by ecancer Clare Sansom

Prostate cancer is a heterogeneous disease that may be indolent or aggressive; some men can live with the disease for several decades, whereas others progress rapidly and die within months of diagnosis.

Many solid tumours, including prostate tumours, are thought to arise in populations of epithelial stem cells which have the capacity to self-renew.

Three different types of epithelial cell have been identified in the prostate gland, and stem cells have been identified in two of these: the luminal cells, which express high levels of the cytokeratins CK8 and CK18 and the androgen receptor, and the basal cells, which express high levels of p63, CK5 and CK14.

It has been suggested that tumours that arise from different populations of stem cells within the same organ may have different optimal treatments and different prognoses.

A group of researchers led by Michael Shen of Columbia University Medical Center, New York, USA has now shown that, at least in a mouse model, both basal and luminal stem cells in the prostate gland, and that these tumours have different characteristics.

Firstly, Shen and his co-workers marked cells from the basal cell compartment of the mouse prostate gland with yellow fluorescent protein (YFP) and proved that these cells tested positive for the basal cell markers p63 and CK14.

Between 3 and 5% of these YFP marked cells were found to form spheres and to generate prostate ducts in renal grafts; the similar percentages suggested that these might represent the same population of cells with stem-like properties, although this could not be proved.

The researchers then induced prostate regression followed by regeneration by modulating androgen levels, and showed that a very small proportion of the basal cells were so-called “bipotential” cells that could acquire luminal cell characteristics.

The proportion of basal cells with bipotential properties was much smaller than the proportion of those cells that displayed the stem-like properties.

Continuous rounds of alternating prostate regression and regeneration caused a very gradual increase in the proportion of luminal cells in the basal cell population.

This rare transformation from the basal to the luminal cell type was also found to occur during prostate homeostasis in vivo at a similar rate to that observed during regression and regeneration.

Shen and his co-workers then investigated whether prostate cancer could arise in these basal cells using a transgenic mouse model in which one of the most common genetic lesions in human prostate cancer, inactivation of the tumour suppressor gene Pten, could be induced only in prostate basal cells using tamoxifen.

These mice developed Grade III and IV lesions in their prostate glands in the six months following Pten inactivation.

Interestingly, these tumours of basal cell origin were found to contain large numbers of luminal cells.

The researchers then compared these tumours to tumours in similar mice that were known to originate in luminal cells.

The tumours arising from the basal cells had a similar histology to those arising from the luminal cells but developed more slowly.

The transcriptomes of tumours arising in basal and luminal cells were compared at time points when the tumours displayed very similar phenotypes, and found to be similarly enriched in genes associated with tumour initiation and progression.

Nevertheless, small but significant differences in gene expression could be observed between the tumours of basal and luminal origin.

Finally, Shen and his co-workers tested these profiles for the mouse homologs of human genes that had been found to be associated with good and poor prognosis in a cohort of Swedish prostate cancer patients.

They found that the mouse tumours of luminal origin over-expressed significant numbers of genes that were homologous to human genes associated with prostate cancer lethality; no such correlation was observed in the mouse basal tumours.

Many of the sixty-eight genes that were upregulated in both the mouse luminal tumours and the Swedish poor prognosis patient group had previously been associated with prostate cancer.

Taken together, these results indicate that tumours arising from basal and luminal stem cells in the prostate are of different molecular subtypes and that the luminal subtype is associated with poor prognosis; however, basal derived cancer cells are plastic and may differentiate into luminal-like ones.  

Reference

Wang, Z.A., Mitrofanova, A., Bergren, S.K., Abate-Shen, C., Cardiff, R.D.,
Califano, A. and She, M.M. (2013). Lineage analysis of basal epithelial cells reveals their unexpected plasticity and supports a cell-of-origin model for prostate cancer heterogeneity. Nature Cell Biology, published online ahead of print 24 February 2013. doi: 10.1038/ncb2697