An interesting hypothesis in the evolutionary genetics of treating infections and cancers is that if the therapeutic agent does not directly target the pathogen or tumor, then the pathogen or tumor will be less likely to evolve resistance to that agent.  While early work on inhibitors of angiogenesis as potential cancer therapeutics suggested that such treatment did not elicit resistance by the tumor cells (Boehm et al., 1997), a recent study by Conley et al. (2012) raises doubts about the reliability of this notion in the context of antiangiogenic therapy for human breast cancer.

Conley and colleagues established tumors in mice using human breast cancer cell lines by implanting the cells in the mammary fat pads of their experimental subjects.  They then treated the mice with either a vehicle control or sunitinib malate, a tyrosine kinase inhibitor that targets among other molecules the receptor for vascular endothelial growth factor (VEGF).  VEGF plays a role in new blood vessel formation.  The sunitinib was administered either one day after tumor cell injection or when the tumor reached 4 mm in diameter. 

As might have been expected on the basis of prior work, treatment at the early time point delayed tumor formation and reduced tumor growth at 70-plus days post-implantation and also inhibited development of new blood vessels.  Treatment of the tumor-bearing mice at the later time point, when the tumors had reached 4 mm in diameter, also suppressed tumor growth.  Treatment of tumor-bearing mice with the VEGF-inhibiting antibody, bevcizumab, elicited similar effects with respect to both inhibition of tumor growth and blood vessel development.  

However, treatment of tumor-bearing mice with sunitinib or with bevacizumab also resulted in increased numbers (relatively and absolutely) of tumor cells expressing aldehyde dehydrogenase, expression of which has been associated with cells that behave like tumor stem cells.  Furthermore, consistent with the preceding finding, the tumors treated with sunitinib were able to more rapidly form secondary tumors when transplanted to immunodeficient NOD/SCID mice.  Sunitinib-treated tumors were also more hypoxic in comparison to tumors only exposed to vehicle and this relative hypoxia was correlated with a greater percentage of cells expressing aldehyde dehydrogenase, i.e. a greater percentage of cancer cells with stem cell-like properties.

Additional studies suggested that the effects of anti-angiogenic treatment were mediated by increased expression of hypoxia-inducible factor 1α (HIF-1α), one representative of a family of transcription factors known to mediate the effects of hypoxia.  The authors also offered indirect evidence that signaling through the Akt/Wnt/β-catenin pathway was critically involved in the elicitation of the response to hypoxia in human breast cancer stem cells and following anti-angiogenic therapy of these cells.

Summarizing, while anti-angiogenic therapy initially limits tumor growth in this murine model of human breast cancer, it also has the effect of eliciting a cellular phenotype, i.e. increased numbers of tumor cells with stem cell-like properties, that can make the tumor cells grow more rapidly later if they survive the first therapeutic onslaught.  The tumor cells do not become resistant to the treatment in the typical manner through mutation such that the therapeutic agent no longer affects the cells; instead they adopt a new and arguably a more aggressive phenotype that ultimately increases the threat to the host’s survival.  So far as we now know, this new and more threatening phenotype arises primarily via epigenetic, not genetic, mechanisms. 

Thus, contrary to the early and hopeful studies of Folkman and colleagues, even though anti-angiogenic molecules do not directly target breast cancer cells, the malignant cells can evade the desired outcome of the therapy.  Conley et al. conclude that to minimize the clinical consequences of such tumor cell adaptation, therapeutic agents targeting cancer stem cells may need to be used in parallel with angiogenesis inhibitors.

In the near future, it will be interesting to watch for similar phenomena in the context of new therapeutic approaches for viruses, where it may make sense to explore inhibiting host pathways that are co-opted for virus replication.  It may be overly anthropomorphic to say so, but intense selection pressure appears to bring out the creativity, or what looks like creativity, in entities capable of evolution. 


Boehm T, Folkman J, Browder T, O’Reilly MS. Antiangiogenic therapy of experimental cancer does not induce acquired drug resistance. Nature. 1997 Nov 27;390(6658):404-7. PubMed PMID: 9389480.

Conley SJ, Gheordunescu E, Kakarala P, Newman B, Korkaya H, Heath AN, Clouthier SG, Wicha MS. Antiangiogenic agents increase breast cancer stem cells via the generation of tumor hypoxia. Proc Natl Acad Sci U S A. 2012 Feb 21;109(8):2784-9. Epub 2012 Jan 23. PubMed PMID: 22308314; PubMed Central PMCID:  PMC3286974.