Tyrosine phosphatase SHP2 regulates the expression of acyl-CoA synthetase ACSL4.

Acyl-CoA synthetase 4 (ACSL4) is implicated in fatty acid metabolism with marked preference for arachidonic acid (AA). ACSL4 plays crucial roles in physiological functions such as steroid synthesis and in pathological processes such as tumorigenesis. However, factors regulating ACSL4 mRNA and/or protein levels are not fully described. Because ACSL4 protein expression requires tyrosine phosphatase activity, in this study we aimed to identify the tyrosine phosphatase involved in ACSL4 expression. NSC87877, a specific inhibitor of the tyrosine phosphatase SHP2, reduced ACSL4 protein levels in ACSL4-rich breast cancer cells and steroidogenic cells. Indeed, overexpression of an active form of SHP2 increased ACSL4 protein levels in MA-10 Leydig steroidogenic cells. SHP2 has to be activated through a cAMP-dependent pathway to exert its effect on ACSL4. The effects could be specifically attributed to SHP2 because knockdown of the phosphatase reduced ACSL4 mRNA and protein levels. Through the action on ACSL4 protein levels, SHP2 affected AA-CoA production and metabolism and, finally, the steroidogenic capacity of MA-10 cells: overexpression (or knockdown) of SHP2 led to increased (or decreased) steroid production. We describe for the first time the involvement of SHP2 activity in the regulation of the expression of the fatty acid-metabolizing enzyme ACSL4.

Because of this background, we postulate that SHP2 could be the tyrosine phosphatase necessary for ACSL4 protein induction. The aim of this study was to analyze the involvement of SHP2 in the regulation of ACSL4 protein levels. Here, we introduce the notion that there are SHP2 -determined high levels of ACSL4 protein in MDA-MB231 cells and that SHP2 action stimulates proliferation of these cultured breast cancer cells. In addition, using plasmidmediated gene transfer and RNAi-mediated gene silencing, we provide evidence for the fi rst time that SHP2 has an obligatory role in steroid production by MA-10 cells, a mouse Leydig cell line, regulating the expression of ACSL4 protein, the rate-limiting enzyme in the AA generation/export system.

Cell cultures
MA-10 mouse Leydig tumor cells were generously provided by Dr. Mario Ascoli (University of Iowa, Iowa City, IA) and cultured in Waymouth MB/752 medium supplemented with 15% horse serum and antibiotics at 37°C and 5% CO 2 as previously described ( 31 ). The culture medium was replaced with serum-free Waymouth medium before experiments.
The human breast cancer cell line MDA-MB-231 was generously provided by Dr. Vasilios Papadopoulos (Research Institute of the McGill University Health Centre, Montreal, Canada) and obtained from the Lombardi Comprehensive Cancer Center (Georgetown University Medical Center, Washington, DC). The cell line was maintained in DMEM medium supplemented with 10% FBS and antibiotics at 37°C and 5% CO 2 as previously described ( 17 ).
Interstitial testicular cells were obtained from adult Wistar male rats. The animals were purchased from the animal facility of the School of Pharmacy and Biochemistry, University of Buenos Aires. The studies were approved by the Institutional Committee for the Care and Use of Laboratory Animals (CICUAL) of the School of Medicine, University of Buenos Aires. Freshly isolated interstitial cells were prepared by collagenase dispersion of testes as previously described ( 32 ). Cells were resuspended in Medium 199 containing 1.2 g/l NaHCO 3 , and 0·5% BSA, at a density of 10 6 cells/ml and were maintained under a carbogen (95% O 2 :5% CO 2 ) atmosphere throughout the procedures.

Plasmids and sequences
Mammalian expression vectors encoding the wild-type and the catalytically inactive SHP2 mutant, Addgene plasmids 8381 and step of the steroidogenic biosynthetic pathway. Regarding the function of ACSL4 in other cell types, sustained downregulation of ACSL4 results in markedly reduced PGE2 release in human arterial smooth muscular cells ( 15 ), indicating that endogenous ACSL4 plays an important role as a regulator of eicosanoid synthesis and secretion, which might regulate smooth muscle cell proliferation, release of infl ammatory mediators, or other processes in the vascular wall ( 15 ). ACSL4 also contributes to the development of human hepatocellular carcinoma, adenocarcinoma, and breast cancer ( 9,10,16,17 ). ACSL4 is signifi cantly upregulated in highly aggressive breast cancer cell lines ( 10,17 ). The effect of ACSL4 on cell proliferation, invasion, and migration is attributed to increased production of lipoxygenase and cycloxygenase metabolites ( 17 ), which are known to potentiate tumor aggressiveness in hepatocellular carcinoma ( 8 ), colon adenocarcinoma ( 9 ), and breast cancer ( 10,17 ). In summary, differences in the abundance and/or activity of ACSL4 may result in variations in the cellular content of free AA and AA-CoA, as well as in lipoxygenated and cycloxygenated metabolites. These compounds in turn play critical roles in steroid production in steroidogenic systems and in cell proliferation, migration, and invasion in cancer progression.
The cAMP-dependent action of ACTH and LH/CG on ACSL4 described above is mediated by the activity of protein tyrosine phosphatases (PTP) ( 18 ). The use of compounds that inhibit PTP enzymatic activity results in a decrease in cAMP-stimulated ACSL4 protein levels, in cAMP-stimulated steroid production (19)(20)(21)(22), and in cAMPstimulated tyrosine dephosphorylation of paxillin and focal adhesion disassembly ( 23 ). Although cAMP-dependent stimulation of steroidogenic cells increases the activity of several PTPs, the PTP(s) responsible for the increase in ACSL4 protein levels and steroid production is (are) still not identifi ed. In this work, we focused on an 80 kDa hormone-activated PTP recognized by commercially available antibodies against src homology 2-containing phosphotyrosine phosphatase 2 ( SHP2), a member of the nonreceptor PTP subfamily. In addition, SHP2 is phosphorylated in vitro by cAMP-dependent protein kinase (PKA) and in vivo by ACTH in bovine adrenal cells ( 24 ). In both cases, phosphorylation increases SHP2 activity. However, no evidence is available on the relationship between this activation and the steroidogenic function of the cell. SHP2 is an ubiquitously expressed nontransmembrane PTP that plays an essential role in many organisms, from lower eukaryotes to mammals ( 25,26 ). Interestingly, SHP2 is one of the PTPs that promote the activation, rather than the downregulation, of intracellular signaling pathways ( 27,28 ), serving multiple hormone receptors and participating in the regulation of mitogenic signaling, cell adhesion, and cell migration ( 28 ). In addition, accumulating evidence suggests that SHP2 might play a pivotal role in carcinogenesis. It is upregulated in breast cancer cells and in approximately 70% of infi ltrating ductal carcinoma of the human breast, implying its involvement in breast oncogenesis ( 29 ). In contrast, inhibition of this PTP leads to reversion from mesenchymal to a normal breast epithelial phenotype ( 30 ). cells) were pretreated with or without NSC87877 (50 µM) for 30 min and then incubated in the presence or absence of 8Br-cAMP 0.5 mM for 3 h at 37°C with gentle shaking. Testosterone production was determined in the incubation media by RIA ( 21 ). At the end of all experiments, MA-10, MDA-MB-231, and testicular interstitial cellular protein content were assayed by the Bradford method using BSA as standard ( 34 ).

PTP assay
SHP2 PTP activity was monitored by an in-gel PTP assay using [ 32 P]poly(Glu:Tyr) as substrate, with minor modifi cations ( 21 ) of the original description ( 35 ). Proteins exhibiting PTP activity are visualized as clear bands on a dark background due to removal of inorganic [ 32 P]phosphate from the radiolabeled substrate incorporated into the matrix gel ( 21,35,36 ).

Western blotting
SHP2 and ACSL4 protein levels were analyzed in a total cell lysate, and StAR protein levels were analyzed in the mitochondrial fraction. The primary antibodies used were purifi ed monoclonal mouse antibody anti-SHP2 (BD-transduction laboratories), polyclonal rabbit antibody anti-ACSL4 ( 13, 37 ) (obtained at our laboratory), rabbit antisera anti-StAR (generously provided by Dr. Doug Stocco), anti-␤ -tubulin (used as cell lysate loading control, Millipore), and anti-Ox-Phos complex III core 2 subunit (used as mitochondrial loading control, Invitrogen). The secondary antibodies were goat anti-mouse or goat anti-rabbit immunoglobulins (according to the primary antibody used) conjugated to peroxidase. The bound primary antibodies were detected using an enhanced chemiluminescence kit (GE Healthcare). The levels of SHP2, ACSL4, StAR, ␤ -tubulin, and complex III were quantitated using a computer-assisted image analyzer (Im-ageQuant 5.2), and the results for each sample were normalized by ␤ -tubulin or complex III.

RNA extraction and semiquantitative RT-PCR
Total RNA from the different treatment groups was extracted using TriReagent following the manufacturer's instructions (MRC, Inc.). The reverse transcription and PCR analyses were made using 2 µg of total RNA. The cDNAs generated were further amplifi ed by PCR under optimized conditions using the primer pairs below. For amplifi cation of Acsl4 (amplicon size 420 bp), the sense primer, 5 ′ -GCCCACTTCAGACAAACCTGG-3 ′ , and the antisense primer, 5 ′ -ACAGCTTCTCTGCCAAGTGTGG-3 ′ , were used. For amplifi cation of StAR (amplicon size 566 bp), the sense primer, 5 ′ -GGGACGAAGTGCTAAGTAAGATGG-3 ′ , and the antisense primer, 5 ′ -GGTCAATGTGGTGGACAGTCC-3 ′ , were used. For amplifi cation of L19 ribosomal protein, used as housekeeping gene (amplicon size 405 bp), the sense primer, 5 ′ -GAAATCGCCAATGCCAACTC-3 ′ , and the antisense primer, 5 ′ -TCTTAGACCTGCGAGCCTCA-3 ′ ( 38 ) were used. The reaction conditions were as follows: for Acsl4 and L19 , one cycle of 94°C for 5 min, followed by 23 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 45 s, 94°C for 5 min, and one cycle of 72°C for 10 min; for StAR , one cycle of 94°C for 5 min, followed by 26 cycles of 94°C for 30 s, 50°C for 30 s, 72°C for 1 min, and one cycle of 72°C for 10 min. The number of cycles was optimized for each gene to fall within the linear range of PCR amplifi cation. PCR products were resolved on a 1.5% agarose gel containing 0.5 µg/ml of ethidium bromide to determine the molecular sizes of the Acsl4 , StAR , and L19 amplicons. To compare the amounts of amplifi ed Acsl4 and StAR produced from different RNA samples, the amplifi ed L19 product of each sample was used as an internal standard. The gel images were acquired with the GelPro analyzer (IPS). The levels of Acsl4 , StAR , and L19 mRNA were quantitated 8382 (referred to as WT-SHP2 and CS-SHP2, respectively), were obtained from Addgene. CS-SHP2 is rendered catalytically inactive by the C459S mutation in its catalytic site. Both forms were obtained in the pRC/CMV plasmid. The amplifi cation procedure was verifi ed by nucleotide sequencing (Macrogen), and the obtained sequence was compared with that deposited in Genbank [ Homo sapiens protein tyrosine phosphatase, nonreceptor type 11 (PTPN11), mRNA, accession number NM_002834].
To obtain the knockdown plasmid, we used a 19-bp DNA fragment of the mouse SHP2 (nucleotides 279-297) ( 33 ) in the adequate frameshift to generate a short hairpin RNA (shRNA), which the cells process to generate a functional and active siRNA directed against the murine SHP2. A BLAST search confi rmed that the sequence specifi cally recognizes mouse SHP2. The insert was cloned in the pSUPER.retro.puro vector (Oligoengine). The obtained SHP2 shRNA construct was verifi ed by nucleotide sequencing (Macrogen), and the obtained sequence was compared with that deposited in Genbank [ Mus musculus protein tyrosine phosphatase, nonreceptor type 11 (Ptpn11), transcript variant 1, mRNA, accession number NM_011202]. Nontargeting control shRNA (named scrambled shRNA) consists of the same 19 nucleotides in a scrambled sequence, which does not target any known mouse cellular mRNA.
Overexpression and knockdown effi ciency was confi rmed by Western blot analysis using specifi c antibodies against SHP2. Overexpression of endogenous SHP2, monitored by immunoblot analysis with anti-SHP2 antibody, was confi rmed in MA-10 transfected with the mammalian expression vectors 24 h posttransfection. Downregulation by shRNA occurred 48 h posttransfection of the pSUPER-retro plasmid.
Transfection effi ciency for enhanced green fl uorescent protein (EGFP) ranged from 30 to 45%. Neither up-nor downexpression of SHP2 affected cell viability, as assessed by the MTT assay. These results indicate that the transfection procedures do not affect mitochondrial function. Moreover, staining of the cells with the nuclear dye (DAPI) showed no change in nuclear morphology, indicating that none of the treatments induced apoptosis in MA-10 cells (data not shown).

Cell treatments
MA-10 cells were pretreated with or without NSC87877 (50 µM) for 3 h and then incubated in the presence or absence of 8Br-cAMP at the submaximal concentration of 0.5 mM for 1 h, with or without addition of AA (200 µM). Transfected MA-10 cells were incubated in the presence or absence of 8Br-cAMP at the submaximal concentration of 0.5 mM for the indicated times (see fi gure legends), with or without addition of AA (200 µM). AA treatments were performed in serum-free medium with 1% BSA fatty acid-free for 30 min prior to the stimulation. At the end of all experiments, progesterone production was determined in the culture media by RIA ( 21 ). MDA-MB-231 was treated with or without addition of PAO (1 µM) or BPA-(AM 2 ) (0.25 mM) (cell permeant derivative of the PTP inhibitor BPA) or NSC87877 (50 µM) for the indicated times (see fi gure legends). When steroid production was evaluated, testicular interstitial cells (10 6 cell line ( Fig. 1A ). NSC87877 also showed an inhibitory effect on ACSL4 protein levels ( Fig. 1A ), directly involving SHP2 in the induction of this acyl-CoA synthetase. Next, we measured the proliferation of MDA-MB-231 cells in the presence of NSC87877, which, because of the reduction in ACSL4 protein levels, should affect AA generation and metabolism. SHP2 inhibition with this compound resulted in a signifi cant decrease in the number of MDA-MB-231 cells in the cultures at the end of the treatment, as determined by the MTT ( Fig. 1B ) and 8BrdU ( Fig. 1C ) assays.
We next used MA-10 Leydig cells to investigate the relationship between SHP2 and steroidogenesis. Cell cultures were treated with or without 8Br-cAMP (0.5 mM) in the presence or absence of NSC87877 (50 µM). ACSL4 protein level and progesterone production were determined under SHP2 inhibition. As previously described ( 11 ), we observed an increase in ACSL4 protein levels after the action of 8Br-cAMP ( Fig. 1D ). The SHP2 inhibitor abolished the increase in ACSL4 protein levels ( Fig. 1D ). Regarding steroidogenesis, cAMP-stimulated progesterone production was reduced by the use of NSC87877 ( Fig. 1E ).
To demonstrate that the effect of NSC87877 on steroidogenesis is due to the reduction in ACSL4, we exogenously added AA. We previously published that the cAMP-dependent signal transduction pathway activates tyrosine phosphatases upstream of ACSL4/Acot2 action ( 18 ) ACSL4 is an enzyme involved in the generation and export of mitochondrial AA ( 11,17,42 ). ACSL4 converts AA into the CoA ester, which is then the substrate for Acot2, a mitochondrial thioesterase. Because of the sequential action of these two enzymes, there is a compartmentalized generation of AA ( 12,13 ) and metabolism through the lipoxygenase pathway ( 17 ). Therefore, the addition of AA should by pass the inhibition of SHP2 on steroidogenesis. As was already known, AA increased steroid production by itself and with cAMP ( 42 ). Together with NSC87877, AA blunted the inhibition exerted by NSC87877 on cAMP-stimulated progesterone production ( Fig. 1E ). In conclusion, the result obtained with the exogenous addition of AA indicates that the inhibition observed with NSC87877 reinforces the notion that SHP2 would be acting on AA generation because of its inducing action on ACSL4.
Because we used cell lines, we questioned whether the observed effects also occurred in a primary cell culture. Thus, we tested the effect of SHP2 inhibition on Leydig cells by incubation of interstitial rat testis cells with NSC87877, followed by evaluation of ACSL4 protein level and testosterone production upon stimulation by 8Br-cAMP. The 8Br-cAMP treatment caused an increase in ACSL4 protein levels ( Fig. 1F ) and testosterone production ( Fig. 1G ), as expected. In the presence of the SHP2 inhibitor, there were lower levels of ACSL4 protein and testosterone production after cAMP stimulation. These results confi rmed the conclusions we reached using cell lines and validated using MA-10 Leydig cells as experimental model throughout the study.
As previously reported, ACSL4 was detected as a duplet ( 1,17,45,46 ). This duplet may be seen in the three cell using a computer-assisted image analyzer (ImageQuant 5.2), and the PCR results for each sample were normalized by L19 mRNA.

Cell proliferation assay
Cell proliferation was measured by the MTT and 5-bromo-2'deoxyuridine assays.
The MTT assay, based on MTT reduction by the mitochondrial diaphorase enzyme, is an index of mitochondrial integrity ( 39 ). It was performed as previously described ( 17,40 ). DNA synthesis was monitored by a colorimetric immunoassay based on the measurement of 5-bromo-2'-deoxyuridine (BrdU) incorporation (BrdU cell proliferation ELISA kit, Roche Diagnostics). The assay was performed according to the manufacturer's instructions. ]AA (New England Nuclear; specifi c activity 53.0 mCi/mmol) was incorporated (0.5 µCi/ml in serum-free Waymouth MB/752 containing 0.5% essentially fatty acid-free BSA) during 3 h at 37°C. The cells were incubated in the presence or absence of 0.5 mM 8Br-cAMP for 1 h, and cell lysates were obtained using a lysis buffer containing 20 mM Tris, pH 7.4, 0.5% Triton X-100, 1 mM EGTA, 1 mM EDTA, and 130 mM NaCl. Lipid extraction was performed twice with ethyl acetate (six volumes of solvent per one volume of cellular fraction) and centrifugation at 800 g for 5 min. The organic phase was discarded, and the [ 14 C]AA-CoA formation was evaluated by extraction from the aqueous phase as previously described ( 42,43 ).

Statistical analysis
All experiments were run in duplicate and repeated at least three times. The reported data are means ± SEM of the three (or more) independent experiments. As appropriate, statistical analysis was performed by Student t -test (comparison of two data sets) or one-way ANOVA. The latter was followed by posttests if P < 0.05 by ANOVA, using the Instat software package from Graph-Pad, Inc.

Effect of inhibition of SHP2 on ACSL4 protein levels and cell proliferation in MDA-MB-231 breast cancer cells and on ACSL4 protein levels and steroidogenesis in MA-10 Leydig cells and testicular rat interstitial cells
To test whether SHP2 could be the PTP involved in the regulation of ACSL4 protein expression, we evaluated cell functions under SHP2 inhibition using NSC87877, a potent inhibitor of this PTP ( 44 ).
MDA-MB-231 cells derive from a highly aggressive human breast tumor and express high levels of ACSL4 mRNA and protein ( 10,17 ). We used MDA-MB-231 cells to investigate the relationship between SHP2 and ACSL4 and cell proliferation. We fi rst evaluated ACSL4 protein levels in cells where PTPs were inhibited. To this end, cell cultures were treated with or without PAO (1 µM), BPA-(AM 2 ) (0.25 mM), or NSC87877 (50 µM). While PAO and BPA act through two different mechanisms on all members of the PTP family, NSC87877 is a specifi c inhibitor of SHP2. The inhibition of PTPs with PAO or BPA reduced ACSL4 protein levels in breast cancer cells of the MDA-MB-231 To analyze the effect of modifying SHP2 protein levels on the fi nal response of steroidogenic stimulation, we collected media from cells transfected with the constructs described above and determined steroid secretion. Overexpression of WT-SHP2 signifi cantly increased cAMPstimulated progesterone production compared with mock-transfected cells ( Fig. 2E ), whereas the expression of the catalytically inactive form caused a consistent and repeated slight decrease (but not statistically signifi cant) in cAMP-stimulated steroidogenesis. The effect could be attributed specifi cally to SHP2, as MA-10 cell cultures transfected with SHP2-targeting shRNA, which downexpresses SHP2, signifi cantly decreased cAMP-stimulated progesterone production compared with empty vector or scrambled shRNA transfections ( Fig. 2F ). Because empty vector or scrambled shRNA-transfected cells exhibited similar SHP2 protein levels and basal and cAMP-stimulated steroid production, both may be used as mock transfections.

Effect of overexpression and knockdown of SHP2 on ACSL4 expression in MA-10 Leydig cells
ACSL4 expression was analyzed in both conditions of SHP2 expression levels and the results are shown in Fig. 3 . Mock transfection shows the already known increase in ACSL4 protein level under cAMP stimulation. WT-SHP2 overexpression resulted in a signifi cant increase in ACSL4 protein levels as compared with mock-transfected cells, an effect potentiated by cAMP, while the levels of ACSL4 were unaffected by CS-SHP2 transfection. Furthermore, CS-SHP2 overexpression blunted the effect of cAMP on ACSL4 protein expression ( Fig. 3A ).
The expression of ACSL4 was also analyzed in cells where the expression of SHP2 was down-regulated. The reduced expression of SHP2 resulted in a signifi cant decrease in the mRNA and protein levels of the synthetase ( Fig. 3B-D ). Acsl4 mRNA level was increased in MA-10 Leydig cells treated with cAMP, an effect impaired by SHP2 shRNA ( Fig. 3B ). As in the previous experiment, cAMP increased ACSL4 protein expression in MA-10 Leydig cells, whereas SHP2 knockdown by shRNA abolished the effect, compared with mock ( Fig. 3C ) or scrambled shRNA ( Fig. 3D ) transfections. The decrease in ACSL4 protein levels in SHP2-down-regulated cells should also be refl ected on the synthetase activity. To determine the endogenous activity of ACSL4, we measured AA-CoA production in MA-10 cells labeled with [1-14 C] AA. The cell cultures were labeled with radioactive AA, which, once incorporated, is the substrate of the endogenous ACSL4 that renders AA-CoA. The radioactive form of this derivative was determined after organic extraction. We reproduced the previously published stimulatory effect of cAMP on the content of the CoA derivative of AA ( 42 ), an effect impaired in SHP2-lacking cells ( Fig. 3E ).
We also analyzed the effect of SHP2 expression on ACOT2 protein levels, the second enzyme involved in AA generation ( 13 ). The results indicate that neither the overexpression nor the silencing of SHP2 affected ACOT2 protein levels (data not shown). types used. However, the difference between the two bands is better seen in MDA-MB-231 cells ( Fig. 1A ) than in MA-10 cells ( Fig. 1B ) or freshly isolated interstitial testis cells ( Fig.  1F ), where the electrophoresis conditions have to be fi nely adjusted to evidence the double band. Both alternative splicing and posttranslational modifi cation mechanisms can explain the detection of a duplet by Western blots. The gene encoding ACLS4 shows alternative splicing ( 45,46 ), and the protein is possibly acetylated. It was also suggested that the different mobility is due to proteolisis or posttranslational modifi cations, such as N-glycosilation ( 46 ).
The results obtained with NSC87877 assured us that SHP2 could be one of the PTPs involved in the increase in ACSL4 protein levels. This fi rst approach encouraged us to further analyze the effect of the modifi cation of SHP2 expression levels on the regulation of ACSL4 mRNA and/ or protein expression and function in one of the cell types, the steroidogenic MA-10 Leydig cells. For that purpose, we used constructions with sequences that either upregulate or downregulate the amount of the endogenous enzyme.

Overexpression and knockdown of SHP2 in MA-10 Leydig cells
We fi rst overexpressed SHP2 in MA-10 cells by means of a transient transfection approach using a pRC/CMV plasmid containing WT-or CS-SHP2 cDNA. The transfection of these plasmids renders the wild-type and catalytically inactive forms, respectively. Then, 24 h after transfection, SHP2 expression levels in basal and cAMP-stimulated conditions were evaluated by Western blot in the cell lysates. As expected, in basal nonstimulated conditions, transfection of both SHP2 cDNAs increased the SHP2 protein signal compared with the endogenous expression ( Fig. 2A ). Transfected WT-SHP2 migrated at the expected molecular weight of 80 kDa in SDS-PAGE and showed conserved PTP activity as evidenced by in-gel PTP assays ( Fig. 2B ). A higher intensity of PTP activity in a band of 80 kDa was obtained in WT-SHP2-tranfected cells compared with the mock-transfected cells. Expectedly, cells overexpressing CS-SHP2 showed PTP activity levels comparable to those of mock-transfected cells ( Fig. 2B ).
Regarding the effect of cAMP on SHP2, neither the SHP2 protein level of mock-transfected cells nor that of WT-or CS-SHP2-transfected cells was affected by treatment with cAMP ( Fig. 2A ). However, the stimulation of the cells with cAMP resulted in an increase in the enzymatic activity of an 80 kDa band in mock-and WT-SHP2transfected cells ( Fig. 2B ). Transfection with the WT form of SHP2 displayed a basal activity that could be further increased by cAMP.
Next, we knocked down the expression of SHP2 using SHP2 shRNA. After transfection of SHP2 shRNA sequences, the effi cacy of SHP2 targeting siRNA was confi rmed by Western blot analysis ( Fig. 2C ) and in-gel PTP activity analysis ( Fig. 2D ). As expected, expression of SHP2-targeting shRNA decreased the level of endogenous SHP2, compared with both, empty, or scrambled shRNA transfections ( Fig.  2C ). A lower intensity of PTP activity in a band of 80 kDa was obtained in SHP2 shRNA-transfected cells ( Fig. 2D ). shRNA-treated cells ( Fig. 5A ). The effects of AA alone or together with cAMP on mock-transfected cells reproduced previously published results ( 42 ).
Finally, additional demonstration of a role for ACSL4 in SHP2-mediated effects was obtained from siRNA experiments to knockdown ACSL4 expression in SHP2 overexpressing cells. Cotransfection of MA-10 Leydig cells with the plasmid that overexpresses the wild-type form of SHP2, together with a plasmid that silences ACSL4, exerted an inhibitory effect on cAMP-stimulated progesterone production and ACSL4 protein levels ( Fig. 5B, C ). Progesterone production by SHP2 overexpression in ACSL4-depleted cells was only partially inhibited, possibly due to incomplete knockdown of ACLSL4. The effect of an excessive expression of SHP2 could not be inhibited totally by the ACLS4 shRNA transfection.

DISCUSSION
The intracellular abundance of ACSL4 protein has been determined in physiological and pathological conditions, demonstrating changes under the action of several factors: fasting/feeding state ( 6 ), ACTH/cAMP ( 11,47 ), and AA action ( 47 ), and transformation processes ( 9,10,16 ). The main fi nding of the present study is the identifi cation of SHP2 as a PTP involved in the regulation of the expression of this previously known AA-metabolizing enzyme in highly proliferative breast cancer cells and in steroidogenic differentiated Leydig cells. This conclusion is based on two facts: i ) overexpression and downregulation of SHP2 in a steroidogenic system such as MA-10 Leydig cells resulted in the upregulation and downregulation of ACSL4, respectively, and ii ) inhibition of SHP2 in a proliferative system such as MDA-MD-231 breast cancer cells reduced ACSL4 expression.
The joint action of SHP2 and ACSL4 could regulate a physiological function such as steroid production or lead to a pathological process such as cell transformation. Regulation of ACSL4 might change the subcellular pools of AA and AA-CoA and thereby alter a broad range of cellular Since the changes in ACSL4 protein levels and activity should be refl ected in downstream steps such as StAR protein induction, we next examined the effect of overexpression and knockdown of SHP2 on StAR protein.

Effect of overexpression and knockdown of SHP2 on StAR expression in MA-10 Leydig cells
As StAR abundance depends on both the induction of ACSL4 protein and the generation of AA ( 11 ), we continued analyzing StAR protein levels in SHP2 up-or downregulated MA-10 Leydig cells. The results are shown in Fig. 4 . The overexpression of WT-SHP2 significantly increased the expression of basal StAR protein compared with mock-transfected cells ( Fig. 4A ). Treatment with cAMP further increased the expression of StAR protein in the WT-SHP2-transfected cells. Overexpression of CS-SHP2 did not change StAR basal protein levels and impaired the stimulation produced by cAMP treatment.
Downregulation of SHP2 also modifi ed StAR expression. Regarding StAR mRNA levels, StAR expression was increased in MA-10 Leydig cells treated with cAMP. This effect was impaired by SHP2 shRNA ( Fig. 4B ). Again, cAMP treatment of the cells caused a signifi cant increase in StAR protein levels in mock-transfected cells, whereas SHP2 shRNA treatment prevented the effect of cAMP on StAR protein in MA-10 cells ( Fig. 4C ). The involvement of AA in this process was reinforced by the fact that the AA addition to SHP2 shRNA-treated cells bypassed the inhibitory effect described above ( Fig. 4C ).

Role of ACSL4 in SHP2-mediated effect on cAMP-stimulated steroidogenesis in MA-10 Leydig cells
There is a correlation between SHP2 and ACSL4 protein levels ( Fig. 3 ). As stated above ( Figs. 1E and 4C ) and according to previous publications, ACSL4 action in steroidogenic cells directs AA to the compartamentalized metabolism through the lipoxygenase pathway. Thus, the observed inhibitory effect of SHP2 shRNA on steroidogenesis should be bypassed by the addition of AA. Indeed, AA was able to restore the steroid production in SHP2 stimulation was performed with or without 8Br-cAMP (0.5 mM) for 1 h. Cells were collected and the cell lysate proteins were used to analyze ACSL4 expression by Western blot with specifi c antibodies. A representative immunoblot is shown. The approximate molecular mass of each immunoblotted protein is indicated. Each experiment was performed three times with similar results. The integrated optical density of each band was quantitated, and ACSL4 values were normalized against the corresponding ␤ -tubulin band and expressed as arbitrary units (AU). The results are means ± SEM of three independent experiments. a P < 0.001 versus cAMP alone. E: MA-10 Leydig cells were pretreated with or without NSC87877 (50 µM) for 3 h and then incubated in the presence or absence of AA (200 µM). The stimulation was performed with or without 8Br-cAMP (0.5 mM) for 1 h. Steroid production was evaluated determining progesterone production in the culture media by radioimmunoassay. Data represent means ± SEM of fi ve independent experiments. a P < 0.001 versus cAMP alone; b P < 0.05 versus cAMP + NSC87877. F, G: Interstitial testicular cells were pretreated with or without NSC87877 (50 µM) for 30 min. The stimulation was performed with or without 8Br-cAMP (0.5 mM) for 3 h. F: Cells were collected and the cell lysate proteins were used to analyze ACSL4 expression by Western blot with specifi c antibodies. A representative immunoblot is shown. The approximate molecular mass of each immunoblotted protein is indicated. Each experiment was performed three times with similar results. The integrated optical density of each band was quantitated, and ACSL4 values were normalized against the corresponding ␤ -tubulin band and expressed as arbitrary units (AU). The results are means ± SEM of three independent experiments. a P < 0.05 versus cAMP alone. G: Steroid production was evaluated determining testosterone production in the culture media by radioimmunoassay. Data represent means ± SEM of three independent experiments. a P < 0.01 versus cAMP alone.    processes that are controlled by this ratio. Here we report that changes in ACSL4 protein levels altered AA-mediated cell functions: steroidogenesis in MA-10 Leydig cells and proliferation in MDA-MD-231 cells. As a result of ACSL4 upregulation, overexpression of SHP2 increased StAR protein in MA-10 Leydig cells. SHP2 has to be present and activated by a cAMP-dependent pathway as the overexpression of a catalytically inactive form of the enzyme (CS-SHP2) did not reproduce the stimulatory effects of the wild-type form on ACSL4 protein levels, and it reduced cAMP action on steroidogenesis. Moreover, the catalytically inactive form slightly reduced the basal levels of progesterone production. This was probably due to an already described substrate-trapping action of the CS-SHP2 mutant ( 48 ), which would reduce the action of the endogenous enzyme. The overexpression of the wild-type form of SHP2 increased the levels of ACSL4 ( Fig. 3 ) and StAR ( Fig. 4 ) proteins in nonstimulated MA-10 cultures. However, there was no stimulatory effect on the basal steroid production ( Fig. 2 ). These results are in line with previous publications that indicate that steroidogenesis is dependent on the activity of PKA ( 49,50 ) and on phosphorylation and activation of StAR by PKA ( 51 ) and ERK ( 52 ). The sole increase in SHP2 and ACSL4 levels was not enough to increase steroid production because PKA and ERK activation are lacking.
The effects observed under overexpression of SHP2 could be due to high levels of this PTP unspecifi cally replacing the function of another endogenous PTP. This possibility was ruled out using NSC87877, a specifi c inhibitor of SHP2, and the SHP2-specifi c shRNA. The results confi rmed the need for an endogenous SHP2 to regulate the levels of ACSL4 and StAR proteins and steroidogenesis. This is the fi rst demonstration that SHP2 upregulates the levels of ACSL4, the rate-limiting enzyme of the AA generation/export system in mitochondria of steroidogenic cells. As AA is necessary for StAR protein induction and, fi nally, for steroidogenesis, this statement is reinforced by the fact that the effect of SHP2 shRNA on StAR and steroidogenesis was overcome by exogenous addition of AA. We propose that SHP2-mediated production of progesterone involves ACSL4. Indeed, silencing ACSL4 from MA-10 Leydig cells signifi cantly reduced progesterone production in response to cAMP. Progesterone production by SHP2 overexpression in ACSL4-depleted cells is partially inhibited, possibly due to incomplete knockdown of ACLSL4 ( Fig. 5B, C ). The results are means ± SEM of three (A) or six (B) independent experiments. C: Cells were collected, and the cell lysate proteins were used to analyze ACSL4 expression by Western blot with specifi c antibodies. Representative immunoblots are shown, and the approximate molecular mass of each immunoblotted protein is indicated. The experiment was performed three times with similar results. The integrated optical density of each band was quantitated. ACSL4 values were normalized against the corresponding ␤ -tubulin band and expressed as arbitrary units (AU). The results are means ± SEM of three independent experiments (C). A: a P < 0.001 versus mock + cAMP; b P < 0.05 versus SHP2 shRNA + cAMP. B: a P < 0.05 versus plasmids + cAMP; b P < 0.01 versus plasmids + cAMP; c P < 0.01 versus WT-SHP2 + cAMP. C: a P < 0.001 versus plasmids + cAMP; b P < 0.05 versus plasmids + cAMP; c P < 0.01 versus WT-SHP2 + cAMP. However, we cannot rule out that SHP2 is exerting its actions via alternative mechanisms.
Here we further demonstrated that SHP2 is activated through a cAMP-dependent pathway ( 24 ), evidenced by the increase in SHP2 activity after cAMP action in overexpression experiments. Our results are in agreement with the only report published on the subject. Rocchi et al. ( 24 ) described that SHP2 is phosphorylated in vitro by PKA and in vivo by ACTH in bovine adrenal cells. In both cases, phosphorylation increases SHP2 activity. Note that this previous report did not describe any physiological function of SHP2 nor did it evaluate the relationship of this modifi cation with the function. Thus, our report is the fi rst to show a relationship between cAMP-dependent activation of SHP2 and its functional role. We cannot state whether the activation is by direct or indirect PKA-dependent phosphorylation of SHP2.
To our knowledge, this is also the fi rst report relating SHP2 to an arachidonic acid-metabolizing enzyme such as ACSL4. Although it is known that ACSL4 induction is regulated by a cAMP/PKA-dependent pathway ( 11 ) and by AA ( 47 ), we still do not know the mechanism by which SHP2 regulates ACSL4 induction in our system. It is possible to state that ACSL4 expression is regulated at transcriptional and translational levels, depending on the length of cAMP action. At periods shorter than 1 h of cAMP action, ACSL4 protein levels are increased ( 11 ), and there is no action on mRNA levels (data not shown), indicating a regulation at the translational level. After 1 h of cAMP action, both protein ( Fig. 3C, D ) and mRNA ( Fig. 3B ) levels are increased, indicating that there is additional regulation at the transcriptional level.
The involvement of SHP2 in AA metabolism may explain its role in cell functions. SHP2 has been related to cell proliferation, survival, migration, and differentiation ( 52,53 ). With regard to the mammalian enzyme, stimulatory mutations of the human gene have been found responsible for several abnormalities, such as the pathological activity of Helicobacter pylori ( 54 ), Noonan syndrome ( 33 ), leukemogenesis ( 55 ), and human breast cancer ( 56,57 ). Although the relationship between SHP2 and hormonal imbalance has been analyzed, particularly in diabetes (58)(59)(60), the role of SHP2 in these disorders is not yet fully described or understood. Here we report a new physiological role of SHP2 in hormone action, particularly in the hormonal activation of steroidogenesis.
Our identifi cation of SHP2 as an endogenous regulator for ACSL4 expression puts one more piece of the puzzle in place. It helps to identify the PTP involved and describes a regulatory pathway of activation of a fatty acid-metabolizing enzyme such as ACSL4. Additionally, it provides an attractive target for the treatment of dysregulated cell growth. Studying the mechanism by which SHP2 exerts its action on ACSL4 induction remains a major challenge for future research.