Glioblastoma cell invasion: Go? Grow? Yes

A key feature of glioblastoma that makes it difficult to treat is the highly invasive nature of glioblastoma cells. These invasive cells lie beyond the reach of acceptable surgical margins and focused radiation therapy, making recurrence inevitable. Further complicating the situation is that cells can exist in a variety of substates, and can interconvert between them, making heterogeneity and plasticity potential means of resisting treatments focused solely on one substate or another. Examples of these substates include those defined not only by transcriptomic features, but also by functional features including cell migration, with cells exhibiting varying degrees of migration speeds.¹ The study of Ratliff et al. in this issue of Neuro-Oncology sheds important new light on the ability of cells to sustain their proliferative capacity while also being highly invasive.²

A major hypothesis that Ratliff et al. address is the “Go or Grow” hypothesis, which posits that cells exist in either of 2 states, a highly migratory-slowly proliferating state or a slowly migrating-rapidly proliferating state. The Go or Grow hypothesis has its origins in the analysis of glioblastoma cell migration in vitro.³ Since then, the Go or Grow hypothesis has been highly influential, not only in neuro-oncology, but across cancer research. In addition, the Go or Grow hypothesis attracted the attention of mathematical modelers, leading to numerous studies of the potential impact of such phenotype switching in tumor progression. However, whether such phenotype switching is occurring in vivo has remained an open question.

Ratliff et al. address this question using an elegant optical imaging approach where a live cell fluorescent reporter provides a readout of the extent of proliferation of the past history of the cell. The main fluorescent reporter that they used was a green fluorescent protein fused to histone H2B under the control of Tet-Off where doxycycline turns off transcription. In their system, tumor cells had high H2B-GFP expression until the mice had doxycycline added to their water. Cells that do not proliferate remain strongly fluorescent, while those that proliferate incorporate endogenous (nonfluorescent) H2B so that their brightness reduces by roughly a factor of 2 with each round of division. Thus, nonproliferative cells remain fluorescent, while proliferative cells progressively dim. Since these cells can also be tracked by fluorescence microscopy, their migratory capacity can be assessed concurrently with their proliferation.

Importantly, Ratliff et al. conducted these experiments in live animals and tracked glioblastoma cell proliferation and migration using intravital 2-photon fluorescence microscopy. They found that highly invasive cells that had migrated far from the tumor mass had in fact undergone multiple rounds of proliferation in their past history, demonstrating that invasive cells can remain proliferative as they invade. Over time, these invasive cells became less migratory and less proliferative, again in contrast to the prediction of the Go or Grow hypothesis. Finally, they also showed that in resected glioblastomas from patients, tumor cells from the invasion zone had higher proliferative potential than cells from the tumor core region. Altogether, Ratliff et al. developed a strong in vivo test of the Go or Grow hypothesis, and found that glioblastoma cells can actually exhibit Go AND Grow.

While perhaps surprising, recent studies have already questioned the generality of the Go or Grow hypothesis both experimentally and theoretically. For example, Ulrich et al. showed that both migration and proliferation of glioblastoma cells in vitro increase with mechanical stiffness of the cell culture surface, demonstrating that “Go” and “Grow” are not always mutually exclusive.⁴ In addition, Vittadello et al. showed that migration speed for 3 melanoma lines in vitro was independent of cell-cycle phase as measured by fluorescent ubiquitination-based cell-cycle indicator technology, and that blocking cell-cycle progression had no effect on cell speed.⁵ From a theoretical perspective, recent modeling showed that complex multilayer structures characteristic of glioblastomas could be recapitulated without having to assume a Go or Grow dichotomy.⁶

From a biophysical perspective, one might intuitively expect that cells cannot truly proliferate and migrate at the same time. However, even for rapidly dividing cells, mitosis takes 1–2 hours, during which time cells typically round up and remain stationary, while the intervening interphase takes 1–2 days. Thus, rapidly proliferating cells spend relatively little time in mitosis, and during interphase are capable of fast migration, so that cells spend a negligible fraction of the time in an immotile state. Cell growth even occurs during mitosis, although it is slower than interphase and, again, mitosis is relatively short.⁷ In addition, the observation of Go and Grow implies that neither proliferation nor migration by itself is operating close to limits dictated solely by energy availability. The bottom line is that Go or Grow is not a hard constraint; cancer cells can do both, even in vivo, as now shown by Ratliff et al. for glioblastoma cells.

A number of important questions remain for future investigation. For example, Ratliff et al. used immunocompromised mice, and one wonders to what extent the Go and Grow behavior would be altered in an immunocompetent mouse model. In addition, the observation by Ratliff et al. that the highly invasive cells are similar to the OPC/NPC (proneural) subtype, while the less migratory cells are similar to the mesenchymal subtype, appears to be counter to the usual association of the mesenchymal subtype with greater invasiveness.⁸ Also, the connection to glioblastoma stem cells is not addressed, although this is not a critical issue, as the nature and identity of such cells remain debated.⁹ It is interesting to speculate what the results could mean for therapy development. It could be good news, in that an effective antimigratory strategy could result in a less invasive phenotype that is also coordinately less proliferative. For example, inhibition of migration could theoretically promote a physically “jammed” state that in turn suppresses proliferation due to lack of space in which to grow.¹⁰ Regardless, the study by Ratliff et al. shows the functional plasticity of glioblastoma cells in vivo, providing important new insights into tumor evolution.

Funding

D.J.O. is supported by National Institutes of Health (grants U54CA210190, P01CA254849, and U54CA268069).

Acknowledgments

The content of this work is solely the responsibility of the author and does not necessarily represent the official views of the National Institutes of Health.

Conflict of interest statement

The author has no financial or business conflicts to declare.

Declaration

The text is the sole product of the author(s) and that no third party had input or gave support to its writing.

References