Aldo-keto Reductase Family 1 B10 Affects Fatty Acid Synthesis by Regulating the Stability of Acetyl-CoA Carboxylase-α in Breast Cancer Cells*

Recent studies have demonstrated that aldo-keto reductase family 1 B10 (AKR1B10), a novel protein overexpressed in human hepatocellular carcinoma and non-small cell lung carcinoma, may facilitate cancer cell growth by detoxifying intracellular reactive carbonyls. This study presents a novel function of AKR1B10 in tumorigenic mammary epithelial cells (RAO-3), regulating fatty acid synthesis. In RAO-3 cells, Sephacryl-S 300 gel filtration and DEAE-Sepharose ion exchange chromatography demonstrated that AKR1B10 exists in two distinct forms, monomers (∼40 kDa) bound to DEAE-Sepharose column and protein complexes (∼300 kDa) remaining in flow-through. Co-immunoprecipitation with AKR1B10 antibody and protein mass spectrometry analysis identified that AKR1B10 associates with acetyl-CoA carboxylase-α (ACCA), a rate-limiting enzyme of de novo fatty acid synthesis. This association between AKR1B10 and ACCA proteins was further confirmed by co-immunoprecipitation with ACCA antibody and pulldown assays with recombinant AKR1B10 protein. Intracellular fluorescent studies showed that AKR1B10 and ACCA proteins co-localize in the cytoplasm of RAO-3 cells. More interestingly, small interfering RNA-mediated AKR1B10 knock down increased ACCA degradation through ubiquitination-proteasome pathway and resulted in >50% decrease of fatty acid synthesis in RAO-3 cells. These data suggest that AKR1B10 is a novel regulator of the biosynthesis of fatty acid, an essential component of the cell membrane, in breast cancer cells.

Aldo-keto reductase family 1 B10 (AKR1B10, 2 also designated aldose reductase-like-1, ARL-1) is a novel protein identified from human hepatocellular carcinoma (1). This protein belongs to the aldo-keto reductase superfamily, a group of proteins implicated in intracellular detoxification, cell carcinogenesis, and cancer therapeutics (2)(3)(4)(5). AKR1B10 is primarily expressed in the colon and small intestine with low levels in the liver, thymus, prostate, and testis (1). However, this gene is overexpressed in 54% of human hepatocellular carcinoma, 84.4% of lung squamous cell carcinoma, and 29.2% of lung adenocarcinoma in smokers, making it a potential diagnostic and/or prognostic marker (1,6,7). AKR1B10 is an enzyme that efficiently catalyzes the reduction of carbonyls to corresponding alcohols with NADPH as a co-enzyme (1). Recent studies demonstrate that AKR1B10 expression facilitates growth of cancer cells, enhances their clonogenic capability, and reduces their susceptibility to reactive carbonyls such as acrolein and crotonaldehyde (8,9). In vitro, AKR1B10 also shows strong enzymatic activity toward all-trans-retinal, 9-cis-retinal, and 13-cis-retinal, reducing them to the corresponding retinols. The diversity of retinal metabolism may diminish intracellular retinoic acid, a signaling molecule regulating cell proliferation and differentiation (4,10).
The current study presents a novel biological function of AKR1B10, regulating long chain fatty acid synthesis, in human breast cancer cells. During tumorigenic transformation of human mammary epithelial cells (HMEC), AKR1B10 is up-regulated and associates with acetyl-CoA carboxylase-␣ (ACCA). ACCA is a rate-limiting enzyme of de novo synthesis of long chain fatty acids, catalyzing the formation of malonyl-CoA by ATP-dependent carboxylation of acetyl-CoA (11,12). Increased lipogenesis is an important characteristic of cancer cells and likely contributes to the development and progression of cancer, but the regulatory mechanisms remain to be elucidated (13,14). Up-regulation of lipogenic enzymes such as fatty acid synthase and ACCA has been documented in a variety of cancers, including breast, prostate, ovary, lung, colon, and endometrial cancers (11, 14 -19). Long chain fatty acids are building blocks of membranes and precursors of lipid second messengers, playing a critical role in cell proliferation and division; ACCA knock-out mice are embryonically lethal (20,21). In prostate and breast cancer cells, RNA interferencemediated silencing of ACCA inhibits fatty acid synthesis, arrests cell cycle, and induces caspase-mediated apoptosis (22)(23)(24). This study identifies AKR1B10 as a novel regulator of fatty acid de novo synthesis, providing a new target for the manipulation of cancer cell growth.
Anion Exchanger Chromatography-Total volume of 2 ml of DEAE-Sepharose (Amersham Biosciences) was packed into a column. RAO-3 cells were lysed in the buffer as described above. Soluble proteins were applied to the column connected to a fast protein liquid chromatography system (Bio-Rad) at 0.25 ml/min. Flow-through was collected. After being washed with 10 column volumes of lysis buffer, binding proteins were eluted with NaCl at a gradient of 10 to 1000 mM and collected at 0.5 ml/fraction.
Co-immunoprecipitation-RAO-3 cells were lysed as above. Soluble proteins (500 g) were incubated with 5 g of anti-AKR1B10 (refer to supplemental data for activity and specificity of AKR1B10 antibody produced in our laboratory) or 5 g of anti-ACCA (Cell Signaling) antibodies at 4°C overnight, followed by incubation with 40 l of slurry-Sepharose protein A/G beads at 4°C for 1 h with gentle shaking. Beads were collected by brief centrifugation and washed five times with lysis buffer (see above). Proteins were eluted with 50 l of 100 mM glycine, pH 2.5, and separated on 8 -12% SDS-PAGE, followed by Coomassie Blue staining or Western blot. Rabbit IgG (5 g) was used as a negative control.
Mass Spectrometry Assay-Co-purified protein by immunoprecipitation was collected from SDS-PAGE. Gel slices were destained and digested in 25 l of sequencing grade trypsin (12.5 ng/l in 25 mM ammonium bicarbonate; G-Biosciences) at 37°C overnight. The digested sample was dried using a Savant SpeedVac concentrator and resuspended in 13 l of 5% acetonitrile containing 0.1% formic acid. 10 l of sample was used for mass spectrometry analysis. A quadruple time-offlight mass spectrometer (Waters DEME-ToF) connected to a Waters nano Acquity UPLC was used for mass spectrometry analysis. Column used was Waters Atlantis C-18 (0.03-mm particle, 0.075 ϫ 150 mm). Flow rate was at 250 nl/min. Peptides were eluted using a linear gradient of water/acetonitrile containing 0.1% formic acid from water to acetonitrile in 60 min. The mass spectrometer was set for data-dependent acquisition; tandem mass spectrometry was performed on the top three peaks at any given time. Data analysis was done using Waters Protein Lynx Global Server 2.2.5, MASCOT (Matrix Sciences) and PEAKS (Bioinformatics Solutions Inc. Waterloo, ON, Canada), and blasted against the NCBI NR data base.
Western Blot Analysis-Western blot was performed as previously described (8). ACCA antibody was used at 1:1000 in blocking buffer (Li-Cor Biosciences).
Recombinant AKR1B10 Protein Preparation and Pulldown Assay-Recombinant AKR1B10 protein with His tag was purified as described previously (8). (Refer to supplemental data for quality of protein.) To perform pulldown assay, 5 g of recombinant AKR1B10 protein was incubated with 500 g of soluble protein (cell lysates) at 4°C for 2 h. Ni-NTA-agarose beads (40 l) were then added and incubated at 4°C for 1 h. After brief centrifugation, Ni-NTA-agarose beads were pelleted and washed with lysis buffer (see above) five times. Proteins were released by heating at 95°C in 5ϫ SDS-PAGE loading buffer for 3 min and subjected to Western blot. A Ni-NTA-agarose bead control was run in parallel.
Transient Transfection and Fluorescent Localization-AKR1B10-EGFP fusion protein expression vector was transfected into RAO-3 cells as previously described (8). After culture for 48 h, cells were fixed in ice-cold methanol for 10 min and then in acetone for 1 min, followed by incubation with ACCA antibody (1:10) at 4°C overnight. Texas Red-labeled secondary antibody (1:1000; GeneTex) was used to visualize ACCA antibody. Dual-color detection by confocal laser scan microscopy (TCS SP2 system; Leica, Bensheim, Germany) was performed for cellular distribution and localization of AKR1B10-EGFP and ACCA proteins.
Fatty Acid Synthesis-Fatty acid synthesis was determined by measuring radiolabeled acetic acid incorporation. In 24-well plates, cells were pulsed with 1 Ci/well of 14 C-labeled acetic acid (Moravek) for 4 h. Cells were trypsinized and washed with phosphate-buffered saline three times. One fourth of the cell suspension was lysed for protein quantitation with protein assay dye (Bio-Rad). The remainder was mixed vigorously with 20 volumes of chloroform/methanol (2:1, v/v). After incubation on ice for 10 min, debris was removed at 21,000 ϫ g for 10 min, and supernatant was washed with 0.2 volume of dH 2 O. Aqueous phase and interface were removed, and organic phase was dried in speed FIGURE 1. Up-regulation of AKR1B10 and ACCA in RAO-3 cells. Soluble proteins (50 g for AKR1B10 or 80 g for ACCA) from RAO-1, RAO-2, and RAO-3 cells were separated on SDS-PAGE (12% for AKR1B10 or 8% for ACCA) and blotted on nitrocellulose membranes (Bio-Rad). Western blot was performed as described under Materials and Methods. ␤-Actin was used as a loading control. Refer to Results for RAO cells. FEBRUARY 8, 2008 • VOLUME 283 • NUMBER 6

AKR1B10 Regulates Fatty Acid Synthesis
vacuum. Lipids were dissolved in 4.0 ml of liquid scintillation mixture, and radioactivity was measured by a scintillation counter (Beckman).
Statistic Analysis-Statistic analysis was performed using Student's t test with the INSTAT statistical analysis package (Graph Pad Software). Significance was defined as p Ͻ 0.05.

RESULTS
AKR1B10 and ACCA are Up-regulated in Tumorigenic HMEC Cells-In a previous study, we generated a panel of HMEC cell models (25). Primary HMEC cells were immortalized by a catalytic subunit of telomerase, generating RAO-1 cells. By introduction of activated H-RasQ61L into RAO-1 cells, RAO-2 and RAO-3 cells were isolated. RAO-2 cells with H-RasQ61L expression show transformational growth but are not tumorigenic in nude mice; RAO-3 cells possess an extra deletion of Rab25 gene besides H-RasQ61L expression and show anchorage-independent growth in soft agar and tumorigenesis in nude mice. In this study, we observed a tumorigenic-related up-regulation of both AKR1B10 and ACCA proteins in RAO-3 cells (Fig. 1).
Two Distinct AKR1B10 Protein Complexes in RAO-3 Cells-To explore the biological function of the induced AKR1B10 protein in the tumorigenic RAO-3 cells, we first examined its biochemical properties. As shown in Fig. 2A, gel filtration chromatography exhibited two AKR1B10 protein peaks with masses of ϳ40 and ϳ300 kDa, respectively. To verify this finding, anion exchange chromatography was conducted using DEAE-Sepharose ion exchange column. In this assay, AKR1B10 protein was present in two distinct portions. A portion of AKR1B10 remained in flow-through while the other bound to the column with elution at 200 -300 mM NaCl (Fig.  2B, upper panel). Reloading the flow-through onto the gel filtration column revealed that AKR1B10 in flow-through represented the peak of ϳ300 kDa (Fig. 2B, lower panel). The AKR1B10 bound to ion exchange column stood for the peak at ϳ40 kDa (data not shown). These data indicate the existence of two distinct AKR1B10 forms in RAO-3 cells, monomers with calculated molecular mass of 36 kDa and complexes with a mass of ϳ300 kDa, that associate with other protein(s).
AKR1B10 Associates with ACCA in RAO-3 Cells-To identify proteins that associate with AKR1B10, immunoprecipitation was performed with a specific anti-AKR1B10 antibody generated in our laboratory. (Supplemental data demonstrate the activity and specificity of AKR1B10 antibody.) After separation on SDS-PAGE, Coomassie Blue staining indicated a candidate protein at ϳ250 kDa, which was not present in the rabbit IgG control (Fig. 3A).   AKR1B10 presence in precipitate was confirmed by Western blot (Fig. 3B). Protein mass spectrometry analysis identified that this candidate protein is ACCA with 265 kDa. Fig. 4 shows an example of mass spectrometry analysis of tryptic digests representing amino acids 244 -266. Interestingly, association with AKR1B10 also altered ACCA binding to DEAE-Sepharose (Fig. 3C). Association of AKR1B10 with ACCA was confirmed by coimmunoprecipitation using ACCA-specific antibody and AKR1B10 protein pulldown assay. In RAO-3 cells, AKR1B10 protein was specifically co-precipitated by ACCA antibody (Fig. 5A) and ACCA was pulled down by histidine-tagged recombinant AKR1B10 protein (Fig. 5B). (Refer to supplementary data for the quality of AKR1B10 recombinant protein.) In addition, intracellular fluorescent studies showed that ectopic AKR1B10-EGFP co-localizes with cellular ACCA protein in the cytoplasm of RAO-3 cells (Fig. 6).

B) Western blot A) Coomassie blue staining
AKR1B10 Affects Fatty Acid Synthesis in RAO-3 Cells by Regulating ACCA Degradation through Ubiquitination-proteasome Pathway-To understand the biological significance of this protein-protein association, we further examined the effect of AKR1B10 knock down on ACCA protein stability and de novo fatty acid synthesis. At first, we checked ACCA protein degradation pathways in RAO-3 cells. Fig. 7A shows that exposing RAO-3 cells to epoxomicin, a proteasome inhibitor, resulted in a dose-dependent accumulation of ACCA protein, indicating its degradation through ubiquitination-proteasome pathway as reported previously (27,28). Thereafter, we downregulated AKR1B10 protein in RAO-3 cells with chemically synthesized siRNA 1 and 2 characterized in our previous studies (9) and examined ACCA protein levels. As shown in Fig. 7B, AKR1B10 knock down caused a notable decrease of ACCA protein levels. This ACCA protein reduction was blocked by proteasome inhibitor epoxomicin, suggesting that by direct association AKR1B10 prevents ACCA degradation through ubiquitination-proteasome pathway. Furthermore, we examined the effect of AKR1B10 knock down on de novo fatty acid synthesis in RAO-3 cells by measuring the incorporation of 14 C-labeled acetic acid. The results showed that AKR1B10 knock down resulted in Ͼ50% decrease of fatty acid syntheses (Fig. 7C). Taken together, these data suggest that AKR1B10 affects fatty acid synthesis via regulating ACCA stability.

DISCUSSION
Lipogenic alterations are early events of cancer development. In many types of cancers, lipogenic genes such as fatty acid synthase and ACCA are up-regulated (14 -16). For instance, fatty acid synthase is overexpressed in the earliest stages of prostate neoplastic transformation (PIN lesions) and expression levels are positively related to the grades of PIN lesions and invasive carcinomas (16,29). Recent studies have demonstrated that in tumor cells newly synthesized lipids are mainly phospholipids, the major components of cell membranes, meeting the need of rapid cell division. More importantly, newly synthesized lipids are enriched with saturated or monounsaturated fatty acids. Saturated fatty acids tend to partition into detergent-resistant membrane microdomains or rafts, which mediate cell migration, signal transduction, and intracellular trafficking (13,14,30,31). In this study, we found that AKR1B10 is  Fig. 3 was identified as acetyl-CoA carboxylase-␣ using matrix-assisted laser desorption ionization mass spectrometry. Ion 1141.6265 (2ϩ) was fragmented by collision-induced dissociation de novo sequencing using the b and y ions confirmed the peptide corresponding to residues 244 -266 (IASSIVAQTAGIPTLPWSGS) for acetyl-CoA carboxylase-␣.

FIGURE 5. Confirmation of AKR1B10-ACCA association in RAO-3 cells.
A, conversely, ACCA antibody was used for co-immunoprecipitation of AKR1B10, and precipitates were detected by Western blot as described in text. Lane 1, immunoprecipitation with ACCA antibody; lane 2, rabbit IgG negative control. Thereafter, AKR1B10 pulldown assay was conducted by incubating 5 g of recombinant AKR1B10 protein (tagged with six histidines) with 500 g of RAO-3 cell lysates to pull down ACCA protein. Protein complexes were collected using Ni-NTA beads (binding to histidines) and subjected to Western blot (B). Lane 1, AKR1B10 pull-down sample; lane 2, Ni-NTA bead control. This functional study of AKR1B10 protein up-regulated in RAO-3 cells started with Sephacyl-S300 gel filtration and DEAE-Sepharose ion exchange chromatography. The results demonstrated that AKR1B10 protein exists in two distinct forms in RAO-3 cells. One (ϳ40 kDa) bound to the DEAE-Sepharose column while the other (ϳ300 kDa) remained in flow-through. The latter is a protein complex with different ion exchange properties (Fig. 2). Using co-immunoprecipitation, pulldown assay, and mass spectrometry analysis, we discovered that ACCA (265 kDa) is the protein associating with AKR1B10 in RAO-3 cells, and intracellularly, ectopic AKR1B10 tagged with EGFP demonstrated a co-localization with ACCA protein. Similar results were observed in HCT-8 cells, a colon adenocarcinoma cell line (data not shown).
ACCA catalyzes the carboxylation of acetyl-CoA, forming malonyl-CoA. It is the rate-limiting step of de novo fatty acid synthesis (12). In normal tissues, ACCA activity is tightly regulated by both metabolite-mediated allosteric and phospho-dependent regulations. The latter is triggered by different hormones, such as insulin, glucagon, and growth factors (32)(33)(34). In MCF7 breast cancer cells, breast cancer 1 (BRCA1) protein, via the BRCT tandem domain at C terminus, specifically associates with ACCA and regulates its enzymatic activity by blocking Ser 79 residue from dephosphorylation, mediating fatty acid synthesis (26,35). To explore the physiological function of AKR1B10-ACCA association, we down-regulated AKR1B10 protein using two siRNAs that target the encoding (siRNA 1) and 3Ј-untranslational (siRNA 2) regions of AKR1B10 mRNA. These two siRNAs were well characterized in our previous studies, resulting in Ͼ60% (siRNA 1) and 90% (siRNA 2) knock down of AKR1B10 protein, respectively (9). In this study, AKR1B10 knock down induced a notable decrease of ACCA protein and subsequent reduction (Ͼ50%) of fatty acid synthesis. The fact that ACCA protein degradation induced by AKR1B10 silencing was blocked by epoxomicin, a proteasome inhibitor, suggests that AKR1B10 binds to ACCA and prevents its degradation through the ubiquitinationproteasome pathway. These data reveal a novel regulatory mechanism of ACCA activity and fatty acid synthesis. Further studies are in progress to identify the functional domains and binding sites in these proteins.
In summary, this study demonstrates for the first time that AKR1B10, induced in tumorigenic HMEC cells, affects fatty acid synthesis by regulating the stability of ACCA protein and thus becomes a novel target for the modulation of de novo fatty acid synthesis in cancer cells. A, ACCA accumulation by epoxomicin (Epo). Cells were treated by epoxomicin, a proteasome inhibitor, for 12 h, and lysates (50 g for AKR1B10 or 100 g for ACCA) were subjected to Western blot as described under Materials and Methods. Epoxomicin induced a dose-dependent accumulation of ACCA protein. B, ACCA reduction induced by AKR1B10 silencing. RAO-3 cells were transiently transfected by two siRNAs as described in text and then exposed or not to epoxomicin at 200 nM for 12 h. Cell lysates were subjected to Western blot. AKR1B10 knock down induced ACCA decrease that was blocked by epoxomicin. Control (Ctrl), scrambled siRNA; R1, siRNA 1; R2, siRNA 2. C, fatty acid synthesis. RAO-3 cells transfected by siRNA 1 and 2 were incubated in 24-well plates for 72 h to trigger AKR1B10 knockdown and then pulsed with 1 Ci/well of 14 C-labeled acetic acid for 4 h. Lipids were extracted as described under Materials and Methods, and acetic acid incorporation was determined by radioactivity (cpm/g protein). Data represent the mean Ϯ S.D. from three independent experiments. *, p Ͻ 0.01 compared with scrambled siRNA control.