TERT Promoter Mutation c.-124C>T Commonly Occurs in Low-Grade Fibromatosis-like Metaplastic Breast Carcinoma

Seven formalin-fixed, paraffin-embedded (FFPE) cases of FLMCs were available for molecular studies, which were performed in accordance with the guidelines and regulations of the ethics committee (ethics committee approval 31/09, Ordensklinikum Linz/Hospital of the Sisters of Charity, Linz, Austria). All cases revealed a uniform spindle cell morphology with irregular and infiltrating borders showing minimal to low-grade nuclear atypia, very low mitotic activity (1–3 mitotic figures per 10 high-power fields), a proliferative index below 5% by Ki-67 immunohistochemistry, areas with thick collagen fibers admixed with the spindle cells (Figure 1, A and B), and diffuse and intense immunoreaction for pancytokeratin (Figure 1, C) (ready to use; clone AE1/AE3, Agilent/DAKO). While all 7 examined cases of FLMC in our study showed diffuse and intense immunoreaction for pancytokeratin, one should acknowledge that not all cases of this entity display strong diffuse staining for CKs. The examined cases also exhibited either multifocal or diffuse positive immunoreaction of spindle cells for basal-type CKs such as CK5/6 (1:20; clone D5/6 B4, Agilent/DAKO) or CK5/14 (Figure 1, D) (ready to use; clone EP1601Y/LL002, Roche). All cases were completely immunonegative for ER (ready to use; clone SP1, Roche), PR (ready to use; clone 1E2, Roche), and HER2 (ready to use; clone 4B5, Roche). Furthermore, all cases were immunonegative for β-catenin (ready to use; Cell Marque), and none of the cases revealed abnormal immunoexpression for p53 (ready to use; clone DO-7, Agilent/DAKO).

After initial analysis of the FLMCs and the detection of TERT promoter mutations (see Results), we included 4 high-grade triple-negative and basal-like breast carcinomas with known TERT copy number gains for comparison for comparative molecular profiling. These 4 cases of high-grade TNBC (first control group; high-grade TNBC) were part of a separate and ongoing molecular-genetic study that revealed TERT copy number gains. They showed high-grade nuclear atypia, high mitotic activity (>15 mitoses per 10 high-power fields), and a high Ki-67 proliferation index (.50%) and were positive for basal-type CK (CK5/6 and/or CK5/14) but negative for ER, PR, and HER2. These 4 cases consisted of 2 with prominent spindle cell morphology (high-grade spindle cell metaplastic carcinoma), 1 of metaplastic carcinoma with prominent myxochondroid changes (“matrix-producing” metaplastic carcinoma), and 1 of carcinoma with medullary features.

After detection of TERT promoter mutations in FLMCs, we also decided to examine 7 additional rare cases of low-grade triple-negative breast carcinomas that did not show any fibromatosis-like or spindle cell morphology (second control group; low-grade TNBC). All these additional cases (second control group) represented very rare TNBCs that showed immunoexpression for CK5/14 (multifocal positivity), low-grade nuclear atypia, low mitotic activity (3 to 4 figures in 10 high-power fields), low Ki-67 proliferation index (between 5% and 10%), and, most importantly, lack of a fibromatosis-like or spindle cell histomorphology.

DNA extraction was performed with the Maxwell RSC DNA FFPE Kit according to manufacturer’s instructions (Promega, Madison, Wisconsin). DNA quantification was performed with the Qubit fluorometer (Thermo Fisher Scientific, Waltham, Massachusetts).

We used the OncoScan CNV assay (Thermo Fisher Scientific) for the genome-wide detection of copy number variations (CNVs) and the detection of loss of heterozygosity. The assay design is based on molecular inversion probes.¹¹  Five of 7 FLMCs and 3 high-grade MBCs for comparison were analyzed according to manufacturer’s instructions. Copy number analysis was not possible in the 2 remaining FLMCs owing to insufficient DNA quantity. In addition, we also performed CNV analysis in 7 cases of low-grade TNBCs with no fibromatosis-like or spindle cell morphology.

CNV data analysis was performed with the OncoScan console software version 1.3 (Thermo Fisher Scientific) using the OSCHIP-SNP-FASST2 algorithm. Data normalization based on normal diploid regions was verified with the Chromosome Analysis Suite (ChAS) software version 3.3 (Thermo Fisher Scientific). Percent genome change was calculated with the Nexus Express software version 3 (BioDiscovery, Hawthorne, California).

Targeted next-generation sequencing for 315 cancer-related genes (gene list provided in the supplemental digital content, containing 3 tables, at https://meridian.allenpress.com/aplm in the December 2023 table of contents) was performed by Foundation Medicine (Cambridge, Massachusetts), providing information on single-nucleotide variants, indels, CNVs, and gene fusions, as well as the tumor mutational burden and microsatellite instability in 7 FLMCs. The fusions examined may not be completely inclusive for tumors of soft tissue or myoepithelial origin.

Reverse transcription was performed with M-MLV Reverse Transcriptase, RNase H Minus, Point Mutant 200 U/μL by random hexamer priming (Promega). Relative quantification of the TERT expression was performed with a TaqMan gene expression assay (Hs00972650_m1) in triplicate reactions, with the PUM1 gene (Hs00472881_m1) as an endogenous control.¹² 

The TERT expression in the 7 FLMC samples was compared to that of normal (nonneoplastic) breast tissue (n = 7). Furthermore, TERT expression in FLMC samples bearing the c.-124C>T TERT promoter mutation was compared to that of 4 high-grade TNBCs with already known TERT copy number gains and unmutated TERT promoter sequences. We evaluated the relative expression differences by using the comparative Cq method, Welch analysis of variance, and Tukey-HSD/Tukey-Kramer post hoc analysis.¹³ 

Copy number analysis was performed for 5 of the 7 FLMC samples and mainly consisted of gains (Figure 2). Recurrent aberrations were detected in 3 of the 5 FLMC samples analyzed. Gains of 5p15.33-5p11, encompassing the TERT locus, were detected in 3 cases. Gains of 13q14.2 and trisomy 7 were detected in 3 and 2 cases, respectively. All abovementioned alterations were not exclusive for FLMCs; gains of 13q14.2 were also detected in 4 low-grade TNBCs (second control group), while 4 high-grade TNBCs showed TERT copy number gains, and 2 of these cases displayed additional trisomy 7.

Additionally, copy number analysis revealed low genomic instability with a mean genome change of 8.34% (95% CI, 2.15%–14.53%). CNV analysis of high-grade TNBCs, performed for comparison (first control group), revealed a mean genome change of 73.8% (95% CI, 64.04%–83.56%). Although the number of samples was small, the difference was highly significant (P < .001). CNV analysis in 7 cases of low-grade TNBCs (second control group) showed no significant difference to that of FLMCs, with a mean genome change of 5.96% (95% CI, 0%–12.24%) (Figure 3, A and B).

Targeted next-generation sequencing for 315 cancer-related genes was performed for 7 FLMCs and revealed a low mutational burden, with 2 to 3 mutations per megabase. An overview of the mutations detected is shown in Figure 2.

PIK3CA/PIK3R1 mutations leading to activation of the PI3K/AKT/mTOR pathway were found in all patients and are presumably the main oncogenic driver in these tumors. We detected a recurrent TERT promoter mutation c.-124C>T in 6 of 7 tumors (85%). Furthermore, loss-of-function of CDKN2A due to point mutations or deletion was observed in 2 cases each. In 1 case a CDKN1B mutation was detected. Microsatellite status was informative in 6 of the 7 samples and was stable in all 6 cases. None of the FLMCs displayed TP53 mutations.

TP53, however, was mutated in all high-grade TNBCs. No mutation in TP53 or TERT could be identified in low-grade TNBCs (second control group).

The expression of TERT in the 6 FLMC samples carrying the TERT promoter mutation c.-124C>T was compared to that of 4 high-grade TNBCs with known TERT copy number gain and 7 nonneoplastic breast tissue samples.

Gene expression analysis revealed significant TERT overexpression in samples with the c.-124C>T TERT promoter mutation (fold change: 62.17; 95% CI, 14.99–257.77; P = .02) and TERT copy number gain (fold change: 117.72; 95% CI, 16.44–842.89; P = .01) when compared to the nonneoplastic breast tissue samples. Gene expression of TERT promoter–mutated samples was equivalent to TERT copy number gain samples, as no significant difference was detected between the 2 cohorts (P = .33) (Figure 3, C through E).

The aim of the present study was the examination of molecular-genetic alterations (mutational profiling and CNV) of FLMC, a very rare subgroup within the spectrum of “triple-negative” and “metaplastic” carcinomas, morphologically closely resembling fibromatosis of the breast, which is currently regarded as a benign (nonneoplastic) mesenchymal (fibroblastic/myofibroblastic) proliferation. Although these cancers display a triple-negative and basal-like immunophenotype, they are low-grade tumors with a propensity for local recurrence (if not completely excised) but a very low risk of lymph node and distant metastases.^3–5 

MBCs as a whole are rare and mostly aggressive cancers with diverse morphologic differentiation patterns including a spindled cell (sarcomatoid) morphology, a bland spindled cell morphology with prominent fibrosis (fibromatosis-like), and squamous, chondroid, osseous, lipomatous, or rhabdomyoid differentiation.^9,14,15  The differential diagnosis of FLMC includes exuberant scar formation, nodular fasciitis, myofibroblastoma, inflammatory myofibroblastic tumor, pseudoangiomatous stromal hyperplasia, solitary fibrous tumor, phyllodes tumor, dermatofibrosarcoma protuberans, melanoma, and angiosarcoma. Previous studies using genomic profiling of MBCs have revealed alterations in the PI3K/AKT pathway in 83.3% of tumors. Additionally MBCs are TP53 mutated in about half of the cases.^9,14,16,17  Seventeen percent of MBCs additionally show a TERT promoter mutation in the study of da Silva et al,¹⁵  including spindle cell MBCs.

Similar to findings in other high-grade triple-negative MBCs, sequencing revealed frequent alterations in the PI3K/AKT/mTOR pathway in our study. All FLMC samples in our cohort had PIK3CA pathway alterations, 6 of 7 (85.7%) in the form of nonsynonymous PIK3CA mutations and 1 PIK3R1 loss-of-function mutation (1 of 7 [14.3%]). CDKN2A and CDKN1B alterations (loss or mutation) affecting the cell cycle control were observed in 5 of 7 cases (71.4%). TP53 mutations were absent in all FLMCs analyzed. Our results are in agreement with previously reported findings of frequent CDKN2A/1B mutations in MBC^16,18  and the absence of TP53 mutations in MBCs with spindle cell morphology in the study of Krings and Chen,¹⁹  which analyzed intermediate- and high-grade “metaplastic” spindle cell carcinomas of the breast but not low-grade FLMCs. CDKN2A loss and PIK3CA mutations were also described in myoepithelial-like metaplastic breast cancer.²⁰ 

Telomerase activity is one of the hallmarks of cancer and plays a major role in carcinogenesis.^21,22  Interestingly, TERT genetic alterations were found in all FLMC samples analyzed, in agreement with few previously reported cases^3,10  that also found the same TERT promoter mutation detected in our study.^3,10  We detected the c.-124 C>T point mutation in 6 of 7 cases.

The c.-124C>T mutation is a hotspot mutation described in various entities.²³  Although TERT promoter mutations are described in only 0.9% of breast cancers,²⁴  they are overrepresented in intermediate- and high-grade spindle cell (metaplastic) carcinomas¹⁹  and phyllodes tumors of the breast.²⁵ 

The TERT promoter harbors binding sites for transcriptional regulators.²³  The mutation is expected to generate a new binding site for ETS-transcription factors.^26,27  Indeed, functional studies demonstrated a significant enhancement of TERT-promoter activity after introduction of a c.-124C>T mutation.^26,27 

We confirmed the promoter-activating effect of this point mutation by gene expression analysis, showing TERT upregulation in mutated FLMC samples compared to nonneoplastic breast tissue (Figure 3, A through D). Furthermore, we demonstrated that the c.-124C>T alteration has an equivalent effect on gene expression, compared to TERT copy number gains, seen in about 15% to 25% of breast cancer samples.¹⁷  Interestingly, in our study, none of the examined low-grade TNBCs without fibromatosis-like or spindle cell morphology (second control group) showed any TERT alterations (mutation or copy number gain).

Copy number analysis of 5 FLMC samples revealed a low mean genome change of 8.34% (95% CI, 2.15%–14.53%), in stark contrast to a nearly 10-fold mean genome change of 73.8% (95% CI, 64.04%–83.56%) in high-grade TNBC samples analyzed for comparison (P < .001). Low-grade TNBC samples analyzed for comparison did not show a significant difference with that of FLMCs (mean genome change: 5.96% [95% CI, 0%–12.24%]). The chromosomal band 13q14.2 is known for gains in invasive ductal breast cancer,²⁸  and recurrent gains of this region were found in 3 of 5 FLMC samples (60%) in our study.

A synopsis of FLMCs in our study with data from the studies of Krings and Chen¹⁹  and da Silva et al¹⁵  and a recent case report³  demonstrates that FLMCs share PIK3CA/PIK3R1 mutations as their main oncogenic drivers with spindle cell (metaplastic) carcinomas without fibromatosis-like morphology. Equally, TERT promoter mutations are the defining alteration in both FLMCs and spindle cell (metaplastic) carcinomas without fibromatosis-like morphology. Seven of 7 FLMCs (100%) in our study, 3 of 4 FLMC cases (75%) in the study of Zhong et al,¹⁰  1 case report,³  and 4 of 5 pure spindle cell (metaplastic) carcinomas (80%) in the study of Krings and Chen¹⁹  harbored TERT promoter alterations (mostly hotspot mutations), a combined 15 of 17 cases (88%) (Figure 4). Krings and Chen¹⁹  also detected the same TERT promoter mutation in 40% of mixed spindle and squamous MBCs as well as in 1 pure squamous MBC, but not in any of 13 matrix-producing or otherwise mixed MBCs. In both data sets, low copy number alterations in low-grade metaplastic carcinomas with spindled morphology with or without fibromatosis-like morphology contrasted with high copy number alterations of matrix-producing, high-grade MBCs. FLMC/spindle cell carcinoma is further set apart from non–spindle cell-type MBCs by the presence of PIK3CA/R1 alterations but the absence of mutant TP53. Genetic alterations affecting TERT in the study of da Silva et al¹⁵  were detected in 10 of 60 MBCs (17%), including a recurrent hotspot mutation affecting the TERT promoter hotspot locus (c.-124C>T, also known as C228T, in 9 of 60 MBCs or 15%) and TERT gene amplification (1 of 60, 2%) in the presence of PIK3CA mutations and the absence of TP53 mutations. It is of note that none of the examined cases by da Silva et al¹⁵  were described as fibromatosis-like carcinoma.

In a recent study of 4 cases, Zhong et al¹⁰  reported common PIK3CA and TERT promoter mutations in fibromatosis-like spindle cell breast carcinomas (PIK3CA p.H1047R mutation and hotspot TERT mutations in c.124C>T were found in 3 of 4 examined cases). In our study, we found the same hotspot TERT mutation in 6 cases of FLMCs. In 1 FLMC case, a TERT copy number gain could be identified. Furthermore, we did not find any TERT alteration (mutation or amplification) in low-grade TNBCs without fibromatosis-like or spindle cell morphology. In contrast to the study of Zhong et al¹⁰  we have correlated the finding of TERT genetic alterations with TERT expression by using quantitative reverse transcription–polymerase chain reaction and found that in all 7 cases TERT was significantly overexpressed. While in our study we analyzed the status of the tumor mutation burden and found very low burden in FLMC, Zhong et al¹⁰  could not determine the tumor mutation burden in 3 of 4 cases.

According to the findings of the literature and our study, pure squamous cell MBCs seem to constitute an intermediate group, with cases either following the FLMC/spindle cell carcinoma pattern (mut-PIK3/wt-TP53/low-CNV) or the matrix-producing pattern (wt-PIK3/mut-TP53/high-CNV), with TERT promoter mutations seen in a subset of cases. Contrary to the spindle cell (metaplastic) carcinoma in the study of Krings and Chen,¹⁹  we did observe frequent loss-of-function alterations in CDKN2A/B in 5 of 7 FLMCs (71%). Despite their fibromatosis-like morphology, CTNNB1 and APC mutations were not seen in any of our FLMCs or in conventional spindle cell (metaplastic) carcinoma,¹⁹  as has been described in fibromatosis of the breast.²⁹  In summary, FLMCs are similar to non–fibromatosis-like/conventional low-grade triple-negative spindle cell metaplastic carcinomas on a genomic level but are distinctly different from high-grade MBCs.

Regarding actionable molecular alterations, PIK3CA-targeting agents such as selective PIK3CA inhibitors (eg, Alpelisib, currently approved for hormone receptor–positive breast cancer) might be of value, but clinical studies to support their use in TNBCs have only recently been initiated. The therapeutic agent Imetelstat (GRN163L), a TERT-targeting inhibitor tested in clinical studies for non–small cell lung cancer, failed to improve progression-free survival in patients with long telomere length (indicative of TERT overexpression).³⁰  Imeteltsat also showed no effect in a small cohort (n = 14) of metastatic breast cancers presented at the annual meeting of the American Society of Clinical Oncology 2010, but no information on the TERT status of these tumors was made available.³¹ 

In summary, low-grade FLMCs are rare triple-negative and basal-like carcinomas showing genomic stability, absence of TP53 mutations, activating mutations in the PI3K pathway, and a recurrent c.-124C>T TERT promoter mutation. Our findings, along with other rare publications on FLMCs, suggest that this TERT promoter mutation is likely to be one of the most characteristic genetic alterations of this rare breast cancer entity. Owing to the small number of cases, this study is hypothesis generating. Future studies are needed to validate the results and to show whether this common hotspot TERT mutation plays a key role in the development of this type of breast carcinoma. At present, based on the molecular profiles, PIK3 and TERT inhibition appear as the most promising actionable targets in cases of refractory disease with multiple relapses of FLMCs.

The authors thank Roche for its financial support.