Multiparametric Classification Links Tumor Microenvironments with Tumor Cell Phenotype

• Bojana Gligorijevic mail,
  
  
• Aviv Bergman ¶,
  
  
• John Condeelis ¶
  
  

• Published: November 11, 2014
• DOI: 10.1371/journal.pbio.1001995

Author Summary

A large proportion of cancer deaths are due to metastasis—the spread of cancer from the primary tumor to other parts of the body. Movement of cells may require the formation of protrusions called invadopodia, which degrade extracellular matrix. Although some studies have reported on locomotion in primary tumors, the presence of invadopodia was not tested. Here, we show that single cells from mouse mammary carcinoma can move using a fast- or slow-locomotion mode depending on different levels of cues present in the tumor microenvironment. Using multiphoton microscopy in vivo combined with a machine-learning algorithm we show how manipulation of microenvironmental conditions can induce predictable changes in the number of locomoting cells or switch between the two locomotion modes. We also demonstrate that only the slower moving cells are associated with invadopodia in vivo, and that only tumor cells from regions rich in invadopodia degrade the surrounding extracellular matrix and disseminate. Specific targeting of invadopodia results in inhibition of lung metastasis. This work proposes a systems biology view of how tumor microenvironments regulate tumor progression and presents insight into the heterogeneity of the treatment response. The ability to define and predict conditions under which tumor cells disseminate offers potential therapeutic benefits in regulating tumor progression.
http://www.plosbiology.org/article/info:doi/10.1371/journal.pbio.1001995#pbio-1001995-g004