Carcinoma of the Breast

The ETV6-NTRK3 gene rearrangement was first characterized as a genetic event in secretory breast carcinoma in the laboratory of Poul Sorensen. Christina Tognon and Stevan Knezevich were first authors of this report.

History behind the insightful discovery

Four years before examining a tumor from a six year old female with breast cancer, the Sorensen Laboratory sequenced the break point of  the gene fusion,  t(12;15)(p13;q25), associated with congenital fibrosarcoma (CFS).  T(12;15)(p13;q25) is short hand for a break in Chromosome 12 at band p13 joined to Chromosome 15 at a break at band q25.  Figure 1B says it very well. Our chromosomes contain fragile sites that are prone to breaking. This gene fusion had been previously  shown to occur in the following  cancers

  • CFS
  • pediatric mesoblastic nephroma
  • adult myeloid leukemia

The first two cancers  are of mesenchymal origin, the third of  hematopoietic origin. Finding this gene rearrangement in a new cancer was alone an accomplishment. What makes this publication truly comprehensive is the work with the  protein product of this ETV6-NTRK3 gene fusion: a chimera of the kinase domain of the neurotrophin-receptor tyrosine kinase (TrkC) and the dimerization domain of the TEL transcription factor.

A six year old girl with invasive edenocarcinoma breast cancer.

This tumor had eosinophilic PAS+ secretory features (Figure 1A). Karyotyping revealed t(12;15)(p13;q25) translocation as the only cytogenetic abnormality in the tumor cells (Figure 1B). The authors’ prior work left them with  finely tuned molecular methods to detect specific chrome 12 breaks (Figure 1C). Cosmid FISH probes were directed against  exons in the ETV6 gene on chromosome 12 (Figure 1D). Arrows indicate regions of ETV6 gene disruption (Figure 1 C).

IHC image of tumor with eosinophilic PAS+ secretory features (Fig. A). Fig B, shows Karyotyping revealed translocation as the only cytogenetic abnormality in the tumor cells. Detection of specific chromosomee 12 breaks, arrows indicate regions of ETV6 gene disruption (Figure 1C). Fig. D shows cosmid FISH probes were directed against exons in the ETV6 gene on chromosome 12.

Figure 1. An ETV6 rearrangement in pediatric breast cancer A. An H&E stained section with PAS+ secretions B. Karyotyping reveals a chromosome 12 – 15 rearrangement C. FISH probes indicate disruption of the ETV6 gene D. Intronic regions of the ETV6 FISH probes, relative to the TEL transcription factor domains.

Is the ETV6-NTRK3 fusion in other secretory breast carcinomas?

The authors initiated their inquiry by asking if the ETV6-NTRK3 fusion was common to a third pediatric cancer. If the ETV6-NTRK3 fusion can occur in adult myeloid leukemia, can it occur in adult secretory breast cancer?

Table showing 18 patients of various ages. Out of 13 secretory carcinoma cases, 12 are positive for the ETV6-NTRK3 fusion as indicated by rtPCR and break apart FISH. Of the five ductal carcinomas, four are negative for the ETV6-NTRK3 fusion.

Table 1 from Tognon 2002

Of the 13 secretory carcinoma cases, 12 are positive for the ETV6-NTRK3 fusion as indicated by rtPCR and break apart FISH. Of the five ductal carcinomas, four  are negative for the ETV6-NTRK3 fusion.

Closer look at rtPCR and FISH

CFS was used as a positive control in rtPCR amplification of the region spanning the fusion junction (Figure 2A).

  • The total RNA from the six year old female with secretory breast cancer (SBC) gave a strong rtPCR signal.
  • The CFS positive control also gave a very strong rtPCR signal.
  • Lane 1  The sample from the 19 year old female gave a strong rtPCR signal and an unambiguous FISH signal.
  • Lane 2  The 75 year old male with gynecomastia with secretory features appears to be a weak positive in this rtPCR image but recorded as a negative in table 1.  Compare this “negative” to the negative in lane 4.  It should be noted that the FISH results for this sample are also not indicative. There may be have been a problem with the RNA quality.
  • Lane 3  The ductal carcinoma from the 63 year old female was weakly positive with rtPCR and positive for FISH.
  • Lane  4  The ductal carcinoma from the 16 year old male is negative for both rtPCR and FISH.
Fig. A, shows results of PCR directed against the ETV6-NTRK3 fusion junction. Fig. B, shows breakpoint ETV6-NTRK3 FISH probes.

Figure 2 Representative results of A. PCR directed against the ETV6-NTRK3 fusion junction B. ETV6-NTRK3 FISH probes.

Yeast artificial chromosome FISH probes were used to detect gene fusions in metaphase cells (Figure 2B). In metaphase, chromosomes separate. Separate color dots indicate separated chromosomes 12p13 (green) and 15q25 (red).  The two colors side by side  or yellow indicate a inter-chromosomal rearrangement.  The authors presented CFS as an example of this technique. The same FISH technique was used to generate the breast cancer results  presented in Table 1.

Is the “EN” protein product of ETV6-NTRK3 an active kinase?

The next question that the authors addressed was whether the “EN” protein product of the ETV6-NTRK3 gene fusion was an active kinase. They examined three tumors

  1. 3T3 fibroblasts transfected with the ETV6-NTRK3 gene such that they expressed the EN protein
  2. congenital fibrosarcoma tumors, a positive control
  3. invasive breast cancer tumors, a negative control

The first major step (panel 3A) was separating the EN protein from all of the other proteins in the tissue or cells. This was done by way of an antibody against TrkC (bright red circles). Beads with an antibody binding protein are used to retrieve the antibody-antigen complex. Anything else bound to EN will come down with the beads when they are “spun down.”

The second major step (Panel 3B) was to elute proteins from the beads and coat them with a detergent called SDS. This breaks up protein complexes and gives the proteins a negative charge. The gel is a polymer matrix that sieves proteins according to size. All of the proteins are negatively charged and move towards the anode when an electrical potential difference is applied across the gel.

The gel is removed from the apparatus and subsequently transferred to a membrane. Skipping a few minor steps, the membrane is placed in a solution containing rabbit antibodies against phosphotyrosine. These antibodies will recognize “autophosphorylation” sites on TrkC and EN as well as any down stream signaling molecule. The rabbit antibodies bound to EN were detected by mouse antibodies against rabbit antibodies. The mouse antibodies were conjugated with an enzyme called horse radish peroxidase. When the ECL reagent is added, HRP produces light, which is recorded on film.

Fig. A cartoon shows protocol to detect kinase activated EN protein. Fig. C shows activation of EN via autophosphorylation.

Figure 3. A common protocol to detect kinase activated EN A. EN is removed from homogenates via immunoprecipitation B. Proteins in this complex are separated by size and reacted with phosphotyrosine antibodies. C. results showing activation of EN via autophosphorylation.


The results (Panel 3C). In order to not confuse us too much, the authors are only showing us the EN molecule. A common lore to explain the double bands is conformational changes in the kinase that heating in SDS will not destroy. Note that nothing at the position of EN is phosphorylated in invasive breast cancer or the 3T3 fibroblasts transfected with the empty vector.

Expression of “EN” results in a cancer phenotype

In addition to tranfecting 3T3 fibroblasts with the ETV6-NTRK3 gene the authors also transfected two breast epithelial cell lines

The Scg6 cell line has stable myoepithelial (more mesenchymal) features.

The Eph4 cells were claimed to show a stable epithelial phenotype in culture.

Fig. A, B and C show how EN transfected cell line can express EN and these cells line when injected in nude mice can lead to tumors.

Figure 4 “EN” can transform cells A. ETV6-NTRK3 transfected cells express tyrosine (auto) phosphorylated “EN” B. When these cell lines are injected in nude mice, they grow tumors. C. A plot of estimated tumor growth in grams as a function of time in weeks

Note that all transfected cell lines contain presumably  active, autophosphorylated EN fusion kinase. It is also assumed that the phosphotyrosine antibodies reacted with autophosphorylation sies of  EN. These assumptions are supported by the ability of these cell lines to grow as tumors in nude mice (Panels 4B, C).

The mesenchymal vs epithelial inquiry  continues

The ETV6-NTRK3 fusion drives CFS, a mesenchymal tumor. The authors asked if epithelial cells expressing the EN protein would also express  mesenchymal or epithelial proteins. The intermediate filament cytokeratin was chosen as a typical epithelial protein whereas vimentin was chosen as a mesenchymal marker (Figure 5).

IHC images showing epithelial cells expressing EN protein could also express  mesenchymal or epithelial proteins.

Figure 5. Immunohistochemistry with epithelial (cytokeratin) and mesenchymal (vimentin) markers

Arrows mark glandular structures in the Eph4 cells. No arrows are needed for the epithelial marker cytokeratin.  Note that no hint of vimentin staining is seen in glandular structures of the En-Eph4 tumor. The arrow in the TrkC immunohistochemistry stain of the EN-Eph4 tumor indicates both glandular and mesenchymal features.

Tognon and coworkers indicated that the Eph6 cells had myoepithelial like features whereas the Eph4 cells were more epithelial.  Since EN expression is driven by the ETV6 promoter, is TEL normally expressed in both cell types in the breast?  The short answer is “yes” to both cell types.

Since EN expression is driven by the ETV6 promoter, TEL is expressed in the normal breast tissue.

Figure 6. TEL expression in normal breast tissue from

The importance of getting tested

While the ETV6 break apart FISH probe was to become a popular aid in the diagnosis of the mammary analog of secretory carcinoma (MASC) of the breast, this early study indicated that it is not always informative.

The positive PCR and FISH results in the ductal carcinoma of case 14 received further mention of Tognon and coworkers. Closer examination revealed glandular features and eosinophilic secretions. Speculation was that case 14 underwent clonal evolution to assume a more differentiated phenotype. In the past 15 years Trk tests for NTRK gene fusions have become far more sophisticated. Another reason to get tested in an open clinical trial for a Trk Treatment. SBC treatment options for patients is also available. An informative video on the Trk inhibitor entrectinib is also available.