Fatty acid synthase is a metabolic oncogene targetable in malignant peripheral nerve sheath tumors

Abstract

Background

Malignant peripheral nerve sheath tumors (MPNSTs) are soft tissue sarcomas with minimal therapeutic opportunities. We observed that lipid droplets (LDs) accumulate in human MPNST cell lines and in primary human tumor samples. The goal of this study was to investigate the relevance of lipid metabolism to MPNST survival and as a possible therapeutic target.

Methods

Based on preliminary findings that MPNSTs accumulate LDs, we hypothesized that a deregulated lipid metabolism supports MPNST cell survival/proliferation rate. To test this, we examined respiration, role of fatty acid oxidation (FAO), and the enzyme fatty acid synthase involved in de novo fatty acid synthesis in MPNSTs using both genetic and pharmacological tools.

Results

We demonstrate that LDs accumulate in MPNST cell lines, primary human and mouse MPNST tumors, and neural crest cells. LDs from MPNST cells disappear on lipid deprivation, indicating that LDs can be oxidized as a source of energy. Inhibition of FAO decreased oxygen consumption and reduced MPNST survival, indicating that MPNST cells likely metabolize LDs through active FAO. FAO inhibition reduced oxygen consumption and survival even in the absence of exogenous lipids, indicating that lipids synthesized de novo can also be oxidized. Consequently, inhibition of de novo fatty acid synthesis, which is overexpressed in human MPNST cell lines, effectively reduced MPNST survival and delayed induction of tumor growth in vivo.

Conclusion

Our results show that MPNSTs depend on lipid metabolic pathways and suggest that disrupting lipid metabolism could be a potential new strategy for the development of MPNST therapeutics.

Malignant peripheral nerve sheath tumor (MPNST) is a rare soft tissue sarcoma that is highly invasive and lethal unless complete resection is feasible. Half of MPNSTs arise spontaneously in adults (sporadic MPNSTs), and 50% of MPNSTs are associated with neurofibromatosis type 1 (NF1); the lifetime risk of MPNST in NF1 patients is 8%–13%.^1,2 Chemotherapy and radiation have minimal advantage in MPNST³; therefore, alternative therapies are urgently needed.

MPNSTs are nerve-associated soft tissue sarcomas. MPNST cells express markers characteristic of the neural crest cells from which they are believed to arise.^4–6 Schwann cell progenitors from the neural crest differentiate into nerve glial cells (Schwann cells). Neural crest cells have considerable self-renewal and differentiation potential, while progenitors identified after the establishment of the dorsal root ganglia and peripheral nerve have more limited self-renewal and differentiation potential.^7,8 We hypothesized that shared features of neural crest cells in defined MPNST cells could be exploited as MPNST therapies.

MPNSTs have a complex karyotype with amplifications and deletions of many alleles.⁹ Sporadic MPNST can have mutations in the NF1 gene or other Ras pathway genes, confirming reliance on the Ras pathway in this tumor type.¹⁰ Inhibition of MEK (mitogen-activated protein kinase kinase) downstream of Ras signaling evoked a modest reduction in growth of human MPNST xenografts.¹¹ The mammalian target of rapamycin (mTOR) signaling pathway is also activated in MPNST, and blocking mTOR signaling with rapamycin or its analogs also transiently delayed tumor growth.^12,13 Several other signaling pathways have been tested over the past decade, but there have been no therapeutics demonstrated to target MPNSTs. Here we began to identify targetable molecules by studying metabolism of MPNST cells.

We focused our efforts on lipid metabolism, as increased lipogenesis is a hallmark of many human cancers and has been associated with poor prognosis in breast, prostate, and colon cancer.^14,15 Two hallmarks of altered lipid metabolism are the increased accumulation of lipid bodies/droplets in cell bodies and an overexpression of enzymes involved in fatty acid synthesis (FASN). Inhibitors of FASN are potential targets for antineoplastic interventions and chemoprevention.¹⁶ Animal cells have 2 sources of fatty acids (FAs), exogenously derived (dietary) FAs and endogenously synthesized FAs. The biosynthesis of the latter is catalyzed by the multifunctional, homodimeric enzyme FASN.^15,17 FASN synthesizes long-chain FAs from acetyl-CoA using malonyl-CoA and NADPH. Cerulenin, a natural antibiotic product of the fungus Cephalosporium caerulens,¹⁸ inhibits FASN activity and can cause apoptotic death of cancer cells in vitro.^19,20 C75, a synthetic cerulenin analog, is also a potent FASN inhibitor²¹ cytotoxic to cancer cells in vitro, and shows significant in vivo antitumor activity against some human cancer xenografts.^22,23

We show here the importance of lipid metabolism in MPNST. Lipid droplets (LDs) accumulate in human MPNST cells in vitro and in human tumor samples, consistent with the stored lipids that we identify in migrating neural crest cells. Inhibiting FASN in vitro caused cytotoxic effects on MPNSTs and significantly delayed tumor growth in a human MPNST xenograft model. Thus, targeting lipid metabolic pathways could be a potential therapeutic strategy in MPNST.

Materials and Methods

Cell Lines and Reagents

MPNST cell lines STS26T, ST8814, ST88-3, S462, T265p21, 90-8, immortalized human Schwann cells (iHSCs), and normal human Schwann cells from autopsy specimens were obtained and maintained as described.^24–26 Postmigratory neural crest cells were isolated from mice on embryonic day (E) 8.5 and plated on poly-l-lysine + fibronectin-coated Labtech chamber slides (Nunc) as described.²⁷

BODIPY Staining

MPNST cells cultured in 8-well chambered slides were washed with ice cold phosphate buffered saline (PBS) and fixed in 4% paraformaldehyde for 5 min at room temperature (RT). After a rinse in PBS the cells were incubated with BODIPY 493/503 (boron-dipyrromethene; Life Techonologies #D-3922) at 10 ng/mL in dimethyl sulfoxide (DMSO) for 15 min in the dark at RT. After rinses in PBS, cells were mounted in Fluoromount-G and imaged using the Zeiss Axiovert 200M microscope.

Oil Red O Staining

Sections were fixed in 10% formalin for 10 min, then rinsed 3 times in 1× PBS. Slides were then placed in oil red O (ORO) solution for 10 min, rinsed in tap water, and counterstained with hematoxylin for 1 min, per kit protocol (Abcam # ab150678). Sections were then mounted in Fluoromount-G and coverglassed.

Isolation and Sectioning of Mouse Embryos

Embryos were isolated from pregnant wild-type C57Bl/6 mice on E8.5–E15.5. Wnt-Cre enhanced green fluorescent protein (EGFP) embryos were from Jeff Robbins at Cincinnati Children's Hospital Medical Center (CCHMC).²⁷ Embryos were decapitated and placed in 4% paraformaldehyde for 12–24 h, transferred to 20% sucrose, and stored at 4°C until embedding and freezing in optimal cutting temperature (OCT) medium. Serial sections (12 µM) were cut using a cryostat (Leica). Tissue sections at the level of the hind limbs were collected and stained with ORO to visualize neutral lipids and cholesteryl esters but not biological membranes.²⁸

Tumor Samples

Human MPNST tumors were collected in accordance with institutional review board–approved protocols from discarded surgical specimens received from the Cincinnati Children's Hospital Bio Bank as flash frozen samples. Geneticaly engineered mouse (GEM)–PNST tumors from Nf1+/−/p53+/− mutations in cis on mouse chromosome 11 (NPcis) have been described.⁶ Xenograft tumors were obtained from nu/nu mice injected with the MPNST cell line STS26T, a non-NF1 (sporadic) MPNST cell line.¹³ Unfixed tumors were embedded in OCT medium and frozen and 20-μM cryostat sections cut.

MPNST Xenograft and Drug Administration

The STS26T human MPNST xenograft model has been described.^13,29 STS26T MPNST cells (1.8 × 10⁶) were injected subcutaneously, in Matrigel, into the flanks of 4- to 5-week-old female nu/nu mice (Harlan). Treatment began when measurable (250 mm³) tumors developed. C75 was dissolved in DMSO at 100 mM and diluted further in Dulbecco's modified Eagle's medium (DMEM) for administration to mice at a dose of 40 mg/kg (first dose) and 30 mg/kg subsequently, once a week i.p. in 0.1 mL total volume as in previous studies.^30,31 Controls were administered vehicle (DMSO/DMEM). Mice were weighed and their tumor volumes measured with digital calipers twice weekly until tumors reached 2500 mm³. Tumor volume was calculated by: L * W2 (π/6), where L is the longest diameter and W is the width. All experiments were conducted following the approved protocol of the Institutional Animal Care and Use Committee.

Lentiviral Transfection

MPNST cells were transduced with lentiviral particles at 50%–60% confluence. Short hairpin (sh)RNAs targeting FASN, acetyl-CoA carboxylase (ACC), and control (nontargeting) were from the Sigma-Aldrich TRC (The RNAi Consortium) library. The CCHMC Viral Vector Core produced virus using a 4-plasmid packaging system (http://www.cincinnatichildrens.org/research/div/exphematology/translational/vpf/vvc/default.html). Lentiviral particles were incubated with MPNST cells in the presence of polybrene (8 μg/mL; Sigma) for 24 h, followed by selection in 2 μg/mL puromycin, which killed uninfected cells within 3 days.

Immunoblot

Cell lysates were created and western blotting conducted as described.⁴ Membranes were probed with anti-FASN (Cell Signaling Technology #3189) and stripped to be reprobed with horseradish peroxidase (HRP)–conjugated anti–β-actin (Cell Signaling Technology #5125) as a loading control. Signals were detected using HRP-conjugated secondary antibodies (BioRad) and the Immobilon Western Chemiluminescent HRP Substrate system (Millipore #WBKLS0500).

Quantitative Real-Time PCR

Total RNA was isolated from cells using the RNeasy kit (Qiagen) and used as a template for cDNA synthesis (High-Capacity cDNA archive kit, Applied Biosystems) and quantitative real-time PCR (ABI 7500 Sequence Detection System) as described.²⁹ For FASN and ACC, prime time pre-validated quantitative PCR primer sets were used from Integrated DNA Technologies assay #Hs.PT.56a.20384174 and #Hs.PT.56a.513712.g (for ACC).

Immunohistochemistry

Ten-micrometer sections from cryoprotected frozen xenograft tumors were fixed in 10% formalin for 10 min and immunohistochemistry conducted as described.²⁹ We used rabbit anti-cleaved caspase-3 (1:8000; Cell Signaling #9661) and rabbit anti-Ki-67 (1:5000 NCL-Ki67-P; Novocastra). Subsequently, sections were stained with the LD stain ORO and hematoxylin. Sections were photographed on a Nikon Eclipse 80i bright field microscope. For nuclear staining, sections were incubated in 4′,6′-diamidino-2-phenylindole (DAPI) (1:10 000; Sigma) for 5 min, rinsed in PBS 3 times, and coverglassed in Fluoromount-G (EM Sciences #17984-25).

In vitro Cytotoxicity Assays

MPNST cells were plated in quadruplicate in serum containing growth medium on 96-well plates at 500 cells/well/dose. Plates were incubated at 37°C and 5% CO₂. After 24 h, cells were treated with carrier, etomoxir (R&D Systems #4539), C75 (Tocris #2489), orlistat (Sigma #O4139), and Irgasan (Sigma #72779). The amount of cell growth was determined using 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxy-phenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt (MTS) assay using the Cell Titer 96 proliferation kit (Promega). Absorbance was read at 490 nm on a Spectramax M2 plate reader (Molecular Devices).

Cell Proliferation/Death Analysis

MPNST cells at 60%–70% confluency were transduced with lentivirus, puromycin selected, and plated at day 3 post infection in 8-well chamber slides in DMEM/10% fetal bovine serum. The next day, cells were incubated with bromodeoxyuridine (BrdU) (BD Pharmingen #550891) 10 μM in cell culture medium for 15 min at 37°C, rinsed with 1× PBS followed by a distilled (d)H₂O rinse, and fixed with 4% paraformaldehyde (in 1× PBS) for 10 min at RT, then with 2N HCl/PBS containing 0.5% Triton and heated to 37°C for 30 min, then neutralized with 0.1 M borate buffer (pH 8.5) for 5 min. Slides were washed in dH₂O thrice. Blocking was with 10 mg/mL bovine serum albumin–PBS–0.1% Triton for 15 min. Incubation was overnight in anti-BrdU (1:500; Abcam # ab6326) or rabbit anti-cleaved caspase-3 (Asp175) (1:200; Cell Signaling #9661). The slides were washed in PBS, then incubated for 2 h at RT in Alexa-488 anti-rat (A11006) or Alexa-647 anti-rabbit (A21244; Invitrogen) at 1:500 in blocking buffer. Cells were stained with DAPI; washed in PBS, then dH₂O; and mounted in Fluoromount-G. Immunolabeled cells were visualized using a Zeiss Axiovert 200 M microscope. Antigen-positive cells were quantified from 5 fields/well for each treatment and plotted as percent of total DAPI+ nuclei per field.

Oxygen Consumption Rate Measurements

Cells (3000 cells/well) were plated into 96-well plates (Flux pack of XFe96, Seahorse Biosciences #102416-001) in 80 μL normal cell culture medium at 37°C, 5% CO₂ overnight. The next day, after 2 rinses with 1× PBS, the medium was changed to extracellular flux minimal assay medium (Seahorse Biosciences #102353-100SB) supplemented with 1× GlutaMAX (Gibco #35050-079), 10 mM glucose (Sigma #G8769), and 0.5 mM carnitine (Sigma #C0158-5) and incubated at 37°C without CO₂ for an hour. The steady-state (baseline) oxygen consumption rate (OCR) was measured with an XF96e Extracellular Flux Analyzer (Seahorse Biosciences). Etomoxir was added 10 min prior to OCR measurement. For OCR measurement in low lipid serum (LLS) (Hyclone #SH30855- 03), cells were plated on day 0, medium changed to DMEM/LLS on day 1, and OCR measurement conducted on day 4.

Results

MPNSTs Show Lipid Droplet Accumulation in Their Cell Bodies

While observing MPNST cells by phase contrast microscopy, we noted the consistent presence of dark speckles in cell cytoplasm. To determine if these were LDs, we stained the MPNST cells with BODIPY 493/503. Imaging in phase and fluorescence and overlapping the images (Fig. 1A–C) confirmed that the dark spots observed were indeed LDs. We then stained cells from 7 human MPNST cell lines with ORO, a red bright field-neutral lipid stain. All MPNST cell lines exhibited LDs (Fig. 1E–G), while normal human Schwann cells (Fig. 1D) and iHSCs (data not shown) did not. To determine if LDs are present in MPNST tumors, we stained tissue sections. Surgically resected human MPNST tumors from 9 patients, and GEM-PNST tumors from the NPcis model (n = 3; data not shown) were stained with ORO. There was considerable variability within each sample, but all showed ORO-positive regions (Fig. 1H–K). Thus, primary human MPNSTs are enriched in LDs.

Fig. 1.

MPNSTs accumulate LDs. (A–C) BODIPY-stained MPNST (STS26T) cells showing dark black spots in bright field imaging overlapping with fluorescent green droplets stained positive with neutral lipid stain BODIPY 493/503. The inset in C shows a higher magnification image of the droplets outlined in white in A, B, and C. Bright field images of human MPNST cell lines are stained with ORO (red) and hematoxylin (blue). (D) No droplets are seen in normal human Schwann cells. (E–G) LDs are numerous in the sporadic MPNST cell line STS26T (F) and 2 NF1-related MPNST cell lines, S462-TY (G) and 8814 (E). (H–K) Bright field images of 4 different human MPNST tumor samples stained with ORO (red) and hematoxylin (blue); the images show fields with few LDs (H) and others with numerous LDs (I–K). Scale bars in G and K are 100 μm.

Migrating Neural Crest Cells Are Enriched in LDs During Development

MPNSTs are peripheral nerve sarcomas that may derive from peripheral nerve Schwann cell lineage cells or their multipotent neural crest cell precursors. Given that MPNSTs express transcripts and proteins characteristic of neural crest cells, we examined whether the cells of the migrating neural crest also show LD accumulation. The neural crest is a transient pluripotent cell population that delaminates from the dorsal neural tube. Migratory neural crest cells present in the trunk at E9 develop into Schwann cell precursors between E11 and E13 in mouse sciatic nerve, and into Schwann cells that ensheath nerve axons by birth.^32,33 To enable identification of neural crest cells, we used a Wnt-Cre mouse model; breeding to a reporter mouse resulted in EGFP expression specifically in neural crest lineage cells in the trunk.²⁷ During neural crest cell migration at E9.5, precursors of peripheral nerve glia (Schwann cells), migrate through the ventrolateral migratory pathway adjacent to the neural tube.³⁴ At this time point we observed many EGFP-labeled cells in the migratory pathway that appeared to contain LDs (Fig. 2A and B). We also isolated postmigratory neural crest cells from E8.5 mouse neural tubes and tested for LDs. Consistent with the in vivo result, most or all cells that migrated from neural tubes showed abundant LDs in their cell bodies (Fig. 2C).

Fig. 2.

Migrating neural crest cells show lipid accumulation. (A–C) Micrograph of sections from an E9.5 mouse embryo showing EGFP-positive neural crest cells. LDs stained with ORO are red and shown at 200× (A) or 670× (B). Postmigratory neural crest cells from E8.5 neural tubes also show abundant LDs (C). (D–G) Sections from wild-type embryos at E12.5 stained with ORO and hematoxylin. Section from an E12.5 embryo showing the location of the neural tube (NT) and dorsal root ganglia (DRG) (D). Few LDs are seen in cells around the DRG (D and E) or in the developing peripheral nerve (PN) (F). While most sections are entirely negative for lipid staining, rare cells in the growing nerve marked by # (G) were positively stained for ORO. Scale bar in G is 20 μm and applies to E and F.

In tissue sections from older embryos (E10.5–E11.5; data not shown), progressively fewer ORO-positive cells were observed. By E12.5, most cells near the dorsal root ganglia and in the developing peripheral nerve were ORO negative (Fig. 2D–F). Only in rare sections were occasional ORO-positive cells present, either near the dorsal root ganglia (data not shown) or within the developing peripheral nerve (Fig. 2G). Thus, LDs are commonly present in migrating neural crest cells at E9.5, but not in the more mature embryonic Schwann cells present in the dorsal root ganglion or peripheral nerve at E12.5.³³

MPNSTs Use Endogenous and Exogenous Lipids

As MPNSTs accumulate LDs in cell lines and in tumor samples, we reasoned that they are used as a source of fuel to support proliferation. To examine if the lipids in LDs are being oxidized through the fatty acid oxidation (FAO) pathway, we used 2 FAO pathway inhibitors, etomoxir and perhexiline (Fig. 3A), which are carnitine-palmitoyl transferase 1 (CPT1) inhibitors that block the rate-limiting step in FAO. Both inhibitors significantly decreased viability of MPNST cells (Fig. 3B, Supplementary Fig. S1A and S1B). Since FAO reduced viability, we reasoned that FAO inhibitors would also reduce O₂ consumption (OCR) if MPNST cells use molecular O₂ to metabolize LDs. Consistent with the reliance of MPNST cells on FAO, etomoxir caused a 20%–30% drop in OCR in all MPNST cell lines tested (Fig. 3C). No toxicity was observed after etomoxir treatment, as judged in a cytotoxicity assay (MTS) after OCR reading (Fig. 3D). Therefore, under normal cell culture conditions (DMEM with 10% fetal bovine serum), MPNST cell lines actively utilize FAO, possibly as a source of energy.

Fig. 3.

MPNSTs are dependent on fatty acid oxidation. (A) The FAO pathway and FAO inhibitors. Perhexiline and etomoxir inhibit CPT1, an enzyme that determines the key rate-limiting step in FAO. (B) Perhexiline and etomoxir have detrimental effects on human MPNST cell survival, as measured by MTS assay; 490-nm absorbance reading on y-axis. (C) Metabolic rate measurement presented as % OCR of untreated cells for 4 human MPNST cell lines ± etomoxir. (D) MTS cell survival assay performed at endpoint on the same samples as OCR measurements in C. (E) Bright field images of MPNST cells cultured in normal fetal bovine serum (FBS) (NS)/DMEM vs low lipid FBS (LLS)/DMEM at day 4 showing depletion in LDs in cell bodies. Scale bar = 50 μm. (F) Reduced viability (30% drop) of MPNST cells in LLS by day 4. (G) Cells cultured in LLS + etomoxir show a drop in OCR. (H) Lipid-starved cells are more sensitive to inhibition of FAO via etomoxir.

We hypothesized that LDs are stored in MPNST cells as a source of energy to be used when nutrients are limited. Since normal MPNST culture medium is rich in lipids, we reasoned that an LLS culture condition would test this hypothesis. MPNST (STS26T) cells cultured in the absence of exogenous lipids showed a depletion of LDs over 4 days of in vitro culture (Fig. 3E). At this time point, MPNST cells cultured in LLS media also showed reduced viability (Fig. 3F). To test if cells cultured in the absence of exogenous lipids were using FAO, OCR was measured in lipid-starved cells at day 4. A significant drop in OCR was observed in LLS cells treated with etomoxir versus untreated LLS cells (Fig. 3G, Supplementary Fig. S1C). Also, STS26T–LLS cells were more sensitive to etomoxir than were untreated STS26T cells (Fig. 3H). The drop in OCR in the LLS cells on day 4 suggests that besides oxidizing exogenous lipids, MPNST cells also oxidize lipids synthesized de novo. Thus, MPNST cells rely on oxidation of lipids when nutrients are available, and they metabolize endogenous lipids under lipid-reduced conditions.

Fatty Acid Synthase Is Overexpressed in MPNST Cell Lines and Contributes to MPNST Cell Survival

The results presented above suggest that de novo synthesized lipids are also oxidized by MPNST cells. Therefore, we examined the expression of FASN, the only mammalian enzyme that makes de novo lipids in mammals. Indeed, FASN was significantly overexpressed in MPNST cell lines at the mRNA and protein levels versus normal human Schwann cells (Fig. 4A and B). To test whether FASN is important for MPNST viability, FASN expression was blocked with 2 validated lentiviral shRNAs. Three days after lentiviral infection of MPNST cells, FASN mRNA expression in FASN-silenced cells (shFASN) was reduced by 3.8-fold relative to cells expressing nontargeting shRNA (shNT) (Fig. 4C). The expression of FASN was also reduced at the protein level, as confirmed by western blot analyses 4 days after lentiviral shRNA infection (Fig. 4D). FASN inhibition significantly decreased MPNST viability (Fig. 4E) and reduced metabolic activity as measured by MTS assay (Fig. 4E and F). ShRNA targeting ACC, an upstream enzyme in the FAO pathway, also diminished MPNST cell metabolic activity and viability (Fig. 4F). We next measured cell proliferation via BrdU incorporation and cell death via cleaved caspase-3 assay. ShFASN–MPNST cells showed a decrease in cell entry into S phase (proliferation) and increased cell death (apoptosis) compared with controls (Fig. 4G). Thus, both FAO and de novo synthesis are necessary for survival of MPNST cells.

Fig. 4.

Fatty acid synthesis is important for MPNST survival. (A) Real-time (RT) PCR analysis of FASN mRNA in iHSCs, the sporadic MPNST cell line 26T, the NF1-related human MPNST cell lines 88-3, 8814, and S462-TY relative to normal human Schwann cells (nHSCs). (B) Western blot analysis showing elevated FASN protein in 26T, 88-3, 8814, and S462-TY human MPNST cells compared with nHSCs and iHSCs, with β-actin shown as a loading control. (C) RT-PCR analysis shows diminished FASN mRNA 3 days posttransduction with shRNA targeting FASN. (D) FASN protein level is efficiently reduced post-infection day 4. (E) By day 5 posttransduction, phase contrast images show that few cells remain in wells infected with shFASN vs controls. (F) Cell viability is decreased in MPNST cells expressing ACC or FASN silencing shRNA as quantified by MTS assay; there is no effect of downregulating FASN on iHSCs; scale bar is 100 μm. (G) Downregulation of FASN reduced the % BrdU-positive cells and increased cell death (cleaved caspase-3 positive cells).

Inhibiting FASN Activity Reduces Cell Growth In vitro

To corroborate our genetic data suggesting a role for FASN activity in MPNST cell survival, we used drugs to inhibit FASN activity.¹⁶ MPNST and iHSC lines were treated with 3 FASN inhibitors: C75, orlistat, and Irgasan. Each significantly reduced MPNST cell survival, with half-maximal inhibitory concentration (IC₅₀) values for orlistat ranging from 9 μM (STS8814) to 25 μM (STS26T) and averaging 50 μM for Irgasan (Fig. 5B and C). All 3 inhibitors also had a detrimental effect on the survival of the iHSCs (Fig. 5A–C). However, iHSCs were consistently less sensitive than MPNST cells to each of the inhibitors. C75 is a widely used FASN inhibitor, studied in cancer models over the last decade. The IC₅₀ for MPNST cells exposed to C75 averaged 28 μM (Fig. 5A), in accordance with reports of an average IC₅₀ of 20 μM for a subset of breast cancer cells³⁵ and 72 μM for lung cancer cell lines.³⁶

Fig. 5.

C75 delays tumor formation in MPNST (26T) xenografts. (A–C) FASN inhibitors C75, orlistat, and Irgasan significantly reduced survival of iHSC, 26T, and S462-TY MPNST cell lines by day 5 of treatment. (D–F) Tumor volume (mm³) in mice treated with vehicle or C75. There is a significant delay in tumor growth in drug-treated mice compared with the vehicle group (F). (G) Images of xenograft tumor section stained for cleaved caspase-3 and ORO. Nuclei were counterstained with hematoxylin (blue). All tumors showed similar morphology, including regions of cell death/necrosis (brown) and LDs (red). Scale bar is 100 μm.

Inhibiting FASN Activity Delays Tumor Growth In vivo

To test the effect of FASN inhibition on MPNST growth in vivo, we implanted STS26T MPNST cells into nu/nu mice and treated established tumors (250 mm³) with C75 (n = 14) or vehicle control (n = 13). C75 was administered at 40 mg/kg on week 1, but some mice lost close to 10% body weight. We therefore dosed at 30 mg/kg on subsequent weeks. Higher levels of C75 are known to induce anorexia and significant weight loss, so this dosing schedule has been used before.^37,38 Inhibiting FASN with C75 significantly delayed tumor growth compared with vehicle controls (Fig. 5D–F). While the entire vehicle-treated mouse group reached 2500 mm³ in 16–21 days (Fig. 5D), C75-treated mice showed a 10-day delay (Fig. 5E). Xenograft tumors were collected when they reached maximum allowable tumor burden. Paraffin sections of tumor showed apoptotic regions as detected by anti-cleaved caspase-3. These apoptotic regions contained cells with LDs, as detected via ORO (Fig. 5G). Notably, these LD-rich, apoptotic domains surrounded the necrotic regions and were absent in actively dividing regions of tumor. Thus, targeting FASN is a potential therapeutic for MPNSTs, with modest single-agent efficacy.

Discussion

Here we have shown that lipid storage and mobilization are features of MPNST cells and neural crest cells. LDs were present in migrating mouse neural crest cells, in human MPNST cells in vitro, and in human/mouse tumor samples. MPNST can use LDs as a source of energy, as observed under exogenous lipid deprivation. Inhibiting the expression or activity of FASN, a key enzyme involved in de novo FASN, caused cytotoxic effects on MPNSTs in vitro and delayed tumor growth in a human MPNST xenograft model. These results suggest that MPNSTs critically depend on lipid metabolism for optimal growth and proliferation.

Our finding that LDs are present in MPNSTs and migrating neural crest cells is consistent with studies showing that human MPNST cells resemble neural crest cells in expression of the markers TWIST1, SOX9 PAX3, N-CAM, and p75.^4,5 Their characteristic LD expression may be useful as a potential marker to distinguish MPNST from synovial sarcoma or other difficult to differentiate tumor types. In particular, we note that during development, most or all Schwann cell precursors and mature Schwann cells lack LDs. Indeed, our unpublished findings show that neurofibromas, which contain Schwann cells and Schwann cell precursor-like cells, lack LDs. Therefore, presence of LD-rich cells in a neurofibroma biopsy could potentially reflect regions of transformation to MPNST. These ideas will require verification.

We found that MPNSTs accumulate LDs in their cytoplasm and metabolize LDs on nutritional deprivation. Inhibition of FAO using several inhibitors showed a drop in OCR readings in these MPNST cells, indicating that these cells rely on FAO to meet their energy demands. The source of MPNST lipids (endogenous vs exogenous) in growing tumors remains to be determined. We show that under limiting exogenous lipid, addition of FASN inhibitor still causes a drop in OCR. This indicates that when exogenous lipid sources are reduced, MPNSTs rely on oxidizing lipids in the LDs (which gradually disappear) but also channel endogenously produced lipids toward FAO when LDs are no longer available. The mechanisms underlying lipid mobilization in MPNST remain to be studied. Several complex mechanisms make lipids available for mitochondrial FAO; these include autophagic digestion of intracellular membranes and of LDs.³⁹

Transformed cells frequently exhibit specific alterations in metabolic activity, which supports their rapid proliferation and survival. A prominent metabolic alteration in cancer cells is an increase in glucose uptake and the use of aerobic glycolysis, called the Warburg effect.⁴⁰ In the past decade there has been a renewed interest in examining mitochondria-dependent functions in cancer. Citrate produced during glucose oxidation in mitochondria is the precursor of cellular lipids in the cytoplasm. In some contexts (as we found in MPNST cells), lipids reenter mitochondria for oxidation—a process important for the survival of these cells. Our results substantiate the importance of mitochondrial metabolism in MPNSTs for proliferation and growth. Besides oxidation in the mitochondria, lipids in LDs also undergo autophagy (macrolithophagy) during energy crisis,⁴¹ and future studies will determine whether macrolithophagy contributes to MPNST survival.

The majority of adult mammalian tissues satisfy their lipid requirements through the uptake of free FAs from the blood. Fatty acid and cholesterol biosynthesis are restricted to select tissues, including liver and adipose and lactating breast tissues. However, reactivation of lipid biosynthesis is frequently observed in cancer.¹⁶ We found this to be true for human MPNSTs, as evidenced by the overexpression of key enzymes of the de novo FASN pathway. Consequently, blocking FASN and ACC expression or activity significantly reduced the survival of MPNST cells. Inhibition of FASN in vivo in human MPNST xenograft models reduced sarcoma burden, consistent with previous observations in breast and prostate cancer xenograft models.³⁵

We posit that FASN inhibition, via lifestyle changes including a lipid-restricted diet or drugs, could be a component of MPNST therapy. However, because the current FASN inhibitors elicit anorexic effects, newer agents without this side effect are needed to test this idea. In addition, combinatorial therapies may be required for maximal benefit to patients. In conclusion, deregulated lipid metabolism is a novel therapeutic avenue for MPNST.

Supplementary material

Supplementary material is available online at Neuro-Oncology (http://neuro-oncology.oxfordjournals.org/).

Funding

This work was supported by a grant from Cancer Free Kids Grant for N.R. R01NS075291-01 to B.D. A.V.P. received funding from the DAMD Program on Neurofibromatosis Postdoctoral Fellowship W81XWH1110144 and a Pelotonia Postdoctoral Award.