Molecular comparison of breast and salivary gland secretory carcinomas

Comparing Secretory Carcinoma of the breast with MASC

Gregor Krings led a multi-center group from the University of California, San Francisco, to compare histology and genetics of nine secretory carcinomas of the breast with six MASC salivary gland tumors. For this post secretory carcinoma of the salivary glands will be referred to as MASC, mammary analog of secretory carcinoma of the breast.   Secretory carcinoma of the breast will simply be referred to as SC.

The genetic profile of the tumors in one  image

Diagram showing genomic alterations in various patients with their sex and age.

Figure 1, adapted from Figure 3 of Krings 2017

  • Note that males are well represented in both MASC and secretory carcinoma of the breast.
  • Young patients are also well represented.
  • The ETV6 chromosomal break, as detected by FISH, was found in all but one case of MASC and all cases of SC.
  • Genomic DNA sequencing revealed the ETV6-NTRK3 fusion in all but one of the tumors. This tumor had a complex fusion that required mRNA sequencing to identify. Reciprocal fusions are simply ETV6-NTRK3 and NTRK3-ETV6.
  • The deletions and duplications around the ETV6 and NTRK3 genes have the potential to affect tumor morphology and behavior.
  • Gains and losses have been associated with cancer, as discussed on companion posts.
  • At least in this study, ETV6-NTRK3 is the sole driver of  MASC and SC. The authors discussed whether their panel of 510 cancer genes was large enough.
  • Single nucleotide variations (SNV)  generally do not affect protein function. Some will be presented on this post

ETV6-NTRK3 Companion genomic alterations discussed on companion posts

Other genomic alterations may contribute to the cancers driven by ETV6-NTRK3 gene fusions. Speculations on what the implications are based on databases such as Cancer Index and the peer reviewed literature are presented on two companion posts.

Secretory Carcinoma of the Breast  A discussion of a complex fusion and chromosomal gains and deletions

Secretory Carcinoma of the Salivary Glands  A discussion of two complex fusions and chromosomal gains and deletions

The above discussions are only meant to generate hypotheses as to why the ETV6-NTRK3 driven secretory cancer tumors sometimes look different.  The short observation is that these “drivers” may have “vocal passengers.”

Histology, a comparison

The following is  histology presented by Krings and coworkers with some reminders of the genetic background.  Scroll through this section, note the histological differences, and then read a few words on genetic differences.

SC, secretory carcinoma of the breast

SC1 came from a 57 year old female. Except for a gain in the q arm of chromosome 8, her tumor is genetically unremarkable. Submitted for consideration is a single SNP gene IRS2, protein S723R. This particular substitution in the insulin receptor substrate occurs between YXXM motif 3 and 4.

Immunohistology slides showing secretory carcinoma of patient SC1 (slide a and b) showing characteristic microcystic growth pattern and foci of tubular infiltration and associated eosinophilic luminal secretions.

Figure 1 a-b, Krings 2017

SC2 came from a 71 year old female who had no evidence of disease 98 months later. One interesting feature of this tumor is the deletion of the interstitial region of chromosome 15q. The gain in chromosome 16 is discussed in the secretory carcinoma of the breast companion post. Compare the vacuoles with SC3.

SC3 came from a 23 year old male. This male lacked evidence of disease after 97 months. The fusion gene chromosome was duplicated in this tumor. There was a gain in distal 12p and a loss of chromosome 15. Book keeping in reciprocal gene fusions is inherently problematic.

The authors examined the SOX10 transcription factor expression. This tumor had a nonsense mutation in the SOX9 transcription factor that resulted in premature termination of translation.

gene SOX9,  protein Q175* While a good third of this gene never gets translated, SOX9 is generally considered a transcription factor in skeletal development.

Immunohistology slides showing secretory carcinoma of patient SC2 (slide c) showing prominant cytoplasmic vacuolization and pseudolactational changes. Slide d of patient SC3 shows microcystic change merging with solid growth.

Figure 1 c-d, Krings 2017

SC4 and SC5 came from females on the younger side. Note that the former has eosinophilic secretions (deep pink), and the latter does not.

Some gains in chromosomes other than 12 and 15 were seen in SC4.

gene SMARCA4,  protein A1448V Transcription activator BRG1 is a largely neuronal transcription factor. This is a conservative substitution. A  V → A substitution at position 1484 had no effect on binding to acetylated lysine on histone H3.

SC5 is not only rather uninteresting genetically but it is also relatively benign. “In situ” in this case refers to ductal carcinoma in situ (DCIS). The cancerous cells are found in the milk ducts, in the expected location. The cells are not considered cancerous until they invade other locations. The SC5 patient had no evidence of disease  after 72 months. Is something more than an ETV6-NTRK3 gene fusion is needed to drive cancer?

Immunohistology slides showing secretory carcinoma of patient SC4 (slide e) showing prominant eosinophilic granules, rounded nucleilarge nucleoli, prominant microcystic. Slide f of patient SC5 shows well developed papillary features and superimposed microcystic growth.

Figure 1e-f, Krings 2017

Tumor SC6 came from a patient who died from metastatic cervical cancer 46 months later with no evidence of SC (of the breast) recurrence. This secretory  cancer was otherwise genetically unremarkable. Some single nucleotide variations resulting in amino acid changes of unknown significance were given

gene CHD4,  protein Q568R,  This chromatin helicase DNA binding protein participates in histone deacetylation.

gene MYH9,  protein E530K  Myosin 9 is one of many myosins that control cell shape in non muscle cells.

The authors did not indicate if the cervical cancer was genetically identical to the secretory cancer.

Immunohistology slides showing secretory carcinoma of patient SC6 (slide g) showing solid microcystic growth. Slide h of patient SC7 mimic fenestrated growth florid usual ductal hyperplasia. Slied i of patient SC9 shows intermediate grade nuclei and eosinophilic luminal secretions.

Figure 1g-i, Krings 2017

SC7 is from a 50 year old female with no evidence of disease after 79 months. No chromosomal gains or deletions were noted with this tumor. Fenestrated growth was noted. The nuclei seem to be more prominent and plentiful. A single nucleotide variation of probably no consequence was noted.

gene RASA1,  protein N577S  Ras GTPase-activating protein 1 stimulates the GTPase of normal but not oncogenic Ras p21. This particular amino acid substitution is fairly conservative.

SC8 is genetically interesting because it was an ETV6-PLEKHA5–NTRK3 fusion and came from a 13 year old female.  Both the ETV6 and the PLEKHA5 genes are located on the p arm of chromosome 12. At 0.5 cm, this tumor was the smallest in the group. Very little IHC was performed. It is uncertain that this cancer was even remotely aggressive.

SC9 came from a 53 year old female whose tumor was only 0.7 cm in diameter. She was lost to follow up. The luminal secretions seemed to be more than the other secretory carcinomas of the breast. Is there some unknown passenger mutation?  Is this phenotype due to some difference in the hormonal environment?

MASC, secretory carcinoma of the salivary glands

MASC1, from a 57 year old female, has some chromosomal gains and deletions that may be associated with cancer that are covered on other posts. The extensive genome sequencing uncovered a small single nucleotide variation that is probably inconsequential.

gene SETD2, protein I1974V This isoleucine to valine substitution in histone-lysine N-methyltransferase is extremely conservative and probably does very little if anything to impact function.

MASC2 is unusual in that it came from a 59 year old male. There were gains in the  chromosome 12 (distal) and 15 (proximal). The ETV6-NTRK3 fusion chromosome is duplicated. The authors claim SC (of the breast) like nuclear features.

Immunohistology slides showing MASC1 and MASC2, nuclear features similar to breast SC.

Figure 1j-k, Krings 2017

MASC3, from a 23 year old male, was the largest in the group at 2.5 cm in diameter. This male was disease free after 97 months.

gene TCF7L2,  protein S301P, This transcription factor 7 like 2 participates in the Wnt pathway and binds to the MYC promoter.

gene BRD4, protein P955fs, This acetylated histone binding bromodomain protein only has a C-terminal domain region taken out by a frame shift. This region is missing in isoform 3.

MASC4 is particular interesting in that it involves a complex fusion with ETV6-HERC2-NTRK3. HERC2 and NTRK3 are located on opposite ends of the long (q arm) of chromosome 15. Krings and coworkers commented on the resemblance of MASC4 to SC2, a tumor in which the interstitial region of 15q has been deleted.

Both HERC2 and NTRK3 reside on the anti-sense strand.  It wasn’t clear how much of the HERC2 gene product, if any, makes its way into the TEL-TRKC fusion protein. For these to genes to become fused, a large portion of  the interstitial part of 15q must be removed. Krings only reported deletion of the proximal region of 15q.

Finally, MASC4 contains the only substitution in a well recognized tumor suppressor, tuberin.

gene TSC2,  protein S838N  There are many natural variants in this general region of the protein.

Immunohistology slides showing MASC3 and MASC4, features similar to breast SC.

Figure 1l-m, Krings 2017

MASC5 is another complex fusion, this time with WDR53. The WDR53 gene is located on the very distal q arm of chromosome 3, 3q29 to be somewhat exact. Some genes in the neighborhood are covered in the Secretory Carcinoma of the Salivary Glands  companion post. For this post, note how less dense the nuclei stain in MASC5 versus MASC6.

Immunohistology slides showing MASC5 and MASC6, features similar to breast SC.

Figure 1n-o, Krings 2017

MASC6 was was one of many tumors found to have  a reciprocal ETV6-NTRK gene fusion. The single amino acid substitution in the intestinal maltase-glucoamylase.

gene MGA,  protein A1529V This is a highly conservative substitution, in close proximity to the proton donor of the active site, D1526.

IHC comparison

This section is very short.  Compare the IHC markers in representative MASC and SC tumors that the authors chose.  Then compare a very broad selection of IHC markers in table 2.

Basal cell markers cytokeratin 5/6 and epidermal growth factor receptor (EGFR) were compared in SC2 and MASC3

Figure 2a-f, Krings 2017

MUC4 and SOX10 are simply MASC markers that the authors found in their secretory carcinoma of the breast counterpart.

Figure 2g-l, Krings 2017

All of these cancers are negative for the estrogen receptor, progesterone receptor, and HER2, i.e. they are triple negative.  S100 is a family of calcium binding proteins. A Dako polyclonal antibody that is/was selective for the S100B isoform has  long history in MASC immunohistochemistry.

Table 2, histology results from Krings 2017


 Testing for TRK

These data were presented to make two points

  1. Histological differences exist between and within these two groups of secretory carcinoma. Genomic alterations traveling with the ETV6-NTRK gene fusion may or may not explain these differences.
  2. IHC markers are helpful but not definitive.

As part of the Trailblaze® molecular diagnostics program, Ignyta is now offering Trailblaze Pharos™, a suite of molecular diagnostic assays and services, including proprietary companion diagnostic tests, to accompany its molecularly targeted oncology programs.

Trailblaze Pharos™ is a multiplex assay for identifying actionable fusions in NTRK1, NTRK2, NTRK3, ROS1, and ALK genes, resulting in solid tumors that can be treated with entrectinib, to support testing for the STARTRK-2 clinical trial.


Krings G, Joseph NM, Bean GR, Solomon D, Onodera C, Talevich E1, Yeh I, Grenert JP, Hosfield E, Crawford ED, Jordan RC, van Zante A, Zaloudek C, Shin SJ, Chen YY. (2017) Genomic profiling of breast secretory carcinomas reveals distinct genetics from other breast cancers and similarity to mammary analog secretory carcinomas. Mod Pathol. 30(8):1086-1099. PubMed