Biomedical Engineering Reference
In-Depth Information
derived from BCCs overexpress Gremlin 1 while non-tumor-derived skin cells
do not (Sneddon et al. 2006), suggesting that the tumor can influence micro-
environmental cues, tipping the balance toward growth of pCSCs and possibly
mCSCs. Other interactions between tumor stroma and environment involving
immune surveillance and hypoxia pathways have been shown to play a role in
metastatic potential (Gupta and Massague 2006; Barnhart and Simon 2007)
and may influence the proliferation of mCSCs.
Some of the most exciting developments pertaining to mCSCs regard their
ability to condition a pre-metastatic niche in distant tissue prior to tumor
dissemination. Tumor tropisms have long been explained by the ''Seed and
Soil'' hypothesis posited by Paget (Paget 1989) in which cancer cells are
dispersed ''in all directions'' but only colonize tissue which happens to provide
a supportive microenvironment ideal for growth. Recent functional genomics
studies have at least partially validated this model by demonstrating subpo-
pulations of tumor cells with tissue-specific metastatic tendencies (Kang 2005;
Minn et al. 2005a). On the other hand, gene signatures identified through this
approach could also be used to predict organ-specific metastasis tendency
based on the overall primary tumor gene expression profile (Minn et al.
2005b). Explanation for these preprogrammed tendencies came in a ground-
breaking report in 2005 showing lung and melanoma cells recruit factors to
future metastatic sites (Kaplan et al. 2005). Secretion of vascular endothelial
growth factor (VEGF) and placental growth factor (PlGF) by primary tumor
cells recruits bone marrow-derived cells (BMDCs) to what is now referred to
as the pre-metastatic niche. Interfering with BMDC induction by blocking
secretion of these factors virtually eliminates metastasis in these cancers.
Injecting animal hosts with media conditioned by tumor cells was sufficient
to re-direct tumor tropism to the organ associated with the source of secreted
factors. The mechanism of this effect is still unclear, but potential candidates
for involvement are the cytokines S100A8 and S100A9 which are expressed by
myeloid cells and upregulated in lungs of pre-metastatic tumor-bearing mice.
While untreated animals develop metastases, inhibiting the overexpression of
these cytokines in lung tissue greatly reduces metastasis to lung tissue. Induc-
tion of the p38 MAP kinase pathway involved in tumor cell migration and
invasion likely helps mediate this cytokine response (Hiratsuka et al. 2006).
Activation signals may also be required for metastatic growth even after
mCSCs reach the secondary site. It is thought that mCSCs, for instance in
breast cancer, can migrate to bone marrow or lung tissue and enter a quiescent
state (Li et al. 2006) thus requiring an activation signal to induce growth of
macrometastatic lesions. While there is little evidence for some steps of the
mCSC-mediated metastatic cascade (such as reactivation), the overall model
provides a working framework of mCSC function as summarized in Fig. 2.
This model highlights new potential interactions of mCSCs with the meta-
static niche and illuminates novel
therapeutic targets to block these
interactions.
