Flower power as human cancer cells compete with normal cells

In multicellular organisms, cells usually communicate with each other in a peaceful manner. But harmony is shattered if abnormal cells appear and battle for space and survival with normal cells in a process called cell competition. This process was identified in the fruit fly Drosophilamelanogaster¹, and it also occurs in mammals². When cancer starts to form in mammals, competition occurs between normal cells and those on a path to tumour formation, but how the molecular differences between such cells trigger cell competition is poorly understood. Writing in Nature, Madan et al.³ report that whether competing mammalian cells win such a battle depends on which version of a membrane protein called Flower they express.

Earlier work⁴ in D. melanogaster revealed that, during cell competition, the cells that survive (winner cells) are distinguished from the cells that die (loser cells) by the version of Flower protein (FWE) that they express. Different versions of this protein are made when the messenger RNA that encodes FWE undergoes a process called alternative splicing. Loser cells express a version of FWE termed FWE^Lose and winner cells express a version termed FWE^Win. When loser cells come into contact with winner cells, the former cells die and the latter divide to fill the empty space left by this cell death⁴. An encounter between cells that express FWE^Lose and those that express FWE^Win is required for cell death to occur because cells that express FWE^Lose survive if cells expressing FWE^Win are absent⁴. However, whether a comparable FWE system functions in mammals was unknown.

Madan and colleagues report that alternative splicing generates four versions of human FWE, which they term FWE1, FWE2, FWE3 and FWE4. Using human breast cancer cells grown in vitro, the authors examined whether any of these versions of FWE behave as winner or loser forms of the protein. They report that, when human cells that express either FWE1 or FWE3 are cultured with cells that express either FWE2 or FWE4, cells that express FWE1 or FWE3 die and those that express FWE2 or FWE4 divide to compensate for the loss of the dying cells. Thus, in this system, cells that express FWE1 and FWE3 are loser cells and those that express FWE2 and FWE4 are winner cells. Loser cells undergo a type of cell death called apoptosis, and the initiation of cell death requires direct contact between winner and loser cells.

The authors examined the expression of winner and loser versions of FWE in samples of human cancers. FWE^Win expression was higher in malignant tumours than in benign tumours. Madan and colleagues found that expression of FWE^Lose in normal cells adjacent to the tumour is higher than in normal cells farther away from it. Moreover, the level of FWE^Lose was higher in normal tissues adjacent to malignant tumours than in normal tissues that surrounded a benign tumour.

When the authors transplanted human breast cancer cells that express FWE^Win into mice, the mouse cells adjacent to the transplanted tumour cells increased their expression of mouse FWE^Lose compared with the levels in animals that had not received a tumour transplant. All these results suggest that FWE^Win expression in tumour cells induces FWE^Loseexpression in neighbouring normal cells (Fig. 1). The mechanism responsible for such induction is unknown.

The authors report that, when human breast cancer cells expressing FWE^Win were transplanted into the breast region of mice engineered to express human FWE^Lose, the transplanted cells generated aggressive tumours. By contrast, less aggressive tumours were generated if FWE^Lose-expressing human breast cancer cells were transplanted into mouse breast tissue that expressed human FWE^Win. This indicates that it is the combination of high expression of FWE^Win in tumours and high expression of FWE^Lose in the tissue that surrounds them that aids cancer growth.

When the authors engineered human cancer cells to block expression of FWE and transplanted these cells into mouse legs, the cancer cells showed diminished growth and reduced capacity for migration (termed metastasis) to a secondary site compared with transplants of human cancer cells in which FWE expression was not blocked. When chemotherapy was also administered, growth of the engineered human cancer cells in the mouse legs was substantially inhibited.

Madan and colleagues suggest that FWE should be investigated as a possible therapeutic target in human tumours and in the tissues that surround them. However, whether human FWE can be selectively targeted using antibodies or chemical compounds should be examined before a clinical approach can be considered.

The authors have demonstrated convincingly that, in addition to its known role in D. melanogaster, FWE also functions in cell competition in mammals. In both mammals and flies, the expression of FWE^Lose is induced in loser cells; cells expressing FWE^Lose die only if they encounter cells that express FWE^Win; and it is the relative rather than the absolute levels of FWE^Lose and FWE^Win that trigger cell competition.

Several issues remain to be addressed. For example, the regulatory proteins that act upstream or downstream of FWE have not been identified. What controls the alternative splicing that generates different forms of FWE is unknown, and understanding this process might reveal other therapeutic targets. Previous work⁴ suggests that membrane proteins of unknown identity can distinguish between winner and loser versions of FWE expressed on neighbouring cells. If such proteins exist, their identification will be necessary to understand how FWE-mediated cell competition functions.

Another key question is whether cancer-promoting mutations trigger FWE-mediated cell competition in mammals, and, if so, which mutations are responsible. There are reports that abnormal expression of the tumour-promoting proteins Myc or Wnt is involved in FWE-related cell competition in D. melanogaster^4–6. Analyses of tumour cells from patients might shed light on whether this also occurs in humans.

Madan and colleagues’ work should motivate researchers to analyse human-tumour samples to determine the involvement of FWE in cell competition and cancer development. If antibodies could be developed to specifically recognize human FWE^Lose proteins, this would greatly aid such studies. However, generating such antibodies is not straightforward, and the authors discuss the technical hurdles that would need to be overcome.

In D. melanogaster, other proteins in addition to FWE can regulate cell competition^7,8, and further studies in human cancer cells will be needed to gain a more complete picture of mammalian cell competition. Such work might offer new perspectives for improving cancer treatments.