Background
Breast cancer (BC) is the most common cancer in women worldwide and the leading cause of cancer deaths for women [
1]. Approximately 15–20% of patients with this tumor overexpress the human epidermal growth factor receptor 2 (HER2) protein [
2]. HER2 can activate downstream signaling cascades that induce cell proliferation through the Ras-mitogen-activated protein kinase (MAPK) pathway and inhibit cell death through the phosphatidylinositol 3′-kinase (PI3K)/protein kinase B (Akt)/mammalian target of the rapamycin (mTOR) pathway [
3,
4]. Characterizing HER2 as a proto-oncogene, a poor prognostic marker, and eventually as a therapeutic target has dramatically changed BC classification, risk assessment, and treatment [
5]. The development of powerful targeted therapies directed specifically at HER2 has improved the survival of patients with both early-stage and metastatic BC.
Trastuzumab (Herceptin®; Genentech, Inc., South San Francisco, CA, USA) is a humanized monoclonal antibody that selectively binds with high affinity to the extracellular domain of human HER2 protein and was the first targeted drug approved for treating HER2+ BC [
6]. Although there are currently other anti-HER2 agents available, trastuzumab remains the gold standard for treating this subtype of BC. Despite the clinical benefits trastuzumab treatment in HER2+ BC brings, a large percentage of patients display primary or acquired resistance to the drug. In the last few years, several studies have focused on identifying the molecular mechanisms of trastuzumab resistance, such as aberrant activation of downstream signaling pathways [
7] or the HER2 carboxyl-terminal fragments (CTF), also known as p95HER2, which are frequently found in HER2-expressing BC cell lines and tumors [
8]. Despite these efforts, the complete picture of the molecular mechanisms triggering trastuzumab resistance in BC remains unclear. Therefore, clarifying these mechanisms and identifying new resistance biomarkers is essential in the advance towards precision oncology in BC and the quest for new treatment options for those patients who do not respond to trastuzumab therapy.
DNA methylation is the most well-known epigenetic modification in humans and has been implicated in regulating the expression of a great variety of critical genes in cancer [
9]. For this reason, DNA methylation status has emerged as one of the most promising epigenetic biomarkers for several types of cancer, including BC [
10]. These epigenetic markers can be useful in detecting tumors earlier or identifying patients with an increased risk of cancer, as well as evaluating disease progression or predicting the response to anticancer drugs [
11]. In the last few years, some significant genes that are inactivated by promoter methylation in BC have been identified, including BRCA1 [
12] and RASSF1A [
13]. However, the analysis of hypermethylated genes in association with trastuzumab resistance is still a largely unexplored field that holds great potential.
The aim of this study was to evaluate the implication DNA methylation has in trastuzumab resistance and to identify epigenetically regulated genes with potential clinical value as biomarkers for trastuzumab resistance in HER2+ BC patients. To this purpose, we employed an integrative approach with genome-wide DNA methylation (450K array) and a transcriptomic analysis (RNA-Seq) in trastuzumab-sensitive and trastuzumab-resistant cell line models (SK and SKTR). In vitro results were validated in a cohort of 24 HER2+ tumor samples from patients experiencing both sensitivity and resistance to trastuzumab-based neoadjuvant therapy settings. Here, we determine epigenetic inactivation of the
TGFBI gene by promoter CpG island hypermethylation, which are CpG-rich regions of DNA that are often associated with the transcription start sites of genes, with possible implications for trastuzumab-resistant BC pathways [
14]. The hypermethylation of
TGFBI also suggest its potential clinical usefulness as a biomarker for trastuzumab resistance in HER2+ BC patients.
Discussion
HER2 is overexpressed in 20–30% of BC and is associated with aggressive phenotype and poor prognosis. Despite the initial good response to the treatment, a large portion of patients present de novo or acquired treatment resistance. Currently, HER2 detection is the only validated biomarker for predicting the benefit of anti-HER2 therapies [
33]. While different cellular and molecular mechanisms involved in trastuzumab resistance have been described, none of them are used to detect, predict, or monitor BC treatment in a clinical routine [
7,
8,
34]. Thus, there is great interest in finding potential biomarkers able to detect, predict, or monitor trastuzumab response which, in turn, would allow us to stratify patients and save costs and toxicities for those who do not respond to treatment.
Epigenetic biomarker studies are focused on analyzing the methylation changes in the promoter regions of candidate genes in specific tumor types for the implication in gene silencing [
9,
34]. In general, this gene repression results in an adaptive advantage for the cells, allowing a more aggressive and invasive phenotype to be adopted. After analyzing the promoter methylation profile of long-term trastuzumab-sensitive and trastuzumab-resistant models developed in our laboratory [
16,
17], three different methylated genes (
TGFBI,
CXCL2, and
SLC38A1) implicated in cancer were identified and subsequently validated using different methylation and expression approaches. Additionally, the selected genes were validated in two other long-term trastuzumab-sensitive (AU) and trastuzumab-resistant (AUTR) human HER2+ BC cell lines developed by our group [
17]. Although both cell lines analyzed derive from the same patient and have similar genetic and transcriptomic profiles, they have some differences in transcriptional levels [
35]. Moreover, SKBr3 and AU565 have been described as having distinct responses to drug treatment depending on culture conditions [
36]. From all the genes selected,
TGFBI (also known as Big-H3 or keratoepithelin) was the only one with a consistent methylation, expression, and protein pattern in both resistant models. This epigenetic silencing of the
TGFBI gene observed was in line with previous studies of other cancers [
37,
38].
As explained before, TGFBI encodes a 68-kDa secretory protein induced by transforming growth factor-β (TGF-β) in human adenocarcinoma cells as well as other human cell types [
39]. According to the bibliography, TGFBI protein is composed by 683 amino acids containing a secretory signal (SP) in the
N-terminal cysteine-rich domain (CRD), and four fasciclin-1 domains (FAS1-1, FAS1-2, FAS1-3, and FAS1-4), which contain several known integrin-binding motifs including NKDIL, YH18, and EPDIM as well as an Arg-Gly-Asp (RGD) domain [
22]. TGFBI has been reported to function as an ECM protein to mediate cell adhesion and migration through interacting with collagen, fibronectin, and laminin and several integrins including α1β1, α3β1, αvβ3, and αvβ5 [
30‐
32]. While TGFBI has been reported to be involved in tumorigenesis, its role is not clear. In this paper, we have linked the promoter DNA methylation-associated silencing of
TGFBI with a possible tumor suppressor function in trastuzumab-resistant models. Several previous reports have indicated that TGFBI plays the role of a tumor suppressor gene in various cancers such as lung, breast, and ovarian [
25,
40‐
42]. However, in other cancers, such as colon or pancreas, TGFBI has been described as having a tumor-promoting function [
43‐
46]. These opposing effects of TGFBI suggest that its expression and function are dependent on cell type [
22].
TGFBI promotes cell adhesion by interacting with different integrins, which, in turn, has been shown to play an important role in tumor progression in humans. TGFBI’s inhibition of invasive capacity correlates with a previous analysis by our group, which revealed a greater invasive capacity for SKTR compared to the SK [
16]. To determine if TGFBI has a potential role in trastuzumab resistance, different functional studies were carried out. Our results show that TGFBI overexpression in SKTR significantly sensitizes the cells to the treatment and activates some HER2 downstream protein pathways (pAKT, pHER1, and pHER2), adopting a similar pattern to SK. In contrast, TGFBI knockdown in TGFBI endogenous expressed in the SK cells did not produce any effect. A direct interaction between the HER2 receptor and different integrins has been described in multiple reports [
47,
48]. Moreover, different integrins have been related to trastuzumab resistance such as β1 or α6β4 [
49,
50]. Therefore, TGFBI mutated vector was performed to investigate the role of the RGD motif and the three integrin-binding motifs NKDIL, EPDIM, and YH18 in trastuzumab response [
22]. Like TGFBI overexpression, the TGFBI mutated form induced the activation of different HER2 downstream proteins, but there were no differences in trastuzumab treatment response. Therefore, it is likely that the four mutated domains of TGFBI are required for its function in trastuzumab response, which is consistent with other studies that described the involvement of FAS1 or RGD motifs in tumor angiogenesis and tumor growth inhibition, as well as in promoting apoptosis [
31,
51‐
53]. Moreover, it has also been described that TGFBI is involved in mesothelioma progression through the AKT/mTOR pathway which could be related to the HER2 pathway changes observed in TGFBI overexpression and mutagenesis [
54]. In summary, this study suggests that TGFBI expression may promote the effectiveness of trastuzumab treatment.
The methylation of
TGFBI showed certain capacity to predict the resistance to trastuzumab according to the Hosmer-Lemeshow test; however, these results should be interpreted with caution because the number of grouped patients was small. Although more in vitro functional analysis are necessary to elucidate the TGFBI role in trastuzumab resistance, we suggest some hypothetical situations taking into account all TGFBI action mechanisms. It has been described that TGFBI could inhibit cell adhesion to various ECM proteins inhibiting cell proliferation and invasion in neuroblastoma [
55]. Moreover, cell adhesion to ECM has been consistently reported as one of the mechanisms used by tumor cells to resist chemotherapy [
56]. These observations are in correlation with previous studies by our group, where it is exposed that our SKTR model presented a high significant adhesion capacity to bind to extracellular matrix proteins like fibronectin, collagen I, collagen IV, or laminin I compared to SK [
16]. Therefore, we hypothesized that in the absence of TGFBI, the integrins could bind to the ECM providing attachment sites for cells inducing more invasion, or acting as a physical drug barrier, restricting drug transport and limiting their efficacy. The intratumoral diffusion and physical masking of HER receptors by ECM proteins have also been described as a mechanism that can affect the therapeutic efficacy of some drugs including trastuzumab [
57].
Finally, we correlated the previous in vitro results in human samples treated with neoadjuvant anthracycline-taxane-based chemotherapy
plus trastuzumab. In our preliminary study, a higher
TGFBI methylation level after treatment was observed in patients who developed treatment resistance. These results are in accordance with other studies which demonstrated an association between
TGFBI hypermethylation and poor prognosis in prostate and lung cancer [
41]. Moreover, decreased
TGFBI expression was identified in advanced stages of BC and NSCLC tumors [
40,
58]. Due to the small size of the patient cohort and the retrospective design, the role of
TGFBI methylation as a biomarker requires further validation in a larger and independent cohort. Nevertheless, the present study brings to light for the first time that
TGFBI methylation has a significant discriminative value between pre- and post-treatment samples, as demonstrated by ROC curve analyses. Interestingly, the
TGFBI promoter methylation analysis in responder patients showed similar methylation levels with pre-treatment samples of non-responders. This observation suggests a possible role of
TGFBI as a monitoring biomarker for trastuzumab response in patients with HER2+ BC.
Although no significant association was found between
TGFBI methylation before and after treatment and patients’ clinical-histopathologic characteristics, a high number ER-positive patients were observed. In this sense, previous studies have shown that there is crosstalk between ER and HER2 pathways and it affects the response to treatment [
59]. In addition, although the type of surgery is not associated with a higher
TGFBI methylation level in post-treatment samples, 73% of our patients were treated with mastectomy versus 27% with lumpectomy, probably due to the high prevalence of stage III. The low histological grade and high clinical stage to treatment were probably due to most of the patients having non-operable cancers including inflammatory tumors. Currently, neoadjuvant treatment is generally used for operable HER2+ BC thanks to an improvement in the efficacy of the drugs used for treatment [
60]. At the time our cohort was treated, none of these drugs were yet available. That said, trastuzumab remains the gold standard treatment for HER2+ BC.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.