Dear Pedro,

Try out spheroid culture that might be of some help. This assay is frequently used to validate the right population of cancer stem cells possessing self renewal property.

The exact definition of GBM stem cells is different from that of normal stem cells due to the complexity of cancer stem cells.  ADAM17 is reported to regulate the self-renewal potential and trans-differentiation of CD44+/CD133+ glioblastoma stem cells in vitro.  Knockdown of ADAM17 inhibits Hes1 and Hes5 and activated Notch1 expression, which explains the ADAM17 shRNA-induced suppression of self-renewal and differentiation of  GBM stem cells. 

The molecular machinery of how tissue stem cells maintain the cellular polarity during the symmetric division (self-renewal) is a complex issue, so that I would like to focus on hemidesmosome-relared proteins. In the case of symmetrical cell division often observed in proliferative stem cell in the niche, the round co-localization of Rac1 and ColXVII (also named BP180) is thought to be necessary. In contrast, in the case of asymmetrical cell division which provides 1 stem cell and 1 progenitor cell (=relatively differentiated cell), the apical aPKC/Rac1/Par3 localization and the linear-pattern localization of ColXVII at the basement membrane is frequently recognized. Knockdown of ColXVII induces the exhaustion of stem cells particularly in the hair follicle stem cells due to the disruption of the cellular polarity. After all, the abnormal polarity of stem cells fails to maintain the stem cell pool providing the precursor cells for the rapid turn-over and wound-healing in response to tissue damage.

Collectively, I recommend you to evaluate the following molecules which regulate the self-renewal potential; Par3, Par6, aPKC, ColXVII, Gli1 (Shh activation hallmark), Cdc42, ADAM17, CD44

First of all, thank you very for the time you took to help me on this. 

Sorry for the confusiing question. I will try to clarify a little bit more:

Our goal is to sensitize GBM to chemotherapy by modulating microRNA levels, and in that process hit the major source of resistence that are the CSCs. miR-21 served that purpose in a previous work with orthotopic GBM mice models but we are now looking for aditional miR whose modulation may sensitize CSCs (miR128 is our main target so far).  

1) We are using two human GBM cell lines, the well known U87 and another called DBTRG, from which we intent to isolate the cancer stem cells.

2)The isolation protocol is indeed cell sorting, but with a magnetic activating rather than fluorescence activating strategy. We use for that CD133 antibodies associated with magnetic beads. So we rely on this CD133 to adquire the cancer stem cells. 

a) After isolation, and as Dr. Singh proposed, we indeed check if the isolated cells grow in spheres, in a media containing bFGF and EGF. And they do form spheres. But also the CD133- are forming some spheres (especialy the ones isolated from U87)

3)So we think we are obtaining an enriching  cancer stem cell population rather than an exclusive CSCs fraction.  

a)Nestin cytometry analysis show little difference between CD133- and CD133+ cells.

b) With qPCR analysis we also didn't find significant differences in Sox 2, Oct4, Nanog and GFAP between the two populations, although some ABC transporters seem to be overexpressed in CD133+ cells.

4) Because in my master thesis I will not have time/money to spend on the ultimate in vivo experiment to validate the cells as CSCs, I was looking for what i beleave to be the 2nd best alternative (evaluate self-renewal/assymetric division).

a) what I meant by microscopic techniques (and I was not clear at all) was live imaging, because i would not have nor the expertise nor time to do that yet again.

Overall:

Thank you Dr. Duntsch for the extensive review and proposed experiments. Especially I think the culture with the FBS containing media will indeed prove my point. Especially when combined with some of the markers Dr. Yoshida proposed. For that and for the amazing lesson on CSC self-renewal I thank you very very much. 

I've already ordered  primers for two more markers: CD15 (presumably a recent GBM stem like-cells marker) and Tuj-1 (neuron marker). Nevertheless, I will ask my supervisors to try at least some of the markers you've proposed. For the most promising ones I will check protein content by WB/imunohistochemistry. Only then I will be sure when analysing possible sensitization effects with our combined strategy.

You should be exceedingly cautious in the use of CD markers and their combinations as identifiers of stem cells (SCs) or cancer stem cells (CSCs). Most of these markers select far too large a fraction of the cells to be SCs or CSCs. One of the best studies using such markers used 3 markers to identify and select a supposed CSC "phenotype" for breast cancer, and collected 2% of the tumor cells with such phenotype. When a limiting dilution assay was done in NOD/SCID mice,100 such cells yielded 60 - 62.5% of the animals growing tumors in their "humanized" breast fat pads (Filmore & Kuperwasser). However, this does not mean that the cells with this phenotype are the breast CSCs. As shown by Hewitt & Wilson in 1958, when injecting spontaneous leukemia cells into syngeneic mice, a single cell may transfer the leukemia, but when a volume of cell suspension containing an average on one cell per injection, the Poisson distribution applies and the probability of zero cells (and therefore no cancer) when the average input is one cell, is 37%. Hence 63% get one or more cells and the cancer. Thus of the 100 cells yielding 60-62.5% animals with tumors, only one cell was the CSC. Therefore, the original CSC frequency in the tumor was 1% of 2%, or 2x10^-4. This is in excellent agreement with our result of 4x10-4 using our patented Hybrid Spheroid Assay (and no marker selections). Consider the CD markers as in a Venn diagram, the cells matching the phenotype may contain amongst all such cells, a (tiny) fraction which are actually the CSCs.The rest share the "phenotype" but are not the cells of interest (99% in this case) and hence studies of such cells will yield misleading results.

Thank you very much Dr. Christopher Lange.

What you are saying is that for CSC isolation one must choose an assay that relies on their functional differences raher than looking for some marker that we don't even know it's functional.

It's correct to assume that all the CSC will grow in spheroid like collonies and none of the more differentiated cancer cells will be able to do so?

Are the (cancer) stem cells capable of growing in spheroids because of some adhesion protein like a cadherin? Does that help them mantain their undifferentiated phenotype?