Background
Targeting cancer specific metabolism represents an opportunity to develop novel, potentially selective and broadly applicable drugs to treat a multiplicity of cancer types. Malignant tissues require large amounts of lipid for membrane biosynthesis, energy, and signal transduction during tumor progression [
1].
De novo fatty acid synthesis is the main means of fatty acid supply in cancers, therefore, enzymes involved in fatty acid metabolism have been implicated in cancer biology [
2]. For example, overexpression of fatty acid synthase results in enhanced lipogenesis, a common feature in a variety of human cancers, including primary brain tumors [
3,
4]; and inhibiting fatty acid synthase or lipogenesis induces cancer cell death [
5]. In addition to fatty acid synthase, several other enzymes involved in lipid metabolism have recently been shown to be involved in tumor growth and malignancy [
6,
7]. These data show that enzymes involved in lipid metabolism are potential therapeutic targets against cancers.
In the lipid metabolism cascade, addition of coenzyme A (CoA) to fatty acids is a fundamental initial step in the utilization of fatty acids for structural and storage lipid biosynthesis, signaling lipid protein acylation, and other metabolic processes [
8]. Acyl-CoA synthetases (ACSs) are key enzymes for this fatty acid activation step [
9]. ACS catalyzes an ATP-dependent multi-substrate reaction, resulting in the formation of fatty acyl-CoA. The overall reaction scheme is:
Human cells contain 26 genes encoding ACSs [
9,
10]. Phylogenetically, ACSs are divided into at least four subfamilies that correlate with the chain length of their fatty acid substrates, although there is considerable overlap. There are short-chain ACS (ACSS), medium-chain ACS (ACSM), long-chain ACS (ACSL) and very long-chain ACS (ACSVL). Both ACSL and ACSVL isozymes are capable of activating fatty acids containing 16–18 carbons, which are among the most abundant in nature, but only the ACSVL family enzymes have significant ability to utilize substrates containing 22 or more carbons. Each ACS has a unique role in lipid metabolism based on tissue expression patterns, subcellular locations, and substrate preferences. For example, ACSL4 is overexpressed in breast, prostate, colon, and liver cancer specimens [
11‐
13]. Among the multiple ACS members, two isozymes ACSL5 and ACSVL3, have been found important in gliomagenesis and malignancy [
14,
15].
Many solid malignancies, including glioblastoma multiforme (GBM), exhibit a cellular hierarchy containing subsets of tumor cells with stem-like features, which are currently believed to disproportionately contribute to tumor growth and recurrence [
16,
17]. These “cancer stem cells” display the capacity for long-term self-renewal, efficient propagation of tumor xenografts in experimental animals, the capacity for multi-lineage differentiation, and resistance to cytotoxic DNA-damaging agents [
18,
19]. Understanding the mechanisms that regulate cancer stem cell self-renewal and tumor-propagating potential could lead to new and more effective anti-cancer strategies.
The influence of lipid metabolism pathways on cancer stem cells has not been explored in great detail. ACSVL3 (alternatively designated as FATP3, SLC27A3) is one of the most recently characterized members of the ACS family [
20]. Mouse ACSVL3 mRNA is found primarily in adrenal, testis, ovary, and developing brain; and ACSVL3 protein mainly localizes to subcellular vesicles that fractionate with mitochondria [
20]. Compared with normal brain tissues, ACSVL3 expression levels are elevated in clinical GBM specimens and induced in GBM cells following the activation of oncogenic receptor tyrosine kinases. We previously reported that ACSVL3 supports tumor promoting capacity in human GBM [
14], a biological property attributed to the cancer stem cell phenotype. This current study examines the expression and function of ACSVL3 in GBM stem cell enriched neurosphere isolates. We show that ACSVL3 functions to support GBM stem cell self-renewal and the capacity of GBM stem cells to propagate tumor xenografts. Our results suggest that targeting ACSVL3-dependent lipid metabolic pathways could be a strategy for inhibiting GBM stem cells and their capacity to support tumor growth and recurrence.
Methods
Reagents
All reagents were purchased from Sigma Chemical Co. (St. Louis, MO) unless otherwise stated. Hepatocyte growth factor (HGF) was a gift from Genentech (San Francisco, CA, USA). Epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF) were purchased from Peprotech (Rocky Hill, NJ, USA). This study utilized discarded human pathological specimens from Johns Hopkins Neurological Operating Suite. Our use of de-identified pathological specimens as described here was reviewed by the John Hopkins IRB and designated to be “not human subjects research”.
GBM neurosphere culture and differentiation
Human glioblastoma neurosphere lines HSR-GBM1A (20913) and HSR-GBM1B (10627) were originally derived by Vescovi and colleagues [
16]. The GBM-DM14602 neurosphere line was derived from a glioblastoma at the University of Freiburg and kindly provided by Dr. Jaroslaw Maciaczy [
21,
22]. The primary neurospheres JHH612, JHH626 and JHH710 were derived from discarded glioblastoma surgical specimens at Johns Hopkins Hospital using the same methods and culture conditions as described in Galli et al. [
16,
23]. The primary neurosphere isolates were used at passage ≤ 10. All human materials were obtained and used in compliance with the Johns Hopkins IRB. GBM neurosphere cells were maintained in serum-free medium containing DMEM/F-12 (Life technologies, Carlsbad, CA), 1% BSA, EGF and FGF [
16,
24,
25]. Cells were incubated in a humidified incubator containing 5% CO
2 and 95% air at 37°C, and passaged every 4–5 days. Forced differentiation was performed according to the method of Galli et al. [
16] with some modifications [
26]. Briefly, the neurosphere cells were cultured on Matrigel (BD Biosciences, Bedford, MA, USA)-coated surfaces in medium containing bFGF (no EGF) for 2 days and then grown in medium containing 1% fetal bovine serum (FBS) without EGF/FGF for 3–5 days.
Neurosphere transfection
Transient ACSVL3 knockdown was achieved using previously described ACSVL3 siRNA3 and ACSVL3 siRNA4 [
20]. Targeted sequences of siRNA 3 and siRNA4 corresponded to the human ACSVL3 coding region (total 2430 bp) at bp1243-1263 and 1855–1875, respectively. Transfections of ACSVL3 siRNAs were performed with Oligofectamine (Life technologies) according to the manufacturer’s instructions. Fifteen nmol/L of siRNA was incubated with GBM neurosphere cells for 72 hours.
Neurosphere cells were plated in six well plates. Cells were cultured in serum-free neurosphere medium for 5 days before being dissociated to single cell suspension and counted. For neurosphere formation assay, cells were grown for 5 days in medium containing EGF and FGF. Agarose (4%, Invitrogen) was then added to cultures to a final concentration of 1%. Immobilized neurospheres were stained with 1% Wright solution. For soft agar clonogenic assays, 1% agarose in DMEM was cast on the bottom of plastic six-well plates. Dissociated neurosphere cells (5 × 103cells/well in 6 well plates) were suspended in neurosphere culture medium containing 0.5% agarose and placed on top of the bottom layer. Cells were incubated in neurosphere culture medium for 7–14 days and colonies were fixed and stained with 1% Wright solution. The number of spheres or colonies (>100 μm in diameter) was measured in three random microscopic fields per well by computer-assisted morphometry (MCID, Linton, Cambridge, England). For serial dilution of sphere-formation assay, cells were incubated with control or ACSVL3 siRNA3 for 48 h and plated at the density of 25, 50 and 100 cells/well in of 48 well/plates. Wells containing neurospheres diameter were counted after 3 days.
Quantitative real time-PCR (qRT-PCR)
Total cellular RNA from GBM neurosphere cells was extracted using the RNeasy Mini kit (Qiagen, Germantown, MD, USA). The primer pairs used for amplifying genes of interest were: (1) ACSVL3: Forward primer 5′-cccagagtttctgtggctct-3′ and reverse primer 5′-ggacacttcagccagcaaat-3′ amplify a 256-bp intron-spanning ACSVL3 fragment; (2) nestin: forward primer 5′-aggatgtggaggtagtgaga-3′ and reverse primer 5′- ggagatctcagtggctctt-3′; (3) Musashi-1: forward primer 5′- gagactgacgcgccccagcc-3′ and reverse primer 5′-cgcctggtccatgaaagtgacg-3′; and (4) Sox-2: forward primer 5′- accggcggcaaccagaagaacag -3′ and reverse primer 5′- gcgccgcggccggtatttat -3′. Reverse transcription utilized MuLV Reverse Transcriptase and Oligo (dT) primers. Quantitative real-time PCR (qRT-PCR) was performed as we described in Ying et al. [
21]. Relative expression of each gene was normalized to 18S RNA.
Flow cytometry
The percentages of neurosphere cells expressing CD133 and ALDH were determined by analytical flow cytometry [
21,
26]. For the cell surface marker CD133, single-cell suspensions in 100 μl assay buffer (phosphate buffered saline pH 7.2, 0.5% bovine serum albumin, 2 mM EDTA) were incubated with 10 μl of phycoerythrin (PE)-conjugated anti-CD133 antibody (clone 293C3, Miltenyi Biotec, Auburn, CA) for 10 min in the dark at 4°C. Alternatively, single-cell suspensions were incubated ± diethylaminobenzaldehyde (DEAB) and then incubated in ALDH substrate (Stem Cell Technologies, Vancouver, Canada). The stained cells were analyzed on a FACScan (BD Biosciences). For sorting CD133+ from CD133− cells, neurosphere cells were incubated with microbead-conjugated CD133 antibodies and isolated with magnetic columns (Miltenyi Biotec).
Immunoblotting and immunofluorescence staining
Immunoblotting analyses were performed as previously described [
27]. The primary antibodies used were: anti-ACSVL3 (1:1000) [
20]; anti-β-actin (1:6000); anti-GFAP (1:500, DAKO, Carpinteria, CA, USA) and anti-Tuj1 (1:1000, EMD).
For immunofluorescence staining, neurosphere cells were collected by cytospin onto glass slides, fixed with 4% paraformaldehyde for 30 min at 4°C, permeabilized with PBS containing 0.5% Triton X-100 for 5 min and stained with anti-GFAP and anti-Tuj1 antibodies according to the manufacturers’ protocols. Secondary antibodies were conjugated with Alexa 488 or Cy3 (Life Technologies). Coverslips were placed with Vectashield antifade solution containing 4′6-diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA, USA). Immunofluorescent images were analyzed using Axiovision software (Carl Zeiss, Microscope, Thornwood, NY, USA).
Intracranial xenograft mouse models
All animal protocols were approved by the Johns Hopkins Animal Care and Use Committee. Orthotopic tumor xenograft formation was assessed in 4- to 6-wk-old female mice as previously described [
21]. HSR-GBM1A or HSR-GBM1B cells were transient transfected with ACSVL3 siRNAs for 3 days. Cell viability was determined by trypan blue dye exclusion. Equal numbers of viable cells (1×10
4 cells/animal) in 5 μL PBS were injected unilaterally into the caudate/putamen of C.B-17 SCID/beige mice (n = 10) under stereotactic control [
21]. The animals were sacrificed on post implantation week 10. Brains were removed, sectioned, and stained with H & E. Maximal tumor cross-sectional areas were measured by computer-assisted image analysis as previously described [
28]. Tumor volumes were estimated according to the following formula: tumor volume = (square root of maximum cross-sectional area)
3.
Statistical analysis
Data were analyzed using Prism software (GraphPad, San Diego, CA, USA). When appropriate, two group comparisons were analyzed with a t test unless otherwise indicated. Multiple group comparisons were analyzed by one-way ANOVA with Bonferroni’s multiple comparison. All data are represented as mean value ± standard error of mean (SEM); n = 3 unless indicated otherwise. Significance was set at P < 0.05.
Discussion
A thorough understanding of cancer cell metabolism is critical to the identification of new targets for therapeutic intervention. Lipid metabolism in cancer is one area that has in general been under-studied. The identification of OA-519, a marker of poor prognosis in breast cancer, as fatty acid synthase two decades ago [
31] sparked new interest in this area of cancer metabolism. Several new synthetic fatty acid synthase inhibitors have shown promise in preclinical studies [
32,
33]. However, to the best of our knowledge there are no current ongoing clinical trials testing drugs that target tumor lipid metabolism.
A significant issue in cancer therapeutics is that of recurrence and subsequent refractoriness to therapy. Tumor cells with stem-like features have been hypothesized to be, at least in part, responsible for these phenomena [
16,
17]. Thus, drugs that target stem-like cells would be an invaluable weapon in the treatment arsenal. Our previous work suggested that the acyl-CoA synthetase ACSVL3 was overproduced in human GBM and GBM cells in culture, and that decreasing the expression of this enzyme in GBM cells reduced both their malignant behavior in culture and their tumorigenicity in nude mice [
14]. In this report, we show that expression of ACSVL3 is even more robust in cancer stem cell enriched neurospheres than in the cell population from which they were derived. Reducing ACSVL3 expression in these cells also decreased tumorigenicity in mice. Furthermore, differentiation of cancer stem cells with all-trans retinoic acid or Trichostatin A reduced ACSVL3 expression. Taken together, these observations indicate that ACSVL3 expression is associated with a highly undifferentiated phenotype and that therapeutic targeting this enzyme may be a promising anti-cancer therapy.
ACSVL3 is one of 26 acyl-CoA synthetases encoded by the human genome [
34]. Acyl-CoA synthetases activate fatty acids to their coenzyme A thioesters, allowing subsequent entry into diverse metabolic pathways. RNA interference studies suggest that ACSVL3 is responsible for up to 30% of long-chain and very-long chain acyl-CoA synthetase activity in cells that endogenously express the enzyme [
9]. Although this enzyme is also known as “fatty acid transport protein 3”, a role in fatty acid uptake could not be demonstrated experimentally [
9]. Results presented here, and our previous work [
14], show a correlation between ACLVL3 levels and cell growth rate, suggesting that this enzyme may provide fatty acid substrates required for bulk membrane phospholipid biosynthesis. Our current studies do not support this hypothesis (Shi and Watkins, unpublished); rather, a role in lipid signaling, possibly via phosphoinositide species and PI3 kinase signaling [
14], seems more likely. The induction of ACSVL3 by RTK oncogenic pathways supports this notion, and indicates the importance of fatty acid metabolism in cancer stem cell maintenance. Activated fatty acid can regulate oncogenic signaling transduction pathways that are necessary for cell survival, proliferation, and differentiation [
35], either directly or indirectly, by functioning as agonists of a number of G protein-coupled receptors, activating RTK downstream targets such as phosphatidylinositol 3-kinase/Akt and p44/42 mitogen-activated protein kinases, and stimulating phospholipase C/protein kinase. Elucidation of the specific downstream lipid metabolism pathways that are “fed” by ACSVL3 will provide new clues as to how this enzyme supports the malignant phenotype, and this is currently an area of active investigation in our laboratory.
Lipid metabolism has been linked to cellular differentiation mechanisms in some in vitro and in vivo models. ACSVL4 (or fatty acid transporter protein 4) has been shown to regulate keratinocyte differentiation [
36]. Fatty acids and their metabolites can modulate stem cell self-renewal, survival, proliferation and differentiation by regulating gene expression, enzyme activity, and G protein-coupled receptor signal transduction [
35]. Recent studies revealed that arachidonic acid (AA), eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA) may regulate the proliferation and differentiation of various types of stem cells. For example, both AA and EPA were the most potent inhibitors of proliferation of promyelocytic leukemic cells [
37,
38]. DHA or AA was found to promote the differentiation of neural stem cells into neurons by promoting cell cycle exit and suppressing cell death [
39,
40]. The role of fatty acid metabolism pathways in cancer stem cell differentiation has not been explored. To our knowledge, this is the first report showing that ACSVL3 regulates cancer stem cell phenotype and that ACSVL3 loss-of-function promotes cancer stem cell differentiation and inhibits tumor-initiation properties of cancer stem cells.
Our findings suggest that ACSVL3 is a potential therapeutic target worthy of further investigation. Findings reported here suggest that if identified, a small molecule inhibitor of ACSVL3 could inhibit the growth of GBM stem cells as well as non-stem tumor cells. Although there have been a few inhibitors of acyl-CoA synthetases reported [
41‐
44], most are non-specific, and none that target ACSVL3 have been described. Research efforts to discover specific ACSVL3 inhibiters are also underway.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
PS, SX: Conception and design, Collection and assembly of data, Data analysis and interpretation, Manuscript writing, Final approval; BL, XS, KY: Collection and assembly of data, Data analysis and interpretation, Final approval; PW, JL: Conception and design, Financial support, Administrative support, Data analysis and interpretation, Manuscript writing, Final approval.