One protein, two partners: a balancing act with cancerous implications

From the Bai, Eisenman and Moens Laboratories (Basic Sciences Division) and the Grady Laboratory (Clinical Research Division)

During cancer metastasis, cellular signaling is re-wired to promote migration and proliferation. “To date, how cells modify signaling pathways to acquire such aggressive behaviors remains unclear,” says Dr. Jihong Bai, an Associate Member in the Basic Sciences Division. Changes in signaling are mediated in part by the endocytic machinery, which transports nutrients and signaling molecules into the cell by engulfing them in membrane-bound compartments. Altered activity of endocytic proteins has been observed in many cancers, but how these changes promote cancer progression is not well understood.

The endophilin A protein family facilitates endocytosis by binding both to membranes and to other proteins. Dr. Raj Poudel, a postdoctoral fellow jointly advised by Dr. Bai and Dr. Bob Eisenman, studies the endophilin A family member EndoA3, which has been linked to cancer invasiveness. In work recently published in Developmental Cell, Dr. Poudel and his colleagues in the Bai and Eisenman labs, in collaboration with the laboratories of Dr. Cecilia Moens and Dr. Bill Grady, report a mechanism by which EndoA3 balances the migratory and proliferative behaviors of colorectal cancer cells.

EndoA3 balances cell growth and migration through mutually exclusive interactions with membranes vs. the guanine nucleotide exchange factor TIAM1. Image provided by Dr. Jihong Bai

To analyze the connection between EndoA3 and colon cancer, the researchers first compared EndoA3 expression levels in cancerous tissue vs. normal tissue using immunohistochemistry. They observed that EndoA3 expression levels were higher in cancerous regions, and that more advanced cancers tended to express the most EndoA3. Next, the authors showed that over-expression of EndoA3 in colon cancer cell lines increases the rates of growth and migration, while siRNA knockdown of EndoA3 attenuated both of these proliferative phenotypes. In addition, EndoA3 expression was found to increase the abundance and dynamics of membrane projections called filopodia and lamellipodia.

Endophilin A proteins each possess a membrane-binding (BAR) domain and a protein-binding (SH3) domain. To determine the role of EndoA3-membrane interactions in promoting cell migration and proliferation, the authors deleted an N-terminal amphipathic helix of EndoA3’s BAR domain (EndoA3ΔN) and asked how this affected the EndoA3-induced phenotypes. While EndoA3ΔN was less effective than full-length EndoA3 at promoting endocytosis and cell growth, the researchers were surprised to find that EndoA3ΔN was more effective at promoting cell migration and filopodia formation, suggesting that EndoA3’s membrane binding activity inhibits the protein’s ability to drive invasive phenotypes.

To investigate how EndoA3 promotes migratory behaviors, the authors performed a yeast two-hybrid screen to identify other proteins that bind EndoA3. They found that the EndoA3 BAR domain strongly binds to TIAM1, a guanine nucleotide exchange factor (GEF) that regulates the activity of the small GTPase Rac1. Because Rac1 is known to drive filopodia formation and cellular motility, the researchers hypothesized that EndoA3 might promote invasiveness by stimulating TIAM1’s ability to activate Rac1. Consistent with this idea, they observed that EndoA3 expression increased the levels of active Rac1 and that EndoA3’s ability to promote filopodia formation requires Rac1.

Because membranes and TIAM1 both bind to EndoA3 via its BAR domain, Dr. Poudel and his colleagues next asked whether binding to both partners can occur simultaneously. By performing a liposome co-sedimentation assay, they found that addition of TIAM1 reduces the proportion of EndoA3 that binds to membranes. Reciprocally, they measured Rac1 activity in the presence or absence of phospholipid membranes and found that membranes interfered with EndoA3’s ability to enhance GTP hydrolysis by Rac1. In addition, artificially anchoring EndoA3 to the membrane via a myristoylation tag also reduced Rac1 activation, suggesting that membrane-bound EndoA3 cannot engage TIAM1. Thus, the researchers concluded that binding of EndoA3 to membranes and TIAM1 is mutually exclusive.

Together, the results indicate that EndoA3-membrane binding promotes cell proliferation, while the EndoA3-TIAM1 interaction promotes cell motility. To determine how these competing interactions balance each other during cancer progression in vivo, the researchers injected EndoA3- or EndoA3ΔN-expressing colon cancer cells into the tail veins of mice and measured their ability to invade the lungs. Consistent with the finding that EndoA3-membrane interactions antagonize EndoA3’s ability to promote cell migration, the authors observed that cells expressing EndoA3ΔN (which does not bind membranes) formed more clusters in the lung compared to cells expressing full-length EndoA3. However, the area of the EndoA3ΔN clusters was significantly smaller, indicative of slower growth. Thus, EndoA3 promotes different aspects of cancer progression (proliferation vs. migration) depending on whether or not it is membrane-bound.

Several mutations have been documented in cancers that are expected to interfere with EndoA3-membrane interactions. Thus, the authors propose that disruption of EndoA3’s ability to balance its competing functions in proliferation and migration might contribute to cancer metastasis. Future work toward manipulating the distribution of EndoA3 between cytosolic and membrane-bound pools may yield a means of reducing cancer invasiveness.


Poudel KR, Roh-Johnson M, Su A, Ho T, Mathsyaraja H, Anderson S, Grady WM, Moens CB, Conacci-Sorrell M, Eisenman RN and Bai J. (2018) Competition between TIAM1 and Membranes Balances Endophilin A3 Activity in Cancer Metastasis. Developmental Cell 45(6):738-752.e6.

This work was supported by the National Institutes of Health and the American Heart Association.