Expression of TTF-1 in breast cancer independently of ER expression: A case report and pathogenic implications

Abstract

Thyroid Transcription Factor 1 (TTF-1) is often used for the immunohistochemical evaluation of lung cancers, as a significant proportion of these cancers are positive while other adenocarcinomas are usually negative. Breast cancers are known to express TTF-1 only in a small minority of cases but this may be problematic when this staining is used for its differential diagnosis from lung cancer. We present a case of ER-positive breast cancer in 30% of tumor cells that was also TTF-1 positive in some areas of the primary tumor but lost ER expression completely in the metastatic recurrence site while retaining TTF-1 positivity. Additionally, a PTEN mutation was present on genomic evaluation of the primary tumor. Diagnostic, pathogenic and therapeutic implications of these findings are discussed.

Introduction

Thyroid Transcription Factor 1 [TTF-1, also called NKX2.1, TITF or, Thyroid-specific Enhancer-Binding Protein (T/EBP)] is a member of the homeodomain transcription factor family expressed in lung, thyroid and parts of the Central Nervous System (CNS) [1]. It has been used clinically for the immunohistochemical (IHC) investigation of metastatic cancers, given that cancers of lung and thyroid origin often express TTF-1, while adenocarcinomas of other origins usually do not. Nevertheless these associations are not always helpful as there is a percentage of cancers of lung origin not expressing the protein and carcinomas of non-lung origin that do express TTF-1. Adenocarcinomas of lung express TTF-1 in two thirds to three fourths of cases depending on the antibody used [2]. On the other hand, breast cancers have been reported to express TTF-1 in 2.4 and 2.8% of cases in two extensive series [3,4]. Moreover, to further complicate differential diagnosis sometimes, about one fifth of primary lung adenocarcinomas may express the Estrogen Receptor (ER) [5].

We report a case of TTF-1-expressing ER-positive primary breast cancer that continued to express TTF-1 when it became metastatic, while losing ER expression. The case suggests that the two transcription factors do not directly regulate the expression of each other in breast cancer. This finding has implications for the evaluation of both newly diagnosed metastatic and recurrent breast cancers. Further, we include an in silico investigation of the TTF-1 promoter regulation that may inform on future investigations for elucidation of TTF-1 expression in breast cancer.

Case presentation

A 53 years-old woman without significant past medical history was diagnosed with ER-positive, PR-negative Her2-negative carcinoma in the outer upper quadrant of the right breast. She underwent a mastectomy and axillary lymph node dissection. Pathology showed T2N1a disease with the tumor being a grade III infiltrating ductal carcinoma (Fig. 1a) of 4.3 cm in major diameter and 1 out of 10 lymph nodes were positive for metastatic deposits. The positive lymph node contained a tumor deposit of 1.1 cm with extranodal extension. ER positivity was observed in about 30% of tumor cell nuclei (Fig. 1b). Ki67 was positive in 50% of cells. No in situ ductal or lobular carcinoma was present. The patient received adjuvant chemotherapy with the FEC-D regimen (3 cycles of 5-FU, Epirubicin, Cyclophosphamide combination followed by 3 cycles of docetaxel) followed by adjuvant radiation therapy. Then she went on to hormonal therapy with tamoxifen and she also participated in a randomized trial of adjuvant denosumab versus placebo. Two years after her diagnosis she underwent hysterectomy with bilateral salpingo-oophorectomy and her hormonal treatment was switched to letrozole. A few months later she also had a prophylactic left mastectomy and bilateral breast reconstruction. Three years and three months after her initial diagnosis, in a scanographic evaluation as a follow-up for the trial, the patient was found to have a soft tissue mass in the right retrosternal area. Biopsy of this mass disclosed an invasive carcinoma (Fig. 2a) that was ER- (Fig. 2b), PR- and Her2-negative. IHC evaluation showed positivity for CK7 and diffuse strong positivity for TTF-1 (Fig. 2(c), Clone used: SPT24). GATA3 was weakly to moderately positive in some cells (Fig. 2d). CK20 was negative. Ki67 was even higher than in the primary at 75%. Given that no lung or thyroid primary tumors were evident, the primary breast cancer was evaluated for TTF-1 and it was indeed found to be positive in some areas, while negative in others (Fig. 1c). In addition, GATA3 was also positive in the primary tumor (Fig. 1d), concordantly with the metastatic focus. A plan was made for palliative chemotherapy followed by local irradiation. The patient was also enrolled in a study evaluating molecular profiling of advanced cancers in the community. Her primary tumor was found to harbor a mutation in exon 8 of the PTEN gene classified as probably modifying protein function. She was started on first line chemotherapy with weakly paclitaxel. She received six cycles with initial stability but then local progression at the retrosternal area. As the recurrent disease was triple negative she was switched to a regimen of cisplatin with gemcitabine. She received four cycles but she did not have a response. Since her disease was still localized, she received stereotactic radiation therapy. Unfortunately, she developed progressive disease with mediastinal hilar and supraclavicular lymphadenopathy soon after this treatment. She refused any further treatment and succumbed to her metastatic cancer four years and nine months after her primary diagnosis and eighteen months after metastatic disease diagnosis.

Discussion

TTF-1 is a member of the homeodomain transcription factor family. It is often expressed in lung adenocarcinomas and is used clinically for investigation of metastatic adenocarcinomas of unknown primary to help with identification of the primary site. TTF-1 gene is located at human chromosome 14q13 and is comprised of 3 exons [1]. It displays a high conservation between human, mouse and rat with 98% sequence similarity [1]. The protein is comprised of 371 aminoacids and has a DNA-binding homeobox domain in the middle part of the protein from aminoacids 161 to 220 that binds the specific sequence tcAAGkGcy (where k represents G or T and y represents C or T). Two activation domains occupy the N-terminal and C-terminal of the TTF-1 protein and a glutamine-rich inhibitory domain lies between the homeobox domain and the C-terminal activation domain [6]. Its normal expression is observed in thyroid and lung fetal development. Target genes of TTF-1 include several crucial lung development proteins such as surfactant proteins A, B and C and Clara cell secretory protein. Deletion of TTF-1 homolog in mice is lethal due to defects in these organs as well as the forebrain [7]. TTF-1 haploinsufficiency in humans from gene mutations of one allele leads to choreoathetosis, hypothyroidism and respiratory insufficiency with frequent infections [1,8].

Expression of TTF-1 is observed in 72.4% of lung adenocarcinomas when using the antibody clone SPT24 and in 65.4% of cases with the use of the less sensitive but more specific clone 8G7G3/1 [2]. In the same study squamous cell lung carcinomas had a significant lower positivity (16.8% and 1% with the two antibodies respectively). Another study that used clone 8G7G3/1 found a positivity of 72% in adenocarcinomas [9]. Despite some variability between studies, the positivity of lung adenocarcinomas appears to be in the range between 50 and 75%. Small cell carcinomas of the lung (SCLCs) as well as large cell neuroendocrine carcinomas (LCNECs), in contrast to carcinoids, express TTF-1 [10]. Thus TTF-1 is useful in diagnosis of adenocarcinomas, SCLCs and LCNECs of the lung. Nevertheless a small percentage of non-lung cancers express TTF-1 and this has to be kept in mind when relying on this marker for formulating a differential diagnosis. For example, in a series of 555 colorectal cancers TTF-1 was positive in 4.3% and 3.2% of cases with the use of SPT24 and 8G7G3/1 antibodies, respectively [11]. Expression in about 5% of bladder carcinomas and in cases of gynecologic cancers has also been observed [2,12]. In breast cancer, two series reported a TTF-1 expression of 2.4% (13 of 546 cases) and 2.8% (seven of 247 cases) [3,4]. Both investigations used the more sensitive SPT24 clone. Regarding sub-types, among these 20 TTF-1 positive cases, seven were ER-positive (five also PR-positive) and Her2-negative, eight were triple negative, one was ER-negative, PR-positive, Her-2-negative, one was hormone receptor-negative, Her2-positive and three were Her2-negative and the hormone receptors were unknown. Histologically, eighteen of the 20 cases were invasive ductal carcinomas and two were invasive lobular carcinomas [3,4]. Interestingly, in a more extensive series of 1132 patients from Asia, only one patient (0.09%) was TTF-1 positive using the 8G7G3/1 antibody [13]. This case was negative by evaluation by the alternative antibody SPT24 and was also positive for the neuroendocrine markers synaptophysin and chromogranin [14]. In fact, a few additional case reports of TTF-1 positivity in primary small cell breast cancers are available in the literature and this has to be considered when using TTF-1 as a tool in the differential diagnosis of primary neuroendocrine breast cancer from a metastatic SCLC [15,16]. Nevertheless, presence of an in situ component in the breast tumor would virtually prove that the breast is the primary site [16].

The patient described in this report displayed TTF-1 positive metastatic retrosternal disease that was ER-, PR- and Her2-negative. Her primary breast cancer, which was ER-positive, was confirmed to be also TTF-1 positive in part of the tumor. The TTF-1 positive primary clone was the one that gave rise to metastatic seeding. The primary tumor was PR-negative, a phenotype observed in two of the 20 patients in the two series, discussed above [3,4]. Overall the evolution of the disease in our patient and the data from other published cases suggest that TTF-1 expression in breast cancer is independent from hormone receptor transcriptional activity. Her2 overexpression or amplification appears also not to play a role in TTF-1 expression as both our case and all but one other reported cases have been Her2 negative. Of note, though, a PTEN mutation classified as probably modifying protein function was observed in our patient and this would be expected to activate the PI3K/Akt pathway and down-stream effectors, as well as the RAF/MEK/ERK pathway, similarly to Her2 pathway activation [17]. In the only other case of TTF-1 positive breast cancer where molecular evaluation is available in the literature, a mutation of PIK3CA (in addition to a TP53 mutation) was present and would be expected to activate the PI3K/Akt pathway [18]. PIK3CA and PTEN mutations are present in about 20% and 2% of breast cancers respectively, most commonly in ER-positive carcinomas and even more commonly in the subset that is PR-negative [19].

In order to further investigate TTF-1 expression regulation we performed an in silico query of the five promoter sequences of TTF-1 listed in the Transcriptional Regulatory Element Database (TRED, https://cb.utdallas.edu/cgi-bin/TRED/) for binding sites of the ER (ESR1), Androgen Receptor (AR, which has identical binding sequence with PR that is not listed in the Jaspar database), Glucocorticoid Receptor (GR, NR3C1), FOXA1 and GATA3 (both steroid receptor pioneer factors) and the FOS-JUN dimer and FOXO4 transcription factors. The two latter are transcription factors activated and inhibited by the KRAS/MEK/ERK and the PI3K/Akt pathways, respectively, down-stream of receptor tyrosine kinases. Promoter sequences were introduced in the query machine of the Jaspar database of vertebrate promoters (http://jaspar.genereg.net) and checked for the above binding sites. Results showed several binding sites for the pioneer factors (clusters of 5–14 sites in 4 out of 5 promoters for FOXA1 and 2–7 sites in 4 out of 5 promoters for GATA3) and for RTKs-dependent factors (5–14 binding sites for FOXO4 and 2–7 sites for FOS-JUN in all 5 promoters for both these transcription factors) but no sites for ER or AR in any of the five promoters. Interestingly a binding site for GR was present in one of the promoters. These results further argue for a lack of TTF-1 regulation by the hormone receptors in breast cancer and a possible role of oncogenic pathways.

Despite absence of Her2 up-regulation in most TTF-1 positive breast cancer cases reported so far, a high percentage of TTF-1 positivity has been shown in lung adenocarcinoma cases with activating mutations of the EGFR1 receptor (Epidermal Growth Factor Receptor 1), belonging to the same family of receptors as Her2 [20]. Indeed, TTF-1 negativity in lung cancer has a high negative predictive value of 94.2% for absence of such EGFR1 mutations and has been proposed as a surrogate marker for them [21]. Thus it is possible that activation of the KRAS/MEK/ERK and the PI3K/Akt pathways is a prerequisite for TTF-1 expression but additional conditions may be required in a specific cellular micro-environment that are fulfilled in lung cancers and some breast cancers with PI3K activating mutations or PTEN inactivating mutations but not in other Her2-positive breast cancers.

ER associated pioneer factor FOXA1 has been reported to be co-operating with TTF-1 in the lung in the transcription of genes such as surfactants [22]. In addition, TTF-1 expression is regulated by FOXA family members [23]. Thus, they may play a similar role in TTF-1 expression in breast under certain conditions, as suggested by the presence of binding sites in TTF-1 promoters. GATA3 is expressed in both ER-positive and -negative breast cancers [24] and it was expressed in our case. Another GATA family member GATA6 is expressed in the lung and induces TTF-1 expression in this organ [25]. The two factors subsequently co-operate in the transcriptional regulation of surfactant protein C [26]. GATA3 may substitute for GATA6 in TTF-1 regulation in breast cancer.

Presence of a GR binding site in one of TTF-1 promoters is intriguing. This steroid receptor is a known TTF-1 regulator in the lung and glucocorticoids are used in the prevention of neonatal respiratory complications of immaturity by induction of surfactants [1,27]. In breast cancer, GR is expressed in 70% and 30% of ER-positive and ER-negative cases respectively and may use FOXA1 as a pioneer factor for the shaping of its transcriptional activity [28]. Thus, similarly to the lung during development, GR could participate in the induction of TTF-1 in breast cancer.

In summary, we describe a case of a TTF-1 positive breast cancer that was also initially ER-positive but became ER-negative when recurred as metastatic. This also represents the first case where TTF-1 positivity is associated with a PTEN mutation in breast cancer. These data argue for the independence of TTF-1 expression from ER, PR and Her2 and a possible dependence on PI3K/Akt activation. A hypothesis on a role of GR is put forward that may have therapeutic implications in TTF-1 positive breast cancers if confirmed and deserves to be further investigated.