mammosphere formation.

A. SUM149 and SUM159 cells were sorted by FACS into either stem (CD44⁺/CD24⁻/EpCAM⁺) or luminal (CD44⁺/CD24⁺/EpCAM⁺) populations. Expression of NRP2 was analysed using qPCR. *p = 0.0129 and **p = 0.00499.

B,C. HMEC, HMLE-PR, MCF10A and MCF10AT cells were plated on low adhesion plates to measure their ability to form mammospheres. NRP2 and NRP1 expression was analysed by immunoblotting. Representative data from three independent experiments are shown. *p = 0.0032 (for B) and *p = 0.0077 (for C).

D,E. Epithelial cells freshly isolated from breast cancer biopsies were sorted by FACS into either α6^high or α6^low populations. Cell extracts were immunoblotted to assess expression of NRP2, α6 and actin (D). These α6^high and α6^low populations were analysed for their ability to form mammospheres in the presence of either control IgG or a NRP2 inhibitory Ab (E). These results are consistent using three different primary tumours (*p = 0.0036).

F. The NRP2^high population was sorted from SUM1315 cells by FACS and analysed for their ability to form mammospheres in the presence of either control IgG or a NRP2 inhibitory Ab (*p = 0.0049, three independent experiments).

G. The NRP2^high and NRP2^low populations of SUM1315 were isolated. The ability of the NRP2^high population to form mammospheres in the presence of a VEGF-NRP inhibitory peptide (C-furSema) or control peptide (C-Sema) was determined. Representative data from three independent experiments are shown (*p = 0.028). Scale bar = 100 µm for all panels. For all panels, error bars represent the mean ± SD. Statistical differences between data groups were determined using Student's t-test.