Salivary AciCC and the CDKN2A/B Locus

In reviewing the literature we asked the question, “What mutations that drive salivary acinic cell carcinoma (AciCC) might cause it to resemble MASC.” Shumei Kato, of MD Anderson Cancer Center Houston, led a team that also included the Center of Personalized Cancer Therapy at the Moores Caner Center in La Jolla, and the private DNA sequencing company N-of-One in Lexington, MA. These authors sequenced a total of 117 salivary gland tumors and found 354 aberrations, many occurring simultaneously in the same patient. In the AciCC tumors that they sequenced, alterations in the DDKN2A/B locus were a common theme, yellow highlight. In this particular study, none of the AciCC cases were really MASC. An aside: Contrary to expectations, two cases of adenocarcinoma, not otherwise specified, were unmasked to be MASC. Note that these two cases of ETV6-NTRK3 rearrangement were not associated with any other genetic alterations that could have been driving the cancer. One case of adenocarcinoma NOS is shown as an example of multiple potential driving mutations showing up in one tumor.

Kato 2015, gene modifications associated with acinic cell carcinoma (AciCC)

Loss of INK4/ARF Locus is a recurring theme.

Kato and coworkers stated that the most common abnormalities involved the

(1) TP53 gene (36/117, 30.8% ),

(2) cyclin pathway of CCND1, CDK4/6 or CDKN2A/B (31/117, 26.5%), and

(3)  PI3K pathway of PIK3CA, PIK3R1, PTEN or AKT1/3 (28/117, 23.9%).

Interestingly, five of the seven cases of AciCC showed loss of the CDKN2A /B genes.  The authors did not reveal whether surrounding genes and/or genes towards the telomer were also deleted. Perhaps a closer look at this locus may shed some light on why AciCC and MASC might resemble one another. The Ink4/ARF locus resides on p21.3 band of chromosome 9.

Fig 1. The INK4/ARF locus A. Chromosome location, these genes are right on top of each other B. Splicing C. Function of Ink4 and ARF proteins. An overview, continue for greater detail


The CDKN2B gene codes for two splice variant proteins, p15Ink4B and p10, that appear to have distinct functions. A regulatory domain (RD) may control transcription from not only CDKN2B but also CDKN2A.  The cell division 6 protein, encoded by the CDC6 gene, appears to be a regulatory factor.

p15INK4b , along with its chromosome 9 relative p16INK4a, belong to a family of inhibitors of cyclin dependent kinases 4 and 6 (Cdk4/6).   

  • The Ink inhibitors complete with Cdks for their binding partners cyclin D effectors (Fig1, panel C).
  • Failure to phosphorylate the retinoblastoma protein (RB) translates into RB binding to and inactivating the E2F transcription factor.
  • When over expressed, p15Ink4B activates the RB pathway.  A stable senescent state is manifested by E2F transcription silencing and alterations in heterochromatin.

Fig. 2.  p15Ink4b functions in the same manner as p16Ink4a.  Both oppose receptor tyrosine kinases like TrkC.

  • E2F is required for the transcriptional activation of genes necessary for entry into the S phase of the cell cycle and DNA synthesis.

Fig. 3.  The Rb:E2F pathway. Kinase complexes Cyclin D: Cdk 4/6 and Cyclin E:Cdk 2 cause conformational changes to the Rb structure and release of E2F. The release of E2F is necessary for the expression of S-phases genes.  The R-point marks the commitment of a cell to divide.

  • The link between TrkC and Cyclin D has been demonstrated in the development of the murine cerebral cortex (Ohtsuka 2013).  These authors examined the timing of the TrkC ligand, neurotrophin 3 (NT3), on S phase newly synthesized DNA as measured by BrdU.  The changes were inhibited by the Mek inhibitor U0125.  Ohtusuka and coworkers also saw NT3 changes in expression of cyclin D1 and D2.

p10 of CDKN2B, a helper of an alternative reading frame (Arf) gene product

The p10 protein of CDKN2B received little comment from Flares and coworkers, except what is shown in Panel B of Figure 1. This splice variant was first reported by Tsubari and coworkers in 1997. Some interesting things they reported are as follows.

  • In contrast to p15Ink4b, p10 did not immunoprecipitate with Cdl4 and Cdk6 when transiently expressed in COS-7 cells.
  • The over expression of pl0, together with p15Ink4b in COS-7 cells, did not interfere with the complex formation of p15Ink with Cdk4 or Cdk6.
  • A somewhat later study (Pérez de Castro 2004) demonstrated p10 cell cycle arrest in a p53 dependent manner in a mouse model.


The CDKN2A gene codes for multiple, alternatively spliced transcripts. The major ones, p16Ink4a and p14Arf, differ in their first exon. Robertson and Jones (1999) also found a third splice variant of this gene, p12.

Two different views of CDKN2A splicing. Top,  Bottom, Robertson and Jones 1999

  • The expression of p12 seems to be restricted to the pancreas.
  • Robertson and Jones found p14Arf in cDNA libraries of the human heart, liver, and kidney but not the pancreas, brain and skeletal muscle.
  • Robertson and Jones found p16INK4a in heart, liver, lung, and the pancreas.
  • None of the three transcripts were found in human brain and skeletal muscle.


Peggy Ozenne and colleagues (2010) at Institut Albert Bonniot in Grenoble, France  reviewed this unique protein and its  role in cancer.

  • In a way, p14Arf could be considered a functional protein product of a frame shift mutation.
  • p14ARF neutralizes two ubiquitin ligases: mdm2 and ARF-BP1 (coded for by the HUWE1 gene).
  • Both ubiquitin ligases target p53 for degradation.
  • p14ARF inhibits the transcription factors Myc and E2F1 independently of p53
  • Exon 1b of  is responsible for nucleolar localization.  It is also reported in the Ozenne review to bind to mdm2.
  • A second nucleolar localization sequence is reported in exon 2, also the site responsible for sumolylation.
  • As a separate function, p14ARF reduces rRNA transcription as well as export of the ribosome from the nucleus.

Failure to inhibit MDM2 will result in degrdation of tumor suppressor p53 via the the ubiquentin degradation pathway. Over expression of S100B in MASC could hypothetically have the same affect.

A short note on promoter methylation

Robertson and Jones (1998) sequenced the promoters of p14Arf and p16INK4a to determine their regulation of methylation, and the feedback control of p14Arf via p53 expression. Al-Kaabi cite references demonstrating hypermethylation of the p16Ink promoter as being anti-oncogenic whereas hypermethylation of the p14Arfpromoter being pro-oncogenic. Ozenne and colleagues (2010) also summarized many cancers in which the ARF promoter is methylated.

In summary, the loss CDKN2A and/or CDKN2A/B in five of the seven AciCC tumors sequenced by Kato and coworkers (2015) strongly suggests overlapping pathways with MASC.

  1. Tel-TrkC activates the cyclins.  Loss of p15Inkb and p16Inka would result in failure to inhibit cyclins 4/6
  2. Loss of p14ARF’s protection of p53 from degradation might be felt as the same way as MASC protein S100b binding to p53.

Important Information

The ETV6-NTRK3 fusion of MASC is only one of many cancer driving NTRK gene fusions. There is an open salivary gland cancer trial testing a specific inhibitor of the Trk family of kinases.  This particular small molecule inhibitor is effective against ROS/Alk/Trk kinase fusion proteins.  A screening test for this trial utilizes IHC/NGS.


Al-Kaabi A, van Bockel LW, Pothen AJ, Willems SM. (2014) p16INK4A and p14ARF gene promoter hypermethylation as prognostic biomarker in oral and oropharyngeal squamous cell carcinoma: a review. Dis Markers. 2014;2014:260549. PubMed

Fares J, Wolff L, Bies J (2011) CDKN2B cyclin-dependent kinase inhibitor 2B p15, inhibits CDK4 Atlas Genet Cytogenet Oncol Haematol. 2011;15(8):652-657.  free link

Kato S, Elkin SK, Schwaederle M, Tomson BN, Helsten T, Carter JL, Kurzrock R. (2015) Genomic landscape of salivary gland tumors. Oncotarget.6(28):25631-45. PubMed

Ohtsuka M, Soumiya H, Hanai M, Furukawa S, Fukumitsu H. (2013) Neurotrophin-3 influences the number and the laminar fate of cortical progenitors in the developing cerebral cortex of mice through the MEK/ERK1/2 signaling pathway.  Biomed Res.34(5):231-9.  PubMed

Pérez de Castro I, Benet M, Jiménez M, Alzabin S, Malumbres M, Pellicer A. (2005) Mouse p10, an alternative spliced form of p15INK4b, inhibits cell cycle progression and malignant transformation. Cancer Res. 65(8):3249-56. PubMed

Robertson KD, Jones PA. (1998) The human ARF cell cycle regulatory gene promoter is a CpG island which can be silenced by DNA methylation and down-regulated by wild-type p53. Mol Cell Biol.18(11):6457-73. PubMed

Robertson KD, Jones PA. (1999) Tissue-specific alternative splicing in the human INK4a/ARF cell cycle regulatory locus.Oncogene. 18(26):3810-20. PubMed

Tsubari M, Tiihonen E, Laiho M. (1997) Cloning and characterization of p10, an alternatively spliced form of p15 cyclin-dependent kinase inhibitor. Cancer Res. 57(14):2966-73. PubMed

Zhang Y, Xiong Y, Yarbrough WG (1998). “ARF promotes MDM2 degradation and stabilizes p53: ARF-INK4a locus deletion impairs both the Rb and p53 tumor suppression pathways”. Cell. 92 (6): 725–34. PubMed

Zhang W, Zhu J, Bai J, Jiang H, Liu F, Liu A, Liu P, Ji G, Guan R, Sun D, Ji W, Yu Y, Jin Y, Meng X, Fu S.(2010) Comparison of the inhibitory effects of three transcriptional variants of CDKN2A in human lung cancer cell line. A549. J Exp Clin Cancer Res. 29:74. PubMed

Leave a Reply