To Iman Loay and others,

The problem with the CD markers is that many cells express them without being stem cells or cancer stem cells. You can use them to enrich for a cancer stem cell (CSC) fraction, but you must realize that the vast majority of these selected cells are still not CSCs. There is a huge difference between seeing a gene expressed in your target population and jumping to the conclusion that ALL of your target cells express this gene and that ONLY your target cells express that gene. It is the difference between detecting the presence of a cell type and counting how many cells of that type you have. Think Venn diagrams. The iPS inducing genes (h Oct3/4, h sox2, h c-Myc, h Klf4, h Nanog, h Lin-28); the first 3 plus either the 4th or 5th should all be expressed) may offer a way to dectect the stem cells, but I don't think that their gene products are on the cell surface, so immunological labeling of live cells may be more difficult. The alternative is to follow the Till functional definition of a stem cell, particularly the requirement for unlimited proliferation capacity (immortality), i.e., a large number of divisions (not the 5-6 of the standard clonogenic assay), i.e., more than the number done by the amplifying transit cells (e.g., 7 for erythroblasts). The other components of the definition are to be able to make more of self (more stem cells), differentiate, and forCSCs, to also be tumorigenic.

A couple of different isoforms. Here's a reference that will get you started on your way.

LJ Bendall et al. Leukemia (2000) 14, 1239–1246

Was not aware CD44 was important in leukemia.

Dear Yoshus Hoosier, Thanks for showing me that reference.

Dear all: I have found out this paper, if you feel interested in, please read it

Jianing Liu and Guosheng Jiang, "CD44 and Hematologic Malignancies" Cellular & Molecular Immunology (2006) 3(5):359-365

CD44 in hematological neoplasias.Ann Hematol. 2011 May;90(5):493-508. Epub 2011 Jan 22,.

This paper may show you what you want .

Dear Dr. Li Jinqing, I can't thank you enough. This is exactly what I want.

I would urge extreme caution in the interpretation of any single CD marker to identify cancer stem cells. We have recently shown (Poster at 14th ICRR in Warsaw, Poland Aug/Sept 2011) that CD24- and CD44+ cells (supposedly markers of breast cancer stem cells) are present in 75% and 85%, respectively of human breast cancer cells taken directly from surgical specimens. CD326/ESA+ cells are 3.3% and the combination of all 3 markers (the supposed "breast cancer stem cell phenotype") yields 1.33+/-0.06% of the initial population. However, only 1-3% of these show stem cell functional properties (e.g., tumorigenesis).

Here you can see some data about the isoforms: http://www.copewithcytokines.org/cope.cgi?key=CD44

Dear Csongor Kolozsi, thanks for your help

It' s really, CD44 are important for the etiopatogenesis and the infiltrate of this cells in leuKemic and in brest cancer can doing single CD a marker to identify cancer stem cells.

As I noted above, one cannot use CD44 to identify breast cancer stem cells (CSCs), and probably not any other CSCs. 85% of breast cancer cells taken directly from the patient were CD44+. This is a ridiculously high number for the CSC fraction of a tumor. The combined triple markers CD24-CD44+CD326/ESA+ identified and separated (magnetic beads linked to said markers) 1.33% of the original tumor cells. This is similar to the 2% Kuperwasser reported for her separations (flow cytometry) of direct from patient cancer cells. However, in limiting dilution assay in NOD/SCID mice, 100 such separated cells were needed to get 60% of such mice to grow a breast tumor. Due to the Poisson distribution of cells/innoculum, an average of one CSC/animal will produce 37% of animals getting no CSC and thus 63% of animals getting 1 or more CSCs and hence growing the tumor (see Hewitt & Wilson classical experiments with transplantable mouse leukemia in the late 1950s); hence the 100 cell requirement to get 60% of the animals to grow a tumor, shows that only 1% of the 1-2% enriched fraction are really CSCs. This means that 99% of the biomarker selected cells were not CSCs. The math is quite simple.

Dear Dr Christopher Lange we are about to study ALDH 1 and Sox 2 to test for cancer stem cells in breast cancer paraffin sections by immunohistochemistry and correlate them with prognostic parameters . we have been told that we should also accompany our research with CD44/CD24 but as i see in your comment you don not suggest use of CD44. what do you think about this research and what do you suggest for us to add to enrich our research taking in consideration that this point has not been tested yet in our department at NCI, Cairo

Overexpression of CD44V6 in the mild to moderate laryngeal dysplasia, ameloblastoma and breast cancer has been shown in several studies, so it seems that this variant of CD44 may have an important role in malignancy.

I think what is important to keep in mind (and I think is Dr. Lange's major point) is that not one "marker" can be relied on to properly segregate the tumorigenic subpopulation from a tumor. The actual fraction of a tumor that is a cancer stem cell can vary widely between patients; this includes marker expression. In glioblastoma, CD133 is often used as a marker but many groups have demonstrated that CD133 is not always a good marker. In fact, there have been studies (the reference escapes me) that have looked at how marker expression changes as a function of time after a tumor has been excised from a patient. In many cases, the marker expression is so fluid that 30 min post-resection the surviving cells have a drastically different expression pattern.

In your case, if the literature has shown that a particular isoform of CD44 is useful in segregating a tumorigenic population then I would say use it but with caution. Always use quantifiable experiments (like tumor formation) to confirm that in your hands, with that sample, in your conditions, sorting by your marker of interest behaves as expected.

To Iman Loay and others,

The problem with the CD markers is that many cells express them without being stem cells or cancer stem cells. You can use them to enrich for a cancer stem cell (CSC) fraction, but you must realize that the vast majority of these selected cells are still not CSCs. There is a huge difference between seeing a gene expressed in your target population and jumping to the conclusion that ALL of your target cells express this gene and that ONLY your target cells express that gene. It is the difference between detecting the presence of a cell type and counting how many cells of that type you have. Think Venn diagrams. The iPS inducing genes (h Oct3/4, h sox2, h c-Myc, h Klf4, h Nanog, h Lin-28); the first 3 plus either the 4th or 5th should all be expressed) may offer a way to dectect the stem cells, but I don't think that their gene products are on the cell surface, so immunological labeling of live cells may be more difficult. The alternative is to follow the Till functional definition of a stem cell, particularly the requirement for unlimited proliferation capacity (immortality), i.e., a large number of divisions (not the 5-6 of the standard clonogenic assay), i.e., more than the number done by the amplifying transit cells (e.g., 7 for erythroblasts). The other components of the definition are to be able to make more of self (more stem cells), differentiate, and forCSCs, to also be tumorigenic.

thanks so much Dr Lange for your reply

Read this:

"Expression of CD44 variant isoforms in peripheral blood leukocytes in malignant lymphoma and leukemia: Inverse correlation between expression and tumor progression", by Sophia Khaldoyanidi, Martin Achtnich, Rüdiger Hehlmann, Margot Zöller. Leukemia Research, Volume 20, Issue 10, October 1996, Pages 839–851

In leukemia and in lymphoma is intresting to know how changes the expression of CD44,for the medical therapy of the cancer.

Over 20 years ago CD44 splice variants came into the focus of molecular biologists. Meanwhile, perhaps thousands of PhD students and postdocs wasted their carreers on such splice variants wasting many millions of funding. At that time, probably anything could be published but not much appears to be useful today for medicine. Now, a few generations of postdocs later, many aspects seem to be re-evaluated with the hope of more than a correlation to cancer stem cells. If CD44 is important in any way, then it seems to be too complex so far - or just a dead end.