Re: “Risk Stratification Using a Novel Genetic Classifier Including PLEKHS1 Promoter Mutations for Differentiated Thyroid Cancer with Distant Metastasis” by Jung et al.

Dear Editor:

We read with interest the article by Jung et al. (1), assessing the role of PLEKHS1 promoter mutations as genetic markers of aggressive differentiated thyroid cancers (DTCs). In their study, the authors found that noncoding mutations at PLEKHS1 hotspots chr10:115,511,590 and 115,511,593 occurred in 6 of 61 (10%) of metastatic DTCs, all of them papillary thyroid cancers (PTCs), but only in 1 of 75 (1%) of their nonmetastatic counterparts. Strikingly, the median variant allele frequency (VAF) for PLEKHS1 mutations among PTCs was only 4%, which is suggestive of a late tumor event in cancer evolution, but also counterintuitive with the idea of tumors selecting genetic alterations that confer them a fitness advantage (i.e., “driver” or “oncogenic” events) versus neutral lesions (“passengers”). Another recent publication did not find PLEKHS1 mutations in the more aggressive anaplastic thyroid cancers (ATCs; 0/18), but showed that high PLEKHS1 expression tracked with metastasis and shorter survival in PTC patients, and that exogenous overexpression of PLEKHS1 in thyroid cells enhances AKT signaling and proliferation (2).

PLEKHS1 noncoding mutations were identified at low frequencies in thyroid and other tumor types in the early cancer whole genome sequencing efforts (3), and recently confirmed by the Pan Cancer Analysis of Whole Genomes (4). However, these studies doubted the oncogenic status of PLEKHS1 mutations because (a) they occur in palindromic sequence contexts believed to form hairpin DNA structures targeted by APOBEC enzymes (associated with passenger events) and (b) there is no reproducible association with PLEKHS1 expression levels, since different mutant cancer cohorts show either slight downregulation or no significant changes.

We surveyed our set of 30 thyroid cancer cell lines to assess whether PLEKHS1 mutations are subjected to positive selection, as previously shown for key thyroid cancer lesions (e.g., BRAF, TERT, TP53) (5) and correlated the presence of PLEKHS1 alterations with our existing expression data. We found that two cell lines harbored clonal heterozygous PLEKHS1 promoter mutations (Fig. 1A): THJ16-T, an ATC-derived cell line, had a PLEKHS1 115,511,593C>T change, whereas TT2609-CO2, derived from a follicular thyroid cancer (FTC) recurrence, showed mutations at both 115,511,590G>C (different than the reported G>A) and 115,511,593C>T, further stressing the idea of this locus as a mutation-prone region in the genome. Interestingly, three additional cell lines maintained subclonal PLEKHS1 mutations (VAFs = 3–15%, detected by digital droplet polymerase chain reaction). Overall, PLEKHS1 messenger RNA (mRNA) levels in cell lines were low and homogeneous, and none of the PLEKHS1-mutant samples were expression outliers (Fig. 1B), in line with the lack of correlation of PLEKHS1 mutations and mRNA expression in thyroid primary tumors (3,4).

OPEN IN VIEWER

FIG. 1.

PLEKHS1 status in thyroid cancer cell lines. (A) Sanger sequencing traces of the PLEKHS1 hotspot region for a wild-type control (top), THJ-16T (middle), and TT2609-CO2 (bottom) cells. (B) Violin plot showing relative levels of PLEKHS1 mRNA (average of 219857_at and 1554190_s_at probes) in thyroid cancer cells from Landa et al. (5). Cell lines with clonal and subclonal PLEKHS1 mutations are shown as red and green dots, respectively, whereas the blue line represents the median value. Clonality of PLEKHS1 mutations was determined by digital droplet polymerase chain reaction. mRNA, messenger RNA. Color images are available online.

Taken together, we show that PLEKHS1 promoter mutations are also present in a subset of FTC- and ATC-derived cell lines, but their consequences on PLEKHS1 gene expression remain elusive. We believe that the inherent difficulty of assigning functional effects to noncoding mutations requires a more thorough examination of the role of PLEKHS1 in thyroid cancer biology. The former will help determine to what extent these lesions and/or other surrogate markers (expression levels, DNA methylation, or others) become useful indicators in thyroid cancer prognosis.