Gene Profile of Myeloid-Derived Suppressive Cells from the Bone Marrow of Lysosomal Acid Lipase Knock-Out Mice

Background Lysosomal acid lipase (LAL) controls development and homeostasis of myeloid lineage cells. Loss of the lysosomal acid lipase (LAL) function leads to expansion of myeloid-derived suppressive cells (MDSCs) that cause myeloproliferative neoplasm. Methodology/Principal Findings Affymetrix GeneChip microarray analysis identified detailed intrinsic defects in Ly6G+ myeloid lineage cells of LAL knock-out (lal−/−) mice. Ingenuity Pathway Analysis revealed activation of the mammalian target of rapamycin (mTOR) signaling, which functions as a nutrient/energy/redox sensor, and controls cell growth, cell cycle entry, cell survival, and cell motility. Loss of the LAL function led to major alteration of large GTPase and small GTPase signal transduction pathways. lal−/− Ly6G+ myeloid cells in the bone marrow showed substantial increase of cell proliferation in association with up-regulation of cyclin and cyclin-dependent kinase (cdk) genes. The epigenetic microenvironment was significantly changed due to the increased expression of multiple histone cluster genes, centromere protein genes and chromosome modification genes. Gene expression of bioenergetic pathways, including glycolysis, aerobic glycolysis, mitochondrial oxidative phosphorylation, and respiratory chain proteins, was also increased, while the mitochondrial function was impaired in lal−/− Ly6G+ myeloid cells. The concentration of reactive oxygen species (ROS) was significantly increased accompanied by up-regulation of nitric oxide/ROS production genes in these cells. Conclusions/Significance This comprehensive gene profile study for the first time identifies and defines important gene pathways involved in the myeloid lineage cells towards MDSCs using lal−/− mouse model.


Introduction
Myeloid-derived suppressive cells (MDSCs) are heterogeneous populations that express CD11b and Gr-1 antigens. MDSCs actively participate in inflammation-induced pathogenic processes in various diseases (i.e. cancer) by suppressing T lymphocytes [1,2,3]. We previously reported that the neutral lipid metabolic pathway controlled by lysosomal acid lipase (LAL) plays a critical role in development and homeostasis of MDSCs [4,5]. LAL hydrolyzes cholesteryl esters and triglycerides in the lysosome of cells to generate free cholesterol and free fatty acids. Ablating LAL (lal2/2) in mice led to aberrant expansion of MDSCs (.40% in the blood, and .70% in the bone marrow) that arise from dysregulated production of myeloid progenitor cells in the bone marrow [4]. Ly6G + MDSCs in lal2/2 mice show strong immunosuppression on T cells, which contributes to impaired T cell proliferation and function in vivo [4,5]. As a consequence of myeloproliferative neoplasm, severe pathogenic phenotypes in multiple organs are observed in lal2/2 mice, including the adult lung, liver, spleen, thymus, adrenal glands, and small intestine, which are all associated with MDSCs infiltration [4,6,7,8,9,10]. Over-expression of LAL downstream genes in myeloid lineage cells driven by the 7.2 kb c-fms promoter/intron 2 induces chronic inflammation, immunosuppression and tumorigenesis in vivo [11,12,13].
Given the important role in inflammation and tissue pathogenesis, it is essential to elucidate the intrinsic molecular mechanisms governing MDSCs development and homeostasis in lal2/2 mice. At the moment, up-regulated genes related to amino acid metabolism (i.e. L-arginine) and production of reactive oxygen (ROS)/nitrogen species are well studied and serve as parameters to define MDSCs [14,15]. In this report, we aim at identifying a comprehensive gene profile to define pathways that are involved in MDSCs development in lal2/2 mice by GeneChip microarray analysis. The results showed that the mammalian target of rapamycin (mTOR) signaling, which functions as a nutrient/ energy/redox sensor, and controls cell growth, cell cycle entry, cell survival, and cell motility, is activated in bone marrow MDSCs during LAL deficiency.

Ethics Statement and Animal Care
All scientific protocols involving the use of animals have been approved by the Institutional Animal Care and Use Committee of Indiana University School of Medicine and followed guidelines established by the Panel on Euthanasia of the American Veterinary Medical Association. Protocols involving the use of recombinant DNA or biohazardous materials have been reviewed by the Biosafety Committee of Indiana University School of Medicine and followed guidelines established by the National Institutes of Health. Animals were housed under Institutional Animal Care and Use Committee-approved conditions in a secured animal facility at Indiana University School of Medicine.

S6 and E4-BP analysis
Fluorescence activated cell sorting (FACS) analysis was performed on single cells from the bone marrow of 5-month-old lal+/+ and lal2/2 mice. Bone marrow cells were prepared as previously described [11]. Approximately 1 to 2610 6 cells from various organs in FACS buffer were blocked with FcR blocking antibodies (BD Pharmingen, San Diego, CA) followed by incubation with APC anti-mouse CD11b and PE rat anti-mouse Ly6G (1A8, BD Bioscience). Cells were fixed and permeabilized using BD Cytofix/Cytoperm TM Fixation/permeabilization Kit according to the manufacture's instruction, followed by labeling with anti-pS6 (ser235/236) and anti-p4E-BP (Thr37/46) antibodies (1:50 dilution, Cell Signaling Technology, Beverly, MA) at 4uC overnight. Cells were analyzed on a LSRII machine (BD Biosciences, San Jose, CA). Data were analyzed using the BD FACStation TM Software (BD Biosciences). Quadrants were assigned using isotype control mAb.

MDSCs RNA isolation
Single cells from bone marrow of 5-month-old lal+/+ and lal2/2 mice (n = 5) were stained with anti-Ly6G + antibody, followed by positive magnetic selection using anti-biotin microbeads following the manufacture's instructions (Miltenyi Biotec, Auburn, CA). The purity of the Ly6G + MDSC population was typically higher than 90%. Total RNAs from isolated Ly6G + MDSCs were purified using the Qiagen total RNA purification kit (Qiagen, Valencia, CA). RNA concentrations were measured with Spectra Max 190 (Molecular Devices, Sunnyvale, CA).

Affymetrix GeneChip microarray analyses
The quality of the total RNA was checked by the Agilent Bioanalyzer 2100 (Hewelett Packard) using the RNA 6000 Pico Assay. To create biotin labeled cDNA products, 20 nanograms of total RNA was used. Double stranded target cDNA was synthesized using a random hexamer with a T7 promoter. Target sense transcript cRNA was generated from the double-stranded cDNA template using the Whole Transcript cDNA Synthesis and Amplification Kit. cDNA was regenerated using a reverse transcription reaction randomly primed with a mix containing dUTP. After hydrolysis of the cRNA with RNase H, the sense strand of cDNA was purified using the Affymetrix sample cleanup module, fragmented by incubation with UDG (uracil DNA glycosylase) and APE 1 (apurinic/apyrimidicendonuclease 1), and terminally biotin-labeled with TdT (terminal deoxynucleotidyl transferase) using the WT Terminal Labeling Kit. Biotinylated sense strand fragments were hybridized to Affymetrix Mouse Gene 1.0 ST GeneChips using the Hybridization Control and Hybridization Wash and Stain kits at 45uC for 18 hrs. The stained array was scanned using an Affymetrix GeneChip Scanner 3000 7G to generate the CEL files. Primary quality control was performed using the Affymetrix Expression Console. The chip data were imported with Partek Genomics Suite 6.5 (Partek, Inc., St Louis, MO), normalized and summarized using the RMA (Robust Multiarray Average) algorithm. The relative log expression was examined to ensure that the data were properly corrected by normalization and that there were no outliers. Mv A plots were generated to examine the reproducibility of the replicates. To identify expression changes between genotypes, a two-way ANOVA was performed by using the methods of moments [16] to partition the effect of tissue and genotype. Genes differentially expressed in lal2/2 mice vs lal+/+ mice were identified at a false positive rate (FDR) of 0.05 and fold change $2. The unprocessed microarray data is available at Gene Expression Ominibus Database (GEO) at NCBI (accession number GSE 29401).

Pathway and functional classification
Significantly affected or differentially expressed genes were subjected to an intensive search to identify biological functions. Functional pathway gene ontology and network analysis were executed using Ingenuity Pathway Analysis (IPA) (Ingenuity Systems, Mountain View, CA), Partek (Partek Inc., St. Louis, MO), public information and literature references. The enriched functional categories were determined by Fisher Exact Test using the corresponding murine genome as a reference dataset. The significance was set at p-value,0.05. The differentially expressed genes were grouped into various categories. To cluster gene expression profiles, Hierarchical cluster analysis of the significantly expressed genes was performed using Partek genomic suite 6.6 (St. Louis, MO) which showed the correlated groups of genes and their expression patterns across all points.

FACS for JC-1 analysis
Mitochondrial membrane potential was measured according to MitoProbe JC-1 Assay Kit (Molecular Probe, Eugene, OR). Bone marrow single cells from lal+/+ and lal2/2 mice were suspended in pre-warmed PBS. Cells were then stained with anti-CD11b, anti-Ly6G antibody and 10 ml of 200 mM JC-1 dye. After incubation at 37uC, 5% CO 2 for 30 minutes, cells were suspended by gently flicking the tubes with the warmed PBS and washed twice. JC-1 was analyzed by flow cytometry in gated CD11b + Ly6G + cells.

ROS detection
Ly6G + cells were isolated from the bone marrow of lal+/+ and lal2/2 mice with ly6G + antibody-coated magnetic beads and MACS-LS columns according to the manufacturer's instructions (Miltenyi Biotec, Auburn, CA). Purified cells were suspended in pre-warmed PBS. Oxidation-sensitive dye 29-79-dichlorodihydrofluorescein diacetate (DCFDA, C6827, Invitrogen, Eugene, OR), was used to measure ROS production by MDSC. Cells were incubated at 37uC in pre-warmed PBS in the presence of 2.5 mM DCFDA for 20 min. Cells were then labeled with APC-conjugated anti-Ly6G + and PE-conjugated anti-CD11b Abs on ice. Analysis was then conducted by flow cytometry.

Cell Cycle
Bone marrow cells from lal+/+ and lal2/2 mice were stained with surface markers anti-CD11b and ly6G antibodies (1:200 dilution, BD Bioscience) for 15 min at 4uC. After washed with PBS, cells were fixed in 1 ml 70% methanol for 1 hour at room temperature. After cells were washed with PBS, 100 ml of 10 mg/ ml RNAase (Rase LS005649, Worthington Biochemicals, Lakewood, NJ) and 1 ml of 50 mg/ml propidium iodide (PI, P4170, Sigma, Saint Louis, MI) were added and incubated for 30 min at 37uC. Cells were washed with PBS. PI-labeled DNA concentrations were analyzed in gated CD11b + Ly6G + cells by FACS Calibur APC machine. The data were analyzed with ModFit LT Tm DNA analysis software (VMFLTMAC3, Verity Software House, Topsham, ME).

ATP assay
Bone marrow cells (1610 6 ) were rinsed with PBS and lysed with ATP-releasing buffer containing 100 mM potassium phosphate buffer (pH 7.8), 2 mM EDTA, 1 mM dithiothreitol, and 1% Triton X-100. The ATP concentration of cell lysate was measured using ATP kit (Invitrogen, Carlsbad, CA) according to the manufacture's instruction.

Results
Affymetrix GeneChip microarray and Ingenuity pathway analyses of Ly6G + cells from the lal2/2 bone marrow In order to identify intrinsic defects of MDSCs in lal2/2 mice, total RNAs were isolated from Ly6G + myeloid cells of the agematched wild type and lal2/2 bone marrow. Purified total RNAs were subjective to the Affymetrix GeneChip microarray study. At a false positive rate (FDR) of 0.05 and fold change $2, there were 3086 changed genes in lal2/2 bone marrow MDSCs. Major changes of lal2/2bone marrow MDSCs were observed in Gprotein signaling, cell cycle, chromatin modification and bioenergetics ( Figure 1A). Gene expression of both Ly6G (9.1 fold) and Ly6C (7.1 fold) was increased in lal2/2 bone marrow MDSCs. Ingenuity Pathway Analysis revealed alteration of several important pathways in lal2/2 bone marrow MDSCs ( Figure 1B). Compared with normal wild-type cells, expression of genes that are positively or negatively regulated by LAL downstream effector PPARc was altered. Due to lack of cholesterol generation through the LAL pathway, genes involved in cholesterol biosynthesis were increased in lal2/2 bone marrow MDSCs to compensate the loss. Importantly, genes involved in the mammalian target of rapamycin (mTOR) singaling pathway were altered in lal2/2 bone marrow MDSCs. The mTOR is a serine/threonine protein kinase that regulates cell growth, cell proliferation, cell motility, cell survival, protein synthesis, and transcription in response to growth factors, and mitogens. mTOR also senses cellular nutrient and energy levels and redox status [17]. Compared with wild type mice, mTOR downstream effectors S6 (S6) and 4E-BP1 were highly phosphorylated in lal2/2 bone marrow MDSCs ( Figure 1C), a strong indication of mTOR activation.

ROS production in lal2/2 bone marrow MDSCs
Since mTOR controls the redox status, and nitric oxide and reactive oxygen species production are known to play important roles in MDSCs, genes involved in nitric oxide and ROS production were investigated in lal2/2 bone marrow MDSCs. As demonstrated in Figure 2A, expression of these genes was significantly altered. As a result, the concentration of ROS was increased 3 times in the Ly6G population of lal2/2 bone marrow MDSCs compared with those in wild type MDSCs ( Figure 2B). Reduced glutathione (GSH) protects the cell by destroying hydrogen peroxide and hydroxyl free radicals. Both glutathione peroxidase (3.0 fold) and glutathione reductase genes (20.1 fold) showed increased expression in lal2/2 bone marrow MDSCs as well (Table S2). The gene expression level of glucose 6-phosphate dehydrogenase that is required for regeneration of GSH from its oxidized form (GSSG) was increased in lal2/2 bone marrow MDSCs as well (Table S2).

Gene expression levels of G-protein family in lal2/2 MDSCs
Guanosine nucleotide-binding proteins (G proteins) are involved in many aspects of cellular functions. They play key roles in signal transduction and regulation of gene expression in almost all cell types including immune cells [18]. This group of proteins can be separated into two classes. The first class is the heterotrimeric G protein superfamily, consisting of a-, b-, and c-subunit, with the asubunit binding the guanosine nucleotides. Many members in this large G-protein superfamily, including receptors, effectors, heterotrimeric Gabc subunits, and the regulators of G-protein signaling proteins were altered in gene expression in lal2/2 bone marrow MDSCs (Table 1, 2). The second class is the monomeric low-molecular-weight Ras GTPase superfamily, which contains more than 100 proteins, and cycles between an inactive guanosine diphosphate (GDP) and an active GTP-bound form [19]. As demonstrated in Table 3, 4, 5, gene expression of many members of the small GTPase superfamily and downstream effectors were altered in lal2/2 bone marrow MDSCs. It is known that G-protein family has a tight relationship with the mTOR activity [20].

Gene expression levels of cell cycle protein family in lal2/2 MDSCs
The cell cycle consists of four distinct phases: G 1 phase, S phase (synthesis), G 2 phase and M phase (mitosis). The G 1 checkpoint control mechanism ensures that everything is ready for DNA synthesis. DNA replication occurs during the S phase. During the interphase (cell grows), nutrients are accumulated and DNA is duplicated in cells. The G 2 checkpoint control mechanism ensures that everything is ready to enter the M phase and divide. During the mitosis (M) phase, the cell splits itself into two distinct daughter cells. The cell cycle is precisely controlled by cyclin-dependent kinases (CDKs). CDK activity requires binding of regulatory subunit cyclins [21]. When activated by a bound cyclin, CDKs perform phosphorylation that activates or inactivates target proteins to orchestrate coordinated entry into the next phase of the cell cycle. Gene expression of many Cdk and cyclin family members of the cell cycle were altered in lal2/2 bone marrow MDSCs (Table 6). In lal2/2 bone marrow MDSCs, gene expression of all four critical cyclins (A, B, D, E-type) as well as Cdk1, Cdk2, Cdk5, Cdk9 were up-regulated compared with wild type MDSCs (Table 6). When the cell cycle was studied, 4 times more lal2/2 MDSCs (20.96%) were accumulated in the G2/M phases than wild type MDSCs (5.35%) ( Figure 3A, B), suggesting that lal2/2 bone marrow  MDSCs are highly proliferative. This is consistent with our previous BrdU proliferation study [4]. The G 1 cyclin-CDK complexes also promote the degradation of molecules (i.e. cyclins after using) that function as S phase inhibitors by targeting them for ubiquitination [21]. Once a protein has been ubiquitinated, it is targeted for proteolytic degradation by the proteasome. In lal2/2 bone marrow MDSCs, gene expression of enzymes/protein factors involved in ubiquitination (Table 7) and proteasome subunits (Table 8) showed up-regulation, indicating an enhanced protein recycling activity.
Gene expression levels of histone cluster protein family and centromere protein family in lal2/2 MDSCs Cell proliferation and cell cycle are also controlled by the chromatin activity. The nucleosomes form the basic repeating units of chromatin in eukaryotes. The composition of the individual nucleosomes is fundamentally similar and consists of an octameric core of four types of histones (H2A, H2B, H3 and H4). The expression of the major histones is tightly regulated during the cell cycle, and the histones are deposited onto DNA in a process that is strictly coupled to DNA replication. However, histone variants are not subject to this stringent regulation, and are expressed throughout the cell cycle. Histone-variant exchange activities contribute to gene expression and other cellular events (i.e. formation of centromeric and telomeric chromatin during cell cycles) [22]. In lal2/2 bone marrow MDSCs, gene expression of multiple histone variants were increased ( Table 9). The function of histone deposition onto DNA is also influenced by interacting with other associated proteins and by posttranslational modification (such as phosphorylation, methylation, and acetylation, polyADP ribosylation and monoubiquitylation) [22]. Alteration of gene expression of these factors was observed in lal2/2 bone marrow MDSCs (Table 7 and Table S1).
During the M phase of cell cycle, the centromeric histones (i.e. CENPA) are required for chromosome segregation. They form centromere that is the site of spindle attachment to the chromosomes during mitosis. Gene expression of multiple centromeric and telomeric histones along with other M phase chromosome structural proteins (i.e. kinetochore protein subunits) was increased in lal2/2 bone marrow MDSCs (Table 10).

Gene expression levels of metabolic proteins involved in glycolysis and citric acid cycles in lal2/2 MDSCs
Metabolism provides the cell with the energy and resources to support cell growth. Glucose serves as a fuel for ATP generation. Glycolysis occurs in the cytosol. The link between glycolysis and citric acid cycle is the oxidative decarboxylation of pyruvate to form acetyl CoA. In the matrix of mitochondrial, the citric acid cycle is the final common pathway for the oxidation of fuel  molecules. It also serves as a source of building blocks for biosyntheses. Many enzymes that are involved in the glycolytic pathway (hexokinase, glucose phosphate isomerase, aldolase, triosephosphate isomerase, glyceraldehyde-3-phosphate dehydrogenase, phosphoglycerate kinase, enolase, pyruvate kinase) and the citric acid cycle (isocitrate dehydrogenase 1, succinate dehydrogenase complex, subunit B, malate dehydrogenase 2) were upregulated in lal2/2 bone marrow MDSCs (Table 11). In addition, gene expression of both lactate dehydrogenase A and B was up-regulated in lal2/2 bone marrow MDSCs (Table 11), a pathway generally used by cancer cells (Warburg effect). Some enzymes in the pentose phosphate pathway (energy conservation for biosynthetic purposes) and glycogen synthesis (storage form of glucose and metabolic energy) were also up-regulated in lal2/2 bone marrow MDSCs (Table S2).

Gene expression levels of mitochondrial respiratory assembly and energy generating proteins in lal2/2 MDSCs
Mitochondria are the powerhouse for cells. In order to support robust cell proliferation, ATP synthesis is required. In mitochondrial oxidative phosphorylation, the synthesis of ATP is coupled to the flow of electrons from NADH or FADH2 to O 2 by a proton gradient across the inner mitochondrial membrane and a series of electron carriers, such as NADH dehydrogenases and the cytochromes [23]. The synthesis of ATP is carried out by mitochondrial ATPase in the inner mitochondrial membrane. In lal2/2 bone marrow MDSCs, the ATP concentration was significantly increased compared with wild type mice ( Figure 4A). As demonstrated in Table 12, 13, expression of multiple NADH dehydrogenase subunits, cytochrome subunits, and mitochondrial ATPase were all up-regulated in lal2/2 bone marrow MDSCs. Of the 85 mitochondrial respiratory chain subunits, 13 are synthesized within the organelle by the mitochondrial ribosome [24]. Gene expression of multiple mitochondrial ribosomal protein subunits were also up-regulated in lal2/2 bone marrow MDSCs (Table S3), indicating the increased activity of protein synthesis within the mitochondria to meet the demand of ATP production. Despite increased mitochondrial activities and ATP synthesis, the membrane potential was reduced ( Figure 4B), an indication of the functional impairment of mitochondria in lal2/2 bone marrow MDSCs.

Discussion
LAL plays an important role in controlling inflammation. Genetic ablation of LAL function leads to systemic expansion of MDSCs. The aggressive behavior of MDSCs in lal2/2 mice relies on their intrinsic malfunction. The gene profile analysis reveals the underneath mechanisms by which LAL regulates MDSCs development and homeostasis. It is surprising that MDSCs in lal2/2 animal model share many characteristics used by cancer cells.

Activation of G-protein superfamily
LAL deficiency leads to vast changes of the G-protein signaling pathways (both large and small GTPases) in lal2/2 bone marrow MDSCs (Table 1, 2,3,4,5). Especially, multiple components of small GTPases are up-regulated. Since G-protein signaling pathways are well known for their diverse functions in cells (such as glucose metabolism, ion channels, transcription, motility, and secretion) [25,26], ups and downs of these genes exert their profound pathological influences on MDSCs development and homeostasis in lal2/2 mice. Among changed small GTPases, 1) Ras GTPases activate the Raf/Mek/Erk pathway, which mediates cell growth and cell-cycle entry by phosphorylation of transcription factors, MNK (MAPK-interacting serine/threonine kinase) family of kinases, and the PI3K/AKT pathway to control cell survival,  (Table 3, 4,5). For example, Erk (Mapk3) was up to 7.3 folds, and p38 (Mapk13) was up to 5.1 folds. We reported previously that the phosphorylation levels of both Erk and p38 are substantially increased in lal2/2 bone marrow MDSCs [4]; 2) Rho GTPases organize actin cytoskeleton, cell adhesion and cell motility. They promote cell-cycle progression through G1 by regulating cyclin D1 and cyclin-dependent inhibitors p21 and p27 [19]; 3) Rab GTPases regulate receptor internalization, vesicle formation and trafficking to various cellular sites, including the nucleus, lysosome and plasma membrane. Through regulation of endocytic trafficking, they integrate multiple signaling pathways that are involved in cellular proliferation, apoptosis and migration. Gene expression of multiple Rab GTPase-regulated membrane vesicle traffic effectors that mediate vesicle intracellular transport, including actin-dependent motors (myosins) or microtubule-dependent motors (kinesins or dyneins) (Table S4), and tethering factors (exocyst, trafficking protein particle complex or TRAPPC, vacuole protein sorting or VPS) (Table S5) [27] were significantly altered in lal2/2 bone marrow MDSCs. Farnesylation on G-proteins is a critical step for membrane binding and biological function. Farnesyltransferase expression that farnesylates Ras on the CAAX motif was increase 2 folds in lal2/2 bone marrow MDSCs (Table 3, 4, 5).

Cell cycle and Histone cluster genes
The second group of disregulation in lal2/2 bone marrow MDSCs belongs to cell cycle genes (Table 6). These cell cycle  related genes control proper progression of the cell cycle by checkpoints that sense possible defects during DNA synthesis and chromosome segregation [21]. It has been demonstrated that CDK1 (6.6 fold increase) alone can drive the cell cycle in most mammalian cells [21]. Up-regulation of Cdk2, Cdk5, Cdk9 suggests that lal2/2 bone marrow MDSCs acquire specific needs for the activities of these Cdks during development and homeostasis. Importantly, all Cdk regulatory cyclins (A, B, D, E-type) that are required for phase entering of the cell cycle were up-regulated, suggesting constitutive mitogenic signaling and defective responses to anti-mitogenic signals under the pressure of Ras GTPases overactivation that contribute to unscheduled proliferation during LAL deficiency. Up-regulation of enzymes/protein factors that are involved in ubiquitination and proteasome (Table 7 & 8) further suggests the higher cell cycle activity in lal2/2 bone marrow MDSCs.
Ubiquitin and ubiquitin-like molecules direct proteins to proteolysis within proteasome for recycling, which is required in a wide variety of cellular processes, including cell cycle and division, DNA transcription and repair, biogenesis of organelles, modulation of cell surface receptors, ion channels and the secretory pathway, response to stress and extracellular modulators, ribosome biogenesis [21,28].
Another evidence of enhanced cell proliferation in lal2/2 bone marrow MDSCs is up-regulation of multiple members of histone cluster genes ( Table 9). Changes of these histone variants and associate factors set up an epigenetic microenvironment that preferentially favor certain gene expression and cellular events to Table 9. Up-regulation of histone cluster genes in lal2/2 bone marrow MDSCs.   (Table 10). This was further supported by the cell cycle analysis, in which lal2/2 bone marrow MDSCs in G1/M phases were around 4 times more compared with wild type MDSCs (Figure 3).

Metabolism and bioenergetics
Another similarity between lal2/2 bone marrow MDSCs and cancer cells is reprogramming of energy metabolism. First, cells rely on mitochondrial oxidative phosphorylation to provide energy (ATP production) for cellular activities in normal conditions. Cancer cells are characterized by increased glycolysis and reduced mitochondrial respiratory function [24]. The electron flow through the respiratory chain is substantially lower in malignant cells leading to oxidative stress and increased ROS production. Similar to cancer cells, multiple enzymes and proteins in glycolysis and citric acid cycles were up-regulated in lal2/2 bone marrow MDSCs (Table 11). Interestingly, the mitochondrial function in lal2/2 bone marrow MDSCs was impaired ( Figure 4B) despite increased expression of respiratory chain proteins (including NADH dehydrogenases, cytochrome proteins, ATPases and mitochondrial ribosomal proteins) ( Table 12, 13) and increased ATP production (Figure 4 A) compared with those in wild type counterparts. Second, pyruvate (the end product of glycolysis) is imported into mitochondria and enters the citric acid cycle in normal cells. Cancer cells use an altered metabolic pattern by taking up much more glucose and mainly process it through aerobic glycolysis, producing large quantities of secreted lactate by lactate dehydrogenase (LDH) with a lower use of oxidative phosphorylation  Mitochondrial membrane potential of CD11b + Ly6G + MDSCs from the bone marrow of lal+/+ mice and lal2/2 mice was analyzed by JC-1 staining. Positive staining of JC-1 red represents healthy mitochondria. Transition from negative staining of JC-1 red to positive staining of JC-1 green represents impaired mitochondrial potential. doi:10.1371/journal.pone.0030701.g004 [23,29]. This metabolic switch by aerobic glycolysis is advantageous to cancers cells to allow them better surviving, producing intermediates for cell growth and division. Similar to cancer cells, lal2/2 bone marrow MDSCs showed increased gene expression of both lactate dehydrogenase A and B (Table 11), which keep pyruvate away from the mitochondria. This observation indicates that LAL-controlled neutral lipid metabolism plays a critical role in preventing metabolic switch from oxidative phosphorylation to aerobic glycolysis in myeloid lineage cells.
In addition to generate energy for the cell (ATP) and produce substrates to synthesize amino acids, nucleosides, and fatty acids, normal cells use both glucose and glutamine as substrates to regulate the redox potential to minimize the effects of reactive oxygen species (ROS) that damage membranes, proteins and cause mutations in a cell. Similar to cancer cells, the concentration of ROS was significantly increased in lal2/2 bone marrow MDSCs ( Figure 2B), which was accompanied by up-regulation of nitric oxide/ROS production genes (Figure 2A), glutathione peroxidase/glutathione reductase genes, and glucose 6-phosphate dehydrogenase gene (Table  S2). High levels of ROS allow for the stimulation of cell proliferation, induction of genetic instability, and evasion from senescence [23].

The mTOR pathway
Ingenuity Pathway Analysis revealed alteration in PI3K/ thymoma viral proto-oncogene (AKT)/mammalian target of rapamycin (mTOR) signaling pathway [30,31,32] in lal2/2 bone marrow MDSCs ( Figure 1B). Activation of the mTOR pathway has been confirmed by highly phosphorylated S6 and 4E-BP, two authentic mTOR downstream effectors ( Figure 1C). mTOR serves as a signal integrator for nutrients, growth factors, energy and stress [20]. Activation of this pathway suppresses apoptosis, promotes an influx of glucose and amino acids into the cells, stimulates ATP production [17], as well as contributes to cell growth, cell cycle entry, cell survival, and cell motility during tumorigenesis [33]. Increasing evidence suggests that membrane trafficking causes mTORC1 to shuttle to lysosomes and regulate mTORC1signalling, enabling it to respond to growth factors [20,34]. The lysosomal surface hosts a molecular machinery for mTORC1 activation that includes the Rag GTPases, the trimeric regulator complex, and possibly GTPase activating proteins (GAPs) and guanine nucleotide exchange factors (GEFs) for the Rag GTPases [20]. Since LAL is a lysosome-associated enzyme, it is conceivable that lack of the LAL activity may change lipid composition and dynamics on the lysosomal membrane that influence endomembrane trafficking and stimulate the mTOR1 activity, which in tune coordinates the  cellular metabolism and growth to increase abnormal proliferation of lal2/2 bone marrow MDSCs. It is important to keep in mind that development and expansion of MDSCs are a complex process. In addition to changes of gene expression, posttranscriptional modification of intracellular signaling pathways also contributes to the lal2/2 MDSCs autonomous defect. For example, although up-regulation of Stats family members was not detected by Affymetrix GeneChip microarray analysis, phosphorylation of Stat3 at Y705 has been detected in expanded lal2/2 MDSCs [4]. Activation of Stat3 directly leads to MDSCs expansion in vivo [35,36]. Phosphorylation of Erk and p38 in the Ras signaling pathway has also been detected in expanded lal2/2 MDSCs.
In summary, studies outlined here demonstrate that the loss of the LAL function leads to myeloproliferative neoplasm. Affymetrix GeneChip microarray analysis provides a detailed map of intrinsic defects existing in lal2/2 bone marrow MDSCs. The interrelationships between these biological processes during lal2/2 MDSCs development and homeostasis remain to be elucidated in the future. The study provides new avenues and interventions clinically to eliminate MDSCs expansion in disease setting.

Supporting Information
Table S1 Changes of histone related genes in MDSCs from the bone marrow of lal2/2 mice. (DOC)