Overcoming resistance to HER2 blockade in breast cancer: AXL as a promising druggable target and prognostic biomarker

Introduction: Breast cancer is a heterogeneous disease consisting of uncontrolled growth of epithelial cells originating in the ducts or breast lobules. It is currently the most diagnosed cancer worldwide and the fifth cause of cancer-related death in general population, but the first cause in women. Breast cancer incidence is higher in developed countries due to increased risk factors and to increased and earlier detection by mammography screening, which considerably reduces mortality. This is the reason why transitioning countries have higher mortality despite having lower incidence. Regarding the origin of the disease, it is evidenced that the pathways implicated in normal development of the mammary gland are deregulated in breast cancer disease. Nevertheless, despite several models have been proposed, the specific mechanism by which breast cancer is originated and how its formation and progression occur is still unknown. Between the risk factors, genetic risk factors are responsible of around 10% of breast cancers, being BRCA1 and BRCA2 the most common inherited mutations. Non-genetic risk factors are related to ethnicity, breast conditions or previous breast pathologies and lifestyle factors, such as hormonal therapy, alcohol consumption, overweight, lack of physical activity and late age at giving birth or not breastfeeding. Diagnosis in the very initial steps of the disease is one of the major issues in the clinic and enables efficacy of the therapies and reduces the risk of breast cancer death. For this issue, mammography screening has been demonstrated to be the best strategy and is implemented in most of the developed countries. Mammography together with clinical examination, imaging and needle biopsy are performed for diagnosis. After diagnosis, breast tumors should be classified according to TNM stage, mainly based of tumor size, lymph node involvement and metastasis; histological grading based on differentiation; histological subtypes that include ductal invasive carcinoma, lobular invasive carcinoma and other special subtypes, and pathological or intrinsic subtypes. Currently pathological subtype, which is based on expression of estrogen and progesterone receptors, HER2 and Ki67 by immunohistochemistry, determines treatment decision. Molecular intrinsic subtyping is based on gene expression in tumor tissue and classifies breast cancer into six subtypes. This classification has demonstrated potential value to predict prognosis and treatment response. However, it is not fully applicable to the clinic nowadays. Regarding treatment, locoregional therapy of early breast cancer consists of a surgery and radiation therapy if necessary. For systemic treatment, neoadjuvant or adjuvant strategies could be performed with endocrine therapy in luminal tumors, HER2 blockade in HER2+ breast cancer and chemotherapy, that could be used in all subtypes but is the standard of care for triple negative breast cancer. HER2+ breast cancer presents overexpression of HER2 receptor which, upon dimerization, leads to activation of downstream pathways involved in proliferation, angiogenesis, survival, differentiation, apoptosis, invasion and metastasis. Between them, RAS/RAF, MAPK, JAK/STAT and PI3K/AKT/mTOR are the most studied. HER2+ breast cancer accounts for 15-20% of the cases and is characterized by an aggressive biological and clinical behaviour which correlates with poor prognosis. However, development of targeted therapies against HER2 has considerably improved HER2+ breast cancer patients’ outcome and changed the paradigm of clinical management. Based on this fact, a variety of anti-HER2 agents have been approved for the treatment of HER2+ breast cancer and others are under evaluation. The antibody trastuzumab was the first anti-HER2 agent to be developed and approved and, in combination with pertuzumab, remains the standard of care for HER2+ breast cancer. Despite the evident improvement of HER2+ breast cancer disease management and the effectiveness of anti-HER2 therapies, still around 20% of HER2+ breast cancer patients experience relapse leading to metastatic disease, that nowadays is considered incurable. Disease relapse may occur due to resistance to anti-HER2 agents. Thus, the main challenge in the last years has been to decipher the mechanisms underlying anti-HER2 resistance, to identify biomarkers to predict treatment response and to develop new therapeutic strategies to overcome resistance. Four major mechanisms have been described to be involved in anti-HER2 therapy resistance: impediments of drug binding to HER2 due to upregulation of mucines or HER2 isoforms; failure to trigger immune response because of Fc receptor dysfunctions; activation PI3K/AKT/mTOR downstream signaling pathway due to PIK3CA mutations, loss of PTEN or AKT deregulation; and upregulation of alternative pathways such as cell cycle, SRC or FASN. Crosstalk with other receptors has also been described to occur and to activate alternative pathways. Between them, crosstalk with hormone receptors, HER family members and other receptors tyrosine kinase have been involved in anti-HER2 resistance. Epithelial to mesenchymal process is implicated in metastasis, invasion, migration and colonization and has been related to drug resistance, including anti-HER2 resistance. Some receptors tyrosine kinase, such as AXL and MET, have been described to be involved in this process and subsequently take part in metastasis and mechanisms of therapy resistance. AXL can be activated by ligand-dependent or independent mechanism and by dimerization, leading to activation of downstream signaling pathways, such as PI3K/AKT/mTOR, JAK/STAT, NF-κB, and RAS/RAF/MEK/ERK that play major roles in tumor cell survival, migration, invasion, anoikis, angiogenesis and drug resistance. AXL is overexpressed in several types of cancer including breast cancer and correlates with worse prognosis. In breast cancer, it was shown to be a driver for metastasis and to be essential for epithelial to mesenchymal process, besides, AXL has been proposed as a therapeutic target in triple negative breast cancer. Regarding drug resistance, associations between AXL and chemotherapy or targeted drugs resistance have been reported in several cancers. In the particular case of breast cancer, AXL is involved in chemo- and radio-resistance and EGFR-targeted therapy resistance in triple negative breast cancer. However, the role of AXL in anti-HER2 therapy resistance has been poorly investigated. Several AXL-targeted therapies have been shown to supress tumor growth and metastasis and to potentiate the effect of chemotherapy and targeted therapies and some of them are currently under clinical evaluation in solid tumors. Given the knowledge of AXL's role in resistance to therapy, future studies will help to determine the translational application of AXL as a biomarker of treatment response and as a therapeutic target in HER2+ breast cancer. Objectives: In the frame of the antecedents described above, we hypothesize that AXL could be involved in acquisition of trastuzumab resistance in HER2+ breast cancer maybe due to crosstalk with HER2 receptor. Thus, the present work is focused on the study of the role of AXL in acquisition of trastuzumab resistance, as well as its applicability as a therapeutic target and prognostic biomarker. Methods: In this study AU565, BT474 and SKBR3 HER2+ breast cancer cell lines were selected for in vitro experiments and acquired trastuzumab-resistant cell lines (AU565R, BT474R and SKBR3R) were generated by exposure to increasing concentrations of the drug for prolonged time. Innate trastuzumab-resistant HCC1954 HER2+ and MDA-MB-231 triple negative breast cancer cell lines were used as controls. Tumor tissue samples were obtained from patients with early-stage HER2+ breast cancer treated at the Hospital Clínico Universitario de Valencia by standard guidelines, after signature of informed consent and ethical approval of the study. Data from neoadjuvant phase II PAMELA study were used to analyze gene expression in this cohort of HER2+ breast cancer patients. To evaluate changes in gene expression, mRNA levels were measured by real time quantitative PCR on RNA isolated from breast cancer cells and patients’ samples. For protein expression analysis, western blot, immunohistochemistry and flow cytometry were performed and co-immunoprecipitation and proximity ligation assay were used to evaluate dimerization between AXL and HER2 proteins. Apoptosis and viability were analyzed with Annexin V and propidium iodide staining by flow cytometry. To study the effect of the loss and gain of function of specific proteins, small-interfering RNAs or cDNA overexpression plasmids were transfected into the cells by lipofectamine reagent and several functional assays were carried out. Proliferation was measured by WST-1 assay and migration and invasion were studied by wound healing and transwell assays. Mammospheres were generated from BT474R cells and number of live cells was measured by flow cytometry with propidium iodide. Patient-derived xenograft model was obtained from a HER2+ breast cancer patient. Two in vitro trastuzumab-resistant patient-derived xenograft cell lines were established by subculture in presence of increasing concentrations of trastuzumab and five trastuzumab-resistant in vivo patient-derived xenografts were generated by transplantation of tumor tissue into NOD-SCID mice treated with trastuzumab and grown with serial passages. 3D models were also generated from patient-derived xenograft cell lines and in vivo models in which viability and apoptosis were measured after treatment. For the in vivo experiment, NOD-SCID mice were injected in mammary fat pad with trastuzumab-resistant patient-derived xenograft tumors (N=30), randomly divided into 4 groups and treated with either vehicle, trastuzumab, TP-0903 or a combination of both. Tumor size and body weight were measured during follow-up and animals were sacrificed at the end of the experiment or when institutional euthanasia criteria for tumor size and health condition were reached. In-silico analysis were performed for gene expression studies in public databases and for transcription factor binding site prediction. The prognostic value of AXL and GAS6 mRNA expression was analyzed in 252 HER2+ breast cancer patients with available data from Kaplan-Meier Plotter Software. To identify predicted putative binding sites of JUNB in AXL promoter JASPAR 2022 and PROMO software were used. Statistical analysis was performed using GraphPad Prism and R software. Results were expressed as means ± SD, means ± SEM for in vivo tumor size and median and interquartile for mRNA expression in patients’ samples. Normality was checked using Shapiro-Wilks’s test. Mean comparison was carried out using two-tailed t-Student test. All experiments were reproduced at least three times. mRNA expression in patients’ cohort was dichotomized by the median and ROC analysis was carried out. Disease-free survival and overall survival were described graphically using Kaplan-Meier curves and differences were evaluated using log-rank test. Selection of the best model was done according to Akaike’s Information Criterion. To identify AXL gene expression changes in PAMELA trial we used two-tailed paired t-tests. Two-way ANOVA with Bonferroni correction post-hoc was used for in vivo statistical analysis. The cut-off for statistical significance in all tests was p-value