Clonal cooperation

Widespread genetic heterogeneity in cells of human tumours poses a question: what prevents the fittest clone from taking over? A demonstration of interdependence between distinct clones might shed light on this puzzle. See Letter p.113

Sequencing of human cancer genomes has revealed a high degree of genetic heterogeneity among cells of a given tumour1. In most cases, tumour growth is thought to be driven by the most 'advanced' cancer-cell subpopulation — that carrying the highest number of cancer-driving mutations. However, the presence of many mutations that occur at only low frequency implies that tumours contain multiple subclones, and the relevance of these is not fully understood. On page 113 of this issue, Cleary et al.2 provide a potential explanation for some forms of intratumoral heterogeneity, by describing a cooperative cellular interaction in mouse mammary tumours in which the presence of two types of clone is required for tumour formation.

Mammary tumours induced by overexpression of the Wnt1 gene are thought to originate in mammary epithelial stem cells3, which can differentiate into both the luminal and basal cells that make up mammary epithelial tissue. These tumour-initiating cells can therefore give rise to tumours composed of cancer cells with either basal or luminal features. It was previously established4 that the coexistence of these two lineages is maintained by paracrine interactions between the two cell types — that is, short-distance interactions mediated by signalling molecules — because only luminal cells produce Wnt1, and basal cells depend on this signalling protein to proliferate. The first report5 of Wnt1-induced mammary tumours noted that many tumours contain more than one clone, but the mechanistic basis of this heterogeneity had not been explored.

Now, Cleary et al. find that, in a subset of Wnt1-induced mammary tumours, the basal cells harbour a potent cancer-driving mutation in the Hras gene that is not detected in the cancerous luminal cells. This suggests that the two differentiation lineages in the tumours represent different tumour clones. Both of these lineage-restricted clones are essential for tumour formation, because neither was able to generate tumours by itself when transplanted into secondary recipient mice, whereas the combined transplantation of both cell types was highly tumorigenic (Fig. 1).

Figure 1: Luminal and basal clones cooperate to form mammary tumours.

Mammary tumours that develop in mice expressing the Wnt1 transgene are composed of heterogeneous mixtures of basal and luminal cancer cells. Cleary et al.2 show that, in some such tumours, neither luminal nor basal cancer cells alone can form a new tumour when injected into the mammary fat pads of tumour-prone mice, whereas mixtures of the two cell populations induce biclonal tumour formation with high efficiency. In such tumours, the basal cells also carry cancer-driving mutations in the Hras gene. Thus, it seems that paracrine interactions between the Hras-mutant basal cells and the luminal cells, which provide the Wnt1 on which the basal cells depend, is required for the development of these cancers.

To study this interclonal cooperation further, the authors used transgenic mice in which Wnt1 expression is regulated by a gene promoter that requires the small molecule doxycycline for activity; in this model, the tumours regress following doxycycline withdrawal. Transplantation of cells from these inducible tumours into the mammary fat pads of wild-type mice or mice in which Wnt1 is continuously expressed led to tumour generation in the presence of doxycycline in both cases. However, the transplanted tumours behaved differently in the two recipient strains after doxycycline withdrawal. Complete regression occurred in the wild-type animals, although small numbers of residual tumour cells survived and led to tumour recurrence after doxycycline readministration. By contrast, only partial regression was observed in mice continuously expressing Wnt1, and the initial regression was followed by rapid tumour regrowth, despite the continued absence of doxycycline.

Most surprisingly, the authors found a large number of Wnt1-transgene-expressing luminal cells from the recipient mice in these recurrent tumours, suggesting that the void left by loss of Wnt expression in tumour luminal cells can be filled by recruiting Wnt-expressing wild-type luminal cells. As with the parental tumours, both Hras-mutant basal cells and Hras-wild-type luminal cells were required for efficient tumour maintenance.

Although this study provides convincing evidence for cooperation between basal and luminal clones and for the requirement of both cell types in Wnt1-driven tumours, extrapolating these findings to explain subclonal intratumoral heterogeneity in human breast cancer is not straightforward. The reasons for this are that, first, the authors used a transgenic mouse model in which the transgene is expressed in a large number of mammary epithelial cells. Thus, biclonal tumours could have evolved following tumour initiation in two independent cells that happened to be next to each other, rather than reflecting two subclones derived from a single initiating cell. The likelihood of two independently initiated transformed cells forming a single tumour mass in humans is very low, but the overall conclusions of the study regarding clonal cooperation are still valid.

Second, the paracrine interdependence between basal and luminal cells, which enables independent evolution within the two differentiation lineages in this mouse cancer model, might represent a special case that is not necessarily relevant to primary human cancers. Although the Wnt1 pathway is one of the main cancer-causing pathways, Wnt1 activation has rarely been observed in human breast cancers, and most such tumours do not depend on Wnt1 activity for growth. Paracrine signalling between basal and luminal cells plays a key part in normal mammary-gland development6,7, but the degree to which these interactions are preserved in tumours is uncertain, especially given that most human breast tumours are thought to originate in luminal progenitor cells. Furthermore, because human breast tumours can take decades to evolve, the extent of intratumoral heterogeneity in such tumours is much higher8,9 than that in animal models. Although the existence of paracrine interdependence relieves competition between the lineage-restricted subclones, lack or loss of strict interdependence is expected to allow the fittest clone to take over the tumour. Thus, other mechanisms besides paracrine interdependence between clones may be responsible for intratumoral heterogeneity in human breast cancers.

Despite these differences between the model and human cancers, Cleary and colleagues' study will fuel the already-intense exploration of the functional and clinical relevance of intratumoral heterogeneity. More importantly, their results raise questions about the validity of assays for identifying tumour-initiating cells (also known as cancer stem cells) in human tumours that rely on transplanting single cells or homogeneous cell populations10. Such assays would inevitably fail to determine the tumorigenic potential of clones that require cooperation with other cells. Many different experimental approaches, in combination with detailed molecular characterization of large cohorts of human tumours at different stages, will be required to answer the many lingering questions regarding intratumoral heterogeneity. Let us hope that acquiring a better understanding of this phenomenon will not only be of scientific interest, but also lead to improved treatment for cancer patients.


  1. 1

    Greenman, C. et al. Nature 446, 153–158 (2007).

  2. 2

    Cleary, A. S., Leonard, T. L., Gestl, S. A. & Gunther, E. J. Nature 508, 113–117 (2014).

  3. 3

    Li, Y. et al. Proc. Natl Acad. Sci. USA 100, 15853–15858 (2003).

  4. 4

    Kim, S., Goel, S. & Alexander, C. M. PLoS ONE 6, e19310 (2011).

  5. 5

    Mester, J., Wagenaar, E., Sluyser, M. & Nusse, R. J. Virol. 61, 1073–1078 (1987).

  6. 6

    Rosen, J. M. & Roarty, K. Breast Cancer Res. 16, 202 (2014).

  7. 7

    Forster, N. et al. Dev. Cell 28, 147–160 (2014).

  8. 8

    Almendro, V. et al. Cell Rep. 6, 514–527 (2014).

  9. 9

    Almendro, V. et al. Cancer Res. 74, 1338–1348 (2014).

  10. 10

    Meacham, C. E. & Morrison, S. J. Nature 501, 328–337 (2013).

Download references

Author information

Correspondence to Kornelia Polyak.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Polyak, K., Marusyk, A. Clonal cooperation. Nature 508, 52–53 (2014).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.