Neutrophil Gelatinase-associated Lipocalin (NGAL) Expression Is Dependent on the Tumor-associated Sigma-2 Receptor S2RPgrmc1*

Background: Tumor cell invasion is important in cancer progression, and the sigma-2 receptor (S2RPgrmc1) contributes to invasion. Results: S2RPgrmc1-knockdown cells had diminished levels of NGAL protein and RNA, corresponding with decreased active NFκB. Conclusion: We propose a model in which S2RPgrmc1 elevates NGAL expression via EGFR and NFκB. Significance: S2RPgrmc1 may be a potent target for inhibiting tumor invasion. Tumor invasion is a critical step in the spread of cancer. S2R (sigma-2 receptor)/Pgrmc1 (progesterone receptor membrane component 1) is a cytochrome b5-related drug-binding orphan receptor essential for tumor formation and invasion. Secretory proteins drive these processes, so we screened for S2RPgrmc1-dependent secreted proteins using antibody arrays. S2RPgrmc1 markedly regulated the expression of NGAL/LCN2 (neutrophil gelatinase-associated lipocalin/lipocalin 2), a secreted glycoprotein that binds to MMP-9 (matrix metalloproteinase 9) and protects it from degradation. S2RPgrmc1 knock-down blocked NGAL/LCN2 expression at the protein and RNA levels and decreased MMP9 activity. NGAL expression was required for MMP-9 activity and tumor formation. S2RPgrmc1 associates with EGFR and increases EGFR levels at the plasma membrane, and the EGFR inhibitors erlotinib and AG1478, as well as Akt and ERK inhibitors, suppressed the NGAL/LCN2 RNA and protein levels. NGAL is transcriptionally regulated by NFκB, and S2RPgrmc1 knock-down decreased the NFκB subunit p65/RelA acetylation, phosphorylation, and activation. In S2RPgrmc1 knock-down cells, p65 acetylation was reversed by inhibitors of histone deacetylase 1, and the inhibitors partially restored NGAL levels. Our results are consistent with a model in which S2RPgrmc1 increases NGAL/LCN2 levels by activating NFκB via EGFR.

Cancers spread to distant sites by tumor invasion, which is driven by the combined action of proteases, adhesion factors, and signaling proteins. Neutrophil gelatinase-associated lipocalin/lipocalin 2 (NGAL/LCN2) 2 is an iron-binding protein (1,2) that complexes and stabilizes the matrix metalloproteinase MMP9 (3,4), promoting survival (5,6), and invasion (4,7,8). The NGAL-MMP9 complex is expressed in tumors and is detectable in the blood and urine of cancer patients (3,9). NGAL expression is driven, at least in part, by a pathway consisting of the HER2/neu-phosphotidylinositol 3-kinase-NFB pathway in breast cancer cells (7,10), where it profoundly increases tumor formation and invasion (7). NFB exists in an inactive complex in the cytoplasm, and is activated by acetylation and phosphorylation (11), although the mechanism through which HER2/neu activates NFB to elevate NGAL is unknown.
In the present study, we report that S2R Pgrmc1 drives the transcription of NGAL and the activation of the NGAL-MMP9 complex. We show that NGAL is required for tumor formation in lung cancer cells; NGAL transcription requires EGFR, and both proteins are activated by S2R Pgrmc1 . The results suggest a model in which S2R Pgrmc1 promotes metalloproteinase activity by activating receptor signaling to NFB.

EXPERIMENTAL PROCEDURES
Cell Lines and Treatments-A549 and NCI-H226 cells were obtained from the ATCC, cultured under the suggested conditions, and their identity was verified by Genetica LLC (Cincinnati, OH). Cells were maintained in DMEM containing 10% serum supreme and antibiotics, except where described. The AG-205 inhibitor (26) has been described. Short hairpin RNA (shRNA) lentiviruses (Sigma-Aldrich) containing shRNAs for NGAL (clone D: TRCN0000060540 and clone E: TRCN0000060549) were transduced and selected in puromycin. The Pgrmc1 expression plasmid pRC40 has been described (27), and EGFR was expressed from the plasmid pcDNA3-EGFR, which was a kind gift from Drs. Penni Black of the University of Kentucky and William Pao of Vanderbilt University. Erlotinib (LC Laboratories, Woburn, MA), LY294002 (Sigma), PD98059 (Sigma), AG1478 (Sigma), SAHA (suberoylanalide hydroxamic acid, Biomol, Plymouth Meeting, MA) and sodium butyrate (Sigma) were used as indicated. Conditioned media was generated by incubating cells in serum-free DMEM media and concentrating the media 10-fold using an Amicon Ultracel 10 kDa molecular mass cut-off filter unit (Millipore, Billerica, MA). For xenografts, athymic nude female mice (5-6 weeks old, Harlan Laboratories, Indianapolis, IN) were injected subcutaneously with either A549/con or A549/ shNGAL cells (3 ϫ 10 6 cells in 100 l of PBS) into the upper flank. After 21 days, tumor volume (mm 3 ) was calculated by using the following formula V ϭ (W 2 ϫ L)/2, where W is width (small diameter) and L is length (long diameter).
Invasion was determined by adding 1 ϫ 10 5 A549 cells to the inner chamber of a Matrigel-coated invasion chamber (BD Biosciences, 8 m pore size). 600 l of culture medium containing 5% FBS was added to the bottom chamber, and the cells were incubated for 16 h 37°C in 5% CO 2 . Cells on the upper surface of the inner chamber were then removed, and cells adhering to the lower surface of the membrane were fixed, stained with 1% toluidine blue in 1% Borax and counted.
Protein Analysis-Human soluble receptor arrays were utilized according to manufacturer's instructions. Briefly, membranes were probed with 100 -250 l of conditioned media and detected with chemiluminescence. Quantification of pixels was performed by densitometry using Adobe CS2 software. Western blots were performed using previously published techniques. The antibodies used were NGAL (MAB1757, R&D Sys- (10E2, Cell Signaling) and MMP9 (#3852, Cell Signaling). For zymography, equal amounts of 10-fold concentrated conditioned media were resolved on a 10% Novex Zymogram Gel (Invitrogen, Carlsbad, CA) and developed per the manufacturer's instructions, except that the developing step was increased to 24 -48 h.
Nuclear fractionation was performed using the NE-PER nuclear and cytoplasmic reagent (Thermo Pierce, Fremont, CA). Thermo-Pierce p65/RelA activation was measured using the NFkB p65 chemiluminescent transcription factor assay kit according to the manufacturer's instructions. Nuclear lysates were prepared using the NE-PER reagent, and antibody incubations were for 1 h, as indicated.
RNA Analysis-For RT-PCR, total RNA was isolated by TRIzol Reagent (Invitrogen; 15596-026) according to manufacturer's protocols. cDNA synthesis from 2 g of RNA was carried out using SuperScript II (Invitrogen) with random hexamers as described (23). Semi-quantitative RT-PCR was performed as described (38). Triplicate samples for quantitative PCR were run in an iCycler (Bio-Rad) using the SYBR Green I system (Bio-Rad). ⌬C t for each gene was determined after normalization to ␤-actin, and ⌬⌬C t. was calculated relative to the control. Gene expression values were then expressed as a fold change, calculated by 2 Ϫ⌬⌬ C t. The primer sequences were NGAL-F, 5Ј-

RESULTS
NGAL Expression Is Dependent on S2R Pgrmc1 -S2R Pgrmc1 promotes tumor cell migration and metastatic colonization (26), and we probed conditioned media from S2R Pgrmc1 -knockdown A549 NSCLC cells (26) for secreted proteins associated with invasion. Using two arrays that included 81 embedded antibodies to secreted proteins, we detected numerous altered proteins (supplemental Fig. S1). NGAL was essentially absent in S2R Pgrmc1 -knockdown cells (Fig. 1A, plotted in Fig. 1B), and loss of NGAL in S2R Pgrmc1 -knockdown cells was confirmed by Western blots in conditioned media ( Fig. 1C) and cell lysates (Fig. 1D). NGAL binds covalently to the MMP9 matrix metalloproteinase, forming a 125 kDa complex (3). In S2R Pgrmc1knockdown cells, the 125 kDa complex decreased 8.5-fold in zymography gels containing gelatin (Fig. 1E), while the lower MMP9 band decreased 4.1-fold. MMP9 levels were essentially unchanged in the same conditioned media (Fig. 1E, middle panel) and lysates (Fig. 1F) from S2R Pgrmc1 -knockdown cells, suggesting that decreased MMP9 activity in media is not due to changes in MMP9 levels.
The shRNA targeting S2R Pgrmc1 binds a sequence in the 3Ј-untranslated region of the S2R Pgrmc1 transcript, and we expressed an exogenous S2R Pgrmc1 transcript lacking this sequence. Exogenous S2R Pgrmc1 restored NGAL levels in conditioned media ( Fig. 2A, lane 4) to levels approximating the control cells ( Fig. 2A, lanes 1 and 2). NGAL expression also decreased after treatment with AG-205, an S2R Pgrmc1 small molecule inhibitor (26, 28) in a dose-dependent manner (Fig.  2B). MMP9 was unchanged (Fig. 2B, third panel). While we focused this study primarily on Ngal, other proteins were altered in the condition media, including activated cathepsin D (Fig. 2C), although Timp2 was unchanged (Fig. 2C).
One potential mechanism governing NGAL levels in S2R Pgrmc1 -knockdown cells is the activation of a general stress response. However, levels of the stress response genes p53, calnexin and hsp90 (39 -41) were unchanged in S2R Pgrmc1knockdown cells (Fig. 2D). Instead, the decline in NGAL protein levels was reflected in decreased NGAL RNA levels in S2R Pgrmc1 -knockdown cells by RT-PCR (Fig. 3A) or real-time quantitative PCR (Fig. 3B). Decreased NGAL RNA levels were also detected after expression of a heme-binding-deficient (S2R-hbd) S2R Pgrmc1 mutant (Fig. 3C).
NGAL Requirement for Tumor Formation-NGAL biological activity has not been reported previously for lung cancer. We inhibited NGAL expression (Fig. 4A) by introducing an NGALtargeting shRNA into A549 NSCLC cells. Zymography of conditioned media from control and NGAL-knockdown cells indicated that both NGAL-MMP9 and MMP9 activity were decreased (26-fold and 52-fold, respectively, Fig. 4B). A 65 kDa band, which could be a cleaved form of MMP9 or uncleaved MMP2, also decreased 11-fold but was detectable in the NGALknockdown cells (Fig. 4B, MMP2/9). Tumor xenograft formation was sharply impaired in the mice infused with NGALknockdown cells (p ϭ 0.002, t test, Fig. 4C), indicating an essential role for NGAL in lung tumor growth. NGAL also promoted motility in A549 cells in a modified Boyden chamber assay (p ϭ 0.0002, t test, Fig. 4D), consistent with its function in regulating metalloproteinases.
NFB has been previously implicated in NGAL transcription, and p65-RelA is activated by acetylation and phosphorylation and inhibited by histone decetylases or HDACs (11,(42)(43)(44). Acetylated p65/RelA (Ac-RelA) decreased 8-fold in the nuclear fraction of S2R Pgrmc1 -knockdown cells (Fig. 6A, top  panel), while p65-phosphoS536 levels diminished by 5.7-fold (Fig. 6A, second panel) and total nuclear p65 by 3.9-fold (Fig.  6A, third panel). In contrast, the nuclear marker E2F5 was unchanged (Fig. 6A, lower panel). The nuclear and cytoplasmic protein separation was validated with markers for each fraction (supplemental Fig. S4). The results are consistent with a model in which p65 phosphorylation, acetylation, and nuclear transport are partially dependent on S2R Pgrmc1 . We detected a similar decrease in acetylated p65 in NCI-H226 cells following infection with the Ad-S2R-hbd mutant (supplemental Fig. S3C).
Because HDACs decrease NFB acetylation, we posited that the HDAC inhibitor SAHA would have the opposite effect. Indeed, control cells treated with SAHA exhibited increased secreted NGAL (Fig. 6B, lane 2), and SAHA restored NGAL to near basal levels in S2R Pgrmc1 -knockdown cells (Fig. 6B, lane 4). Furthermore, NGAL transcription increased following treatment with the HDAC inhibitor sodium butyrate in control and shPGR cells by 2.2-and 8-fold, respectively (Fig. 6C, lanes 2 and  5). SAHA treatment slightly repressed NGAL transcription in control cells (Fig. 6C, lane 3) but induced NGAL transcription by 4-fold in S2R Pgrmc1 -knockdown cells. As expected, SAHA treatment had little effect on p65 acetylation in control cells (Fig. 6D, lane 2) but increased p65 acetylation by 2.8-fold in S2R Pgrmc1 -knockdown cells (Fig. 6D, lane 4). In contrast, E2F5 levels decreased slightly after SAHA treatment and were  unchanged in control and S2R Pgrmc1 -knockdown cells (Fig. 6D,  lower panel). The results suggest that defects in p65/RelA activation in S2R Pgrmc1 -knockdown cells contribute to loss of NGAL transcription.

DISCUSSION
S2R Pgrmc1 is elevated in lung cancers and contributes to tumor growth, metastasis and invasion in lung cancer cells. In the present study, we demonstrate that S2R Pgrmc1 elevates the transcription and protein levels of NGAL, a secreted glycoprotein that complexes with the MMP9 metalloproteinase, stabilizing the NGAL-MMP9 complex. NGAL also binds to iron (2), and the yeast S2R Pgrmc1 homologue is regulated by iron (45) and is essential for iron storage (24). Our results suggest a model in which S2R Pgrmc1 activates NGAL transcription. The role of NGAL in promoting or suppressing tumor growth is highly dependent on the tissue of origin, and our results support a role for NGAL in cancer cell invasion and tumor formation in lung cancer.
In addition to increasing NGAL expression, S2R Pgrmc1 elevated NGAL-MMP9 activity. Surprisingly, uncomplexed MMP9 activity was also increased by S2R Pgrmc1 (Fig. 1), even though MMP9 levels were largely unchanged in S2R Pgrmc1knockdown cells. One potential mechanism is the induction of Timp1, -2, and -3 expression in S2R Pgrmc1 -knockdown cells (supplemental Fig. S1), which can inhibit MMP9 activation (46). However, Timp2 was not induced in conditioned media by Western blot (Fig. 2), suggesting that the proteome array results may be tightly dependent on specific epitopes recognized by the antibodies embedded in the arrays. Alternately, MMP9 can be activated by multiple proteases (46), and any of them could be altered by loss of S2R Pgrmc1 . Loss of S2R Pgrmc1 reduced activated cathepsin D (Fig. 2), which in turn suggests alterations in cathepsins L and B (47). Interestingly, cathepsin G is important in MMP9 processing (46), and one potential model is that multiple cathepsins are activated by S2R Pgrmc1 . This model, which is purely speculative, is currently under investigation. The results are important because MMP9 plays a key role in tumor growth and progression (48), and NGAL is a potential mediator of tumor invasion by stabilizing MMP9.
One enigmatic finding in the present study is that S2R Pgrmc1 has little effect on MMP9 levels ( Fig. 1), while NGAL expression increased MMP9 levels (Fig. 4). In the simplest sense, the expected result is that both S2R Pgrmc1 and NGAL inhibition should have the same phenotype, decreased MMP9, reflecting a common pathway. The results suggest that S2R Pgrmc1 may have NGAL-independent functions, increasing MMP9, that are uncharacterized.
Based on previously published studies, we propose that the S2R Pgrmc1 -EGFR complex (28) drives NGAL expression. According to this model, NFB p65/RelA, an essential protein in K-ras-driven lung cancer (49), is subsequently phosphorylated, acetylated and transported to the nucleus, driving NGAL transcription (Fig. 7). One caveat to this model is that the S2R Pgrmc1 -knockdown affected NGAL levels more potently than EGFR inhibitors (Fig. 5), suggesting that S2R Pgrmc1 may utilize pathways other than EGFR in activating NGAL. Exogenous EGFR expression restored NGAL levels (Fig. 5), but only partially, likely because of the important role of S2R Pgrmc1 in stabilizing EGFR pools at the plasma membrane (28). The  2 and 4). The blot was probed for NGAL and showed a partial restoration of NGAL expression in S2R Pgrmc1 -knockdown cells following EGFR expression. mechanism linking S2R Pgrmc1 and EGFR to p65/RelA includes Akt (for maintaining nuclear, acetylated p65/RelA and for activation) and ERK activation (for nuclear, acetylated p65/RelA, Figs. 5 and 6), but does not include altered HDAC1 expression (Fig. 6D). An alternate model is that HDAC1 can be activated by non-EGFR metabolic pathways. For example, S2R Pgrmc1 ligands have been implicated in oxidative stress induction (50), suggesting that S2R Pgrmc1 normally increases oxidative stress. In turn, oxidative stress inhibits HDAC activity (51), which is predicted to elevate p65/RelA acetylation. The role of S2R Pgrmc1 in oxidative stress is a subject of ongoing investigation.
The significance of the findings is that tumor invasion is a hallmark of aggressive cancers and is driven, at least in part, by MMPs. Therapeutic targeting of MMPs has had mixed results (48), and the current findings suggest a new approach for targeting MMP-NGAL complexes. Indeed, a small molecule S2R Pgrmc1 inhibitor suppressed NGAL expression, and other S2R ligands have activity against cancer cells (29,52) and may indirectly inhibit MMPs. Our results suggest that strategies inhibiting S2R Pgrmc1 could potentially target MMP9 activity in lung cancer and perhaps other cancer types.  6) for 72 h. NGAL/actin ratios, relative to vehicle-treated control cells, are shown between the two panels. D, nuclear lysates from control (lanes 1 and 2) and S2R Pgrmc1 -knockdown cells (lanes 3 and 4), untreated (lanes 1 and 3) or treated with the HDAC inhibitor SAHA (lanes 2 and 4), were analyzed by Western blot for acetylated p65. The lower panel shows loading controls for the same samples. E, nuclear lysates from A549 cells treated with vehicle (odd lanes) or 10 M of the EGFR inhibitor erlotinib (lane 2), 20 M of the ERK inhibitor PD98059 (lane 4) or 20 M of the Akt inhibitor LY294002 (lanes 2 and 4), were analyzed by Western blot for acetylated p65. The lower panel shows a loading control (E2F5) for the same samples. The numbers between the panels indicate the ratio acetylated p65 in treated samples divided by the untreated sample, relative to the loading control [(p65 treated / p65 vehicle )/E2F5 control ]. F, inhibition of activated p65/RelA levels by erlotinib, AG1478, and LY294002 (at the concentrations indicated in panel E). Surprisingly, PD98059 did not have a significant activity in this assay.