Identification Of AKR1B10 As A Novel Secretory Protein And Its Application As A Tumor Serum Marker In Breast Cancer Diagnosis

Aldo-keto reductase family 1 member B10 (AKR1B10) is primarily expressed in the normal human colon and small intestine. AKR1B10 efficiently catalyzes the reduction of aldehydes and ketones to form corresponding alcohols to protect cells from carbonyls-induced DNA damage and carcinogenesis. Our previous studies have shown that AKR1B10 mediates the ubiquitin-dependent degradation of acetyl-CoA carboxylase-a (ACCa). ACCa is a rate-limiting enzyme in long chain fatty acid synthesis. Therefore, AKR1B10 plays important roles in lipid synthesis, cell growth and proliferation. Futhermore, AKR1B10 is also highly related to tumorigenesis since it is overexpressed in liver, lung, cervices cancers.In this study, we developed a novel AKR1B10 detection ELISA kit, with purified AKR1B10 as standard samples and found not only that endogenous AKR1B10 is secreted from colon cancer cell lines HCT8, HT29, breat cancer cell lines H460, A549, MDA-MB-468, hepatic cancer cell line HepG2, but also exogenous AKR1B10 from 293T transfected with AKR1B10 gene. AKR1B10's secretion was stimulated by serum and decreased in serum-free or low serum conditions. Serum dialyzed with 5.0 kDa cut-off filters had less stimulatory activity at beginning, but the same plateau was eventually reached, suggesting that certain small molecules (such as ions) in serum were responsible for the immediate early stimulation, and large molecules (>5.0 kDa) participated in the later phase of stimulation. Cell number positively and cell density negatively affected AKR1B10's secretion. AKR1B10 is secreted in a functional form, which is identified by enzymatic activity assay to DL-glyceraldehyde.The potential mechanisms of AKR1B10 secretion were subsequently explored in current study. We found that AKRIBIO lacks a secretion signal peptide in its amino acid sequence by signal peptide prediction analysis. At the same time, AKR1B10's secretion was not affected by the protein synthesis inhibitor cycloheximide and the classical protein secretion pathway inhibitor brefeldin A. Therefore AKRIBIO is released through an endoplasmic reticulum (ER)-to-Golgi-independent non-classical secretion pathway.Lysosome-mediated protein secretion is one of most important non-classical secretion pathways. We isolated lysosomes and found AKR1B10 located in lysosomal fraction with Cathepsin D and protected from protease digestion. Fluorescence protein protection assay showed that EGFP-AKR1B10 was located in lysosomes marked with LysoTracker(?)Red DND-99, and this location can protect AKRIBIO from protease digestion by exogenous trypsin. AKR1B10's secretion is stimulated by many factors of stimulating lysosome exocytosis such as temperature, ATP, Ca2+and Ca2+carrier ionomycin, Mg2+chelator EDTA. Lysosomotropic NH4Cl induced AKR1B10 accumulation in lysosome and lysosome exocytosis to increase AKR1B10 release. Lysosomal exocytosis is regulated by several signal pathways modulating intracellular Ca2+ concentration, such as ADP-ribosylation factor (ARF), phospholipase C (PLC) and G-protein. AKR1B10 secretion is inhibited by Exo-1, an ARF inhibitor, and U73122, a PLC inhibitor, but stimulated by GTPyS, an activator of G-protein, and fMLP, a ligand of G-protein coupling receptor, suggesting AKR1B10 secretion is modulated by the signal pathways regulating secretory lysosomes exocytosis. AKR1B10 is secreted together with Cathepsin D, a lysosomal marker. Therefore, AKR1B10 is secreted through lysosome-mediated non-classical pathway. We also determined the translocation of AKR1B10 into lysosomes. ABC transporter inhibitors Glib and DIDS suppressed AKR1B10 secretion by reducing its accumulation in lysosomes, suggesting the role of ABC transporters in the entry of AKR1B10 into lysosomes.In this study, we also identify AKR1B10 as a secretory protein in vivo. AKR1B10 is primarily expressed in the normal human colon and small intestine, so we detected AKR1B10 concentration in ileal fluid from normal subjects, and found that in the intestine, AKR1B10 is specifically expressed in mature epithelial cells and secreted into the ileal fluids of lumen at the concentration of 188.6?535.7ng/ml (average= 298.1 ng/ml, n=11). Then we further explored the mechanism on the transportation of AKR1B10 to lysosome. HSP90a, a chaperone molecular, plays an important role in protein folding, transportation. We found that cytosolic AKR1B10 interacts with HSP90a by immunoprecipitation and pulldown assays. Geldamycin (GA), a HSP90a-specific inhibitor, decreased AKR1B10 accumulation in lysosomes and AKR1B10 secretion. Therefore HSP90a transports AKR1B10 to lysosomes.A functional domain is often recognized in the proteins secreted through a nonclassical secretion pathway. We tested the secretory activity of different N-and C-terminal peptides. GST-tagged N/C-terminal peptides were expressed in 293T cells and tested for their secretory activity. The N-terminus up to 231aa had not secretory activity, whereas C-terminal peptides (C204-261 and C204-316) were secreted with efficiencies of 74.2±3.0%and 81.5±5.5% compared to full length AKR1B10 (316aa), suggesting that the functional domain mediating AKR1B10 secretion may be located in the C-terminal peptide C204-261. We replaced the amino acids with a long, charged residue in helix 10 (233-240aa) with alanine, producing mutants K233A, E236A, or K240A. A single mutation of K233A, E236A, or K240A decreased secretory efficiency of AKR1B10 to 59.2±2.1%,63.3±1.1%, and 46.9±3.1%, respectively. Combinations of two-site mutations further reduced the secretory activity to 43.15±1.5% for K233A plus E236A,33.3O±1% for K233A plus K240A, and 33.52±3.1% for E236A plus K240A, and all three-site mutations lowered down AKR1B10 secretory activity to 18.2±3.2%. These secretory data were supported by the decrease of the mutant AKR1B10 in lysosome. The mutant AKR1B10 in lost the capability of interacting with HSP90. These data suggest that the helix 10 acts as a functional domain for the secretion of AKR1B10 and the amino acids K233, E236, and K240 in this helix are the key residues.Recently, my colleagues found that overexpression of AKR1B10 in breast cancer is associated with tumor growth, lymph node metastasis, and patient survival, suggesting that AKR1B10 may promote tumor growth and progression, thus being a novel prognostic marker. In this study, we detected AKR1B10 expression in tissue microarray of breast cancer by immunohistochemistry and found AKR1B10 was undetectable or at a very low level in normal breast lobules and ducts, but overexpressed in 20 of 28 (71.4%) ductal carcinomas in situ,184 of 220 (83.6%) infiltrating carcinomas, and 28 of 32 (87.5%) recurrent tumors. Moreover, we collected 50 cases of tumor tissue and blood samples of breast cancer patients and found that AKRIB10 was highly expressed in tumor tissue than adjacent noncancerous tissue in 48 cases, and the AKRIB10 concentration was 20.72 ng/ml, which was the higher than the concentration of 5.19 ng/ml in normal subjects. There is a positive correlation between AKR1B10 concentration in serum and AKR1B10 expression in tissue of the breast cancer of patients, suggesting that AKR1B10 may be a novel serum marker for breast cancer.In summary, this study demonstrated that AKR1B10 is secreted through a HSP90a-mediated lysosomal nonclassical pathway and may act as a serum biomarker for the diagnosis of breast cancer.