NTRK, RET, ROS1, and ALK Gene Fusion in HER2 Immunohistochemistry 2+ Breast Carcinoma

HER2 immunohistochemistry (IHC) 2+ breast cancer patients need to determine the nal HER2 status by uorescence in situ hybridization (FISH) for selection of suitable treatment options. Although the HER2-positive cases can benet from the anti-HER2 targeted therapy, it only made a small proportion of this group, so nding more targeted therapy methods is necessary. NTRK, RET, ROS1 and RET gene fusions have been fully investigated in non-small cell lung carcinoma and are subject to targeted therapy in clinical practice and trials. However, there are only few reports investigating these four fusion genes in breast cancer. Our study is designed to evaluate the four fusion genes in HER2 IHC 2+ breast cancer patients to nd an alternative treatment option. One hundred and seventy-seven tissue samples were included. IHC was employed to assess ALK and NTRK protein levels. FISH probes specic for HER2, ALK, NTRK1, NTRK2, NTRK3, ROS1 and RET were used. The HER2-positivity rate of all HER2 IHC 2+ cases were 5.7%. The total fusion rate of the four oncogenes was 3.95% in HER2 IHC 2+ breast cancer patients. The fusion-positive patients were prone to be ER/PR/HER2 IHC triple negative (P=0.01) and were associated with poorly differentiated tumor (P=0.005). The NTRK, RET, ROS1, and ALK fusion rate was 0.56%, 1.13%, 1.13%, 1.13%, respectively. NRTK, RET, ROS1, and ALK fusion rearrangements were detected in triple-negative breast carcinoma patients which can provide patients with alternate treatment opportunities in clinical practice.


Introduction
Breast cancer is classi ed into molecular subtypes according to the expression of markers such as estrogen receptor-alpha (ER-α), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2). Each subtype is characterized by distinct clinical features, including response to treatment, pattern of recurrence, and survival. The determination of HER2 expression or ampli cation is pivotal during clinical practice, as it is a predictive marker for potential responsiveness to targeted therapies in breast cancer patients. The expression of HER2 is routinely determined by immunohistochemistry (IHC) and IHC results are divided into three categories based on scores: HER2-negative (0, 1+), HER2-equivocal (2+), and HER2-positive (3+). According to the recommendation of the American Society of Clinical Oncology (ASCO)/College of American Pathologists (CAP), samples of HER2 IHC 2+ form a special group that need further FISH analysis for the nal determination of HER2 status [1]. The nal hormone and HER2 status is pivotal for the selection of suitable treatment methods in the HER2 IHC 2+ patients. For example the hormone positive and HER2 negative patients in HER2 IHC 2+ group can receive the endocrine therapy, the HER2 positive patients in this group can accept the anti-HER2 targeted therapy. While the only approved treatment method for the hormone negative and HER2 negative patients (triple negative) patients is cytotoxic chemotherapy. Approximately 17% of HER2 IHC 2+ patients are classi ed as HER2-positive cases after the con rmatory uorescence in situ hybridization (FISH) test and this subgroup of patients can bene t from targeted therapy [2]. The other patients are devoid of targeted therapy methods in the HER-2 IHC 2+ patients. Additionally, advanced or metastatic hormone receptor-negative and HER2-negative breast cancer patients have no effective choice of treatment, development of novel drug targets offers alternative treatment opportunities. RTK (receptor-tyrosine kinase) gene rearrangement can lead to ligand-independent activation of the oncogenic signaling pathway and increase proliferation and survival of tumor cells. RTK rearrangement is a rare event in solid tumors. Approximately 250,000 new cases of invasive breast cancer are diagnosed in the U.S. annually, there are approximately 5000 new cases of RTK fusion-positive patients each year that may bene t from the corresponding targeted therapy [3]. As these biomarkers provide alternative treatment options for advanced or refractory breast cancer patients, it is worth the effort to investigate RTK fusion genes in breast carcinoma.
RTKs such as neurotrophic tropomyosin-related kinases (NTRKs) play an important role in neuronal development and exhibit limited expression in the nervous system after embryogenesis. NTRK codes for a family of genes: NTRK1 (also known as TRKA), NTRK2 (also known as TRKB), and NTRK3 (also known as TRKC). NTRK is a widespread cancer biomarker targeted for therapeutics, and oncogenic fusions involving NTRK1, NTRK2 and NTRK3 have been reported in a variety of malignancies and represent key mechanisms of oncogenic TRK activation [4]. Multiple fusion partners of NTRK genes have been described. The best known form is ETV6-NTRK3, which is a typical gene alteration in secretory carcinomas of the breast, salivary glands, and congenital brosarcoma [5][6][7]. However, the reported NTRK fusion rates in the non-secretory type breast carcinoma are variable and not fully investigated. ROS1, a proto-oncogene expressed in various tumor cell lines, is a regulator of cellular signal transduction pathway that mediates cell proliferation and migration and cell-to-cell communication. ALK gene fusion was rst reported in anaplastic large cell lymphoma in 1994 [8,9]. Thereafter, this fusion oncogene was found to be associated with the development of diverse tumor types of different lineages, including lung cancer, in am matory myo broblastic tumors, Spitz tumors, renal carcinoma, thyroid cancer, digestive tract cancer, breast cancer, leukemia, and ovarian cancer [10].
The new U. S. Food and Drug Administration (FDA)-approved NTRK-targeted drug elecintinib is also effective in ALK and ROS1 fusion-positive cancer patients. RET (Rearranged during transfection) is proto-oncogene which can be found in 6.8% of papillary carcinoma and 1-2% of NSCLC patients [11,12]. Additionally, cancers harboring RET fusion are collectively termed as 'REToma', similar to the termination of 'ALKoma' for tumors with ALK alterations [13,14]. Thus, considering the opinion that cancer may be treated with TKIs (tyrosine kinase inhibitors) irrespective of the tissue of origin, it is worthy to explore the speci c fusion rates of different cancer types before applying targeted therapy. In our study, we focused on evaluating the rearrangement of four different RTKs, NTRK, RET, ROS1 and ALK, in patients with HER2 IHC 2+ breast cancer.

Patient information
One hundred seventy-seven patients were enrolled in our study, who underwent surgery between October 2016 and January 2018 at Peking Union Medical College Hospital (Beijing, China). All the cases were diagnosed according to the morphology of the hematoxylin and eosin (HE) staining of the tumor sample and immunohistochemistry results. The secretory type of breast carcinoma was excluded from our cohort. Two experienced pathologists reviewed the HE slides to con rm the nal diagnosis.
The clinical and pathological information of the subjects, including age, sex, tumor sizes, ER-α, PR, HER2 and Ki-67 index, were collected from digital clinical archives and pathological reports.
This study was approved by the institutional review board of Peking Union Medical College Hospital.

Tissue microarray construction
The selective areas of representative morphology of the HE slides were labeled. The corresponding formalin-xed para n-embedded (FFPE) primary tumor samples were obtained from the Department of Pathology. The tissue microarray construction machine (Quick-Ray UT-06, UNITMA) was used and two coretissue biopsies of 2.0mm diameter were collected for each sample. Each block contained 6 cases of HER2-negative and 2 cases of HER2-positive cases, serving as the negative and positive control, respectively. This tissue microarray block was evaluated by FISH analysis. Apart Probe (ZytoVision GmbH, Bremerhaven, Germany). One hundred tumor nuclei per case were calculated and the case was considered positive (rearranged) if 15% or greater tumor cell nuclei were rearranged. Different evaluation criteria were set for different probes. For ALK, cell was considered positive if there was a split of two or more signal widths apart between the orange and green signals or there was a single orange signal without a corresponding green signal in combination together with a fused and/or split signal. For ROS1, cell was considered positive if there was a split of two or more signal widths apart between the orange and green signals or there was a single green signal without a corresponding orange signal in combination with a fused and/or split signal. For RET, cell was considered positive if there was a split of one or more signal widths apart between the orange and green signals or there was a single green signal without a corresponding orange signal in combination with a fused and/or split signal. For NTRK1/2/3, cell was considered positive if there was a split of one or more signal widths apart between the orange and green signals.
HER-2 FISH test was performed with a PathVysion HER2 DNA probe kit (Abbott Molecular, Des Plaines, IL, USA) according to the standard protocol. FISH results were evaluated according to the ASCO/CAP HER2 testing guidelines: it was considered positive when the ratio of HER2/CEP17 ≥2.0 or the average HER2 signal/tumor cell ≥6.0 with a ratio of HER2/CEP17 <2.0; HER2 negative was determined when the ratio of HER2/CEP17<2.0.
A total of 177 invasive breast carcinomas of HER2 IHC 2+ status was included in our study. The age of the patients ranged from 26 to 88 years (median 54 years). The histological subtypes included 171 (97.2%) cases of invasive carcinomas of no special types and 5 (2.8%) cases of special type breast carcinoma, which included 3 cases of micropapillary carcinomas, 1 case of papillary carcinoma, and 1 case of mucinous carcinoma. According to the morphology and IHC results, none of them were secretory type breast carcinoma. The median tumor size was 2.0 cm (ranging from 0.5 cm to 9 cm).
The total fusion rate of NTRK, ROS1, RET, and ALK genes was 3.95%. NTRK IHC result was positive in 5 cases, and 1 of them was further con rmed to be NTRK1-rearranged, using FISH analysis. The rest of the cases were devoid of break apart signal in the FISH analysis. The nal NTRK fusion rate was 0.56%. Two of the three ALK fusion-positive samples detected by IHC were con rmed to be positive by FISH test. The nal ALK fusion rate was 1.13%. There were two (1.13%) RET fusion-positive and two (1.13%) ROS1 fusion-positive patients in our cohort (Figure 1). Detailed information regarding the fusion-positive patients is summarized in Table 1.
The fusion rearrangement was associated with ER negative (p=0.011), PR negative (p=0.002), and triple negative (p=0.01) status. The fusion-positive tumors were poorly differentiated compared with the fusion-negative group (p=0.005). There was no signi cant difference between groups with respect to patients' age, tumor size, tumor type, HER-2 status, or Ki-67 index ( Table 2). Discussion NRTK, RET, ROS1, and RET fusion rearrangements are the most common genomic aberrations and have been fully investigated in non-small cell lung carcinoma (NSCLC). RTK-targeted therapy is available in clinical practice or trials. However, information regarding RTK fusion in breast carcinoma is limited. In our study, we found a fusion-positive rearrangement rate of 3.95% for the above genes in HER2 2+ breast cancer patients, and these patients were associated with ER/PR/HER2-negative status and poorly differentiated tumor type compared with the fusion-negative group. To our knowledge, this is the rst report showing the presence of RTK-fusion oncogenic alterations in breast cancer. The detection of the four RTK fusion genes in the HER2 2+ patients can provide this special group of patients an alternative treatment option, especially for the triple negative breast carcinoma patients who lack an effective and less toxic treatment options in this group.
NTRK is proved to be a pan cancer marker and NTRK-targeted therapy has been e cient regardless of tumor types or fusion partners. Moreover, the NTRK inhibitor entrectinib is also effective in ROS1 and ALK fusion-positive tumors. The RET, ROS1 and ALK fusion-positive NSCLC patients have similar clinical characteristics and were likely to be young and female and have non-smoker adenocarcinoma without EGFR, BRAF, and KRAS mutations. With the popularity of the terms 'ALKoma' and 'REToma' and effective treatment outcomes in fusion-positive tumors other than NSCLC, it is important to note that apart from fusion rates of these genes, clinical features of fusion-positive patients should be taken into consideration before testing. As there are no effective treatment methods for triple-negative breast cancer patients currently, the nding that fusion-positive breast cancer patients are prone to be triple-negative is valuable to clinical practitioners. Furthermore, the gene fusion in the breast carcinoma is a rare event, enrichment of the targeted population is pivotal, and testing the fusion genes in the triple negative patients could be a practical choice. Validation in a larger group of breast carcinoma patients, especially in the triplenegative category, will be needed in the future.
Larotrectinib and entrectinib are FDA-approved drugs that can be used in the treatment of solid tumors harboring NTRK gene fusion. It is pivotal to identify the fusion status of individuals to apply the correct targeted therapy. In our study, a positive NTRK-fusion rate of 0.56% was found in the HER2 2+ patients. The reported NTRK-fusion rates in literature range from 0% to 0.34% in the non-secretory type of breast carcinoma [15,16]. Despite the differences in the selected HER2 2+ patient population, in terms of ethnicity, the result of the present study is similar to the previous ones. Further studies on examining the rate of NTRK fusion in larger number of breast cancer patients of Asian origin are needed before a decision of rate of NTRK-fusion is made.
RET gene, a driver oncogene, is a therapeutic target and can be found in 6.8% of papillary carcinoma and 1-2% of NSCLC patients [11,12]. There are a few studies reporting an incidence of 0.17% of RET gene fusion in breast carcinoma. The main reason for the higher RET fusion rate in the present study (1.13%) may be due to differences in the enrolled patients. HER2-positive patients were the focus of the present study while in previous study, patients with available ER information were included and the RET fusion positive patients were prone to have a negative ER status. Furthermore, in the present study, two patients were ER positive and ER negative, respectively. The chromosomal rearrangement of RET leads to activation and formation of RET fusion protein which is capable of auto dimerization. Other RET gene alterations include mutation and ampli cation. The RET mutations in breast cancer patients are reported to be prone in young female patients and confer resistance to hormone therapy [17]. The RET fusion-positive patients in our study are all young patients of age 48 all young patients of age 48 and 26 at diagnosis, similar to the clinical characteristics of RET mutation patients. There was a preliminary report which showed the use of a combination of aromatase inhibitor and RET inhibitor resulted in better e cacy as opposed to aromatase inhibitor alone in breast cancer cell lines [18]. The RET inhibitor also showed promising results in refractory ER+/HER2+ breast cancer patients [19]. Although the fusion rate of RET gene in breast cancer is much lower than that of papillary thyroid carcinoma (PTC), it is comparable with that of NSCLC. Therefore, it is important to identify RET fusionpositive patients and provide alternative treatment options.
The average frequency of ALK fusion is approximately 5% in NSCLC and several FDA approved targeted therapy drugs have been used in NSCLC as a rst line treatment option [20].According to a report on gene rearrangements in solid tumors, EML4-ALK fusion was found in 5 of 209 breast carcinomas (2.4%) [21]. The fusion rate of ALK in our study is 1.13%, wherein the two individuals are triple negative breast cancer patients. The ALK kinase domain shares a similar sequence identity with that of the proto-oncogene ROS1 (50% similarity). The ALK inhibitor, crizotinib, is also effective in ROS1 fusion-positive NSCLC patients. Inhibition of ROS1 and its phosphorylation level lowers the in vitro alcohol-induced breast cancer cell proliferation and growth [22]. Another study showed that ROS1 inhibitors produce synthetic lethality in E-cadherin-de cient cell line and anti-tumor effects in in vivo models of breast cancer [23]. ROS1 gene rearrangements rate is approximately 1-2% in NSCLC patients [24,25], and presence of ROS1 fusion-positive patients in cholangiocarcinoma, glioblastoma, ovarian, gastric, and colorectal cancers is also reported [26]. Similarly, in the present study, two ROS1 fusion-positive patients (1.13%) were identi ed. The two tumors were poorly differentiated with invasion ductal carcinoma, not otherwise speci ed. One patient was ER+/HER2+ while the other was ER+/HER2-.
Our cohort included cases with HER2 expression with an IHC score of 2+, thereby narrowing down the study to speci c subtypes of breast cancer. Thus, this study may not re ect the true incidence of RTK fusion in the general population. Owing to the retrospective nature of this study, we analyzed the fusion rate of RTK genes independent of drug e cacy with regard to gene rearrangement characteristics. Nevertheless, this study shows that RTK fusion-positive status can provide breast cancer patients with alternate treatment opportunities in clinical practice.

Statistical analysis
The Statistical Package for the Social Sciences (SPSS) Version 20.0 Software (SPSS Inc., Chicago, IL) was used to analyze the data. Chi-square or Fisher's exact tests were employed to analyze the categorical variables or continuous variables as appropriate. All p values are reported as two-sided with the p values < 0.05 being considered statistically signi cant. ZL conceptualized the study design and paper writing. All authors read and approved the nal manuscript.
Data availability statement All datasets analyzed for this study are included in the article Compliance with ethical standards Con ict of interest The authors declare that they have no competing interests.
Ethics statement All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The retrospective study was approved by the institutional review board of Peking Union Medical College Hospital with a waiver for the need to obtain informed consent.