Currently we are using E-cadherin and beta-cantenin expression for epithelial and N-cadherin, vimentin, fibronectin expression showing the mesenchymal transformation.

You can identified these EMT cancer cells by performing of immunohistochemistry or immunofluorescence of EMT markers (Snail, Smooth muscle actin, vimentin, N-Cadherin).

Currently we are using E-cadherin and beta-cantenin expression for epithelial and N-cadherin, vimentin, fibronectin expression showing the mesenchymal transformation.

No markers. Patient samples! What antigens or other markers would you suggest?

IHC for smooth muscle actin is the most reliable EMT (myofibroblast) marker for formalin-fixed paraffin-embedded (FFPE) tissues. Other markers mentioned by Bilal and Godfrey are tricky in terms of antigen retrieval.

α-smooth muscle actin (α-SMA)) marker are the most reliable EMT marker for IHC.

Depends what your samples are: fixed & paraffin embedded, fresh frozen excision biopsies, FNA biopsies? I would design my strategy around the type of samples.

In principal you test for an EMT-signature by either IHC or gene-expression profiling (either targeted -qRT-PCR- or global approach -microarray or RNA-seq-). Important in your case is that you also need a/or a set of markers that distinguish your tumor cells from normal cells and they can be detected with the same methods.

Christer, you ask how to distinguish EMT cancer cells from authentic mesenchymal cells. The purpose of IHC is not only to stain cells based on characteristic markers, but also to view their morphology. In this regard, the choice of counterstain is also important. Epithelial cells are always arranged in a normal histological architecture. EMT cancer cells show pathological abnormalities. Combined morphology and marker staining is the only way to be certain.

This is why IHC grading by an experienced pathologist is needed. Gene-expression profiling may give false positives if a biopsy contains too much mesenchymal/adipose normal tissue, but this is way better than a false negative. Gene expression profiling gives much more information, and doesn't depend on human judgement - no matter how experienced the pathologist is. Now that molecular profiling has become commonplace, there is no excuse not to use it in combination with IHC.

I am following all the information. I agree with Gerald that antigen retrieval is tricky. Although Gene expression profiles are important, they are still difficult to implement in the clinic in our institute. Following microdissection of FFPE tissue, our lab is profiling a set of genes including EMT. What is the best set of genes to check with qPCR. Here again RNa quality is an issue for FFPE material.

In colorectal cancer samples an experienced pathologist should be able to help you to identify areas of tumour budding at the invasive front, which is likely a morphological correlate of EMT. The small tumour nests are still recognisably tumour cells, but show nuclear beta-catenin translocation, loss of E-cadherin and upregulation of vimentin by IHC. However broad spectrum cytokeratin expression may be retained. Looking at molecular markers alone risks assessing the stromal/inflammatory cell population.

Yes friends, as per our experience, first resort to dual immunostaining for pan-cytokeratin and vimentin using contrasting chromogens and the cells in transition, would express both the markers simultaneously. Secondly, look at the loss of of E-cadherin and gain of N-cadherin expession in cells undergoing EMT. Thirdly, look at the expression of snail slug and twist that mediate EMT. I hope that manages to resolve the problem to a great extent.

Certainly the current markers for EMT are enough. IHC for ECadherin, Vimentin, and depending on the tissue of origin, specific Keratins (K5, K8, etc). Also there are some reliable antibodies for EMT specific transcription factors (Snail, Twist). Also, if available, you can monitor by RTqPCR the expression of these and other markers. In general with appropriate primer design you can even analyze them in FFPE samples if fresh specimens are not.

Actually, you cant except for genomic markers for transformation. That is all the beauty of the EMT concept... Classic markers mentioned above are indeed useful to characterize cell phenotype and local dedifferentiation process, which maybe a clinically relevant phenotype (progression, invasion) . EMT master-genes are more tricky: few reliable antibodies, frequent co-expression with e-cadherin and cytokeratins in tumor cells and frequent stromal expression in stroma fibroblasts (that may be of tumor origin or not ). Their expression means something is happening, but is not at all meaning a full EMT with ck down and vimentin up (like in transformed cells in vitro). My advice is to focus on functional markers (proliferation, migration, differentiation, stemness) to evaluate tumor cell fate.

Loss of epithelial markers as suggested by Jesus Paramio and upregulation of vimentins and TWIST, SNAIL by qPCR.

I agree with David regarding colorectal cancer cells in EMT. By IHC, beta-catenin expression can be reduced or lost or translocated into the nucleus. E-Cadherin can be reduced or lost. Keratin 8 is still strongly expressed also in the budding area, but Vimentin, usually negative in normal colon epithelial cells, is expressed in EMT cancer cells.

In reviewing all the helpful suggestions made by contributors, a few things come to mind. Based on the wording of Christer's question, any strategy to identify EMT tumour cells in patient samples will depend on the quality of the biopsies, which may not capture all the architectural features of the growing tumour unless the tissue is from debulking surgical resection. The strategy will also depend on the type of tumour, because primary tumours from different tissues have unique features, as do metastatic nodes lodged in secondary sites. The choice of methods used for analysis will depend on the capabilities of the lab, and some of the methods suggested by contributors are not routinely performed in diagnostic clinical pathology labs. Also, some of the methods work well for in vitro studies, but are difficult to apply to clinical samples. There is a difference between what is possible and what is practical.

What about in hematological cancers? Is there a particular method for identifying the EMT cancer cells?

I am also interested in looking at this in human breast tumor tissues. Does anybody know of good human Snail and Twist antibodies that work well for IHC on FFPE samples? Thank you!

Dear Dr. Ericsson, I have been working with something similar and I have found decreased CDH1 to be a good reference for the loss of Epithelial Characteristics and an increase in CDH2, Vimentin plus a spindle shaped morphology to know that the cells have mesenchymal characteristics. Moreover, I have also found the expression of Snail and Slug to be highly increased in the cells showing the mesenchymal characteristics since they control CDH1 gene. But in all of my experiments, I have found Twist to be downregulated. So, I would suggest that a decrease in CHD1, Increase in CHD2, Vimentin and an increase in any of these transcription factors (Snail, Slug and/or Twist) followed by an elongated/spindle morphology should be a good observation to conclude that your cancer cells have transitioned from Epithelial to Mesenchymal state.