Abstract

   Breast ductal carcinoma in situ (DCIS) has been typically recognized by
   pathologists on the basis of aberrant mammary duct morphology. Thus,
   there are increasing efforts to detect DCIS biomarkers and druggable
   targets. In this study we focused on the molecular mechanism involving
   Annexin A8 (ANXA8), a Ca^2+ and phospholipid binding protein, which is
   regulated by all-trans Retinoic Acid (RA), and it is highly expressed
   in breast DCIS tissue samples relative to atypical ductal hyperplasia,
   and normal breast tissue. Using a panel of human mammary epithelial
   HME1 cell lines that share a common protein signature, and develop in
   vitro three dimensional (3D) “DCIS-like” amorphous structures, we
   identified by bioinformatics analysis protein-miRNA pairs, potentially
   involved in mammary morphogenetic mechanisms, including the ANXA8
   mechanism. HME1 cells with genetic mutations hampering the
   physiological RA regulation of the RA receptor alpha (RARA)
   transcriptional function, but retain the RARA function controlling the
   PI3KCA-AKT signaling, develop 3D “DCIS-like” amorphous structures with
   upregulated ANXA8. Consistently, ectopic ANXA8 expression, by affecting
   the RARA transcriptional function, induced HME1 DCIS-like amorphous
   acini expressing phosphorylated AKT (P-AKT). Apparently, a
   RA-RARA-ANXA8 feedback loop fosters a vicious circle of aberrant
   morphogenesis. Interestingly, a few miRNAs regulated by RA are
   predicted to target ANXA8 mRNA. These miRNAs are candidate components
   of the RA-RARA-ANXA8 mechanism, and their deregulation might induce
   DCIS initiation.

   Keywords: ductal carcinoma in situ (DCIS), 3D HME1 models,
   protein-miRNA pairs, DCIS biomarkers and morphogenetic mechanisms

1. Introduction

   Early breast cancer stages, such as ductal carcinoma in situ (DCIS),
   are characterized by confined breast lesions with aberrant ductal
   morphological features. When left untreated, these in situ lesions can
   become invasive [[26]1,[27]2,[28]3]. Detecting and treating breast
   cancer at its earlier stages can greatly increase survivability. To
   identify early stage breast cancer and prevent disease progression,
   research efforts have been focused on the identification of DCIS
   biomarkers and druggable targets. However, despite these efforts, the
   heterogeneity of breast DCIS lesions has made it extremely challenging
   to identify biomarkers with high diagnostic and prognostic value
   [[29]4] that can be used also in liquid biopsies [[30]5]. In an attempt
   to identify biomarkers with high sensitivity and specificity, we set
   out to use an unconventional experimental approach based on three
   dimensional (3D) mammary epithelial cell models. When grown in basement
   membrane culture non tumorigenic mammary epithelial cells, including
   HME1 cells, form 3D structures with a lumen-enclosing epithelial
   monolayer of apicobasal polarized cells, which is typical of normal
   mammary ducts [[31]1,[32]6,[33]7,[34]8]. In contrast, HME1 cells
   carrying different genetic mutations- hereafter referred to as HME1
   “DCIS-precursor” cell lines- develop 3D DCIS-like structures with a
   luminal space filled with proliferating cells and loss of apicobasal
   polarity. HME1“DCIS-precursor” cell lines when grown in 3D basement
   membrane culture evade the growth inhibitory and pro-apoptotic effects
   regulated by physiological Retinoic Acid (RA) via the RA receptor alpha
   (RARA) transcriptional function, which is indispensable for lumen
   formation [[35]7,[36]9,[37]10,[38]11,[39]12], but retain the RARA
   function that enables the physiological RA induction of AKT signaling
   by the phosphatidylinositol-3 kinase catalytic (PI3KCA) subunit
   ([[40]13] and references within).