FADS2 Function Loss at the Cancer Hotspot 11q13 Locus Diverts Lipid Signaling Precursor Synthesis to Unusual Eicosanoid Fatty Acids

Background Genes coding for the fatty acid desaturases (FADS1, 2, 3) localized at the cancer genomic hotspot 11q13 locus are required for the biosynthesis of 20 carbon polyunsaturated fatty acids (PUFA) that are direct eicosanoid precursors. In several cancer cell lines, FADS2 encoded Δ6 and Δ8 desaturation is not functional. Methodology/Principal Findings Analyzing MCF7 cell fatty acids with detailed structural mass spectrometry, we show that in the absence of FADS2 activity, the FADS1 product Δ5-desaturase operates to produce 5,11,14–20∶3 and 5,11,14,17–20∶4. These PUFA are missing the 8–9 double bond of the eicosanoid signaling precursors arachidonic acid (5,8,11,14–20∶4) and eicosapentaenoic acid (5,8,11,14,17–20∶5). Heterologous expression of FADS2 restores Δ6 and Δ8-desaturase activity and normal eicosanoid precursor synthesis. Conclusions/Significance The loss of FADS2-encoded activities in cancer cells shuts down normal PUFA biosynthesis, deleting the endogenous supply of eicosanoid and downstream docosanoid precursors, and replacing them with unusual butylene-interrupted fatty acids. If recapitulated in vivo, the normal eicosanoid and docosanoid cell signaling milieu would be depleted and altered due to reduction and substitution of normal substrates with unusual substrates, with unpredictable consequences for cellular communication.

The FADS2 encoded D6-desaturase catalyzes the first and ratelimiting step in the biosynthesis of LCPUFA. In several cancer cells desaturation does not occur, which may be due to inactivation of the desaturase or of upstream (e.g. CoA production) or downstream (e.g. acyl transferase) steps. The reason for this defect has been suggested to be due to extensive chromosomal deletions, but no molecular evidence is available [11,12]. The D8desaturation substrates 20:2n26 and 20:3n23 are often detected, thus suggesting that the elongation step is functional [12,13]. Until recently, the D8-desaturation substrates 20:2n26 and 20:3n23 were widely considered dead-end products, even though they are found in human plasma and red cells as well as other tissues. We showed the first molecular evidence that FADS2 catalyzes D8desaturation for both of these substrates, representing an alternative route to LCPUFA biosynthesis in mammals where they are converted to eicosanoid precursor LCPUFA [14]. This pathway may be available when D6-desaturase activity is compromised.
The unusual loss of PUFA desaturase activity in several cancer cells prompted us to hypothesize that the primary defect resides in the desaturase itself, and not elsewhere such as in the activation of PUFA by CoA synthesis or synthesis of phosphoglycerides as acceptors of nascent LCPUFA. We confirmed that human breast cancer MCF7 cells do not possess any capacity for biosynthesis of LCPUFA because of absence of D6-desaturase activity [12]. Here we show the restoration of this metabolic defect by heterologous expression of FADS2, evidence of novel substrate specificity for FADS1, and competition between FADS1 and FADS2 for the same substrates, leading to the production of unusual LCPUFA which are not substrates for eicosanoid production.

Results and Discussion
FADS2 and FADS1 transiently transfected MCF7 cells were probed for evidence of bioactivity toward 18:2n26 and 18:3n23. FADS2 transfected cells showed activity towards both the substrates; gas chromatography-covalent adduct chemical ionization tandem mass spectrometry (GC-CACI-MS/MS) confirmed the new peaks to be 18:3n26 and 18:4n23 respectively ( Figure 1). As expected, no products were observed when either 18:2n26 or 18:3n23 were incubated with the FADS1 and the empty vector controls. These data provide the first unambiguous molecular evidence that the metabolic defect in these cells can be restored by replacing the rate limiting D6-desaturase enzyme encoded by FADS2.
FADS genes, encoding enzymes required for PUFA biosynthesis arose evolutionarily by gene duplication events. Despite key importance of FADS2 and the loss of its function in several cancer cells, the molecular details and consequences of FADS2 loss has not been described or investigated. FADS2 is known to have activity towards at least seven substrates (18:2n26, 18:3n23, 20:2n26, 20:3n23, 24:4n26, 24:5n23, 16:0). The loss of FADS2-encoded activities in cancer cells completely shuts down the classical and alternative PUFA pathways, eliminating eicosanoid and docosanoid precursor biosynthesis from the plant-based PUFA linoleic and linolenic acids and thus limiting cell-cell signaling (Figure 3). Many studies have found strong associations between single nucleotide polymorphisms within the FADS gene cluster and complex phenotypes related to chronic disease, as well as to long chain PUFA levels [19,20]. Further studies are needed to fully understand the consequences of FADS gene function loss and/or modulation based on SNPs in neoplasm. Early studies demonstrated that micro-cell mediated transfer of HSA11 into MCF7 cells reduced tumorigenicity, suggestive of potential tumor suppressor genes on this chromosome [21].
Our findings show that a primary molecular defect in MCF7 cells lies within the FADS2 gene causing loss of FADS2-encoded D6-desaturase activity. Compensation for FADS2 function by FADS1 leads to the production of 5,11,14-20:3 and 5,11,14,17-20:4, both primarily dead-end fatty acid products that cannot be precursors for most eicosanoids and are likely to be competitive  inhibitors. The physiological significance of butylene interrupted PUFA in cancer cells calls for further detailed investigation. Our present results provide an impetus to better understanding the role of fatty acid desaturases, especially FADS2 as a tumor suppressor in neoplastic disorders.

Plasmid vector construction
The protein coding sequences of FADS1 (GenBank Accession# EF531577) and FADS2 (GenBank Accession# EU780003) were cloned into pcDNA3.1 expression vector using cDNA from neonate baboon liver tissue. The cDNA was obtained from banked samples drawn from a study approved by the Cornell University Institutional Animal Care and Use Committee (IACUC, protocol # 02-105). Analysis and comparison of amino acid sequence of baboon FADS1 showed 95% identities and 97% positives with human FADS1 (AF084558), whereas, baboon FADS2 showed 98% identities and 99% positives with human FADS2 (NM_004265). Earlier, we have shown novel D8desaturation function using baboon FADS2 [14].

MCF7 cell culture and fatty acid supplementation
For transfection experiments MCF7 cells were grown in MEMa media with 10% Lipid-reduced FBS (media and serum obtained from HyClone) at 37uC in a humidified environment with 5% CO 2 . Human breast cancer MCF7 cells (American Type Culture Collection, ATCC, Rockville, MD) were the kind gift of Dr. Rui Hai Liu, Cornell University. The FADS1 and FADS2 constructs were transfected into MCF7 cells using Lipofectamine LTX (Invitrogen, USA) as per the manufacturer's recommendations. Twenty four hours after transfection, the MCF7 cells were supplemented with 100 mM of albumin bound 18:2n26, 18:3n23, 20:2n26 and 20:3n23 fatty acids and were incubated for additional 24 hours.

Fatty acid analysis
After incubation with fatty acids, the cells were washed twice with 1XPBS and removed by trypsinization. The cells were harvested by centrifugation and the cell pellet was processed for lipid extraction. Fatty acid methyl ester preparation was carried out using a modified one-step method of Garces and Mancha [22]. Analysis was by gas chromatography-flame ionization detection (GC-FID) [8] and peak identification was confirmed by GCcovalent adduct chemical ionization tandem mass spectrometry (GC-CACI-MS/MS) [14,23].